Index: head/sbin/pfctl/parse.y =================================================================== --- head/sbin/pfctl/parse.y (revision 306683) +++ head/sbin/pfctl/parse.y (revision 306684) @@ -1,6269 +1,6270 @@ /* $OpenBSD: parse.y,v 1.554 2008/10/17 12:59:53 henning Exp $ */ /* * Copyright (c) 2001 Markus Friedl. All rights reserved. * Copyright (c) 2001 Daniel Hartmeier. All rights reserved. * Copyright (c) 2001 Theo de Raadt. All rights reserved. * Copyright (c) 2002,2003 Henning Brauer. All rights reserved. * * 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. * * THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 AUTHOR 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 __FBSDID("$FreeBSD$"); #include #include #include #ifdef __FreeBSD__ #include #endif #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "pfctl_parser.h" #include "pfctl.h" static struct pfctl *pf = NULL; static int debug = 0; static int rulestate = 0; static u_int16_t returnicmpdefault = (ICMP_UNREACH << 8) | ICMP_UNREACH_PORT; static u_int16_t returnicmp6default = (ICMP6_DST_UNREACH << 8) | ICMP6_DST_UNREACH_NOPORT; static int blockpolicy = PFRULE_DROP; static int require_order = 1; static int default_statelock; static TAILQ_HEAD(files, file) files = TAILQ_HEAD_INITIALIZER(files); static struct file { TAILQ_ENTRY(file) entry; FILE *stream; char *name; int lineno; int errors; } *file; struct file *pushfile(const char *, int); int popfile(void); int check_file_secrecy(int, const char *); int yyparse(void); int yylex(void); int yyerror(const char *, ...); int kw_cmp(const void *, const void *); int lookup(char *); int lgetc(int); int lungetc(int); int findeol(void); static TAILQ_HEAD(symhead, sym) symhead = TAILQ_HEAD_INITIALIZER(symhead); struct sym { TAILQ_ENTRY(sym) entry; int used; int persist; char *nam; char *val; }; int symset(const char *, const char *, int); char *symget(const char *); int atoul(char *, u_long *); enum { PFCTL_STATE_NONE, PFCTL_STATE_OPTION, PFCTL_STATE_SCRUB, PFCTL_STATE_QUEUE, PFCTL_STATE_NAT, PFCTL_STATE_FILTER }; struct node_proto { u_int8_t proto; struct node_proto *next; struct node_proto *tail; }; struct node_port { u_int16_t port[2]; u_int8_t op; struct node_port *next; struct node_port *tail; }; struct node_uid { uid_t uid[2]; u_int8_t op; struct node_uid *next; struct node_uid *tail; }; struct node_gid { gid_t gid[2]; u_int8_t op; struct node_gid *next; struct node_gid *tail; }; struct node_icmp { u_int8_t code; u_int8_t type; u_int8_t proto; struct node_icmp *next; struct node_icmp *tail; }; enum { PF_STATE_OPT_MAX, PF_STATE_OPT_NOSYNC, PF_STATE_OPT_SRCTRACK, PF_STATE_OPT_MAX_SRC_STATES, PF_STATE_OPT_MAX_SRC_CONN, PF_STATE_OPT_MAX_SRC_CONN_RATE, PF_STATE_OPT_MAX_SRC_NODES, PF_STATE_OPT_OVERLOAD, PF_STATE_OPT_STATELOCK, PF_STATE_OPT_TIMEOUT, PF_STATE_OPT_SLOPPY, }; enum { PF_SRCTRACK_NONE, PF_SRCTRACK, PF_SRCTRACK_GLOBAL, PF_SRCTRACK_RULE }; struct node_state_opt { int type; union { u_int32_t max_states; u_int32_t max_src_states; u_int32_t max_src_conn; struct { u_int32_t limit; u_int32_t seconds; } max_src_conn_rate; struct { u_int8_t flush; char tblname[PF_TABLE_NAME_SIZE]; } overload; u_int32_t max_src_nodes; u_int8_t src_track; u_int32_t statelock; struct { int number; u_int32_t seconds; } timeout; } data; struct node_state_opt *next; struct node_state_opt *tail; }; struct peer { struct node_host *host; struct node_port *port; }; static struct node_queue { char queue[PF_QNAME_SIZE]; char parent[PF_QNAME_SIZE]; char ifname[IFNAMSIZ]; int scheduler; struct node_queue *next; struct node_queue *tail; } *queues = NULL; struct node_qassign { char *qname; char *pqname; }; static struct filter_opts { int marker; #define FOM_FLAGS 0x01 #define FOM_ICMP 0x02 #define FOM_TOS 0x04 #define FOM_KEEP 0x08 #define FOM_SRCTRACK 0x10 #define FOM_SETPRIO 0x0400 #define FOM_PRIO 0x2000 struct node_uid *uid; struct node_gid *gid; struct { u_int8_t b1; u_int8_t b2; u_int16_t w; u_int16_t w2; } flags; struct node_icmp *icmpspec; u_int32_t tos; u_int32_t prob; struct { int action; struct node_state_opt *options; } keep; int fragment; int allowopts; char *label; struct node_qassign queues; char *tag; char *match_tag; u_int8_t match_tag_not; u_int rtableid; u_int8_t prio; u_int8_t set_prio[2]; struct { struct node_host *addr; u_int16_t port; } divert; } filter_opts; static struct antispoof_opts { char *label; u_int rtableid; } antispoof_opts; static struct scrub_opts { int marker; #define SOM_MINTTL 0x01 #define SOM_MAXMSS 0x02 #define SOM_FRAGCACHE 0x04 #define SOM_SETTOS 0x08 int nodf; int minttl; int maxmss; int settos; int fragcache; int randomid; int reassemble_tcp; char *match_tag; u_int8_t match_tag_not; u_int rtableid; } scrub_opts; static struct queue_opts { int marker; #define QOM_BWSPEC 0x01 #define QOM_SCHEDULER 0x02 #define QOM_PRIORITY 0x04 #define QOM_TBRSIZE 0x08 #define QOM_QLIMIT 0x10 struct node_queue_bw queue_bwspec; struct node_queue_opt scheduler; int priority; int tbrsize; int qlimit; } queue_opts; static struct table_opts { int flags; int init_addr; struct node_tinithead init_nodes; } table_opts; static struct pool_opts { int marker; #define POM_TYPE 0x01 #define POM_STICKYADDRESS 0x02 u_int8_t opts; int type; int staticport; struct pf_poolhashkey *key; } pool_opts; static struct codel_opts codel_opts; static struct node_hfsc_opts hfsc_opts; static struct node_fairq_opts fairq_opts; static struct node_state_opt *keep_state_defaults = NULL; int disallow_table(struct node_host *, const char *); int disallow_urpf_failed(struct node_host *, const char *); int disallow_alias(struct node_host *, const char *); int rule_consistent(struct pf_rule *, int); int filter_consistent(struct pf_rule *, int); int nat_consistent(struct pf_rule *); int rdr_consistent(struct pf_rule *); int process_tabledef(char *, struct table_opts *); void expand_label_str(char *, size_t, const char *, const char *); void expand_label_if(const char *, char *, size_t, const char *); void expand_label_addr(const char *, char *, size_t, u_int8_t, struct node_host *); void expand_label_port(const char *, char *, size_t, struct node_port *); void expand_label_proto(const char *, char *, size_t, u_int8_t); void expand_label_nr(const char *, char *, size_t); void expand_label(char *, size_t, const char *, u_int8_t, struct node_host *, struct node_port *, struct node_host *, struct node_port *, u_int8_t); void expand_rule(struct pf_rule *, struct node_if *, struct node_host *, struct node_proto *, struct node_os *, struct node_host *, struct node_port *, struct node_host *, struct node_port *, struct node_uid *, struct node_gid *, struct node_icmp *, const char *); int expand_altq(struct pf_altq *, struct node_if *, struct node_queue *, struct node_queue_bw bwspec, struct node_queue_opt *); int expand_queue(struct pf_altq *, struct node_if *, struct node_queue *, struct node_queue_bw, struct node_queue_opt *); int expand_skip_interface(struct node_if *); int check_rulestate(int); int getservice(char *); int rule_label(struct pf_rule *, char *); int rt_tableid_max(void); void mv_rules(struct pf_ruleset *, struct pf_ruleset *); void decide_address_family(struct node_host *, sa_family_t *); void remove_invalid_hosts(struct node_host **, sa_family_t *); int invalid_redirect(struct node_host *, sa_family_t); u_int16_t parseicmpspec(char *, sa_family_t); static TAILQ_HEAD(loadanchorshead, loadanchors) loadanchorshead = TAILQ_HEAD_INITIALIZER(loadanchorshead); struct loadanchors { TAILQ_ENTRY(loadanchors) entries; char *anchorname; char *filename; }; typedef struct { union { int64_t number; double probability; int i; char *string; u_int rtableid; struct { u_int8_t b1; u_int8_t b2; u_int16_t w; u_int16_t w2; } b; struct range { int a; int b; int t; } range; struct node_if *interface; struct node_proto *proto; struct node_icmp *icmp; struct node_host *host; struct node_os *os; struct node_port *port; struct node_uid *uid; struct node_gid *gid; struct node_state_opt *state_opt; struct peer peer; struct { struct peer src, dst; struct node_os *src_os; } fromto; struct { struct node_host *host; u_int8_t rt; u_int8_t pool_opts; sa_family_t af; struct pf_poolhashkey *key; } route; struct redirection { struct node_host *host; struct range rport; } *redirection; struct { int action; struct node_state_opt *options; } keep_state; struct { u_int8_t log; u_int8_t logif; u_int8_t quick; } logquick; struct { int neg; char *name; } tagged; struct pf_poolhashkey *hashkey; struct node_queue *queue; struct node_queue_opt queue_options; struct node_queue_bw queue_bwspec; struct node_qassign qassign; struct filter_opts filter_opts; struct antispoof_opts antispoof_opts; struct queue_opts queue_opts; struct scrub_opts scrub_opts; struct table_opts table_opts; struct pool_opts pool_opts; struct node_hfsc_opts hfsc_opts; struct node_fairq_opts fairq_opts; struct codel_opts codel_opts; } v; int lineno; } YYSTYPE; #define PPORT_RANGE 1 #define PPORT_STAR 2 int parseport(char *, struct range *r, int); #define DYNIF_MULTIADDR(addr) ((addr).type == PF_ADDR_DYNIFTL && \ (!((addr).iflags & PFI_AFLAG_NOALIAS) || \ !isdigit((addr).v.ifname[strlen((addr).v.ifname)-1]))) %} %token PASS BLOCK SCRUB RETURN IN OS OUT LOG QUICK ON FROM TO FLAGS %token RETURNRST RETURNICMP RETURNICMP6 PROTO INET INET6 ALL ANY ICMPTYPE %token ICMP6TYPE CODE KEEP MODULATE STATE PORT RDR NAT BINAT ARROW NODF %token MINTTL ERROR ALLOWOPTS FASTROUTE FILENAME ROUTETO DUPTO REPLYTO NO LABEL %token NOROUTE URPFFAILED FRAGMENT USER GROUP MAXMSS MAXIMUM TTL TOS DROP TABLE %token REASSEMBLE FRAGDROP FRAGCROP ANCHOR NATANCHOR RDRANCHOR BINATANCHOR %token SET OPTIMIZATION TIMEOUT LIMIT LOGINTERFACE BLOCKPOLICY RANDOMID %token REQUIREORDER SYNPROXY FINGERPRINTS NOSYNC DEBUG SKIP HOSTID %token ANTISPOOF FOR INCLUDE %token BITMASK RANDOM SOURCEHASH ROUNDROBIN STATICPORT PROBABILITY %token ALTQ CBQ CODEL PRIQ HFSC FAIRQ BANDWIDTH TBRSIZE LINKSHARE REALTIME %token UPPERLIMIT QUEUE PRIORITY QLIMIT HOGS BUCKETS RTABLE TARGET INTERVAL %token LOAD RULESET_OPTIMIZATION PRIO %token STICKYADDRESS MAXSRCSTATES MAXSRCNODES SOURCETRACK GLOBAL RULE %token MAXSRCCONN MAXSRCCONNRATE OVERLOAD FLUSH SLOPPY %token TAGGED TAG IFBOUND FLOATING STATEPOLICY STATEDEFAULTS ROUTE SETTOS %token DIVERTTO DIVERTREPLY %token STRING %token NUMBER %token PORTBINARY %type interface if_list if_item_not if_item %type number icmptype icmp6type uid gid %type tos not yesno %type probability %type no dir af fragcache optimizer %type sourcetrack flush unaryop statelock %type action nataction natpasslog scrubaction %type flags flag blockspec prio %type portplain portstar portrange %type hashkey %type proto proto_list proto_item %type protoval %type icmpspec %type icmp_list icmp_item %type icmp6_list icmp6_item %type reticmpspec reticmp6spec %type fromto %type ipportspec from to %type ipspec toipspec xhost host dynaddr host_list %type redir_host_list redirspec %type route_host route_host_list routespec %type os xos os_list %type portspec port_list port_item %type uids uid_list uid_item %type gids gid_list gid_item %type route %type redirection redirpool %type label stringall tag anchorname %type string varstring numberstring %type keep %type state_opt_spec state_opt_list state_opt_item %type logquick quick log logopts logopt %type antispoof_ifspc antispoof_iflst antispoof_if %type qname %type qassign qassign_list qassign_item %type scheduler %type cbqflags_list cbqflags_item %type priqflags_list priqflags_item %type hfscopts_list hfscopts_item hfsc_opts %type fairqopts_list fairqopts_item fairq_opts %type codelopts_list codelopts_item codel_opts %type bandwidth %type filter_opts filter_opt filter_opts_l %type filter_sets filter_set filter_sets_l %type antispoof_opts antispoof_opt antispoof_opts_l %type queue_opts queue_opt queue_opts_l %type scrub_opts scrub_opt scrub_opts_l %type table_opts table_opt table_opts_l %type pool_opts pool_opt pool_opts_l %type tagged %type rtable %% ruleset : /* empty */ | ruleset include '\n' | ruleset '\n' | ruleset option '\n' | ruleset scrubrule '\n' | ruleset natrule '\n' | ruleset binatrule '\n' | ruleset pfrule '\n' | ruleset anchorrule '\n' | ruleset loadrule '\n' | ruleset altqif '\n' | ruleset queuespec '\n' | ruleset varset '\n' | ruleset antispoof '\n' | ruleset tabledef '\n' | '{' fakeanchor '}' '\n'; | ruleset error '\n' { file->errors++; } ; include : INCLUDE STRING { struct file *nfile; if ((nfile = pushfile($2, 0)) == NULL) { yyerror("failed to include file %s", $2); free($2); YYERROR; } free($2); file = nfile; lungetc('\n'); } ; /* * apply to previouslys specified rule: must be careful to note * what that is: pf or nat or binat or rdr */ fakeanchor : fakeanchor '\n' | fakeanchor anchorrule '\n' | fakeanchor binatrule '\n' | fakeanchor natrule '\n' | fakeanchor pfrule '\n' | fakeanchor error '\n' ; optimizer : string { if (!strcmp($1, "none")) $$ = 0; else if (!strcmp($1, "basic")) $$ = PF_OPTIMIZE_BASIC; else if (!strcmp($1, "profile")) $$ = PF_OPTIMIZE_BASIC | PF_OPTIMIZE_PROFILE; else { yyerror("unknown ruleset-optimization %s", $1); YYERROR; } } ; option : SET OPTIMIZATION STRING { if (check_rulestate(PFCTL_STATE_OPTION)) { free($3); YYERROR; } if (pfctl_set_optimization(pf, $3) != 0) { yyerror("unknown optimization %s", $3); free($3); YYERROR; } free($3); } | SET RULESET_OPTIMIZATION optimizer { if (!(pf->opts & PF_OPT_OPTIMIZE)) { pf->opts |= PF_OPT_OPTIMIZE; pf->optimize = $3; } } | SET TIMEOUT timeout_spec | SET TIMEOUT '{' optnl timeout_list '}' | SET LIMIT limit_spec | SET LIMIT '{' optnl limit_list '}' | SET LOGINTERFACE stringall { if (check_rulestate(PFCTL_STATE_OPTION)) { free($3); YYERROR; } if (pfctl_set_logif(pf, $3) != 0) { yyerror("error setting loginterface %s", $3); free($3); YYERROR; } free($3); } | SET HOSTID number { if ($3 == 0 || $3 > UINT_MAX) { yyerror("hostid must be non-zero"); YYERROR; } if (pfctl_set_hostid(pf, $3) != 0) { yyerror("error setting hostid %08x", $3); YYERROR; } } | SET BLOCKPOLICY DROP { if (pf->opts & PF_OPT_VERBOSE) printf("set block-policy drop\n"); if (check_rulestate(PFCTL_STATE_OPTION)) YYERROR; blockpolicy = PFRULE_DROP; } | SET BLOCKPOLICY RETURN { if (pf->opts & PF_OPT_VERBOSE) printf("set block-policy return\n"); if (check_rulestate(PFCTL_STATE_OPTION)) YYERROR; blockpolicy = PFRULE_RETURN; } | SET REQUIREORDER yesno { if (pf->opts & PF_OPT_VERBOSE) printf("set require-order %s\n", $3 == 1 ? "yes" : "no"); require_order = $3; } | SET FINGERPRINTS STRING { if (pf->opts & PF_OPT_VERBOSE) printf("set fingerprints \"%s\"\n", $3); if (check_rulestate(PFCTL_STATE_OPTION)) { free($3); YYERROR; } if (!pf->anchor->name[0]) { if (pfctl_file_fingerprints(pf->dev, pf->opts, $3)) { yyerror("error loading " "fingerprints %s", $3); free($3); YYERROR; } } free($3); } | SET STATEPOLICY statelock { if (pf->opts & PF_OPT_VERBOSE) switch ($3) { case 0: printf("set state-policy floating\n"); break; case PFRULE_IFBOUND: printf("set state-policy if-bound\n"); break; } default_statelock = $3; } | SET DEBUG STRING { if (check_rulestate(PFCTL_STATE_OPTION)) { free($3); YYERROR; } if (pfctl_set_debug(pf, $3) != 0) { yyerror("error setting debuglevel %s", $3); free($3); YYERROR; } free($3); } | SET SKIP interface { if (expand_skip_interface($3) != 0) { yyerror("error setting skip interface(s)"); YYERROR; } } | SET STATEDEFAULTS state_opt_list { if (keep_state_defaults != NULL) { yyerror("cannot redefine state-defaults"); YYERROR; } keep_state_defaults = $3; } ; stringall : STRING { $$ = $1; } | ALL { if (($$ = strdup("all")) == NULL) { err(1, "stringall: strdup"); } } ; string : STRING string { if (asprintf(&$$, "%s %s", $1, $2) == -1) err(1, "string: asprintf"); free($1); free($2); } | STRING ; varstring : numberstring varstring { if (asprintf(&$$, "%s %s", $1, $2) == -1) err(1, "string: asprintf"); free($1); free($2); } | numberstring ; numberstring : NUMBER { char *s; if (asprintf(&s, "%lld", (long long)$1) == -1) { yyerror("string: asprintf"); YYERROR; } $$ = s; } | STRING ; varset : STRING '=' varstring { if (pf->opts & PF_OPT_VERBOSE) printf("%s = \"%s\"\n", $1, $3); if (symset($1, $3, 0) == -1) err(1, "cannot store variable %s", $1); free($1); free($3); } ; anchorname : STRING { $$ = $1; } | /* empty */ { $$ = NULL; } ; pfa_anchorlist : /* empty */ | pfa_anchorlist '\n' | pfa_anchorlist pfrule '\n' | pfa_anchorlist anchorrule '\n' ; pfa_anchor : '{' { char ta[PF_ANCHOR_NAME_SIZE]; struct pf_ruleset *rs; /* steping into a brace anchor */ pf->asd++; pf->bn++; pf->brace = 1; /* create a holding ruleset in the root */ snprintf(ta, PF_ANCHOR_NAME_SIZE, "_%d", pf->bn); rs = pf_find_or_create_ruleset(ta); if (rs == NULL) err(1, "pfa_anchor: pf_find_or_create_ruleset"); pf->astack[pf->asd] = rs->anchor; pf->anchor = rs->anchor; } '\n' pfa_anchorlist '}' { pf->alast = pf->anchor; pf->asd--; pf->anchor = pf->astack[pf->asd]; } | /* empty */ ; anchorrule : ANCHOR anchorname dir quick interface af proto fromto filter_opts pfa_anchor { struct pf_rule r; struct node_proto *proto; if (check_rulestate(PFCTL_STATE_FILTER)) { if ($2) free($2); YYERROR; } if ($2 && ($2[0] == '_' || strstr($2, "/_") != NULL)) { free($2); yyerror("anchor names beginning with '_' " "are reserved for internal use"); YYERROR; } memset(&r, 0, sizeof(r)); if (pf->astack[pf->asd + 1]) { /* move inline rules into relative location */ pf_anchor_setup(&r, &pf->astack[pf->asd]->ruleset, $2 ? $2 : pf->alast->name); if (r.anchor == NULL) err(1, "anchorrule: unable to " "create ruleset"); if (pf->alast != r.anchor) { if (r.anchor->match) { yyerror("inline anchor '%s' " "already exists", r.anchor->name); YYERROR; } mv_rules(&pf->alast->ruleset, &r.anchor->ruleset); } pf_remove_if_empty_ruleset(&pf->alast->ruleset); pf->alast = r.anchor; } else { if (!$2) { yyerror("anchors without explicit " "rules must specify a name"); YYERROR; } } r.direction = $3; r.quick = $4.quick; r.af = $6; r.prob = $9.prob; r.rtableid = $9.rtableid; if ($9.tag) if (strlcpy(r.tagname, $9.tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } if ($9.match_tag) if (strlcpy(r.match_tagname, $9.match_tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } r.match_tag_not = $9.match_tag_not; if (rule_label(&r, $9.label)) YYERROR; free($9.label); r.flags = $9.flags.b1; r.flagset = $9.flags.b2; if (($9.flags.b1 & $9.flags.b2) != $9.flags.b1) { yyerror("flags always false"); YYERROR; } if ($9.flags.b1 || $9.flags.b2 || $8.src_os) { for (proto = $7; proto != NULL && proto->proto != IPPROTO_TCP; proto = proto->next) ; /* nothing */ if (proto == NULL && $7 != NULL) { if ($9.flags.b1 || $9.flags.b2) yyerror( "flags only apply to tcp"); if ($8.src_os) yyerror( "OS fingerprinting only " "applies to tcp"); YYERROR; } } r.tos = $9.tos; if ($9.keep.action) { yyerror("cannot specify state handling " "on anchors"); YYERROR; } if ($9.match_tag) if (strlcpy(r.match_tagname, $9.match_tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } r.match_tag_not = $9.match_tag_not; if ($9.marker & FOM_PRIO) { if ($9.prio == 0) r.prio = PF_PRIO_ZERO; else r.prio = $9.prio; } if ($9.marker & FOM_SETPRIO) { r.set_prio[0] = $9.set_prio[0]; r.set_prio[1] = $9.set_prio[1]; r.scrub_flags |= PFSTATE_SETPRIO; } decide_address_family($8.src.host, &r.af); decide_address_family($8.dst.host, &r.af); expand_rule(&r, $5, NULL, $7, $8.src_os, $8.src.host, $8.src.port, $8.dst.host, $8.dst.port, $9.uid, $9.gid, $9.icmpspec, pf->astack[pf->asd + 1] ? pf->alast->name : $2); free($2); pf->astack[pf->asd + 1] = NULL; } | NATANCHOR string interface af proto fromto rtable { struct pf_rule r; if (check_rulestate(PFCTL_STATE_NAT)) { free($2); YYERROR; } memset(&r, 0, sizeof(r)); r.action = PF_NAT; r.af = $4; r.rtableid = $7; decide_address_family($6.src.host, &r.af); decide_address_family($6.dst.host, &r.af); expand_rule(&r, $3, NULL, $5, $6.src_os, $6.src.host, $6.src.port, $6.dst.host, $6.dst.port, 0, 0, 0, $2); free($2); } | RDRANCHOR string interface af proto fromto rtable { struct pf_rule r; if (check_rulestate(PFCTL_STATE_NAT)) { free($2); YYERROR; } memset(&r, 0, sizeof(r)); r.action = PF_RDR; r.af = $4; r.rtableid = $7; decide_address_family($6.src.host, &r.af); decide_address_family($6.dst.host, &r.af); if ($6.src.port != NULL) { yyerror("source port parameter not supported" " in rdr-anchor"); YYERROR; } if ($6.dst.port != NULL) { if ($6.dst.port->next != NULL) { yyerror("destination port list " "expansion not supported in " "rdr-anchor"); YYERROR; } else if ($6.dst.port->op != PF_OP_EQ) { yyerror("destination port operators" " not supported in rdr-anchor"); YYERROR; } r.dst.port[0] = $6.dst.port->port[0]; r.dst.port[1] = $6.dst.port->port[1]; r.dst.port_op = $6.dst.port->op; } expand_rule(&r, $3, NULL, $5, $6.src_os, $6.src.host, $6.src.port, $6.dst.host, $6.dst.port, 0, 0, 0, $2); free($2); } | BINATANCHOR string interface af proto fromto rtable { struct pf_rule r; if (check_rulestate(PFCTL_STATE_NAT)) { free($2); YYERROR; } memset(&r, 0, sizeof(r)); r.action = PF_BINAT; r.af = $4; r.rtableid = $7; if ($5 != NULL) { if ($5->next != NULL) { yyerror("proto list expansion" " not supported in binat-anchor"); YYERROR; } r.proto = $5->proto; free($5); } if ($6.src.host != NULL || $6.src.port != NULL || $6.dst.host != NULL || $6.dst.port != NULL) { yyerror("fromto parameter not supported" " in binat-anchor"); YYERROR; } decide_address_family($6.src.host, &r.af); decide_address_family($6.dst.host, &r.af); pfctl_add_rule(pf, &r, $2); free($2); } ; loadrule : LOAD ANCHOR string FROM string { struct loadanchors *loadanchor; if (strlen(pf->anchor->name) + 1 + strlen($3) >= MAXPATHLEN) { yyerror("anchorname %s too long, max %u\n", $3, MAXPATHLEN - 1); free($3); YYERROR; } loadanchor = calloc(1, sizeof(struct loadanchors)); if (loadanchor == NULL) err(1, "loadrule: calloc"); if ((loadanchor->anchorname = malloc(MAXPATHLEN)) == NULL) err(1, "loadrule: malloc"); if (pf->anchor->name[0]) snprintf(loadanchor->anchorname, MAXPATHLEN, "%s/%s", pf->anchor->name, $3); else strlcpy(loadanchor->anchorname, $3, MAXPATHLEN); if ((loadanchor->filename = strdup($5)) == NULL) err(1, "loadrule: strdup"); TAILQ_INSERT_TAIL(&loadanchorshead, loadanchor, entries); free($3); free($5); }; scrubaction : no SCRUB { $$.b2 = $$.w = 0; if ($1) $$.b1 = PF_NOSCRUB; else $$.b1 = PF_SCRUB; } ; scrubrule : scrubaction dir logquick interface af proto fromto scrub_opts { struct pf_rule r; if (check_rulestate(PFCTL_STATE_SCRUB)) YYERROR; memset(&r, 0, sizeof(r)); r.action = $1.b1; r.direction = $2; r.log = $3.log; r.logif = $3.logif; if ($3.quick) { yyerror("scrub rules do not support 'quick'"); YYERROR; } r.af = $5; if ($8.nodf) r.rule_flag |= PFRULE_NODF; if ($8.randomid) r.rule_flag |= PFRULE_RANDOMID; if ($8.reassemble_tcp) { if (r.direction != PF_INOUT) { yyerror("reassemble tcp rules can not " "specify direction"); YYERROR; } r.rule_flag |= PFRULE_REASSEMBLE_TCP; } if ($8.minttl) r.min_ttl = $8.minttl; if ($8.maxmss) r.max_mss = $8.maxmss; if ($8.marker & SOM_SETTOS) { r.rule_flag |= PFRULE_SET_TOS; r.set_tos = $8.settos; } if ($8.fragcache) r.rule_flag |= $8.fragcache; if ($8.match_tag) if (strlcpy(r.match_tagname, $8.match_tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } r.match_tag_not = $8.match_tag_not; r.rtableid = $8.rtableid; expand_rule(&r, $4, NULL, $6, $7.src_os, $7.src.host, $7.src.port, $7.dst.host, $7.dst.port, NULL, NULL, NULL, ""); } ; scrub_opts : { bzero(&scrub_opts, sizeof scrub_opts); scrub_opts.rtableid = -1; } scrub_opts_l { $$ = scrub_opts; } | /* empty */ { bzero(&scrub_opts, sizeof scrub_opts); scrub_opts.rtableid = -1; $$ = scrub_opts; } ; scrub_opts_l : scrub_opts_l scrub_opt | scrub_opt ; scrub_opt : NODF { if (scrub_opts.nodf) { yyerror("no-df cannot be respecified"); YYERROR; } scrub_opts.nodf = 1; } | MINTTL NUMBER { if (scrub_opts.marker & SOM_MINTTL) { yyerror("min-ttl cannot be respecified"); YYERROR; } if ($2 < 0 || $2 > 255) { yyerror("illegal min-ttl value %d", $2); YYERROR; } scrub_opts.marker |= SOM_MINTTL; scrub_opts.minttl = $2; } | MAXMSS NUMBER { if (scrub_opts.marker & SOM_MAXMSS) { yyerror("max-mss cannot be respecified"); YYERROR; } if ($2 < 0 || $2 > 65535) { yyerror("illegal max-mss value %d", $2); YYERROR; } scrub_opts.marker |= SOM_MAXMSS; scrub_opts.maxmss = $2; } | SETTOS tos { if (scrub_opts.marker & SOM_SETTOS) { yyerror("set-tos cannot be respecified"); YYERROR; } scrub_opts.marker |= SOM_SETTOS; scrub_opts.settos = $2; } | fragcache { if (scrub_opts.marker & SOM_FRAGCACHE) { yyerror("fragcache cannot be respecified"); YYERROR; } scrub_opts.marker |= SOM_FRAGCACHE; scrub_opts.fragcache = $1; } | REASSEMBLE STRING { if (strcasecmp($2, "tcp") != 0) { yyerror("scrub reassemble supports only tcp, " "not '%s'", $2); free($2); YYERROR; } free($2); if (scrub_opts.reassemble_tcp) { yyerror("reassemble tcp cannot be respecified"); YYERROR; } scrub_opts.reassemble_tcp = 1; } | RANDOMID { if (scrub_opts.randomid) { yyerror("random-id cannot be respecified"); YYERROR; } scrub_opts.randomid = 1; } | RTABLE NUMBER { if ($2 < 0 || $2 > rt_tableid_max()) { yyerror("invalid rtable id"); YYERROR; } scrub_opts.rtableid = $2; } | not TAGGED string { scrub_opts.match_tag = $3; scrub_opts.match_tag_not = $1; } ; fragcache : FRAGMENT REASSEMBLE { $$ = 0; /* default */ } | FRAGMENT FRAGCROP { $$ = 0; } | FRAGMENT FRAGDROP { $$ = 0; } ; antispoof : ANTISPOOF logquick antispoof_ifspc af antispoof_opts { struct pf_rule r; struct node_host *h = NULL, *hh; struct node_if *i, *j; if (check_rulestate(PFCTL_STATE_FILTER)) YYERROR; for (i = $3; i; i = i->next) { bzero(&r, sizeof(r)); r.action = PF_DROP; r.direction = PF_IN; r.log = $2.log; r.logif = $2.logif; r.quick = $2.quick; r.af = $4; if (rule_label(&r, $5.label)) YYERROR; r.rtableid = $5.rtableid; j = calloc(1, sizeof(struct node_if)); if (j == NULL) err(1, "antispoof: calloc"); if (strlcpy(j->ifname, i->ifname, sizeof(j->ifname)) >= sizeof(j->ifname)) { free(j); yyerror("interface name too long"); YYERROR; } j->not = 1; if (i->dynamic) { h = calloc(1, sizeof(*h)); if (h == NULL) err(1, "address: calloc"); h->addr.type = PF_ADDR_DYNIFTL; set_ipmask(h, 128); if (strlcpy(h->addr.v.ifname, i->ifname, sizeof(h->addr.v.ifname)) >= sizeof(h->addr.v.ifname)) { free(h); yyerror( "interface name too long"); YYERROR; } hh = malloc(sizeof(*hh)); if (hh == NULL) err(1, "address: malloc"); bcopy(h, hh, sizeof(*hh)); h->addr.iflags = PFI_AFLAG_NETWORK; } else { h = ifa_lookup(j->ifname, PFI_AFLAG_NETWORK); hh = NULL; } if (h != NULL) expand_rule(&r, j, NULL, NULL, NULL, h, NULL, NULL, NULL, NULL, NULL, NULL, ""); if ((i->ifa_flags & IFF_LOOPBACK) == 0) { bzero(&r, sizeof(r)); r.action = PF_DROP; r.direction = PF_IN; r.log = $2.log; r.logif = $2.logif; r.quick = $2.quick; r.af = $4; if (rule_label(&r, $5.label)) YYERROR; r.rtableid = $5.rtableid; if (hh != NULL) h = hh; else h = ifa_lookup(i->ifname, 0); if (h != NULL) expand_rule(&r, NULL, NULL, NULL, NULL, h, NULL, NULL, NULL, NULL, NULL, NULL, ""); } else free(hh); } free($5.label); } ; antispoof_ifspc : FOR antispoof_if { $$ = $2; } | FOR '{' optnl antispoof_iflst '}' { $$ = $4; } ; antispoof_iflst : antispoof_if optnl { $$ = $1; } | antispoof_iflst comma antispoof_if optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; antispoof_if : if_item { $$ = $1; } | '(' if_item ')' { $2->dynamic = 1; $$ = $2; } ; antispoof_opts : { bzero(&antispoof_opts, sizeof antispoof_opts); antispoof_opts.rtableid = -1; } antispoof_opts_l { $$ = antispoof_opts; } | /* empty */ { bzero(&antispoof_opts, sizeof antispoof_opts); antispoof_opts.rtableid = -1; $$ = antispoof_opts; } ; antispoof_opts_l : antispoof_opts_l antispoof_opt | antispoof_opt ; antispoof_opt : label { if (antispoof_opts.label) { yyerror("label cannot be redefined"); YYERROR; } antispoof_opts.label = $1; } | RTABLE NUMBER { if ($2 < 0 || $2 > rt_tableid_max()) { yyerror("invalid rtable id"); YYERROR; } antispoof_opts.rtableid = $2; } ; not : '!' { $$ = 1; } | /* empty */ { $$ = 0; } ; tabledef : TABLE '<' STRING '>' table_opts { struct node_host *h, *nh; struct node_tinit *ti, *nti; if (strlen($3) >= PF_TABLE_NAME_SIZE) { yyerror("table name too long, max %d chars", PF_TABLE_NAME_SIZE - 1); free($3); YYERROR; } if (pf->loadopt & PFCTL_FLAG_TABLE) if (process_tabledef($3, &$5)) { free($3); YYERROR; } free($3); for (ti = SIMPLEQ_FIRST(&$5.init_nodes); ti != SIMPLEQ_END(&$5.init_nodes); ti = nti) { if (ti->file) free(ti->file); for (h = ti->host; h != NULL; h = nh) { nh = h->next; free(h); } nti = SIMPLEQ_NEXT(ti, entries); free(ti); } } ; table_opts : { bzero(&table_opts, sizeof table_opts); SIMPLEQ_INIT(&table_opts.init_nodes); } table_opts_l { $$ = table_opts; } | /* empty */ { bzero(&table_opts, sizeof table_opts); SIMPLEQ_INIT(&table_opts.init_nodes); $$ = table_opts; } ; table_opts_l : table_opts_l table_opt | table_opt ; table_opt : STRING { if (!strcmp($1, "const")) table_opts.flags |= PFR_TFLAG_CONST; else if (!strcmp($1, "persist")) table_opts.flags |= PFR_TFLAG_PERSIST; else if (!strcmp($1, "counters")) table_opts.flags |= PFR_TFLAG_COUNTERS; else { yyerror("invalid table option '%s'", $1); free($1); YYERROR; } free($1); } | '{' optnl '}' { table_opts.init_addr = 1; } | '{' optnl host_list '}' { struct node_host *n; struct node_tinit *ti; for (n = $3; n != NULL; n = n->next) { switch (n->addr.type) { case PF_ADDR_ADDRMASK: continue; /* ok */ case PF_ADDR_RANGE: yyerror("address ranges are not " "permitted inside tables"); break; case PF_ADDR_DYNIFTL: yyerror("dynamic addresses are not " "permitted inside tables"); break; case PF_ADDR_TABLE: yyerror("tables cannot contain tables"); break; case PF_ADDR_NOROUTE: yyerror("\"no-route\" is not permitted " "inside tables"); break; case PF_ADDR_URPFFAILED: yyerror("\"urpf-failed\" is not " "permitted inside tables"); break; default: yyerror("unknown address type %d", n->addr.type); } YYERROR; } if (!(ti = calloc(1, sizeof(*ti)))) err(1, "table_opt: calloc"); ti->host = $3; SIMPLEQ_INSERT_TAIL(&table_opts.init_nodes, ti, entries); table_opts.init_addr = 1; } | FILENAME STRING { struct node_tinit *ti; if (!(ti = calloc(1, sizeof(*ti)))) err(1, "table_opt: calloc"); ti->file = $2; SIMPLEQ_INSERT_TAIL(&table_opts.init_nodes, ti, entries); table_opts.init_addr = 1; } ; altqif : ALTQ interface queue_opts QUEUE qassign { struct pf_altq a; if (check_rulestate(PFCTL_STATE_QUEUE)) YYERROR; memset(&a, 0, sizeof(a)); if ($3.scheduler.qtype == ALTQT_NONE) { yyerror("no scheduler specified!"); YYERROR; } a.scheduler = $3.scheduler.qtype; a.qlimit = $3.qlimit; a.tbrsize = $3.tbrsize; if ($5 == NULL && $3.scheduler.qtype != ALTQT_CODEL) { yyerror("no child queues specified"); YYERROR; } if (expand_altq(&a, $2, $5, $3.queue_bwspec, &$3.scheduler)) YYERROR; } ; queuespec : QUEUE STRING interface queue_opts qassign { struct pf_altq a; if (check_rulestate(PFCTL_STATE_QUEUE)) { free($2); YYERROR; } memset(&a, 0, sizeof(a)); if (strlcpy(a.qname, $2, sizeof(a.qname)) >= sizeof(a.qname)) { yyerror("queue name too long (max " "%d chars)", PF_QNAME_SIZE-1); free($2); YYERROR; } free($2); if ($4.tbrsize) { yyerror("cannot specify tbrsize for queue"); YYERROR; } if ($4.priority > 255) { yyerror("priority out of range: max 255"); YYERROR; } a.priority = $4.priority; a.qlimit = $4.qlimit; a.scheduler = $4.scheduler.qtype; if (expand_queue(&a, $3, $5, $4.queue_bwspec, &$4.scheduler)) { yyerror("errors in queue definition"); YYERROR; } } ; queue_opts : { bzero(&queue_opts, sizeof queue_opts); queue_opts.priority = DEFAULT_PRIORITY; queue_opts.qlimit = DEFAULT_QLIMIT; queue_opts.scheduler.qtype = ALTQT_NONE; queue_opts.queue_bwspec.bw_percent = 100; } queue_opts_l { $$ = queue_opts; } | /* empty */ { bzero(&queue_opts, sizeof queue_opts); queue_opts.priority = DEFAULT_PRIORITY; queue_opts.qlimit = DEFAULT_QLIMIT; queue_opts.scheduler.qtype = ALTQT_NONE; queue_opts.queue_bwspec.bw_percent = 100; $$ = queue_opts; } ; queue_opts_l : queue_opts_l queue_opt | queue_opt ; queue_opt : BANDWIDTH bandwidth { if (queue_opts.marker & QOM_BWSPEC) { yyerror("bandwidth cannot be respecified"); YYERROR; } queue_opts.marker |= QOM_BWSPEC; queue_opts.queue_bwspec = $2; } | PRIORITY NUMBER { if (queue_opts.marker & QOM_PRIORITY) { yyerror("priority cannot be respecified"); YYERROR; } if ($2 < 0 || $2 > 255) { yyerror("priority out of range: max 255"); YYERROR; } queue_opts.marker |= QOM_PRIORITY; queue_opts.priority = $2; } | QLIMIT NUMBER { if (queue_opts.marker & QOM_QLIMIT) { yyerror("qlimit cannot be respecified"); YYERROR; } if ($2 < 0 || $2 > 65535) { yyerror("qlimit out of range: max 65535"); YYERROR; } queue_opts.marker |= QOM_QLIMIT; queue_opts.qlimit = $2; } | scheduler { if (queue_opts.marker & QOM_SCHEDULER) { yyerror("scheduler cannot be respecified"); YYERROR; } queue_opts.marker |= QOM_SCHEDULER; queue_opts.scheduler = $1; } | TBRSIZE NUMBER { if (queue_opts.marker & QOM_TBRSIZE) { yyerror("tbrsize cannot be respecified"); YYERROR; } if ($2 < 0 || $2 > 65535) { yyerror("tbrsize too big: max 65535"); YYERROR; } queue_opts.marker |= QOM_TBRSIZE; queue_opts.tbrsize = $2; } ; bandwidth : STRING { double bps; char *cp; $$.bw_percent = 0; bps = strtod($1, &cp); if (cp != NULL) { if (strlen(cp) > 1) { char *cu = cp + 1; if (!strcmp(cu, "Bit") || !strcmp(cu, "B") || !strcmp(cu, "bit") || !strcmp(cu, "b")) { *cu = 0; } } if (!strcmp(cp, "b")) ; /* nothing */ else if (!strcmp(cp, "K")) bps *= 1000; else if (!strcmp(cp, "M")) bps *= 1000 * 1000; else if (!strcmp(cp, "G")) bps *= 1000 * 1000 * 1000; else if (!strcmp(cp, "%")) { if (bps < 0 || bps > 100) { yyerror("bandwidth spec " "out of range"); free($1); YYERROR; } $$.bw_percent = bps; bps = 0; } else { yyerror("unknown unit %s", cp); free($1); YYERROR; } } free($1); $$.bw_absolute = (u_int32_t)bps; } | NUMBER { if ($1 < 0 || $1 > UINT_MAX) { yyerror("bandwidth number too big"); YYERROR; } $$.bw_percent = 0; $$.bw_absolute = $1; } ; scheduler : CBQ { $$.qtype = ALTQT_CBQ; $$.data.cbq_opts.flags = 0; } | CBQ '(' cbqflags_list ')' { $$.qtype = ALTQT_CBQ; $$.data.cbq_opts.flags = $3; } | PRIQ { $$.qtype = ALTQT_PRIQ; $$.data.priq_opts.flags = 0; } | PRIQ '(' priqflags_list ')' { $$.qtype = ALTQT_PRIQ; $$.data.priq_opts.flags = $3; } | HFSC { $$.qtype = ALTQT_HFSC; bzero(&$$.data.hfsc_opts, sizeof(struct node_hfsc_opts)); } | HFSC '(' hfsc_opts ')' { $$.qtype = ALTQT_HFSC; $$.data.hfsc_opts = $3; } | FAIRQ { $$.qtype = ALTQT_FAIRQ; bzero(&$$.data.fairq_opts, sizeof(struct node_fairq_opts)); } | FAIRQ '(' fairq_opts ')' { $$.qtype = ALTQT_FAIRQ; $$.data.fairq_opts = $3; } | CODEL { $$.qtype = ALTQT_CODEL; bzero(&$$.data.codel_opts, sizeof(struct codel_opts)); } | CODEL '(' codel_opts ')' { $$.qtype = ALTQT_CODEL; $$.data.codel_opts = $3; } ; cbqflags_list : cbqflags_item { $$ |= $1; } | cbqflags_list comma cbqflags_item { $$ |= $3; } ; cbqflags_item : STRING { if (!strcmp($1, "default")) $$ = CBQCLF_DEFCLASS; else if (!strcmp($1, "borrow")) $$ = CBQCLF_BORROW; else if (!strcmp($1, "red")) $$ = CBQCLF_RED; else if (!strcmp($1, "ecn")) $$ = CBQCLF_RED|CBQCLF_ECN; else if (!strcmp($1, "rio")) $$ = CBQCLF_RIO; else if (!strcmp($1, "codel")) $$ = CBQCLF_CODEL; else { yyerror("unknown cbq flag \"%s\"", $1); free($1); YYERROR; } free($1); } ; priqflags_list : priqflags_item { $$ |= $1; } | priqflags_list comma priqflags_item { $$ |= $3; } ; priqflags_item : STRING { if (!strcmp($1, "default")) $$ = PRCF_DEFAULTCLASS; else if (!strcmp($1, "red")) $$ = PRCF_RED; else if (!strcmp($1, "ecn")) $$ = PRCF_RED|PRCF_ECN; else if (!strcmp($1, "rio")) $$ = PRCF_RIO; else if (!strcmp($1, "codel")) $$ = PRCF_CODEL; else { yyerror("unknown priq flag \"%s\"", $1); free($1); YYERROR; } free($1); } ; hfsc_opts : { bzero(&hfsc_opts, sizeof(struct node_hfsc_opts)); } hfscopts_list { $$ = hfsc_opts; } ; hfscopts_list : hfscopts_item | hfscopts_list comma hfscopts_item ; hfscopts_item : LINKSHARE bandwidth { if (hfsc_opts.linkshare.used) { yyerror("linkshare already specified"); YYERROR; } hfsc_opts.linkshare.m2 = $2; hfsc_opts.linkshare.used = 1; } | LINKSHARE '(' bandwidth comma NUMBER comma bandwidth ')' { if ($5 < 0 || $5 > INT_MAX) { yyerror("timing in curve out of range"); YYERROR; } if (hfsc_opts.linkshare.used) { yyerror("linkshare already specified"); YYERROR; } hfsc_opts.linkshare.m1 = $3; hfsc_opts.linkshare.d = $5; hfsc_opts.linkshare.m2 = $7; hfsc_opts.linkshare.used = 1; } | REALTIME bandwidth { if (hfsc_opts.realtime.used) { yyerror("realtime already specified"); YYERROR; } hfsc_opts.realtime.m2 = $2; hfsc_opts.realtime.used = 1; } | REALTIME '(' bandwidth comma NUMBER comma bandwidth ')' { if ($5 < 0 || $5 > INT_MAX) { yyerror("timing in curve out of range"); YYERROR; } if (hfsc_opts.realtime.used) { yyerror("realtime already specified"); YYERROR; } hfsc_opts.realtime.m1 = $3; hfsc_opts.realtime.d = $5; hfsc_opts.realtime.m2 = $7; hfsc_opts.realtime.used = 1; } | UPPERLIMIT bandwidth { if (hfsc_opts.upperlimit.used) { yyerror("upperlimit already specified"); YYERROR; } hfsc_opts.upperlimit.m2 = $2; hfsc_opts.upperlimit.used = 1; } | UPPERLIMIT '(' bandwidth comma NUMBER comma bandwidth ')' { if ($5 < 0 || $5 > INT_MAX) { yyerror("timing in curve out of range"); YYERROR; } if (hfsc_opts.upperlimit.used) { yyerror("upperlimit already specified"); YYERROR; } hfsc_opts.upperlimit.m1 = $3; hfsc_opts.upperlimit.d = $5; hfsc_opts.upperlimit.m2 = $7; hfsc_opts.upperlimit.used = 1; } | STRING { if (!strcmp($1, "default")) hfsc_opts.flags |= HFCF_DEFAULTCLASS; else if (!strcmp($1, "red")) hfsc_opts.flags |= HFCF_RED; else if (!strcmp($1, "ecn")) hfsc_opts.flags |= HFCF_RED|HFCF_ECN; else if (!strcmp($1, "rio")) hfsc_opts.flags |= HFCF_RIO; else if (!strcmp($1, "codel")) hfsc_opts.flags |= HFCF_CODEL; else { yyerror("unknown hfsc flag \"%s\"", $1); free($1); YYERROR; } free($1); } ; fairq_opts : { bzero(&fairq_opts, sizeof(struct node_fairq_opts)); } fairqopts_list { $$ = fairq_opts; } ; fairqopts_list : fairqopts_item | fairqopts_list comma fairqopts_item ; fairqopts_item : LINKSHARE bandwidth { if (fairq_opts.linkshare.used) { yyerror("linkshare already specified"); YYERROR; } fairq_opts.linkshare.m2 = $2; fairq_opts.linkshare.used = 1; } | LINKSHARE '(' bandwidth number bandwidth ')' { if (fairq_opts.linkshare.used) { yyerror("linkshare already specified"); YYERROR; } fairq_opts.linkshare.m1 = $3; fairq_opts.linkshare.d = $4; fairq_opts.linkshare.m2 = $5; fairq_opts.linkshare.used = 1; } | HOGS bandwidth { fairq_opts.hogs_bw = $2; } | BUCKETS number { fairq_opts.nbuckets = $2; } | STRING { if (!strcmp($1, "default")) fairq_opts.flags |= FARF_DEFAULTCLASS; else if (!strcmp($1, "red")) fairq_opts.flags |= FARF_RED; else if (!strcmp($1, "ecn")) fairq_opts.flags |= FARF_RED|FARF_ECN; else if (!strcmp($1, "rio")) fairq_opts.flags |= FARF_RIO; else if (!strcmp($1, "codel")) fairq_opts.flags |= FARF_CODEL; else { yyerror("unknown fairq flag \"%s\"", $1); free($1); YYERROR; } free($1); } ; codel_opts : { bzero(&codel_opts, sizeof(struct codel_opts)); } codelopts_list { $$ = codel_opts; } ; codelopts_list : codelopts_item | codelopts_list comma codelopts_item ; codelopts_item : INTERVAL number { if (codel_opts.interval) { yyerror("interval already specified"); YYERROR; } codel_opts.interval = $2; } | TARGET number { if (codel_opts.target) { yyerror("target already specified"); YYERROR; } codel_opts.target = $2; } | STRING { if (!strcmp($1, "ecn")) codel_opts.ecn = 1; else { yyerror("unknown codel option \"%s\"", $1); free($1); YYERROR; } free($1); } ; qassign : /* empty */ { $$ = NULL; } | qassign_item { $$ = $1; } | '{' optnl qassign_list '}' { $$ = $3; } ; qassign_list : qassign_item optnl { $$ = $1; } | qassign_list comma qassign_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; qassign_item : STRING { $$ = calloc(1, sizeof(struct node_queue)); if ($$ == NULL) err(1, "qassign_item: calloc"); if (strlcpy($$->queue, $1, sizeof($$->queue)) >= sizeof($$->queue)) { yyerror("queue name '%s' too long (max " "%d chars)", $1, sizeof($$->queue)-1); free($1); free($$); YYERROR; } free($1); $$->next = NULL; $$->tail = $$; } ; pfrule : action dir logquick interface route af proto fromto filter_opts { struct pf_rule r; struct node_state_opt *o; struct node_proto *proto; int srctrack = 0; int statelock = 0; int adaptive = 0; int defaults = 0; if (check_rulestate(PFCTL_STATE_FILTER)) YYERROR; memset(&r, 0, sizeof(r)); r.action = $1.b1; switch ($1.b2) { case PFRULE_RETURNRST: r.rule_flag |= PFRULE_RETURNRST; r.return_ttl = $1.w; break; case PFRULE_RETURNICMP: r.rule_flag |= PFRULE_RETURNICMP; r.return_icmp = $1.w; r.return_icmp6 = $1.w2; break; case PFRULE_RETURN: r.rule_flag |= PFRULE_RETURN; r.return_icmp = $1.w; r.return_icmp6 = $1.w2; break; } r.direction = $2; r.log = $3.log; r.logif = $3.logif; r.quick = $3.quick; r.prob = $9.prob; r.rtableid = $9.rtableid; if ($9.marker & FOM_PRIO) { if ($9.prio == 0) r.prio = PF_PRIO_ZERO; else r.prio = $9.prio; } if ($9.marker & FOM_SETPRIO) { r.set_prio[0] = $9.set_prio[0]; r.set_prio[1] = $9.set_prio[1]; r.scrub_flags |= PFSTATE_SETPRIO; } r.af = $6; if ($9.tag) if (strlcpy(r.tagname, $9.tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } if ($9.match_tag) if (strlcpy(r.match_tagname, $9.match_tag, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } r.match_tag_not = $9.match_tag_not; if (rule_label(&r, $9.label)) YYERROR; free($9.label); r.flags = $9.flags.b1; r.flagset = $9.flags.b2; if (($9.flags.b1 & $9.flags.b2) != $9.flags.b1) { yyerror("flags always false"); YYERROR; } if ($9.flags.b1 || $9.flags.b2 || $8.src_os) { for (proto = $7; proto != NULL && proto->proto != IPPROTO_TCP; proto = proto->next) ; /* nothing */ if (proto == NULL && $7 != NULL) { if ($9.flags.b1 || $9.flags.b2) yyerror( "flags only apply to tcp"); if ($8.src_os) yyerror( "OS fingerprinting only " "apply to tcp"); YYERROR; } #if 0 if (($9.flags.b1 & parse_flags("S")) == 0 && $8.src_os) { yyerror("OS fingerprinting requires " "the SYN TCP flag (flags S/SA)"); YYERROR; } #endif } r.tos = $9.tos; r.keep_state = $9.keep.action; o = $9.keep.options; /* 'keep state' by default on pass rules. */ if (!r.keep_state && !r.action && !($9.marker & FOM_KEEP)) { r.keep_state = PF_STATE_NORMAL; o = keep_state_defaults; defaults = 1; } while (o) { struct node_state_opt *p = o; switch (o->type) { case PF_STATE_OPT_MAX: if (r.max_states) { yyerror("state option 'max' " "multiple definitions"); YYERROR; } r.max_states = o->data.max_states; break; case PF_STATE_OPT_NOSYNC: if (r.rule_flag & PFRULE_NOSYNC) { yyerror("state option 'sync' " "multiple definitions"); YYERROR; } r.rule_flag |= PFRULE_NOSYNC; break; case PF_STATE_OPT_SRCTRACK: if (srctrack) { yyerror("state option " "'source-track' " "multiple definitions"); YYERROR; } srctrack = o->data.src_track; r.rule_flag |= PFRULE_SRCTRACK; break; case PF_STATE_OPT_MAX_SRC_STATES: if (r.max_src_states) { yyerror("state option " "'max-src-states' " "multiple definitions"); YYERROR; } if (o->data.max_src_states == 0) { yyerror("'max-src-states' must " "be > 0"); YYERROR; } r.max_src_states = o->data.max_src_states; r.rule_flag |= PFRULE_SRCTRACK; break; case PF_STATE_OPT_OVERLOAD: if (r.overload_tblname[0]) { yyerror("multiple 'overload' " "table definitions"); YYERROR; } if (strlcpy(r.overload_tblname, o->data.overload.tblname, PF_TABLE_NAME_SIZE) >= PF_TABLE_NAME_SIZE) { yyerror("state option: " "strlcpy"); YYERROR; } r.flush = o->data.overload.flush; break; case PF_STATE_OPT_MAX_SRC_CONN: if (r.max_src_conn) { yyerror("state option " "'max-src-conn' " "multiple definitions"); YYERROR; } if (o->data.max_src_conn == 0) { yyerror("'max-src-conn' " "must be > 0"); YYERROR; } r.max_src_conn = o->data.max_src_conn; r.rule_flag |= PFRULE_SRCTRACK | PFRULE_RULESRCTRACK; break; case PF_STATE_OPT_MAX_SRC_CONN_RATE: if (r.max_src_conn_rate.limit) { yyerror("state option " "'max-src-conn-rate' " "multiple definitions"); YYERROR; } if (!o->data.max_src_conn_rate.limit || !o->data.max_src_conn_rate.seconds) { yyerror("'max-src-conn-rate' " "values must be > 0"); YYERROR; } if (o->data.max_src_conn_rate.limit > PF_THRESHOLD_MAX) { yyerror("'max-src-conn-rate' " "maximum rate must be < %u", PF_THRESHOLD_MAX); YYERROR; } r.max_src_conn_rate.limit = o->data.max_src_conn_rate.limit; r.max_src_conn_rate.seconds = o->data.max_src_conn_rate.seconds; r.rule_flag |= PFRULE_SRCTRACK | PFRULE_RULESRCTRACK; break; case PF_STATE_OPT_MAX_SRC_NODES: if (r.max_src_nodes) { yyerror("state option " "'max-src-nodes' " "multiple definitions"); YYERROR; } if (o->data.max_src_nodes == 0) { yyerror("'max-src-nodes' must " "be > 0"); YYERROR; } r.max_src_nodes = o->data.max_src_nodes; r.rule_flag |= PFRULE_SRCTRACK | PFRULE_RULESRCTRACK; break; case PF_STATE_OPT_STATELOCK: if (statelock) { yyerror("state locking option: " "multiple definitions"); YYERROR; } statelock = 1; r.rule_flag |= o->data.statelock; break; case PF_STATE_OPT_SLOPPY: if (r.rule_flag & PFRULE_STATESLOPPY) { yyerror("state sloppy option: " "multiple definitions"); YYERROR; } r.rule_flag |= PFRULE_STATESLOPPY; break; case PF_STATE_OPT_TIMEOUT: if (o->data.timeout.number == PFTM_ADAPTIVE_START || o->data.timeout.number == PFTM_ADAPTIVE_END) adaptive = 1; if (r.timeout[o->data.timeout.number]) { yyerror("state timeout %s " "multiple definitions", pf_timeouts[o->data. timeout.number].name); YYERROR; } r.timeout[o->data.timeout.number] = o->data.timeout.seconds; } o = o->next; if (!defaults) free(p); } /* 'flags S/SA' by default on stateful rules */ if (!r.action && !r.flags && !r.flagset && !$9.fragment && !($9.marker & FOM_FLAGS) && r.keep_state) { r.flags = parse_flags("S"); r.flagset = parse_flags("SA"); } if (!adaptive && r.max_states) { r.timeout[PFTM_ADAPTIVE_START] = (r.max_states / 10) * 6; r.timeout[PFTM_ADAPTIVE_END] = (r.max_states / 10) * 12; } if (r.rule_flag & PFRULE_SRCTRACK) { if (srctrack == PF_SRCTRACK_GLOBAL && r.max_src_nodes) { yyerror("'max-src-nodes' is " "incompatible with " "'source-track global'"); YYERROR; } if (srctrack == PF_SRCTRACK_GLOBAL && r.max_src_conn) { yyerror("'max-src-conn' is " "incompatible with " "'source-track global'"); YYERROR; } if (srctrack == PF_SRCTRACK_GLOBAL && r.max_src_conn_rate.seconds) { yyerror("'max-src-conn-rate' is " "incompatible with " "'source-track global'"); YYERROR; } if (r.timeout[PFTM_SRC_NODE] < r.max_src_conn_rate.seconds) r.timeout[PFTM_SRC_NODE] = r.max_src_conn_rate.seconds; r.rule_flag |= PFRULE_SRCTRACK; if (srctrack == PF_SRCTRACK_RULE) r.rule_flag |= PFRULE_RULESRCTRACK; } if (r.keep_state && !statelock) r.rule_flag |= default_statelock; if ($9.fragment) r.rule_flag |= PFRULE_FRAGMENT; r.allow_opts = $9.allowopts; decide_address_family($8.src.host, &r.af); decide_address_family($8.dst.host, &r.af); if ($5.rt) { if (!r.direction) { yyerror("direction must be explicit " "with rules that specify routing"); YYERROR; } r.rt = $5.rt; r.rpool.opts = $5.pool_opts; if ($5.key != NULL) memcpy(&r.rpool.key, $5.key, sizeof(struct pf_poolhashkey)); } - if (r.rt && r.rt != PF_FASTROUTE) { + if (r.rt) { decide_address_family($5.host, &r.af); remove_invalid_hosts(&$5.host, &r.af); if ($5.host == NULL) { yyerror("no routing address with " "matching address family found."); YYERROR; } if ((r.rpool.opts & PF_POOL_TYPEMASK) == PF_POOL_NONE && ($5.host->next != NULL || $5.host->addr.type == PF_ADDR_TABLE || DYNIF_MULTIADDR($5.host->addr))) r.rpool.opts |= PF_POOL_ROUNDROBIN; if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN && disallow_table($5.host, "tables are only " "supported in round-robin routing pools")) YYERROR; if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN && disallow_alias($5.host, "interface (%s) " "is only supported in round-robin " "routing pools")) YYERROR; if ($5.host->next != NULL) { if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN) { yyerror("r.rpool.opts must " "be PF_POOL_ROUNDROBIN"); YYERROR; } } } if ($9.queues.qname != NULL) { if (strlcpy(r.qname, $9.queues.qname, sizeof(r.qname)) >= sizeof(r.qname)) { yyerror("rule qname too long (max " "%d chars)", sizeof(r.qname)-1); YYERROR; } free($9.queues.qname); } if ($9.queues.pqname != NULL) { if (strlcpy(r.pqname, $9.queues.pqname, sizeof(r.pqname)) >= sizeof(r.pqname)) { yyerror("rule pqname too long (max " "%d chars)", sizeof(r.pqname)-1); YYERROR; } free($9.queues.pqname); } #ifdef __FreeBSD__ r.divert.port = $9.divert.port; #else if ((r.divert.port = $9.divert.port)) { if (r.direction == PF_OUT) { if ($9.divert.addr) { yyerror("address specified " "for outgoing divert"); YYERROR; } bzero(&r.divert.addr, sizeof(r.divert.addr)); } else { if (!$9.divert.addr) { yyerror("no address specified " "for incoming divert"); YYERROR; } if ($9.divert.addr->af != r.af) { yyerror("address family " "mismatch for divert"); YYERROR; } r.divert.addr = $9.divert.addr->addr.v.a.addr; } } #endif expand_rule(&r, $4, $5.host, $7, $8.src_os, $8.src.host, $8.src.port, $8.dst.host, $8.dst.port, $9.uid, $9.gid, $9.icmpspec, ""); } ; filter_opts : { bzero(&filter_opts, sizeof filter_opts); filter_opts.rtableid = -1; } filter_opts_l { $$ = filter_opts; } | /* empty */ { bzero(&filter_opts, sizeof filter_opts); filter_opts.rtableid = -1; $$ = filter_opts; } ; filter_opts_l : filter_opts_l filter_opt | filter_opt ; filter_opt : USER uids { if (filter_opts.uid) $2->tail->next = filter_opts.uid; filter_opts.uid = $2; } | GROUP gids { if (filter_opts.gid) $2->tail->next = filter_opts.gid; filter_opts.gid = $2; } | flags { if (filter_opts.marker & FOM_FLAGS) { yyerror("flags cannot be redefined"); YYERROR; } filter_opts.marker |= FOM_FLAGS; filter_opts.flags.b1 |= $1.b1; filter_opts.flags.b2 |= $1.b2; filter_opts.flags.w |= $1.w; filter_opts.flags.w2 |= $1.w2; } | icmpspec { if (filter_opts.marker & FOM_ICMP) { yyerror("icmp-type cannot be redefined"); YYERROR; } filter_opts.marker |= FOM_ICMP; filter_opts.icmpspec = $1; } | PRIO NUMBER { if (filter_opts.marker & FOM_PRIO) { yyerror("prio cannot be redefined"); YYERROR; } if ($2 < 0 || $2 > PF_PRIO_MAX) { yyerror("prio must be 0 - %u", PF_PRIO_MAX); YYERROR; } filter_opts.marker |= FOM_PRIO; filter_opts.prio = $2; } | TOS tos { if (filter_opts.marker & FOM_TOS) { yyerror("tos cannot be redefined"); YYERROR; } filter_opts.marker |= FOM_TOS; filter_opts.tos = $2; } | keep { if (filter_opts.marker & FOM_KEEP) { yyerror("modulate or keep cannot be redefined"); YYERROR; } filter_opts.marker |= FOM_KEEP; filter_opts.keep.action = $1.action; filter_opts.keep.options = $1.options; } | FRAGMENT { filter_opts.fragment = 1; } | ALLOWOPTS { filter_opts.allowopts = 1; } | label { if (filter_opts.label) { yyerror("label cannot be redefined"); YYERROR; } filter_opts.label = $1; } | qname { if (filter_opts.queues.qname) { yyerror("queue cannot be redefined"); YYERROR; } filter_opts.queues = $1; } | TAG string { filter_opts.tag = $2; } | not TAGGED string { filter_opts.match_tag = $3; filter_opts.match_tag_not = $1; } | PROBABILITY probability { double p; p = floor($2 * UINT_MAX + 0.5); if (p < 0.0 || p > UINT_MAX) { yyerror("invalid probability: %lf", p); YYERROR; } filter_opts.prob = (u_int32_t)p; if (filter_opts.prob == 0) filter_opts.prob = 1; } | RTABLE NUMBER { if ($2 < 0 || $2 > rt_tableid_max()) { yyerror("invalid rtable id"); YYERROR; } filter_opts.rtableid = $2; } | DIVERTTO portplain { #ifdef __FreeBSD__ filter_opts.divert.port = $2.a; if (!filter_opts.divert.port) { yyerror("invalid divert port: %u", ntohs($2.a)); YYERROR; } #endif } | DIVERTTO STRING PORT portplain { #ifndef __FreeBSD__ if ((filter_opts.divert.addr = host($2)) == NULL) { yyerror("could not parse divert address: %s", $2); free($2); YYERROR; } #else if ($2) #endif free($2); filter_opts.divert.port = $4.a; if (!filter_opts.divert.port) { yyerror("invalid divert port: %u", ntohs($4.a)); YYERROR; } } | DIVERTREPLY { #ifdef __FreeBSD__ yyerror("divert-reply has no meaning in FreeBSD pf(4)"); YYERROR; #else filter_opts.divert.port = 1; /* some random value */ #endif } | filter_sets ; filter_sets : SET '(' filter_sets_l ')' { $$ = filter_opts; } | SET filter_set { $$ = filter_opts; } ; filter_sets_l : filter_sets_l comma filter_set | filter_set ; filter_set : prio { if (filter_opts.marker & FOM_SETPRIO) { yyerror("prio cannot be redefined"); YYERROR; } filter_opts.marker |= FOM_SETPRIO; filter_opts.set_prio[0] = $1.b1; filter_opts.set_prio[1] = $1.b2; } prio : PRIO NUMBER { if ($2 < 0 || $2 > PF_PRIO_MAX) { yyerror("prio must be 0 - %u", PF_PRIO_MAX); YYERROR; } $$.b1 = $$.b2 = $2; } | PRIO '(' NUMBER comma NUMBER ')' { if ($3 < 0 || $3 > PF_PRIO_MAX || $5 < 0 || $5 > PF_PRIO_MAX) { yyerror("prio must be 0 - %u", PF_PRIO_MAX); YYERROR; } $$.b1 = $3; $$.b2 = $5; } ; probability : STRING { char *e; double p = strtod($1, &e); if (*e == '%') { p *= 0.01; e++; } if (*e) { yyerror("invalid probability: %s", $1); free($1); YYERROR; } free($1); $$ = p; } | NUMBER { $$ = (double)$1; } ; action : PASS { $$.b1 = PF_PASS; $$.b2 = $$.w = 0; } | BLOCK blockspec { $$ = $2; $$.b1 = PF_DROP; } ; blockspec : /* empty */ { $$.b2 = blockpolicy; $$.w = returnicmpdefault; $$.w2 = returnicmp6default; } | DROP { $$.b2 = PFRULE_DROP; $$.w = 0; $$.w2 = 0; } | RETURNRST { $$.b2 = PFRULE_RETURNRST; $$.w = 0; $$.w2 = 0; } | RETURNRST '(' TTL NUMBER ')' { if ($4 < 0 || $4 > 255) { yyerror("illegal ttl value %d", $4); YYERROR; } $$.b2 = PFRULE_RETURNRST; $$.w = $4; $$.w2 = 0; } | RETURNICMP { $$.b2 = PFRULE_RETURNICMP; $$.w = returnicmpdefault; $$.w2 = returnicmp6default; } | RETURNICMP6 { $$.b2 = PFRULE_RETURNICMP; $$.w = returnicmpdefault; $$.w2 = returnicmp6default; } | RETURNICMP '(' reticmpspec ')' { $$.b2 = PFRULE_RETURNICMP; $$.w = $3; $$.w2 = returnicmpdefault; } | RETURNICMP6 '(' reticmp6spec ')' { $$.b2 = PFRULE_RETURNICMP; $$.w = returnicmpdefault; $$.w2 = $3; } | RETURNICMP '(' reticmpspec comma reticmp6spec ')' { $$.b2 = PFRULE_RETURNICMP; $$.w = $3; $$.w2 = $5; } | RETURN { $$.b2 = PFRULE_RETURN; $$.w = returnicmpdefault; $$.w2 = returnicmp6default; } ; reticmpspec : STRING { if (!($$ = parseicmpspec($1, AF_INET))) { free($1); YYERROR; } free($1); } | NUMBER { u_int8_t icmptype; if ($1 < 0 || $1 > 255) { yyerror("invalid icmp code %lu", $1); YYERROR; } icmptype = returnicmpdefault >> 8; $$ = (icmptype << 8 | $1); } ; reticmp6spec : STRING { if (!($$ = parseicmpspec($1, AF_INET6))) { free($1); YYERROR; } free($1); } | NUMBER { u_int8_t icmptype; if ($1 < 0 || $1 > 255) { yyerror("invalid icmp code %lu", $1); YYERROR; } icmptype = returnicmp6default >> 8; $$ = (icmptype << 8 | $1); } ; dir : /* empty */ { $$ = PF_INOUT; } | IN { $$ = PF_IN; } | OUT { $$ = PF_OUT; } ; quick : /* empty */ { $$.quick = 0; } | QUICK { $$.quick = 1; } ; logquick : /* empty */ { $$.log = 0; $$.quick = 0; $$.logif = 0; } | log { $$ = $1; $$.quick = 0; } | QUICK { $$.quick = 1; $$.log = 0; $$.logif = 0; } | log QUICK { $$ = $1; $$.quick = 1; } | QUICK log { $$ = $2; $$.quick = 1; } ; log : LOG { $$.log = PF_LOG; $$.logif = 0; } | LOG '(' logopts ')' { $$.log = PF_LOG | $3.log; $$.logif = $3.logif; } ; logopts : logopt { $$ = $1; } | logopts comma logopt { $$.log = $1.log | $3.log; $$.logif = $3.logif; if ($$.logif == 0) $$.logif = $1.logif; } ; logopt : ALL { $$.log = PF_LOG_ALL; $$.logif = 0; } | USER { $$.log = PF_LOG_SOCKET_LOOKUP; $$.logif = 0; } | GROUP { $$.log = PF_LOG_SOCKET_LOOKUP; $$.logif = 0; } | TO string { const char *errstr; u_int i; $$.log = 0; if (strncmp($2, "pflog", 5)) { yyerror("%s: should be a pflog interface", $2); free($2); YYERROR; } i = strtonum($2 + 5, 0, 255, &errstr); if (errstr) { yyerror("%s: %s", $2, errstr); free($2); YYERROR; } free($2); $$.logif = i; } ; interface : /* empty */ { $$ = NULL; } | ON if_item_not { $$ = $2; } | ON '{' optnl if_list '}' { $$ = $4; } ; if_list : if_item_not optnl { $$ = $1; } | if_list comma if_item_not optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; if_item_not : not if_item { $$ = $2; $$->not = $1; } ; if_item : STRING { struct node_host *n; $$ = calloc(1, sizeof(struct node_if)); if ($$ == NULL) err(1, "if_item: calloc"); if (strlcpy($$->ifname, $1, sizeof($$->ifname)) >= sizeof($$->ifname)) { free($1); free($$); yyerror("interface name too long"); YYERROR; } if ((n = ifa_exists($1)) != NULL) $$->ifa_flags = n->ifa_flags; free($1); $$->not = 0; $$->next = NULL; $$->tail = $$; } ; af : /* empty */ { $$ = 0; } | INET { $$ = AF_INET; } | INET6 { $$ = AF_INET6; } ; proto : /* empty */ { $$ = NULL; } | PROTO proto_item { $$ = $2; } | PROTO '{' optnl proto_list '}' { $$ = $4; } ; proto_list : proto_item optnl { $$ = $1; } | proto_list comma proto_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; proto_item : protoval { u_int8_t pr; pr = (u_int8_t)$1; if (pr == 0) { yyerror("proto 0 cannot be used"); YYERROR; } $$ = calloc(1, sizeof(struct node_proto)); if ($$ == NULL) err(1, "proto_item: calloc"); $$->proto = pr; $$->next = NULL; $$->tail = $$; } ; protoval : STRING { struct protoent *p; p = getprotobyname($1); if (p == NULL) { yyerror("unknown protocol %s", $1); free($1); YYERROR; } $$ = p->p_proto; free($1); } | NUMBER { if ($1 < 0 || $1 > 255) { yyerror("protocol outside range"); YYERROR; } } ; fromto : ALL { $$.src.host = NULL; $$.src.port = NULL; $$.dst.host = NULL; $$.dst.port = NULL; $$.src_os = NULL; } | from os to { $$.src = $1; $$.src_os = $2; $$.dst = $3; } ; os : /* empty */ { $$ = NULL; } | OS xos { $$ = $2; } | OS '{' optnl os_list '}' { $$ = $4; } ; xos : STRING { $$ = calloc(1, sizeof(struct node_os)); if ($$ == NULL) err(1, "os: calloc"); $$->os = $1; $$->tail = $$; } ; os_list : xos optnl { $$ = $1; } | os_list comma xos optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; from : /* empty */ { $$.host = NULL; $$.port = NULL; } | FROM ipportspec { $$ = $2; } ; to : /* empty */ { $$.host = NULL; $$.port = NULL; } | TO ipportspec { if (disallow_urpf_failed($2.host, "\"urpf-failed\" is " "not permitted in a destination address")) YYERROR; $$ = $2; } ; ipportspec : ipspec { $$.host = $1; $$.port = NULL; } | ipspec PORT portspec { $$.host = $1; $$.port = $3; } | PORT portspec { $$.host = NULL; $$.port = $2; } ; optnl : '\n' optnl | ; ipspec : ANY { $$ = NULL; } | xhost { $$ = $1; } | '{' optnl host_list '}' { $$ = $3; } ; toipspec : TO ipspec { $$ = $2; } | /* empty */ { $$ = NULL; } ; host_list : ipspec optnl { $$ = $1; } | host_list comma ipspec optnl { if ($3 == NULL) $$ = $1; else if ($1 == NULL) $$ = $3; else { $1->tail->next = $3; $1->tail = $3->tail; $$ = $1; } } ; xhost : not host { struct node_host *n; for (n = $2; n != NULL; n = n->next) n->not = $1; $$ = $2; } | not NOROUTE { $$ = calloc(1, sizeof(struct node_host)); if ($$ == NULL) err(1, "xhost: calloc"); $$->addr.type = PF_ADDR_NOROUTE; $$->next = NULL; $$->not = $1; $$->tail = $$; } | not URPFFAILED { $$ = calloc(1, sizeof(struct node_host)); if ($$ == NULL) err(1, "xhost: calloc"); $$->addr.type = PF_ADDR_URPFFAILED; $$->next = NULL; $$->not = $1; $$->tail = $$; } ; host : STRING { if (($$ = host($1)) == NULL) { /* error. "any" is handled elsewhere */ free($1); yyerror("could not parse host specification"); YYERROR; } free($1); } | STRING '-' STRING { struct node_host *b, *e; if ((b = host($1)) == NULL || (e = host($3)) == NULL) { free($1); free($3); yyerror("could not parse host specification"); YYERROR; } if (b->af != e->af || b->addr.type != PF_ADDR_ADDRMASK || e->addr.type != PF_ADDR_ADDRMASK || unmask(&b->addr.v.a.mask, b->af) != (b->af == AF_INET ? 32 : 128) || unmask(&e->addr.v.a.mask, e->af) != (e->af == AF_INET ? 32 : 128) || b->next != NULL || b->not || e->next != NULL || e->not) { free(b); free(e); free($1); free($3); yyerror("invalid address range"); YYERROR; } memcpy(&b->addr.v.a.mask, &e->addr.v.a.addr, sizeof(b->addr.v.a.mask)); b->addr.type = PF_ADDR_RANGE; $$ = b; free(e); free($1); free($3); } | STRING '/' NUMBER { char *buf; if (asprintf(&buf, "%s/%lld", $1, (long long)$3) == -1) err(1, "host: asprintf"); free($1); if (($$ = host(buf)) == NULL) { /* error. "any" is handled elsewhere */ free(buf); yyerror("could not parse host specification"); YYERROR; } free(buf); } | NUMBER '/' NUMBER { char *buf; /* ie. for 10/8 parsing */ #ifdef __FreeBSD__ if (asprintf(&buf, "%lld/%lld", (long long)$1, (long long)$3) == -1) #else if (asprintf(&buf, "%lld/%lld", $1, $3) == -1) #endif err(1, "host: asprintf"); if (($$ = host(buf)) == NULL) { /* error. "any" is handled elsewhere */ free(buf); yyerror("could not parse host specification"); YYERROR; } free(buf); } | dynaddr | dynaddr '/' NUMBER { struct node_host *n; if ($3 < 0 || $3 > 128) { yyerror("bit number too big"); YYERROR; } $$ = $1; for (n = $1; n != NULL; n = n->next) set_ipmask(n, $3); } | '<' STRING '>' { if (strlen($2) >= PF_TABLE_NAME_SIZE) { yyerror("table name '%s' too long", $2); free($2); YYERROR; } $$ = calloc(1, sizeof(struct node_host)); if ($$ == NULL) err(1, "host: calloc"); $$->addr.type = PF_ADDR_TABLE; if (strlcpy($$->addr.v.tblname, $2, sizeof($$->addr.v.tblname)) >= sizeof($$->addr.v.tblname)) errx(1, "host: strlcpy"); free($2); $$->next = NULL; $$->tail = $$; } ; number : NUMBER | STRING { u_long ulval; if (atoul($1, &ulval) == -1) { yyerror("%s is not a number", $1); free($1); YYERROR; } else $$ = ulval; free($1); } ; dynaddr : '(' STRING ')' { int flags = 0; char *p, *op; op = $2; if (!isalpha(op[0])) { yyerror("invalid interface name '%s'", op); free(op); YYERROR; } while ((p = strrchr($2, ':')) != NULL) { if (!strcmp(p+1, "network")) flags |= PFI_AFLAG_NETWORK; else if (!strcmp(p+1, "broadcast")) flags |= PFI_AFLAG_BROADCAST; else if (!strcmp(p+1, "peer")) flags |= PFI_AFLAG_PEER; else if (!strcmp(p+1, "0")) flags |= PFI_AFLAG_NOALIAS; else { yyerror("interface %s has bad modifier", $2); free(op); YYERROR; } *p = '\0'; } if (flags & (flags - 1) & PFI_AFLAG_MODEMASK) { free(op); yyerror("illegal combination of " "interface modifiers"); YYERROR; } $$ = calloc(1, sizeof(struct node_host)); if ($$ == NULL) err(1, "address: calloc"); $$->af = 0; set_ipmask($$, 128); $$->addr.type = PF_ADDR_DYNIFTL; $$->addr.iflags = flags; if (strlcpy($$->addr.v.ifname, $2, sizeof($$->addr.v.ifname)) >= sizeof($$->addr.v.ifname)) { free(op); free($$); yyerror("interface name too long"); YYERROR; } free(op); $$->next = NULL; $$->tail = $$; } ; portspec : port_item { $$ = $1; } | '{' optnl port_list '}' { $$ = $3; } ; port_list : port_item optnl { $$ = $1; } | port_list comma port_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; port_item : portrange { $$ = calloc(1, sizeof(struct node_port)); if ($$ == NULL) err(1, "port_item: calloc"); $$->port[0] = $1.a; $$->port[1] = $1.b; if ($1.t) $$->op = PF_OP_RRG; else $$->op = PF_OP_EQ; $$->next = NULL; $$->tail = $$; } | unaryop portrange { if ($2.t) { yyerror("':' cannot be used with an other " "port operator"); YYERROR; } $$ = calloc(1, sizeof(struct node_port)); if ($$ == NULL) err(1, "port_item: calloc"); $$->port[0] = $2.a; $$->port[1] = $2.b; $$->op = $1; $$->next = NULL; $$->tail = $$; } | portrange PORTBINARY portrange { if ($1.t || $3.t) { yyerror("':' cannot be used with an other " "port operator"); YYERROR; } $$ = calloc(1, sizeof(struct node_port)); if ($$ == NULL) err(1, "port_item: calloc"); $$->port[0] = $1.a; $$->port[1] = $3.a; $$->op = $2; $$->next = NULL; $$->tail = $$; } ; portplain : numberstring { if (parseport($1, &$$, 0) == -1) { free($1); YYERROR; } free($1); } ; portrange : numberstring { if (parseport($1, &$$, PPORT_RANGE) == -1) { free($1); YYERROR; } free($1); } ; uids : uid_item { $$ = $1; } | '{' optnl uid_list '}' { $$ = $3; } ; uid_list : uid_item optnl { $$ = $1; } | uid_list comma uid_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; uid_item : uid { $$ = calloc(1, sizeof(struct node_uid)); if ($$ == NULL) err(1, "uid_item: calloc"); $$->uid[0] = $1; $$->uid[1] = $1; $$->op = PF_OP_EQ; $$->next = NULL; $$->tail = $$; } | unaryop uid { if ($2 == UID_MAX && $1 != PF_OP_EQ && $1 != PF_OP_NE) { yyerror("user unknown requires operator = or " "!="); YYERROR; } $$ = calloc(1, sizeof(struct node_uid)); if ($$ == NULL) err(1, "uid_item: calloc"); $$->uid[0] = $2; $$->uid[1] = $2; $$->op = $1; $$->next = NULL; $$->tail = $$; } | uid PORTBINARY uid { if ($1 == UID_MAX || $3 == UID_MAX) { yyerror("user unknown requires operator = or " "!="); YYERROR; } $$ = calloc(1, sizeof(struct node_uid)); if ($$ == NULL) err(1, "uid_item: calloc"); $$->uid[0] = $1; $$->uid[1] = $3; $$->op = $2; $$->next = NULL; $$->tail = $$; } ; uid : STRING { if (!strcmp($1, "unknown")) $$ = UID_MAX; else { struct passwd *pw; if ((pw = getpwnam($1)) == NULL) { yyerror("unknown user %s", $1); free($1); YYERROR; } $$ = pw->pw_uid; } free($1); } | NUMBER { if ($1 < 0 || $1 >= UID_MAX) { yyerror("illegal uid value %lu", $1); YYERROR; } $$ = $1; } ; gids : gid_item { $$ = $1; } | '{' optnl gid_list '}' { $$ = $3; } ; gid_list : gid_item optnl { $$ = $1; } | gid_list comma gid_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; gid_item : gid { $$ = calloc(1, sizeof(struct node_gid)); if ($$ == NULL) err(1, "gid_item: calloc"); $$->gid[0] = $1; $$->gid[1] = $1; $$->op = PF_OP_EQ; $$->next = NULL; $$->tail = $$; } | unaryop gid { if ($2 == GID_MAX && $1 != PF_OP_EQ && $1 != PF_OP_NE) { yyerror("group unknown requires operator = or " "!="); YYERROR; } $$ = calloc(1, sizeof(struct node_gid)); if ($$ == NULL) err(1, "gid_item: calloc"); $$->gid[0] = $2; $$->gid[1] = $2; $$->op = $1; $$->next = NULL; $$->tail = $$; } | gid PORTBINARY gid { if ($1 == GID_MAX || $3 == GID_MAX) { yyerror("group unknown requires operator = or " "!="); YYERROR; } $$ = calloc(1, sizeof(struct node_gid)); if ($$ == NULL) err(1, "gid_item: calloc"); $$->gid[0] = $1; $$->gid[1] = $3; $$->op = $2; $$->next = NULL; $$->tail = $$; } ; gid : STRING { if (!strcmp($1, "unknown")) $$ = GID_MAX; else { struct group *grp; if ((grp = getgrnam($1)) == NULL) { yyerror("unknown group %s", $1); free($1); YYERROR; } $$ = grp->gr_gid; } free($1); } | NUMBER { if ($1 < 0 || $1 >= GID_MAX) { yyerror("illegal gid value %lu", $1); YYERROR; } $$ = $1; } ; flag : STRING { int f; if ((f = parse_flags($1)) < 0) { yyerror("bad flags %s", $1); free($1); YYERROR; } free($1); $$.b1 = f; } ; flags : FLAGS flag '/' flag { $$.b1 = $2.b1; $$.b2 = $4.b1; } | FLAGS '/' flag { $$.b1 = 0; $$.b2 = $3.b1; } | FLAGS ANY { $$.b1 = 0; $$.b2 = 0; } ; icmpspec : ICMPTYPE icmp_item { $$ = $2; } | ICMPTYPE '{' optnl icmp_list '}' { $$ = $4; } | ICMP6TYPE icmp6_item { $$ = $2; } | ICMP6TYPE '{' optnl icmp6_list '}' { $$ = $4; } ; icmp_list : icmp_item optnl { $$ = $1; } | icmp_list comma icmp_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; icmp6_list : icmp6_item optnl { $$ = $1; } | icmp6_list comma icmp6_item optnl { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; icmp_item : icmptype { $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = 0; $$->proto = IPPROTO_ICMP; $$->next = NULL; $$->tail = $$; } | icmptype CODE STRING { const struct icmpcodeent *p; if ((p = geticmpcodebyname($1-1, $3, AF_INET)) == NULL) { yyerror("unknown icmp-code %s", $3); free($3); YYERROR; } free($3); $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = p->code + 1; $$->proto = IPPROTO_ICMP; $$->next = NULL; $$->tail = $$; } | icmptype CODE NUMBER { if ($3 < 0 || $3 > 255) { yyerror("illegal icmp-code %lu", $3); YYERROR; } $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = $3 + 1; $$->proto = IPPROTO_ICMP; $$->next = NULL; $$->tail = $$; } ; icmp6_item : icmp6type { $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = 0; $$->proto = IPPROTO_ICMPV6; $$->next = NULL; $$->tail = $$; } | icmp6type CODE STRING { const struct icmpcodeent *p; if ((p = geticmpcodebyname($1-1, $3, AF_INET6)) == NULL) { yyerror("unknown icmp6-code %s", $3); free($3); YYERROR; } free($3); $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = p->code + 1; $$->proto = IPPROTO_ICMPV6; $$->next = NULL; $$->tail = $$; } | icmp6type CODE NUMBER { if ($3 < 0 || $3 > 255) { yyerror("illegal icmp-code %lu", $3); YYERROR; } $$ = calloc(1, sizeof(struct node_icmp)); if ($$ == NULL) err(1, "icmp_item: calloc"); $$->type = $1; $$->code = $3 + 1; $$->proto = IPPROTO_ICMPV6; $$->next = NULL; $$->tail = $$; } ; icmptype : STRING { const struct icmptypeent *p; if ((p = geticmptypebyname($1, AF_INET)) == NULL) { yyerror("unknown icmp-type %s", $1); free($1); YYERROR; } $$ = p->type + 1; free($1); } | NUMBER { if ($1 < 0 || $1 > 255) { yyerror("illegal icmp-type %lu", $1); YYERROR; } $$ = $1 + 1; } ; icmp6type : STRING { const struct icmptypeent *p; if ((p = geticmptypebyname($1, AF_INET6)) == NULL) { yyerror("unknown icmp6-type %s", $1); free($1); YYERROR; } $$ = p->type + 1; free($1); } | NUMBER { if ($1 < 0 || $1 > 255) { yyerror("illegal icmp6-type %lu", $1); YYERROR; } $$ = $1 + 1; } ; tos : STRING { if (!strcmp($1, "lowdelay")) $$ = IPTOS_LOWDELAY; else if (!strcmp($1, "throughput")) $$ = IPTOS_THROUGHPUT; else if (!strcmp($1, "reliability")) $$ = IPTOS_RELIABILITY; else if ($1[0] == '0' && $1[1] == 'x') $$ = strtoul($1, NULL, 16); else $$ = 256; /* flag bad argument */ if ($$ < 0 || $$ > 255) { yyerror("illegal tos value %s", $1); free($1); YYERROR; } free($1); } | NUMBER { $$ = $1; if ($$ < 0 || $$ > 255) { yyerror("illegal tos value %s", $1); YYERROR; } } ; sourcetrack : SOURCETRACK { $$ = PF_SRCTRACK; } | SOURCETRACK GLOBAL { $$ = PF_SRCTRACK_GLOBAL; } | SOURCETRACK RULE { $$ = PF_SRCTRACK_RULE; } ; statelock : IFBOUND { $$ = PFRULE_IFBOUND; } | FLOATING { $$ = 0; } ; keep : NO STATE { $$.action = 0; $$.options = NULL; } | KEEP STATE state_opt_spec { $$.action = PF_STATE_NORMAL; $$.options = $3; } | MODULATE STATE state_opt_spec { $$.action = PF_STATE_MODULATE; $$.options = $3; } | SYNPROXY STATE state_opt_spec { $$.action = PF_STATE_SYNPROXY; $$.options = $3; } ; flush : /* empty */ { $$ = 0; } | FLUSH { $$ = PF_FLUSH; } | FLUSH GLOBAL { $$ = PF_FLUSH | PF_FLUSH_GLOBAL; } ; state_opt_spec : '(' state_opt_list ')' { $$ = $2; } | /* empty */ { $$ = NULL; } ; state_opt_list : state_opt_item { $$ = $1; } | state_opt_list comma state_opt_item { $1->tail->next = $3; $1->tail = $3; $$ = $1; } ; state_opt_item : MAXIMUM NUMBER { if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_MAX; $$->data.max_states = $2; $$->next = NULL; $$->tail = $$; } | NOSYNC { $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_NOSYNC; $$->next = NULL; $$->tail = $$; } | MAXSRCSTATES NUMBER { if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_MAX_SRC_STATES; $$->data.max_src_states = $2; $$->next = NULL; $$->tail = $$; } | MAXSRCCONN NUMBER { if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_MAX_SRC_CONN; $$->data.max_src_conn = $2; $$->next = NULL; $$->tail = $$; } | MAXSRCCONNRATE NUMBER '/' NUMBER { if ($2 < 0 || $2 > UINT_MAX || $4 < 0 || $4 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_MAX_SRC_CONN_RATE; $$->data.max_src_conn_rate.limit = $2; $$->data.max_src_conn_rate.seconds = $4; $$->next = NULL; $$->tail = $$; } | OVERLOAD '<' STRING '>' flush { if (strlen($3) >= PF_TABLE_NAME_SIZE) { yyerror("table name '%s' too long", $3); free($3); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); if (strlcpy($$->data.overload.tblname, $3, PF_TABLE_NAME_SIZE) >= PF_TABLE_NAME_SIZE) errx(1, "state_opt_item: strlcpy"); free($3); $$->type = PF_STATE_OPT_OVERLOAD; $$->data.overload.flush = $5; $$->next = NULL; $$->tail = $$; } | MAXSRCNODES NUMBER { if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_MAX_SRC_NODES; $$->data.max_src_nodes = $2; $$->next = NULL; $$->tail = $$; } | sourcetrack { $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_SRCTRACK; $$->data.src_track = $1; $$->next = NULL; $$->tail = $$; } | statelock { $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_STATELOCK; $$->data.statelock = $1; $$->next = NULL; $$->tail = $$; } | SLOPPY { $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_SLOPPY; $$->next = NULL; $$->tail = $$; } | STRING NUMBER { int i; if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } for (i = 0; pf_timeouts[i].name && strcmp(pf_timeouts[i].name, $1); ++i) ; /* nothing */ if (!pf_timeouts[i].name) { yyerror("illegal timeout name %s", $1); free($1); YYERROR; } if (strchr(pf_timeouts[i].name, '.') == NULL) { yyerror("illegal state timeout %s", $1); free($1); YYERROR; } free($1); $$ = calloc(1, sizeof(struct node_state_opt)); if ($$ == NULL) err(1, "state_opt_item: calloc"); $$->type = PF_STATE_OPT_TIMEOUT; $$->data.timeout.number = pf_timeouts[i].timeout; $$->data.timeout.seconds = $2; $$->next = NULL; $$->tail = $$; } ; label : LABEL STRING { $$ = $2; } ; qname : QUEUE STRING { $$.qname = $2; $$.pqname = NULL; } | QUEUE '(' STRING ')' { $$.qname = $3; $$.pqname = NULL; } | QUEUE '(' STRING comma STRING ')' { $$.qname = $3; $$.pqname = $5; } ; no : /* empty */ { $$ = 0; } | NO { $$ = 1; } ; portstar : numberstring { if (parseport($1, &$$, PPORT_RANGE|PPORT_STAR) == -1) { free($1); YYERROR; } free($1); } ; redirspec : host { $$ = $1; } | '{' optnl redir_host_list '}' { $$ = $3; } ; redir_host_list : host optnl { $$ = $1; } | redir_host_list comma host optnl { $1->tail->next = $3; $1->tail = $3->tail; $$ = $1; } ; redirpool : /* empty */ { $$ = NULL; } | ARROW redirspec { $$ = calloc(1, sizeof(struct redirection)); if ($$ == NULL) err(1, "redirection: calloc"); $$->host = $2; $$->rport.a = $$->rport.b = $$->rport.t = 0; } | ARROW redirspec PORT portstar { $$ = calloc(1, sizeof(struct redirection)); if ($$ == NULL) err(1, "redirection: calloc"); $$->host = $2; $$->rport = $4; } ; hashkey : /* empty */ { $$ = calloc(1, sizeof(struct pf_poolhashkey)); if ($$ == NULL) err(1, "hashkey: calloc"); $$->key32[0] = arc4random(); $$->key32[1] = arc4random(); $$->key32[2] = arc4random(); $$->key32[3] = arc4random(); } | string { if (!strncmp($1, "0x", 2)) { if (strlen($1) != 34) { free($1); yyerror("hex key must be 128 bits " "(32 hex digits) long"); YYERROR; } $$ = calloc(1, sizeof(struct pf_poolhashkey)); if ($$ == NULL) err(1, "hashkey: calloc"); if (sscanf($1, "0x%8x%8x%8x%8x", &$$->key32[0], &$$->key32[1], &$$->key32[2], &$$->key32[3]) != 4) { free($$); free($1); yyerror("invalid hex key"); YYERROR; } } else { MD5_CTX context; $$ = calloc(1, sizeof(struct pf_poolhashkey)); if ($$ == NULL) err(1, "hashkey: calloc"); MD5Init(&context); MD5Update(&context, (unsigned char *)$1, strlen($1)); MD5Final((unsigned char *)$$, &context); HTONL($$->key32[0]); HTONL($$->key32[1]); HTONL($$->key32[2]); HTONL($$->key32[3]); } free($1); } ; pool_opts : { bzero(&pool_opts, sizeof pool_opts); } pool_opts_l { $$ = pool_opts; } | /* empty */ { bzero(&pool_opts, sizeof pool_opts); $$ = pool_opts; } ; pool_opts_l : pool_opts_l pool_opt | pool_opt ; pool_opt : BITMASK { if (pool_opts.type) { yyerror("pool type cannot be redefined"); YYERROR; } pool_opts.type = PF_POOL_BITMASK; } | RANDOM { if (pool_opts.type) { yyerror("pool type cannot be redefined"); YYERROR; } pool_opts.type = PF_POOL_RANDOM; } | SOURCEHASH hashkey { if (pool_opts.type) { yyerror("pool type cannot be redefined"); YYERROR; } pool_opts.type = PF_POOL_SRCHASH; pool_opts.key = $2; } | ROUNDROBIN { if (pool_opts.type) { yyerror("pool type cannot be redefined"); YYERROR; } pool_opts.type = PF_POOL_ROUNDROBIN; } | STATICPORT { if (pool_opts.staticport) { yyerror("static-port cannot be redefined"); YYERROR; } pool_opts.staticport = 1; } | STICKYADDRESS { if (filter_opts.marker & POM_STICKYADDRESS) { yyerror("sticky-address cannot be redefined"); YYERROR; } pool_opts.marker |= POM_STICKYADDRESS; pool_opts.opts |= PF_POOL_STICKYADDR; } ; redirection : /* empty */ { $$ = NULL; } | ARROW host { $$ = calloc(1, sizeof(struct redirection)); if ($$ == NULL) err(1, "redirection: calloc"); $$->host = $2; $$->rport.a = $$->rport.b = $$->rport.t = 0; } | ARROW host PORT portstar { $$ = calloc(1, sizeof(struct redirection)); if ($$ == NULL) err(1, "redirection: calloc"); $$->host = $2; $$->rport = $4; } ; natpasslog : /* empty */ { $$.b1 = $$.b2 = 0; $$.w2 = 0; } | PASS { $$.b1 = 1; $$.b2 = 0; $$.w2 = 0; } | PASS log { $$.b1 = 1; $$.b2 = $2.log; $$.w2 = $2.logif; } | log { $$.b1 = 0; $$.b2 = $1.log; $$.w2 = $1.logif; } ; nataction : no NAT natpasslog { if ($1 && $3.b1) { yyerror("\"pass\" not valid with \"no\""); YYERROR; } if ($1) $$.b1 = PF_NONAT; else $$.b1 = PF_NAT; $$.b2 = $3.b1; $$.w = $3.b2; $$.w2 = $3.w2; } | no RDR natpasslog { if ($1 && $3.b1) { yyerror("\"pass\" not valid with \"no\""); YYERROR; } if ($1) $$.b1 = PF_NORDR; else $$.b1 = PF_RDR; $$.b2 = $3.b1; $$.w = $3.b2; $$.w2 = $3.w2; } ; natrule : nataction interface af proto fromto tag tagged rtable redirpool pool_opts { struct pf_rule r; if (check_rulestate(PFCTL_STATE_NAT)) YYERROR; memset(&r, 0, sizeof(r)); r.action = $1.b1; r.natpass = $1.b2; r.log = $1.w; r.logif = $1.w2; r.af = $3; if (!r.af) { if ($5.src.host && $5.src.host->af && !$5.src.host->ifindex) r.af = $5.src.host->af; else if ($5.dst.host && $5.dst.host->af && !$5.dst.host->ifindex) r.af = $5.dst.host->af; } if ($6 != NULL) if (strlcpy(r.tagname, $6, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } if ($7.name) if (strlcpy(r.match_tagname, $7.name, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } r.match_tag_not = $7.neg; r.rtableid = $8; if (r.action == PF_NONAT || r.action == PF_NORDR) { if ($9 != NULL) { yyerror("translation rule with 'no' " "does not need '->'"); YYERROR; } } else { if ($9 == NULL || $9->host == NULL) { yyerror("translation rule requires '-> " "address'"); YYERROR; } if (!r.af && ! $9->host->ifindex) r.af = $9->host->af; remove_invalid_hosts(&$9->host, &r.af); if (invalid_redirect($9->host, r.af)) YYERROR; if (check_netmask($9->host, r.af)) YYERROR; r.rpool.proxy_port[0] = ntohs($9->rport.a); switch (r.action) { case PF_RDR: if (!$9->rport.b && $9->rport.t && $5.dst.port != NULL) { r.rpool.proxy_port[1] = ntohs($9->rport.a) + (ntohs( $5.dst.port->port[1]) - ntohs( $5.dst.port->port[0])); } else r.rpool.proxy_port[1] = ntohs($9->rport.b); break; case PF_NAT: r.rpool.proxy_port[1] = ntohs($9->rport.b); if (!r.rpool.proxy_port[0] && !r.rpool.proxy_port[1]) { r.rpool.proxy_port[0] = PF_NAT_PROXY_PORT_LOW; r.rpool.proxy_port[1] = PF_NAT_PROXY_PORT_HIGH; } else if (!r.rpool.proxy_port[1]) r.rpool.proxy_port[1] = r.rpool.proxy_port[0]; break; default: break; } r.rpool.opts = $10.type; if ((r.rpool.opts & PF_POOL_TYPEMASK) == PF_POOL_NONE && ($9->host->next != NULL || $9->host->addr.type == PF_ADDR_TABLE || DYNIF_MULTIADDR($9->host->addr))) r.rpool.opts = PF_POOL_ROUNDROBIN; if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN && disallow_table($9->host, "tables are only " "supported in round-robin redirection " "pools")) YYERROR; if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN && disallow_alias($9->host, "interface (%s) " "is only supported in round-robin " "redirection pools")) YYERROR; if ($9->host->next != NULL) { if ((r.rpool.opts & PF_POOL_TYPEMASK) != PF_POOL_ROUNDROBIN) { yyerror("only round-robin " "valid for multiple " "redirection addresses"); YYERROR; } } } if ($10.key != NULL) memcpy(&r.rpool.key, $10.key, sizeof(struct pf_poolhashkey)); if ($10.opts) r.rpool.opts |= $10.opts; if ($10.staticport) { if (r.action != PF_NAT) { yyerror("the 'static-port' option is " "only valid with nat rules"); YYERROR; } if (r.rpool.proxy_port[0] != PF_NAT_PROXY_PORT_LOW && r.rpool.proxy_port[1] != PF_NAT_PROXY_PORT_HIGH) { yyerror("the 'static-port' option can't" " be used when specifying a port" " range"); YYERROR; } r.rpool.proxy_port[0] = 0; r.rpool.proxy_port[1] = 0; } expand_rule(&r, $2, $9 == NULL ? NULL : $9->host, $4, $5.src_os, $5.src.host, $5.src.port, $5.dst.host, $5.dst.port, 0, 0, 0, ""); free($9); } ; binatrule : no BINAT natpasslog interface af proto FROM host toipspec tag tagged rtable redirection { struct pf_rule binat; struct pf_pooladdr *pa; if (check_rulestate(PFCTL_STATE_NAT)) YYERROR; if (disallow_urpf_failed($9, "\"urpf-failed\" is not " "permitted as a binat destination")) YYERROR; memset(&binat, 0, sizeof(binat)); if ($1 && $3.b1) { yyerror("\"pass\" not valid with \"no\""); YYERROR; } if ($1) binat.action = PF_NOBINAT; else binat.action = PF_BINAT; binat.natpass = $3.b1; binat.log = $3.b2; binat.logif = $3.w2; binat.af = $5; if (!binat.af && $8 != NULL && $8->af) binat.af = $8->af; if (!binat.af && $9 != NULL && $9->af) binat.af = $9->af; if (!binat.af && $13 != NULL && $13->host) binat.af = $13->host->af; if (!binat.af) { yyerror("address family (inet/inet6) " "undefined"); YYERROR; } if ($4 != NULL) { memcpy(binat.ifname, $4->ifname, sizeof(binat.ifname)); binat.ifnot = $4->not; free($4); } if ($10 != NULL) if (strlcpy(binat.tagname, $10, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } if ($11.name) if (strlcpy(binat.match_tagname, $11.name, PF_TAG_NAME_SIZE) >= PF_TAG_NAME_SIZE) { yyerror("tag too long, max %u chars", PF_TAG_NAME_SIZE - 1); YYERROR; } binat.match_tag_not = $11.neg; binat.rtableid = $12; if ($6 != NULL) { binat.proto = $6->proto; free($6); } if ($8 != NULL && disallow_table($8, "invalid use of " "table <%s> as the source address of a binat rule")) YYERROR; if ($8 != NULL && disallow_alias($8, "invalid use of " "interface (%s) as the source address of a binat " "rule")) YYERROR; if ($13 != NULL && $13->host != NULL && disallow_table( $13->host, "invalid use of table <%s> as the " "redirect address of a binat rule")) YYERROR; if ($13 != NULL && $13->host != NULL && disallow_alias( $13->host, "invalid use of interface (%s) as the " "redirect address of a binat rule")) YYERROR; if ($8 != NULL) { if ($8->next) { yyerror("multiple binat ip addresses"); YYERROR; } if ($8->addr.type == PF_ADDR_DYNIFTL) $8->af = binat.af; if ($8->af != binat.af) { yyerror("binat ip versions must match"); YYERROR; } if (check_netmask($8, binat.af)) YYERROR; memcpy(&binat.src.addr, &$8->addr, sizeof(binat.src.addr)); free($8); } if ($9 != NULL) { if ($9->next) { yyerror("multiple binat ip addresses"); YYERROR; } if ($9->af != binat.af && $9->af) { yyerror("binat ip versions must match"); YYERROR; } if (check_netmask($9, binat.af)) YYERROR; memcpy(&binat.dst.addr, &$9->addr, sizeof(binat.dst.addr)); binat.dst.neg = $9->not; free($9); } if (binat.action == PF_NOBINAT) { if ($13 != NULL) { yyerror("'no binat' rule does not need" " '->'"); YYERROR; } } else { if ($13 == NULL || $13->host == NULL) { yyerror("'binat' rule requires" " '-> address'"); YYERROR; } remove_invalid_hosts(&$13->host, &binat.af); if (invalid_redirect($13->host, binat.af)) YYERROR; if ($13->host->next != NULL) { yyerror("binat rule must redirect to " "a single address"); YYERROR; } if (check_netmask($13->host, binat.af)) YYERROR; if (!PF_AZERO(&binat.src.addr.v.a.mask, binat.af) && !PF_AEQ(&binat.src.addr.v.a.mask, &$13->host->addr.v.a.mask, binat.af)) { yyerror("'binat' source mask and " "redirect mask must be the same"); YYERROR; } TAILQ_INIT(&binat.rpool.list); pa = calloc(1, sizeof(struct pf_pooladdr)); if (pa == NULL) err(1, "binat: calloc"); pa->addr = $13->host->addr; pa->ifname[0] = 0; TAILQ_INSERT_TAIL(&binat.rpool.list, pa, entries); free($13); } pfctl_add_rule(pf, &binat, ""); } ; tag : /* empty */ { $$ = NULL; } | TAG STRING { $$ = $2; } ; tagged : /* empty */ { $$.neg = 0; $$.name = NULL; } | not TAGGED string { $$.neg = $1; $$.name = $3; } ; rtable : /* empty */ { $$ = -1; } | RTABLE NUMBER { if ($2 < 0 || $2 > rt_tableid_max()) { yyerror("invalid rtable id"); YYERROR; } $$ = $2; } ; route_host : STRING { $$ = calloc(1, sizeof(struct node_host)); if ($$ == NULL) err(1, "route_host: calloc"); $$->ifname = $1; set_ipmask($$, 128); $$->next = NULL; $$->tail = $$; } | '(' STRING host ')' { $$ = $3; $$->ifname = $2; } ; route_host_list : route_host optnl { $$ = $1; } | route_host_list comma route_host optnl { if ($1->af == 0) $1->af = $3->af; if ($1->af != $3->af) { yyerror("all pool addresses must be in the " "same address family"); YYERROR; } $1->tail->next = $3; $1->tail = $3->tail; $$ = $1; } ; routespec : route_host { $$ = $1; } | '{' optnl route_host_list '}' { $$ = $3; } ; route : /* empty */ { $$.host = NULL; $$.rt = 0; $$.pool_opts = 0; } | FASTROUTE { + /* backwards-compat */ $$.host = NULL; - $$.rt = PF_FASTROUTE; + $$.rt = 0; $$.pool_opts = 0; } | ROUTETO routespec pool_opts { $$.host = $2; $$.rt = PF_ROUTETO; $$.pool_opts = $3.type | $3.opts; if ($3.key != NULL) $$.key = $3.key; } | REPLYTO routespec pool_opts { $$.host = $2; $$.rt = PF_REPLYTO; $$.pool_opts = $3.type | $3.opts; if ($3.key != NULL) $$.key = $3.key; } | DUPTO routespec pool_opts { $$.host = $2; $$.rt = PF_DUPTO; $$.pool_opts = $3.type | $3.opts; if ($3.key != NULL) $$.key = $3.key; } ; timeout_spec : STRING NUMBER { if (check_rulestate(PFCTL_STATE_OPTION)) { free($1); YYERROR; } if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } if (pfctl_set_timeout(pf, $1, $2, 0) != 0) { yyerror("unknown timeout %s", $1); free($1); YYERROR; } free($1); } | INTERVAL NUMBER { if (check_rulestate(PFCTL_STATE_OPTION)) YYERROR; if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } if (pfctl_set_timeout(pf, "interval", $2, 0) != 0) YYERROR; } ; timeout_list : timeout_list comma timeout_spec optnl | timeout_spec optnl ; limit_spec : STRING NUMBER { if (check_rulestate(PFCTL_STATE_OPTION)) { free($1); YYERROR; } if ($2 < 0 || $2 > UINT_MAX) { yyerror("only positive values permitted"); YYERROR; } if (pfctl_set_limit(pf, $1, $2) != 0) { yyerror("unable to set limit %s %u", $1, $2); free($1); YYERROR; } free($1); } ; limit_list : limit_list comma limit_spec optnl | limit_spec optnl ; comma : ',' | /* empty */ ; yesno : NO { $$ = 0; } | STRING { if (!strcmp($1, "yes")) $$ = 1; else { yyerror("invalid value '%s', expected 'yes' " "or 'no'", $1); free($1); YYERROR; } free($1); } ; unaryop : '=' { $$ = PF_OP_EQ; } | '!' '=' { $$ = PF_OP_NE; } | '<' '=' { $$ = PF_OP_LE; } | '<' { $$ = PF_OP_LT; } | '>' '=' { $$ = PF_OP_GE; } | '>' { $$ = PF_OP_GT; } ; %% int yyerror(const char *fmt, ...) { va_list ap; file->errors++; va_start(ap, fmt); fprintf(stderr, "%s:%d: ", file->name, yylval.lineno); vfprintf(stderr, fmt, ap); fprintf(stderr, "\n"); va_end(ap); return (0); } int disallow_table(struct node_host *h, const char *fmt) { for (; h != NULL; h = h->next) if (h->addr.type == PF_ADDR_TABLE) { yyerror(fmt, h->addr.v.tblname); return (1); } return (0); } int disallow_urpf_failed(struct node_host *h, const char *fmt) { for (; h != NULL; h = h->next) if (h->addr.type == PF_ADDR_URPFFAILED) { yyerror(fmt); return (1); } return (0); } int disallow_alias(struct node_host *h, const char *fmt) { for (; h != NULL; h = h->next) if (DYNIF_MULTIADDR(h->addr)) { yyerror(fmt, h->addr.v.tblname); return (1); } return (0); } int rule_consistent(struct pf_rule *r, int anchor_call) { int problems = 0; switch (r->action) { case PF_PASS: case PF_DROP: case PF_SCRUB: case PF_NOSCRUB: problems = filter_consistent(r, anchor_call); break; case PF_NAT: case PF_NONAT: problems = nat_consistent(r); break; case PF_RDR: case PF_NORDR: problems = rdr_consistent(r); break; case PF_BINAT: case PF_NOBINAT: default: break; } return (problems); } int filter_consistent(struct pf_rule *r, int anchor_call) { int problems = 0; if (r->proto != IPPROTO_TCP && r->proto != IPPROTO_UDP && (r->src.port_op || r->dst.port_op)) { yyerror("port only applies to tcp/udp"); problems++; } if (r->proto != IPPROTO_ICMP && r->proto != IPPROTO_ICMPV6 && (r->type || r->code)) { yyerror("icmp-type/code only applies to icmp"); problems++; } if (!r->af && (r->type || r->code)) { yyerror("must indicate address family with icmp-type/code"); problems++; } if (r->overload_tblname[0] && r->max_src_conn == 0 && r->max_src_conn_rate.seconds == 0) { yyerror("'overload' requires 'max-src-conn' " "or 'max-src-conn-rate'"); problems++; } if ((r->proto == IPPROTO_ICMP && r->af == AF_INET6) || (r->proto == IPPROTO_ICMPV6 && r->af == AF_INET)) { yyerror("proto %s doesn't match address family %s", r->proto == IPPROTO_ICMP ? "icmp" : "icmp6", r->af == AF_INET ? "inet" : "inet6"); problems++; } if (r->allow_opts && r->action != PF_PASS) { yyerror("allow-opts can only be specified for pass rules"); problems++; } if (r->rule_flag & PFRULE_FRAGMENT && (r->src.port_op || r->dst.port_op || r->flagset || r->type || r->code)) { yyerror("fragments can be filtered only on IP header fields"); problems++; } if (r->rule_flag & PFRULE_RETURNRST && r->proto != IPPROTO_TCP) { yyerror("return-rst can only be applied to TCP rules"); problems++; } if (r->max_src_nodes && !(r->rule_flag & PFRULE_RULESRCTRACK)) { yyerror("max-src-nodes requires 'source-track rule'"); problems++; } if (r->action == PF_DROP && r->keep_state) { yyerror("keep state on block rules doesn't make sense"); problems++; } if (r->rule_flag & PFRULE_STATESLOPPY && (r->keep_state == PF_STATE_MODULATE || r->keep_state == PF_STATE_SYNPROXY)) { yyerror("sloppy state matching cannot be used with " "synproxy state or modulate state"); problems++; } return (-problems); } int nat_consistent(struct pf_rule *r) { return (0); /* yeah! */ } int rdr_consistent(struct pf_rule *r) { int problems = 0; if (r->proto != IPPROTO_TCP && r->proto != IPPROTO_UDP) { if (r->src.port_op) { yyerror("src port only applies to tcp/udp"); problems++; } if (r->dst.port_op) { yyerror("dst port only applies to tcp/udp"); problems++; } if (r->rpool.proxy_port[0]) { yyerror("rpool port only applies to tcp/udp"); problems++; } } if (r->dst.port_op && r->dst.port_op != PF_OP_EQ && r->dst.port_op != PF_OP_RRG) { yyerror("invalid port operator for rdr destination port"); problems++; } return (-problems); } int process_tabledef(char *name, struct table_opts *opts) { struct pfr_buffer ab; struct node_tinit *ti; bzero(&ab, sizeof(ab)); ab.pfrb_type = PFRB_ADDRS; SIMPLEQ_FOREACH(ti, &opts->init_nodes, entries) { if (ti->file) if (pfr_buf_load(&ab, ti->file, 0, append_addr)) { if (errno) yyerror("cannot load \"%s\": %s", ti->file, strerror(errno)); else yyerror("file \"%s\" contains bad data", ti->file); goto _error; } if (ti->host) if (append_addr_host(&ab, ti->host, 0, 0)) { yyerror("cannot create address buffer: %s", strerror(errno)); goto _error; } } if (pf->opts & PF_OPT_VERBOSE) print_tabledef(name, opts->flags, opts->init_addr, &opts->init_nodes); if (!(pf->opts & PF_OPT_NOACTION) && pfctl_define_table(name, opts->flags, opts->init_addr, pf->anchor->name, &ab, pf->anchor->ruleset.tticket)) { yyerror("cannot define table %s: %s", name, pfr_strerror(errno)); goto _error; } pf->tdirty = 1; pfr_buf_clear(&ab); return (0); _error: pfr_buf_clear(&ab); return (-1); } struct keywords { const char *k_name; int k_val; }; /* macro gore, but you should've seen the prior indentation nightmare... */ #define FREE_LIST(T,r) \ do { \ T *p, *node = r; \ while (node != NULL) { \ p = node; \ node = node->next; \ free(p); \ } \ } while (0) #define LOOP_THROUGH(T,n,r,C) \ do { \ T *n; \ if (r == NULL) { \ r = calloc(1, sizeof(T)); \ if (r == NULL) \ err(1, "LOOP: calloc"); \ r->next = NULL; \ } \ n = r; \ while (n != NULL) { \ do { \ C; \ } while (0); \ n = n->next; \ } \ } while (0) void expand_label_str(char *label, size_t len, const char *srch, const char *repl) { char *tmp; char *p, *q; if ((tmp = calloc(1, len)) == NULL) err(1, "expand_label_str: calloc"); p = q = label; while ((q = strstr(p, srch)) != NULL) { *q = '\0'; if ((strlcat(tmp, p, len) >= len) || (strlcat(tmp, repl, len) >= len)) errx(1, "expand_label: label too long"); q += strlen(srch); p = q; } if (strlcat(tmp, p, len) >= len) errx(1, "expand_label: label too long"); strlcpy(label, tmp, len); /* always fits */ free(tmp); } void expand_label_if(const char *name, char *label, size_t len, const char *ifname) { if (strstr(label, name) != NULL) { if (!*ifname) expand_label_str(label, len, name, "any"); else expand_label_str(label, len, name, ifname); } } void expand_label_addr(const char *name, char *label, size_t len, sa_family_t af, struct node_host *h) { char tmp[64], tmp_not[66]; if (strstr(label, name) != NULL) { switch (h->addr.type) { case PF_ADDR_DYNIFTL: snprintf(tmp, sizeof(tmp), "(%s)", h->addr.v.ifname); break; case PF_ADDR_TABLE: snprintf(tmp, sizeof(tmp), "<%s>", h->addr.v.tblname); break; case PF_ADDR_NOROUTE: snprintf(tmp, sizeof(tmp), "no-route"); break; case PF_ADDR_URPFFAILED: snprintf(tmp, sizeof(tmp), "urpf-failed"); break; case PF_ADDR_ADDRMASK: if (!af || (PF_AZERO(&h->addr.v.a.addr, af) && PF_AZERO(&h->addr.v.a.mask, af))) snprintf(tmp, sizeof(tmp), "any"); else { char a[48]; int bits; if (inet_ntop(af, &h->addr.v.a.addr, a, sizeof(a)) == NULL) snprintf(tmp, sizeof(tmp), "?"); else { bits = unmask(&h->addr.v.a.mask, af); if ((af == AF_INET && bits < 32) || (af == AF_INET6 && bits < 128)) snprintf(tmp, sizeof(tmp), "%s/%d", a, bits); else snprintf(tmp, sizeof(tmp), "%s", a); } } break; default: snprintf(tmp, sizeof(tmp), "?"); break; } if (h->not) { snprintf(tmp_not, sizeof(tmp_not), "! %s", tmp); expand_label_str(label, len, name, tmp_not); } else expand_label_str(label, len, name, tmp); } } void expand_label_port(const char *name, char *label, size_t len, struct node_port *port) { char a1[6], a2[6], op[13] = ""; if (strstr(label, name) != NULL) { snprintf(a1, sizeof(a1), "%u", ntohs(port->port[0])); snprintf(a2, sizeof(a2), "%u", ntohs(port->port[1])); if (!port->op) ; else if (port->op == PF_OP_IRG) snprintf(op, sizeof(op), "%s><%s", a1, a2); else if (port->op == PF_OP_XRG) snprintf(op, sizeof(op), "%s<>%s", a1, a2); else if (port->op == PF_OP_EQ) snprintf(op, sizeof(op), "%s", a1); else if (port->op == PF_OP_NE) snprintf(op, sizeof(op), "!=%s", a1); else if (port->op == PF_OP_LT) snprintf(op, sizeof(op), "<%s", a1); else if (port->op == PF_OP_LE) snprintf(op, sizeof(op), "<=%s", a1); else if (port->op == PF_OP_GT) snprintf(op, sizeof(op), ">%s", a1); else if (port->op == PF_OP_GE) snprintf(op, sizeof(op), ">=%s", a1); expand_label_str(label, len, name, op); } } void expand_label_proto(const char *name, char *label, size_t len, u_int8_t proto) { struct protoent *pe; char n[4]; if (strstr(label, name) != NULL) { pe = getprotobynumber(proto); if (pe != NULL) expand_label_str(label, len, name, pe->p_name); else { snprintf(n, sizeof(n), "%u", proto); expand_label_str(label, len, name, n); } } } void expand_label_nr(const char *name, char *label, size_t len) { char n[11]; if (strstr(label, name) != NULL) { snprintf(n, sizeof(n), "%u", pf->anchor->match); expand_label_str(label, len, name, n); } } void expand_label(char *label, size_t len, const char *ifname, sa_family_t af, struct node_host *src_host, struct node_port *src_port, struct node_host *dst_host, struct node_port *dst_port, u_int8_t proto) { expand_label_if("$if", label, len, ifname); expand_label_addr("$srcaddr", label, len, af, src_host); expand_label_addr("$dstaddr", label, len, af, dst_host); expand_label_port("$srcport", label, len, src_port); expand_label_port("$dstport", label, len, dst_port); expand_label_proto("$proto", label, len, proto); expand_label_nr("$nr", label, len); } int expand_altq(struct pf_altq *a, struct node_if *interfaces, struct node_queue *nqueues, struct node_queue_bw bwspec, struct node_queue_opt *opts) { struct pf_altq pa, pb; char qname[PF_QNAME_SIZE]; struct node_queue *n; struct node_queue_bw bw; int errs = 0; if ((pf->loadopt & PFCTL_FLAG_ALTQ) == 0) { FREE_LIST(struct node_if, interfaces); if (nqueues) FREE_LIST(struct node_queue, nqueues); return (0); } LOOP_THROUGH(struct node_if, interface, interfaces, memcpy(&pa, a, sizeof(struct pf_altq)); if (strlcpy(pa.ifname, interface->ifname, sizeof(pa.ifname)) >= sizeof(pa.ifname)) errx(1, "expand_altq: strlcpy"); if (interface->not) { yyerror("altq on ! is not supported"); errs++; } else { if (eval_pfaltq(pf, &pa, &bwspec, opts)) errs++; else if (pfctl_add_altq(pf, &pa)) errs++; if (pf->opts & PF_OPT_VERBOSE) { print_altq(&pf->paltq->altq, 0, &bwspec, opts); if (nqueues && nqueues->tail) { printf("queue { "); LOOP_THROUGH(struct node_queue, queue, nqueues, printf("%s ", queue->queue); ); printf("}"); } printf("\n"); } if (pa.scheduler == ALTQT_CBQ || pa.scheduler == ALTQT_HFSC) { /* now create a root queue */ memset(&pb, 0, sizeof(struct pf_altq)); if (strlcpy(qname, "root_", sizeof(qname)) >= sizeof(qname)) errx(1, "expand_altq: strlcpy"); if (strlcat(qname, interface->ifname, sizeof(qname)) >= sizeof(qname)) errx(1, "expand_altq: strlcat"); if (strlcpy(pb.qname, qname, sizeof(pb.qname)) >= sizeof(pb.qname)) errx(1, "expand_altq: strlcpy"); if (strlcpy(pb.ifname, interface->ifname, sizeof(pb.ifname)) >= sizeof(pb.ifname)) errx(1, "expand_altq: strlcpy"); pb.qlimit = pa.qlimit; pb.scheduler = pa.scheduler; bw.bw_absolute = pa.ifbandwidth; bw.bw_percent = 0; if (eval_pfqueue(pf, &pb, &bw, opts)) errs++; else if (pfctl_add_altq(pf, &pb)) errs++; } LOOP_THROUGH(struct node_queue, queue, nqueues, n = calloc(1, sizeof(struct node_queue)); if (n == NULL) err(1, "expand_altq: calloc"); if (pa.scheduler == ALTQT_CBQ || pa.scheduler == ALTQT_HFSC) if (strlcpy(n->parent, qname, sizeof(n->parent)) >= sizeof(n->parent)) errx(1, "expand_altq: strlcpy"); if (strlcpy(n->queue, queue->queue, sizeof(n->queue)) >= sizeof(n->queue)) errx(1, "expand_altq: strlcpy"); if (strlcpy(n->ifname, interface->ifname, sizeof(n->ifname)) >= sizeof(n->ifname)) errx(1, "expand_altq: strlcpy"); n->scheduler = pa.scheduler; n->next = NULL; n->tail = n; if (queues == NULL) queues = n; else { queues->tail->next = n; queues->tail = n; } ); } ); FREE_LIST(struct node_if, interfaces); if (nqueues) FREE_LIST(struct node_queue, nqueues); return (errs); } int expand_queue(struct pf_altq *a, struct node_if *interfaces, struct node_queue *nqueues, struct node_queue_bw bwspec, struct node_queue_opt *opts) { struct node_queue *n, *nq; struct pf_altq pa; u_int8_t found = 0; u_int8_t errs = 0; if ((pf->loadopt & PFCTL_FLAG_ALTQ) == 0) { FREE_LIST(struct node_queue, nqueues); return (0); } if (queues == NULL) { yyerror("queue %s has no parent", a->qname); FREE_LIST(struct node_queue, nqueues); return (1); } LOOP_THROUGH(struct node_if, interface, interfaces, LOOP_THROUGH(struct node_queue, tqueue, queues, if (!strncmp(a->qname, tqueue->queue, PF_QNAME_SIZE) && (interface->ifname[0] == 0 || (!interface->not && !strncmp(interface->ifname, tqueue->ifname, IFNAMSIZ)) || (interface->not && strncmp(interface->ifname, tqueue->ifname, IFNAMSIZ)))) { /* found ourself in queues */ found++; memcpy(&pa, a, sizeof(struct pf_altq)); if (pa.scheduler != ALTQT_NONE && pa.scheduler != tqueue->scheduler) { yyerror("exactly one scheduler type " "per interface allowed"); return (1); } pa.scheduler = tqueue->scheduler; /* scheduler dependent error checking */ switch (pa.scheduler) { case ALTQT_PRIQ: if (nqueues != NULL) { yyerror("priq queues cannot " "have child queues"); return (1); } if (bwspec.bw_absolute > 0 || bwspec.bw_percent < 100) { yyerror("priq doesn't take " "bandwidth"); return (1); } break; default: break; } if (strlcpy(pa.ifname, tqueue->ifname, sizeof(pa.ifname)) >= sizeof(pa.ifname)) errx(1, "expand_queue: strlcpy"); if (strlcpy(pa.parent, tqueue->parent, sizeof(pa.parent)) >= sizeof(pa.parent)) errx(1, "expand_queue: strlcpy"); if (eval_pfqueue(pf, &pa, &bwspec, opts)) errs++; else if (pfctl_add_altq(pf, &pa)) errs++; for (nq = nqueues; nq != NULL; nq = nq->next) { if (!strcmp(a->qname, nq->queue)) { yyerror("queue cannot have " "itself as child"); errs++; continue; } n = calloc(1, sizeof(struct node_queue)); if (n == NULL) err(1, "expand_queue: calloc"); if (strlcpy(n->parent, a->qname, sizeof(n->parent)) >= sizeof(n->parent)) errx(1, "expand_queue strlcpy"); if (strlcpy(n->queue, nq->queue, sizeof(n->queue)) >= sizeof(n->queue)) errx(1, "expand_queue strlcpy"); if (strlcpy(n->ifname, tqueue->ifname, sizeof(n->ifname)) >= sizeof(n->ifname)) errx(1, "expand_queue strlcpy"); n->scheduler = tqueue->scheduler; n->next = NULL; n->tail = n; if (queues == NULL) queues = n; else { queues->tail->next = n; queues->tail = n; } } if ((pf->opts & PF_OPT_VERBOSE) && ( (found == 1 && interface->ifname[0] == 0) || (found > 0 && interface->ifname[0] != 0))) { print_queue(&pf->paltq->altq, 0, &bwspec, interface->ifname[0] != 0, opts); if (nqueues && nqueues->tail) { printf("{ "); LOOP_THROUGH(struct node_queue, queue, nqueues, printf("%s ", queue->queue); ); printf("}"); } printf("\n"); } } ); ); FREE_LIST(struct node_queue, nqueues); FREE_LIST(struct node_if, interfaces); if (!found) { yyerror("queue %s has no parent", a->qname); errs++; } if (errs) return (1); else return (0); } void expand_rule(struct pf_rule *r, struct node_if *interfaces, struct node_host *rpool_hosts, struct node_proto *protos, struct node_os *src_oses, struct node_host *src_hosts, struct node_port *src_ports, struct node_host *dst_hosts, struct node_port *dst_ports, struct node_uid *uids, struct node_gid *gids, struct node_icmp *icmp_types, const char *anchor_call) { sa_family_t af = r->af; int added = 0, error = 0; char ifname[IF_NAMESIZE]; char label[PF_RULE_LABEL_SIZE]; char tagname[PF_TAG_NAME_SIZE]; char match_tagname[PF_TAG_NAME_SIZE]; struct pf_pooladdr *pa; struct node_host *h; u_int8_t flags, flagset, keep_state; if (strlcpy(label, r->label, sizeof(label)) >= sizeof(label)) errx(1, "expand_rule: strlcpy"); if (strlcpy(tagname, r->tagname, sizeof(tagname)) >= sizeof(tagname)) errx(1, "expand_rule: strlcpy"); if (strlcpy(match_tagname, r->match_tagname, sizeof(match_tagname)) >= sizeof(match_tagname)) errx(1, "expand_rule: strlcpy"); flags = r->flags; flagset = r->flagset; keep_state = r->keep_state; LOOP_THROUGH(struct node_if, interface, interfaces, LOOP_THROUGH(struct node_proto, proto, protos, LOOP_THROUGH(struct node_icmp, icmp_type, icmp_types, LOOP_THROUGH(struct node_host, src_host, src_hosts, LOOP_THROUGH(struct node_port, src_port, src_ports, LOOP_THROUGH(struct node_os, src_os, src_oses, LOOP_THROUGH(struct node_host, dst_host, dst_hosts, LOOP_THROUGH(struct node_port, dst_port, dst_ports, LOOP_THROUGH(struct node_uid, uid, uids, LOOP_THROUGH(struct node_gid, gid, gids, r->af = af; /* for link-local IPv6 address, interface must match up */ if ((r->af && src_host->af && r->af != src_host->af) || (r->af && dst_host->af && r->af != dst_host->af) || (src_host->af && dst_host->af && src_host->af != dst_host->af) || (src_host->ifindex && dst_host->ifindex && src_host->ifindex != dst_host->ifindex) || (src_host->ifindex && *interface->ifname && src_host->ifindex != if_nametoindex(interface->ifname)) || (dst_host->ifindex && *interface->ifname && dst_host->ifindex != if_nametoindex(interface->ifname))) continue; if (!r->af && src_host->af) r->af = src_host->af; else if (!r->af && dst_host->af) r->af = dst_host->af; if (*interface->ifname) strlcpy(r->ifname, interface->ifname, sizeof(r->ifname)); else if (if_indextoname(src_host->ifindex, ifname)) strlcpy(r->ifname, ifname, sizeof(r->ifname)); else if (if_indextoname(dst_host->ifindex, ifname)) strlcpy(r->ifname, ifname, sizeof(r->ifname)); else memset(r->ifname, '\0', sizeof(r->ifname)); if (strlcpy(r->label, label, sizeof(r->label)) >= sizeof(r->label)) errx(1, "expand_rule: strlcpy"); if (strlcpy(r->tagname, tagname, sizeof(r->tagname)) >= sizeof(r->tagname)) errx(1, "expand_rule: strlcpy"); if (strlcpy(r->match_tagname, match_tagname, sizeof(r->match_tagname)) >= sizeof(r->match_tagname)) errx(1, "expand_rule: strlcpy"); expand_label(r->label, PF_RULE_LABEL_SIZE, r->ifname, r->af, src_host, src_port, dst_host, dst_port, proto->proto); expand_label(r->tagname, PF_TAG_NAME_SIZE, r->ifname, r->af, src_host, src_port, dst_host, dst_port, proto->proto); expand_label(r->match_tagname, PF_TAG_NAME_SIZE, r->ifname, r->af, src_host, src_port, dst_host, dst_port, proto->proto); error += check_netmask(src_host, r->af); error += check_netmask(dst_host, r->af); r->ifnot = interface->not; r->proto = proto->proto; r->src.addr = src_host->addr; r->src.neg = src_host->not; r->src.port[0] = src_port->port[0]; r->src.port[1] = src_port->port[1]; r->src.port_op = src_port->op; r->dst.addr = dst_host->addr; r->dst.neg = dst_host->not; r->dst.port[0] = dst_port->port[0]; r->dst.port[1] = dst_port->port[1]; r->dst.port_op = dst_port->op; r->uid.op = uid->op; r->uid.uid[0] = uid->uid[0]; r->uid.uid[1] = uid->uid[1]; r->gid.op = gid->op; r->gid.gid[0] = gid->gid[0]; r->gid.gid[1] = gid->gid[1]; r->type = icmp_type->type; r->code = icmp_type->code; if ((keep_state == PF_STATE_MODULATE || keep_state == PF_STATE_SYNPROXY) && r->proto && r->proto != IPPROTO_TCP) r->keep_state = PF_STATE_NORMAL; else r->keep_state = keep_state; if (r->proto && r->proto != IPPROTO_TCP) { r->flags = 0; r->flagset = 0; } else { r->flags = flags; r->flagset = flagset; } if (icmp_type->proto && r->proto != icmp_type->proto) { yyerror("icmp-type mismatch"); error++; } if (src_os && src_os->os) { r->os_fingerprint = pfctl_get_fingerprint(src_os->os); if ((pf->opts & PF_OPT_VERBOSE2) && r->os_fingerprint == PF_OSFP_NOMATCH) fprintf(stderr, "warning: unknown '%s' OS fingerprint\n", src_os->os); } else { r->os_fingerprint = PF_OSFP_ANY; } TAILQ_INIT(&r->rpool.list); for (h = rpool_hosts; h != NULL; h = h->next) { pa = calloc(1, sizeof(struct pf_pooladdr)); if (pa == NULL) err(1, "expand_rule: calloc"); pa->addr = h->addr; if (h->ifname != NULL) { if (strlcpy(pa->ifname, h->ifname, sizeof(pa->ifname)) >= sizeof(pa->ifname)) errx(1, "expand_rule: strlcpy"); } else pa->ifname[0] = 0; TAILQ_INSERT_TAIL(&r->rpool.list, pa, entries); } if (rule_consistent(r, anchor_call[0]) < 0 || error) yyerror("skipping rule due to errors"); else { r->nr = pf->astack[pf->asd]->match++; pfctl_add_rule(pf, r, anchor_call); added++; } )))))))))); FREE_LIST(struct node_if, interfaces); FREE_LIST(struct node_proto, protos); FREE_LIST(struct node_host, src_hosts); FREE_LIST(struct node_port, src_ports); FREE_LIST(struct node_os, src_oses); FREE_LIST(struct node_host, dst_hosts); FREE_LIST(struct node_port, dst_ports); FREE_LIST(struct node_uid, uids); FREE_LIST(struct node_gid, gids); FREE_LIST(struct node_icmp, icmp_types); FREE_LIST(struct node_host, rpool_hosts); if (!added) yyerror("rule expands to no valid combination"); } int expand_skip_interface(struct node_if *interfaces) { int errs = 0; if (!interfaces || (!interfaces->next && !interfaces->not && !strcmp(interfaces->ifname, "none"))) { if (pf->opts & PF_OPT_VERBOSE) printf("set skip on none\n"); errs = pfctl_set_interface_flags(pf, "", PFI_IFLAG_SKIP, 0); return (errs); } if (pf->opts & PF_OPT_VERBOSE) printf("set skip on {"); LOOP_THROUGH(struct node_if, interface, interfaces, if (pf->opts & PF_OPT_VERBOSE) printf(" %s", interface->ifname); if (interface->not) { yyerror("skip on ! is not supported"); errs++; } else errs += pfctl_set_interface_flags(pf, interface->ifname, PFI_IFLAG_SKIP, 1); ); if (pf->opts & PF_OPT_VERBOSE) printf(" }\n"); FREE_LIST(struct node_if, interfaces); if (errs) return (1); else return (0); } #undef FREE_LIST #undef LOOP_THROUGH int check_rulestate(int desired_state) { if (require_order && (rulestate > desired_state)) { yyerror("Rules must be in order: options, normalization, " "queueing, translation, filtering"); return (1); } rulestate = desired_state; return (0); } int kw_cmp(const void *k, const void *e) { return (strcmp(k, ((const struct keywords *)e)->k_name)); } int lookup(char *s) { /* this has to be sorted always */ static const struct keywords keywords[] = { { "all", ALL}, { "allow-opts", ALLOWOPTS}, { "altq", ALTQ}, { "anchor", ANCHOR}, { "antispoof", ANTISPOOF}, { "any", ANY}, { "bandwidth", BANDWIDTH}, { "binat", BINAT}, { "binat-anchor", BINATANCHOR}, { "bitmask", BITMASK}, { "block", BLOCK}, { "block-policy", BLOCKPOLICY}, { "buckets", BUCKETS}, { "cbq", CBQ}, { "code", CODE}, { "codelq", CODEL}, { "crop", FRAGCROP}, { "debug", DEBUG}, { "divert-reply", DIVERTREPLY}, { "divert-to", DIVERTTO}, { "drop", DROP}, { "drop-ovl", FRAGDROP}, { "dup-to", DUPTO}, { "fairq", FAIRQ}, { "fastroute", FASTROUTE}, { "file", FILENAME}, { "fingerprints", FINGERPRINTS}, { "flags", FLAGS}, { "floating", FLOATING}, { "flush", FLUSH}, { "for", FOR}, { "fragment", FRAGMENT}, { "from", FROM}, { "global", GLOBAL}, { "group", GROUP}, { "hfsc", HFSC}, { "hogs", HOGS}, { "hostid", HOSTID}, { "icmp-type", ICMPTYPE}, { "icmp6-type", ICMP6TYPE}, { "if-bound", IFBOUND}, { "in", IN}, { "include", INCLUDE}, { "inet", INET}, { "inet6", INET6}, { "interval", INTERVAL}, { "keep", KEEP}, { "label", LABEL}, { "limit", LIMIT}, { "linkshare", LINKSHARE}, { "load", LOAD}, { "log", LOG}, { "loginterface", LOGINTERFACE}, { "max", MAXIMUM}, { "max-mss", MAXMSS}, { "max-src-conn", MAXSRCCONN}, { "max-src-conn-rate", MAXSRCCONNRATE}, { "max-src-nodes", MAXSRCNODES}, { "max-src-states", MAXSRCSTATES}, { "min-ttl", MINTTL}, { "modulate", MODULATE}, { "nat", NAT}, { "nat-anchor", NATANCHOR}, { "no", NO}, { "no-df", NODF}, { "no-route", NOROUTE}, { "no-sync", NOSYNC}, { "on", ON}, { "optimization", OPTIMIZATION}, { "os", OS}, { "out", OUT}, { "overload", OVERLOAD}, { "pass", PASS}, { "port", PORT}, { "prio", PRIO}, { "priority", PRIORITY}, { "priq", PRIQ}, { "probability", PROBABILITY}, { "proto", PROTO}, { "qlimit", QLIMIT}, { "queue", QUEUE}, { "quick", QUICK}, { "random", RANDOM}, { "random-id", RANDOMID}, { "rdr", RDR}, { "rdr-anchor", RDRANCHOR}, { "realtime", REALTIME}, { "reassemble", REASSEMBLE}, { "reply-to", REPLYTO}, { "require-order", REQUIREORDER}, { "return", RETURN}, { "return-icmp", RETURNICMP}, { "return-icmp6", RETURNICMP6}, { "return-rst", RETURNRST}, { "round-robin", ROUNDROBIN}, { "route", ROUTE}, { "route-to", ROUTETO}, { "rtable", RTABLE}, { "rule", RULE}, { "ruleset-optimization", RULESET_OPTIMIZATION}, { "scrub", SCRUB}, { "set", SET}, { "set-tos", SETTOS}, { "skip", SKIP}, { "sloppy", SLOPPY}, { "source-hash", SOURCEHASH}, { "source-track", SOURCETRACK}, { "state", STATE}, { "state-defaults", STATEDEFAULTS}, { "state-policy", STATEPOLICY}, { "static-port", STATICPORT}, { "sticky-address", STICKYADDRESS}, { "synproxy", SYNPROXY}, { "table", TABLE}, { "tag", TAG}, { "tagged", TAGGED}, { "target", TARGET}, { "tbrsize", TBRSIZE}, { "timeout", TIMEOUT}, { "to", TO}, { "tos", TOS}, { "ttl", TTL}, { "upperlimit", UPPERLIMIT}, { "urpf-failed", URPFFAILED}, { "user", USER}, }; const struct keywords *p; p = bsearch(s, keywords, sizeof(keywords)/sizeof(keywords[0]), sizeof(keywords[0]), kw_cmp); if (p) { if (debug > 1) fprintf(stderr, "%s: %d\n", s, p->k_val); return (p->k_val); } else { if (debug > 1) fprintf(stderr, "string: %s\n", s); return (STRING); } } #define MAXPUSHBACK 128 static char *parsebuf; static int parseindex; static char pushback_buffer[MAXPUSHBACK]; static int pushback_index = 0; int lgetc(int quotec) { int c, next; if (parsebuf) { /* Read character from the parsebuffer instead of input. */ if (parseindex >= 0) { c = parsebuf[parseindex++]; if (c != '\0') return (c); parsebuf = NULL; } else parseindex++; } if (pushback_index) return (pushback_buffer[--pushback_index]); if (quotec) { if ((c = getc(file->stream)) == EOF) { yyerror("reached end of file while parsing quoted string"); if (popfile() == EOF) return (EOF); return (quotec); } return (c); } while ((c = getc(file->stream)) == '\\') { next = getc(file->stream); if (next != '\n') { c = next; break; } yylval.lineno = file->lineno; file->lineno++; } while (c == EOF) { if (popfile() == EOF) return (EOF); c = getc(file->stream); } return (c); } int lungetc(int c) { if (c == EOF) return (EOF); if (parsebuf) { parseindex--; if (parseindex >= 0) return (c); } if (pushback_index < MAXPUSHBACK-1) return (pushback_buffer[pushback_index++] = c); else return (EOF); } int findeol(void) { int c; parsebuf = NULL; /* skip to either EOF or the first real EOL */ while (1) { if (pushback_index) c = pushback_buffer[--pushback_index]; else c = lgetc(0); if (c == '\n') { file->lineno++; break; } if (c == EOF) break; } return (ERROR); } int yylex(void) { char buf[8096]; char *p, *val; int quotec, next, c; int token; top: p = buf; while ((c = lgetc(0)) == ' ' || c == '\t') ; /* nothing */ yylval.lineno = file->lineno; if (c == '#') while ((c = lgetc(0)) != '\n' && c != EOF) ; /* nothing */ if (c == '$' && parsebuf == NULL) { while (1) { if ((c = lgetc(0)) == EOF) return (0); if (p + 1 >= buf + sizeof(buf) - 1) { yyerror("string too long"); return (findeol()); } if (isalnum(c) || c == '_') { *p++ = (char)c; continue; } *p = '\0'; lungetc(c); break; } val = symget(buf); if (val == NULL) { yyerror("macro '%s' not defined", buf); return (findeol()); } parsebuf = val; parseindex = 0; goto top; } switch (c) { case '\'': case '"': quotec = c; while (1) { if ((c = lgetc(quotec)) == EOF) return (0); if (c == '\n') { file->lineno++; continue; } else if (c == '\\') { if ((next = lgetc(quotec)) == EOF) return (0); if (next == quotec || c == ' ' || c == '\t') c = next; else if (next == '\n') continue; else lungetc(next); } else if (c == quotec) { *p = '\0'; break; } if (p + 1 >= buf + sizeof(buf) - 1) { yyerror("string too long"); return (findeol()); } *p++ = (char)c; } yylval.v.string = strdup(buf); if (yylval.v.string == NULL) err(1, "yylex: strdup"); return (STRING); case '<': next = lgetc(0); if (next == '>') { yylval.v.i = PF_OP_XRG; return (PORTBINARY); } lungetc(next); break; case '>': next = lgetc(0); if (next == '<') { yylval.v.i = PF_OP_IRG; return (PORTBINARY); } lungetc(next); break; case '-': next = lgetc(0); if (next == '>') return (ARROW); lungetc(next); break; } #define allowed_to_end_number(x) \ (isspace(x) || x == ')' || x ==',' || x == '/' || x == '}' || x == '=') if (c == '-' || isdigit(c)) { do { *p++ = c; if ((unsigned)(p-buf) >= sizeof(buf)) { yyerror("string too long"); return (findeol()); } } while ((c = lgetc(0)) != EOF && isdigit(c)); lungetc(c); if (p == buf + 1 && buf[0] == '-') goto nodigits; if (c == EOF || allowed_to_end_number(c)) { const char *errstr = NULL; *p = '\0'; yylval.v.number = strtonum(buf, LLONG_MIN, LLONG_MAX, &errstr); if (errstr) { yyerror("\"%s\" invalid number: %s", buf, errstr); return (findeol()); } return (NUMBER); } else { nodigits: while (p > buf + 1) lungetc(*--p); c = *--p; if (c == '-') return (c); } } #define allowed_in_string(x) \ (isalnum(x) || (ispunct(x) && x != '(' && x != ')' && \ x != '{' && x != '}' && x != '<' && x != '>' && \ x != '!' && x != '=' && x != '/' && x != '#' && \ x != ',')) if (isalnum(c) || c == ':' || c == '_') { do { *p++ = c; if ((unsigned)(p-buf) >= sizeof(buf)) { yyerror("string too long"); return (findeol()); } } while ((c = lgetc(0)) != EOF && (allowed_in_string(c))); lungetc(c); *p = '\0'; if ((token = lookup(buf)) == STRING) if ((yylval.v.string = strdup(buf)) == NULL) err(1, "yylex: strdup"); return (token); } if (c == '\n') { yylval.lineno = file->lineno; file->lineno++; } if (c == EOF) return (0); return (c); } int check_file_secrecy(int fd, const char *fname) { struct stat st; if (fstat(fd, &st)) { warn("cannot stat %s", fname); return (-1); } if (st.st_uid != 0 && st.st_uid != getuid()) { warnx("%s: owner not root or current user", fname); return (-1); } if (st.st_mode & (S_IRWXG | S_IRWXO)) { warnx("%s: group/world readable/writeable", fname); return (-1); } return (0); } struct file * pushfile(const char *name, int secret) { struct file *nfile; if ((nfile = calloc(1, sizeof(struct file))) == NULL || (nfile->name = strdup(name)) == NULL) { warn("malloc"); return (NULL); } if (TAILQ_FIRST(&files) == NULL && strcmp(nfile->name, "-") == 0) { nfile->stream = stdin; free(nfile->name); if ((nfile->name = strdup("stdin")) == NULL) { warn("strdup"); free(nfile); return (NULL); } } else if ((nfile->stream = fopen(nfile->name, "r")) == NULL) { warn("%s", nfile->name); free(nfile->name); free(nfile); return (NULL); } else if (secret && check_file_secrecy(fileno(nfile->stream), nfile->name)) { fclose(nfile->stream); free(nfile->name); free(nfile); return (NULL); } nfile->lineno = 1; TAILQ_INSERT_TAIL(&files, nfile, entry); return (nfile); } int popfile(void) { struct file *prev; if ((prev = TAILQ_PREV(file, files, entry)) != NULL) { prev->errors += file->errors; TAILQ_REMOVE(&files, file, entry); fclose(file->stream); free(file->name); free(file); file = prev; return (0); } return (EOF); } int parse_config(char *filename, struct pfctl *xpf) { int errors = 0; struct sym *sym; pf = xpf; errors = 0; rulestate = PFCTL_STATE_NONE; returnicmpdefault = (ICMP_UNREACH << 8) | ICMP_UNREACH_PORT; returnicmp6default = (ICMP6_DST_UNREACH << 8) | ICMP6_DST_UNREACH_NOPORT; blockpolicy = PFRULE_DROP; require_order = 1; if ((file = pushfile(filename, 0)) == NULL) { warn("cannot open the main config file!"); return (-1); } yyparse(); errors = file->errors; popfile(); /* Free macros and check which have not been used. */ while ((sym = TAILQ_FIRST(&symhead))) { if ((pf->opts & PF_OPT_VERBOSE2) && !sym->used) fprintf(stderr, "warning: macro '%s' not " "used\n", sym->nam); free(sym->nam); free(sym->val); TAILQ_REMOVE(&symhead, sym, entry); free(sym); } return (errors ? -1 : 0); } int symset(const char *nam, const char *val, int persist) { struct sym *sym; for (sym = TAILQ_FIRST(&symhead); sym && strcmp(nam, sym->nam); sym = TAILQ_NEXT(sym, entry)) ; /* nothing */ if (sym != NULL) { if (sym->persist == 1) return (0); else { free(sym->nam); free(sym->val); TAILQ_REMOVE(&symhead, sym, entry); free(sym); } } if ((sym = calloc(1, sizeof(*sym))) == NULL) return (-1); sym->nam = strdup(nam); if (sym->nam == NULL) { free(sym); return (-1); } sym->val = strdup(val); if (sym->val == NULL) { free(sym->nam); free(sym); return (-1); } sym->used = 0; sym->persist = persist; TAILQ_INSERT_TAIL(&symhead, sym, entry); return (0); } int pfctl_cmdline_symset(char *s) { char *sym, *val; int ret; if ((val = strrchr(s, '=')) == NULL) return (-1); if ((sym = malloc(strlen(s) - strlen(val) + 1)) == NULL) err(1, "pfctl_cmdline_symset: malloc"); strlcpy(sym, s, strlen(s) - strlen(val) + 1); ret = symset(sym, val + 1, 1); free(sym); return (ret); } char * symget(const char *nam) { struct sym *sym; TAILQ_FOREACH(sym, &symhead, entry) if (strcmp(nam, sym->nam) == 0) { sym->used = 1; return (sym->val); } return (NULL); } void mv_rules(struct pf_ruleset *src, struct pf_ruleset *dst) { int i; struct pf_rule *r; for (i = 0; i < PF_RULESET_MAX; ++i) { while ((r = TAILQ_FIRST(src->rules[i].active.ptr)) != NULL) { TAILQ_REMOVE(src->rules[i].active.ptr, r, entries); TAILQ_INSERT_TAIL(dst->rules[i].active.ptr, r, entries); dst->anchor->match++; } src->anchor->match = 0; while ((r = TAILQ_FIRST(src->rules[i].inactive.ptr)) != NULL) { TAILQ_REMOVE(src->rules[i].inactive.ptr, r, entries); TAILQ_INSERT_TAIL(dst->rules[i].inactive.ptr, r, entries); } } } void decide_address_family(struct node_host *n, sa_family_t *af) { if (*af != 0 || n == NULL) return; *af = n->af; while ((n = n->next) != NULL) { if (n->af != *af) { *af = 0; return; } } } void remove_invalid_hosts(struct node_host **nh, sa_family_t *af) { struct node_host *n = *nh, *prev = NULL; while (n != NULL) { if (*af && n->af && n->af != *af) { /* unlink and free n */ struct node_host *next = n->next; /* adjust tail pointer */ if (n == (*nh)->tail) (*nh)->tail = prev; /* adjust previous node's next pointer */ if (prev == NULL) *nh = next; else prev->next = next; /* free node */ if (n->ifname != NULL) free(n->ifname); free(n); n = next; } else { if (n->af && !*af) *af = n->af; prev = n; n = n->next; } } } int invalid_redirect(struct node_host *nh, sa_family_t af) { if (!af) { struct node_host *n; /* tables and dyniftl are ok without an address family */ for (n = nh; n != NULL; n = n->next) { if (n->addr.type != PF_ADDR_TABLE && n->addr.type != PF_ADDR_DYNIFTL) { yyerror("address family not given and " "translation address expands to multiple " "address families"); return (1); } } } if (nh == NULL) { yyerror("no translation address with matching address family " "found."); return (1); } return (0); } int atoul(char *s, u_long *ulvalp) { u_long ulval; char *ep; errno = 0; ulval = strtoul(s, &ep, 0); if (s[0] == '\0' || *ep != '\0') return (-1); if (errno == ERANGE && ulval == ULONG_MAX) return (-1); *ulvalp = ulval; return (0); } int getservice(char *n) { struct servent *s; u_long ulval; if (atoul(n, &ulval) == 0) { if (ulval > 65535) { yyerror("illegal port value %lu", ulval); return (-1); } return (htons(ulval)); } else { s = getservbyname(n, "tcp"); if (s == NULL) s = getservbyname(n, "udp"); if (s == NULL) { yyerror("unknown port %s", n); return (-1); } return (s->s_port); } } int rule_label(struct pf_rule *r, char *s) { if (s) { if (strlcpy(r->label, s, sizeof(r->label)) >= sizeof(r->label)) { yyerror("rule label too long (max %d chars)", sizeof(r->label)-1); return (-1); } } return (0); } u_int16_t parseicmpspec(char *w, sa_family_t af) { const struct icmpcodeent *p; u_long ulval; u_int8_t icmptype; if (af == AF_INET) icmptype = returnicmpdefault >> 8; else icmptype = returnicmp6default >> 8; if (atoul(w, &ulval) == -1) { if ((p = geticmpcodebyname(icmptype, w, af)) == NULL) { yyerror("unknown icmp code %s", w); return (0); } ulval = p->code; } if (ulval > 255) { yyerror("invalid icmp code %lu", ulval); return (0); } return (icmptype << 8 | ulval); } int parseport(char *port, struct range *r, int extensions) { char *p = strchr(port, ':'); if (p == NULL) { if ((r->a = getservice(port)) == -1) return (-1); r->b = 0; r->t = PF_OP_NONE; return (0); } if ((extensions & PPORT_STAR) && !strcmp(p+1, "*")) { *p = 0; if ((r->a = getservice(port)) == -1) return (-1); r->b = 0; r->t = PF_OP_IRG; return (0); } if ((extensions & PPORT_RANGE)) { *p++ = 0; if ((r->a = getservice(port)) == -1 || (r->b = getservice(p)) == -1) return (-1); if (r->a == r->b) { r->b = 0; r->t = PF_OP_NONE; } else r->t = PF_OP_RRG; return (0); } return (-1); } int pfctl_load_anchors(int dev, struct pfctl *pf, struct pfr_buffer *trans) { struct loadanchors *la; TAILQ_FOREACH(la, &loadanchorshead, entries) { if (pf->opts & PF_OPT_VERBOSE) fprintf(stderr, "\nLoading anchor %s from %s\n", la->anchorname, la->filename); if (pfctl_rules(dev, la->filename, pf->opts, pf->optimize, la->anchorname, trans) == -1) return (-1); } return (0); } int rt_tableid_max(void) { #ifdef __FreeBSD__ int fibs; size_t l = sizeof(fibs); if (sysctlbyname("net.fibs", &fibs, &l, NULL, 0) == -1) fibs = 16; /* XXX RT_MAXFIBS, at least limit it some. */ /* * As the OpenBSD code only compares > and not >= we need to adjust * here given we only accept values of 0..n and want to avoid #ifdefs * in the grammar. */ return (fibs - 1); #else return (RT_TABLEID_MAX); #endif } Index: head/sbin/pfctl/pfctl_parser.c =================================================================== --- head/sbin/pfctl/pfctl_parser.c (revision 306683) +++ head/sbin/pfctl/pfctl_parser.c (revision 306684) @@ -1,1768 +1,1764 @@ /* $OpenBSD: pfctl_parser.c,v 1.240 2008/06/10 20:55:02 mcbride Exp $ */ /* * Copyright (c) 2001 Daniel Hartmeier * Copyright (c) 2002,2003 Henning Brauer * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * - Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * - 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. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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 * COPYRIGHT HOLDERS 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 __FBSDID("$FreeBSD$"); #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include "pfctl_parser.h" #include "pfctl.h" void print_op (u_int8_t, const char *, const char *); void print_port (u_int8_t, u_int16_t, u_int16_t, const char *, int); void print_ugid (u_int8_t, unsigned, unsigned, const char *, unsigned); void print_flags (u_int8_t); void print_fromto(struct pf_rule_addr *, pf_osfp_t, struct pf_rule_addr *, u_int8_t, u_int8_t, int, int); int ifa_skip_if(const char *filter, struct node_host *p); struct node_host *ifa_grouplookup(const char *, int); struct node_host *host_if(const char *, int); struct node_host *host_v4(const char *, int); struct node_host *host_v6(const char *, int); struct node_host *host_dns(const char *, int, int); const char * const tcpflags = "FSRPAUEW"; static const struct icmptypeent icmp_type[] = { { "echoreq", ICMP_ECHO }, { "echorep", ICMP_ECHOREPLY }, { "unreach", ICMP_UNREACH }, { "squench", ICMP_SOURCEQUENCH }, { "redir", ICMP_REDIRECT }, { "althost", ICMP_ALTHOSTADDR }, { "routeradv", ICMP_ROUTERADVERT }, { "routersol", ICMP_ROUTERSOLICIT }, { "timex", ICMP_TIMXCEED }, { "paramprob", ICMP_PARAMPROB }, { "timereq", ICMP_TSTAMP }, { "timerep", ICMP_TSTAMPREPLY }, { "inforeq", ICMP_IREQ }, { "inforep", ICMP_IREQREPLY }, { "maskreq", ICMP_MASKREQ }, { "maskrep", ICMP_MASKREPLY }, { "trace", ICMP_TRACEROUTE }, { "dataconv", ICMP_DATACONVERR }, { "mobredir", ICMP_MOBILE_REDIRECT }, { "ipv6-where", ICMP_IPV6_WHEREAREYOU }, { "ipv6-here", ICMP_IPV6_IAMHERE }, { "mobregreq", ICMP_MOBILE_REGREQUEST }, { "mobregrep", ICMP_MOBILE_REGREPLY }, { "skip", ICMP_SKIP }, { "photuris", ICMP_PHOTURIS } }; static const struct icmptypeent icmp6_type[] = { { "unreach", ICMP6_DST_UNREACH }, { "toobig", ICMP6_PACKET_TOO_BIG }, { "timex", ICMP6_TIME_EXCEEDED }, { "paramprob", ICMP6_PARAM_PROB }, { "echoreq", ICMP6_ECHO_REQUEST }, { "echorep", ICMP6_ECHO_REPLY }, { "groupqry", ICMP6_MEMBERSHIP_QUERY }, { "listqry", MLD_LISTENER_QUERY }, { "grouprep", ICMP6_MEMBERSHIP_REPORT }, { "listenrep", MLD_LISTENER_REPORT }, { "groupterm", ICMP6_MEMBERSHIP_REDUCTION }, { "listendone", MLD_LISTENER_DONE }, { "routersol", ND_ROUTER_SOLICIT }, { "routeradv", ND_ROUTER_ADVERT }, { "neighbrsol", ND_NEIGHBOR_SOLICIT }, { "neighbradv", ND_NEIGHBOR_ADVERT }, { "redir", ND_REDIRECT }, { "routrrenum", ICMP6_ROUTER_RENUMBERING }, { "wrureq", ICMP6_WRUREQUEST }, { "wrurep", ICMP6_WRUREPLY }, { "fqdnreq", ICMP6_FQDN_QUERY }, { "fqdnrep", ICMP6_FQDN_REPLY }, { "niqry", ICMP6_NI_QUERY }, { "nirep", ICMP6_NI_REPLY }, { "mtraceresp", MLD_MTRACE_RESP }, { "mtrace", MLD_MTRACE } }; static const struct icmpcodeent icmp_code[] = { { "net-unr", ICMP_UNREACH, ICMP_UNREACH_NET }, { "host-unr", ICMP_UNREACH, ICMP_UNREACH_HOST }, { "proto-unr", ICMP_UNREACH, ICMP_UNREACH_PROTOCOL }, { "port-unr", ICMP_UNREACH, ICMP_UNREACH_PORT }, { "needfrag", ICMP_UNREACH, ICMP_UNREACH_NEEDFRAG }, { "srcfail", ICMP_UNREACH, ICMP_UNREACH_SRCFAIL }, { "net-unk", ICMP_UNREACH, ICMP_UNREACH_NET_UNKNOWN }, { "host-unk", ICMP_UNREACH, ICMP_UNREACH_HOST_UNKNOWN }, { "isolate", ICMP_UNREACH, ICMP_UNREACH_ISOLATED }, { "net-prohib", ICMP_UNREACH, ICMP_UNREACH_NET_PROHIB }, { "host-prohib", ICMP_UNREACH, ICMP_UNREACH_HOST_PROHIB }, { "net-tos", ICMP_UNREACH, ICMP_UNREACH_TOSNET }, { "host-tos", ICMP_UNREACH, ICMP_UNREACH_TOSHOST }, { "filter-prohib", ICMP_UNREACH, ICMP_UNREACH_FILTER_PROHIB }, { "host-preced", ICMP_UNREACH, ICMP_UNREACH_HOST_PRECEDENCE }, { "cutoff-preced", ICMP_UNREACH, ICMP_UNREACH_PRECEDENCE_CUTOFF }, { "redir-net", ICMP_REDIRECT, ICMP_REDIRECT_NET }, { "redir-host", ICMP_REDIRECT, ICMP_REDIRECT_HOST }, { "redir-tos-net", ICMP_REDIRECT, ICMP_REDIRECT_TOSNET }, { "redir-tos-host", ICMP_REDIRECT, ICMP_REDIRECT_TOSHOST }, { "normal-adv", ICMP_ROUTERADVERT, ICMP_ROUTERADVERT_NORMAL }, { "common-adv", ICMP_ROUTERADVERT, ICMP_ROUTERADVERT_NOROUTE_COMMON }, { "transit", ICMP_TIMXCEED, ICMP_TIMXCEED_INTRANS }, { "reassemb", ICMP_TIMXCEED, ICMP_TIMXCEED_REASS }, { "badhead", ICMP_PARAMPROB, ICMP_PARAMPROB_ERRATPTR }, { "optmiss", ICMP_PARAMPROB, ICMP_PARAMPROB_OPTABSENT }, { "badlen", ICMP_PARAMPROB, ICMP_PARAMPROB_LENGTH }, { "unknown-ind", ICMP_PHOTURIS, ICMP_PHOTURIS_UNKNOWN_INDEX }, { "auth-fail", ICMP_PHOTURIS, ICMP_PHOTURIS_AUTH_FAILED }, { "decrypt-fail", ICMP_PHOTURIS, ICMP_PHOTURIS_DECRYPT_FAILED } }; static const struct icmpcodeent icmp6_code[] = { { "admin-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADMIN }, { "noroute-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_NOROUTE }, { "notnbr-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_NOTNEIGHBOR }, { "beyond-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_BEYONDSCOPE }, { "addr-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_ADDR }, { "port-unr", ICMP6_DST_UNREACH, ICMP6_DST_UNREACH_NOPORT }, { "transit", ICMP6_TIME_EXCEEDED, ICMP6_TIME_EXCEED_TRANSIT }, { "reassemb", ICMP6_TIME_EXCEEDED, ICMP6_TIME_EXCEED_REASSEMBLY }, { "badhead", ICMP6_PARAM_PROB, ICMP6_PARAMPROB_HEADER }, { "nxthdr", ICMP6_PARAM_PROB, ICMP6_PARAMPROB_NEXTHEADER }, { "redironlink", ND_REDIRECT, ND_REDIRECT_ONLINK }, { "redirrouter", ND_REDIRECT, ND_REDIRECT_ROUTER } }; const struct pf_timeout pf_timeouts[] = { { "tcp.first", PFTM_TCP_FIRST_PACKET }, { "tcp.opening", PFTM_TCP_OPENING }, { "tcp.established", PFTM_TCP_ESTABLISHED }, { "tcp.closing", PFTM_TCP_CLOSING }, { "tcp.finwait", PFTM_TCP_FIN_WAIT }, { "tcp.closed", PFTM_TCP_CLOSED }, { "tcp.tsdiff", PFTM_TS_DIFF }, { "udp.first", PFTM_UDP_FIRST_PACKET }, { "udp.single", PFTM_UDP_SINGLE }, { "udp.multiple", PFTM_UDP_MULTIPLE }, { "icmp.first", PFTM_ICMP_FIRST_PACKET }, { "icmp.error", PFTM_ICMP_ERROR_REPLY }, { "other.first", PFTM_OTHER_FIRST_PACKET }, { "other.single", PFTM_OTHER_SINGLE }, { "other.multiple", PFTM_OTHER_MULTIPLE }, { "frag", PFTM_FRAG }, { "interval", PFTM_INTERVAL }, { "adaptive.start", PFTM_ADAPTIVE_START }, { "adaptive.end", PFTM_ADAPTIVE_END }, { "src.track", PFTM_SRC_NODE }, { NULL, 0 } }; const struct icmptypeent * geticmptypebynumber(u_int8_t type, sa_family_t af) { unsigned int i; if (af != AF_INET6) { for (i=0; i < nitems(icmp_type); i++) { if (type == icmp_type[i].type) return (&icmp_type[i]); } } else { for (i=0; i < nitems(icmp6_type); i++) { if (type == icmp6_type[i].type) return (&icmp6_type[i]); } } return (NULL); } const struct icmptypeent * geticmptypebyname(char *w, sa_family_t af) { unsigned int i; if (af != AF_INET6) { for (i=0; i < nitems(icmp_type); i++) { if (!strcmp(w, icmp_type[i].name)) return (&icmp_type[i]); } } else { for (i=0; i < nitems(icmp6_type); i++) { if (!strcmp(w, icmp6_type[i].name)) return (&icmp6_type[i]); } } return (NULL); } const struct icmpcodeent * geticmpcodebynumber(u_int8_t type, u_int8_t code, sa_family_t af) { unsigned int i; if (af != AF_INET6) { for (i=0; i < nitems(icmp_code); i++) { if (type == icmp_code[i].type && code == icmp_code[i].code) return (&icmp_code[i]); } } else { for (i=0; i < nitems(icmp6_code); i++) { if (type == icmp6_code[i].type && code == icmp6_code[i].code) return (&icmp6_code[i]); } } return (NULL); } const struct icmpcodeent * geticmpcodebyname(u_long type, char *w, sa_family_t af) { unsigned int i; if (af != AF_INET6) { for (i=0; i < nitems(icmp_code); i++) { if (type == icmp_code[i].type && !strcmp(w, icmp_code[i].name)) return (&icmp_code[i]); } } else { for (i=0; i < nitems(icmp6_code); i++) { if (type == icmp6_code[i].type && !strcmp(w, icmp6_code[i].name)) return (&icmp6_code[i]); } } return (NULL); } void print_op(u_int8_t op, const char *a1, const char *a2) { if (op == PF_OP_IRG) printf(" %s >< %s", a1, a2); else if (op == PF_OP_XRG) printf(" %s <> %s", a1, a2); else if (op == PF_OP_EQ) printf(" = %s", a1); else if (op == PF_OP_NE) printf(" != %s", a1); else if (op == PF_OP_LT) printf(" < %s", a1); else if (op == PF_OP_LE) printf(" <= %s", a1); else if (op == PF_OP_GT) printf(" > %s", a1); else if (op == PF_OP_GE) printf(" >= %s", a1); else if (op == PF_OP_RRG) printf(" %s:%s", a1, a2); } void print_port(u_int8_t op, u_int16_t p1, u_int16_t p2, const char *proto, int numeric) { char a1[6], a2[6]; struct servent *s; if (!numeric) s = getservbyport(p1, proto); else s = NULL; p1 = ntohs(p1); p2 = ntohs(p2); snprintf(a1, sizeof(a1), "%u", p1); snprintf(a2, sizeof(a2), "%u", p2); printf(" port"); if (s != NULL && (op == PF_OP_EQ || op == PF_OP_NE)) print_op(op, s->s_name, a2); else print_op(op, a1, a2); } void print_ugid(u_int8_t op, unsigned u1, unsigned u2, const char *t, unsigned umax) { char a1[11], a2[11]; snprintf(a1, sizeof(a1), "%u", u1); snprintf(a2, sizeof(a2), "%u", u2); printf(" %s", t); if (u1 == umax && (op == PF_OP_EQ || op == PF_OP_NE)) print_op(op, "unknown", a2); else print_op(op, a1, a2); } void print_flags(u_int8_t f) { int i; for (i = 0; tcpflags[i]; ++i) if (f & (1 << i)) printf("%c", tcpflags[i]); } void print_fromto(struct pf_rule_addr *src, pf_osfp_t osfp, struct pf_rule_addr *dst, sa_family_t af, u_int8_t proto, int verbose, int numeric) { char buf[PF_OSFP_LEN*3]; if (src->addr.type == PF_ADDR_ADDRMASK && dst->addr.type == PF_ADDR_ADDRMASK && PF_AZERO(&src->addr.v.a.addr, AF_INET6) && PF_AZERO(&src->addr.v.a.mask, AF_INET6) && PF_AZERO(&dst->addr.v.a.addr, AF_INET6) && PF_AZERO(&dst->addr.v.a.mask, AF_INET6) && !src->neg && !dst->neg && !src->port_op && !dst->port_op && osfp == PF_OSFP_ANY) printf(" all"); else { printf(" from "); if (src->neg) printf("! "); print_addr(&src->addr, af, verbose); if (src->port_op) print_port(src->port_op, src->port[0], src->port[1], proto == IPPROTO_TCP ? "tcp" : "udp", numeric); if (osfp != PF_OSFP_ANY) printf(" os \"%s\"", pfctl_lookup_fingerprint(osfp, buf, sizeof(buf))); printf(" to "); if (dst->neg) printf("! "); print_addr(&dst->addr, af, verbose); if (dst->port_op) print_port(dst->port_op, dst->port[0], dst->port[1], proto == IPPROTO_TCP ? "tcp" : "udp", numeric); } } void print_pool(struct pf_pool *pool, u_int16_t p1, u_int16_t p2, sa_family_t af, int id) { struct pf_pooladdr *pooladdr; if ((TAILQ_FIRST(&pool->list) != NULL) && TAILQ_NEXT(TAILQ_FIRST(&pool->list), entries) != NULL) printf("{ "); TAILQ_FOREACH(pooladdr, &pool->list, entries){ switch (id) { case PF_NAT: case PF_RDR: case PF_BINAT: print_addr(&pooladdr->addr, af, 0); break; case PF_PASS: if (PF_AZERO(&pooladdr->addr.v.a.addr, af)) printf("%s", pooladdr->ifname); else { printf("(%s ", pooladdr->ifname); print_addr(&pooladdr->addr, af, 0); printf(")"); } break; default: break; } if (TAILQ_NEXT(pooladdr, entries) != NULL) printf(", "); else if (TAILQ_NEXT(TAILQ_FIRST(&pool->list), entries) != NULL) printf(" }"); } switch (id) { case PF_NAT: if ((p1 != PF_NAT_PROXY_PORT_LOW || p2 != PF_NAT_PROXY_PORT_HIGH) && (p1 != 0 || p2 != 0)) { if (p1 == p2) printf(" port %u", p1); else printf(" port %u:%u", p1, p2); } break; case PF_RDR: if (p1) { printf(" port %u", p1); if (p2 && (p2 != p1)) printf(":%u", p2); } break; default: break; } switch (pool->opts & PF_POOL_TYPEMASK) { case PF_POOL_NONE: break; case PF_POOL_BITMASK: printf(" bitmask"); break; case PF_POOL_RANDOM: printf(" random"); break; case PF_POOL_SRCHASH: printf(" source-hash 0x%08x%08x%08x%08x", pool->key.key32[0], pool->key.key32[1], pool->key.key32[2], pool->key.key32[3]); break; case PF_POOL_ROUNDROBIN: printf(" round-robin"); break; } if (pool->opts & PF_POOL_STICKYADDR) printf(" sticky-address"); if (id == PF_NAT && p1 == 0 && p2 == 0) printf(" static-port"); } const char * const pf_reasons[PFRES_MAX+1] = PFRES_NAMES; const char * const pf_lcounters[LCNT_MAX+1] = LCNT_NAMES; const char * const pf_fcounters[FCNT_MAX+1] = FCNT_NAMES; const char * const pf_scounters[FCNT_MAX+1] = FCNT_NAMES; void print_status(struct pf_status *s, int opts) { char statline[80], *running; time_t runtime; int i; char buf[PF_MD5_DIGEST_LENGTH * 2 + 1]; static const char hex[] = "0123456789abcdef"; runtime = time(NULL) - s->since; running = s->running ? "Enabled" : "Disabled"; if (s->since) { unsigned int sec, min, hrs, day = runtime; sec = day % 60; day /= 60; min = day % 60; day /= 60; hrs = day % 24; day /= 24; snprintf(statline, sizeof(statline), "Status: %s for %u days %.2u:%.2u:%.2u", running, day, hrs, min, sec); } else snprintf(statline, sizeof(statline), "Status: %s", running); printf("%-44s", statline); switch (s->debug) { case PF_DEBUG_NONE: printf("%15s\n\n", "Debug: None"); break; case PF_DEBUG_URGENT: printf("%15s\n\n", "Debug: Urgent"); break; case PF_DEBUG_MISC: printf("%15s\n\n", "Debug: Misc"); break; case PF_DEBUG_NOISY: printf("%15s\n\n", "Debug: Loud"); break; } if (opts & PF_OPT_VERBOSE) { printf("Hostid: 0x%08x\n", ntohl(s->hostid)); for (i = 0; i < PF_MD5_DIGEST_LENGTH; i++) { buf[i + i] = hex[s->pf_chksum[i] >> 4]; buf[i + i + 1] = hex[s->pf_chksum[i] & 0x0f]; } buf[i + i] = '\0'; printf("Checksum: 0x%s\n\n", buf); } if (s->ifname[0] != 0) { printf("Interface Stats for %-16s %5s %16s\n", s->ifname, "IPv4", "IPv6"); printf(" %-25s %14llu %16llu\n", "Bytes In", (unsigned long long)s->bcounters[0][0], (unsigned long long)s->bcounters[1][0]); printf(" %-25s %14llu %16llu\n", "Bytes Out", (unsigned long long)s->bcounters[0][1], (unsigned long long)s->bcounters[1][1]); printf(" Packets In\n"); printf(" %-23s %14llu %16llu\n", "Passed", (unsigned long long)s->pcounters[0][0][PF_PASS], (unsigned long long)s->pcounters[1][0][PF_PASS]); printf(" %-23s %14llu %16llu\n", "Blocked", (unsigned long long)s->pcounters[0][0][PF_DROP], (unsigned long long)s->pcounters[1][0][PF_DROP]); printf(" Packets Out\n"); printf(" %-23s %14llu %16llu\n", "Passed", (unsigned long long)s->pcounters[0][1][PF_PASS], (unsigned long long)s->pcounters[1][1][PF_PASS]); printf(" %-23s %14llu %16llu\n\n", "Blocked", (unsigned long long)s->pcounters[0][1][PF_DROP], (unsigned long long)s->pcounters[1][1][PF_DROP]); } printf("%-27s %14s %16s\n", "State Table", "Total", "Rate"); printf(" %-25s %14u %14s\n", "current entries", s->states, ""); for (i = 0; i < FCNT_MAX; i++) { printf(" %-25s %14llu ", pf_fcounters[i], (unsigned long long)s->fcounters[i]); if (runtime > 0) printf("%14.1f/s\n", (double)s->fcounters[i] / (double)runtime); else printf("%14s\n", ""); } if (opts & PF_OPT_VERBOSE) { printf("Source Tracking Table\n"); printf(" %-25s %14u %14s\n", "current entries", s->src_nodes, ""); for (i = 0; i < SCNT_MAX; i++) { printf(" %-25s %14lld ", pf_scounters[i], #ifdef __FreeBSD__ (long long)s->scounters[i]); #else s->scounters[i]); #endif if (runtime > 0) printf("%14.1f/s\n", (double)s->scounters[i] / (double)runtime); else printf("%14s\n", ""); } } printf("Counters\n"); for (i = 0; i < PFRES_MAX; i++) { printf(" %-25s %14llu ", pf_reasons[i], (unsigned long long)s->counters[i]); if (runtime > 0) printf("%14.1f/s\n", (double)s->counters[i] / (double)runtime); else printf("%14s\n", ""); } if (opts & PF_OPT_VERBOSE) { printf("Limit Counters\n"); for (i = 0; i < LCNT_MAX; i++) { printf(" %-25s %14lld ", pf_lcounters[i], #ifdef __FreeBSD__ (unsigned long long)s->lcounters[i]); #else s->lcounters[i]); #endif if (runtime > 0) printf("%14.1f/s\n", (double)s->lcounters[i] / (double)runtime); else printf("%14s\n", ""); } } } void print_src_node(struct pf_src_node *sn, int opts) { struct pf_addr_wrap aw; int min, sec; memset(&aw, 0, sizeof(aw)); if (sn->af == AF_INET) aw.v.a.mask.addr32[0] = 0xffffffff; else memset(&aw.v.a.mask, 0xff, sizeof(aw.v.a.mask)); aw.v.a.addr = sn->addr; print_addr(&aw, sn->af, opts & PF_OPT_VERBOSE2); printf(" -> "); aw.v.a.addr = sn->raddr; print_addr(&aw, sn->af, opts & PF_OPT_VERBOSE2); printf(" ( states %u, connections %u, rate %u.%u/%us )\n", sn->states, sn->conn, sn->conn_rate.count / 1000, (sn->conn_rate.count % 1000) / 100, sn->conn_rate.seconds); if (opts & PF_OPT_VERBOSE) { sec = sn->creation % 60; sn->creation /= 60; min = sn->creation % 60; sn->creation /= 60; printf(" age %.2u:%.2u:%.2u", sn->creation, min, sec); if (sn->states == 0) { sec = sn->expire % 60; sn->expire /= 60; min = sn->expire % 60; sn->expire /= 60; printf(", expires in %.2u:%.2u:%.2u", sn->expire, min, sec); } printf(", %llu pkts, %llu bytes", #ifdef __FreeBSD__ (unsigned long long)(sn->packets[0] + sn->packets[1]), (unsigned long long)(sn->bytes[0] + sn->bytes[1])); #else sn->packets[0] + sn->packets[1], sn->bytes[0] + sn->bytes[1]); #endif switch (sn->ruletype) { case PF_NAT: if (sn->rule.nr != -1) printf(", nat rule %u", sn->rule.nr); break; case PF_RDR: if (sn->rule.nr != -1) printf(", rdr rule %u", sn->rule.nr); break; case PF_PASS: if (sn->rule.nr != -1) printf(", filter rule %u", sn->rule.nr); break; } printf("\n"); } } void print_rule(struct pf_rule *r, const char *anchor_call, int verbose, int numeric) { static const char *actiontypes[] = { "pass", "block", "scrub", "no scrub", "nat", "no nat", "binat", "no binat", "rdr", "no rdr" }; static const char *anchortypes[] = { "anchor", "anchor", "anchor", "anchor", "nat-anchor", "nat-anchor", "binat-anchor", "binat-anchor", "rdr-anchor", "rdr-anchor" }; int i, opts; if (verbose) printf("@%d ", r->nr); if (r->action > PF_NORDR) printf("action(%d)", r->action); else if (anchor_call[0]) { if (anchor_call[0] == '_') { printf("%s", anchortypes[r->action]); } else printf("%s \"%s\"", anchortypes[r->action], anchor_call); } else { printf("%s", actiontypes[r->action]); if (r->natpass) printf(" pass"); } if (r->action == PF_DROP) { if (r->rule_flag & PFRULE_RETURN) printf(" return"); else if (r->rule_flag & PFRULE_RETURNRST) { if (!r->return_ttl) printf(" return-rst"); else printf(" return-rst(ttl %d)", r->return_ttl); } else if (r->rule_flag & PFRULE_RETURNICMP) { const struct icmpcodeent *ic, *ic6; ic = geticmpcodebynumber(r->return_icmp >> 8, r->return_icmp & 255, AF_INET); ic6 = geticmpcodebynumber(r->return_icmp6 >> 8, r->return_icmp6 & 255, AF_INET6); switch (r->af) { case AF_INET: printf(" return-icmp"); if (ic == NULL) printf("(%u)", r->return_icmp & 255); else printf("(%s)", ic->name); break; case AF_INET6: printf(" return-icmp6"); if (ic6 == NULL) printf("(%u)", r->return_icmp6 & 255); else printf("(%s)", ic6->name); break; default: printf(" return-icmp"); if (ic == NULL) printf("(%u, ", r->return_icmp & 255); else printf("(%s, ", ic->name); if (ic6 == NULL) printf("%u)", r->return_icmp6 & 255); else printf("%s)", ic6->name); break; } } else printf(" drop"); } if (r->direction == PF_IN) printf(" in"); else if (r->direction == PF_OUT) printf(" out"); if (r->log) { printf(" log"); if (r->log & ~PF_LOG || r->logif) { int count = 0; printf(" ("); if (r->log & PF_LOG_ALL) printf("%sall", count++ ? ", " : ""); if (r->log & PF_LOG_SOCKET_LOOKUP) printf("%suser", count++ ? ", " : ""); if (r->logif) printf("%sto pflog%u", count++ ? ", " : "", r->logif); printf(")"); } } if (r->quick) printf(" quick"); if (r->ifname[0]) { if (r->ifnot) printf(" on ! %s", r->ifname); else printf(" on %s", r->ifname); } if (r->rt) { if (r->rt == PF_ROUTETO) printf(" route-to"); else if (r->rt == PF_REPLYTO) printf(" reply-to"); else if (r->rt == PF_DUPTO) printf(" dup-to"); - else if (r->rt == PF_FASTROUTE) - printf(" fastroute"); - if (r->rt != PF_FASTROUTE) { - printf(" "); - print_pool(&r->rpool, 0, 0, r->af, PF_PASS); - } + printf(" "); + print_pool(&r->rpool, 0, 0, r->af, PF_PASS); } if (r->af) { if (r->af == AF_INET) printf(" inet"); else printf(" inet6"); } if (r->proto) { struct protoent *p; if ((p = getprotobynumber(r->proto)) != NULL) printf(" proto %s", p->p_name); else printf(" proto %u", r->proto); } print_fromto(&r->src, r->os_fingerprint, &r->dst, r->af, r->proto, verbose, numeric); if (r->uid.op) print_ugid(r->uid.op, r->uid.uid[0], r->uid.uid[1], "user", UID_MAX); if (r->gid.op) print_ugid(r->gid.op, r->gid.gid[0], r->gid.gid[1], "group", GID_MAX); if (r->flags || r->flagset) { printf(" flags "); print_flags(r->flags); printf("/"); print_flags(r->flagset); } else if (r->action == PF_PASS && (!r->proto || r->proto == IPPROTO_TCP) && !(r->rule_flag & PFRULE_FRAGMENT) && !anchor_call[0] && r->keep_state) printf(" flags any"); if (r->type) { const struct icmptypeent *it; it = geticmptypebynumber(r->type-1, r->af); if (r->af != AF_INET6) printf(" icmp-type"); else printf(" icmp6-type"); if (it != NULL) printf(" %s", it->name); else printf(" %u", r->type-1); if (r->code) { const struct icmpcodeent *ic; ic = geticmpcodebynumber(r->type-1, r->code-1, r->af); if (ic != NULL) printf(" code %s", ic->name); else printf(" code %u", r->code-1); } } if (r->tos) printf(" tos 0x%2.2x", r->tos); if (r->prio) printf(" prio %u", r->prio == PF_PRIO_ZERO ? 0 : r->prio); if (r->scrub_flags & PFSTATE_SETMASK) { char *comma = ""; printf(" set ("); if (r->scrub_flags & PFSTATE_SETPRIO) { if (r->set_prio[0] == r->set_prio[1]) printf("%s prio %u", comma, r->set_prio[0]); else printf("%s prio(%u, %u)", comma, r->set_prio[0], r->set_prio[1]); comma = ","; } printf(" )"); } if (!r->keep_state && r->action == PF_PASS && !anchor_call[0]) printf(" no state"); else if (r->keep_state == PF_STATE_NORMAL) printf(" keep state"); else if (r->keep_state == PF_STATE_MODULATE) printf(" modulate state"); else if (r->keep_state == PF_STATE_SYNPROXY) printf(" synproxy state"); if (r->prob) { char buf[20]; snprintf(buf, sizeof(buf), "%f", r->prob*100.0/(UINT_MAX+1.0)); for (i = strlen(buf)-1; i > 0; i--) { if (buf[i] == '0') buf[i] = '\0'; else { if (buf[i] == '.') buf[i] = '\0'; break; } } printf(" probability %s%%", buf); } opts = 0; if (r->max_states || r->max_src_nodes || r->max_src_states) opts = 1; if (r->rule_flag & PFRULE_NOSYNC) opts = 1; if (r->rule_flag & PFRULE_SRCTRACK) opts = 1; if (r->rule_flag & PFRULE_IFBOUND) opts = 1; if (r->rule_flag & PFRULE_STATESLOPPY) opts = 1; for (i = 0; !opts && i < PFTM_MAX; ++i) if (r->timeout[i]) opts = 1; if (opts) { printf(" ("); if (r->max_states) { printf("max %u", r->max_states); opts = 0; } if (r->rule_flag & PFRULE_NOSYNC) { if (!opts) printf(", "); printf("no-sync"); opts = 0; } if (r->rule_flag & PFRULE_SRCTRACK) { if (!opts) printf(", "); printf("source-track"); if (r->rule_flag & PFRULE_RULESRCTRACK) printf(" rule"); else printf(" global"); opts = 0; } if (r->max_src_states) { if (!opts) printf(", "); printf("max-src-states %u", r->max_src_states); opts = 0; } if (r->max_src_conn) { if (!opts) printf(", "); printf("max-src-conn %u", r->max_src_conn); opts = 0; } if (r->max_src_conn_rate.limit) { if (!opts) printf(", "); printf("max-src-conn-rate %u/%u", r->max_src_conn_rate.limit, r->max_src_conn_rate.seconds); opts = 0; } if (r->max_src_nodes) { if (!opts) printf(", "); printf("max-src-nodes %u", r->max_src_nodes); opts = 0; } if (r->overload_tblname[0]) { if (!opts) printf(", "); printf("overload <%s>", r->overload_tblname); if (r->flush) printf(" flush"); if (r->flush & PF_FLUSH_GLOBAL) printf(" global"); } if (r->rule_flag & PFRULE_IFBOUND) { if (!opts) printf(", "); printf("if-bound"); opts = 0; } if (r->rule_flag & PFRULE_STATESLOPPY) { if (!opts) printf(", "); printf("sloppy"); opts = 0; } for (i = 0; i < PFTM_MAX; ++i) if (r->timeout[i]) { int j; if (!opts) printf(", "); opts = 0; for (j = 0; pf_timeouts[j].name != NULL; ++j) if (pf_timeouts[j].timeout == i) break; printf("%s %u", pf_timeouts[j].name == NULL ? "inv.timeout" : pf_timeouts[j].name, r->timeout[i]); } printf(")"); } if (r->rule_flag & PFRULE_FRAGMENT) printf(" fragment"); if (r->rule_flag & PFRULE_NODF) printf(" no-df"); if (r->rule_flag & PFRULE_RANDOMID) printf(" random-id"); if (r->min_ttl) printf(" min-ttl %d", r->min_ttl); if (r->max_mss) printf(" max-mss %d", r->max_mss); if (r->rule_flag & PFRULE_SET_TOS) printf(" set-tos 0x%2.2x", r->set_tos); if (r->allow_opts) printf(" allow-opts"); if (r->action == PF_SCRUB) { if (r->rule_flag & PFRULE_REASSEMBLE_TCP) printf(" reassemble tcp"); printf(" fragment reassemble"); } if (r->label[0]) printf(" label \"%s\"", r->label); if (r->qname[0] && r->pqname[0]) printf(" queue(%s, %s)", r->qname, r->pqname); else if (r->qname[0]) printf(" queue %s", r->qname); if (r->tagname[0]) printf(" tag %s", r->tagname); if (r->match_tagname[0]) { if (r->match_tag_not) printf(" !"); printf(" tagged %s", r->match_tagname); } if (r->rtableid != -1) printf(" rtable %u", r->rtableid); if (r->divert.port) { #ifdef __FreeBSD__ printf(" divert-to %u", ntohs(r->divert.port)); #else if (PF_AZERO(&r->divert.addr, r->af)) { printf(" divert-reply"); } else { /* XXX cut&paste from print_addr */ char buf[48]; printf(" divert-to "); if (inet_ntop(r->af, &r->divert.addr, buf, sizeof(buf)) == NULL) printf("?"); else printf("%s", buf); printf(" port %u", ntohs(r->divert.port)); } #endif } if (!anchor_call[0] && (r->action == PF_NAT || r->action == PF_BINAT || r->action == PF_RDR)) { printf(" -> "); print_pool(&r->rpool, r->rpool.proxy_port[0], r->rpool.proxy_port[1], r->af, r->action); } } void print_tabledef(const char *name, int flags, int addrs, struct node_tinithead *nodes) { struct node_tinit *ti, *nti; struct node_host *h; printf("table <%s>", name); if (flags & PFR_TFLAG_CONST) printf(" const"); if (flags & PFR_TFLAG_PERSIST) printf(" persist"); if (flags & PFR_TFLAG_COUNTERS) printf(" counters"); SIMPLEQ_FOREACH(ti, nodes, entries) { if (ti->file) { printf(" file \"%s\"", ti->file); continue; } printf(" {"); for (;;) { for (h = ti->host; h != NULL; h = h->next) { printf(h->not ? " !" : " "); print_addr(&h->addr, h->af, 0); } nti = SIMPLEQ_NEXT(ti, entries); if (nti != NULL && nti->file == NULL) ti = nti; /* merge lists */ else break; } printf(" }"); } if (addrs && SIMPLEQ_EMPTY(nodes)) printf(" { }"); printf("\n"); } int parse_flags(char *s) { char *p, *q; u_int8_t f = 0; for (p = s; *p; p++) { if ((q = strchr(tcpflags, *p)) == NULL) return -1; else f |= 1 << (q - tcpflags); } return (f ? f : PF_TH_ALL); } void set_ipmask(struct node_host *h, u_int8_t b) { struct pf_addr *m, *n; int i, j = 0; m = &h->addr.v.a.mask; memset(m, 0, sizeof(*m)); while (b >= 32) { m->addr32[j++] = 0xffffffff; b -= 32; } for (i = 31; i > 31-b; --i) m->addr32[j] |= (1 << i); if (b) m->addr32[j] = htonl(m->addr32[j]); /* Mask off bits of the address that will never be used. */ n = &h->addr.v.a.addr; if (h->addr.type == PF_ADDR_ADDRMASK) for (i = 0; i < 4; i++) n->addr32[i] = n->addr32[i] & m->addr32[i]; } int check_netmask(struct node_host *h, sa_family_t af) { struct node_host *n = NULL; struct pf_addr *m; for (n = h; n != NULL; n = n->next) { if (h->addr.type == PF_ADDR_TABLE) continue; m = &h->addr.v.a.mask; /* fix up netmask for dynaddr */ if (af == AF_INET && h->addr.type == PF_ADDR_DYNIFTL && unmask(m, AF_INET6) > 32) set_ipmask(n, 32); /* netmasks > 32 bit are invalid on v4 */ if (af == AF_INET && (m->addr32[1] || m->addr32[2] || m->addr32[3])) { fprintf(stderr, "netmask %u invalid for IPv4 address\n", unmask(m, AF_INET6)); return (1); } } return (0); } /* interface lookup routines */ static struct node_host *iftab; void ifa_load(void) { struct ifaddrs *ifap, *ifa; struct node_host *n = NULL, *h = NULL; if (getifaddrs(&ifap) < 0) err(1, "getifaddrs"); for (ifa = ifap; ifa; ifa = ifa->ifa_next) { if (!(ifa->ifa_addr->sa_family == AF_INET || ifa->ifa_addr->sa_family == AF_INET6 || ifa->ifa_addr->sa_family == AF_LINK)) continue; n = calloc(1, sizeof(struct node_host)); if (n == NULL) err(1, "address: calloc"); n->af = ifa->ifa_addr->sa_family; n->ifa_flags = ifa->ifa_flags; #ifdef __KAME__ if (n->af == AF_INET6 && IN6_IS_ADDR_LINKLOCAL(&((struct sockaddr_in6 *) ifa->ifa_addr)->sin6_addr) && ((struct sockaddr_in6 *)ifa->ifa_addr)->sin6_scope_id == 0) { struct sockaddr_in6 *sin6; sin6 = (struct sockaddr_in6 *)ifa->ifa_addr; sin6->sin6_scope_id = sin6->sin6_addr.s6_addr[2] << 8 | sin6->sin6_addr.s6_addr[3]; sin6->sin6_addr.s6_addr[2] = 0; sin6->sin6_addr.s6_addr[3] = 0; } #endif n->ifindex = 0; if (n->af == AF_INET) { memcpy(&n->addr.v.a.addr, &((struct sockaddr_in *) ifa->ifa_addr)->sin_addr.s_addr, sizeof(struct in_addr)); memcpy(&n->addr.v.a.mask, &((struct sockaddr_in *) ifa->ifa_netmask)->sin_addr.s_addr, sizeof(struct in_addr)); if (ifa->ifa_broadaddr != NULL) memcpy(&n->bcast, &((struct sockaddr_in *) ifa->ifa_broadaddr)->sin_addr.s_addr, sizeof(struct in_addr)); if (ifa->ifa_dstaddr != NULL) memcpy(&n->peer, &((struct sockaddr_in *) ifa->ifa_dstaddr)->sin_addr.s_addr, sizeof(struct in_addr)); } else if (n->af == AF_INET6) { memcpy(&n->addr.v.a.addr, &((struct sockaddr_in6 *) ifa->ifa_addr)->sin6_addr.s6_addr, sizeof(struct in6_addr)); memcpy(&n->addr.v.a.mask, &((struct sockaddr_in6 *) ifa->ifa_netmask)->sin6_addr.s6_addr, sizeof(struct in6_addr)); if (ifa->ifa_broadaddr != NULL) memcpy(&n->bcast, &((struct sockaddr_in6 *) ifa->ifa_broadaddr)->sin6_addr.s6_addr, sizeof(struct in6_addr)); if (ifa->ifa_dstaddr != NULL) memcpy(&n->peer, &((struct sockaddr_in6 *) ifa->ifa_dstaddr)->sin6_addr.s6_addr, sizeof(struct in6_addr)); n->ifindex = ((struct sockaddr_in6 *) ifa->ifa_addr)->sin6_scope_id; } if ((n->ifname = strdup(ifa->ifa_name)) == NULL) err(1, "ifa_load: strdup"); n->next = NULL; n->tail = n; if (h == NULL) h = n; else { h->tail->next = n; h->tail = n; } } iftab = h; freeifaddrs(ifap); } int get_socket_domain(void) { int sdom; sdom = AF_UNSPEC; #ifdef WITH_INET6 if (sdom == AF_UNSPEC && feature_present("inet6")) sdom = AF_INET6; #endif #ifdef WITH_INET if (sdom == AF_UNSPEC && feature_present("inet")) sdom = AF_INET; #endif if (sdom == AF_UNSPEC) sdom = AF_LINK; return (sdom); } struct node_host * ifa_exists(const char *ifa_name) { struct node_host *n; struct ifgroupreq ifgr; int s; if (iftab == NULL) ifa_load(); /* check wether this is a group */ if ((s = socket(get_socket_domain(), SOCK_DGRAM, 0)) == -1) err(1, "socket"); bzero(&ifgr, sizeof(ifgr)); strlcpy(ifgr.ifgr_name, ifa_name, sizeof(ifgr.ifgr_name)); if (ioctl(s, SIOCGIFGMEMB, (caddr_t)&ifgr) == 0) { /* fake a node_host */ if ((n = calloc(1, sizeof(*n))) == NULL) err(1, "calloc"); if ((n->ifname = strdup(ifa_name)) == NULL) err(1, "strdup"); close(s); return (n); } close(s); for (n = iftab; n; n = n->next) { if (n->af == AF_LINK && !strncmp(n->ifname, ifa_name, IFNAMSIZ)) return (n); } return (NULL); } struct node_host * ifa_grouplookup(const char *ifa_name, int flags) { struct ifg_req *ifg; struct ifgroupreq ifgr; int s, len; struct node_host *n, *h = NULL; if ((s = socket(get_socket_domain(), SOCK_DGRAM, 0)) == -1) err(1, "socket"); bzero(&ifgr, sizeof(ifgr)); strlcpy(ifgr.ifgr_name, ifa_name, sizeof(ifgr.ifgr_name)); if (ioctl(s, SIOCGIFGMEMB, (caddr_t)&ifgr) == -1) { close(s); return (NULL); } len = ifgr.ifgr_len; if ((ifgr.ifgr_groups = calloc(1, len)) == NULL) err(1, "calloc"); if (ioctl(s, SIOCGIFGMEMB, (caddr_t)&ifgr) == -1) err(1, "SIOCGIFGMEMB"); for (ifg = ifgr.ifgr_groups; ifg && len >= sizeof(struct ifg_req); ifg++) { len -= sizeof(struct ifg_req); if ((n = ifa_lookup(ifg->ifgrq_member, flags)) == NULL) continue; if (h == NULL) h = n; else { h->tail->next = n; h->tail = n->tail; } } free(ifgr.ifgr_groups); close(s); return (h); } struct node_host * ifa_lookup(const char *ifa_name, int flags) { struct node_host *p = NULL, *h = NULL, *n = NULL; int got4 = 0, got6 = 0; const char *last_if = NULL; if ((h = ifa_grouplookup(ifa_name, flags)) != NULL) return (h); if (!strncmp(ifa_name, "self", IFNAMSIZ)) ifa_name = NULL; if (iftab == NULL) ifa_load(); for (p = iftab; p; p = p->next) { if (ifa_skip_if(ifa_name, p)) continue; if ((flags & PFI_AFLAG_BROADCAST) && p->af != AF_INET) continue; if ((flags & PFI_AFLAG_BROADCAST) && !(p->ifa_flags & IFF_BROADCAST)) continue; if ((flags & PFI_AFLAG_PEER) && !(p->ifa_flags & IFF_POINTOPOINT)) continue; if ((flags & PFI_AFLAG_NETWORK) && p->ifindex > 0) continue; if (last_if == NULL || strcmp(last_if, p->ifname)) got4 = got6 = 0; last_if = p->ifname; if ((flags & PFI_AFLAG_NOALIAS) && p->af == AF_INET && got4) continue; if ((flags & PFI_AFLAG_NOALIAS) && p->af == AF_INET6 && got6) continue; if (p->af == AF_INET) got4 = 1; else got6 = 1; n = calloc(1, sizeof(struct node_host)); if (n == NULL) err(1, "address: calloc"); n->af = p->af; if (flags & PFI_AFLAG_BROADCAST) memcpy(&n->addr.v.a.addr, &p->bcast, sizeof(struct pf_addr)); else if (flags & PFI_AFLAG_PEER) memcpy(&n->addr.v.a.addr, &p->peer, sizeof(struct pf_addr)); else memcpy(&n->addr.v.a.addr, &p->addr.v.a.addr, sizeof(struct pf_addr)); if (flags & PFI_AFLAG_NETWORK) set_ipmask(n, unmask(&p->addr.v.a.mask, n->af)); else { if (n->af == AF_INET) { if (p->ifa_flags & IFF_LOOPBACK && p->ifa_flags & IFF_LINK1) memcpy(&n->addr.v.a.mask, &p->addr.v.a.mask, sizeof(struct pf_addr)); else set_ipmask(n, 32); } else set_ipmask(n, 128); } n->ifindex = p->ifindex; n->next = NULL; n->tail = n; if (h == NULL) h = n; else { h->tail->next = n; h->tail = n; } } return (h); } int ifa_skip_if(const char *filter, struct node_host *p) { int n; if (p->af != AF_INET && p->af != AF_INET6) return (1); if (filter == NULL || !*filter) return (0); if (!strcmp(p->ifname, filter)) return (0); /* exact match */ n = strlen(filter); if (n < 1 || n >= IFNAMSIZ) return (1); /* sanity check */ if (filter[n-1] >= '0' && filter[n-1] <= '9') return (1); /* only do exact match in that case */ if (strncmp(p->ifname, filter, n)) return (1); /* prefix doesn't match */ return (p->ifname[n] < '0' || p->ifname[n] > '9'); } struct node_host * host(const char *s) { struct node_host *h = NULL; int mask, v4mask, v6mask, cont = 1; char *p, *q, *ps; if ((p = strrchr(s, '/')) != NULL) { mask = strtol(p+1, &q, 0); if (!q || *q || mask > 128 || q == (p+1)) { fprintf(stderr, "invalid netmask '%s'\n", p); return (NULL); } if ((ps = malloc(strlen(s) - strlen(p) + 1)) == NULL) err(1, "host: malloc"); strlcpy(ps, s, strlen(s) - strlen(p) + 1); v4mask = v6mask = mask; } else { if ((ps = strdup(s)) == NULL) err(1, "host: strdup"); v4mask = 32; v6mask = 128; mask = -1; } /* interface with this name exists? */ if (cont && (h = host_if(ps, mask)) != NULL) cont = 0; /* IPv4 address? */ if (cont && (h = host_v4(s, mask)) != NULL) cont = 0; /* IPv6 address? */ if (cont && (h = host_v6(ps, v6mask)) != NULL) cont = 0; /* dns lookup */ if (cont && (h = host_dns(ps, v4mask, v6mask)) != NULL) cont = 0; free(ps); if (h == NULL || cont == 1) { fprintf(stderr, "no IP address found for %s\n", s); return (NULL); } return (h); } struct node_host * host_if(const char *s, int mask) { struct node_host *n, *h = NULL; char *p, *ps; int flags = 0; if ((ps = strdup(s)) == NULL) err(1, "host_if: strdup"); while ((p = strrchr(ps, ':')) != NULL) { if (!strcmp(p+1, "network")) flags |= PFI_AFLAG_NETWORK; else if (!strcmp(p+1, "broadcast")) flags |= PFI_AFLAG_BROADCAST; else if (!strcmp(p+1, "peer")) flags |= PFI_AFLAG_PEER; else if (!strcmp(p+1, "0")) flags |= PFI_AFLAG_NOALIAS; else { free(ps); return (NULL); } *p = '\0'; } if (flags & (flags - 1) & PFI_AFLAG_MODEMASK) { /* Yep! */ fprintf(stderr, "illegal combination of interface modifiers\n"); free(ps); return (NULL); } if ((flags & (PFI_AFLAG_NETWORK|PFI_AFLAG_BROADCAST)) && mask > -1) { fprintf(stderr, "network or broadcast lookup, but " "extra netmask given\n"); free(ps); return (NULL); } if (ifa_exists(ps) || !strncmp(ps, "self", IFNAMSIZ)) { /* interface with this name exists */ h = ifa_lookup(ps, flags); for (n = h; n != NULL && mask > -1; n = n->next) set_ipmask(n, mask); } free(ps); return (h); } struct node_host * host_v4(const char *s, int mask) { struct node_host *h = NULL; struct in_addr ina; int bits = 32; memset(&ina, 0, sizeof(struct in_addr)); if (strrchr(s, '/') != NULL) { if ((bits = inet_net_pton(AF_INET, s, &ina, sizeof(ina))) == -1) return (NULL); } else { if (inet_pton(AF_INET, s, &ina) != 1) return (NULL); } h = calloc(1, sizeof(struct node_host)); if (h == NULL) err(1, "address: calloc"); h->ifname = NULL; h->af = AF_INET; h->addr.v.a.addr.addr32[0] = ina.s_addr; set_ipmask(h, bits); h->next = NULL; h->tail = h; return (h); } struct node_host * host_v6(const char *s, int mask) { struct addrinfo hints, *res; struct node_host *h = NULL; memset(&hints, 0, sizeof(hints)); hints.ai_family = AF_INET6; hints.ai_socktype = SOCK_DGRAM; /*dummy*/ hints.ai_flags = AI_NUMERICHOST; if (getaddrinfo(s, "0", &hints, &res) == 0) { h = calloc(1, sizeof(struct node_host)); if (h == NULL) err(1, "address: calloc"); h->ifname = NULL; h->af = AF_INET6; memcpy(&h->addr.v.a.addr, &((struct sockaddr_in6 *)res->ai_addr)->sin6_addr, sizeof(h->addr.v.a.addr)); h->ifindex = ((struct sockaddr_in6 *)res->ai_addr)->sin6_scope_id; set_ipmask(h, mask); freeaddrinfo(res); h->next = NULL; h->tail = h; } return (h); } struct node_host * host_dns(const char *s, int v4mask, int v6mask) { struct addrinfo hints, *res0, *res; struct node_host *n, *h = NULL; int error, noalias = 0; int got4 = 0, got6 = 0; char *p, *ps; if ((ps = strdup(s)) == NULL) err(1, "host_dns: strdup"); if ((p = strrchr(ps, ':')) != NULL && !strcmp(p, ":0")) { noalias = 1; *p = '\0'; } memset(&hints, 0, sizeof(hints)); hints.ai_family = PF_UNSPEC; hints.ai_socktype = SOCK_STREAM; /* DUMMY */ error = getaddrinfo(ps, NULL, &hints, &res0); if (error) { free(ps); return (h); } for (res = res0; res; res = res->ai_next) { if (res->ai_family != AF_INET && res->ai_family != AF_INET6) continue; if (noalias) { if (res->ai_family == AF_INET) { if (got4) continue; got4 = 1; } else { if (got6) continue; got6 = 1; } } n = calloc(1, sizeof(struct node_host)); if (n == NULL) err(1, "host_dns: calloc"); n->ifname = NULL; n->af = res->ai_family; if (res->ai_family == AF_INET) { memcpy(&n->addr.v.a.addr, &((struct sockaddr_in *) res->ai_addr)->sin_addr.s_addr, sizeof(struct in_addr)); set_ipmask(n, v4mask); } else { memcpy(&n->addr.v.a.addr, &((struct sockaddr_in6 *) res->ai_addr)->sin6_addr.s6_addr, sizeof(struct in6_addr)); n->ifindex = ((struct sockaddr_in6 *) res->ai_addr)->sin6_scope_id; set_ipmask(n, v6mask); } n->next = NULL; n->tail = n; if (h == NULL) h = n; else { h->tail->next = n; h->tail = n; } } freeaddrinfo(res0); free(ps); return (h); } /* * convert a hostname to a list of addresses and put them in the given buffer. * test: * if set to 1, only simple addresses are accepted (no netblock, no "!"). */ int append_addr(struct pfr_buffer *b, char *s, int test) { char *r; struct node_host *h, *n; int rv, not = 0; for (r = s; *r == '!'; r++) not = !not; if ((n = host(r)) == NULL) { errno = 0; return (-1); } rv = append_addr_host(b, n, test, not); do { h = n; n = n->next; free(h); } while (n != NULL); return (rv); } /* * same as previous function, but with a pre-parsed input and the ability * to "negate" the result. Does not free the node_host list. * not: * setting it to 1 is equivalent to adding "!" in front of parameter s. */ int append_addr_host(struct pfr_buffer *b, struct node_host *n, int test, int not) { int bits; struct pfr_addr addr; do { bzero(&addr, sizeof(addr)); addr.pfra_not = n->not ^ not; addr.pfra_af = n->af; addr.pfra_net = unmask(&n->addr.v.a.mask, n->af); switch (n->af) { case AF_INET: addr.pfra_ip4addr.s_addr = n->addr.v.a.addr.addr32[0]; bits = 32; break; case AF_INET6: memcpy(&addr.pfra_ip6addr, &n->addr.v.a.addr.v6, sizeof(struct in6_addr)); bits = 128; break; default: errno = EINVAL; return (-1); } if ((test && (not || addr.pfra_net != bits)) || addr.pfra_net > bits) { errno = EINVAL; return (-1); } if (pfr_buf_add(b, &addr)) return (-1); } while ((n = n->next) != NULL); return (0); } int pfctl_add_trans(struct pfr_buffer *buf, int rs_num, const char *anchor) { struct pfioc_trans_e trans; bzero(&trans, sizeof(trans)); trans.rs_num = rs_num; if (strlcpy(trans.anchor, anchor, sizeof(trans.anchor)) >= sizeof(trans.anchor)) errx(1, "pfctl_add_trans: strlcpy"); return pfr_buf_add(buf, &trans); } u_int32_t pfctl_get_ticket(struct pfr_buffer *buf, int rs_num, const char *anchor) { struct pfioc_trans_e *p; PFRB_FOREACH(p, buf) if (rs_num == p->rs_num && !strcmp(anchor, p->anchor)) return (p->ticket); errx(1, "pfctl_get_ticket: assertion failed"); } int pfctl_trans(int dev, struct pfr_buffer *buf, u_long cmd, int from) { struct pfioc_trans trans; bzero(&trans, sizeof(trans)); trans.size = buf->pfrb_size - from; trans.esize = sizeof(struct pfioc_trans_e); trans.array = ((struct pfioc_trans_e *)buf->pfrb_caddr) + from; return ioctl(dev, cmd, &trans); } Index: head/share/man/man5/pf.conf.5 =================================================================== --- head/share/man/man5/pf.conf.5 (revision 306683) +++ head/share/man/man5/pf.conf.5 (revision 306684) @@ -1,3056 +1,3052 @@ .\" $FreeBSD$ .\" $OpenBSD: pf.conf.5,v 1.406 2009/01/31 19:37:12 sobrado Exp $ .\" .\" Copyright (c) 2002, Daniel Hartmeier .\" All rights reserved. .\" .\" Redistribution and use in source and binary forms, with or without .\" modification, are permitted provided that the following conditions .\" are met: .\" .\" - Redistributions of source code must retain the above copyright .\" notice, this list of conditions and the following disclaimer. .\" - 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. .\" .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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 .\" COPYRIGHT HOLDERS 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. .\" -.Dd June 9, 2016 +.Dd September 28, 2016 .Dt PF.CONF 5 .Os .Sh NAME .Nm pf.conf .Nd packet filter configuration file .Sh DESCRIPTION The .Xr pf 4 packet filter modifies, drops or passes packets according to rules or definitions specified in .Nm pf.conf . .Sh STATEMENT ORDER There are seven types of statements in .Nm pf.conf : .Bl -tag -width xxxx .It Cm Macros User-defined variables may be defined and used later, simplifying the configuration file. Macros must be defined before they are referenced in .Nm pf.conf . .It Cm Tables Tables provide a mechanism for increasing the performance and flexibility of rules with large numbers of source or destination addresses. .It Cm Options Options tune the behaviour of the packet filtering engine. .It Cm Traffic Normalization Li (e.g. Em scrub ) Traffic normalization protects internal machines against inconsistencies in Internet protocols and implementations. .It Cm Queueing Queueing provides rule-based bandwidth control. .It Cm Translation Li (Various forms of NAT) Translation rules specify how addresses are to be mapped or redirected to other addresses. .It Cm Packet Filtering Packet filtering provides rule-based blocking or passing of packets. .El .Pp With the exception of .Cm macros and .Cm tables , the types of statements should be grouped and appear in .Nm pf.conf in the order shown above, as this matches the operation of the underlying packet filtering engine. By default .Xr pfctl 8 enforces this order (see .Ar set require-order below). .Pp Comments can be put anywhere in the file using a hash mark .Pq Sq # , and extend to the end of the current line. .Pp Additional configuration files can be included with the .Ic include keyword, for example: .Bd -literal -offset indent include "/etc/pf/sub.filter.conf" .Ed .Sh MACROS Macros can be defined that will later be expanded in context. Macro names must start with a letter, and may contain letters, digits and underscores. Macro names may not be reserved words (for example .Ar pass , .Ar in , .Ar out ) . Macros are not expanded inside quotes. .Pp For example, .Bd -literal -offset indent ext_if = \&"kue0\&" all_ifs = \&"{\&" $ext_if lo0 \&"}\&" pass out on $ext_if from any to any pass in on $ext_if proto tcp from any to any port 25 .Ed .Sh TABLES Tables are named structures which can hold a collection of addresses and networks. Lookups against tables in .Xr pf 4 are relatively fast, making a single rule with tables much more efficient, in terms of processor usage and memory consumption, than a large number of rules which differ only in IP address (either created explicitly or automatically by rule expansion). .Pp Tables can be used as the source or destination of filter rules, .Ar scrub rules or translation rules such as .Ar nat or .Ar rdr (see below for details on the various rule types). Tables can also be used for the redirect address of .Ar nat and .Ar rdr rules and in the routing options of filter rules, but only for .Ar round-robin pools. .Pp Tables can be defined with any of the following .Xr pfctl 8 mechanisms. As with macros, reserved words may not be used as table names. .Bl -tag -width "manually" .It Ar manually Persistent tables can be manually created with the .Ar add or .Ar replace option of .Xr pfctl 8 , before or after the ruleset has been loaded. .It Pa pf.conf Table definitions can be placed directly in this file, and loaded at the same time as other rules are loaded, atomically. Table definitions inside .Nm pf.conf use the .Ar table statement, and are especially useful to define non-persistent tables. The contents of a pre-existing table defined without a list of addresses to initialize it is not altered when .Nm pf.conf is loaded. A table initialized with the empty list, .Li { } , will be cleared on load. .El .Pp Tables may be defined with the following attributes: .Bl -tag -width persist .It Ar persist The .Ar persist flag forces the kernel to keep the table even when no rules refer to it. If the flag is not set, the kernel will automatically remove the table when the last rule referring to it is flushed. .It Ar const The .Ar const flag prevents the user from altering the contents of the table once it has been created. Without that flag, .Xr pfctl 8 can be used to add or remove addresses from the table at any time, even when running with .Xr securelevel 7 = 2. .It Ar counters The .Ar counters flag enables per-address packet and byte counters which can be displayed with .Xr pfctl 8 . .El .Pp For example, .Bd -literal -offset indent table \*(Ltprivate\*(Gt const { 10/8, 172.16/12, 192.168/16 } table \*(Ltbadhosts\*(Gt persist block on fxp0 from { \*(Ltprivate\*(Gt, \*(Ltbadhosts\*(Gt } to any .Ed .Pp creates a table called private, to hold RFC 1918 private network blocks, and a table called badhosts, which is initially empty. A filter rule is set up to block all traffic coming from addresses listed in either table. The private table cannot have its contents changed and the badhosts table will exist even when no active filter rules reference it. Addresses may later be added to the badhosts table, so that traffic from these hosts can be blocked by using .Bd -literal -offset indent # pfctl -t badhosts -Tadd 204.92.77.111 .Ed .Pp A table can also be initialized with an address list specified in one or more external files, using the following syntax: .Bd -literal -offset indent table \*(Ltspam\*(Gt persist file \&"/etc/spammers\&" file \&"/etc/openrelays\&" block on fxp0 from \*(Ltspam\*(Gt to any .Ed .Pp The files .Pa /etc/spammers and .Pa /etc/openrelays list IP addresses, one per line. Any lines beginning with a # are treated as comments and ignored. In addition to being specified by IP address, hosts may also be specified by their hostname. When the resolver is called to add a hostname to a table, .Em all resulting IPv4 and IPv6 addresses are placed into the table. IP addresses can also be entered in a table by specifying a valid interface name, a valid interface group or the .Em self keyword, in which case all addresses assigned to the interface(s) will be added to the table. .Sh OPTIONS .Xr pf 4 may be tuned for various situations using the .Ar set command. .Bl -tag -width xxxx .It Ar set timeout .Pp .Bl -tag -width "src.track" -compact .It Ar interval Interval between purging expired states and fragments. .It Ar frag Seconds before an unassembled fragment is expired. .It Ar src.track Length of time to retain a source tracking entry after the last state expires. .El .Pp When a packet matches a stateful connection, the seconds to live for the connection will be updated to that of the .Ar proto.modifier which corresponds to the connection state. Each packet which matches this state will reset the TTL. Tuning these values may improve the performance of the firewall at the risk of dropping valid idle connections. .Pp .Bl -tag -width xxxx -compact .It Ar tcp.first The state after the first packet. .It Ar tcp.opening The state before the destination host ever sends a packet. .It Ar tcp.established The fully established state. .It Ar tcp.closing The state after the first FIN has been sent. .It Ar tcp.finwait The state after both FINs have been exchanged and the connection is closed. Some hosts (notably web servers on Solaris) send TCP packets even after closing the connection. Increasing .Ar tcp.finwait (and possibly .Ar tcp.closing ) can prevent blocking of such packets. .It Ar tcp.closed The state after one endpoint sends an RST. .El .Pp ICMP and UDP are handled in a fashion similar to TCP, but with a much more limited set of states: .Pp .Bl -tag -width xxxx -compact .It Ar udp.first The state after the first packet. .It Ar udp.single The state if the source host sends more than one packet but the destination host has never sent one back. .It Ar udp.multiple The state if both hosts have sent packets. .It Ar icmp.first The state after the first packet. .It Ar icmp.error The state after an ICMP error came back in response to an ICMP packet. .El .Pp Other protocols are handled similarly to UDP: .Pp .Bl -tag -width xxxx -compact .It Ar other.first .It Ar other.single .It Ar other.multiple .El .Pp Timeout values can be reduced adaptively as the number of state table entries grows. .Pp .Bl -tag -width xxxx -compact .It Ar adaptive.start When the number of state entries exceeds this value, adaptive scaling begins. All timeout values are scaled linearly with factor (adaptive.end - number of states) / (adaptive.end - adaptive.start). .It Ar adaptive.end When reaching this number of state entries, all timeout values become zero, effectively purging all state entries immediately. This value is used to define the scale factor, it should not actually be reached (set a lower state limit, see below). .El .Pp Adaptive timeouts are enabled by default, with an adaptive.start value equal to 60% of the state limit, and an adaptive.end value equal to 120% of the state limit. They can be disabled by setting both adaptive.start and adaptive.end to 0. .Pp The adaptive timeout values can be defined both globally and for each rule. When used on a per-rule basis, the values relate to the number of states created by the rule, otherwise to the total number of states. .Pp For example: .Bd -literal -offset indent set timeout tcp.first 120 set timeout tcp.established 86400 set timeout { adaptive.start 6000, adaptive.end 12000 } set limit states 10000 .Ed .Pp With 9000 state table entries, the timeout values are scaled to 50% (tcp.first 60, tcp.established 43200). .It Ar set loginterface Enable collection of packet and byte count statistics for the given interface or interface group. These statistics can be viewed using .Bd -literal -offset indent # pfctl -s info .Ed .Pp In this example .Xr pf 4 collects statistics on the interface named dc0: .Bd -literal -offset indent set loginterface dc0 .Ed .Pp One can disable the loginterface using: .Bd -literal -offset indent set loginterface none .Ed .It Ar set limit Sets hard limits on the memory pools used by the packet filter. See .Xr zone 9 for an explanation of memory pools. .Pp For example, .Bd -literal -offset indent set limit states 20000 .Ed .Pp sets the maximum number of entries in the memory pool used by state table entries (generated by .Ar pass rules which do not specify .Ar no state ) to 20000. Using .Bd -literal -offset indent set limit frags 20000 .Ed .Pp sets the maximum number of entries in the memory pool used for fragment reassembly (generated by .Ar scrub rules) to 20000. Using .Bd -literal -offset indent set limit src-nodes 2000 .Ed .Pp sets the maximum number of entries in the memory pool used for tracking source IP addresses (generated by the .Ar sticky-address and .Ar src.track options) to 2000. Using .Bd -literal -offset indent set limit tables 1000 set limit table-entries 100000 .Ed .Pp sets limits on the memory pools used by tables. The first limits the number of tables that can exist to 1000. The second limits the overall number of addresses that can be stored in tables to 100000. .Pp Various limits can be combined on a single line: .Bd -literal -offset indent set limit { states 20000, frags 20000, src-nodes 2000 } .Ed .It Ar set ruleset-optimization .Bl -tag -width xxxxxxxx -compact .It Ar none Disable the ruleset optimizer. .It Ar basic Enable basic ruleset optimization. This is the default behaviour. Basic ruleset optimization does four things to improve the performance of ruleset evaluations: .Pp .Bl -enum -compact .It remove duplicate rules .It remove rules that are a subset of another rule .It combine multiple rules into a table when advantageous .It re-order the rules to improve evaluation performance .El .Pp .It Ar profile Uses the currently loaded ruleset as a feedback profile to tailor the ordering of quick rules to actual network traffic. .El .Pp It is important to note that the ruleset optimizer will modify the ruleset to improve performance. A side effect of the ruleset modification is that per-rule accounting statistics will have different meanings than before. If per-rule accounting is important for billing purposes or whatnot, either the ruleset optimizer should not be used or a label field should be added to all of the accounting rules to act as optimization barriers. .Pp Optimization can also be set as a command-line argument to .Xr pfctl 8 , overriding the settings in .Nm . .It Ar set optimization Optimize state timeouts for one of the following network environments: .Pp .Bl -tag -width xxxx -compact .It Ar normal A normal network environment. Suitable for almost all networks. .It Ar high-latency A high-latency environment (such as a satellite connection). .It Ar satellite Alias for .Ar high-latency . .It Ar aggressive Aggressively expire connections. This can greatly reduce the memory usage of the firewall at the cost of dropping idle connections early. .It Ar conservative Extremely conservative settings. Avoid dropping legitimate connections at the expense of greater memory utilization (possibly much greater on a busy network) and slightly increased processor utilization. .El .Pp For example: .Bd -literal -offset indent set optimization aggressive .Ed .It Ar set block-policy The .Ar block-policy option sets the default behaviour for the packet .Ar block action: .Pp .Bl -tag -width xxxxxxxx -compact .It Ar drop Packet is silently dropped. .It Ar return A TCP RST is returned for blocked TCP packets, an ICMP UNREACHABLE is returned for blocked UDP packets, and all other packets are silently dropped. .El .Pp For example: .Bd -literal -offset indent set block-policy return .Ed .It Ar set state-policy The .Ar state-policy option sets the default behaviour for states: .Pp .Bl -tag -width group-bound -compact .It Ar if-bound States are bound to interface. .It Ar floating States can match packets on any interfaces (the default). .El .Pp For example: .Bd -literal -offset indent set state-policy if-bound .Ed .It Ar set state-defaults The .Ar state-defaults option sets the state options for states created from rules without an explicit .Ar keep state . For example: .Bd -literal -offset indent set state-defaults pflow, no-sync .Ed .It Ar set hostid The 32-bit .Ar hostid identifies this firewall's state table entries to other firewalls in a .Xr pfsync 4 failover cluster. By default the hostid is set to a pseudo-random value, however it may be desirable to manually configure it, for example to more easily identify the source of state table entries. .Bd -literal -offset indent set hostid 1 .Ed .Pp The hostid may be specified in either decimal or hexadecimal. .It Ar set require-order By default .Xr pfctl 8 enforces an ordering of the statement types in the ruleset to: .Em options , .Em normalization , .Em queueing , .Em translation , .Em filtering . Setting this option to .Ar no disables this enforcement. There may be non-trivial and non-obvious implications to an out of order ruleset. Consider carefully before disabling the order enforcement. .It Ar set fingerprints Load fingerprints of known operating systems from the given filename. By default fingerprints of known operating systems are automatically loaded from .Xr pf.os 5 in .Pa /etc but can be overridden via this option. Setting this option may leave a small period of time where the fingerprints referenced by the currently active ruleset are inconsistent until the new ruleset finishes loading. .Pp For example: .Pp .Dl set fingerprints \&"/etc/pf.os.devel\&" .It Ar set skip on Aq Ar ifspec List interfaces for which packets should not be filtered. Packets passing in or out on such interfaces are passed as if pf was disabled, i.e. pf does not process them in any way. This can be useful on loopback and other virtual interfaces, when packet filtering is not desired and can have unexpected effects. For example: .Pp .Dl set skip on lo0 .It Ar set debug Set the debug .Ar level to one of the following: .Pp .Bl -tag -width xxxxxxxxxxxx -compact .It Ar none Don't generate debug messages. .It Ar urgent Generate debug messages only for serious errors. .It Ar misc Generate debug messages for various errors. .It Ar loud Generate debug messages for common conditions. .El .El .Sh TRAFFIC NORMALIZATION Traffic normalization is used to sanitize packet content in such a way that there are no ambiguities in packet interpretation on the receiving side. The normalizer does IP fragment reassembly to prevent attacks that confuse intrusion detection systems by sending overlapping IP fragments. Packet normalization is invoked with the .Ar scrub directive. .Pp .Ar scrub has the following options: .Bl -tag -width xxxx .It Ar no-df Clears the .Ar dont-fragment bit from a matching IP packet. Some operating systems are known to generate fragmented packets with the .Ar dont-fragment bit set. This is particularly true with NFS. .Ar Scrub will drop such fragmented .Ar dont-fragment packets unless .Ar no-df is specified. .Pp Unfortunately some operating systems also generate their .Ar dont-fragment packets with a zero IP identification field. Clearing the .Ar dont-fragment bit on packets with a zero IP ID may cause deleterious results if an upstream router later fragments the packet. Using the .Ar random-id modifier (see below) is recommended in combination with the .Ar no-df modifier to ensure unique IP identifiers. .It Ar min-ttl Aq Ar number Enforces a minimum TTL for matching IP packets. .It Ar max-mss Aq Ar number Enforces a maximum MSS for matching TCP packets. .It Xo Ar set-tos Aq Ar string .No \*(Ba Aq Ar number .Xc Enforces a .Em TOS for matching IP packets. .Em TOS may be given as one of .Ar lowdelay , .Ar throughput , .Ar reliability , or as either hex or decimal. .It Ar random-id Replaces the IP identification field with random values to compensate for predictable values generated by many hosts. This option only applies to packets that are not fragmented after the optional fragment reassembly. .It Ar fragment reassemble Using .Ar scrub rules, fragments can be reassembled by normalization. In this case, fragments are buffered until they form a complete packet, and only the completed packet is passed on to the filter. The advantage is that filter rules have to deal only with complete packets, and can ignore fragments. The drawback of caching fragments is the additional memory cost. .It Ar reassemble tcp Statefully normalizes TCP connections. .Ar scrub reassemble tcp rules may not have the direction (in/out) specified. .Ar reassemble tcp performs the following normalizations: .Pp .Bl -tag -width timeout -compact .It ttl Neither side of the connection is allowed to reduce their IP TTL. An attacker may send a packet such that it reaches the firewall, affects the firewall state, and expires before reaching the destination host. .Ar reassemble tcp will raise the TTL of all packets back up to the highest value seen on the connection. .It timestamp modulation Modern TCP stacks will send a timestamp on every TCP packet and echo the other endpoint's timestamp back to them. Many operating systems will merely start the timestamp at zero when first booted, and increment it several times a second. The uptime of the host can be deduced by reading the timestamp and multiplying by a constant. Also observing several different timestamps can be used to count hosts behind a NAT device. And spoofing TCP packets into a connection requires knowing or guessing valid timestamps. Timestamps merely need to be monotonically increasing and not derived off a guessable base time. .Ar reassemble tcp will cause .Ar scrub to modulate the TCP timestamps with a random number. .It extended PAWS checks There is a problem with TCP on long fat pipes, in that a packet might get delayed for longer than it takes the connection to wrap its 32-bit sequence space. In such an occurrence, the old packet would be indistinguishable from a new packet and would be accepted as such. The solution to this is called PAWS: Protection Against Wrapped Sequence numbers. It protects against it by making sure the timestamp on each packet does not go backwards. .Ar reassemble tcp also makes sure the timestamp on the packet does not go forward more than the RFC allows. By doing this, .Xr pf 4 artificially extends the security of TCP sequence numbers by 10 to 18 bits when the host uses appropriately randomized timestamps, since a blind attacker would have to guess the timestamp as well. .El .El .Pp For example, .Bd -literal -offset indent scrub in on $ext_if all fragment reassemble .Ed .Pp The .Ar no option prefixed to a scrub rule causes matching packets to remain unscrubbed, much in the same way as .Ar drop quick works in the packet filter (see below). This mechanism should be used when it is necessary to exclude specific packets from broader scrub rules. .Sh QUEUEING The ALTQ system is currently not available in the GENERIC kernel nor as loadable modules. In order to use the herein after called queueing options one has to use a custom built kernel. Please refer to .Xr altq 4 to learn about the related kernel options. .Pp Packets can be assigned to queues for the purpose of bandwidth control. At least two declarations are required to configure queues, and later any packet filtering rule can reference the defined queues by name. During the filtering component of .Nm pf.conf , the last referenced .Ar queue name is where any packets from .Ar pass rules will be queued, while for .Ar block rules it specifies where any resulting ICMP or TCP RST packets should be queued. The .Ar scheduler defines the algorithm used to decide which packets get delayed, dropped, or sent out immediately. There are three .Ar schedulers currently supported. .Bl -tag -width xxxx .It Ar cbq Class Based Queueing. .Ar Queues attached to an interface build a tree, thus each .Ar queue can have further child .Ar queues . Each queue can have a .Ar priority and a .Ar bandwidth assigned. .Ar Priority mainly controls the time packets take to get sent out, while .Ar bandwidth has primarily effects on throughput. .Ar cbq achieves both partitioning and sharing of link bandwidth by hierarchically structured classes. Each class has its own .Ar queue and is assigned its share of .Ar bandwidth . A child class can borrow bandwidth from its parent class as long as excess bandwidth is available (see the option .Ar borrow , below). .It Ar priq Priority Queueing. .Ar Queues are flat attached to the interface, thus, .Ar queues cannot have further child .Ar queues . Each .Ar queue has a unique .Ar priority assigned, ranging from 0 to 15. Packets in the .Ar queue with the highest .Ar priority are processed first. .It Ar hfsc Hierarchical Fair Service Curve. .Ar Queues attached to an interface build a tree, thus each .Ar queue can have further child .Ar queues . Each queue can have a .Ar priority and a .Ar bandwidth assigned. .Ar Priority mainly controls the time packets take to get sent out, while .Ar bandwidth primarily affects throughput. .Ar hfsc supports both link-sharing and guaranteed real-time services. It employs a service curve based QoS model, and its unique feature is an ability to decouple .Ar delay and .Ar bandwidth allocation. .El .Pp The interfaces on which queueing should be activated are declared using the .Ar altq on declaration. .Ar altq on has the following keywords: .Bl -tag -width xxxx .It Aq Ar interface Queueing is enabled on the named interface. .It Aq Ar scheduler Specifies which queueing scheduler to use. Currently supported values are .Ar cbq for Class Based Queueing, .Ar priq for Priority Queueing and .Ar hfsc for the Hierarchical Fair Service Curve scheduler. .It Ar bandwidth Aq Ar bw The maximum bitrate for all queues on an interface may be specified using the .Ar bandwidth keyword. The value can be specified as an absolute value or as a percentage of the interface bandwidth. When using an absolute value, the suffixes .Ar b , .Ar Kb , .Ar Mb , and .Ar Gb are used to represent bits, kilobits, megabits, and gigabits per second, respectively. The value must not exceed the interface bandwidth. If .Ar bandwidth is not specified, the interface bandwidth is used (but take note that some interfaces do not know their bandwidth, or can adapt their bandwidth rates). .It Ar qlimit Aq Ar limit The maximum number of packets held in the queue. The default is 50. .It Ar tbrsize Aq Ar size Adjusts the size, in bytes, of the token bucket regulator. If not specified, heuristics based on the interface bandwidth are used to determine the size. .It Ar queue Aq Ar list Defines a list of subqueues to create on an interface. .El .Pp In the following example, the interface dc0 should queue up to 5Mbps in four second-level queues using Class Based Queueing. Those four queues will be shown in a later example. .Bd -literal -offset indent altq on dc0 cbq bandwidth 5Mb queue { std, http, mail, ssh } .Ed .Pp Once interfaces are activated for queueing using the .Ar altq directive, a sequence of .Ar queue directives may be defined. The name associated with a .Ar queue must match a queue defined in the .Ar altq directive (e.g. mail), or, except for the .Ar priq .Ar scheduler , in a parent .Ar queue declaration. The following keywords can be used: .Bl -tag -width xxxx .It Ar on Aq Ar interface Specifies the interface the queue operates on. If not given, it operates on all matching interfaces. .It Ar bandwidth Aq Ar bw Specifies the maximum bitrate to be processed by the queue. This value must not exceed the value of the parent .Ar queue and can be specified as an absolute value or a percentage of the parent queue's bandwidth. If not specified, defaults to 100% of the parent queue's bandwidth. The .Ar priq scheduler does not support bandwidth specification. .It Ar priority Aq Ar level Between queues a priority level can be set. For .Ar cbq and .Ar hfsc , the range is 0 to 7 and for .Ar priq , the range is 0 to 15. The default for all is 1. .Ar Priq queues with a higher priority are always served first. .Ar Cbq and .Ar Hfsc queues with a higher priority are preferred in the case of overload. .It Ar qlimit Aq Ar limit The maximum number of packets held in the queue. The default is 50. .El .Pp The .Ar scheduler can get additional parameters with .Xo Aq Ar scheduler .Pf ( Aq Ar parameters ) . .Xc Parameters are as follows: .Bl -tag -width Fl .It Ar default Packets not matched by another queue are assigned to this one. Exactly one default queue is required. .It Ar red Enable RED (Random Early Detection) on this queue. RED drops packets with a probability proportional to the average queue length. .It Ar rio Enables RIO on this queue. RIO is RED with IN/OUT, thus running RED two times more than RIO would achieve the same effect. RIO is currently not supported in the GENERIC kernel. .It Ar ecn Enables ECN (Explicit Congestion Notification) on this queue. ECN implies RED. .El .Pp The .Ar cbq .Ar scheduler supports an additional option: .Bl -tag -width Fl .It Ar borrow The queue can borrow bandwidth from the parent. .El .Pp The .Ar hfsc .Ar scheduler supports some additional options: .Bl -tag -width Fl .It Ar realtime Aq Ar sc The minimum required bandwidth for the queue. .It Ar upperlimit Aq Ar sc The maximum allowed bandwidth for the queue. .It Ar linkshare Aq Ar sc The bandwidth share of a backlogged queue. .El .Pp .Aq Ar sc is an acronym for .Ar service curve . .Pp The format for service curve specifications is .Ar ( m1 , d , m2 ) . .Ar m2 controls the bandwidth assigned to the queue. .Ar m1 and .Ar d are optional and can be used to control the initial bandwidth assignment. For the first .Ar d milliseconds the queue gets the bandwidth given as .Ar m1 , afterwards the value given in .Ar m2 . .Pp Furthermore, with .Ar cbq and .Ar hfsc , child queues can be specified as in an .Ar altq declaration, thus building a tree of queues using a part of their parent's bandwidth. .Pp Packets can be assigned to queues based on filter rules by using the .Ar queue keyword. Normally only one .Ar queue is specified; when a second one is specified it will instead be used for packets which have a .Em TOS of .Em lowdelay and for TCP ACKs with no data payload. .Pp To continue the previous example, the examples below would specify the four referenced queues, plus a few child queues. Interactive .Xr ssh 1 sessions get priority over bulk transfers like .Xr scp 1 and .Xr sftp 1 . The queues may then be referenced by filtering rules (see .Sx PACKET FILTERING below). .Bd -literal queue std bandwidth 10% cbq(default) queue http bandwidth 60% priority 2 cbq(borrow red) \e { employees, developers } queue developers bandwidth 75% cbq(borrow) queue employees bandwidth 15% queue mail bandwidth 10% priority 0 cbq(borrow ecn) queue ssh bandwidth 20% cbq(borrow) { ssh_interactive, ssh_bulk } queue ssh_interactive bandwidth 50% priority 7 cbq(borrow) queue ssh_bulk bandwidth 50% priority 0 cbq(borrow) block return out on dc0 inet all queue std pass out on dc0 inet proto tcp from $developerhosts to any port 80 \e queue developers pass out on dc0 inet proto tcp from $employeehosts to any port 80 \e queue employees pass out on dc0 inet proto tcp from any to any port 22 \e queue(ssh_bulk, ssh_interactive) pass out on dc0 inet proto tcp from any to any port 25 \e queue mail .Ed .Sh TRANSLATION Translation rules modify either the source or destination address of the packets associated with a stateful connection. A stateful connection is automatically created to track packets matching such a rule as long as they are not blocked by the filtering section of .Nm pf.conf . The translation engine modifies the specified address and/or port in the packet, recalculates IP, TCP and UDP checksums as necessary, and passes it to the packet filter for evaluation. .Pp Since translation occurs before filtering the filter engine will see packets as they look after any addresses and ports have been translated. Filter rules will therefore have to filter based on the translated address and port number. Packets that match a translation rule are only automatically passed if the .Ar pass modifier is given, otherwise they are still subject to .Ar block and .Ar pass rules. .Pp The state entry created permits .Xr pf 4 to keep track of the original address for traffic associated with that state and correctly direct return traffic for that connection. .Pp Various types of translation are possible with pf: .Bl -tag -width xxxx .It Ar binat A .Ar binat rule specifies a bidirectional mapping between an external IP netblock and an internal IP netblock. .It Ar nat A .Ar nat rule specifies that IP addresses are to be changed as the packet traverses the given interface. This technique allows one or more IP addresses on the translating host to support network traffic for a larger range of machines on an "inside" network. Although in theory any IP address can be used on the inside, it is strongly recommended that one of the address ranges defined by RFC 1918 be used. These netblocks are: .Bd -literal 10.0.0.0 - 10.255.255.255 (all of net 10, i.e., 10/8) 172.16.0.0 - 172.31.255.255 (i.e., 172.16/12) 192.168.0.0 - 192.168.255.255 (i.e., 192.168/16) .Ed .It Pa rdr The packet is redirected to another destination and possibly a different port. .Ar rdr rules can optionally specify port ranges instead of single ports. rdr ... port 2000:2999 -\*(Gt ... port 4000 redirects ports 2000 to 2999 (inclusive) to port 4000. rdr ... port 2000:2999 -\*(Gt ... port 4000:* redirects port 2000 to 4000, 2001 to 4001, ..., 2999 to 4999. .El .Pp In addition to modifying the address, some translation rules may modify source or destination ports for .Xr tcp 4 or .Xr udp 4 connections; implicitly in the case of .Ar nat rules and explicitly in the case of .Ar rdr rules. Port numbers are never translated with a .Ar binat rule. .Pp Evaluation order of the translation rules is dependent on the type of the translation rules and of the direction of a packet. .Ar binat rules are always evaluated first. Then either the .Ar rdr rules are evaluated on an inbound packet or the .Ar nat rules on an outbound packet. Rules of the same type are evaluated in the same order in which they appear in the ruleset. The first matching rule decides what action is taken. .Pp The .Ar no option prefixed to a translation rule causes packets to remain untranslated, much in the same way as .Ar drop quick works in the packet filter (see below). If no rule matches the packet it is passed to the filter engine unmodified. .Pp Translation rules apply only to packets that pass through the specified interface, and if no interface is specified, translation is applied to packets on all interfaces. For instance, redirecting port 80 on an external interface to an internal web server will only work for connections originating from the outside. Connections to the address of the external interface from local hosts will not be redirected, since such packets do not actually pass through the external interface. Redirections cannot reflect packets back through the interface they arrive on, they can only be redirected to hosts connected to different interfaces or to the firewall itself. .Pp Note that redirecting external incoming connections to the loopback address, as in .Bd -literal -offset indent rdr on ne3 inet proto tcp to port smtp -\*(Gt 127.0.0.1 port spamd .Ed .Pp will effectively allow an external host to connect to daemons bound solely to the loopback address, circumventing the traditional blocking of such connections on a real interface. Unless this effect is desired, any of the local non-loopback addresses should be used as redirection target instead, which allows external connections only to daemons bound to this address or not bound to any address. .Pp See .Sx TRANSLATION EXAMPLES below. .Sh PACKET FILTERING .Xr pf 4 has the ability to .Ar block and .Ar pass packets based on attributes of their layer 3 (see .Xr ip 4 and .Xr ip6 4 ) and layer 4 (see .Xr icmp 4 , .Xr icmp6 4 , .Xr tcp 4 , .Xr udp 4 ) headers. In addition, packets may also be assigned to queues for the purpose of bandwidth control. .Pp For each packet processed by the packet filter, the filter rules are evaluated in sequential order, from first to last. The last matching rule decides what action is taken. If no rule matches the packet, the default action is to pass the packet. .Pp The following actions can be used in the filter: .Bl -tag -width xxxx .It Ar block The packet is blocked. There are a number of ways in which a .Ar block rule can behave when blocking a packet. The default behaviour is to .Ar drop packets silently, however this can be overridden or made explicit either globally, by setting the .Ar block-policy option, or on a per-rule basis with one of the following options: .Pp .Bl -tag -width xxxx -compact .It Ar drop The packet is silently dropped. .It Ar return-rst This applies only to .Xr tcp 4 packets, and issues a TCP RST which closes the connection. .It Ar return-icmp .It Ar return-icmp6 This causes ICMP messages to be returned for packets which match the rule. By default this is an ICMP UNREACHABLE message, however this can be overridden by specifying a message as a code or number. .It Ar return This causes a TCP RST to be returned for .Xr tcp 4 packets and an ICMP UNREACHABLE for UDP and other packets. .El .Pp Options returning ICMP packets currently have no effect if .Xr pf 4 operates on a .Xr if_bridge 4 , as the code to support this feature has not yet been implemented. .Pp The simplest mechanism to block everything by default and only pass packets that match explicit rules is specify a first filter rule of: .Bd -literal -offset indent block all .Ed .It Ar pass The packet is passed; state is created unless the .Ar no state option is specified. .El .Pp By default .Xr pf 4 filters packets statefully; the first time a packet matches a .Ar pass rule, a state entry is created; for subsequent packets the filter checks whether the packet matches any state. If it does, the packet is passed without evaluation of any rules. After the connection is closed or times out, the state entry is automatically removed. .Pp This has several advantages. For TCP connections, comparing a packet to a state involves checking its sequence numbers, as well as TCP timestamps if a .Ar scrub reassemble tcp rule applies to the connection. If these values are outside the narrow windows of expected values, the packet is dropped. This prevents spoofing attacks, such as when an attacker sends packets with a fake source address/port but does not know the connection's sequence numbers. Similarly, .Xr pf 4 knows how to match ICMP replies to states. For example, .Bd -literal -offset indent pass out inet proto icmp all icmp-type echoreq .Ed .Pp allows echo requests (such as those created by .Xr ping 8 ) out statefully, and matches incoming echo replies correctly to states. .Pp Also, looking up states is usually faster than evaluating rules. If there are 50 rules, all of them are evaluated sequentially in O(n). Even with 50000 states, only 16 comparisons are needed to match a state, since states are stored in a binary search tree that allows searches in O(log2 n). .Pp Furthermore, correct handling of ICMP error messages is critical to many protocols, particularly TCP. .Xr pf 4 matches ICMP error messages to the correct connection, checks them against connection parameters, and passes them if appropriate. For example if an ICMP source quench message referring to a stateful TCP connection arrives, it will be matched to the state and get passed. .Pp Finally, state tracking is required for .Ar nat , binat No and Ar rdr rules, in order to track address and port translations and reverse the translation on returning packets. .Pp .Xr pf 4 will also create state for other protocols which are effectively stateless by nature. UDP packets are matched to states using only host addresses and ports, and other protocols are matched to states using only the host addresses. .Pp If stateless filtering of individual packets is desired, the .Ar no state keyword can be used to specify that state will not be created if this is the last matching rule. A number of parameters can also be set to affect how .Xr pf 4 handles state tracking. See .Sx STATEFUL TRACKING OPTIONS below for further details. .Sh PARAMETERS The rule parameters specify the packets to which a rule applies. A packet always comes in on, or goes out through, one interface. Most parameters are optional. If a parameter is specified, the rule only applies to packets with matching attributes. Certain parameters can be expressed as lists, in which case .Xr pfctl 8 generates all needed rule combinations. .Bl -tag -width xxxx .It Ar in No or Ar out This rule applies to incoming or outgoing packets. If neither .Ar in nor .Ar out are specified, the rule will match packets in both directions. .It Ar log In addition to the action specified, a log message is generated. Only the packet that establishes the state is logged, unless the .Ar no state option is specified. The logged packets are sent to a .Xr pflog 4 interface, by default .Ar pflog0 . This interface is monitored by the .Xr pflogd 8 logging daemon, which dumps the logged packets to the file .Pa /var/log/pflog in .Xr pcap 3 binary format. .It Ar log (all) Used to force logging of all packets for a connection. This is not necessary when .Ar no state is explicitly specified. As with .Ar log , packets are logged to .Xr pflog 4 . .It Ar log (user) Logs the .Ux user ID of the user that owns the socket and the PID of the process that has the socket open where the packet is sourced from or destined to (depending on which socket is local). This is in addition to the normal information logged. .Pp Only the first packet logged via .Ar log (all, user) will have the user credentials logged when using stateful matching. .It Ar log (to Aq Ar interface ) Send logs to the specified .Xr pflog 4 interface instead of .Ar pflog0 . .It Ar quick If a packet matches a rule which has the .Ar quick option set, this rule is considered the last matching rule, and evaluation of subsequent rules is skipped. .It Ar on Aq Ar interface This rule applies only to packets coming in on, or going out through, this particular interface or interface group. For more information on interface groups, see the .Ic group keyword in .Xr ifconfig 8 . .It Aq Ar af This rule applies only to packets of this address family. Supported values are .Ar inet and .Ar inet6 . .It Ar proto Aq Ar protocol This rule applies only to packets of this protocol. Common protocols are .Xr icmp 4 , .Xr icmp6 4 , .Xr tcp 4 , and .Xr udp 4 . For a list of all the protocol name to number mappings used by .Xr pfctl 8 , see the file .Pa /etc/protocols . .It Xo .Ar from Aq Ar source .Ar port Aq Ar source .Ar os Aq Ar source .Ar to Aq Ar dest .Ar port Aq Ar dest .Xc This rule applies only to packets with the specified source and destination addresses and ports. .Pp Addresses can be specified in CIDR notation (matching netblocks), as symbolic host names, interface names or interface group names, or as any of the following keywords: .Pp .Bl -tag -width xxxxxxxxxxxxxx -compact .It Ar any Any address. .It Ar no-route Any address which is not currently routable. .It Ar urpf-failed Any source address that fails a unicast reverse path forwarding (URPF) check, i.e. packets coming in on an interface other than that which holds the route back to the packet's source address. .It Aq Ar table Any address that matches the given table. .El .Pp Ranges of addresses are specified by using the .Sq - operator. For instance: .Dq 10.1.1.10 - 10.1.1.12 means all addresses from 10.1.1.10 to 10.1.1.12, hence addresses 10.1.1.10, 10.1.1.11, and 10.1.1.12. .Pp Interface names and interface group names can have modifiers appended: .Pp .Bl -tag -width xxxxxxxxxxxx -compact .It Ar :network Translates to the network(s) attached to the interface. .It Ar :broadcast Translates to the interface's broadcast address(es). .It Ar :peer Translates to the point-to-point interface's peer address(es). .It Ar :0 Do not include interface aliases. .El .Pp Host names may also have the .Ar :0 option appended to restrict the name resolution to the first of each v4 and v6 address found. .Pp Host name resolution and interface to address translation are done at ruleset load-time. When the address of an interface (or host name) changes (under DHCP or PPP, for instance), the ruleset must be reloaded for the change to be reflected in the kernel. Surrounding the interface name (and optional modifiers) in parentheses changes this behaviour. When the interface name is surrounded by parentheses, the rule is automatically updated whenever the interface changes its address. The ruleset does not need to be reloaded. This is especially useful with .Ar nat . .Pp Ports can be specified either by number or by name. For example, port 80 can be specified as .Em www . For a list of all port name to number mappings used by .Xr pfctl 8 , see the file .Pa /etc/services . .Pp Ports and ranges of ports are specified by using these operators: .Bd -literal -offset indent = (equal) != (unequal) \*(Lt (less than) \*(Le (less than or equal) \*(Gt (greater than) \*(Ge (greater than or equal) : (range including boundaries) \*(Gt\*(Lt (range excluding boundaries) \*(Lt\*(Gt (except range) .Ed .Pp .Sq \*(Gt\*(Lt , .Sq \*(Lt\*(Gt and .Sq \&: are binary operators (they take two arguments). For instance: .Bl -tag -width Fl .It Ar port 2000:2004 means .Sq all ports \*(Ge 2000 and \*(Le 2004 , hence ports 2000, 2001, 2002, 2003 and 2004. .It Ar port 2000 \*(Gt\*(Lt 2004 means .Sq all ports \*(Gt 2000 and \*(Lt 2004 , hence ports 2001, 2002 and 2003. .It Ar port 2000 \*(Lt\*(Gt 2004 means .Sq all ports \*(Lt 2000 or \*(Gt 2004 , hence ports 1-1999 and 2005-65535. .El .Pp The operating system of the source host can be specified in the case of TCP rules with the .Ar OS modifier. See the .Sx OPERATING SYSTEM FINGERPRINTING section for more information. .Pp The host, port and OS specifications are optional, as in the following examples: .Bd -literal -offset indent pass in all pass in from any to any pass in proto tcp from any port \*(Le 1024 to any pass in proto tcp from any to any port 25 pass in proto tcp from 10.0.0.0/8 port \*(Gt 1024 \e to ! 10.1.2.3 port != ssh pass in proto tcp from any os "OpenBSD" .Ed .It Ar all This is equivalent to "from any to any". .It Ar group Aq Ar group Similar to .Ar user , this rule only applies to packets of sockets owned by the specified group. .It Ar user Aq Ar user This rule only applies to packets of sockets owned by the specified user. For outgoing connections initiated from the firewall, this is the user that opened the connection. For incoming connections to the firewall itself, this is the user that listens on the destination port. For forwarded connections, where the firewall is not a connection endpoint, the user and group are .Em unknown . .Pp All packets, both outgoing and incoming, of one connection are associated with the same user and group. Only TCP and UDP packets can be associated with users; for other protocols these parameters are ignored. .Pp User and group refer to the effective (as opposed to the real) IDs, in case the socket is created by a setuid/setgid process. User and group IDs are stored when a socket is created; when a process creates a listening socket as root (for instance, by binding to a privileged port) and subsequently changes to another user ID (to drop privileges), the credentials will remain root. .Pp User and group IDs can be specified as either numbers or names. The syntax is similar to the one for ports. The value .Em unknown matches packets of forwarded connections. .Em unknown can only be used with the operators .Cm = and .Cm != . Other constructs like .Cm user \*(Ge unknown are invalid. Forwarded packets with unknown user and group ID match only rules that explicitly compare against .Em unknown with the operators .Cm = or .Cm != . For instance .Cm user \*(Ge 0 does not match forwarded packets. The following example allows only selected users to open outgoing connections: .Bd -literal -offset indent block out proto { tcp, udp } all pass out proto { tcp, udp } all user { \*(Lt 1000, dhartmei } .Ed .It Xo Ar flags Aq Ar a .Pf / Ns Aq Ar b .No \*(Ba / Ns Aq Ar b .No \*(Ba any .Xc This rule only applies to TCP packets that have the flags .Aq Ar a set out of set .Aq Ar b . Flags not specified in .Aq Ar b are ignored. For stateful connections, the default is .Ar flags S/SA . To indicate that flags should not be checked at all, specify .Ar flags any . The flags are: (F)IN, (S)YN, (R)ST, (P)USH, (A)CK, (U)RG, (E)CE, and C(W)R. .Bl -tag -width Fl .It Ar flags S/S Flag SYN is set. The other flags are ignored. .It Ar flags S/SA This is the default setting for stateful connections. Out of SYN and ACK, exactly SYN may be set. SYN, SYN+PSH and SYN+RST match, but SYN+ACK, ACK and ACK+RST do not. This is more restrictive than the previous example. .It Ar flags /SFRA If the first set is not specified, it defaults to none. All of SYN, FIN, RST and ACK must be unset. .El .Pp Because .Ar flags S/SA is applied by default (unless .Ar no state is specified), only the initial SYN packet of a TCP handshake will create a state for a TCP connection. It is possible to be less restrictive, and allow state creation from intermediate .Pq non-SYN packets, by specifying .Ar flags any . This will cause .Xr pf 4 to synchronize to existing connections, for instance if one flushes the state table. However, states created from such intermediate packets may be missing connection details such as the TCP window scaling factor. States which modify the packet flow, such as those affected by .Ar nat , binat No or Ar rdr rules, .Ar modulate No or Ar synproxy state options, or scrubbed with .Ar reassemble tcp will also not be recoverable from intermediate packets. Such connections will stall and time out. .It Xo Ar icmp-type Aq Ar type .Ar code Aq Ar code .Xc .It Xo Ar icmp6-type Aq Ar type .Ar code Aq Ar code .Xc This rule only applies to ICMP or ICMPv6 packets with the specified type and code. Text names for ICMP types and codes are listed in .Xr icmp 4 and .Xr icmp6 4 . This parameter is only valid for rules that cover protocols ICMP or ICMP6. The protocol and the ICMP type indicator .Po .Ar icmp-type or .Ar icmp6-type .Pc must match. .It Xo Ar tos Aq Ar string .No \*(Ba Aq Ar number .Xc This rule applies to packets with the specified .Em TOS bits set. .Em TOS may be given as one of .Ar lowdelay , .Ar throughput , .Ar reliability , or as either hex or decimal. .Pp For example, the following rules are identical: .Bd -literal -offset indent pass all tos lowdelay pass all tos 0x10 pass all tos 16 .Ed .It Ar allow-opts By default, IPv4 packets with IP options or IPv6 packets with routing extension headers are blocked. When .Ar allow-opts is specified for a .Ar pass rule, packets that pass the filter based on that rule (last matching) do so even if they contain IP options or routing extension headers. For packets that match state, the rule that initially created the state is used. The implicit .Ar pass rule that is used when a packet does not match any rules does not allow IP options. .It Ar label Aq Ar string Adds a label (name) to the rule, which can be used to identify the rule. For instance, pfctl -s labels shows per-rule statistics for rules that have labels. .Pp The following macros can be used in labels: .Pp .Bl -tag -width $srcaddr -compact -offset indent .It Ar $if The interface. .It Ar $srcaddr The source IP address. .It Ar $dstaddr The destination IP address. .It Ar $srcport The source port specification. .It Ar $dstport The destination port specification. .It Ar $proto The protocol name. .It Ar $nr The rule number. .El .Pp For example: .Bd -literal -offset indent ips = \&"{ 1.2.3.4, 1.2.3.5 }\&" pass in proto tcp from any to $ips \e port \*(Gt 1023 label \&"$dstaddr:$dstport\&" .Ed .Pp expands to .Bd -literal -offset indent pass in inet proto tcp from any to 1.2.3.4 \e port \*(Gt 1023 label \&"1.2.3.4:\*(Gt1023\&" pass in inet proto tcp from any to 1.2.3.5 \e port \*(Gt 1023 label \&"1.2.3.5:\*(Gt1023\&" .Ed .Pp The macro expansion for the .Ar label directive occurs only at configuration file parse time, not during runtime. .It Xo Ar queue Aq Ar queue .No \*(Ba ( Aq Ar queue , .Aq Ar queue ) .Xc Packets matching this rule will be assigned to the specified queue. If two queues are given, packets which have a .Em TOS of .Em lowdelay and TCP ACKs with no data payload will be assigned to the second one. See .Sx QUEUEING for setup details. .Pp For example: .Bd -literal -offset indent pass in proto tcp to port 25 queue mail pass in proto tcp to port 22 queue(ssh_bulk, ssh_prio) .Ed .Pp .It Cm set prio Ar priority | Pq Ar priority , priority Packets matching this rule will be assigned a specific queueing priority. Priorities are assigned as integers 0 through 7. If the packet is transmitted on a .Xr vlan 4 interface, the queueing priority will be written as the priority code point in the 802.1Q VLAN header. If two priorities are given, packets which have a TOS of .Cm lowdelay and TCP ACKs with no data payload will be assigned to the second one. .Pp For example: .Bd -literal -offset indent pass in proto tcp to port 25 set prio 2 pass in proto tcp to port 22 set prio (2, 5) .Ed .Pp .It Ar tag Aq Ar string Packets matching this rule will be tagged with the specified string. The tag acts as an internal marker that can be used to identify these packets later on. This can be used, for example, to provide trust between interfaces and to determine if packets have been processed by translation rules. Tags are .Qq sticky , meaning that the packet will be tagged even if the rule is not the last matching rule. Further matching rules can replace the tag with a new one but will not remove a previously applied tag. A packet is only ever assigned one tag at a time. Packet tagging can be done during .Ar nat , .Ar rdr , or .Ar binat rules in addition to filter rules. Tags take the same macros as labels (see above). .It Ar tagged Aq Ar string Used with filter, translation or scrub rules to specify that packets must already be tagged with the given tag in order to match the rule. Inverse tag matching can also be done by specifying the .Cm !\& operator before the .Ar tagged keyword. .It Ar rtable Aq Ar number Used to select an alternate routing table for the routing lookup. Only effective before the route lookup happened, i.e. when filtering inbound. .It Xo Ar divert-to Aq Ar host .Ar port Aq Ar port .Xc Used to redirect packets to a local socket bound to .Ar host and .Ar port . The packets will not be modified, so .Xr getsockname 2 on the socket will return the original destination address of the packet. .It Ar divert-reply Used to receive replies for sockets that are bound to addresses which are not local to the machine. See .Xr setsockopt 2 for information on how to bind these sockets. .It Ar probability Aq Ar number A probability attribute can be attached to a rule, with a value set between 0 and 1, bounds not included. In that case, the rule will be honoured using the given probability value only. For example, the following rule will drop 20% of incoming ICMP packets: .Bd -literal -offset indent block in proto icmp probability 20% .Ed .It Ar prio Aq Ar number Only match packets which have the given queueing priority assigned. .Pp .El .Sh ROUTING If a packet matches a rule with a route option set, the packet filter will route the packet according to the type of route option. When such a rule creates state, the route option is also applied to all packets matching the same connection. .Bl -tag -width xxxx -.It Ar fastroute -The -.Ar fastroute -option does a normal route lookup to find the next hop for the packet. .It Ar route-to The .Ar route-to option routes the packet to the specified interface with an optional address for the next hop. When a .Ar route-to rule creates state, only packets that pass in the same direction as the filter rule specifies will be routed in this way. Packets passing in the opposite direction (replies) are not affected and are routed normally. .It Ar reply-to The .Ar reply-to option is similar to .Ar route-to , but routes packets that pass in the opposite direction (replies) to the specified interface. Opposite direction is only defined in the context of a state entry, and .Ar reply-to is useful only in rules that create state. It can be used on systems with multiple external connections to route all outgoing packets of a connection through the interface the incoming connection arrived through (symmetric routing enforcement). .It Ar dup-to The .Ar dup-to option creates a duplicate of the packet and routes it like .Ar route-to . The original packet gets routed as it normally would. .El .Sh POOL OPTIONS For .Ar nat and .Ar rdr rules, (as well as for the .Ar route-to , .Ar reply-to and .Ar dup-to rule options) for which there is a single redirection address which has a subnet mask smaller than 32 for IPv4 or 128 for IPv6 (more than one IP address), a variety of different methods for assigning this address can be used: .Bl -tag -width xxxx .It Ar bitmask The .Ar bitmask option applies the network portion of the redirection address to the address to be modified (source with .Ar nat , destination with .Ar rdr ) . .It Ar random The .Ar random option selects an address at random within the defined block of addresses. .It Ar source-hash The .Ar source-hash option uses a hash of the source address to determine the redirection address, ensuring that the redirection address is always the same for a given source. An optional key can be specified after this keyword either in hex or as a string; by default .Xr pfctl 8 randomly generates a key for source-hash every time the ruleset is reloaded. .It Ar round-robin The .Ar round-robin option loops through the redirection address(es). .Pp When more than one redirection address is specified, .Ar round-robin is the only permitted pool type. .It Ar static-port With .Ar nat rules, the .Ar static-port option prevents .Xr pf 4 from modifying the source port on TCP and UDP packets. .El .Pp Additionally, the .Ar sticky-address option can be specified to help ensure that multiple connections from the same source are mapped to the same redirection address. This option can be used with the .Ar random and .Ar round-robin pool options. Note that by default these associations are destroyed as soon as there are no longer states which refer to them; in order to make the mappings last beyond the lifetime of the states, increase the global options with .Ar set timeout src.track . See .Sx STATEFUL TRACKING OPTIONS for more ways to control the source tracking. .Sh STATE MODULATION Much of the security derived from TCP is attributable to how well the initial sequence numbers (ISNs) are chosen. Some popular stack implementations choose .Em very poor ISNs and thus are normally susceptible to ISN prediction exploits. By applying a .Ar modulate state rule to a TCP connection, .Xr pf 4 will create a high quality random sequence number for each connection endpoint. .Pp The .Ar modulate state directive implicitly keeps state on the rule and is only applicable to TCP connections. .Pp For instance: .Bd -literal -offset indent block all pass out proto tcp from any to any modulate state pass in proto tcp from any to any port 25 flags S/SFRA modulate state .Ed .Pp Note that modulated connections will not recover when the state table is lost (firewall reboot, flushing the state table, etc...). .Xr pf 4 will not be able to infer a connection again after the state table flushes the connection's modulator. When the state is lost, the connection may be left dangling until the respective endpoints time out the connection. It is possible on a fast local network for the endpoints to start an ACK storm while trying to resynchronize after the loss of the modulator. The default .Ar flags settings (or a more strict equivalent) should be used on .Ar modulate state rules to prevent ACK storms. .Pp Note that alternative methods are available to prevent loss of the state table and allow for firewall failover. See .Xr carp 4 and .Xr pfsync 4 for further information. .Sh SYN PROXY By default, .Xr pf 4 passes packets that are part of a .Xr tcp 4 handshake between the endpoints. The .Ar synproxy state option can be used to cause .Xr pf 4 itself to complete the handshake with the active endpoint, perform a handshake with the passive endpoint, and then forward packets between the endpoints. .Pp No packets are sent to the passive endpoint before the active endpoint has completed the handshake, hence so-called SYN floods with spoofed source addresses will not reach the passive endpoint, as the sender can't complete the handshake. .Pp The proxy is transparent to both endpoints, they each see a single connection from/to the other endpoint. .Xr pf 4 chooses random initial sequence numbers for both handshakes. Once the handshakes are completed, the sequence number modulators (see previous section) are used to translate further packets of the connection. .Ar synproxy state includes .Ar modulate state . .Pp Rules with .Ar synproxy will not work if .Xr pf 4 operates on a .Xr bridge 4 . .Pp Example: .Bd -literal -offset indent pass in proto tcp from any to any port www synproxy state .Ed .Sh STATEFUL TRACKING OPTIONS A number of options related to stateful tracking can be applied on a per-rule basis. .Ar keep state , .Ar modulate state and .Ar synproxy state support these options, and .Ar keep state must be specified explicitly to apply options to a rule. .Pp .Bl -tag -width xxxx -compact .It Ar max Aq Ar number Limits the number of concurrent states the rule may create. When this limit is reached, further packets that would create state will not match this rule until existing states time out. .It Ar no-sync Prevent state changes for states created by this rule from appearing on the .Xr pfsync 4 interface. .It Xo Aq Ar timeout .Aq Ar seconds .Xc Changes the timeout values used for states created by this rule. For a list of all valid timeout names, see .Sx OPTIONS above. .It Ar sloppy Uses a sloppy TCP connection tracker that does not check sequence numbers at all, which makes insertion and ICMP teardown attacks way easier. This is intended to be used in situations where one does not see all packets of a connection, e.g. in asymmetric routing situations. Cannot be used with modulate or synproxy state. .It Ar pflow States created by this rule are exported on the .Xr pflow 4 interface. .El .Pp Multiple options can be specified, separated by commas: .Bd -literal -offset indent pass in proto tcp from any to any \e port www keep state \e (max 100, source-track rule, max-src-nodes 75, \e max-src-states 3, tcp.established 60, tcp.closing 5) .Ed .Pp When the .Ar source-track keyword is specified, the number of states per source IP is tracked. .Pp .Bl -tag -width xxxx -compact .It Ar source-track rule The maximum number of states created by this rule is limited by the rule's .Ar max-src-nodes and .Ar max-src-states options. Only state entries created by this particular rule count toward the rule's limits. .It Ar source-track global The number of states created by all rules that use this option is limited. Each rule can specify different .Ar max-src-nodes and .Ar max-src-states options, however state entries created by any participating rule count towards each individual rule's limits. .El .Pp The following limits can be set: .Pp .Bl -tag -width xxxx -compact .It Ar max-src-nodes Aq Ar number Limits the maximum number of source addresses which can simultaneously have state table entries. .It Ar max-src-states Aq Ar number Limits the maximum number of simultaneous state entries that a single source address can create with this rule. .El .Pp For stateful TCP connections, limits on established connections (connections which have completed the TCP 3-way handshake) can also be enforced per source IP. .Pp .Bl -tag -width xxxx -compact .It Ar max-src-conn Aq Ar number Limits the maximum number of simultaneous TCP connections which have completed the 3-way handshake that a single host can make. .It Xo Ar max-src-conn-rate Aq Ar number .No / Aq Ar seconds .Xc Limit the rate of new connections over a time interval. The connection rate is an approximation calculated as a moving average. .El .Pp Because the 3-way handshake ensures that the source address is not being spoofed, more aggressive action can be taken based on these limits. With the .Ar overload Aq Ar table state option, source IP addresses which hit either of the limits on established connections will be added to the named table. This table can be used in the ruleset to block further activity from the offending host, redirect it to a tarpit process, or restrict its bandwidth. .Pp The optional .Ar flush keyword kills all states created by the matching rule which originate from the host which exceeds these limits. The .Ar global modifier to the flush command kills all states originating from the offending host, regardless of which rule created the state. .Pp For example, the following rules will protect the webserver against hosts making more than 100 connections in 10 seconds. Any host which connects faster than this rate will have its address added to the .Aq bad_hosts table and have all states originating from it flushed. Any new packets arriving from this host will be dropped unconditionally by the block rule. .Bd -literal -offset indent block quick from \*(Ltbad_hosts\*(Gt pass in on $ext_if proto tcp to $webserver port www keep state \e (max-src-conn-rate 100/10, overload \*(Ltbad_hosts\*(Gt flush global) .Ed .Sh OPERATING SYSTEM FINGERPRINTING Passive OS Fingerprinting is a mechanism to inspect nuances of a TCP connection's initial SYN packet and guess at the host's operating system. Unfortunately these nuances are easily spoofed by an attacker so the fingerprint is not useful in making security decisions. But the fingerprint is typically accurate enough to make policy decisions upon. .Pp The fingerprints may be specified by operating system class, by version, or by subtype/patchlevel. The class of an operating system is typically the vendor or genre and would be .Ox for the .Xr pf 4 firewall itself. The version of the oldest available .Ox release on the main FTP site would be 2.6 and the fingerprint would be written .Pp .Dl \&"OpenBSD 2.6\&" .Pp The subtype of an operating system is typically used to describe the patchlevel if that patch led to changes in the TCP stack behavior. In the case of .Ox , the only subtype is for a fingerprint that was normalized by the .Ar no-df scrub option and would be specified as .Pp .Dl \&"OpenBSD 3.3 no-df\&" .Pp Fingerprints for most popular operating systems are provided by .Xr pf.os 5 . Once .Xr pf 4 is running, a complete list of known operating system fingerprints may be listed by running: .Pp .Dl # pfctl -so .Pp Filter rules can enforce policy at any level of operating system specification assuming a fingerprint is present. Policy could limit traffic to approved operating systems or even ban traffic from hosts that aren't at the latest service pack. .Pp The .Ar unknown class can also be used as the fingerprint which will match packets for which no operating system fingerprint is known. .Pp Examples: .Bd -literal -offset indent pass out proto tcp from any os OpenBSD block out proto tcp from any os Doors block out proto tcp from any os "Doors PT" block out proto tcp from any os "Doors PT SP3" block out from any os "unknown" pass on lo0 proto tcp from any os "OpenBSD 3.3 lo0" .Ed .Pp Operating system fingerprinting is limited only to the TCP SYN packet. This means that it will not work on other protocols and will not match a currently established connection. .Pp Caveat: operating system fingerprints are occasionally wrong. There are three problems: an attacker can trivially craft his packets to appear as any operating system he chooses; an operating system patch could change the stack behavior and no fingerprints will match it until the database is updated; and multiple operating systems may have the same fingerprint. .Sh BLOCKING SPOOFED TRAFFIC "Spoofing" is the faking of IP addresses, typically for malicious purposes. The .Ar antispoof directive expands to a set of filter rules which will block all traffic with a source IP from the network(s) directly connected to the specified interface(s) from entering the system through any other interface. .Pp For example, the line .Bd -literal -offset indent antispoof for lo0 .Ed .Pp expands to .Bd -literal -offset indent block drop in on ! lo0 inet from 127.0.0.1/8 to any block drop in on ! lo0 inet6 from ::1 to any .Ed .Pp For non-loopback interfaces, there are additional rules to block incoming packets with a source IP address identical to the interface's IP(s). For example, assuming the interface wi0 had an IP address of 10.0.0.1 and a netmask of 255.255.255.0, the line .Bd -literal -offset indent antispoof for wi0 inet .Ed .Pp expands to .Bd -literal -offset indent block drop in on ! wi0 inet from 10.0.0.0/24 to any block drop in inet from 10.0.0.1 to any .Ed .Pp Caveat: Rules created by the .Ar antispoof directive interfere with packets sent over loopback interfaces to local addresses. One should pass these explicitly. .Sh FRAGMENT HANDLING The size of IP datagrams (packets) can be significantly larger than the maximum transmission unit (MTU) of the network. In cases when it is necessary or more efficient to send such large packets, the large packet will be fragmented into many smaller packets that will each fit onto the wire. Unfortunately for a firewalling device, only the first logical fragment will contain the necessary header information for the subprotocol that allows .Xr pf 4 to filter on things such as TCP ports or to perform NAT. .Pp Besides the use of .Ar scrub rules as described in .Sx TRAFFIC NORMALIZATION above, there are three options for handling fragments in the packet filter. .Pp One alternative is to filter individual fragments with filter rules. If no .Ar scrub rule applies to a fragment, it is passed to the filter. Filter rules with matching IP header parameters decide whether the fragment is passed or blocked, in the same way as complete packets are filtered. Without reassembly, fragments can only be filtered based on IP header fields (source/destination address, protocol), since subprotocol header fields are not available (TCP/UDP port numbers, ICMP code/type). The .Ar fragment option can be used to restrict filter rules to apply only to fragments, but not complete packets. Filter rules without the .Ar fragment option still apply to fragments, if they only specify IP header fields. For instance, the rule .Bd -literal -offset indent pass in proto tcp from any to any port 80 .Ed .Pp never applies to a fragment, even if the fragment is part of a TCP packet with destination port 80, because without reassembly this information is not available for each fragment. This also means that fragments cannot create new or match existing state table entries, which makes stateful filtering and address translation (NAT, redirection) for fragments impossible. .Pp It's also possible to reassemble only certain fragments by specifying source or destination addresses or protocols as parameters in .Ar scrub rules. .Pp In most cases, the benefits of reassembly outweigh the additional memory cost, and it's recommended to use .Ar scrub rules to reassemble all fragments via the .Ar fragment reassemble modifier. .Pp The memory allocated for fragment caching can be limited using .Xr pfctl 8 . Once this limit is reached, fragments that would have to be cached are dropped until other entries time out. The timeout value can also be adjusted. .Pp When forwarding reassembled IPv6 packets, pf refragments them with the original maximum fragment size. This allows the sender to determine the optimal fragment size by path MTU discovery. .Sh ANCHORS Besides the main ruleset, .Xr pfctl 8 can load rulesets into .Ar anchor attachment points. An .Ar anchor is a container that can hold rules, address tables, and other anchors. .Pp An .Ar anchor has a name which specifies the path where .Xr pfctl 8 can be used to access the anchor to perform operations on it, such as attaching child anchors to it or loading rules into it. Anchors may be nested, with components separated by .Sq / characters, similar to how file system hierarchies are laid out. The main ruleset is actually the default anchor, so filter and translation rules, for example, may also be contained in any anchor. .Pp An anchor can reference another .Ar anchor attachment point using the following kinds of rules: .Bl -tag -width xxxx .It Ar nat-anchor Aq Ar name Evaluates the .Ar nat rules in the specified .Ar anchor . .It Ar rdr-anchor Aq Ar name Evaluates the .Ar rdr rules in the specified .Ar anchor . .It Ar binat-anchor Aq Ar name Evaluates the .Ar binat rules in the specified .Ar anchor . .It Ar anchor Aq Ar name Evaluates the filter rules in the specified .Ar anchor . .It Xo Ar load anchor .Aq Ar name .Ar from Aq Ar file .Xc Loads the rules from the specified file into the anchor .Ar name . .El .Pp When evaluation of the main ruleset reaches an .Ar anchor rule, .Xr pf 4 will proceed to evaluate all rules specified in that anchor. .Pp Matching filter and translation rules marked with the .Ar quick option are final and abort the evaluation of the rules in other anchors and the main ruleset. If the .Ar anchor itself is marked with the .Ar quick option, ruleset evaluation will terminate when the anchor is exited if the packet is matched by any rule within the anchor. .Pp .Ar anchor rules are evaluated relative to the anchor in which they are contained. For example, all .Ar anchor rules specified in the main ruleset will reference anchor attachment points underneath the main ruleset, and .Ar anchor rules specified in a file loaded from a .Ar load anchor rule will be attached under that anchor point. .Pp Rules may be contained in .Ar anchor attachment points which do not contain any rules when the main ruleset is loaded, and later such anchors can be manipulated through .Xr pfctl 8 without reloading the main ruleset or other anchors. For example, .Bd -literal -offset indent ext_if = \&"kue0\&" block on $ext_if all anchor spam pass out on $ext_if all pass in on $ext_if proto tcp from any \e to $ext_if port smtp .Ed .Pp blocks all packets on the external interface by default, then evaluates all rules in the .Ar anchor named "spam", and finally passes all outgoing connections and incoming connections to port 25. .Bd -literal -offset indent # echo \&"block in quick from 1.2.3.4 to any\&" \&| \e pfctl -a spam -f - .Ed .Pp This loads a single rule into the .Ar anchor , which blocks all packets from a specific address. .Pp The anchor can also be populated by adding a .Ar load anchor rule after the .Ar anchor rule: .Bd -literal -offset indent anchor spam load anchor spam from "/etc/pf-spam.conf" .Ed .Pp When .Xr pfctl 8 loads .Nm pf.conf , it will also load all the rules from the file .Pa /etc/pf-spam.conf into the anchor. .Pp Optionally, .Ar anchor rules can specify packet filtering parameters using the same syntax as filter rules. When parameters are used, the .Ar anchor rule is only evaluated for matching packets. This allows conditional evaluation of anchors, like: .Bd -literal -offset indent block on $ext_if all anchor spam proto tcp from any to any port smtp pass out on $ext_if all pass in on $ext_if proto tcp from any to $ext_if port smtp .Ed .Pp The rules inside .Ar anchor spam are only evaluated for .Ar tcp packets with destination port 25. Hence, .Bd -literal -offset indent # echo \&"block in quick from 1.2.3.4 to any" \&| \e pfctl -a spam -f - .Ed .Pp will only block connections from 1.2.3.4 to port 25. .Pp Anchors may end with the asterisk .Pq Sq * character, which signifies that all anchors attached at that point should be evaluated in the alphabetical ordering of their anchor name. For example, .Bd -literal -offset indent anchor "spam/*" .Ed .Pp will evaluate each rule in each anchor attached to the .Li spam anchor. Note that it will only evaluate anchors that are directly attached to the .Li spam anchor, and will not descend to evaluate anchors recursively. .Pp Since anchors are evaluated relative to the anchor in which they are contained, there is a mechanism for accessing the parent and ancestor anchors of a given anchor. Similar to file system path name resolution, if the sequence .Dq .. appears as an anchor path component, the parent anchor of the current anchor in the path evaluation at that point will become the new current anchor. As an example, consider the following: .Bd -literal -offset indent # echo ' anchor "spam/allowed" ' | pfctl -f - # echo -e ' anchor "../banned" \en pass' | \e pfctl -a spam/allowed -f - .Ed .Pp Evaluation of the main ruleset will lead into the .Li spam/allowed anchor, which will evaluate the rules in the .Li spam/banned anchor, if any, before finally evaluating the .Ar pass rule. .Pp Filter rule .Ar anchors can also be loaded inline in the ruleset within a brace ('{' '}') delimited block. Brace delimited blocks may contain rules or other brace-delimited blocks. When anchors are loaded this way the anchor name becomes optional. .Bd -literal -offset indent anchor "external" on egress { block anchor out { pass proto tcp from any to port { 25, 80, 443 } } pass in proto tcp to any port 22 } .Ed .Pp Since the parser specification for anchor names is a string, any reference to an anchor name containing .Sq / characters will require double quote .Pq Sq \&" characters around the anchor name. .Sh TRANSLATION EXAMPLES This example maps incoming requests on port 80 to port 8080, on which a daemon is running (because, for example, it is not run as root, and therefore lacks permission to bind to port 80). .Bd -literal # use a macro for the interface name, so it can be changed easily ext_if = \&"ne3\&" # map daemon on 8080 to appear to be on 80 rdr on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 port 8080 .Ed .Pp If the .Ar pass modifier is given, packets matching the translation rule are passed without inspecting the filter rules: .Bd -literal rdr pass on $ext_if proto tcp from any to any port 80 -\*(Gt 127.0.0.1 \e port 8080 .Ed .Pp In the example below, vlan12 is configured as 192.168.168.1; the machine translates all packets coming from 192.168.168.0/24 to 204.92.77.111 when they are going out any interface except vlan12. This has the net effect of making traffic from the 192.168.168.0/24 network appear as though it is the Internet routable address 204.92.77.111 to nodes behind any interface on the router except for the nodes on vlan12. (Thus, 192.168.168.1 can talk to the 192.168.168.0/24 nodes.) .Bd -literal nat on ! vlan12 from 192.168.168.0/24 to any -\*(Gt 204.92.77.111 .Ed .Pp In the example below, the machine sits between a fake internal 144.19.74.* network, and a routable external IP of 204.92.77.100. The .Ar no nat rule excludes protocol AH from being translated. .Bd -literal # NO NAT no nat on $ext_if proto ah from 144.19.74.0/24 to any nat on $ext_if from 144.19.74.0/24 to any -\*(Gt 204.92.77.100 .Ed .Pp In the example below, packets bound for one specific server, as well as those generated by the sysadmins are not proxied; all other connections are. .Bd -literal # NO RDR no rdr on $int_if proto { tcp, udp } from any to $server port 80 no rdr on $int_if proto { tcp, udp } from $sysadmins to any port 80 rdr on $int_if proto { tcp, udp } from any to any port 80 -\*(Gt 127.0.0.1 \e port 80 .Ed .Pp This longer example uses both a NAT and a redirection. The external interface has the address 157.161.48.183. On localhost, we are running .Xr ftp-proxy 8 , waiting for FTP sessions to be redirected to it. The three mandatory anchors for .Xr ftp-proxy 8 are omitted from this example; see the .Xr ftp-proxy 8 manpage. .Bd -literal # NAT # Translate outgoing packets' source addresses (any protocol). # In this case, any address but the gateway's external address is mapped. nat on $ext_if inet from ! ($ext_if) to any -\*(Gt ($ext_if) # NAT PROXYING # Map outgoing packets' source port to an assigned proxy port instead of # an arbitrary port. # In this case, proxy outgoing isakmp with port 500 on the gateway. nat on $ext_if inet proto udp from any port = isakmp to any -\*(Gt ($ext_if) \e port 500 # BINAT # Translate outgoing packets' source address (any protocol). # Translate incoming packets' destination address to an internal machine # (bidirectional). binat on $ext_if from 10.1.2.150 to any -\*(Gt $ext_if # RDR # Translate incoming packets' destination addresses. # As an example, redirect a TCP and UDP port to an internal machine. rdr on $ext_if inet proto tcp from any to ($ext_if) port 8080 \e -\*(Gt 10.1.2.151 port 22 rdr on $ext_if inet proto udp from any to ($ext_if) port 8080 \e -\*(Gt 10.1.2.151 port 53 # RDR # Translate outgoing ftp control connections to send them to localhost # for proxying with ftp-proxy(8) running on port 8021. rdr on $int_if proto tcp from any to any port 21 -\*(Gt 127.0.0.1 port 8021 .Ed .Pp In this example, a NAT gateway is set up to translate internal addresses using a pool of public addresses (192.0.2.16/28) and to redirect incoming web server connections to a group of web servers on the internal network. .Bd -literal # NAT LOAD BALANCE # Translate outgoing packets' source addresses using an address pool. # A given source address is always translated to the same pool address by # using the source-hash keyword. nat on $ext_if inet from any to any -\*(Gt 192.0.2.16/28 source-hash # RDR ROUND ROBIN # Translate incoming web server connections to a group of web servers on # the internal network. rdr on $ext_if proto tcp from any to any port 80 \e -\*(Gt { 10.1.2.155, 10.1.2.160, 10.1.2.161 } round-robin .Ed .Sh FILTER EXAMPLES .Bd -literal # The external interface is kue0 # (157.161.48.183, the only routable address) # and the private network is 10.0.0.0/8, for which we are doing NAT. # use a macro for the interface name, so it can be changed easily ext_if = \&"kue0\&" # normalize all incoming traffic scrub in on $ext_if all fragment reassemble # block and log everything by default block return log on $ext_if all # block anything coming from source we have no back routes for block in from no-route to any # block packets whose ingress interface does not match the one in # the route back to their source address block in from urpf-failed to any # block and log outgoing packets that do not have our address as source, # they are either spoofed or something is misconfigured (NAT disabled, # for instance), we want to be nice and do not send out garbage. block out log quick on $ext_if from ! 157.161.48.183 to any # silently drop broadcasts (cable modem noise) block in quick on $ext_if from any to 255.255.255.255 # block and log incoming packets from reserved address space and invalid # addresses, they are either spoofed or misconfigured, we cannot reply to # them anyway (hence, no return-rst). block in log quick on $ext_if from { 10.0.0.0/8, 172.16.0.0/12, \e 192.168.0.0/16, 255.255.255.255/32 } to any # ICMP # pass out/in certain ICMP queries and keep state (ping) # state matching is done on host addresses and ICMP id (not type/code), # so replies (like 0/0 for 8/0) will match queries # ICMP error messages (which always refer to a TCP/UDP packet) are # handled by the TCP/UDP states pass on $ext_if inet proto icmp all icmp-type 8 code 0 # UDP # pass out all UDP connections and keep state pass out on $ext_if proto udp all # pass in certain UDP connections and keep state (DNS) pass in on $ext_if proto udp from any to any port domain # TCP # pass out all TCP connections and modulate state pass out on $ext_if proto tcp all modulate state # pass in certain TCP connections and keep state (SSH, SMTP, DNS, IDENT) pass in on $ext_if proto tcp from any to any port { ssh, smtp, domain, \e auth } # Do not allow Windows 9x SMTP connections since they are typically # a viral worm. Alternately we could limit these OSes to 1 connection each. block in on $ext_if proto tcp from any os {"Windows 95", "Windows 98"} \e to any port smtp # IPv6 # pass in/out all IPv6 traffic: note that we have to enable this in two # different ways, on both our physical interface and our tunnel pass quick on gif0 inet6 pass quick on $ext_if proto ipv6 # Packet Tagging # three interfaces: $int_if, $ext_if, and $wifi_if (wireless). NAT is # being done on $ext_if for all outgoing packets. tag packets in on # $int_if and pass those tagged packets out on $ext_if. all other # outgoing packets (i.e., packets from the wireless network) are only # permitted to access port 80. pass in on $int_if from any to any tag INTNET pass in on $wifi_if from any to any block out on $ext_if from any to any pass out quick on $ext_if tagged INTNET pass out on $ext_if proto tcp from any to any port 80 # tag incoming packets as they are redirected to spamd(8). use the tag # to pass those packets through the packet filter. rdr on $ext_if inet proto tcp from \*(Ltspammers\*(Gt to port smtp \e tag SPAMD -\*(Gt 127.0.0.1 port spamd block in on $ext_if pass in on $ext_if inet proto tcp tagged SPAMD .Ed .Sh GRAMMAR Syntax for .Nm in BNF: .Bd -literal line = ( option | pf-rule | nat-rule | binat-rule | rdr-rule | antispoof-rule | altq-rule | queue-rule | trans-anchors | anchor-rule | anchor-close | load-anchor | table-rule | include ) option = "set" ( [ "timeout" ( timeout | "{" timeout-list "}" ) ] | [ "ruleset-optimization" [ "none" | "basic" | "profile" ]] | [ "optimization" [ "default" | "normal" | "high-latency" | "satellite" | "aggressive" | "conservative" ] ] [ "limit" ( limit-item | "{" limit-list "}" ) ] | [ "loginterface" ( interface-name | "none" ) ] | [ "block-policy" ( "drop" | "return" ) ] | [ "state-policy" ( "if-bound" | "floating" ) ] [ "state-defaults" state-opts ] [ "require-order" ( "yes" | "no" ) ] [ "fingerprints" filename ] | [ "skip on" ifspec ] | [ "debug" ( "none" | "urgent" | "misc" | "loud" ) ] ) pf-rule = action [ ( "in" | "out" ) ] [ "log" [ "(" logopts ")"] ] [ "quick" ] - [ "on" ifspec ] [ "fastroute" | route ] [ af ] [ protospec ] + [ "on" ifspec ] [ route ] [ af ] [ protospec ] hosts [ filteropt-list ] logopts = logopt [ "," logopts ] logopt = "all" | "user" | "to" interface-name filteropt-list = filteropt-list filteropt | filteropt filteropt = user | group | flags | icmp-type | icmp6-type | "tos" tos | ( "no" | "keep" | "modulate" | "synproxy" ) "state" [ "(" state-opts ")" ] | "fragment" | "no-df" | "min-ttl" number | "set-tos" tos | "max-mss" number | "random-id" | "reassemble tcp" | fragmentation | "allow-opts" | "label" string | "tag" string | [ ! ] "tagged" string | "set prio" ( number | "(" number [ [ "," ] number ] ")" ) | "queue" ( string | "(" string [ [ "," ] string ] ")" ) | "rtable" number | "probability" number"%" | "prio" number nat-rule = [ "no" ] "nat" [ "pass" [ "log" [ "(" logopts ")" ] ] ] [ "on" ifspec ] [ af ] [ protospec ] hosts [ "tag" string ] [ "tagged" string ] [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" ) [ portspec ] [ pooltype ] [ "static-port" ] ] binat-rule = [ "no" ] "binat" [ "pass" [ "log" [ "(" logopts ")" ] ] ] [ "on" interface-name ] [ af ] [ "proto" ( proto-name | proto-number ) ] "from" address [ "/" mask-bits ] "to" ipspec [ "tag" string ] [ "tagged" string ] [ "-\*(Gt" address [ "/" mask-bits ] ] rdr-rule = [ "no" ] "rdr" [ "pass" [ "log" [ "(" logopts ")" ] ] ] [ "on" ifspec ] [ af ] [ protospec ] hosts [ "tag" string ] [ "tagged" string ] [ "-\*(Gt" ( redirhost | "{" redirhost-list "}" ) [ portspec ] [ pooltype ] ] antispoof-rule = "antispoof" [ "log" ] [ "quick" ] "for" ifspec [ af ] [ "label" string ] table-rule = "table" "\*(Lt" string "\*(Gt" [ tableopts-list ] tableopts-list = tableopts-list tableopts | tableopts tableopts = "persist" | "const" | "counters" | "file" string | "{" [ tableaddr-list ] "}" tableaddr-list = tableaddr-list [ "," ] tableaddr-spec | tableaddr-spec tableaddr-spec = [ "!" ] tableaddr [ "/" mask-bits ] tableaddr = hostname | ifspec | "self" | ipv4-dotted-quad | ipv6-coloned-hex altq-rule = "altq on" interface-name queueopts-list "queue" subqueue queue-rule = "queue" string [ "on" interface-name ] queueopts-list subqueue anchor-rule = "anchor" [ string ] [ ( "in" | "out" ) ] [ "on" ifspec ] [ af ] [ protospec ] [ hosts ] [ filteropt-list ] [ "{" ] anchor-close = "}" trans-anchors = ( "nat-anchor" | "rdr-anchor" | "binat-anchor" ) string [ "on" ifspec ] [ af ] [ "proto" ] [ protospec ] [ hosts ] load-anchor = "load anchor" string "from" filename queueopts-list = queueopts-list queueopts | queueopts queueopts = [ "bandwidth" bandwidth-spec ] | [ "qlimit" number ] | [ "tbrsize" number ] | [ "priority" number ] | [ schedulers ] schedulers = ( cbq-def | priq-def | hfsc-def ) bandwidth-spec = "number" ( "b" | "Kb" | "Mb" | "Gb" | "%" ) action = "pass" | "block" [ return ] | [ "no" ] "scrub" return = "drop" | "return" | "return-rst" [ "( ttl" number ")" ] | "return-icmp" [ "(" icmpcode [ [ "," ] icmp6code ] ")" ] | "return-icmp6" [ "(" icmp6code ")" ] icmpcode = ( icmp-code-name | icmp-code-number ) icmp6code = ( icmp6-code-name | icmp6-code-number ) ifspec = ( [ "!" ] ( interface-name | interface-group ) ) | "{" interface-list "}" interface-list = [ "!" ] ( interface-name | interface-group ) [ [ "," ] interface-list ] route = ( "route-to" | "reply-to" | "dup-to" ) ( routehost | "{" routehost-list "}" ) [ pooltype ] af = "inet" | "inet6" protospec = "proto" ( proto-name | proto-number | "{" proto-list "}" ) proto-list = ( proto-name | proto-number ) [ [ "," ] proto-list ] hosts = "all" | "from" ( "any" | "no-route" | "urpf-failed" | "self" | host | "{" host-list "}" ) [ port ] [ os ] "to" ( "any" | "no-route" | "self" | host | "{" host-list "}" ) [ port ] ipspec = "any" | host | "{" host-list "}" host = [ "!" ] ( address [ "/" mask-bits ] | "\*(Lt" string "\*(Gt" ) redirhost = address [ "/" mask-bits ] routehost = "(" interface-name [ address [ "/" mask-bits ] ] ")" address = ( interface-name | interface-group | "(" ( interface-name | interface-group ) ")" | hostname | ipv4-dotted-quad | ipv6-coloned-hex ) host-list = host [ [ "," ] host-list ] redirhost-list = redirhost [ [ "," ] redirhost-list ] routehost-list = routehost [ [ "," ] routehost-list ] port = "port" ( unary-op | binary-op | "{" op-list "}" ) portspec = "port" ( number | name ) [ ":" ( "*" | number | name ) ] os = "os" ( os-name | "{" os-list "}" ) user = "user" ( unary-op | binary-op | "{" op-list "}" ) group = "group" ( unary-op | binary-op | "{" op-list "}" ) unary-op = [ "=" | "!=" | "\*(Lt" | "\*(Le" | "\*(Gt" | "\*(Ge" ] ( name | number ) binary-op = number ( "\*(Lt\*(Gt" | "\*(Gt\*(Lt" | ":" ) number op-list = ( unary-op | binary-op ) [ [ "," ] op-list ] os-name = operating-system-name os-list = os-name [ [ "," ] os-list ] flags = "flags" ( [ flag-set ] "/" flag-set | "any" ) flag-set = [ "F" ] [ "S" ] [ "R" ] [ "P" ] [ "A" ] [ "U" ] [ "E" ] [ "W" ] icmp-type = "icmp-type" ( icmp-type-code | "{" icmp-list "}" ) icmp6-type = "icmp6-type" ( icmp-type-code | "{" icmp-list "}" ) icmp-type-code = ( icmp-type-name | icmp-type-number ) [ "code" ( icmp-code-name | icmp-code-number ) ] icmp-list = icmp-type-code [ [ "," ] icmp-list ] tos = ( "lowdelay" | "throughput" | "reliability" | [ "0x" ] number ) state-opts = state-opt [ [ "," ] state-opts ] state-opt = ( "max" number | "no-sync" | timeout | "sloppy" | "pflow" | "source-track" [ ( "rule" | "global" ) ] | "max-src-nodes" number | "max-src-states" number | "max-src-conn" number | "max-src-conn-rate" number "/" number | "overload" "\*(Lt" string "\*(Gt" [ "flush" ] | "if-bound" | "floating" ) fragmentation = [ "fragment reassemble" ] timeout-list = timeout [ [ "," ] timeout-list ] timeout = ( "tcp.first" | "tcp.opening" | "tcp.established" | "tcp.closing" | "tcp.finwait" | "tcp.closed" | "udp.first" | "udp.single" | "udp.multiple" | "icmp.first" | "icmp.error" | "other.first" | "other.single" | "other.multiple" | "frag" | "interval" | "src.track" | "adaptive.start" | "adaptive.end" ) number limit-list = limit-item [ [ "," ] limit-list ] limit-item = ( "states" | "frags" | "src-nodes" ) number pooltype = ( "bitmask" | "random" | "source-hash" [ ( hex-key | string-key ) ] | "round-robin" ) [ sticky-address ] subqueue = string | "{" queue-list "}" queue-list = string [ [ "," ] string ] cbq-def = "cbq" [ "(" cbq-opt [ [ "," ] cbq-opt ] ")" ] priq-def = "priq" [ "(" priq-opt [ [ "," ] priq-opt ] ")" ] hfsc-def = "hfsc" [ "(" hfsc-opt [ [ "," ] hfsc-opt ] ")" ] cbq-opt = ( "default" | "borrow" | "red" | "ecn" | "rio" ) priq-opt = ( "default" | "red" | "ecn" | "rio" ) hfsc-opt = ( "default" | "red" | "ecn" | "rio" | linkshare-sc | realtime-sc | upperlimit-sc ) linkshare-sc = "linkshare" sc-spec realtime-sc = "realtime" sc-spec upperlimit-sc = "upperlimit" sc-spec sc-spec = ( bandwidth-spec | "(" bandwidth-spec number bandwidth-spec ")" ) include = "include" filename .Ed .Sh FILES .Bl -tag -width "/etc/protocols" -compact .It Pa /etc/hosts Host name database. .It Pa /etc/pf.conf Default location of the ruleset file. .It Pa /etc/pf.os Default location of OS fingerprints. .It Pa /etc/protocols Protocol name database. .It Pa /etc/services Service name database. .El .Sh SEE ALSO .Xr altq 4 , .Xr carp 4 , .Xr icmp 4 , .Xr icmp6 4 , .Xr ip 4 , .Xr ip6 4 , .Xr pf 4 , .Xr pflow 4 , .Xr pfsync 4 , .Xr tcp 4 , .Xr udp 4 , .Xr hosts 5 , .Xr pf.os 5 , .Xr protocols 5 , .Xr services 5 , .Xr ftp-proxy 8 , .Xr pfctl 8 , .Xr pflogd 8 .Sh HISTORY The .Nm file format first appeared in .Ox 3.0 . Index: head/sys/netpfil/pf/pf.c =================================================================== --- head/sys/netpfil/pf/pf.c (revision 306683) +++ head/sys/netpfil/pf/pf.c (revision 306684) @@ -1,6645 +1,6618 @@ /*- * Copyright (c) 2001 Daniel Hartmeier * Copyright (c) 2002 - 2008 Henning Brauer * Copyright (c) 2012 Gleb Smirnoff * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * - Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * - 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. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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 * COPYRIGHT HOLDERS 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. * * Effort sponsored in part by the Defense Advanced Research Projects * Agency (DARPA) and Air Force Research Laboratory, Air Force * Materiel Command, USAF, under agreement number F30602-01-2-0537. * * $OpenBSD: pf.c,v 1.634 2009/02/27 12:37:45 henning Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_bpf.h" #include "opt_pf.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include /* XXX: only for DIR_IN/DIR_OUT */ #ifdef INET6 #include #include #include #include #include #include #include #endif /* INET6 */ #include #include #define DPFPRINTF(n, x) if (V_pf_status.debug >= (n)) printf x /* * Global variables */ /* state tables */ VNET_DEFINE(struct pf_altqqueue, pf_altqs[2]); VNET_DEFINE(struct pf_palist, pf_pabuf); VNET_DEFINE(struct pf_altqqueue *, pf_altqs_active); VNET_DEFINE(struct pf_altqqueue *, pf_altqs_inactive); VNET_DEFINE(struct pf_kstatus, pf_status); VNET_DEFINE(u_int32_t, ticket_altqs_active); VNET_DEFINE(u_int32_t, ticket_altqs_inactive); VNET_DEFINE(int, altqs_inactive_open); VNET_DEFINE(u_int32_t, ticket_pabuf); VNET_DEFINE(MD5_CTX, pf_tcp_secret_ctx); #define V_pf_tcp_secret_ctx VNET(pf_tcp_secret_ctx) VNET_DEFINE(u_char, pf_tcp_secret[16]); #define V_pf_tcp_secret VNET(pf_tcp_secret) VNET_DEFINE(int, pf_tcp_secret_init); #define V_pf_tcp_secret_init VNET(pf_tcp_secret_init) VNET_DEFINE(int, pf_tcp_iss_off); #define V_pf_tcp_iss_off VNET(pf_tcp_iss_off) /* * Queue for pf_intr() sends. */ static MALLOC_DEFINE(M_PFTEMP, "pf_temp", "pf(4) temporary allocations"); struct pf_send_entry { STAILQ_ENTRY(pf_send_entry) pfse_next; struct mbuf *pfse_m; enum { PFSE_IP, PFSE_IP6, PFSE_ICMP, PFSE_ICMP6, } pfse_type; struct { int type; int code; int mtu; } icmpopts; }; STAILQ_HEAD(pf_send_head, pf_send_entry); static VNET_DEFINE(struct pf_send_head, pf_sendqueue); #define V_pf_sendqueue VNET(pf_sendqueue) static struct mtx pf_sendqueue_mtx; MTX_SYSINIT(pf_sendqueue_mtx, &pf_sendqueue_mtx, "pf send queue", MTX_DEF); #define PF_SENDQ_LOCK() mtx_lock(&pf_sendqueue_mtx) #define PF_SENDQ_UNLOCK() mtx_unlock(&pf_sendqueue_mtx) /* * Queue for pf_overload_task() tasks. */ struct pf_overload_entry { SLIST_ENTRY(pf_overload_entry) next; struct pf_addr addr; sa_family_t af; uint8_t dir; struct pf_rule *rule; }; SLIST_HEAD(pf_overload_head, pf_overload_entry); static VNET_DEFINE(struct pf_overload_head, pf_overloadqueue); #define V_pf_overloadqueue VNET(pf_overloadqueue) static VNET_DEFINE(struct task, pf_overloadtask); #define V_pf_overloadtask VNET(pf_overloadtask) static struct mtx pf_overloadqueue_mtx; MTX_SYSINIT(pf_overloadqueue_mtx, &pf_overloadqueue_mtx, "pf overload/flush queue", MTX_DEF); #define PF_OVERLOADQ_LOCK() mtx_lock(&pf_overloadqueue_mtx) #define PF_OVERLOADQ_UNLOCK() mtx_unlock(&pf_overloadqueue_mtx) VNET_DEFINE(struct pf_rulequeue, pf_unlinked_rules); struct mtx pf_unlnkdrules_mtx; MTX_SYSINIT(pf_unlnkdrules_mtx, &pf_unlnkdrules_mtx, "pf unlinked rules", MTX_DEF); static VNET_DEFINE(uma_zone_t, pf_sources_z); #define V_pf_sources_z VNET(pf_sources_z) uma_zone_t pf_mtag_z; VNET_DEFINE(uma_zone_t, pf_state_z); VNET_DEFINE(uma_zone_t, pf_state_key_z); VNET_DEFINE(uint64_t, pf_stateid[MAXCPU]); #define PFID_CPUBITS 8 #define PFID_CPUSHIFT (sizeof(uint64_t) * NBBY - PFID_CPUBITS) #define PFID_CPUMASK ((uint64_t)((1 << PFID_CPUBITS) - 1) << PFID_CPUSHIFT) #define PFID_MAXID (~PFID_CPUMASK) CTASSERT((1 << PFID_CPUBITS) >= MAXCPU); static void pf_src_tree_remove_state(struct pf_state *); static void pf_init_threshold(struct pf_threshold *, u_int32_t, u_int32_t); static void pf_add_threshold(struct pf_threshold *); static int pf_check_threshold(struct pf_threshold *); static void pf_change_ap(struct mbuf *, struct pf_addr *, u_int16_t *, u_int16_t *, u_int16_t *, struct pf_addr *, u_int16_t, u_int8_t, sa_family_t); static int pf_modulate_sack(struct mbuf *, int, struct pf_pdesc *, struct tcphdr *, struct pf_state_peer *); static void pf_change_icmp(struct pf_addr *, u_int16_t *, struct pf_addr *, struct pf_addr *, u_int16_t, u_int16_t *, u_int16_t *, u_int16_t *, u_int16_t *, u_int8_t, sa_family_t); static void pf_send_tcp(struct mbuf *, const struct pf_rule *, sa_family_t, const struct pf_addr *, const struct pf_addr *, u_int16_t, u_int16_t, u_int32_t, u_int32_t, u_int8_t, u_int16_t, u_int16_t, u_int8_t, int, u_int16_t, struct ifnet *); static void pf_send_icmp(struct mbuf *, u_int8_t, u_int8_t, sa_family_t, struct pf_rule *); static void pf_detach_state(struct pf_state *); static int pf_state_key_attach(struct pf_state_key *, struct pf_state_key *, struct pf_state *); static void pf_state_key_detach(struct pf_state *, int); static int pf_state_key_ctor(void *, int, void *, int); static u_int32_t pf_tcp_iss(struct pf_pdesc *); static int pf_test_rule(struct pf_rule **, struct pf_state **, int, struct pfi_kif *, struct mbuf *, int, struct pf_pdesc *, struct pf_rule **, struct pf_ruleset **, struct inpcb *); static int pf_create_state(struct pf_rule *, struct pf_rule *, struct pf_rule *, struct pf_pdesc *, struct pf_src_node *, struct pf_state_key *, struct pf_state_key *, struct mbuf *, int, u_int16_t, u_int16_t, int *, struct pfi_kif *, struct pf_state **, int, u_int16_t, u_int16_t, int); static int pf_test_fragment(struct pf_rule **, int, struct pfi_kif *, struct mbuf *, void *, struct pf_pdesc *, struct pf_rule **, struct pf_ruleset **); static int pf_tcp_track_full(struct pf_state_peer *, struct pf_state_peer *, struct pf_state **, struct pfi_kif *, struct mbuf *, int, struct pf_pdesc *, u_short *, int *); static int pf_tcp_track_sloppy(struct pf_state_peer *, struct pf_state_peer *, struct pf_state **, struct pf_pdesc *, u_short *); static int pf_test_state_tcp(struct pf_state **, int, struct pfi_kif *, struct mbuf *, int, void *, struct pf_pdesc *, u_short *); static int pf_test_state_udp(struct pf_state **, int, struct pfi_kif *, struct mbuf *, int, void *, struct pf_pdesc *); static int pf_test_state_icmp(struct pf_state **, int, struct pfi_kif *, struct mbuf *, int, void *, struct pf_pdesc *, u_short *); static int pf_test_state_other(struct pf_state **, int, struct pfi_kif *, struct mbuf *, struct pf_pdesc *); static u_int8_t pf_get_wscale(struct mbuf *, int, u_int16_t, sa_family_t); static u_int16_t pf_get_mss(struct mbuf *, int, u_int16_t, sa_family_t); static u_int16_t pf_calc_mss(struct pf_addr *, sa_family_t, int, u_int16_t); static int pf_check_proto_cksum(struct mbuf *, int, int, u_int8_t, sa_family_t); static void pf_print_state_parts(struct pf_state *, struct pf_state_key *, struct pf_state_key *); static int pf_addr_wrap_neq(struct pf_addr_wrap *, struct pf_addr_wrap *); static struct pf_state *pf_find_state(struct pfi_kif *, struct pf_state_key_cmp *, u_int); static int pf_src_connlimit(struct pf_state **); static void pf_overload_task(void *v, int pending); static int pf_insert_src_node(struct pf_src_node **, struct pf_rule *, struct pf_addr *, sa_family_t); static u_int pf_purge_expired_states(u_int, int); static void pf_purge_unlinked_rules(void); static int pf_mtag_uminit(void *, int, int); static void pf_mtag_free(struct m_tag *); #ifdef INET static void pf_route(struct mbuf **, struct pf_rule *, int, struct ifnet *, struct pf_state *, struct pf_pdesc *); #endif /* INET */ #ifdef INET6 static void pf_change_a6(struct pf_addr *, u_int16_t *, struct pf_addr *, u_int8_t); static void pf_route6(struct mbuf **, struct pf_rule *, int, struct ifnet *, struct pf_state *, struct pf_pdesc *); #endif /* INET6 */ int in4_cksum(struct mbuf *m, u_int8_t nxt, int off, int len); extern int pf_end_threads; VNET_DEFINE(struct pf_limit, pf_limits[PF_LIMIT_MAX]); #define PACKET_LOOPED(pd) ((pd)->pf_mtag && \ (pd)->pf_mtag->flags & PF_PACKET_LOOPED) #define STATE_LOOKUP(i, k, d, s, pd) \ do { \ (s) = pf_find_state((i), (k), (d)); \ if ((s) == NULL) \ return (PF_DROP); \ if (PACKET_LOOPED(pd)) \ return (PF_PASS); \ if ((d) == PF_OUT && \ (((s)->rule.ptr->rt == PF_ROUTETO && \ (s)->rule.ptr->direction == PF_OUT) || \ ((s)->rule.ptr->rt == PF_REPLYTO && \ (s)->rule.ptr->direction == PF_IN)) && \ (s)->rt_kif != NULL && \ (s)->rt_kif != (i)) \ return (PF_PASS); \ } while (0) #define BOUND_IFACE(r, k) \ ((r)->rule_flag & PFRULE_IFBOUND) ? (k) : V_pfi_all #define STATE_INC_COUNTERS(s) \ do { \ counter_u64_add(s->rule.ptr->states_cur, 1); \ counter_u64_add(s->rule.ptr->states_tot, 1); \ if (s->anchor.ptr != NULL) { \ counter_u64_add(s->anchor.ptr->states_cur, 1); \ counter_u64_add(s->anchor.ptr->states_tot, 1); \ } \ if (s->nat_rule.ptr != NULL) { \ counter_u64_add(s->nat_rule.ptr->states_cur, 1);\ counter_u64_add(s->nat_rule.ptr->states_tot, 1);\ } \ } while (0) #define STATE_DEC_COUNTERS(s) \ do { \ if (s->nat_rule.ptr != NULL) \ counter_u64_add(s->nat_rule.ptr->states_cur, -1);\ if (s->anchor.ptr != NULL) \ counter_u64_add(s->anchor.ptr->states_cur, -1); \ counter_u64_add(s->rule.ptr->states_cur, -1); \ } while (0) static MALLOC_DEFINE(M_PFHASH, "pf_hash", "pf(4) hash header structures"); VNET_DEFINE(struct pf_keyhash *, pf_keyhash); VNET_DEFINE(struct pf_idhash *, pf_idhash); VNET_DEFINE(struct pf_srchash *, pf_srchash); SYSCTL_NODE(_net, OID_AUTO, pf, CTLFLAG_RW, 0, "pf(4)"); u_long pf_hashmask; u_long pf_srchashmask; static u_long pf_hashsize; static u_long pf_srchashsize; SYSCTL_ULONG(_net_pf, OID_AUTO, states_hashsize, CTLFLAG_RDTUN, &pf_hashsize, 0, "Size of pf(4) states hashtable"); SYSCTL_ULONG(_net_pf, OID_AUTO, source_nodes_hashsize, CTLFLAG_RDTUN, &pf_srchashsize, 0, "Size of pf(4) source nodes hashtable"); VNET_DEFINE(void *, pf_swi_cookie); VNET_DEFINE(uint32_t, pf_hashseed); #define V_pf_hashseed VNET(pf_hashseed) int pf_addr_cmp(struct pf_addr *a, struct pf_addr *b, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: if (a->addr32[0] > b->addr32[0]) return (1); if (a->addr32[0] < b->addr32[0]) return (-1); break; #endif /* INET */ #ifdef INET6 case AF_INET6: if (a->addr32[3] > b->addr32[3]) return (1); if (a->addr32[3] < b->addr32[3]) return (-1); if (a->addr32[2] > b->addr32[2]) return (1); if (a->addr32[2] < b->addr32[2]) return (-1); if (a->addr32[1] > b->addr32[1]) return (1); if (a->addr32[1] < b->addr32[1]) return (-1); if (a->addr32[0] > b->addr32[0]) return (1); if (a->addr32[0] < b->addr32[0]) return (-1); break; #endif /* INET6 */ default: panic("%s: unknown address family %u", __func__, af); } return (0); } static __inline uint32_t pf_hashkey(struct pf_state_key *sk) { uint32_t h; h = murmur3_32_hash32((uint32_t *)sk, sizeof(struct pf_state_key_cmp)/sizeof(uint32_t), V_pf_hashseed); return (h & pf_hashmask); } static __inline uint32_t pf_hashsrc(struct pf_addr *addr, sa_family_t af) { uint32_t h; switch (af) { case AF_INET: h = murmur3_32_hash32((uint32_t *)&addr->v4, sizeof(addr->v4)/sizeof(uint32_t), V_pf_hashseed); break; case AF_INET6: h = murmur3_32_hash32((uint32_t *)&addr->v6, sizeof(addr->v6)/sizeof(uint32_t), V_pf_hashseed); break; default: panic("%s: unknown address family %u", __func__, af); } return (h & pf_srchashmask); } #ifdef ALTQ static int pf_state_hash(struct pf_state *s) { u_int32_t hv = (intptr_t)s / sizeof(*s); hv ^= crc32(&s->src, sizeof(s->src)); hv ^= crc32(&s->dst, sizeof(s->dst)); if (hv == 0) hv = 1; return (hv); } #endif #ifdef INET6 void pf_addrcpy(struct pf_addr *dst, struct pf_addr *src, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: dst->addr32[0] = src->addr32[0]; break; #endif /* INET */ case AF_INET6: dst->addr32[0] = src->addr32[0]; dst->addr32[1] = src->addr32[1]; dst->addr32[2] = src->addr32[2]; dst->addr32[3] = src->addr32[3]; break; } } #endif /* INET6 */ static void pf_init_threshold(struct pf_threshold *threshold, u_int32_t limit, u_int32_t seconds) { threshold->limit = limit * PF_THRESHOLD_MULT; threshold->seconds = seconds; threshold->count = 0; threshold->last = time_uptime; } static void pf_add_threshold(struct pf_threshold *threshold) { u_int32_t t = time_uptime, diff = t - threshold->last; if (diff >= threshold->seconds) threshold->count = 0; else threshold->count -= threshold->count * diff / threshold->seconds; threshold->count += PF_THRESHOLD_MULT; threshold->last = t; } static int pf_check_threshold(struct pf_threshold *threshold) { return (threshold->count > threshold->limit); } static int pf_src_connlimit(struct pf_state **state) { struct pf_overload_entry *pfoe; int bad = 0; PF_STATE_LOCK_ASSERT(*state); (*state)->src_node->conn++; (*state)->src.tcp_est = 1; pf_add_threshold(&(*state)->src_node->conn_rate); if ((*state)->rule.ptr->max_src_conn && (*state)->rule.ptr->max_src_conn < (*state)->src_node->conn) { counter_u64_add(V_pf_status.lcounters[LCNT_SRCCONN], 1); bad++; } if ((*state)->rule.ptr->max_src_conn_rate.limit && pf_check_threshold(&(*state)->src_node->conn_rate)) { counter_u64_add(V_pf_status.lcounters[LCNT_SRCCONNRATE], 1); bad++; } if (!bad) return (0); /* Kill this state. */ (*state)->timeout = PFTM_PURGE; (*state)->src.state = (*state)->dst.state = TCPS_CLOSED; if ((*state)->rule.ptr->overload_tbl == NULL) return (1); /* Schedule overloading and flushing task. */ pfoe = malloc(sizeof(*pfoe), M_PFTEMP, M_NOWAIT); if (pfoe == NULL) return (1); /* too bad :( */ bcopy(&(*state)->src_node->addr, &pfoe->addr, sizeof(pfoe->addr)); pfoe->af = (*state)->key[PF_SK_WIRE]->af; pfoe->rule = (*state)->rule.ptr; pfoe->dir = (*state)->direction; PF_OVERLOADQ_LOCK(); SLIST_INSERT_HEAD(&V_pf_overloadqueue, pfoe, next); PF_OVERLOADQ_UNLOCK(); taskqueue_enqueue(taskqueue_swi, &V_pf_overloadtask); return (1); } static void pf_overload_task(void *v, int pending) { struct pf_overload_head queue; struct pfr_addr p; struct pf_overload_entry *pfoe, *pfoe1; uint32_t killed = 0; CURVNET_SET((struct vnet *)v); PF_OVERLOADQ_LOCK(); queue = V_pf_overloadqueue; SLIST_INIT(&V_pf_overloadqueue); PF_OVERLOADQ_UNLOCK(); bzero(&p, sizeof(p)); SLIST_FOREACH(pfoe, &queue, next) { counter_u64_add(V_pf_status.lcounters[LCNT_OVERLOAD_TABLE], 1); if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("%s: blocking address ", __func__); pf_print_host(&pfoe->addr, 0, pfoe->af); printf("\n"); } p.pfra_af = pfoe->af; switch (pfoe->af) { #ifdef INET case AF_INET: p.pfra_net = 32; p.pfra_ip4addr = pfoe->addr.v4; break; #endif #ifdef INET6 case AF_INET6: p.pfra_net = 128; p.pfra_ip6addr = pfoe->addr.v6; break; #endif } PF_RULES_WLOCK(); pfr_insert_kentry(pfoe->rule->overload_tbl, &p, time_second); PF_RULES_WUNLOCK(); } /* * Remove those entries, that don't need flushing. */ SLIST_FOREACH_SAFE(pfoe, &queue, next, pfoe1) if (pfoe->rule->flush == 0) { SLIST_REMOVE(&queue, pfoe, pf_overload_entry, next); free(pfoe, M_PFTEMP); } else counter_u64_add( V_pf_status.lcounters[LCNT_OVERLOAD_FLUSH], 1); /* If nothing to flush, return. */ if (SLIST_EMPTY(&queue)) { CURVNET_RESTORE(); return; } for (int i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; struct pf_state_key *sk; struct pf_state *s; PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { sk = s->key[PF_SK_WIRE]; SLIST_FOREACH(pfoe, &queue, next) if (sk->af == pfoe->af && ((pfoe->rule->flush & PF_FLUSH_GLOBAL) || pfoe->rule == s->rule.ptr) && ((pfoe->dir == PF_OUT && PF_AEQ(&pfoe->addr, &sk->addr[1], sk->af)) || (pfoe->dir == PF_IN && PF_AEQ(&pfoe->addr, &sk->addr[0], sk->af)))) { s->timeout = PFTM_PURGE; s->src.state = s->dst.state = TCPS_CLOSED; killed++; } } PF_HASHROW_UNLOCK(ih); } SLIST_FOREACH_SAFE(pfoe, &queue, next, pfoe1) free(pfoe, M_PFTEMP); if (V_pf_status.debug >= PF_DEBUG_MISC) printf("%s: %u states killed", __func__, killed); CURVNET_RESTORE(); } /* * Can return locked on failure, so that we can consistently * allocate and insert a new one. */ struct pf_src_node * pf_find_src_node(struct pf_addr *src, struct pf_rule *rule, sa_family_t af, int returnlocked) { struct pf_srchash *sh; struct pf_src_node *n; counter_u64_add(V_pf_status.scounters[SCNT_SRC_NODE_SEARCH], 1); sh = &V_pf_srchash[pf_hashsrc(src, af)]; PF_HASHROW_LOCK(sh); LIST_FOREACH(n, &sh->nodes, entry) if (n->rule.ptr == rule && n->af == af && ((af == AF_INET && n->addr.v4.s_addr == src->v4.s_addr) || (af == AF_INET6 && bcmp(&n->addr, src, sizeof(*src)) == 0))) break; if (n != NULL) { n->states++; PF_HASHROW_UNLOCK(sh); } else if (returnlocked == 0) PF_HASHROW_UNLOCK(sh); return (n); } static int pf_insert_src_node(struct pf_src_node **sn, struct pf_rule *rule, struct pf_addr *src, sa_family_t af) { KASSERT((rule->rule_flag & PFRULE_RULESRCTRACK || rule->rpool.opts & PF_POOL_STICKYADDR), ("%s for non-tracking rule %p", __func__, rule)); if (*sn == NULL) *sn = pf_find_src_node(src, rule, af, 1); if (*sn == NULL) { struct pf_srchash *sh = &V_pf_srchash[pf_hashsrc(src, af)]; PF_HASHROW_ASSERT(sh); if (!rule->max_src_nodes || counter_u64_fetch(rule->src_nodes) < rule->max_src_nodes) (*sn) = uma_zalloc(V_pf_sources_z, M_NOWAIT | M_ZERO); else counter_u64_add(V_pf_status.lcounters[LCNT_SRCNODES], 1); if ((*sn) == NULL) { PF_HASHROW_UNLOCK(sh); return (-1); } pf_init_threshold(&(*sn)->conn_rate, rule->max_src_conn_rate.limit, rule->max_src_conn_rate.seconds); (*sn)->af = af; (*sn)->rule.ptr = rule; PF_ACPY(&(*sn)->addr, src, af); LIST_INSERT_HEAD(&sh->nodes, *sn, entry); (*sn)->creation = time_uptime; (*sn)->ruletype = rule->action; (*sn)->states = 1; if ((*sn)->rule.ptr != NULL) counter_u64_add((*sn)->rule.ptr->src_nodes, 1); PF_HASHROW_UNLOCK(sh); counter_u64_add(V_pf_status.scounters[SCNT_SRC_NODE_INSERT], 1); } else { if (rule->max_src_states && (*sn)->states >= rule->max_src_states) { counter_u64_add(V_pf_status.lcounters[LCNT_SRCSTATES], 1); return (-1); } } return (0); } void pf_unlink_src_node(struct pf_src_node *src) { PF_HASHROW_ASSERT(&V_pf_srchash[pf_hashsrc(&src->addr, src->af)]); LIST_REMOVE(src, entry); if (src->rule.ptr) counter_u64_add(src->rule.ptr->src_nodes, -1); } u_int pf_free_src_nodes(struct pf_src_node_list *head) { struct pf_src_node *sn, *tmp; u_int count = 0; LIST_FOREACH_SAFE(sn, head, entry, tmp) { uma_zfree(V_pf_sources_z, sn); count++; } counter_u64_add(V_pf_status.scounters[SCNT_SRC_NODE_REMOVALS], count); return (count); } void pf_mtag_initialize() { pf_mtag_z = uma_zcreate("pf mtags", sizeof(struct m_tag) + sizeof(struct pf_mtag), NULL, NULL, pf_mtag_uminit, NULL, UMA_ALIGN_PTR, 0); } /* Per-vnet data storage structures initialization. */ void pf_initialize() { struct pf_keyhash *kh; struct pf_idhash *ih; struct pf_srchash *sh; u_int i; if (pf_hashsize == 0 || !powerof2(pf_hashsize)) pf_hashsize = PF_HASHSIZ; if (pf_srchashsize == 0 || !powerof2(pf_srchashsize)) pf_srchashsize = PF_HASHSIZ / 4; V_pf_hashseed = arc4random(); /* States and state keys storage. */ V_pf_state_z = uma_zcreate("pf states", sizeof(struct pf_state), NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); V_pf_limits[PF_LIMIT_STATES].zone = V_pf_state_z; uma_zone_set_max(V_pf_state_z, PFSTATE_HIWAT); uma_zone_set_warning(V_pf_state_z, "PF states limit reached"); V_pf_state_key_z = uma_zcreate("pf state keys", sizeof(struct pf_state_key), pf_state_key_ctor, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); V_pf_keyhash = malloc(pf_hashsize * sizeof(struct pf_keyhash), M_PFHASH, M_WAITOK | M_ZERO); V_pf_idhash = malloc(pf_hashsize * sizeof(struct pf_idhash), M_PFHASH, M_WAITOK | M_ZERO); pf_hashmask = pf_hashsize - 1; for (i = 0, kh = V_pf_keyhash, ih = V_pf_idhash; i <= pf_hashmask; i++, kh++, ih++) { mtx_init(&kh->lock, "pf_keyhash", NULL, MTX_DEF | MTX_DUPOK); mtx_init(&ih->lock, "pf_idhash", NULL, MTX_DEF); } /* Source nodes. */ V_pf_sources_z = uma_zcreate("pf source nodes", sizeof(struct pf_src_node), NULL, NULL, NULL, NULL, UMA_ALIGN_PTR, 0); V_pf_limits[PF_LIMIT_SRC_NODES].zone = V_pf_sources_z; uma_zone_set_max(V_pf_sources_z, PFSNODE_HIWAT); uma_zone_set_warning(V_pf_sources_z, "PF source nodes limit reached"); V_pf_srchash = malloc(pf_srchashsize * sizeof(struct pf_srchash), M_PFHASH, M_WAITOK|M_ZERO); pf_srchashmask = pf_srchashsize - 1; for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) mtx_init(&sh->lock, "pf_srchash", NULL, MTX_DEF); /* ALTQ */ TAILQ_INIT(&V_pf_altqs[0]); TAILQ_INIT(&V_pf_altqs[1]); TAILQ_INIT(&V_pf_pabuf); V_pf_altqs_active = &V_pf_altqs[0]; V_pf_altqs_inactive = &V_pf_altqs[1]; /* Send & overload+flush queues. */ STAILQ_INIT(&V_pf_sendqueue); SLIST_INIT(&V_pf_overloadqueue); TASK_INIT(&V_pf_overloadtask, 0, pf_overload_task, curvnet); /* Unlinked, but may be referenced rules. */ TAILQ_INIT(&V_pf_unlinked_rules); } void pf_mtag_cleanup() { uma_zdestroy(pf_mtag_z); } void pf_cleanup() { struct pf_keyhash *kh; struct pf_idhash *ih; struct pf_srchash *sh; struct pf_send_entry *pfse, *next; u_int i; for (i = 0, kh = V_pf_keyhash, ih = V_pf_idhash; i <= pf_hashmask; i++, kh++, ih++) { KASSERT(LIST_EMPTY(&kh->keys), ("%s: key hash not empty", __func__)); KASSERT(LIST_EMPTY(&ih->states), ("%s: id hash not empty", __func__)); mtx_destroy(&kh->lock); mtx_destroy(&ih->lock); } free(V_pf_keyhash, M_PFHASH); free(V_pf_idhash, M_PFHASH); for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) { KASSERT(LIST_EMPTY(&sh->nodes), ("%s: source node hash not empty", __func__)); mtx_destroy(&sh->lock); } free(V_pf_srchash, M_PFHASH); STAILQ_FOREACH_SAFE(pfse, &V_pf_sendqueue, pfse_next, next) { m_freem(pfse->pfse_m); free(pfse, M_PFTEMP); } uma_zdestroy(V_pf_sources_z); uma_zdestroy(V_pf_state_z); uma_zdestroy(V_pf_state_key_z); } static int pf_mtag_uminit(void *mem, int size, int how) { struct m_tag *t; t = (struct m_tag *)mem; t->m_tag_cookie = MTAG_ABI_COMPAT; t->m_tag_id = PACKET_TAG_PF; t->m_tag_len = sizeof(struct pf_mtag); t->m_tag_free = pf_mtag_free; return (0); } static void pf_mtag_free(struct m_tag *t) { uma_zfree(pf_mtag_z, t); } struct pf_mtag * pf_get_mtag(struct mbuf *m) { struct m_tag *mtag; if ((mtag = m_tag_find(m, PACKET_TAG_PF, NULL)) != NULL) return ((struct pf_mtag *)(mtag + 1)); mtag = uma_zalloc(pf_mtag_z, M_NOWAIT); if (mtag == NULL) return (NULL); bzero(mtag + 1, sizeof(struct pf_mtag)); m_tag_prepend(m, mtag); return ((struct pf_mtag *)(mtag + 1)); } static int pf_state_key_attach(struct pf_state_key *skw, struct pf_state_key *sks, struct pf_state *s) { struct pf_keyhash *khs, *khw, *kh; struct pf_state_key *sk, *cur; struct pf_state *si, *olds = NULL; int idx; KASSERT(s->refs == 0, ("%s: state not pristine", __func__)); KASSERT(s->key[PF_SK_WIRE] == NULL, ("%s: state has key", __func__)); KASSERT(s->key[PF_SK_STACK] == NULL, ("%s: state has key", __func__)); /* * We need to lock hash slots of both keys. To avoid deadlock * we always lock the slot with lower address first. Unlock order * isn't important. * * We also need to lock ID hash slot before dropping key * locks. On success we return with ID hash slot locked. */ if (skw == sks) { khs = khw = &V_pf_keyhash[pf_hashkey(skw)]; PF_HASHROW_LOCK(khs); } else { khs = &V_pf_keyhash[pf_hashkey(sks)]; khw = &V_pf_keyhash[pf_hashkey(skw)]; if (khs == khw) { PF_HASHROW_LOCK(khs); } else if (khs < khw) { PF_HASHROW_LOCK(khs); PF_HASHROW_LOCK(khw); } else { PF_HASHROW_LOCK(khw); PF_HASHROW_LOCK(khs); } } #define KEYS_UNLOCK() do { \ if (khs != khw) { \ PF_HASHROW_UNLOCK(khs); \ PF_HASHROW_UNLOCK(khw); \ } else \ PF_HASHROW_UNLOCK(khs); \ } while (0) /* * First run: start with wire key. */ sk = skw; kh = khw; idx = PF_SK_WIRE; keyattach: LIST_FOREACH(cur, &kh->keys, entry) if (bcmp(cur, sk, sizeof(struct pf_state_key_cmp)) == 0) break; if (cur != NULL) { /* Key exists. Check for same kif, if none, add to key. */ TAILQ_FOREACH(si, &cur->states[idx], key_list[idx]) { struct pf_idhash *ih = &V_pf_idhash[PF_IDHASH(si)]; PF_HASHROW_LOCK(ih); if (si->kif == s->kif && si->direction == s->direction) { if (sk->proto == IPPROTO_TCP && si->src.state >= TCPS_FIN_WAIT_2 && si->dst.state >= TCPS_FIN_WAIT_2) { /* * New state matches an old >FIN_WAIT_2 * state. We can't drop key hash locks, * thus we can't unlink it properly. * * As a workaround we drop it into * TCPS_CLOSED state, schedule purge * ASAP and push it into the very end * of the slot TAILQ, so that it won't * conflict with our new state. */ si->src.state = si->dst.state = TCPS_CLOSED; si->timeout = PFTM_PURGE; olds = si; } else { if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: %s key attach " "failed on %s: ", (idx == PF_SK_WIRE) ? "wire" : "stack", s->kif->pfik_name); pf_print_state_parts(s, (idx == PF_SK_WIRE) ? sk : NULL, (idx == PF_SK_STACK) ? sk : NULL); printf(", existing: "); pf_print_state_parts(si, (idx == PF_SK_WIRE) ? sk : NULL, (idx == PF_SK_STACK) ? sk : NULL); printf("\n"); } PF_HASHROW_UNLOCK(ih); KEYS_UNLOCK(); uma_zfree(V_pf_state_key_z, sk); if (idx == PF_SK_STACK) pf_detach_state(s); return (EEXIST); /* collision! */ } } PF_HASHROW_UNLOCK(ih); } uma_zfree(V_pf_state_key_z, sk); s->key[idx] = cur; } else { LIST_INSERT_HEAD(&kh->keys, sk, entry); s->key[idx] = sk; } stateattach: /* List is sorted, if-bound states before floating. */ if (s->kif == V_pfi_all) TAILQ_INSERT_TAIL(&s->key[idx]->states[idx], s, key_list[idx]); else TAILQ_INSERT_HEAD(&s->key[idx]->states[idx], s, key_list[idx]); if (olds) { TAILQ_REMOVE(&s->key[idx]->states[idx], olds, key_list[idx]); TAILQ_INSERT_TAIL(&s->key[idx]->states[idx], olds, key_list[idx]); olds = NULL; } /* * Attach done. See how should we (or should not?) * attach a second key. */ if (sks == skw) { s->key[PF_SK_STACK] = s->key[PF_SK_WIRE]; idx = PF_SK_STACK; sks = NULL; goto stateattach; } else if (sks != NULL) { /* * Continue attaching with stack key. */ sk = sks; kh = khs; idx = PF_SK_STACK; sks = NULL; goto keyattach; } PF_STATE_LOCK(s); KEYS_UNLOCK(); KASSERT(s->key[PF_SK_WIRE] != NULL && s->key[PF_SK_STACK] != NULL, ("%s failure", __func__)); return (0); #undef KEYS_UNLOCK } static void pf_detach_state(struct pf_state *s) { struct pf_state_key *sks = s->key[PF_SK_STACK]; struct pf_keyhash *kh; if (sks != NULL) { kh = &V_pf_keyhash[pf_hashkey(sks)]; PF_HASHROW_LOCK(kh); if (s->key[PF_SK_STACK] != NULL) pf_state_key_detach(s, PF_SK_STACK); /* * If both point to same key, then we are done. */ if (sks == s->key[PF_SK_WIRE]) { pf_state_key_detach(s, PF_SK_WIRE); PF_HASHROW_UNLOCK(kh); return; } PF_HASHROW_UNLOCK(kh); } if (s->key[PF_SK_WIRE] != NULL) { kh = &V_pf_keyhash[pf_hashkey(s->key[PF_SK_WIRE])]; PF_HASHROW_LOCK(kh); if (s->key[PF_SK_WIRE] != NULL) pf_state_key_detach(s, PF_SK_WIRE); PF_HASHROW_UNLOCK(kh); } } static void pf_state_key_detach(struct pf_state *s, int idx) { struct pf_state_key *sk = s->key[idx]; #ifdef INVARIANTS struct pf_keyhash *kh = &V_pf_keyhash[pf_hashkey(sk)]; PF_HASHROW_ASSERT(kh); #endif TAILQ_REMOVE(&sk->states[idx], s, key_list[idx]); s->key[idx] = NULL; if (TAILQ_EMPTY(&sk->states[0]) && TAILQ_EMPTY(&sk->states[1])) { LIST_REMOVE(sk, entry); uma_zfree(V_pf_state_key_z, sk); } } static int pf_state_key_ctor(void *mem, int size, void *arg, int flags) { struct pf_state_key *sk = mem; bzero(sk, sizeof(struct pf_state_key_cmp)); TAILQ_INIT(&sk->states[PF_SK_WIRE]); TAILQ_INIT(&sk->states[PF_SK_STACK]); return (0); } struct pf_state_key * pf_state_key_setup(struct pf_pdesc *pd, struct pf_addr *saddr, struct pf_addr *daddr, u_int16_t sport, u_int16_t dport) { struct pf_state_key *sk; sk = uma_zalloc(V_pf_state_key_z, M_NOWAIT); if (sk == NULL) return (NULL); PF_ACPY(&sk->addr[pd->sidx], saddr, pd->af); PF_ACPY(&sk->addr[pd->didx], daddr, pd->af); sk->port[pd->sidx] = sport; sk->port[pd->didx] = dport; sk->proto = pd->proto; sk->af = pd->af; return (sk); } struct pf_state_key * pf_state_key_clone(struct pf_state_key *orig) { struct pf_state_key *sk; sk = uma_zalloc(V_pf_state_key_z, M_NOWAIT); if (sk == NULL) return (NULL); bcopy(orig, sk, sizeof(struct pf_state_key_cmp)); return (sk); } int pf_state_insert(struct pfi_kif *kif, struct pf_state_key *skw, struct pf_state_key *sks, struct pf_state *s) { struct pf_idhash *ih; struct pf_state *cur; int error; KASSERT(TAILQ_EMPTY(&sks->states[0]) && TAILQ_EMPTY(&sks->states[1]), ("%s: sks not pristine", __func__)); KASSERT(TAILQ_EMPTY(&skw->states[0]) && TAILQ_EMPTY(&skw->states[1]), ("%s: skw not pristine", __func__)); KASSERT(s->refs == 0, ("%s: state not pristine", __func__)); s->kif = kif; if (s->id == 0 && s->creatorid == 0) { /* XXX: should be atomic, but probability of collision low */ if ((s->id = V_pf_stateid[curcpu]++) == PFID_MAXID) V_pf_stateid[curcpu] = 1; s->id |= (uint64_t )curcpu << PFID_CPUSHIFT; s->id = htobe64(s->id); s->creatorid = V_pf_status.hostid; } /* Returns with ID locked on success. */ if ((error = pf_state_key_attach(skw, sks, s)) != 0) return (error); ih = &V_pf_idhash[PF_IDHASH(s)]; PF_HASHROW_ASSERT(ih); LIST_FOREACH(cur, &ih->states, entry) if (cur->id == s->id && cur->creatorid == s->creatorid) break; if (cur != NULL) { PF_HASHROW_UNLOCK(ih); if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: state ID collision: " "id: %016llx creatorid: %08x\n", (unsigned long long)be64toh(s->id), ntohl(s->creatorid)); } pf_detach_state(s); return (EEXIST); } LIST_INSERT_HEAD(&ih->states, s, entry); /* One for keys, one for ID hash. */ refcount_init(&s->refs, 2); counter_u64_add(V_pf_status.fcounters[FCNT_STATE_INSERT], 1); if (pfsync_insert_state_ptr != NULL) pfsync_insert_state_ptr(s); /* Returns locked. */ return (0); } /* * Find state by ID: returns with locked row on success. */ struct pf_state * pf_find_state_byid(uint64_t id, uint32_t creatorid) { struct pf_idhash *ih; struct pf_state *s; counter_u64_add(V_pf_status.fcounters[FCNT_STATE_SEARCH], 1); ih = &V_pf_idhash[(be64toh(id) % (pf_hashmask + 1))]; PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) if (s->id == id && s->creatorid == creatorid) break; if (s == NULL) PF_HASHROW_UNLOCK(ih); return (s); } /* * Find state by key. * Returns with ID hash slot locked on success. */ static struct pf_state * pf_find_state(struct pfi_kif *kif, struct pf_state_key_cmp *key, u_int dir) { struct pf_keyhash *kh; struct pf_state_key *sk; struct pf_state *s; int idx; counter_u64_add(V_pf_status.fcounters[FCNT_STATE_SEARCH], 1); kh = &V_pf_keyhash[pf_hashkey((struct pf_state_key *)key)]; PF_HASHROW_LOCK(kh); LIST_FOREACH(sk, &kh->keys, entry) if (bcmp(sk, key, sizeof(struct pf_state_key_cmp)) == 0) break; if (sk == NULL) { PF_HASHROW_UNLOCK(kh); return (NULL); } idx = (dir == PF_IN ? PF_SK_WIRE : PF_SK_STACK); /* List is sorted, if-bound states before floating ones. */ TAILQ_FOREACH(s, &sk->states[idx], key_list[idx]) if (s->kif == V_pfi_all || s->kif == kif) { PF_STATE_LOCK(s); PF_HASHROW_UNLOCK(kh); if (s->timeout >= PFTM_MAX) { /* * State is either being processed by * pf_unlink_state() in an other thread, or * is scheduled for immediate expiry. */ PF_STATE_UNLOCK(s); return (NULL); } return (s); } PF_HASHROW_UNLOCK(kh); return (NULL); } struct pf_state * pf_find_state_all(struct pf_state_key_cmp *key, u_int dir, int *more) { struct pf_keyhash *kh; struct pf_state_key *sk; struct pf_state *s, *ret = NULL; int idx, inout = 0; counter_u64_add(V_pf_status.fcounters[FCNT_STATE_SEARCH], 1); kh = &V_pf_keyhash[pf_hashkey((struct pf_state_key *)key)]; PF_HASHROW_LOCK(kh); LIST_FOREACH(sk, &kh->keys, entry) if (bcmp(sk, key, sizeof(struct pf_state_key_cmp)) == 0) break; if (sk == NULL) { PF_HASHROW_UNLOCK(kh); return (NULL); } switch (dir) { case PF_IN: idx = PF_SK_WIRE; break; case PF_OUT: idx = PF_SK_STACK; break; case PF_INOUT: idx = PF_SK_WIRE; inout = 1; break; default: panic("%s: dir %u", __func__, dir); } second_run: TAILQ_FOREACH(s, &sk->states[idx], key_list[idx]) { if (more == NULL) { PF_HASHROW_UNLOCK(kh); return (s); } if (ret) (*more)++; else ret = s; } if (inout == 1) { inout = 0; idx = PF_SK_STACK; goto second_run; } PF_HASHROW_UNLOCK(kh); return (ret); } /* END state table stuff */ static void pf_send(struct pf_send_entry *pfse) { PF_SENDQ_LOCK(); STAILQ_INSERT_TAIL(&V_pf_sendqueue, pfse, pfse_next); PF_SENDQ_UNLOCK(); swi_sched(V_pf_swi_cookie, 0); } void pf_intr(void *v) { struct pf_send_head queue; struct pf_send_entry *pfse, *next; CURVNET_SET((struct vnet *)v); PF_SENDQ_LOCK(); queue = V_pf_sendqueue; STAILQ_INIT(&V_pf_sendqueue); PF_SENDQ_UNLOCK(); STAILQ_FOREACH_SAFE(pfse, &queue, pfse_next, next) { switch (pfse->pfse_type) { #ifdef INET case PFSE_IP: ip_output(pfse->pfse_m, NULL, NULL, 0, NULL, NULL); break; case PFSE_ICMP: icmp_error(pfse->pfse_m, pfse->icmpopts.type, pfse->icmpopts.code, 0, pfse->icmpopts.mtu); break; #endif /* INET */ #ifdef INET6 case PFSE_IP6: ip6_output(pfse->pfse_m, NULL, NULL, 0, NULL, NULL, NULL); break; case PFSE_ICMP6: icmp6_error(pfse->pfse_m, pfse->icmpopts.type, pfse->icmpopts.code, pfse->icmpopts.mtu); break; #endif /* INET6 */ default: panic("%s: unknown type", __func__); } free(pfse, M_PFTEMP); } CURVNET_RESTORE(); } void pf_purge_thread(void *unused __unused) { VNET_ITERATOR_DECL(vnet_iter); u_int idx = 0; for (;;) { PF_RULES_RLOCK(); rw_sleep(pf_purge_thread, &pf_rules_lock, 0, "pftm", hz / 10); PF_RULES_RUNLOCK(); VNET_LIST_RLOCK(); VNET_FOREACH(vnet_iter) { CURVNET_SET(vnet_iter); if (pf_end_threads) { pf_end_threads++; wakeup(pf_purge_thread); kproc_exit(0); } /* Process 1/interval fraction of the state table every run. */ idx = pf_purge_expired_states(idx, pf_hashmask / (V_pf_default_rule.timeout[PFTM_INTERVAL] * 10)); /* Purge other expired types every PFTM_INTERVAL seconds. */ if (idx == 0) { /* * Order is important: * - states and src nodes reference rules * - states and rules reference kifs */ pf_purge_expired_fragments(); pf_purge_expired_src_nodes(); pf_purge_unlinked_rules(); pfi_kif_purge(); } CURVNET_RESTORE(); } VNET_LIST_RUNLOCK(); } /* not reached */ } void pf_unload_vnet_purge(void) { /* * To cleanse up all kifs and rules we need * two runs: first one clears reference flags, * then pf_purge_expired_states() doesn't * raise them, and then second run frees. */ pf_purge_unlinked_rules(); pfi_kif_purge(); /* * Now purge everything. */ pf_purge_expired_states(0, pf_hashmask); pf_purge_expired_fragments(); pf_purge_expired_src_nodes(); /* * Now all kifs & rules should be unreferenced, * thus should be successfully freed. */ pf_purge_unlinked_rules(); pfi_kif_purge(); } u_int32_t pf_state_expires(const struct pf_state *state) { u_int32_t timeout; u_int32_t start; u_int32_t end; u_int32_t states; /* handle all PFTM_* > PFTM_MAX here */ if (state->timeout == PFTM_PURGE) return (time_uptime); KASSERT(state->timeout != PFTM_UNLINKED, ("pf_state_expires: timeout == PFTM_UNLINKED")); KASSERT((state->timeout < PFTM_MAX), ("pf_state_expires: timeout > PFTM_MAX")); timeout = state->rule.ptr->timeout[state->timeout]; if (!timeout) timeout = V_pf_default_rule.timeout[state->timeout]; start = state->rule.ptr->timeout[PFTM_ADAPTIVE_START]; if (start) { end = state->rule.ptr->timeout[PFTM_ADAPTIVE_END]; states = counter_u64_fetch(state->rule.ptr->states_cur); } else { start = V_pf_default_rule.timeout[PFTM_ADAPTIVE_START]; end = V_pf_default_rule.timeout[PFTM_ADAPTIVE_END]; states = V_pf_status.states; } if (end && states > start && start < end) { if (states < end) return (state->expire + timeout * (end - states) / (end - start)); else return (time_uptime); } return (state->expire + timeout); } void pf_purge_expired_src_nodes() { struct pf_src_node_list freelist; struct pf_srchash *sh; struct pf_src_node *cur, *next; int i; LIST_INIT(&freelist); for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) { PF_HASHROW_LOCK(sh); LIST_FOREACH_SAFE(cur, &sh->nodes, entry, next) if (cur->states == 0 && cur->expire <= time_uptime) { pf_unlink_src_node(cur); LIST_INSERT_HEAD(&freelist, cur, entry); } else if (cur->rule.ptr != NULL) cur->rule.ptr->rule_flag |= PFRULE_REFS; PF_HASHROW_UNLOCK(sh); } pf_free_src_nodes(&freelist); V_pf_status.src_nodes = uma_zone_get_cur(V_pf_sources_z); } static void pf_src_tree_remove_state(struct pf_state *s) { struct pf_src_node *sn; struct pf_srchash *sh; uint32_t timeout; timeout = s->rule.ptr->timeout[PFTM_SRC_NODE] ? s->rule.ptr->timeout[PFTM_SRC_NODE] : V_pf_default_rule.timeout[PFTM_SRC_NODE]; if (s->src_node != NULL) { sn = s->src_node; sh = &V_pf_srchash[pf_hashsrc(&sn->addr, sn->af)]; PF_HASHROW_LOCK(sh); if (s->src.tcp_est) --sn->conn; if (--sn->states == 0) sn->expire = time_uptime + timeout; PF_HASHROW_UNLOCK(sh); } if (s->nat_src_node != s->src_node && s->nat_src_node != NULL) { sn = s->nat_src_node; sh = &V_pf_srchash[pf_hashsrc(&sn->addr, sn->af)]; PF_HASHROW_LOCK(sh); if (--sn->states == 0) sn->expire = time_uptime + timeout; PF_HASHROW_UNLOCK(sh); } s->src_node = s->nat_src_node = NULL; } /* * Unlink and potentilly free a state. Function may be * called with ID hash row locked, but always returns * unlocked, since it needs to go through key hash locking. */ int pf_unlink_state(struct pf_state *s, u_int flags) { struct pf_idhash *ih = &V_pf_idhash[PF_IDHASH(s)]; if ((flags & PF_ENTER_LOCKED) == 0) PF_HASHROW_LOCK(ih); else PF_HASHROW_ASSERT(ih); if (s->timeout == PFTM_UNLINKED) { /* * State is being processed * by pf_unlink_state() in * an other thread. */ PF_HASHROW_UNLOCK(ih); return (0); /* XXXGL: undefined actually */ } if (s->src.state == PF_TCPS_PROXY_DST) { /* XXX wire key the right one? */ pf_send_tcp(NULL, s->rule.ptr, s->key[PF_SK_WIRE]->af, &s->key[PF_SK_WIRE]->addr[1], &s->key[PF_SK_WIRE]->addr[0], s->key[PF_SK_WIRE]->port[1], s->key[PF_SK_WIRE]->port[0], s->src.seqhi, s->src.seqlo + 1, TH_RST|TH_ACK, 0, 0, 0, 1, s->tag, NULL); } LIST_REMOVE(s, entry); pf_src_tree_remove_state(s); if (pfsync_delete_state_ptr != NULL) pfsync_delete_state_ptr(s); STATE_DEC_COUNTERS(s); s->timeout = PFTM_UNLINKED; PF_HASHROW_UNLOCK(ih); pf_detach_state(s); refcount_release(&s->refs); return (pf_release_state(s)); } void pf_free_state(struct pf_state *cur) { KASSERT(cur->refs == 0, ("%s: %p has refs", __func__, cur)); KASSERT(cur->timeout == PFTM_UNLINKED, ("%s: timeout %u", __func__, cur->timeout)); pf_normalize_tcp_cleanup(cur); uma_zfree(V_pf_state_z, cur); counter_u64_add(V_pf_status.fcounters[FCNT_STATE_REMOVALS], 1); } /* * Called only from pf_purge_thread(), thus serialized. */ static u_int pf_purge_expired_states(u_int i, int maxcheck) { struct pf_idhash *ih; struct pf_state *s; V_pf_status.states = uma_zone_get_cur(V_pf_state_z); /* * Go through hash and unlink states that expire now. */ while (maxcheck > 0) { ih = &V_pf_idhash[i]; relock: PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { if (pf_state_expires(s) <= time_uptime) { V_pf_status.states -= pf_unlink_state(s, PF_ENTER_LOCKED); goto relock; } s->rule.ptr->rule_flag |= PFRULE_REFS; if (s->nat_rule.ptr != NULL) s->nat_rule.ptr->rule_flag |= PFRULE_REFS; if (s->anchor.ptr != NULL) s->anchor.ptr->rule_flag |= PFRULE_REFS; s->kif->pfik_flags |= PFI_IFLAG_REFS; if (s->rt_kif) s->rt_kif->pfik_flags |= PFI_IFLAG_REFS; } PF_HASHROW_UNLOCK(ih); /* Return when we hit end of hash. */ if (++i > pf_hashmask) { V_pf_status.states = uma_zone_get_cur(V_pf_state_z); return (0); } maxcheck--; } V_pf_status.states = uma_zone_get_cur(V_pf_state_z); return (i); } static void pf_purge_unlinked_rules() { struct pf_rulequeue tmpq; struct pf_rule *r, *r1; /* * If we have overloading task pending, then we'd * better skip purging this time. There is a tiny * probability that overloading task references * an already unlinked rule. */ PF_OVERLOADQ_LOCK(); if (!SLIST_EMPTY(&V_pf_overloadqueue)) { PF_OVERLOADQ_UNLOCK(); return; } PF_OVERLOADQ_UNLOCK(); /* * Do naive mark-and-sweep garbage collecting of old rules. * Reference flag is raised by pf_purge_expired_states() * and pf_purge_expired_src_nodes(). * * To avoid LOR between PF_UNLNKDRULES_LOCK/PF_RULES_WLOCK, * use a temporary queue. */ TAILQ_INIT(&tmpq); PF_UNLNKDRULES_LOCK(); TAILQ_FOREACH_SAFE(r, &V_pf_unlinked_rules, entries, r1) { if (!(r->rule_flag & PFRULE_REFS)) { TAILQ_REMOVE(&V_pf_unlinked_rules, r, entries); TAILQ_INSERT_TAIL(&tmpq, r, entries); } else r->rule_flag &= ~PFRULE_REFS; } PF_UNLNKDRULES_UNLOCK(); if (!TAILQ_EMPTY(&tmpq)) { PF_RULES_WLOCK(); TAILQ_FOREACH_SAFE(r, &tmpq, entries, r1) { TAILQ_REMOVE(&tmpq, r, entries); pf_free_rule(r); } PF_RULES_WUNLOCK(); } } void pf_print_host(struct pf_addr *addr, u_int16_t p, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: { u_int32_t a = ntohl(addr->addr32[0]); printf("%u.%u.%u.%u", (a>>24)&255, (a>>16)&255, (a>>8)&255, a&255); if (p) { p = ntohs(p); printf(":%u", p); } break; } #endif /* INET */ #ifdef INET6 case AF_INET6: { u_int16_t b; u_int8_t i, curstart, curend, maxstart, maxend; curstart = curend = maxstart = maxend = 255; for (i = 0; i < 8; i++) { if (!addr->addr16[i]) { if (curstart == 255) curstart = i; curend = i; } else { if ((curend - curstart) > (maxend - maxstart)) { maxstart = curstart; maxend = curend; } curstart = curend = 255; } } if ((curend - curstart) > (maxend - maxstart)) { maxstart = curstart; maxend = curend; } for (i = 0; i < 8; i++) { if (i >= maxstart && i <= maxend) { if (i == 0) printf(":"); if (i == maxend) printf(":"); } else { b = ntohs(addr->addr16[i]); printf("%x", b); if (i < 7) printf(":"); } } if (p) { p = ntohs(p); printf("[%u]", p); } break; } #endif /* INET6 */ } } void pf_print_state(struct pf_state *s) { pf_print_state_parts(s, NULL, NULL); } static void pf_print_state_parts(struct pf_state *s, struct pf_state_key *skwp, struct pf_state_key *sksp) { struct pf_state_key *skw, *sks; u_int8_t proto, dir; /* Do our best to fill these, but they're skipped if NULL */ skw = skwp ? skwp : (s ? s->key[PF_SK_WIRE] : NULL); sks = sksp ? sksp : (s ? s->key[PF_SK_STACK] : NULL); proto = skw ? skw->proto : (sks ? sks->proto : 0); dir = s ? s->direction : 0; switch (proto) { case IPPROTO_IPV4: printf("IPv4"); break; case IPPROTO_IPV6: printf("IPv6"); break; case IPPROTO_TCP: printf("TCP"); break; case IPPROTO_UDP: printf("UDP"); break; case IPPROTO_ICMP: printf("ICMP"); break; case IPPROTO_ICMPV6: printf("ICMPv6"); break; default: printf("%u", proto); break; } switch (dir) { case PF_IN: printf(" in"); break; case PF_OUT: printf(" out"); break; } if (skw) { printf(" wire: "); pf_print_host(&skw->addr[0], skw->port[0], skw->af); printf(" "); pf_print_host(&skw->addr[1], skw->port[1], skw->af); } if (sks) { printf(" stack: "); if (sks != skw) { pf_print_host(&sks->addr[0], sks->port[0], sks->af); printf(" "); pf_print_host(&sks->addr[1], sks->port[1], sks->af); } else printf("-"); } if (s) { if (proto == IPPROTO_TCP) { printf(" [lo=%u high=%u win=%u modulator=%u", s->src.seqlo, s->src.seqhi, s->src.max_win, s->src.seqdiff); if (s->src.wscale && s->dst.wscale) printf(" wscale=%u", s->src.wscale & PF_WSCALE_MASK); printf("]"); printf(" [lo=%u high=%u win=%u modulator=%u", s->dst.seqlo, s->dst.seqhi, s->dst.max_win, s->dst.seqdiff); if (s->src.wscale && s->dst.wscale) printf(" wscale=%u", s->dst.wscale & PF_WSCALE_MASK); printf("]"); } printf(" %u:%u", s->src.state, s->dst.state); } } void pf_print_flags(u_int8_t f) { if (f) printf(" "); if (f & TH_FIN) printf("F"); if (f & TH_SYN) printf("S"); if (f & TH_RST) printf("R"); if (f & TH_PUSH) printf("P"); if (f & TH_ACK) printf("A"); if (f & TH_URG) printf("U"); if (f & TH_ECE) printf("E"); if (f & TH_CWR) printf("W"); } #define PF_SET_SKIP_STEPS(i) \ do { \ while (head[i] != cur) { \ head[i]->skip[i].ptr = cur; \ head[i] = TAILQ_NEXT(head[i], entries); \ } \ } while (0) void pf_calc_skip_steps(struct pf_rulequeue *rules) { struct pf_rule *cur, *prev, *head[PF_SKIP_COUNT]; int i; cur = TAILQ_FIRST(rules); prev = cur; for (i = 0; i < PF_SKIP_COUNT; ++i) head[i] = cur; while (cur != NULL) { if (cur->kif != prev->kif || cur->ifnot != prev->ifnot) PF_SET_SKIP_STEPS(PF_SKIP_IFP); if (cur->direction != prev->direction) PF_SET_SKIP_STEPS(PF_SKIP_DIR); if (cur->af != prev->af) PF_SET_SKIP_STEPS(PF_SKIP_AF); if (cur->proto != prev->proto) PF_SET_SKIP_STEPS(PF_SKIP_PROTO); if (cur->src.neg != prev->src.neg || pf_addr_wrap_neq(&cur->src.addr, &prev->src.addr)) PF_SET_SKIP_STEPS(PF_SKIP_SRC_ADDR); if (cur->src.port[0] != prev->src.port[0] || cur->src.port[1] != prev->src.port[1] || cur->src.port_op != prev->src.port_op) PF_SET_SKIP_STEPS(PF_SKIP_SRC_PORT); if (cur->dst.neg != prev->dst.neg || pf_addr_wrap_neq(&cur->dst.addr, &prev->dst.addr)) PF_SET_SKIP_STEPS(PF_SKIP_DST_ADDR); if (cur->dst.port[0] != prev->dst.port[0] || cur->dst.port[1] != prev->dst.port[1] || cur->dst.port_op != prev->dst.port_op) PF_SET_SKIP_STEPS(PF_SKIP_DST_PORT); prev = cur; cur = TAILQ_NEXT(cur, entries); } for (i = 0; i < PF_SKIP_COUNT; ++i) PF_SET_SKIP_STEPS(i); } static int pf_addr_wrap_neq(struct pf_addr_wrap *aw1, struct pf_addr_wrap *aw2) { if (aw1->type != aw2->type) return (1); switch (aw1->type) { case PF_ADDR_ADDRMASK: case PF_ADDR_RANGE: if (PF_ANEQ(&aw1->v.a.addr, &aw2->v.a.addr, AF_INET6)) return (1); if (PF_ANEQ(&aw1->v.a.mask, &aw2->v.a.mask, AF_INET6)) return (1); return (0); case PF_ADDR_DYNIFTL: return (aw1->p.dyn->pfid_kt != aw2->p.dyn->pfid_kt); case PF_ADDR_NOROUTE: case PF_ADDR_URPFFAILED: return (0); case PF_ADDR_TABLE: return (aw1->p.tbl != aw2->p.tbl); default: printf("invalid address type: %d\n", aw1->type); return (1); } } /** * Checksum updates are a little complicated because the checksum in the TCP/UDP * header isn't always a full checksum. In some cases (i.e. output) it's a * pseudo-header checksum, which is a partial checksum over src/dst IP * addresses, protocol number and length. * * That means we have the following cases: * * Input or forwarding: we don't have TSO, the checksum fields are full * checksums, we need to update the checksum whenever we change anything. * * Output (i.e. the checksum is a pseudo-header checksum): * x The field being updated is src/dst address or affects the length of * the packet. We need to update the pseudo-header checksum (note that this * checksum is not ones' complement). * x Some other field is being modified (e.g. src/dst port numbers): We * don't have to update anything. **/ u_int16_t pf_cksum_fixup(u_int16_t cksum, u_int16_t old, u_int16_t new, u_int8_t udp) { u_int32_t l; if (udp && !cksum) return (0x0000); l = cksum + old - new; l = (l >> 16) + (l & 65535); l = l & 65535; if (udp && !l) return (0xFFFF); return (l); } u_int16_t pf_proto_cksum_fixup(struct mbuf *m, u_int16_t cksum, u_int16_t old, u_int16_t new, u_int8_t udp) { if (m->m_pkthdr.csum_flags & (CSUM_DELAY_DATA | CSUM_DELAY_DATA_IPV6)) return (cksum); return (pf_cksum_fixup(cksum, old, new, udp)); } static void pf_change_ap(struct mbuf *m, struct pf_addr *a, u_int16_t *p, u_int16_t *ic, u_int16_t *pc, struct pf_addr *an, u_int16_t pn, u_int8_t u, sa_family_t af) { struct pf_addr ao; u_int16_t po = *p; PF_ACPY(&ao, a, af); PF_ACPY(a, an, af); if (m->m_pkthdr.csum_flags & (CSUM_DELAY_DATA | CSUM_DELAY_DATA_IPV6)) *pc = ~*pc; *p = pn; switch (af) { #ifdef INET case AF_INET: *ic = pf_cksum_fixup(pf_cksum_fixup(*ic, ao.addr16[0], an->addr16[0], 0), ao.addr16[1], an->addr16[1], 0); *p = pn; *pc = pf_cksum_fixup(pf_cksum_fixup(*pc, ao.addr16[0], an->addr16[0], u), ao.addr16[1], an->addr16[1], u); *pc = pf_proto_cksum_fixup(m, *pc, po, pn, u); break; #endif /* INET */ #ifdef INET6 case AF_INET6: *pc = pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(*pc, ao.addr16[0], an->addr16[0], u), ao.addr16[1], an->addr16[1], u), ao.addr16[2], an->addr16[2], u), ao.addr16[3], an->addr16[3], u), ao.addr16[4], an->addr16[4], u), ao.addr16[5], an->addr16[5], u), ao.addr16[6], an->addr16[6], u), ao.addr16[7], an->addr16[7], u); *pc = pf_proto_cksum_fixup(m, *pc, po, pn, u); break; #endif /* INET6 */ } if (m->m_pkthdr.csum_flags & (CSUM_DELAY_DATA | CSUM_DELAY_DATA_IPV6)) { *pc = ~*pc; if (! *pc) *pc = 0xffff; } } /* Changes a u_int32_t. Uses a void * so there are no align restrictions */ void pf_change_a(void *a, u_int16_t *c, u_int32_t an, u_int8_t u) { u_int32_t ao; memcpy(&ao, a, sizeof(ao)); memcpy(a, &an, sizeof(u_int32_t)); *c = pf_cksum_fixup(pf_cksum_fixup(*c, ao / 65536, an / 65536, u), ao % 65536, an % 65536, u); } void pf_change_proto_a(struct mbuf *m, void *a, u_int16_t *c, u_int32_t an, u_int8_t udp) { u_int32_t ao; memcpy(&ao, a, sizeof(ao)); memcpy(a, &an, sizeof(u_int32_t)); *c = pf_proto_cksum_fixup(m, pf_proto_cksum_fixup(m, *c, ao / 65536, an / 65536, udp), ao % 65536, an % 65536, udp); } #ifdef INET6 static void pf_change_a6(struct pf_addr *a, u_int16_t *c, struct pf_addr *an, u_int8_t u) { struct pf_addr ao; PF_ACPY(&ao, a, AF_INET6); PF_ACPY(a, an, AF_INET6); *c = pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(*c, ao.addr16[0], an->addr16[0], u), ao.addr16[1], an->addr16[1], u), ao.addr16[2], an->addr16[2], u), ao.addr16[3], an->addr16[3], u), ao.addr16[4], an->addr16[4], u), ao.addr16[5], an->addr16[5], u), ao.addr16[6], an->addr16[6], u), ao.addr16[7], an->addr16[7], u); } #endif /* INET6 */ static void pf_change_icmp(struct pf_addr *ia, u_int16_t *ip, struct pf_addr *oa, struct pf_addr *na, u_int16_t np, u_int16_t *pc, u_int16_t *h2c, u_int16_t *ic, u_int16_t *hc, u_int8_t u, sa_family_t af) { struct pf_addr oia, ooa; PF_ACPY(&oia, ia, af); if (oa) PF_ACPY(&ooa, oa, af); /* Change inner protocol port, fix inner protocol checksum. */ if (ip != NULL) { u_int16_t oip = *ip; u_int32_t opc; if (pc != NULL) opc = *pc; *ip = np; if (pc != NULL) *pc = pf_cksum_fixup(*pc, oip, *ip, u); *ic = pf_cksum_fixup(*ic, oip, *ip, 0); if (pc != NULL) *ic = pf_cksum_fixup(*ic, opc, *pc, 0); } /* Change inner ip address, fix inner ip and icmp checksums. */ PF_ACPY(ia, na, af); switch (af) { #ifdef INET case AF_INET: { u_int32_t oh2c = *h2c; *h2c = pf_cksum_fixup(pf_cksum_fixup(*h2c, oia.addr16[0], ia->addr16[0], 0), oia.addr16[1], ia->addr16[1], 0); *ic = pf_cksum_fixup(pf_cksum_fixup(*ic, oia.addr16[0], ia->addr16[0], 0), oia.addr16[1], ia->addr16[1], 0); *ic = pf_cksum_fixup(*ic, oh2c, *h2c, 0); break; } #endif /* INET */ #ifdef INET6 case AF_INET6: *ic = pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(*ic, oia.addr16[0], ia->addr16[0], u), oia.addr16[1], ia->addr16[1], u), oia.addr16[2], ia->addr16[2], u), oia.addr16[3], ia->addr16[3], u), oia.addr16[4], ia->addr16[4], u), oia.addr16[5], ia->addr16[5], u), oia.addr16[6], ia->addr16[6], u), oia.addr16[7], ia->addr16[7], u); break; #endif /* INET6 */ } /* Outer ip address, fix outer ip or icmpv6 checksum, if necessary. */ if (oa) { PF_ACPY(oa, na, af); switch (af) { #ifdef INET case AF_INET: *hc = pf_cksum_fixup(pf_cksum_fixup(*hc, ooa.addr16[0], oa->addr16[0], 0), ooa.addr16[1], oa->addr16[1], 0); break; #endif /* INET */ #ifdef INET6 case AF_INET6: *ic = pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(pf_cksum_fixup( pf_cksum_fixup(pf_cksum_fixup(*ic, ooa.addr16[0], oa->addr16[0], u), ooa.addr16[1], oa->addr16[1], u), ooa.addr16[2], oa->addr16[2], u), ooa.addr16[3], oa->addr16[3], u), ooa.addr16[4], oa->addr16[4], u), ooa.addr16[5], oa->addr16[5], u), ooa.addr16[6], oa->addr16[6], u), ooa.addr16[7], oa->addr16[7], u); break; #endif /* INET6 */ } } } /* * Need to modulate the sequence numbers in the TCP SACK option * (credits to Krzysztof Pfaff for report and patch) */ static int pf_modulate_sack(struct mbuf *m, int off, struct pf_pdesc *pd, struct tcphdr *th, struct pf_state_peer *dst) { int hlen = (th->th_off << 2) - sizeof(*th), thoptlen = hlen; u_int8_t opts[TCP_MAXOLEN], *opt = opts; int copyback = 0, i, olen; struct sackblk sack; #define TCPOLEN_SACKLEN (TCPOLEN_SACK + 2) if (hlen < TCPOLEN_SACKLEN || !pf_pull_hdr(m, off + sizeof(*th), opts, hlen, NULL, NULL, pd->af)) return 0; while (hlen >= TCPOLEN_SACKLEN) { olen = opt[1]; switch (*opt) { case TCPOPT_EOL: /* FALLTHROUGH */ case TCPOPT_NOP: opt++; hlen--; break; case TCPOPT_SACK: if (olen > hlen) olen = hlen; if (olen >= TCPOLEN_SACKLEN) { for (i = 2; i + TCPOLEN_SACK <= olen; i += TCPOLEN_SACK) { memcpy(&sack, &opt[i], sizeof(sack)); pf_change_proto_a(m, &sack.start, &th->th_sum, htonl(ntohl(sack.start) - dst->seqdiff), 0); pf_change_proto_a(m, &sack.end, &th->th_sum, htonl(ntohl(sack.end) - dst->seqdiff), 0); memcpy(&opt[i], &sack, sizeof(sack)); } copyback = 1; } /* FALLTHROUGH */ default: if (olen < 2) olen = 2; hlen -= olen; opt += olen; } } if (copyback) m_copyback(m, off + sizeof(*th), thoptlen, (caddr_t)opts); return (copyback); } static void pf_send_tcp(struct mbuf *replyto, const struct pf_rule *r, sa_family_t af, const struct pf_addr *saddr, const struct pf_addr *daddr, u_int16_t sport, u_int16_t dport, u_int32_t seq, u_int32_t ack, u_int8_t flags, u_int16_t win, u_int16_t mss, u_int8_t ttl, int tag, u_int16_t rtag, struct ifnet *ifp) { struct pf_send_entry *pfse; struct mbuf *m; int len, tlen; #ifdef INET struct ip *h = NULL; #endif /* INET */ #ifdef INET6 struct ip6_hdr *h6 = NULL; #endif /* INET6 */ struct tcphdr *th; char *opt; struct pf_mtag *pf_mtag; len = 0; th = NULL; /* maximum segment size tcp option */ tlen = sizeof(struct tcphdr); if (mss) tlen += 4; switch (af) { #ifdef INET case AF_INET: len = sizeof(struct ip) + tlen; break; #endif /* INET */ #ifdef INET6 case AF_INET6: len = sizeof(struct ip6_hdr) + tlen; break; #endif /* INET6 */ default: panic("%s: unsupported af %d", __func__, af); } /* Allocate outgoing queue entry, mbuf and mbuf tag. */ pfse = malloc(sizeof(*pfse), M_PFTEMP, M_NOWAIT); if (pfse == NULL) return; m = m_gethdr(M_NOWAIT, MT_DATA); if (m == NULL) { free(pfse, M_PFTEMP); return; } #ifdef MAC mac_netinet_firewall_send(m); #endif if ((pf_mtag = pf_get_mtag(m)) == NULL) { free(pfse, M_PFTEMP); m_freem(m); return; } if (tag) m->m_flags |= M_SKIP_FIREWALL; pf_mtag->tag = rtag; if (r != NULL && r->rtableid >= 0) M_SETFIB(m, r->rtableid); #ifdef ALTQ if (r != NULL && r->qid) { pf_mtag->qid = r->qid; /* add hints for ecn */ pf_mtag->hdr = mtod(m, struct ip *); } #endif /* ALTQ */ m->m_data += max_linkhdr; m->m_pkthdr.len = m->m_len = len; m->m_pkthdr.rcvif = NULL; bzero(m->m_data, len); switch (af) { #ifdef INET case AF_INET: h = mtod(m, struct ip *); /* IP header fields included in the TCP checksum */ h->ip_p = IPPROTO_TCP; h->ip_len = htons(tlen); h->ip_src.s_addr = saddr->v4.s_addr; h->ip_dst.s_addr = daddr->v4.s_addr; th = (struct tcphdr *)((caddr_t)h + sizeof(struct ip)); break; #endif /* INET */ #ifdef INET6 case AF_INET6: h6 = mtod(m, struct ip6_hdr *); /* IP header fields included in the TCP checksum */ h6->ip6_nxt = IPPROTO_TCP; h6->ip6_plen = htons(tlen); memcpy(&h6->ip6_src, &saddr->v6, sizeof(struct in6_addr)); memcpy(&h6->ip6_dst, &daddr->v6, sizeof(struct in6_addr)); th = (struct tcphdr *)((caddr_t)h6 + sizeof(struct ip6_hdr)); break; #endif /* INET6 */ } /* TCP header */ th->th_sport = sport; th->th_dport = dport; th->th_seq = htonl(seq); th->th_ack = htonl(ack); th->th_off = tlen >> 2; th->th_flags = flags; th->th_win = htons(win); if (mss) { opt = (char *)(th + 1); opt[0] = TCPOPT_MAXSEG; opt[1] = 4; HTONS(mss); bcopy((caddr_t)&mss, (caddr_t)(opt + 2), 2); } switch (af) { #ifdef INET case AF_INET: /* TCP checksum */ th->th_sum = in_cksum(m, len); /* Finish the IP header */ h->ip_v = 4; h->ip_hl = sizeof(*h) >> 2; h->ip_tos = IPTOS_LOWDELAY; h->ip_off = htons(V_path_mtu_discovery ? IP_DF : 0); h->ip_len = htons(len); h->ip_ttl = ttl ? ttl : V_ip_defttl; h->ip_sum = 0; pfse->pfse_type = PFSE_IP; break; #endif /* INET */ #ifdef INET6 case AF_INET6: /* TCP checksum */ th->th_sum = in6_cksum(m, IPPROTO_TCP, sizeof(struct ip6_hdr), tlen); h6->ip6_vfc |= IPV6_VERSION; h6->ip6_hlim = IPV6_DEFHLIM; pfse->pfse_type = PFSE_IP6; break; #endif /* INET6 */ } pfse->pfse_m = m; pf_send(pfse); } static int pf_ieee8021q_setpcp(struct mbuf *m, u_int8_t prio) { struct m_tag *mtag; KASSERT(prio <= PF_PRIO_MAX, ("%s with invalid pcp", __func__)); mtag = m_tag_locate(m, MTAG_8021Q, MTAG_8021Q_PCP_OUT, NULL); if (mtag == NULL) { mtag = m_tag_alloc(MTAG_8021Q, MTAG_8021Q_PCP_OUT, sizeof(uint8_t), M_NOWAIT); if (mtag == NULL) return (ENOMEM); m_tag_prepend(m, mtag); } *(uint8_t *)(mtag + 1) = prio; return (0); } static int pf_match_ieee8021q_pcp(u_int8_t prio, struct mbuf *m) { struct m_tag *mtag; u_int8_t mpcp; mtag = m_tag_locate(m, MTAG_8021Q, MTAG_8021Q_PCP_IN, NULL); if (mtag == NULL) return (0); if (prio == PF_PRIO_ZERO) prio = 0; mpcp = *(uint8_t *)(mtag + 1); return (mpcp == prio); } static void pf_send_icmp(struct mbuf *m, u_int8_t type, u_int8_t code, sa_family_t af, struct pf_rule *r) { struct pf_send_entry *pfse; struct mbuf *m0; struct pf_mtag *pf_mtag; /* Allocate outgoing queue entry, mbuf and mbuf tag. */ pfse = malloc(sizeof(*pfse), M_PFTEMP, M_NOWAIT); if (pfse == NULL) return; if ((m0 = m_copypacket(m, M_NOWAIT)) == NULL) { free(pfse, M_PFTEMP); return; } if ((pf_mtag = pf_get_mtag(m0)) == NULL) { free(pfse, M_PFTEMP); return; } /* XXX: revisit */ m0->m_flags |= M_SKIP_FIREWALL; if (r->rtableid >= 0) M_SETFIB(m0, r->rtableid); #ifdef ALTQ if (r->qid) { pf_mtag->qid = r->qid; /* add hints for ecn */ pf_mtag->hdr = mtod(m0, struct ip *); } #endif /* ALTQ */ switch (af) { #ifdef INET case AF_INET: pfse->pfse_type = PFSE_ICMP; break; #endif /* INET */ #ifdef INET6 case AF_INET6: pfse->pfse_type = PFSE_ICMP6; break; #endif /* INET6 */ } pfse->pfse_m = m0; pfse->icmpopts.type = type; pfse->icmpopts.code = code; pf_send(pfse); } /* * Return 1 if the addresses a and b match (with mask m), otherwise return 0. * If n is 0, they match if they are equal. If n is != 0, they match if they * are different. */ int pf_match_addr(u_int8_t n, struct pf_addr *a, struct pf_addr *m, struct pf_addr *b, sa_family_t af) { int match = 0; switch (af) { #ifdef INET case AF_INET: if ((a->addr32[0] & m->addr32[0]) == (b->addr32[0] & m->addr32[0])) match++; break; #endif /* INET */ #ifdef INET6 case AF_INET6: if (((a->addr32[0] & m->addr32[0]) == (b->addr32[0] & m->addr32[0])) && ((a->addr32[1] & m->addr32[1]) == (b->addr32[1] & m->addr32[1])) && ((a->addr32[2] & m->addr32[2]) == (b->addr32[2] & m->addr32[2])) && ((a->addr32[3] & m->addr32[3]) == (b->addr32[3] & m->addr32[3]))) match++; break; #endif /* INET6 */ } if (match) { if (n) return (0); else return (1); } else { if (n) return (1); else return (0); } } /* * Return 1 if b <= a <= e, otherwise return 0. */ int pf_match_addr_range(struct pf_addr *b, struct pf_addr *e, struct pf_addr *a, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: if ((ntohl(a->addr32[0]) < ntohl(b->addr32[0])) || (ntohl(a->addr32[0]) > ntohl(e->addr32[0]))) return (0); break; #endif /* INET */ #ifdef INET6 case AF_INET6: { int i; /* check a >= b */ for (i = 0; i < 4; ++i) if (ntohl(a->addr32[i]) > ntohl(b->addr32[i])) break; else if (ntohl(a->addr32[i]) < ntohl(b->addr32[i])) return (0); /* check a <= e */ for (i = 0; i < 4; ++i) if (ntohl(a->addr32[i]) < ntohl(e->addr32[i])) break; else if (ntohl(a->addr32[i]) > ntohl(e->addr32[i])) return (0); break; } #endif /* INET6 */ } return (1); } static int pf_match(u_int8_t op, u_int32_t a1, u_int32_t a2, u_int32_t p) { switch (op) { case PF_OP_IRG: return ((p > a1) && (p < a2)); case PF_OP_XRG: return ((p < a1) || (p > a2)); case PF_OP_RRG: return ((p >= a1) && (p <= a2)); case PF_OP_EQ: return (p == a1); case PF_OP_NE: return (p != a1); case PF_OP_LT: return (p < a1); case PF_OP_LE: return (p <= a1); case PF_OP_GT: return (p > a1); case PF_OP_GE: return (p >= a1); } return (0); /* never reached */ } int pf_match_port(u_int8_t op, u_int16_t a1, u_int16_t a2, u_int16_t p) { NTOHS(a1); NTOHS(a2); NTOHS(p); return (pf_match(op, a1, a2, p)); } static int pf_match_uid(u_int8_t op, uid_t a1, uid_t a2, uid_t u) { if (u == UID_MAX && op != PF_OP_EQ && op != PF_OP_NE) return (0); return (pf_match(op, a1, a2, u)); } static int pf_match_gid(u_int8_t op, gid_t a1, gid_t a2, gid_t g) { if (g == GID_MAX && op != PF_OP_EQ && op != PF_OP_NE) return (0); return (pf_match(op, a1, a2, g)); } int pf_match_tag(struct mbuf *m, struct pf_rule *r, int *tag, int mtag) { if (*tag == -1) *tag = mtag; return ((!r->match_tag_not && r->match_tag == *tag) || (r->match_tag_not && r->match_tag != *tag)); } int pf_tag_packet(struct mbuf *m, struct pf_pdesc *pd, int tag) { KASSERT(tag > 0, ("%s: tag %d", __func__, tag)); if (pd->pf_mtag == NULL && ((pd->pf_mtag = pf_get_mtag(m)) == NULL)) return (ENOMEM); pd->pf_mtag->tag = tag; return (0); } #define PF_ANCHOR_STACKSIZE 32 struct pf_anchor_stackframe { struct pf_ruleset *rs; struct pf_rule *r; /* XXX: + match bit */ struct pf_anchor *child; }; /* * XXX: We rely on malloc(9) returning pointer aligned addresses. */ #define PF_ANCHORSTACK_MATCH 0x00000001 #define PF_ANCHORSTACK_MASK (PF_ANCHORSTACK_MATCH) #define PF_ANCHOR_MATCH(f) ((uintptr_t)(f)->r & PF_ANCHORSTACK_MATCH) #define PF_ANCHOR_RULE(f) (struct pf_rule *) \ ((uintptr_t)(f)->r & ~PF_ANCHORSTACK_MASK) #define PF_ANCHOR_SET_MATCH(f) do { (f)->r = (void *) \ ((uintptr_t)(f)->r | PF_ANCHORSTACK_MATCH); \ } while (0) void pf_step_into_anchor(struct pf_anchor_stackframe *stack, int *depth, struct pf_ruleset **rs, int n, struct pf_rule **r, struct pf_rule **a, int *match) { struct pf_anchor_stackframe *f; PF_RULES_RASSERT(); if (match) *match = 0; if (*depth >= PF_ANCHOR_STACKSIZE) { printf("%s: anchor stack overflow on %s\n", __func__, (*r)->anchor->name); *r = TAILQ_NEXT(*r, entries); return; } else if (*depth == 0 && a != NULL) *a = *r; f = stack + (*depth)++; f->rs = *rs; f->r = *r; if ((*r)->anchor_wildcard) { struct pf_anchor_node *parent = &(*r)->anchor->children; if ((f->child = RB_MIN(pf_anchor_node, parent)) == NULL) { *r = NULL; return; } *rs = &f->child->ruleset; } else { f->child = NULL; *rs = &(*r)->anchor->ruleset; } *r = TAILQ_FIRST((*rs)->rules[n].active.ptr); } int pf_step_out_of_anchor(struct pf_anchor_stackframe *stack, int *depth, struct pf_ruleset **rs, int n, struct pf_rule **r, struct pf_rule **a, int *match) { struct pf_anchor_stackframe *f; struct pf_rule *fr; int quick = 0; PF_RULES_RASSERT(); do { if (*depth <= 0) break; f = stack + *depth - 1; fr = PF_ANCHOR_RULE(f); if (f->child != NULL) { struct pf_anchor_node *parent; /* * This block traverses through * a wildcard anchor. */ parent = &fr->anchor->children; if (match != NULL && *match) { /* * If any of "*" matched, then * "foo/ *" matched, mark frame * appropriately. */ PF_ANCHOR_SET_MATCH(f); *match = 0; } f->child = RB_NEXT(pf_anchor_node, parent, f->child); if (f->child != NULL) { *rs = &f->child->ruleset; *r = TAILQ_FIRST((*rs)->rules[n].active.ptr); if (*r == NULL) continue; else break; } } (*depth)--; if (*depth == 0 && a != NULL) *a = NULL; *rs = f->rs; if (PF_ANCHOR_MATCH(f) || (match != NULL && *match)) quick = fr->quick; *r = TAILQ_NEXT(fr, entries); } while (*r == NULL); return (quick); } #ifdef INET6 void pf_poolmask(struct pf_addr *naddr, struct pf_addr *raddr, struct pf_addr *rmask, struct pf_addr *saddr, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: naddr->addr32[0] = (raddr->addr32[0] & rmask->addr32[0]) | ((rmask->addr32[0] ^ 0xffffffff ) & saddr->addr32[0]); break; #endif /* INET */ case AF_INET6: naddr->addr32[0] = (raddr->addr32[0] & rmask->addr32[0]) | ((rmask->addr32[0] ^ 0xffffffff ) & saddr->addr32[0]); naddr->addr32[1] = (raddr->addr32[1] & rmask->addr32[1]) | ((rmask->addr32[1] ^ 0xffffffff ) & saddr->addr32[1]); naddr->addr32[2] = (raddr->addr32[2] & rmask->addr32[2]) | ((rmask->addr32[2] ^ 0xffffffff ) & saddr->addr32[2]); naddr->addr32[3] = (raddr->addr32[3] & rmask->addr32[3]) | ((rmask->addr32[3] ^ 0xffffffff ) & saddr->addr32[3]); break; } } void pf_addr_inc(struct pf_addr *addr, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: addr->addr32[0] = htonl(ntohl(addr->addr32[0]) + 1); break; #endif /* INET */ case AF_INET6: if (addr->addr32[3] == 0xffffffff) { addr->addr32[3] = 0; if (addr->addr32[2] == 0xffffffff) { addr->addr32[2] = 0; if (addr->addr32[1] == 0xffffffff) { addr->addr32[1] = 0; addr->addr32[0] = htonl(ntohl(addr->addr32[0]) + 1); } else addr->addr32[1] = htonl(ntohl(addr->addr32[1]) + 1); } else addr->addr32[2] = htonl(ntohl(addr->addr32[2]) + 1); } else addr->addr32[3] = htonl(ntohl(addr->addr32[3]) + 1); break; } } #endif /* INET6 */ int pf_socket_lookup(int direction, struct pf_pdesc *pd, struct mbuf *m) { struct pf_addr *saddr, *daddr; u_int16_t sport, dport; struct inpcbinfo *pi; struct inpcb *inp; pd->lookup.uid = UID_MAX; pd->lookup.gid = GID_MAX; switch (pd->proto) { case IPPROTO_TCP: if (pd->hdr.tcp == NULL) return (-1); sport = pd->hdr.tcp->th_sport; dport = pd->hdr.tcp->th_dport; pi = &V_tcbinfo; break; case IPPROTO_UDP: if (pd->hdr.udp == NULL) return (-1); sport = pd->hdr.udp->uh_sport; dport = pd->hdr.udp->uh_dport; pi = &V_udbinfo; break; default: return (-1); } if (direction == PF_IN) { saddr = pd->src; daddr = pd->dst; } else { u_int16_t p; p = sport; sport = dport; dport = p; saddr = pd->dst; daddr = pd->src; } switch (pd->af) { #ifdef INET case AF_INET: inp = in_pcblookup_mbuf(pi, saddr->v4, sport, daddr->v4, dport, INPLOOKUP_RLOCKPCB, NULL, m); if (inp == NULL) { inp = in_pcblookup_mbuf(pi, saddr->v4, sport, daddr->v4, dport, INPLOOKUP_WILDCARD | INPLOOKUP_RLOCKPCB, NULL, m); if (inp == NULL) return (-1); } break; #endif /* INET */ #ifdef INET6 case AF_INET6: inp = in6_pcblookup_mbuf(pi, &saddr->v6, sport, &daddr->v6, dport, INPLOOKUP_RLOCKPCB, NULL, m); if (inp == NULL) { inp = in6_pcblookup_mbuf(pi, &saddr->v6, sport, &daddr->v6, dport, INPLOOKUP_WILDCARD | INPLOOKUP_RLOCKPCB, NULL, m); if (inp == NULL) return (-1); } break; #endif /* INET6 */ default: return (-1); } INP_RLOCK_ASSERT(inp); pd->lookup.uid = inp->inp_cred->cr_uid; pd->lookup.gid = inp->inp_cred->cr_groups[0]; INP_RUNLOCK(inp); return (1); } static u_int8_t pf_get_wscale(struct mbuf *m, int off, u_int16_t th_off, sa_family_t af) { int hlen; u_int8_t hdr[60]; u_int8_t *opt, optlen; u_int8_t wscale = 0; hlen = th_off << 2; /* hlen <= sizeof(hdr) */ if (hlen <= sizeof(struct tcphdr)) return (0); if (!pf_pull_hdr(m, off, hdr, hlen, NULL, NULL, af)) return (0); opt = hdr + sizeof(struct tcphdr); hlen -= sizeof(struct tcphdr); while (hlen >= 3) { switch (*opt) { case TCPOPT_EOL: case TCPOPT_NOP: ++opt; --hlen; break; case TCPOPT_WINDOW: wscale = opt[2]; if (wscale > TCP_MAX_WINSHIFT) wscale = TCP_MAX_WINSHIFT; wscale |= PF_WSCALE_FLAG; /* FALLTHROUGH */ default: optlen = opt[1]; if (optlen < 2) optlen = 2; hlen -= optlen; opt += optlen; break; } } return (wscale); } static u_int16_t pf_get_mss(struct mbuf *m, int off, u_int16_t th_off, sa_family_t af) { int hlen; u_int8_t hdr[60]; u_int8_t *opt, optlen; u_int16_t mss = V_tcp_mssdflt; hlen = th_off << 2; /* hlen <= sizeof(hdr) */ if (hlen <= sizeof(struct tcphdr)) return (0); if (!pf_pull_hdr(m, off, hdr, hlen, NULL, NULL, af)) return (0); opt = hdr + sizeof(struct tcphdr); hlen -= sizeof(struct tcphdr); while (hlen >= TCPOLEN_MAXSEG) { switch (*opt) { case TCPOPT_EOL: case TCPOPT_NOP: ++opt; --hlen; break; case TCPOPT_MAXSEG: bcopy((caddr_t)(opt + 2), (caddr_t)&mss, 2); NTOHS(mss); /* FALLTHROUGH */ default: optlen = opt[1]; if (optlen < 2) optlen = 2; hlen -= optlen; opt += optlen; break; } } return (mss); } static u_int16_t pf_calc_mss(struct pf_addr *addr, sa_family_t af, int rtableid, u_int16_t offer) { #ifdef INET struct nhop4_basic nh4; #endif /* INET */ #ifdef INET6 struct nhop6_basic nh6; struct in6_addr dst6; uint32_t scopeid; #endif /* INET6 */ int hlen = 0; uint16_t mss = 0; switch (af) { #ifdef INET case AF_INET: hlen = sizeof(struct ip); if (fib4_lookup_nh_basic(rtableid, addr->v4, 0, 0, &nh4) == 0) mss = nh4.nh_mtu - hlen - sizeof(struct tcphdr); break; #endif /* INET */ #ifdef INET6 case AF_INET6: hlen = sizeof(struct ip6_hdr); in6_splitscope(&addr->v6, &dst6, &scopeid); if (fib6_lookup_nh_basic(rtableid, &dst6, scopeid, 0,0,&nh6)==0) mss = nh6.nh_mtu - hlen - sizeof(struct tcphdr); break; #endif /* INET6 */ } mss = max(V_tcp_mssdflt, mss); mss = min(mss, offer); mss = max(mss, 64); /* sanity - at least max opt space */ return (mss); } static u_int32_t pf_tcp_iss(struct pf_pdesc *pd) { MD5_CTX ctx; u_int32_t digest[4]; if (V_pf_tcp_secret_init == 0) { read_random(&V_pf_tcp_secret, sizeof(V_pf_tcp_secret)); MD5Init(&V_pf_tcp_secret_ctx); MD5Update(&V_pf_tcp_secret_ctx, V_pf_tcp_secret, sizeof(V_pf_tcp_secret)); V_pf_tcp_secret_init = 1; } ctx = V_pf_tcp_secret_ctx; MD5Update(&ctx, (char *)&pd->hdr.tcp->th_sport, sizeof(u_short)); MD5Update(&ctx, (char *)&pd->hdr.tcp->th_dport, sizeof(u_short)); if (pd->af == AF_INET6) { MD5Update(&ctx, (char *)&pd->src->v6, sizeof(struct in6_addr)); MD5Update(&ctx, (char *)&pd->dst->v6, sizeof(struct in6_addr)); } else { MD5Update(&ctx, (char *)&pd->src->v4, sizeof(struct in_addr)); MD5Update(&ctx, (char *)&pd->dst->v4, sizeof(struct in_addr)); } MD5Final((u_char *)digest, &ctx); V_pf_tcp_iss_off += 4096; #define ISN_RANDOM_INCREMENT (4096 - 1) return (digest[0] + (arc4random() & ISN_RANDOM_INCREMENT) + V_pf_tcp_iss_off); #undef ISN_RANDOM_INCREMENT } static int pf_test_rule(struct pf_rule **rm, struct pf_state **sm, int direction, struct pfi_kif *kif, struct mbuf *m, int off, struct pf_pdesc *pd, struct pf_rule **am, struct pf_ruleset **rsm, struct inpcb *inp) { struct pf_rule *nr = NULL; struct pf_addr * const saddr = pd->src; struct pf_addr * const daddr = pd->dst; sa_family_t af = pd->af; struct pf_rule *r, *a = NULL; struct pf_ruleset *ruleset = NULL; struct pf_src_node *nsn = NULL; struct tcphdr *th = pd->hdr.tcp; struct pf_state_key *sk = NULL, *nk = NULL; u_short reason; int rewrite = 0, hdrlen = 0; int tag = -1, rtableid = -1; int asd = 0; int match = 0; int state_icmp = 0; u_int16_t sport = 0, dport = 0; u_int16_t bproto_sum = 0, bip_sum = 0; u_int8_t icmptype = 0, icmpcode = 0; struct pf_anchor_stackframe anchor_stack[PF_ANCHOR_STACKSIZE]; PF_RULES_RASSERT(); if (inp != NULL) { INP_LOCK_ASSERT(inp); pd->lookup.uid = inp->inp_cred->cr_uid; pd->lookup.gid = inp->inp_cred->cr_groups[0]; pd->lookup.done = 1; } switch (pd->proto) { case IPPROTO_TCP: sport = th->th_sport; dport = th->th_dport; hdrlen = sizeof(*th); break; case IPPROTO_UDP: sport = pd->hdr.udp->uh_sport; dport = pd->hdr.udp->uh_dport; hdrlen = sizeof(*pd->hdr.udp); break; #ifdef INET case IPPROTO_ICMP: if (pd->af != AF_INET) break; sport = dport = pd->hdr.icmp->icmp_id; hdrlen = sizeof(*pd->hdr.icmp); icmptype = pd->hdr.icmp->icmp_type; icmpcode = pd->hdr.icmp->icmp_code; if (icmptype == ICMP_UNREACH || icmptype == ICMP_SOURCEQUENCH || icmptype == ICMP_REDIRECT || icmptype == ICMP_TIMXCEED || icmptype == ICMP_PARAMPROB) state_icmp++; break; #endif /* INET */ #ifdef INET6 case IPPROTO_ICMPV6: if (af != AF_INET6) break; sport = dport = pd->hdr.icmp6->icmp6_id; hdrlen = sizeof(*pd->hdr.icmp6); icmptype = pd->hdr.icmp6->icmp6_type; icmpcode = pd->hdr.icmp6->icmp6_code; if (icmptype == ICMP6_DST_UNREACH || icmptype == ICMP6_PACKET_TOO_BIG || icmptype == ICMP6_TIME_EXCEEDED || icmptype == ICMP6_PARAM_PROB) state_icmp++; break; #endif /* INET6 */ default: sport = dport = hdrlen = 0; break; } r = TAILQ_FIRST(pf_main_ruleset.rules[PF_RULESET_FILTER].active.ptr); /* check packet for BINAT/NAT/RDR */ if ((nr = pf_get_translation(pd, m, off, direction, kif, &nsn, &sk, &nk, saddr, daddr, sport, dport, anchor_stack)) != NULL) { KASSERT(sk != NULL, ("%s: null sk", __func__)); KASSERT(nk != NULL, ("%s: null nk", __func__)); if (pd->ip_sum) bip_sum = *pd->ip_sum; switch (pd->proto) { case IPPROTO_TCP: bproto_sum = th->th_sum; pd->proto_sum = &th->th_sum; if (PF_ANEQ(saddr, &nk->addr[pd->sidx], af) || nk->port[pd->sidx] != sport) { pf_change_ap(m, saddr, &th->th_sport, pd->ip_sum, &th->th_sum, &nk->addr[pd->sidx], nk->port[pd->sidx], 0, af); pd->sport = &th->th_sport; sport = th->th_sport; } if (PF_ANEQ(daddr, &nk->addr[pd->didx], af) || nk->port[pd->didx] != dport) { pf_change_ap(m, daddr, &th->th_dport, pd->ip_sum, &th->th_sum, &nk->addr[pd->didx], nk->port[pd->didx], 0, af); dport = th->th_dport; pd->dport = &th->th_dport; } rewrite++; break; case IPPROTO_UDP: bproto_sum = pd->hdr.udp->uh_sum; pd->proto_sum = &pd->hdr.udp->uh_sum; if (PF_ANEQ(saddr, &nk->addr[pd->sidx], af) || nk->port[pd->sidx] != sport) { pf_change_ap(m, saddr, &pd->hdr.udp->uh_sport, pd->ip_sum, &pd->hdr.udp->uh_sum, &nk->addr[pd->sidx], nk->port[pd->sidx], 1, af); sport = pd->hdr.udp->uh_sport; pd->sport = &pd->hdr.udp->uh_sport; } if (PF_ANEQ(daddr, &nk->addr[pd->didx], af) || nk->port[pd->didx] != dport) { pf_change_ap(m, daddr, &pd->hdr.udp->uh_dport, pd->ip_sum, &pd->hdr.udp->uh_sum, &nk->addr[pd->didx], nk->port[pd->didx], 1, af); dport = pd->hdr.udp->uh_dport; pd->dport = &pd->hdr.udp->uh_dport; } rewrite++; break; #ifdef INET case IPPROTO_ICMP: nk->port[0] = nk->port[1]; if (PF_ANEQ(saddr, &nk->addr[pd->sidx], AF_INET)) pf_change_a(&saddr->v4.s_addr, pd->ip_sum, nk->addr[pd->sidx].v4.s_addr, 0); if (PF_ANEQ(daddr, &nk->addr[pd->didx], AF_INET)) pf_change_a(&daddr->v4.s_addr, pd->ip_sum, nk->addr[pd->didx].v4.s_addr, 0); if (nk->port[1] != pd->hdr.icmp->icmp_id) { pd->hdr.icmp->icmp_cksum = pf_cksum_fixup( pd->hdr.icmp->icmp_cksum, sport, nk->port[1], 0); pd->hdr.icmp->icmp_id = nk->port[1]; pd->sport = &pd->hdr.icmp->icmp_id; } m_copyback(m, off, ICMP_MINLEN, (caddr_t)pd->hdr.icmp); break; #endif /* INET */ #ifdef INET6 case IPPROTO_ICMPV6: nk->port[0] = nk->port[1]; if (PF_ANEQ(saddr, &nk->addr[pd->sidx], AF_INET6)) pf_change_a6(saddr, &pd->hdr.icmp6->icmp6_cksum, &nk->addr[pd->sidx], 0); if (PF_ANEQ(daddr, &nk->addr[pd->didx], AF_INET6)) pf_change_a6(daddr, &pd->hdr.icmp6->icmp6_cksum, &nk->addr[pd->didx], 0); rewrite++; break; #endif /* INET */ default: switch (af) { #ifdef INET case AF_INET: if (PF_ANEQ(saddr, &nk->addr[pd->sidx], AF_INET)) pf_change_a(&saddr->v4.s_addr, pd->ip_sum, nk->addr[pd->sidx].v4.s_addr, 0); if (PF_ANEQ(daddr, &nk->addr[pd->didx], AF_INET)) pf_change_a(&daddr->v4.s_addr, pd->ip_sum, nk->addr[pd->didx].v4.s_addr, 0); break; #endif /* INET */ #ifdef INET6 case AF_INET6: if (PF_ANEQ(saddr, &nk->addr[pd->sidx], AF_INET6)) PF_ACPY(saddr, &nk->addr[pd->sidx], af); if (PF_ANEQ(daddr, &nk->addr[pd->didx], AF_INET6)) PF_ACPY(saddr, &nk->addr[pd->didx], af); break; #endif /* INET */ } break; } if (nr->natpass) r = NULL; pd->nat_rule = nr; } while (r != NULL) { r->evaluations++; if (pfi_kif_match(r->kif, kif) == r->ifnot) r = r->skip[PF_SKIP_IFP].ptr; else if (r->direction && r->direction != direction) r = r->skip[PF_SKIP_DIR].ptr; else if (r->af && r->af != af) r = r->skip[PF_SKIP_AF].ptr; else if (r->proto && r->proto != pd->proto) r = r->skip[PF_SKIP_PROTO].ptr; else if (PF_MISMATCHAW(&r->src.addr, saddr, af, r->src.neg, kif, M_GETFIB(m))) r = r->skip[PF_SKIP_SRC_ADDR].ptr; /* tcp/udp only. port_op always 0 in other cases */ else if (r->src.port_op && !pf_match_port(r->src.port_op, r->src.port[0], r->src.port[1], sport)) r = r->skip[PF_SKIP_SRC_PORT].ptr; else if (PF_MISMATCHAW(&r->dst.addr, daddr, af, r->dst.neg, NULL, M_GETFIB(m))) r = r->skip[PF_SKIP_DST_ADDR].ptr; /* tcp/udp only. port_op always 0 in other cases */ else if (r->dst.port_op && !pf_match_port(r->dst.port_op, r->dst.port[0], r->dst.port[1], dport)) r = r->skip[PF_SKIP_DST_PORT].ptr; /* icmp only. type always 0 in other cases */ else if (r->type && r->type != icmptype + 1) r = TAILQ_NEXT(r, entries); /* icmp only. type always 0 in other cases */ else if (r->code && r->code != icmpcode + 1) r = TAILQ_NEXT(r, entries); else if (r->tos && !(r->tos == pd->tos)) r = TAILQ_NEXT(r, entries); else if (r->rule_flag & PFRULE_FRAGMENT) r = TAILQ_NEXT(r, entries); else if (pd->proto == IPPROTO_TCP && (r->flagset & th->th_flags) != r->flags) r = TAILQ_NEXT(r, entries); /* tcp/udp only. uid.op always 0 in other cases */ else if (r->uid.op && (pd->lookup.done || (pd->lookup.done = pf_socket_lookup(direction, pd, m), 1)) && !pf_match_uid(r->uid.op, r->uid.uid[0], r->uid.uid[1], pd->lookup.uid)) r = TAILQ_NEXT(r, entries); /* tcp/udp only. gid.op always 0 in other cases */ else if (r->gid.op && (pd->lookup.done || (pd->lookup.done = pf_socket_lookup(direction, pd, m), 1)) && !pf_match_gid(r->gid.op, r->gid.gid[0], r->gid.gid[1], pd->lookup.gid)) r = TAILQ_NEXT(r, entries); else if (r->prio && !pf_match_ieee8021q_pcp(r->prio, m)) r = TAILQ_NEXT(r, entries); else if (r->prob && r->prob <= arc4random()) r = TAILQ_NEXT(r, entries); else if (r->match_tag && !pf_match_tag(m, r, &tag, pd->pf_mtag ? pd->pf_mtag->tag : 0)) r = TAILQ_NEXT(r, entries); else if (r->os_fingerprint != PF_OSFP_ANY && (pd->proto != IPPROTO_TCP || !pf_osfp_match( pf_osfp_fingerprint(pd, m, off, th), r->os_fingerprint))) r = TAILQ_NEXT(r, entries); else { if (r->tag) tag = r->tag; if (r->rtableid >= 0) rtableid = r->rtableid; if (r->anchor == NULL) { match = 1; *rm = r; *am = a; *rsm = ruleset; if ((*rm)->quick) break; r = TAILQ_NEXT(r, entries); } else pf_step_into_anchor(anchor_stack, &asd, &ruleset, PF_RULESET_FILTER, &r, &a, &match); } if (r == NULL && pf_step_out_of_anchor(anchor_stack, &asd, &ruleset, PF_RULESET_FILTER, &r, &a, &match)) break; } r = *rm; a = *am; ruleset = *rsm; REASON_SET(&reason, PFRES_MATCH); if (r->log || (nr != NULL && nr->log)) { if (rewrite) m_copyback(m, off, hdrlen, pd->hdr.any); PFLOG_PACKET(kif, m, af, direction, reason, r->log ? r : nr, a, ruleset, pd, 1); } if ((r->action == PF_DROP) && ((r->rule_flag & PFRULE_RETURNRST) || (r->rule_flag & PFRULE_RETURNICMP) || (r->rule_flag & PFRULE_RETURN))) { /* undo NAT changes, if they have taken place */ if (nr != NULL) { PF_ACPY(saddr, &sk->addr[pd->sidx], af); PF_ACPY(daddr, &sk->addr[pd->didx], af); if (pd->sport) *pd->sport = sk->port[pd->sidx]; if (pd->dport) *pd->dport = sk->port[pd->didx]; if (pd->proto_sum) *pd->proto_sum = bproto_sum; if (pd->ip_sum) *pd->ip_sum = bip_sum; m_copyback(m, off, hdrlen, pd->hdr.any); } if (pd->proto == IPPROTO_TCP && ((r->rule_flag & PFRULE_RETURNRST) || (r->rule_flag & PFRULE_RETURN)) && !(th->th_flags & TH_RST)) { u_int32_t ack = ntohl(th->th_seq) + pd->p_len; int len = 0; #ifdef INET struct ip *h4; #endif #ifdef INET6 struct ip6_hdr *h6; #endif switch (af) { #ifdef INET case AF_INET: h4 = mtod(m, struct ip *); len = ntohs(h4->ip_len) - off; break; #endif #ifdef INET6 case AF_INET6: h6 = mtod(m, struct ip6_hdr *); len = ntohs(h6->ip6_plen) - (off - sizeof(*h6)); break; #endif } if (pf_check_proto_cksum(m, off, len, IPPROTO_TCP, af)) REASON_SET(&reason, PFRES_PROTCKSUM); else { if (th->th_flags & TH_SYN) ack++; if (th->th_flags & TH_FIN) ack++; pf_send_tcp(m, r, af, pd->dst, pd->src, th->th_dport, th->th_sport, ntohl(th->th_ack), ack, TH_RST|TH_ACK, 0, 0, r->return_ttl, 1, 0, kif->pfik_ifp); } } else if (pd->proto != IPPROTO_ICMP && af == AF_INET && r->return_icmp) pf_send_icmp(m, r->return_icmp >> 8, r->return_icmp & 255, af, r); else if (pd->proto != IPPROTO_ICMPV6 && af == AF_INET6 && r->return_icmp6) pf_send_icmp(m, r->return_icmp6 >> 8, r->return_icmp6 & 255, af, r); } if (r->action == PF_DROP) goto cleanup; if (tag > 0 && pf_tag_packet(m, pd, tag)) { REASON_SET(&reason, PFRES_MEMORY); goto cleanup; } if (rtableid >= 0) M_SETFIB(m, rtableid); if (!state_icmp && (r->keep_state || nr != NULL || (pd->flags & PFDESC_TCP_NORM))) { int action; action = pf_create_state(r, nr, a, pd, nsn, nk, sk, m, off, sport, dport, &rewrite, kif, sm, tag, bproto_sum, bip_sum, hdrlen); if (action != PF_PASS) return (action); } else { if (sk != NULL) uma_zfree(V_pf_state_key_z, sk); if (nk != NULL) uma_zfree(V_pf_state_key_z, nk); } /* copy back packet headers if we performed NAT operations */ if (rewrite) m_copyback(m, off, hdrlen, pd->hdr.any); if (*sm != NULL && !((*sm)->state_flags & PFSTATE_NOSYNC) && direction == PF_OUT && pfsync_defer_ptr != NULL && pfsync_defer_ptr(*sm, m)) /* * We want the state created, but we dont * want to send this in case a partner * firewall has to know about it to allow * replies through it. */ return (PF_DEFER); return (PF_PASS); cleanup: if (sk != NULL) uma_zfree(V_pf_state_key_z, sk); if (nk != NULL) uma_zfree(V_pf_state_key_z, nk); return (PF_DROP); } static int pf_create_state(struct pf_rule *r, struct pf_rule *nr, struct pf_rule *a, struct pf_pdesc *pd, struct pf_src_node *nsn, struct pf_state_key *nk, struct pf_state_key *sk, struct mbuf *m, int off, u_int16_t sport, u_int16_t dport, int *rewrite, struct pfi_kif *kif, struct pf_state **sm, int tag, u_int16_t bproto_sum, u_int16_t bip_sum, int hdrlen) { struct pf_state *s = NULL; struct pf_src_node *sn = NULL; struct tcphdr *th = pd->hdr.tcp; u_int16_t mss = V_tcp_mssdflt; u_short reason; /* check maximums */ if (r->max_states && (counter_u64_fetch(r->states_cur) >= r->max_states)) { counter_u64_add(V_pf_status.lcounters[LCNT_STATES], 1); REASON_SET(&reason, PFRES_MAXSTATES); return (PF_DROP); } /* src node for filter rule */ if ((r->rule_flag & PFRULE_SRCTRACK || r->rpool.opts & PF_POOL_STICKYADDR) && pf_insert_src_node(&sn, r, pd->src, pd->af) != 0) { REASON_SET(&reason, PFRES_SRCLIMIT); goto csfailed; } /* src node for translation rule */ if (nr != NULL && (nr->rpool.opts & PF_POOL_STICKYADDR) && pf_insert_src_node(&nsn, nr, &sk->addr[pd->sidx], pd->af)) { REASON_SET(&reason, PFRES_SRCLIMIT); goto csfailed; } s = uma_zalloc(V_pf_state_z, M_NOWAIT | M_ZERO); if (s == NULL) { REASON_SET(&reason, PFRES_MEMORY); goto csfailed; } s->rule.ptr = r; s->nat_rule.ptr = nr; s->anchor.ptr = a; STATE_INC_COUNTERS(s); if (r->allow_opts) s->state_flags |= PFSTATE_ALLOWOPTS; if (r->rule_flag & PFRULE_STATESLOPPY) s->state_flags |= PFSTATE_SLOPPY; s->log = r->log & PF_LOG_ALL; s->sync_state = PFSYNC_S_NONE; if (nr != NULL) s->log |= nr->log & PF_LOG_ALL; switch (pd->proto) { case IPPROTO_TCP: s->src.seqlo = ntohl(th->th_seq); s->src.seqhi = s->src.seqlo + pd->p_len + 1; if ((th->th_flags & (TH_SYN|TH_ACK)) == TH_SYN && r->keep_state == PF_STATE_MODULATE) { /* Generate sequence number modulator */ if ((s->src.seqdiff = pf_tcp_iss(pd) - s->src.seqlo) == 0) s->src.seqdiff = 1; pf_change_proto_a(m, &th->th_seq, &th->th_sum, htonl(s->src.seqlo + s->src.seqdiff), 0); *rewrite = 1; } else s->src.seqdiff = 0; if (th->th_flags & TH_SYN) { s->src.seqhi++; s->src.wscale = pf_get_wscale(m, off, th->th_off, pd->af); } s->src.max_win = MAX(ntohs(th->th_win), 1); if (s->src.wscale & PF_WSCALE_MASK) { /* Remove scale factor from initial window */ int win = s->src.max_win; win += 1 << (s->src.wscale & PF_WSCALE_MASK); s->src.max_win = (win - 1) >> (s->src.wscale & PF_WSCALE_MASK); } if (th->th_flags & TH_FIN) s->src.seqhi++; s->dst.seqhi = 1; s->dst.max_win = 1; s->src.state = TCPS_SYN_SENT; s->dst.state = TCPS_CLOSED; s->timeout = PFTM_TCP_FIRST_PACKET; break; case IPPROTO_UDP: s->src.state = PFUDPS_SINGLE; s->dst.state = PFUDPS_NO_TRAFFIC; s->timeout = PFTM_UDP_FIRST_PACKET; break; case IPPROTO_ICMP: #ifdef INET6 case IPPROTO_ICMPV6: #endif s->timeout = PFTM_ICMP_FIRST_PACKET; break; default: s->src.state = PFOTHERS_SINGLE; s->dst.state = PFOTHERS_NO_TRAFFIC; s->timeout = PFTM_OTHER_FIRST_PACKET; } - if (r->rt && r->rt != PF_FASTROUTE) { + if (r->rt) { if (pf_map_addr(pd->af, r, pd->src, &s->rt_addr, NULL, &sn)) { REASON_SET(&reason, PFRES_MAPFAILED); pf_src_tree_remove_state(s); STATE_DEC_COUNTERS(s); uma_zfree(V_pf_state_z, s); goto csfailed; } s->rt_kif = r->rpool.cur->kif; } s->creation = time_uptime; s->expire = time_uptime; if (sn != NULL) s->src_node = sn; if (nsn != NULL) { /* XXX We only modify one side for now. */ PF_ACPY(&nsn->raddr, &nk->addr[1], pd->af); s->nat_src_node = nsn; } if (pd->proto == IPPROTO_TCP) { if ((pd->flags & PFDESC_TCP_NORM) && pf_normalize_tcp_init(m, off, pd, th, &s->src, &s->dst)) { REASON_SET(&reason, PFRES_MEMORY); pf_src_tree_remove_state(s); STATE_DEC_COUNTERS(s); uma_zfree(V_pf_state_z, s); return (PF_DROP); } if ((pd->flags & PFDESC_TCP_NORM) && s->src.scrub && pf_normalize_tcp_stateful(m, off, pd, &reason, th, s, &s->src, &s->dst, rewrite)) { /* This really shouldn't happen!!! */ DPFPRINTF(PF_DEBUG_URGENT, ("pf_normalize_tcp_stateful failed on first pkt")); pf_normalize_tcp_cleanup(s); pf_src_tree_remove_state(s); STATE_DEC_COUNTERS(s); uma_zfree(V_pf_state_z, s); return (PF_DROP); } } s->direction = pd->dir; /* * sk/nk could already been setup by pf_get_translation(). */ if (nr == NULL) { KASSERT((sk == NULL && nk == NULL), ("%s: nr %p sk %p, nk %p", __func__, nr, sk, nk)); sk = pf_state_key_setup(pd, pd->src, pd->dst, sport, dport); if (sk == NULL) goto csfailed; nk = sk; } else KASSERT((sk != NULL && nk != NULL), ("%s: nr %p sk %p, nk %p", __func__, nr, sk, nk)); /* Swap sk/nk for PF_OUT. */ if (pf_state_insert(BOUND_IFACE(r, kif), (pd->dir == PF_IN) ? sk : nk, (pd->dir == PF_IN) ? nk : sk, s)) { if (pd->proto == IPPROTO_TCP) pf_normalize_tcp_cleanup(s); REASON_SET(&reason, PFRES_STATEINS); pf_src_tree_remove_state(s); STATE_DEC_COUNTERS(s); uma_zfree(V_pf_state_z, s); return (PF_DROP); } else *sm = s; if (tag > 0) s->tag = tag; if (pd->proto == IPPROTO_TCP && (th->th_flags & (TH_SYN|TH_ACK)) == TH_SYN && r->keep_state == PF_STATE_SYNPROXY) { s->src.state = PF_TCPS_PROXY_SRC; /* undo NAT changes, if they have taken place */ if (nr != NULL) { struct pf_state_key *skt = s->key[PF_SK_WIRE]; if (pd->dir == PF_OUT) skt = s->key[PF_SK_STACK]; PF_ACPY(pd->src, &skt->addr[pd->sidx], pd->af); PF_ACPY(pd->dst, &skt->addr[pd->didx], pd->af); if (pd->sport) *pd->sport = skt->port[pd->sidx]; if (pd->dport) *pd->dport = skt->port[pd->didx]; if (pd->proto_sum) *pd->proto_sum = bproto_sum; if (pd->ip_sum) *pd->ip_sum = bip_sum; m_copyback(m, off, hdrlen, pd->hdr.any); } s->src.seqhi = htonl(arc4random()); /* Find mss option */ int rtid = M_GETFIB(m); mss = pf_get_mss(m, off, th->th_off, pd->af); mss = pf_calc_mss(pd->src, pd->af, rtid, mss); mss = pf_calc_mss(pd->dst, pd->af, rtid, mss); s->src.mss = mss; pf_send_tcp(NULL, r, pd->af, pd->dst, pd->src, th->th_dport, th->th_sport, s->src.seqhi, ntohl(th->th_seq) + 1, TH_SYN|TH_ACK, 0, s->src.mss, 0, 1, 0, NULL); REASON_SET(&reason, PFRES_SYNPROXY); return (PF_SYNPROXY_DROP); } return (PF_PASS); csfailed: if (sk != NULL) uma_zfree(V_pf_state_key_z, sk); if (nk != NULL) uma_zfree(V_pf_state_key_z, nk); if (sn != NULL) { struct pf_srchash *sh; sh = &V_pf_srchash[pf_hashsrc(&sn->addr, sn->af)]; PF_HASHROW_LOCK(sh); if (--sn->states == 0 && sn->expire == 0) { pf_unlink_src_node(sn); uma_zfree(V_pf_sources_z, sn); counter_u64_add( V_pf_status.scounters[SCNT_SRC_NODE_REMOVALS], 1); } PF_HASHROW_UNLOCK(sh); } if (nsn != sn && nsn != NULL) { struct pf_srchash *sh; sh = &V_pf_srchash[pf_hashsrc(&nsn->addr, nsn->af)]; PF_HASHROW_LOCK(sh); if (--nsn->states == 0 && nsn->expire == 0) { pf_unlink_src_node(nsn); uma_zfree(V_pf_sources_z, nsn); counter_u64_add( V_pf_status.scounters[SCNT_SRC_NODE_REMOVALS], 1); } PF_HASHROW_UNLOCK(sh); } return (PF_DROP); } static int pf_test_fragment(struct pf_rule **rm, int direction, struct pfi_kif *kif, struct mbuf *m, void *h, struct pf_pdesc *pd, struct pf_rule **am, struct pf_ruleset **rsm) { struct pf_rule *r, *a = NULL; struct pf_ruleset *ruleset = NULL; sa_family_t af = pd->af; u_short reason; int tag = -1; int asd = 0; int match = 0; struct pf_anchor_stackframe anchor_stack[PF_ANCHOR_STACKSIZE]; PF_RULES_RASSERT(); r = TAILQ_FIRST(pf_main_ruleset.rules[PF_RULESET_FILTER].active.ptr); while (r != NULL) { r->evaluations++; if (pfi_kif_match(r->kif, kif) == r->ifnot) r = r->skip[PF_SKIP_IFP].ptr; else if (r->direction && r->direction != direction) r = r->skip[PF_SKIP_DIR].ptr; else if (r->af && r->af != af) r = r->skip[PF_SKIP_AF].ptr; else if (r->proto && r->proto != pd->proto) r = r->skip[PF_SKIP_PROTO].ptr; else if (PF_MISMATCHAW(&r->src.addr, pd->src, af, r->src.neg, kif, M_GETFIB(m))) r = r->skip[PF_SKIP_SRC_ADDR].ptr; else if (PF_MISMATCHAW(&r->dst.addr, pd->dst, af, r->dst.neg, NULL, M_GETFIB(m))) r = r->skip[PF_SKIP_DST_ADDR].ptr; else if (r->tos && !(r->tos == pd->tos)) r = TAILQ_NEXT(r, entries); else if (r->os_fingerprint != PF_OSFP_ANY) r = TAILQ_NEXT(r, entries); else if (pd->proto == IPPROTO_UDP && (r->src.port_op || r->dst.port_op)) r = TAILQ_NEXT(r, entries); else if (pd->proto == IPPROTO_TCP && (r->src.port_op || r->dst.port_op || r->flagset)) r = TAILQ_NEXT(r, entries); else if ((pd->proto == IPPROTO_ICMP || pd->proto == IPPROTO_ICMPV6) && (r->type || r->code)) r = TAILQ_NEXT(r, entries); else if (r->prio && !pf_match_ieee8021q_pcp(r->prio, m)) r = TAILQ_NEXT(r, entries); else if (r->prob && r->prob <= (arc4random() % (UINT_MAX - 1) + 1)) r = TAILQ_NEXT(r, entries); else if (r->match_tag && !pf_match_tag(m, r, &tag, pd->pf_mtag ? pd->pf_mtag->tag : 0)) r = TAILQ_NEXT(r, entries); else { if (r->anchor == NULL) { match = 1; *rm = r; *am = a; *rsm = ruleset; if ((*rm)->quick) break; r = TAILQ_NEXT(r, entries); } else pf_step_into_anchor(anchor_stack, &asd, &ruleset, PF_RULESET_FILTER, &r, &a, &match); } if (r == NULL && pf_step_out_of_anchor(anchor_stack, &asd, &ruleset, PF_RULESET_FILTER, &r, &a, &match)) break; } r = *rm; a = *am; ruleset = *rsm; REASON_SET(&reason, PFRES_MATCH); if (r->log) PFLOG_PACKET(kif, m, af, direction, reason, r, a, ruleset, pd, 1); if (r->action != PF_PASS) return (PF_DROP); if (tag > 0 && pf_tag_packet(m, pd, tag)) { REASON_SET(&reason, PFRES_MEMORY); return (PF_DROP); } return (PF_PASS); } static int pf_tcp_track_full(struct pf_state_peer *src, struct pf_state_peer *dst, struct pf_state **state, struct pfi_kif *kif, struct mbuf *m, int off, struct pf_pdesc *pd, u_short *reason, int *copyback) { struct tcphdr *th = pd->hdr.tcp; u_int16_t win = ntohs(th->th_win); u_int32_t ack, end, seq, orig_seq; u_int8_t sws, dws; int ackskew; if (src->wscale && dst->wscale && !(th->th_flags & TH_SYN)) { sws = src->wscale & PF_WSCALE_MASK; dws = dst->wscale & PF_WSCALE_MASK; } else sws = dws = 0; /* * Sequence tracking algorithm from Guido van Rooij's paper: * http://www.madison-gurkha.com/publications/tcp_filtering/ * tcp_filtering.ps */ orig_seq = seq = ntohl(th->th_seq); if (src->seqlo == 0) { /* First packet from this end. Set its state */ if ((pd->flags & PFDESC_TCP_NORM || dst->scrub) && src->scrub == NULL) { if (pf_normalize_tcp_init(m, off, pd, th, src, dst)) { REASON_SET(reason, PFRES_MEMORY); return (PF_DROP); } } /* Deferred generation of sequence number modulator */ if (dst->seqdiff && !src->seqdiff) { /* use random iss for the TCP server */ while ((src->seqdiff = arc4random() - seq) == 0) ; ack = ntohl(th->th_ack) - dst->seqdiff; pf_change_proto_a(m, &th->th_seq, &th->th_sum, htonl(seq + src->seqdiff), 0); pf_change_proto_a(m, &th->th_ack, &th->th_sum, htonl(ack), 0); *copyback = 1; } else { ack = ntohl(th->th_ack); } end = seq + pd->p_len; if (th->th_flags & TH_SYN) { end++; if (dst->wscale & PF_WSCALE_FLAG) { src->wscale = pf_get_wscale(m, off, th->th_off, pd->af); if (src->wscale & PF_WSCALE_FLAG) { /* Remove scale factor from initial * window */ sws = src->wscale & PF_WSCALE_MASK; win = ((u_int32_t)win + (1 << sws) - 1) >> sws; dws = dst->wscale & PF_WSCALE_MASK; } else { /* fixup other window */ dst->max_win <<= dst->wscale & PF_WSCALE_MASK; /* in case of a retrans SYN|ACK */ dst->wscale = 0; } } } if (th->th_flags & TH_FIN) end++; src->seqlo = seq; if (src->state < TCPS_SYN_SENT) src->state = TCPS_SYN_SENT; /* * May need to slide the window (seqhi may have been set by * the crappy stack check or if we picked up the connection * after establishment) */ if (src->seqhi == 1 || SEQ_GEQ(end + MAX(1, dst->max_win << dws), src->seqhi)) src->seqhi = end + MAX(1, dst->max_win << dws); if (win > src->max_win) src->max_win = win; } else { ack = ntohl(th->th_ack) - dst->seqdiff; if (src->seqdiff) { /* Modulate sequence numbers */ pf_change_proto_a(m, &th->th_seq, &th->th_sum, htonl(seq + src->seqdiff), 0); pf_change_proto_a(m, &th->th_ack, &th->th_sum, htonl(ack), 0); *copyback = 1; } end = seq + pd->p_len; if (th->th_flags & TH_SYN) end++; if (th->th_flags & TH_FIN) end++; } if ((th->th_flags & TH_ACK) == 0) { /* Let it pass through the ack skew check */ ack = dst->seqlo; } else if ((ack == 0 && (th->th_flags & (TH_ACK|TH_RST)) == (TH_ACK|TH_RST)) || /* broken tcp stacks do not set ack */ (dst->state < TCPS_SYN_SENT)) { /* * Many stacks (ours included) will set the ACK number in an * FIN|ACK if the SYN times out -- no sequence to ACK. */ ack = dst->seqlo; } if (seq == end) { /* Ease sequencing restrictions on no data packets */ seq = src->seqlo; end = seq; } ackskew = dst->seqlo - ack; /* * Need to demodulate the sequence numbers in any TCP SACK options * (Selective ACK). We could optionally validate the SACK values * against the current ACK window, either forwards or backwards, but * I'm not confident that SACK has been implemented properly * everywhere. It wouldn't surprise me if several stacks accidentally * SACK too far backwards of previously ACKed data. There really aren't * any security implications of bad SACKing unless the target stack * doesn't validate the option length correctly. Someone trying to * spoof into a TCP connection won't bother blindly sending SACK * options anyway. */ if (dst->seqdiff && (th->th_off << 2) > sizeof(struct tcphdr)) { if (pf_modulate_sack(m, off, pd, th, dst)) *copyback = 1; } #define MAXACKWINDOW (0xffff + 1500) /* 1500 is an arbitrary fudge factor */ if (SEQ_GEQ(src->seqhi, end) && /* Last octet inside other's window space */ SEQ_GEQ(seq, src->seqlo - (dst->max_win << dws)) && /* Retrans: not more than one window back */ (ackskew >= -MAXACKWINDOW) && /* Acking not more than one reassembled fragment backwards */ (ackskew <= (MAXACKWINDOW << sws)) && /* Acking not more than one window forward */ ((th->th_flags & TH_RST) == 0 || orig_seq == src->seqlo || (orig_seq == src->seqlo + 1) || (orig_seq + 1 == src->seqlo) || (pd->flags & PFDESC_IP_REAS) == 0)) { /* Require an exact/+1 sequence match on resets when possible */ if (dst->scrub || src->scrub) { if (pf_normalize_tcp_stateful(m, off, pd, reason, th, *state, src, dst, copyback)) return (PF_DROP); } /* update max window */ if (src->max_win < win) src->max_win = win; /* synchronize sequencing */ if (SEQ_GT(end, src->seqlo)) src->seqlo = end; /* slide the window of what the other end can send */ if (SEQ_GEQ(ack + (win << sws), dst->seqhi)) dst->seqhi = ack + MAX((win << sws), 1); /* update states */ if (th->th_flags & TH_SYN) if (src->state < TCPS_SYN_SENT) src->state = TCPS_SYN_SENT; if (th->th_flags & TH_FIN) if (src->state < TCPS_CLOSING) src->state = TCPS_CLOSING; if (th->th_flags & TH_ACK) { if (dst->state == TCPS_SYN_SENT) { dst->state = TCPS_ESTABLISHED; if (src->state == TCPS_ESTABLISHED && (*state)->src_node != NULL && pf_src_connlimit(state)) { REASON_SET(reason, PFRES_SRCLIMIT); return (PF_DROP); } } else if (dst->state == TCPS_CLOSING) dst->state = TCPS_FIN_WAIT_2; } if (th->th_flags & TH_RST) src->state = dst->state = TCPS_TIME_WAIT; /* update expire time */ (*state)->expire = time_uptime; if (src->state >= TCPS_FIN_WAIT_2 && dst->state >= TCPS_FIN_WAIT_2) (*state)->timeout = PFTM_TCP_CLOSED; else if (src->state >= TCPS_CLOSING && dst->state >= TCPS_CLOSING) (*state)->timeout = PFTM_TCP_FIN_WAIT; else if (src->state < TCPS_ESTABLISHED || dst->state < TCPS_ESTABLISHED) (*state)->timeout = PFTM_TCP_OPENING; else if (src->state >= TCPS_CLOSING || dst->state >= TCPS_CLOSING) (*state)->timeout = PFTM_TCP_CLOSING; else (*state)->timeout = PFTM_TCP_ESTABLISHED; /* Fall through to PASS packet */ } else if ((dst->state < TCPS_SYN_SENT || dst->state >= TCPS_FIN_WAIT_2 || src->state >= TCPS_FIN_WAIT_2) && SEQ_GEQ(src->seqhi + MAXACKWINDOW, end) && /* Within a window forward of the originating packet */ SEQ_GEQ(seq, src->seqlo - MAXACKWINDOW)) { /* Within a window backward of the originating packet */ /* * This currently handles three situations: * 1) Stupid stacks will shotgun SYNs before their peer * replies. * 2) When PF catches an already established stream (the * firewall rebooted, the state table was flushed, routes * changed...) * 3) Packets get funky immediately after the connection * closes (this should catch Solaris spurious ACK|FINs * that web servers like to spew after a close) * * This must be a little more careful than the above code * since packet floods will also be caught here. We don't * update the TTL here to mitigate the damage of a packet * flood and so the same code can handle awkward establishment * and a loosened connection close. * In the establishment case, a correct peer response will * validate the connection, go through the normal state code * and keep updating the state TTL. */ if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: loose state match: "); pf_print_state(*state); pf_print_flags(th->th_flags); printf(" seq=%u (%u) ack=%u len=%u ackskew=%d " "pkts=%llu:%llu dir=%s,%s\n", seq, orig_seq, ack, pd->p_len, ackskew, (unsigned long long)(*state)->packets[0], (unsigned long long)(*state)->packets[1], pd->dir == PF_IN ? "in" : "out", pd->dir == (*state)->direction ? "fwd" : "rev"); } if (dst->scrub || src->scrub) { if (pf_normalize_tcp_stateful(m, off, pd, reason, th, *state, src, dst, copyback)) return (PF_DROP); } /* update max window */ if (src->max_win < win) src->max_win = win; /* synchronize sequencing */ if (SEQ_GT(end, src->seqlo)) src->seqlo = end; /* slide the window of what the other end can send */ if (SEQ_GEQ(ack + (win << sws), dst->seqhi)) dst->seqhi = ack + MAX((win << sws), 1); /* * Cannot set dst->seqhi here since this could be a shotgunned * SYN and not an already established connection. */ if (th->th_flags & TH_FIN) if (src->state < TCPS_CLOSING) src->state = TCPS_CLOSING; if (th->th_flags & TH_RST) src->state = dst->state = TCPS_TIME_WAIT; /* Fall through to PASS packet */ } else { if ((*state)->dst.state == TCPS_SYN_SENT && (*state)->src.state == TCPS_SYN_SENT) { /* Send RST for state mismatches during handshake */ if (!(th->th_flags & TH_RST)) pf_send_tcp(NULL, (*state)->rule.ptr, pd->af, pd->dst, pd->src, th->th_dport, th->th_sport, ntohl(th->th_ack), 0, TH_RST, 0, 0, (*state)->rule.ptr->return_ttl, 1, 0, kif->pfik_ifp); src->seqlo = 0; src->seqhi = 1; src->max_win = 1; } else if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: BAD state: "); pf_print_state(*state); pf_print_flags(th->th_flags); printf(" seq=%u (%u) ack=%u len=%u ackskew=%d " "pkts=%llu:%llu dir=%s,%s\n", seq, orig_seq, ack, pd->p_len, ackskew, (unsigned long long)(*state)->packets[0], (unsigned long long)(*state)->packets[1], pd->dir == PF_IN ? "in" : "out", pd->dir == (*state)->direction ? "fwd" : "rev"); printf("pf: State failure on: %c %c %c %c | %c %c\n", SEQ_GEQ(src->seqhi, end) ? ' ' : '1', SEQ_GEQ(seq, src->seqlo - (dst->max_win << dws)) ? ' ': '2', (ackskew >= -MAXACKWINDOW) ? ' ' : '3', (ackskew <= (MAXACKWINDOW << sws)) ? ' ' : '4', SEQ_GEQ(src->seqhi + MAXACKWINDOW, end) ?' ' :'5', SEQ_GEQ(seq, src->seqlo - MAXACKWINDOW) ?' ' :'6'); } REASON_SET(reason, PFRES_BADSTATE); return (PF_DROP); } return (PF_PASS); } static int pf_tcp_track_sloppy(struct pf_state_peer *src, struct pf_state_peer *dst, struct pf_state **state, struct pf_pdesc *pd, u_short *reason) { struct tcphdr *th = pd->hdr.tcp; if (th->th_flags & TH_SYN) if (src->state < TCPS_SYN_SENT) src->state = TCPS_SYN_SENT; if (th->th_flags & TH_FIN) if (src->state < TCPS_CLOSING) src->state = TCPS_CLOSING; if (th->th_flags & TH_ACK) { if (dst->state == TCPS_SYN_SENT) { dst->state = TCPS_ESTABLISHED; if (src->state == TCPS_ESTABLISHED && (*state)->src_node != NULL && pf_src_connlimit(state)) { REASON_SET(reason, PFRES_SRCLIMIT); return (PF_DROP); } } else if (dst->state == TCPS_CLOSING) { dst->state = TCPS_FIN_WAIT_2; } else if (src->state == TCPS_SYN_SENT && dst->state < TCPS_SYN_SENT) { /* * Handle a special sloppy case where we only see one * half of the connection. If there is a ACK after * the initial SYN without ever seeing a packet from * the destination, set the connection to established. */ dst->state = src->state = TCPS_ESTABLISHED; if ((*state)->src_node != NULL && pf_src_connlimit(state)) { REASON_SET(reason, PFRES_SRCLIMIT); return (PF_DROP); } } else if (src->state == TCPS_CLOSING && dst->state == TCPS_ESTABLISHED && dst->seqlo == 0) { /* * Handle the closing of half connections where we * don't see the full bidirectional FIN/ACK+ACK * handshake. */ dst->state = TCPS_CLOSING; } } if (th->th_flags & TH_RST) src->state = dst->state = TCPS_TIME_WAIT; /* update expire time */ (*state)->expire = time_uptime; if (src->state >= TCPS_FIN_WAIT_2 && dst->state >= TCPS_FIN_WAIT_2) (*state)->timeout = PFTM_TCP_CLOSED; else if (src->state >= TCPS_CLOSING && dst->state >= TCPS_CLOSING) (*state)->timeout = PFTM_TCP_FIN_WAIT; else if (src->state < TCPS_ESTABLISHED || dst->state < TCPS_ESTABLISHED) (*state)->timeout = PFTM_TCP_OPENING; else if (src->state >= TCPS_CLOSING || dst->state >= TCPS_CLOSING) (*state)->timeout = PFTM_TCP_CLOSING; else (*state)->timeout = PFTM_TCP_ESTABLISHED; return (PF_PASS); } static int pf_test_state_tcp(struct pf_state **state, int direction, struct pfi_kif *kif, struct mbuf *m, int off, void *h, struct pf_pdesc *pd, u_short *reason) { struct pf_state_key_cmp key; struct tcphdr *th = pd->hdr.tcp; int copyback = 0; struct pf_state_peer *src, *dst; struct pf_state_key *sk; bzero(&key, sizeof(key)); key.af = pd->af; key.proto = IPPROTO_TCP; if (direction == PF_IN) { /* wire side, straight */ PF_ACPY(&key.addr[0], pd->src, key.af); PF_ACPY(&key.addr[1], pd->dst, key.af); key.port[0] = th->th_sport; key.port[1] = th->th_dport; } else { /* stack side, reverse */ PF_ACPY(&key.addr[1], pd->src, key.af); PF_ACPY(&key.addr[0], pd->dst, key.af); key.port[1] = th->th_sport; key.port[0] = th->th_dport; } STATE_LOOKUP(kif, &key, direction, *state, pd); if (direction == (*state)->direction) { src = &(*state)->src; dst = &(*state)->dst; } else { src = &(*state)->dst; dst = &(*state)->src; } sk = (*state)->key[pd->didx]; if ((*state)->src.state == PF_TCPS_PROXY_SRC) { if (direction != (*state)->direction) { REASON_SET(reason, PFRES_SYNPROXY); return (PF_SYNPROXY_DROP); } if (th->th_flags & TH_SYN) { if (ntohl(th->th_seq) != (*state)->src.seqlo) { REASON_SET(reason, PFRES_SYNPROXY); return (PF_DROP); } pf_send_tcp(NULL, (*state)->rule.ptr, pd->af, pd->dst, pd->src, th->th_dport, th->th_sport, (*state)->src.seqhi, ntohl(th->th_seq) + 1, TH_SYN|TH_ACK, 0, (*state)->src.mss, 0, 1, 0, NULL); REASON_SET(reason, PFRES_SYNPROXY); return (PF_SYNPROXY_DROP); } else if (!(th->th_flags & TH_ACK) || (ntohl(th->th_ack) != (*state)->src.seqhi + 1) || (ntohl(th->th_seq) != (*state)->src.seqlo + 1)) { REASON_SET(reason, PFRES_SYNPROXY); return (PF_DROP); } else if ((*state)->src_node != NULL && pf_src_connlimit(state)) { REASON_SET(reason, PFRES_SRCLIMIT); return (PF_DROP); } else (*state)->src.state = PF_TCPS_PROXY_DST; } if ((*state)->src.state == PF_TCPS_PROXY_DST) { if (direction == (*state)->direction) { if (((th->th_flags & (TH_SYN|TH_ACK)) != TH_ACK) || (ntohl(th->th_ack) != (*state)->src.seqhi + 1) || (ntohl(th->th_seq) != (*state)->src.seqlo + 1)) { REASON_SET(reason, PFRES_SYNPROXY); return (PF_DROP); } (*state)->src.max_win = MAX(ntohs(th->th_win), 1); if ((*state)->dst.seqhi == 1) (*state)->dst.seqhi = htonl(arc4random()); pf_send_tcp(NULL, (*state)->rule.ptr, pd->af, &sk->addr[pd->sidx], &sk->addr[pd->didx], sk->port[pd->sidx], sk->port[pd->didx], (*state)->dst.seqhi, 0, TH_SYN, 0, (*state)->src.mss, 0, 0, (*state)->tag, NULL); REASON_SET(reason, PFRES_SYNPROXY); return (PF_SYNPROXY_DROP); } else if (((th->th_flags & (TH_SYN|TH_ACK)) != (TH_SYN|TH_ACK)) || (ntohl(th->th_ack) != (*state)->dst.seqhi + 1)) { REASON_SET(reason, PFRES_SYNPROXY); return (PF_DROP); } else { (*state)->dst.max_win = MAX(ntohs(th->th_win), 1); (*state)->dst.seqlo = ntohl(th->th_seq); pf_send_tcp(NULL, (*state)->rule.ptr, pd->af, pd->dst, pd->src, th->th_dport, th->th_sport, ntohl(th->th_ack), ntohl(th->th_seq) + 1, TH_ACK, (*state)->src.max_win, 0, 0, 0, (*state)->tag, NULL); pf_send_tcp(NULL, (*state)->rule.ptr, pd->af, &sk->addr[pd->sidx], &sk->addr[pd->didx], sk->port[pd->sidx], sk->port[pd->didx], (*state)->src.seqhi + 1, (*state)->src.seqlo + 1, TH_ACK, (*state)->dst.max_win, 0, 0, 1, 0, NULL); (*state)->src.seqdiff = (*state)->dst.seqhi - (*state)->src.seqlo; (*state)->dst.seqdiff = (*state)->src.seqhi - (*state)->dst.seqlo; (*state)->src.seqhi = (*state)->src.seqlo + (*state)->dst.max_win; (*state)->dst.seqhi = (*state)->dst.seqlo + (*state)->src.max_win; (*state)->src.wscale = (*state)->dst.wscale = 0; (*state)->src.state = (*state)->dst.state = TCPS_ESTABLISHED; REASON_SET(reason, PFRES_SYNPROXY); return (PF_SYNPROXY_DROP); } } if (((th->th_flags & (TH_SYN|TH_ACK)) == TH_SYN) && dst->state >= TCPS_FIN_WAIT_2 && src->state >= TCPS_FIN_WAIT_2) { if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: state reuse "); pf_print_state(*state); pf_print_flags(th->th_flags); printf("\n"); } /* XXX make sure it's the same direction ?? */ (*state)->src.state = (*state)->dst.state = TCPS_CLOSED; pf_unlink_state(*state, PF_ENTER_LOCKED); *state = NULL; return (PF_DROP); } if ((*state)->state_flags & PFSTATE_SLOPPY) { if (pf_tcp_track_sloppy(src, dst, state, pd, reason) == PF_DROP) return (PF_DROP); } else { if (pf_tcp_track_full(src, dst, state, kif, m, off, pd, reason, ©back) == PF_DROP) return (PF_DROP); } /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], pd->af) || nk->port[pd->sidx] != th->th_sport) pf_change_ap(m, pd->src, &th->th_sport, pd->ip_sum, &th->th_sum, &nk->addr[pd->sidx], nk->port[pd->sidx], 0, pd->af); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], pd->af) || nk->port[pd->didx] != th->th_dport) pf_change_ap(m, pd->dst, &th->th_dport, pd->ip_sum, &th->th_sum, &nk->addr[pd->didx], nk->port[pd->didx], 0, pd->af); copyback = 1; } /* Copyback sequence modulation or stateful scrub changes if needed */ if (copyback) m_copyback(m, off, sizeof(*th), (caddr_t)th); return (PF_PASS); } static int pf_test_state_udp(struct pf_state **state, int direction, struct pfi_kif *kif, struct mbuf *m, int off, void *h, struct pf_pdesc *pd) { struct pf_state_peer *src, *dst; struct pf_state_key_cmp key; struct udphdr *uh = pd->hdr.udp; bzero(&key, sizeof(key)); key.af = pd->af; key.proto = IPPROTO_UDP; if (direction == PF_IN) { /* wire side, straight */ PF_ACPY(&key.addr[0], pd->src, key.af); PF_ACPY(&key.addr[1], pd->dst, key.af); key.port[0] = uh->uh_sport; key.port[1] = uh->uh_dport; } else { /* stack side, reverse */ PF_ACPY(&key.addr[1], pd->src, key.af); PF_ACPY(&key.addr[0], pd->dst, key.af); key.port[1] = uh->uh_sport; key.port[0] = uh->uh_dport; } STATE_LOOKUP(kif, &key, direction, *state, pd); if (direction == (*state)->direction) { src = &(*state)->src; dst = &(*state)->dst; } else { src = &(*state)->dst; dst = &(*state)->src; } /* update states */ if (src->state < PFUDPS_SINGLE) src->state = PFUDPS_SINGLE; if (dst->state == PFUDPS_SINGLE) dst->state = PFUDPS_MULTIPLE; /* update expire time */ (*state)->expire = time_uptime; if (src->state == PFUDPS_MULTIPLE && dst->state == PFUDPS_MULTIPLE) (*state)->timeout = PFTM_UDP_MULTIPLE; else (*state)->timeout = PFTM_UDP_SINGLE; /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], pd->af) || nk->port[pd->sidx] != uh->uh_sport) pf_change_ap(m, pd->src, &uh->uh_sport, pd->ip_sum, &uh->uh_sum, &nk->addr[pd->sidx], nk->port[pd->sidx], 1, pd->af); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], pd->af) || nk->port[pd->didx] != uh->uh_dport) pf_change_ap(m, pd->dst, &uh->uh_dport, pd->ip_sum, &uh->uh_sum, &nk->addr[pd->didx], nk->port[pd->didx], 1, pd->af); m_copyback(m, off, sizeof(*uh), (caddr_t)uh); } return (PF_PASS); } static int pf_test_state_icmp(struct pf_state **state, int direction, struct pfi_kif *kif, struct mbuf *m, int off, void *h, struct pf_pdesc *pd, u_short *reason) { struct pf_addr *saddr = pd->src, *daddr = pd->dst; u_int16_t icmpid = 0, *icmpsum; u_int8_t icmptype; int state_icmp = 0; struct pf_state_key_cmp key; bzero(&key, sizeof(key)); switch (pd->proto) { #ifdef INET case IPPROTO_ICMP: icmptype = pd->hdr.icmp->icmp_type; icmpid = pd->hdr.icmp->icmp_id; icmpsum = &pd->hdr.icmp->icmp_cksum; if (icmptype == ICMP_UNREACH || icmptype == ICMP_SOURCEQUENCH || icmptype == ICMP_REDIRECT || icmptype == ICMP_TIMXCEED || icmptype == ICMP_PARAMPROB) state_icmp++; break; #endif /* INET */ #ifdef INET6 case IPPROTO_ICMPV6: icmptype = pd->hdr.icmp6->icmp6_type; icmpid = pd->hdr.icmp6->icmp6_id; icmpsum = &pd->hdr.icmp6->icmp6_cksum; if (icmptype == ICMP6_DST_UNREACH || icmptype == ICMP6_PACKET_TOO_BIG || icmptype == ICMP6_TIME_EXCEEDED || icmptype == ICMP6_PARAM_PROB) state_icmp++; break; #endif /* INET6 */ } if (!state_icmp) { /* * ICMP query/reply message not related to a TCP/UDP packet. * Search for an ICMP state. */ key.af = pd->af; key.proto = pd->proto; key.port[0] = key.port[1] = icmpid; if (direction == PF_IN) { /* wire side, straight */ PF_ACPY(&key.addr[0], pd->src, key.af); PF_ACPY(&key.addr[1], pd->dst, key.af); } else { /* stack side, reverse */ PF_ACPY(&key.addr[1], pd->src, key.af); PF_ACPY(&key.addr[0], pd->dst, key.af); } STATE_LOOKUP(kif, &key, direction, *state, pd); (*state)->expire = time_uptime; (*state)->timeout = PFTM_ICMP_ERROR_REPLY; /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; switch (pd->af) { #ifdef INET case AF_INET: if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], AF_INET)) pf_change_a(&saddr->v4.s_addr, pd->ip_sum, nk->addr[pd->sidx].v4.s_addr, 0); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], AF_INET)) pf_change_a(&daddr->v4.s_addr, pd->ip_sum, nk->addr[pd->didx].v4.s_addr, 0); if (nk->port[0] != pd->hdr.icmp->icmp_id) { pd->hdr.icmp->icmp_cksum = pf_cksum_fixup( pd->hdr.icmp->icmp_cksum, icmpid, nk->port[pd->sidx], 0); pd->hdr.icmp->icmp_id = nk->port[pd->sidx]; } m_copyback(m, off, ICMP_MINLEN, (caddr_t )pd->hdr.icmp); break; #endif /* INET */ #ifdef INET6 case AF_INET6: if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], AF_INET6)) pf_change_a6(saddr, &pd->hdr.icmp6->icmp6_cksum, &nk->addr[pd->sidx], 0); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], AF_INET6)) pf_change_a6(daddr, &pd->hdr.icmp6->icmp6_cksum, &nk->addr[pd->didx], 0); m_copyback(m, off, sizeof(struct icmp6_hdr), (caddr_t )pd->hdr.icmp6); break; #endif /* INET6 */ } } return (PF_PASS); } else { /* * ICMP error message in response to a TCP/UDP packet. * Extract the inner TCP/UDP header and search for that state. */ struct pf_pdesc pd2; bzero(&pd2, sizeof pd2); #ifdef INET struct ip h2; #endif /* INET */ #ifdef INET6 struct ip6_hdr h2_6; int terminal = 0; #endif /* INET6 */ int ipoff2 = 0; int off2 = 0; pd2.af = pd->af; /* Payload packet is from the opposite direction. */ pd2.sidx = (direction == PF_IN) ? 1 : 0; pd2.didx = (direction == PF_IN) ? 0 : 1; switch (pd->af) { #ifdef INET case AF_INET: /* offset of h2 in mbuf chain */ ipoff2 = off + ICMP_MINLEN; if (!pf_pull_hdr(m, ipoff2, &h2, sizeof(h2), NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short " "(ip)\n")); return (PF_DROP); } /* * ICMP error messages don't refer to non-first * fragments */ if (h2.ip_off & htons(IP_OFFMASK)) { REASON_SET(reason, PFRES_FRAG); return (PF_DROP); } /* offset of protocol header that follows h2 */ off2 = ipoff2 + (h2.ip_hl << 2); pd2.proto = h2.ip_p; pd2.src = (struct pf_addr *)&h2.ip_src; pd2.dst = (struct pf_addr *)&h2.ip_dst; pd2.ip_sum = &h2.ip_sum; break; #endif /* INET */ #ifdef INET6 case AF_INET6: ipoff2 = off + sizeof(struct icmp6_hdr); if (!pf_pull_hdr(m, ipoff2, &h2_6, sizeof(h2_6), NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short " "(ip6)\n")); return (PF_DROP); } pd2.proto = h2_6.ip6_nxt; pd2.src = (struct pf_addr *)&h2_6.ip6_src; pd2.dst = (struct pf_addr *)&h2_6.ip6_dst; pd2.ip_sum = NULL; off2 = ipoff2 + sizeof(h2_6); do { switch (pd2.proto) { case IPPROTO_FRAGMENT: /* * ICMPv6 error messages for * non-first fragments */ REASON_SET(reason, PFRES_FRAG); return (PF_DROP); case IPPROTO_AH: case IPPROTO_HOPOPTS: case IPPROTO_ROUTING: case IPPROTO_DSTOPTS: { /* get next header and header length */ struct ip6_ext opt6; if (!pf_pull_hdr(m, off2, &opt6, sizeof(opt6), NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMPv6 short opt\n")); return (PF_DROP); } if (pd2.proto == IPPROTO_AH) off2 += (opt6.ip6e_len + 2) * 4; else off2 += (opt6.ip6e_len + 1) * 8; pd2.proto = opt6.ip6e_nxt; /* goto the next header */ break; } default: terminal++; break; } } while (!terminal); break; #endif /* INET6 */ } switch (pd2.proto) { case IPPROTO_TCP: { struct tcphdr th; u_int32_t seq; struct pf_state_peer *src, *dst; u_int8_t dws; int copyback = 0; /* * Only the first 8 bytes of the TCP header can be * expected. Don't access any TCP header fields after * th_seq, an ackskew test is not possible. */ if (!pf_pull_hdr(m, off2, &th, 8, NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short " "(tcp)\n")); return (PF_DROP); } key.af = pd2.af; key.proto = IPPROTO_TCP; PF_ACPY(&key.addr[pd2.sidx], pd2.src, key.af); PF_ACPY(&key.addr[pd2.didx], pd2.dst, key.af); key.port[pd2.sidx] = th.th_sport; key.port[pd2.didx] = th.th_dport; STATE_LOOKUP(kif, &key, direction, *state, pd); if (direction == (*state)->direction) { src = &(*state)->dst; dst = &(*state)->src; } else { src = &(*state)->src; dst = &(*state)->dst; } if (src->wscale && dst->wscale) dws = dst->wscale & PF_WSCALE_MASK; else dws = 0; /* Demodulate sequence number */ seq = ntohl(th.th_seq) - src->seqdiff; if (src->seqdiff) { pf_change_a(&th.th_seq, icmpsum, htonl(seq), 0); copyback = 1; } if (!((*state)->state_flags & PFSTATE_SLOPPY) && (!SEQ_GEQ(src->seqhi, seq) || !SEQ_GEQ(seq, src->seqlo - (dst->max_win << dws)))) { if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: BAD ICMP %d:%d ", icmptype, pd->hdr.icmp->icmp_code); pf_print_host(pd->src, 0, pd->af); printf(" -> "); pf_print_host(pd->dst, 0, pd->af); printf(" state: "); pf_print_state(*state); printf(" seq=%u\n", seq); } REASON_SET(reason, PFRES_BADSTATE); return (PF_DROP); } else { if (V_pf_status.debug >= PF_DEBUG_MISC) { printf("pf: OK ICMP %d:%d ", icmptype, pd->hdr.icmp->icmp_code); pf_print_host(pd->src, 0, pd->af); printf(" -> "); pf_print_host(pd->dst, 0, pd->af); printf(" state: "); pf_print_state(*state); printf(" seq=%u\n", seq); } } /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd2.src, &nk->addr[pd2.sidx], pd2.af) || nk->port[pd2.sidx] != th.th_sport) pf_change_icmp(pd2.src, &th.th_sport, daddr, &nk->addr[pd2.sidx], nk->port[pd2.sidx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, pd2.af); if (PF_ANEQ(pd2.dst, &nk->addr[pd2.didx], pd2.af) || nk->port[pd2.didx] != th.th_dport) pf_change_icmp(pd2.dst, &th.th_dport, saddr, &nk->addr[pd2.didx], nk->port[pd2.didx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, pd2.af); copyback = 1; } if (copyback) { switch (pd2.af) { #ifdef INET case AF_INET: m_copyback(m, off, ICMP_MINLEN, (caddr_t )pd->hdr.icmp); m_copyback(m, ipoff2, sizeof(h2), (caddr_t )&h2); break; #endif /* INET */ #ifdef INET6 case AF_INET6: m_copyback(m, off, sizeof(struct icmp6_hdr), (caddr_t )pd->hdr.icmp6); m_copyback(m, ipoff2, sizeof(h2_6), (caddr_t )&h2_6); break; #endif /* INET6 */ } m_copyback(m, off2, 8, (caddr_t)&th); } return (PF_PASS); break; } case IPPROTO_UDP: { struct udphdr uh; if (!pf_pull_hdr(m, off2, &uh, sizeof(uh), NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short " "(udp)\n")); return (PF_DROP); } key.af = pd2.af; key.proto = IPPROTO_UDP; PF_ACPY(&key.addr[pd2.sidx], pd2.src, key.af); PF_ACPY(&key.addr[pd2.didx], pd2.dst, key.af); key.port[pd2.sidx] = uh.uh_sport; key.port[pd2.didx] = uh.uh_dport; STATE_LOOKUP(kif, &key, direction, *state, pd); /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd2.src, &nk->addr[pd2.sidx], pd2.af) || nk->port[pd2.sidx] != uh.uh_sport) pf_change_icmp(pd2.src, &uh.uh_sport, daddr, &nk->addr[pd2.sidx], nk->port[pd2.sidx], &uh.uh_sum, pd2.ip_sum, icmpsum, pd->ip_sum, 1, pd2.af); if (PF_ANEQ(pd2.dst, &nk->addr[pd2.didx], pd2.af) || nk->port[pd2.didx] != uh.uh_dport) pf_change_icmp(pd2.dst, &uh.uh_dport, saddr, &nk->addr[pd2.didx], nk->port[pd2.didx], &uh.uh_sum, pd2.ip_sum, icmpsum, pd->ip_sum, 1, pd2.af); switch (pd2.af) { #ifdef INET case AF_INET: m_copyback(m, off, ICMP_MINLEN, (caddr_t )pd->hdr.icmp); m_copyback(m, ipoff2, sizeof(h2), (caddr_t)&h2); break; #endif /* INET */ #ifdef INET6 case AF_INET6: m_copyback(m, off, sizeof(struct icmp6_hdr), (caddr_t )pd->hdr.icmp6); m_copyback(m, ipoff2, sizeof(h2_6), (caddr_t )&h2_6); break; #endif /* INET6 */ } m_copyback(m, off2, sizeof(uh), (caddr_t)&uh); } return (PF_PASS); break; } #ifdef INET case IPPROTO_ICMP: { struct icmp iih; if (!pf_pull_hdr(m, off2, &iih, ICMP_MINLEN, NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short i" "(icmp)\n")); return (PF_DROP); } key.af = pd2.af; key.proto = IPPROTO_ICMP; PF_ACPY(&key.addr[pd2.sidx], pd2.src, key.af); PF_ACPY(&key.addr[pd2.didx], pd2.dst, key.af); key.port[0] = key.port[1] = iih.icmp_id; STATE_LOOKUP(kif, &key, direction, *state, pd); /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd2.src, &nk->addr[pd2.sidx], pd2.af) || nk->port[pd2.sidx] != iih.icmp_id) pf_change_icmp(pd2.src, &iih.icmp_id, daddr, &nk->addr[pd2.sidx], nk->port[pd2.sidx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, AF_INET); if (PF_ANEQ(pd2.dst, &nk->addr[pd2.didx], pd2.af) || nk->port[pd2.didx] != iih.icmp_id) pf_change_icmp(pd2.dst, &iih.icmp_id, saddr, &nk->addr[pd2.didx], nk->port[pd2.didx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, AF_INET); m_copyback(m, off, ICMP_MINLEN, (caddr_t)pd->hdr.icmp); m_copyback(m, ipoff2, sizeof(h2), (caddr_t)&h2); m_copyback(m, off2, ICMP_MINLEN, (caddr_t)&iih); } return (PF_PASS); break; } #endif /* INET */ #ifdef INET6 case IPPROTO_ICMPV6: { struct icmp6_hdr iih; if (!pf_pull_hdr(m, off2, &iih, sizeof(struct icmp6_hdr), NULL, reason, pd2.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: ICMP error message too short " "(icmp6)\n")); return (PF_DROP); } key.af = pd2.af; key.proto = IPPROTO_ICMPV6; PF_ACPY(&key.addr[pd2.sidx], pd2.src, key.af); PF_ACPY(&key.addr[pd2.didx], pd2.dst, key.af); key.port[0] = key.port[1] = iih.icmp6_id; STATE_LOOKUP(kif, &key, direction, *state, pd); /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd2.src, &nk->addr[pd2.sidx], pd2.af) || nk->port[pd2.sidx] != iih.icmp6_id) pf_change_icmp(pd2.src, &iih.icmp6_id, daddr, &nk->addr[pd2.sidx], nk->port[pd2.sidx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, AF_INET6); if (PF_ANEQ(pd2.dst, &nk->addr[pd2.didx], pd2.af) || nk->port[pd2.didx] != iih.icmp6_id) pf_change_icmp(pd2.dst, &iih.icmp6_id, saddr, &nk->addr[pd2.didx], nk->port[pd2.didx], NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, AF_INET6); m_copyback(m, off, sizeof(struct icmp6_hdr), (caddr_t)pd->hdr.icmp6); m_copyback(m, ipoff2, sizeof(h2_6), (caddr_t)&h2_6); m_copyback(m, off2, sizeof(struct icmp6_hdr), (caddr_t)&iih); } return (PF_PASS); break; } #endif /* INET6 */ default: { key.af = pd2.af; key.proto = pd2.proto; PF_ACPY(&key.addr[pd2.sidx], pd2.src, key.af); PF_ACPY(&key.addr[pd2.didx], pd2.dst, key.af); key.port[0] = key.port[1] = 0; STATE_LOOKUP(kif, &key, direction, *state, pd); /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; if (PF_ANEQ(pd2.src, &nk->addr[pd2.sidx], pd2.af)) pf_change_icmp(pd2.src, NULL, daddr, &nk->addr[pd2.sidx], 0, NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, pd2.af); if (PF_ANEQ(pd2.dst, &nk->addr[pd2.didx], pd2.af)) pf_change_icmp(pd2.dst, NULL, saddr, &nk->addr[pd2.didx], 0, NULL, pd2.ip_sum, icmpsum, pd->ip_sum, 0, pd2.af); switch (pd2.af) { #ifdef INET case AF_INET: m_copyback(m, off, ICMP_MINLEN, (caddr_t)pd->hdr.icmp); m_copyback(m, ipoff2, sizeof(h2), (caddr_t)&h2); break; #endif /* INET */ #ifdef INET6 case AF_INET6: m_copyback(m, off, sizeof(struct icmp6_hdr), (caddr_t )pd->hdr.icmp6); m_copyback(m, ipoff2, sizeof(h2_6), (caddr_t )&h2_6); break; #endif /* INET6 */ } } return (PF_PASS); break; } } } } static int pf_test_state_other(struct pf_state **state, int direction, struct pfi_kif *kif, struct mbuf *m, struct pf_pdesc *pd) { struct pf_state_peer *src, *dst; struct pf_state_key_cmp key; bzero(&key, sizeof(key)); key.af = pd->af; key.proto = pd->proto; if (direction == PF_IN) { PF_ACPY(&key.addr[0], pd->src, key.af); PF_ACPY(&key.addr[1], pd->dst, key.af); key.port[0] = key.port[1] = 0; } else { PF_ACPY(&key.addr[1], pd->src, key.af); PF_ACPY(&key.addr[0], pd->dst, key.af); key.port[1] = key.port[0] = 0; } STATE_LOOKUP(kif, &key, direction, *state, pd); if (direction == (*state)->direction) { src = &(*state)->src; dst = &(*state)->dst; } else { src = &(*state)->dst; dst = &(*state)->src; } /* update states */ if (src->state < PFOTHERS_SINGLE) src->state = PFOTHERS_SINGLE; if (dst->state == PFOTHERS_SINGLE) dst->state = PFOTHERS_MULTIPLE; /* update expire time */ (*state)->expire = time_uptime; if (src->state == PFOTHERS_MULTIPLE && dst->state == PFOTHERS_MULTIPLE) (*state)->timeout = PFTM_OTHER_MULTIPLE; else (*state)->timeout = PFTM_OTHER_SINGLE; /* translate source/destination address, if necessary */ if ((*state)->key[PF_SK_WIRE] != (*state)->key[PF_SK_STACK]) { struct pf_state_key *nk = (*state)->key[pd->didx]; KASSERT(nk, ("%s: nk is null", __func__)); KASSERT(pd, ("%s: pd is null", __func__)); KASSERT(pd->src, ("%s: pd->src is null", __func__)); KASSERT(pd->dst, ("%s: pd->dst is null", __func__)); switch (pd->af) { #ifdef INET case AF_INET: if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], AF_INET)) pf_change_a(&pd->src->v4.s_addr, pd->ip_sum, nk->addr[pd->sidx].v4.s_addr, 0); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], AF_INET)) pf_change_a(&pd->dst->v4.s_addr, pd->ip_sum, nk->addr[pd->didx].v4.s_addr, 0); break; #endif /* INET */ #ifdef INET6 case AF_INET6: if (PF_ANEQ(pd->src, &nk->addr[pd->sidx], AF_INET)) PF_ACPY(pd->src, &nk->addr[pd->sidx], pd->af); if (PF_ANEQ(pd->dst, &nk->addr[pd->didx], AF_INET)) PF_ACPY(pd->dst, &nk->addr[pd->didx], pd->af); #endif /* INET6 */ } } return (PF_PASS); } /* * ipoff and off are measured from the start of the mbuf chain. * h must be at "ipoff" on the mbuf chain. */ void * pf_pull_hdr(struct mbuf *m, int off, void *p, int len, u_short *actionp, u_short *reasonp, sa_family_t af) { switch (af) { #ifdef INET case AF_INET: { struct ip *h = mtod(m, struct ip *); u_int16_t fragoff = (ntohs(h->ip_off) & IP_OFFMASK) << 3; if (fragoff) { if (fragoff >= len) ACTION_SET(actionp, PF_PASS); else { ACTION_SET(actionp, PF_DROP); REASON_SET(reasonp, PFRES_FRAG); } return (NULL); } if (m->m_pkthdr.len < off + len || ntohs(h->ip_len) < off + len) { ACTION_SET(actionp, PF_DROP); REASON_SET(reasonp, PFRES_SHORT); return (NULL); } break; } #endif /* INET */ #ifdef INET6 case AF_INET6: { struct ip6_hdr *h = mtod(m, struct ip6_hdr *); if (m->m_pkthdr.len < off + len || (ntohs(h->ip6_plen) + sizeof(struct ip6_hdr)) < (unsigned)(off + len)) { ACTION_SET(actionp, PF_DROP); REASON_SET(reasonp, PFRES_SHORT); return (NULL); } break; } #endif /* INET6 */ } m_copydata(m, off, len, p); return (p); } #ifdef RADIX_MPATH static int pf_routable_oldmpath(struct pf_addr *addr, sa_family_t af, struct pfi_kif *kif, int rtableid) { struct radix_node_head *rnh; struct sockaddr_in *dst; int ret = 1; int check_mpath; #ifdef INET6 struct sockaddr_in6 *dst6; struct route_in6 ro; #else struct route ro; #endif struct radix_node *rn; struct rtentry *rt; struct ifnet *ifp; check_mpath = 0; /* XXX: stick to table 0 for now */ rnh = rt_tables_get_rnh(0, af); if (rnh != NULL && rn_mpath_capable(rnh)) check_mpath = 1; bzero(&ro, sizeof(ro)); switch (af) { case AF_INET: dst = satosin(&ro.ro_dst); dst->sin_family = AF_INET; dst->sin_len = sizeof(*dst); dst->sin_addr = addr->v4; break; #ifdef INET6 case AF_INET6: /* * Skip check for addresses with embedded interface scope, * as they would always match anyway. */ if (IN6_IS_SCOPE_EMBED(&addr->v6)) goto out; dst6 = (struct sockaddr_in6 *)&ro.ro_dst; dst6->sin6_family = AF_INET6; dst6->sin6_len = sizeof(*dst6); dst6->sin6_addr = addr->v6; break; #endif /* INET6 */ default: return (0); } /* Skip checks for ipsec interfaces */ if (kif != NULL && kif->pfik_ifp->if_type == IFT_ENC) goto out; switch (af) { #ifdef INET6 case AF_INET6: in6_rtalloc_ign(&ro, 0, rtableid); break; #endif #ifdef INET case AF_INET: in_rtalloc_ign((struct route *)&ro, 0, rtableid); break; #endif } if (ro.ro_rt != NULL) { /* No interface given, this is a no-route check */ if (kif == NULL) goto out; if (kif->pfik_ifp == NULL) { ret = 0; goto out; } /* Perform uRPF check if passed input interface */ ret = 0; rn = (struct radix_node *)ro.ro_rt; do { rt = (struct rtentry *)rn; ifp = rt->rt_ifp; if (kif->pfik_ifp == ifp) ret = 1; rn = rn_mpath_next(rn); } while (check_mpath == 1 && rn != NULL && ret == 0); } else ret = 0; out: if (ro.ro_rt != NULL) RTFREE(ro.ro_rt); return (ret); } #endif int pf_routable(struct pf_addr *addr, sa_family_t af, struct pfi_kif *kif, int rtableid) { #ifdef INET struct nhop4_basic nh4; #endif #ifdef INET6 struct nhop6_basic nh6; #endif struct ifnet *ifp; #ifdef RADIX_MPATH struct radix_node_head *rnh; /* XXX: stick to table 0 for now */ rnh = rt_tables_get_rnh(0, af); if (rnh != NULL && rn_mpath_capable(rnh)) return (pf_routable_oldmpath(addr, af, kif, rtableid)); #endif /* * Skip check for addresses with embedded interface scope, * as they would always match anyway. */ if (af == AF_INET6 && IN6_IS_SCOPE_EMBED(&addr->v6)) return (1); if (af != AF_INET && af != AF_INET6) return (0); /* Skip checks for ipsec interfaces */ if (kif != NULL && kif->pfik_ifp->if_type == IFT_ENC) return (1); ifp = NULL; switch (af) { #ifdef INET6 case AF_INET6: if (fib6_lookup_nh_basic(rtableid, &addr->v6, 0, 0, 0, &nh6)!=0) return (0); ifp = nh6.nh_ifp; break; #endif #ifdef INET case AF_INET: if (fib4_lookup_nh_basic(rtableid, addr->v4, 0, 0, &nh4) != 0) return (0); ifp = nh4.nh_ifp; break; #endif } /* No interface given, this is a no-route check */ if (kif == NULL) return (1); if (kif->pfik_ifp == NULL) return (0); /* Perform uRPF check if passed input interface */ if (kif->pfik_ifp == ifp) return (1); return (0); } #ifdef INET static void pf_route(struct mbuf **m, struct pf_rule *r, int dir, struct ifnet *oifp, struct pf_state *s, struct pf_pdesc *pd) { struct mbuf *m0, *m1; struct sockaddr_in dst; struct ip *ip; struct ifnet *ifp = NULL; struct pf_addr naddr; struct pf_src_node *sn = NULL; int error = 0; uint16_t ip_len, ip_off; KASSERT(m && *m && r && oifp, ("%s: invalid parameters", __func__)); KASSERT(dir == PF_IN || dir == PF_OUT, ("%s: invalid direction", __func__)); if ((pd->pf_mtag == NULL && ((pd->pf_mtag = pf_get_mtag(*m)) == NULL)) || pd->pf_mtag->routed++ > 3) { m0 = *m; *m = NULL; goto bad_locked; } if (r->rt == PF_DUPTO) { if ((m0 = m_dup(*m, M_NOWAIT)) == NULL) { if (s) PF_STATE_UNLOCK(s); return; } } else { if ((r->rt == PF_REPLYTO) == (r->direction == dir)) { if (s) PF_STATE_UNLOCK(s); return; } m0 = *m; } ip = mtod(m0, struct ip *); bzero(&dst, sizeof(dst)); dst.sin_family = AF_INET; dst.sin_len = sizeof(dst); dst.sin_addr = ip->ip_dst; - if (r->rt == PF_FASTROUTE) { - struct nhop4_basic nh4; - - if (s) - PF_STATE_UNLOCK(s); - - if (fib4_lookup_nh_basic(M_GETFIB(m0), ip->ip_dst, 0, - m0->m_pkthdr.flowid, &nh4) != 0) { - KMOD_IPSTAT_INC(ips_noroute); - error = EHOSTUNREACH; - goto bad; - } - - ifp = nh4.nh_ifp; - dst.sin_addr = nh4.nh_addr; + if (TAILQ_EMPTY(&r->rpool.list)) { + DPFPRINTF(PF_DEBUG_URGENT, + ("%s: TAILQ_EMPTY(&r->rpool.list)\n", __func__)); + goto bad_locked; + } + if (s == NULL) { + pf_map_addr(AF_INET, r, (struct pf_addr *)&ip->ip_src, + &naddr, NULL, &sn); + if (!PF_AZERO(&naddr, AF_INET)) + dst.sin_addr.s_addr = naddr.v4.s_addr; + ifp = r->rpool.cur->kif ? + r->rpool.cur->kif->pfik_ifp : NULL; } else { - if (TAILQ_EMPTY(&r->rpool.list)) { - DPFPRINTF(PF_DEBUG_URGENT, - ("%s: TAILQ_EMPTY(&r->rpool.list)\n", __func__)); - goto bad_locked; - } - if (s == NULL) { - pf_map_addr(AF_INET, r, (struct pf_addr *)&ip->ip_src, - &naddr, NULL, &sn); - if (!PF_AZERO(&naddr, AF_INET)) - dst.sin_addr.s_addr = naddr.v4.s_addr; - ifp = r->rpool.cur->kif ? - r->rpool.cur->kif->pfik_ifp : NULL; - } else { - if (!PF_AZERO(&s->rt_addr, AF_INET)) - dst.sin_addr.s_addr = - s->rt_addr.v4.s_addr; - ifp = s->rt_kif ? s->rt_kif->pfik_ifp : NULL; - PF_STATE_UNLOCK(s); - } + if (!PF_AZERO(&s->rt_addr, AF_INET)) + dst.sin_addr.s_addr = + s->rt_addr.v4.s_addr; + ifp = s->rt_kif ? s->rt_kif->pfik_ifp : NULL; + PF_STATE_UNLOCK(s); } if (ifp == NULL) goto bad; if (oifp != ifp) { if (pf_test(PF_OUT, ifp, &m0, NULL) != PF_PASS) goto bad; else if (m0 == NULL) goto done; if (m0->m_len < sizeof(struct ip)) { DPFPRINTF(PF_DEBUG_URGENT, ("%s: m0->m_len < sizeof(struct ip)\n", __func__)); goto bad; } ip = mtod(m0, struct ip *); } if (ifp->if_flags & IFF_LOOPBACK) m0->m_flags |= M_SKIP_FIREWALL; ip_len = ntohs(ip->ip_len); ip_off = ntohs(ip->ip_off); /* Copied from FreeBSD 10.0-CURRENT ip_output. */ m0->m_pkthdr.csum_flags |= CSUM_IP; if (m0->m_pkthdr.csum_flags & CSUM_DELAY_DATA & ~ifp->if_hwassist) { in_delayed_cksum(m0); m0->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA; } #ifdef SCTP if (m0->m_pkthdr.csum_flags & CSUM_SCTP & ~ifp->if_hwassist) { sctp_delayed_cksum(m, (uint32_t)(ip->ip_hl << 2)); m0->m_pkthdr.csum_flags &= ~CSUM_SCTP; } #endif /* * If small enough for interface, or the interface will take * care of the fragmentation for us, we can just send directly. */ if (ip_len <= ifp->if_mtu || (m0->m_pkthdr.csum_flags & ifp->if_hwassist & CSUM_TSO) != 0) { ip->ip_sum = 0; if (m0->m_pkthdr.csum_flags & CSUM_IP & ~ifp->if_hwassist) { ip->ip_sum = in_cksum(m0, ip->ip_hl << 2); m0->m_pkthdr.csum_flags &= ~CSUM_IP; } m_clrprotoflags(m0); /* Avoid confusing lower layers. */ error = (*ifp->if_output)(ifp, m0, sintosa(&dst), NULL); goto done; } /* Balk when DF bit is set or the interface didn't support TSO. */ if ((ip_off & IP_DF) || (m0->m_pkthdr.csum_flags & CSUM_TSO)) { error = EMSGSIZE; KMOD_IPSTAT_INC(ips_cantfrag); if (r->rt != PF_DUPTO) { icmp_error(m0, ICMP_UNREACH, ICMP_UNREACH_NEEDFRAG, 0, ifp->if_mtu); goto done; } else goto bad; } error = ip_fragment(ip, &m0, ifp->if_mtu, ifp->if_hwassist); if (error) goto bad; for (; m0; m0 = m1) { m1 = m0->m_nextpkt; m0->m_nextpkt = NULL; if (error == 0) { m_clrprotoflags(m0); error = (*ifp->if_output)(ifp, m0, sintosa(&dst), NULL); } else m_freem(m0); } if (error == 0) KMOD_IPSTAT_INC(ips_fragmented); done: if (r->rt != PF_DUPTO) *m = NULL; return; bad_locked: if (s) PF_STATE_UNLOCK(s); bad: m_freem(m0); goto done; } #endif /* INET */ #ifdef INET6 static void pf_route6(struct mbuf **m, struct pf_rule *r, int dir, struct ifnet *oifp, struct pf_state *s, struct pf_pdesc *pd) { struct mbuf *m0; struct sockaddr_in6 dst; struct ip6_hdr *ip6; struct ifnet *ifp = NULL; struct pf_addr naddr; struct pf_src_node *sn = NULL; KASSERT(m && *m && r && oifp, ("%s: invalid parameters", __func__)); KASSERT(dir == PF_IN || dir == PF_OUT, ("%s: invalid direction", __func__)); if ((pd->pf_mtag == NULL && ((pd->pf_mtag = pf_get_mtag(*m)) == NULL)) || pd->pf_mtag->routed++ > 3) { m0 = *m; *m = NULL; goto bad_locked; } if (r->rt == PF_DUPTO) { if ((m0 = m_dup(*m, M_NOWAIT)) == NULL) { if (s) PF_STATE_UNLOCK(s); return; } } else { if ((r->rt == PF_REPLYTO) == (r->direction == dir)) { if (s) PF_STATE_UNLOCK(s); return; } m0 = *m; } ip6 = mtod(m0, struct ip6_hdr *); bzero(&dst, sizeof(dst)); dst.sin6_family = AF_INET6; dst.sin6_len = sizeof(dst); dst.sin6_addr = ip6->ip6_dst; - - /* Cheat. XXX why only in the v6 case??? */ - if (r->rt == PF_FASTROUTE) { - if (s) - PF_STATE_UNLOCK(s); - m0->m_flags |= M_SKIP_FIREWALL; - ip6_output(m0, NULL, NULL, 0, NULL, NULL, NULL); - *m = NULL; - return; - } if (TAILQ_EMPTY(&r->rpool.list)) { DPFPRINTF(PF_DEBUG_URGENT, ("%s: TAILQ_EMPTY(&r->rpool.list)\n", __func__)); goto bad_locked; } if (s == NULL) { pf_map_addr(AF_INET6, r, (struct pf_addr *)&ip6->ip6_src, &naddr, NULL, &sn); if (!PF_AZERO(&naddr, AF_INET6)) PF_ACPY((struct pf_addr *)&dst.sin6_addr, &naddr, AF_INET6); ifp = r->rpool.cur->kif ? r->rpool.cur->kif->pfik_ifp : NULL; } else { if (!PF_AZERO(&s->rt_addr, AF_INET6)) PF_ACPY((struct pf_addr *)&dst.sin6_addr, &s->rt_addr, AF_INET6); ifp = s->rt_kif ? s->rt_kif->pfik_ifp : NULL; } if (s) PF_STATE_UNLOCK(s); if (ifp == NULL) goto bad; if (oifp != ifp) { if (pf_test6(PF_FWD, ifp, &m0, NULL) != PF_PASS) goto bad; else if (m0 == NULL) goto done; if (m0->m_len < sizeof(struct ip6_hdr)) { DPFPRINTF(PF_DEBUG_URGENT, ("%s: m0->m_len < sizeof(struct ip6_hdr)\n", __func__)); goto bad; } ip6 = mtod(m0, struct ip6_hdr *); } if (ifp->if_flags & IFF_LOOPBACK) m0->m_flags |= M_SKIP_FIREWALL; if (m0->m_pkthdr.csum_flags & CSUM_DELAY_DATA_IPV6 & ~ifp->if_hwassist) { uint32_t plen = m0->m_pkthdr.len - sizeof(*ip6); in6_delayed_cksum(m0, plen, sizeof(struct ip6_hdr)); m0->m_pkthdr.csum_flags &= ~CSUM_DELAY_DATA_IPV6; } /* * If the packet is too large for the outgoing interface, * send back an icmp6 error. */ if (IN6_IS_SCOPE_EMBED(&dst.sin6_addr)) dst.sin6_addr.s6_addr16[1] = htons(ifp->if_index); if ((u_long)m0->m_pkthdr.len <= ifp->if_mtu) nd6_output_ifp(ifp, ifp, m0, &dst, NULL); else { in6_ifstat_inc(ifp, ifs6_in_toobig); if (r->rt != PF_DUPTO) icmp6_error(m0, ICMP6_PACKET_TOO_BIG, 0, ifp->if_mtu); else goto bad; } done: if (r->rt != PF_DUPTO) *m = NULL; return; bad_locked: if (s) PF_STATE_UNLOCK(s); bad: m_freem(m0); goto done; } #endif /* INET6 */ /* * FreeBSD supports cksum offloads for the following drivers. * em(4), fxp(4), ixgb(4), lge(4), ndis(4), nge(4), re(4), * ti(4), txp(4), xl(4) * * CSUM_DATA_VALID | CSUM_PSEUDO_HDR : * network driver performed cksum including pseudo header, need to verify * csum_data * CSUM_DATA_VALID : * network driver performed cksum, needs to additional pseudo header * cksum computation with partial csum_data(i.e. lack of H/W support for * pseudo header, for instance hme(4), sk(4) and possibly gem(4)) * * After validating the cksum of packet, set both flag CSUM_DATA_VALID and * CSUM_PSEUDO_HDR in order to avoid recomputation of the cksum in upper * TCP/UDP layer. * Also, set csum_data to 0xffff to force cksum validation. */ static int pf_check_proto_cksum(struct mbuf *m, int off, int len, u_int8_t p, sa_family_t af) { u_int16_t sum = 0; int hw_assist = 0; struct ip *ip; if (off < sizeof(struct ip) || len < sizeof(struct udphdr)) return (1); if (m->m_pkthdr.len < off + len) return (1); switch (p) { case IPPROTO_TCP: if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) { if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR) { sum = m->m_pkthdr.csum_data; } else { ip = mtod(m, struct ip *); sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr, htonl((u_short)len + m->m_pkthdr.csum_data + IPPROTO_TCP)); } sum ^= 0xffff; ++hw_assist; } break; case IPPROTO_UDP: if (m->m_pkthdr.csum_flags & CSUM_DATA_VALID) { if (m->m_pkthdr.csum_flags & CSUM_PSEUDO_HDR) { sum = m->m_pkthdr.csum_data; } else { ip = mtod(m, struct ip *); sum = in_pseudo(ip->ip_src.s_addr, ip->ip_dst.s_addr, htonl((u_short)len + m->m_pkthdr.csum_data + IPPROTO_UDP)); } sum ^= 0xffff; ++hw_assist; } break; case IPPROTO_ICMP: #ifdef INET6 case IPPROTO_ICMPV6: #endif /* INET6 */ break; default: return (1); } if (!hw_assist) { switch (af) { case AF_INET: if (p == IPPROTO_ICMP) { if (m->m_len < off) return (1); m->m_data += off; m->m_len -= off; sum = in_cksum(m, len); m->m_data -= off; m->m_len += off; } else { if (m->m_len < sizeof(struct ip)) return (1); sum = in4_cksum(m, p, off, len); } break; #ifdef INET6 case AF_INET6: if (m->m_len < sizeof(struct ip6_hdr)) return (1); sum = in6_cksum(m, p, off, len); break; #endif /* INET6 */ default: return (1); } } if (sum) { switch (p) { case IPPROTO_TCP: { KMOD_TCPSTAT_INC(tcps_rcvbadsum); break; } case IPPROTO_UDP: { KMOD_UDPSTAT_INC(udps_badsum); break; } #ifdef INET case IPPROTO_ICMP: { KMOD_ICMPSTAT_INC(icps_checksum); break; } #endif #ifdef INET6 case IPPROTO_ICMPV6: { KMOD_ICMP6STAT_INC(icp6s_checksum); break; } #endif /* INET6 */ } return (1); } else { if (p == IPPROTO_TCP || p == IPPROTO_UDP) { m->m_pkthdr.csum_flags |= (CSUM_DATA_VALID | CSUM_PSEUDO_HDR); m->m_pkthdr.csum_data = 0xffff; } } return (0); } #ifdef INET int pf_test(int dir, struct ifnet *ifp, struct mbuf **m0, struct inpcb *inp) { struct pfi_kif *kif; u_short action, reason = 0, log = 0; struct mbuf *m = *m0; struct ip *h = NULL; struct m_tag *ipfwtag; struct pf_rule *a = NULL, *r = &V_pf_default_rule, *tr, *nr; struct pf_state *s = NULL; struct pf_ruleset *ruleset = NULL; struct pf_pdesc pd; int off, dirndx, pqid = 0; M_ASSERTPKTHDR(m); if (!V_pf_status.running) return (PF_PASS); memset(&pd, 0, sizeof(pd)); kif = (struct pfi_kif *)ifp->if_pf_kif; if (kif == NULL) { DPFPRINTF(PF_DEBUG_URGENT, ("pf_test: kif == NULL, if_xname %s\n", ifp->if_xname)); return (PF_DROP); } if (kif->pfik_flags & PFI_IFLAG_SKIP) return (PF_PASS); if (m->m_flags & M_SKIP_FIREWALL) return (PF_PASS); pd.pf_mtag = pf_find_mtag(m); PF_RULES_RLOCK(); if (ip_divert_ptr != NULL && ((ipfwtag = m_tag_locate(m, MTAG_IPFW_RULE, 0, NULL)) != NULL)) { struct ipfw_rule_ref *rr = (struct ipfw_rule_ref *)(ipfwtag+1); if (rr->info & IPFW_IS_DIVERT && rr->rulenum == 0) { if (pd.pf_mtag == NULL && ((pd.pf_mtag = pf_get_mtag(m)) == NULL)) { action = PF_DROP; goto done; } pd.pf_mtag->flags |= PF_PACKET_LOOPED; m_tag_delete(m, ipfwtag); } if (pd.pf_mtag && pd.pf_mtag->flags & PF_FASTFWD_OURS_PRESENT) { m->m_flags |= M_FASTFWD_OURS; pd.pf_mtag->flags &= ~PF_FASTFWD_OURS_PRESENT; } } else if (pf_normalize_ip(m0, dir, kif, &reason, &pd) != PF_PASS) { /* We do IP header normalization and packet reassembly here */ action = PF_DROP; goto done; } m = *m0; /* pf_normalize messes with m0 */ h = mtod(m, struct ip *); off = h->ip_hl << 2; if (off < (int)sizeof(struct ip)) { action = PF_DROP; REASON_SET(&reason, PFRES_SHORT); log = 1; goto done; } pd.src = (struct pf_addr *)&h->ip_src; pd.dst = (struct pf_addr *)&h->ip_dst; pd.sport = pd.dport = NULL; pd.ip_sum = &h->ip_sum; pd.proto_sum = NULL; pd.proto = h->ip_p; pd.dir = dir; pd.sidx = (dir == PF_IN) ? 0 : 1; pd.didx = (dir == PF_IN) ? 1 : 0; pd.af = AF_INET; pd.tos = h->ip_tos; pd.tot_len = ntohs(h->ip_len); /* handle fragments that didn't get reassembled by normalization */ if (h->ip_off & htons(IP_MF | IP_OFFMASK)) { action = pf_test_fragment(&r, dir, kif, m, h, &pd, &a, &ruleset); goto done; } switch (h->ip_p) { case IPPROTO_TCP: { struct tcphdr th; pd.hdr.tcp = &th; if (!pf_pull_hdr(m, off, &th, sizeof(th), &action, &reason, AF_INET)) { log = action != PF_PASS; goto done; } pd.p_len = pd.tot_len - off - (th.th_off << 2); if ((th.th_flags & TH_ACK) && pd.p_len == 0) pqid = 1; action = pf_normalize_tcp(dir, kif, m, 0, off, h, &pd); if (action == PF_DROP) goto done; action = pf_test_state_tcp(&s, dir, kif, m, off, h, &pd, &reason); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } case IPPROTO_UDP: { struct udphdr uh; pd.hdr.udp = &uh; if (!pf_pull_hdr(m, off, &uh, sizeof(uh), &action, &reason, AF_INET)) { log = action != PF_PASS; goto done; } if (uh.uh_dport == 0 || ntohs(uh.uh_ulen) > m->m_pkthdr.len - off || ntohs(uh.uh_ulen) < sizeof(struct udphdr)) { action = PF_DROP; REASON_SET(&reason, PFRES_SHORT); goto done; } action = pf_test_state_udp(&s, dir, kif, m, off, h, &pd); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } case IPPROTO_ICMP: { struct icmp ih; pd.hdr.icmp = &ih; if (!pf_pull_hdr(m, off, &ih, ICMP_MINLEN, &action, &reason, AF_INET)) { log = action != PF_PASS; goto done; } action = pf_test_state_icmp(&s, dir, kif, m, off, h, &pd, &reason); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } #ifdef INET6 case IPPROTO_ICMPV6: { action = PF_DROP; DPFPRINTF(PF_DEBUG_MISC, ("pf: dropping IPv4 packet with ICMPv6 payload\n")); goto done; } #endif default: action = pf_test_state_other(&s, dir, kif, m, &pd); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } done: PF_RULES_RUNLOCK(); if (action == PF_PASS && h->ip_hl > 5 && !((s && s->state_flags & PFSTATE_ALLOWOPTS) || r->allow_opts)) { action = PF_DROP; REASON_SET(&reason, PFRES_IPOPTIONS); log = r->log; DPFPRINTF(PF_DEBUG_MISC, ("pf: dropping packet with ip options\n")); } if (s && s->tag > 0 && pf_tag_packet(m, &pd, s->tag)) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); } if (r->rtableid >= 0) M_SETFIB(m, r->rtableid); if (r->scrub_flags & PFSTATE_SETPRIO) { if (pd.tos & IPTOS_LOWDELAY) pqid = 1; if (pf_ieee8021q_setpcp(m, r->set_prio[pqid])) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); log = 1; DPFPRINTF(PF_DEBUG_MISC, ("pf: failed to allocate 802.1q mtag\n")); } } #ifdef ALTQ if (action == PF_PASS && r->qid) { if (pd.pf_mtag == NULL && ((pd.pf_mtag = pf_get_mtag(m)) == NULL)) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); } else { if (s != NULL) pd.pf_mtag->qid_hash = pf_state_hash(s); if (pqid || (pd.tos & IPTOS_LOWDELAY)) pd.pf_mtag->qid = r->pqid; else pd.pf_mtag->qid = r->qid; /* Add hints for ecn. */ pd.pf_mtag->hdr = h; } } #endif /* ALTQ */ /* * connections redirected to loopback should not match sockets * bound specifically to loopback due to security implications, * see tcp_input() and in_pcblookup_listen(). */ if (dir == PF_IN && action == PF_PASS && (pd.proto == IPPROTO_TCP || pd.proto == IPPROTO_UDP) && s != NULL && s->nat_rule.ptr != NULL && (s->nat_rule.ptr->action == PF_RDR || s->nat_rule.ptr->action == PF_BINAT) && (ntohl(pd.dst->v4.s_addr) >> IN_CLASSA_NSHIFT) == IN_LOOPBACKNET) m->m_flags |= M_SKIP_FIREWALL; if (action == PF_PASS && r->divert.port && ip_divert_ptr != NULL && !PACKET_LOOPED(&pd)) { ipfwtag = m_tag_alloc(MTAG_IPFW_RULE, 0, sizeof(struct ipfw_rule_ref), M_NOWAIT | M_ZERO); if (ipfwtag != NULL) { ((struct ipfw_rule_ref *)(ipfwtag+1))->info = ntohs(r->divert.port); ((struct ipfw_rule_ref *)(ipfwtag+1))->rulenum = dir; if (s) PF_STATE_UNLOCK(s); m_tag_prepend(m, ipfwtag); if (m->m_flags & M_FASTFWD_OURS) { if (pd.pf_mtag == NULL && ((pd.pf_mtag = pf_get_mtag(m)) == NULL)) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); log = 1; DPFPRINTF(PF_DEBUG_MISC, ("pf: failed to allocate tag\n")); } else { pd.pf_mtag->flags |= PF_FASTFWD_OURS_PRESENT; m->m_flags &= ~M_FASTFWD_OURS; } } ip_divert_ptr(*m0, dir == PF_IN ? DIR_IN : DIR_OUT); *m0 = NULL; return (action); } else { /* XXX: ipfw has the same behaviour! */ action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); log = 1; DPFPRINTF(PF_DEBUG_MISC, ("pf: failed to allocate divert tag\n")); } } if (log) { struct pf_rule *lr; if (s != NULL && s->nat_rule.ptr != NULL && s->nat_rule.ptr->log & PF_LOG_ALL) lr = s->nat_rule.ptr; else lr = r; PFLOG_PACKET(kif, m, AF_INET, dir, reason, lr, a, ruleset, &pd, (s == NULL)); } kif->pfik_bytes[0][dir == PF_OUT][action != PF_PASS] += pd.tot_len; kif->pfik_packets[0][dir == PF_OUT][action != PF_PASS]++; if (action == PF_PASS || r->action == PF_DROP) { dirndx = (dir == PF_OUT); r->packets[dirndx]++; r->bytes[dirndx] += pd.tot_len; if (a != NULL) { a->packets[dirndx]++; a->bytes[dirndx] += pd.tot_len; } if (s != NULL) { if (s->nat_rule.ptr != NULL) { s->nat_rule.ptr->packets[dirndx]++; s->nat_rule.ptr->bytes[dirndx] += pd.tot_len; } if (s->src_node != NULL) { s->src_node->packets[dirndx]++; s->src_node->bytes[dirndx] += pd.tot_len; } if (s->nat_src_node != NULL) { s->nat_src_node->packets[dirndx]++; s->nat_src_node->bytes[dirndx] += pd.tot_len; } dirndx = (dir == s->direction) ? 0 : 1; s->packets[dirndx]++; s->bytes[dirndx] += pd.tot_len; } tr = r; nr = (s != NULL) ? s->nat_rule.ptr : pd.nat_rule; if (nr != NULL && r == &V_pf_default_rule) tr = nr; if (tr->src.addr.type == PF_ADDR_TABLE) pfr_update_stats(tr->src.addr.p.tbl, (s == NULL) ? pd.src : &s->key[(s->direction == PF_IN)]-> addr[(s->direction == PF_OUT)], pd.af, pd.tot_len, dir == PF_OUT, r->action == PF_PASS, tr->src.neg); if (tr->dst.addr.type == PF_ADDR_TABLE) pfr_update_stats(tr->dst.addr.p.tbl, (s == NULL) ? pd.dst : &s->key[(s->direction == PF_IN)]-> addr[(s->direction == PF_IN)], pd.af, pd.tot_len, dir == PF_OUT, r->action == PF_PASS, tr->dst.neg); } switch (action) { case PF_SYNPROXY_DROP: m_freem(*m0); case PF_DEFER: *m0 = NULL; action = PF_PASS; break; case PF_DROP: m_freem(*m0); *m0 = NULL; break; default: /* pf_route() returns unlocked. */ if (r->rt) { pf_route(m0, r, dir, kif->pfik_ifp, s, &pd); return (action); } break; } if (s) PF_STATE_UNLOCK(s); return (action); } #endif /* INET */ #ifdef INET6 int pf_test6(int dir, struct ifnet *ifp, struct mbuf **m0, struct inpcb *inp) { struct pfi_kif *kif; u_short action, reason = 0, log = 0; struct mbuf *m = *m0, *n = NULL; struct m_tag *mtag; struct ip6_hdr *h = NULL; struct pf_rule *a = NULL, *r = &V_pf_default_rule, *tr, *nr; struct pf_state *s = NULL; struct pf_ruleset *ruleset = NULL; struct pf_pdesc pd; int off, terminal = 0, dirndx, rh_cnt = 0, pqid = 0; int fwdir = dir; M_ASSERTPKTHDR(m); /* Detect packet forwarding. * If the input interface is different from the output interface we're * forwarding. * We do need to be careful about bridges. If the * net.link.bridge.pfil_bridge sysctl is set we can be filtering on a * bridge, so if the input interface is a bridge member and the output * interface is its bridge or a member of the same bridge we're not * actually forwarding but bridging. */ if (dir == PF_OUT && m->m_pkthdr.rcvif && ifp != m->m_pkthdr.rcvif && (m->m_pkthdr.rcvif->if_bridge == NULL || (m->m_pkthdr.rcvif->if_bridge != ifp->if_softc && m->m_pkthdr.rcvif->if_bridge != ifp->if_bridge))) fwdir = PF_FWD; if (!V_pf_status.running) return (PF_PASS); memset(&pd, 0, sizeof(pd)); pd.pf_mtag = pf_find_mtag(m); if (pd.pf_mtag && pd.pf_mtag->flags & PF_TAG_GENERATED) return (PF_PASS); kif = (struct pfi_kif *)ifp->if_pf_kif; if (kif == NULL) { DPFPRINTF(PF_DEBUG_URGENT, ("pf_test6: kif == NULL, if_xname %s\n", ifp->if_xname)); return (PF_DROP); } if (kif->pfik_flags & PFI_IFLAG_SKIP) return (PF_PASS); if (m->m_flags & M_SKIP_FIREWALL) return (PF_PASS); PF_RULES_RLOCK(); /* We do IP header normalization and packet reassembly here */ if (pf_normalize_ip6(m0, dir, kif, &reason, &pd) != PF_PASS) { action = PF_DROP; goto done; } m = *m0; /* pf_normalize messes with m0 */ h = mtod(m, struct ip6_hdr *); #if 1 /* * we do not support jumbogram yet. if we keep going, zero ip6_plen * will do something bad, so drop the packet for now. */ if (htons(h->ip6_plen) == 0) { action = PF_DROP; REASON_SET(&reason, PFRES_NORM); /*XXX*/ goto done; } #endif pd.src = (struct pf_addr *)&h->ip6_src; pd.dst = (struct pf_addr *)&h->ip6_dst; pd.sport = pd.dport = NULL; pd.ip_sum = NULL; pd.proto_sum = NULL; pd.dir = dir; pd.sidx = (dir == PF_IN) ? 0 : 1; pd.didx = (dir == PF_IN) ? 1 : 0; pd.af = AF_INET6; pd.tos = 0; pd.tot_len = ntohs(h->ip6_plen) + sizeof(struct ip6_hdr); off = ((caddr_t)h - m->m_data) + sizeof(struct ip6_hdr); pd.proto = h->ip6_nxt; do { switch (pd.proto) { case IPPROTO_FRAGMENT: action = pf_test_fragment(&r, dir, kif, m, h, &pd, &a, &ruleset); if (action == PF_DROP) REASON_SET(&reason, PFRES_FRAG); goto done; case IPPROTO_ROUTING: { struct ip6_rthdr rthdr; if (rh_cnt++) { DPFPRINTF(PF_DEBUG_MISC, ("pf: IPv6 more than one rthdr\n")); action = PF_DROP; REASON_SET(&reason, PFRES_IPOPTIONS); log = 1; goto done; } if (!pf_pull_hdr(m, off, &rthdr, sizeof(rthdr), NULL, &reason, pd.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: IPv6 short rthdr\n")); action = PF_DROP; REASON_SET(&reason, PFRES_SHORT); log = 1; goto done; } if (rthdr.ip6r_type == IPV6_RTHDR_TYPE_0) { DPFPRINTF(PF_DEBUG_MISC, ("pf: IPv6 rthdr0\n")); action = PF_DROP; REASON_SET(&reason, PFRES_IPOPTIONS); log = 1; goto done; } /* FALLTHROUGH */ } case IPPROTO_AH: case IPPROTO_HOPOPTS: case IPPROTO_DSTOPTS: { /* get next header and header length */ struct ip6_ext opt6; if (!pf_pull_hdr(m, off, &opt6, sizeof(opt6), NULL, &reason, pd.af)) { DPFPRINTF(PF_DEBUG_MISC, ("pf: IPv6 short opt\n")); action = PF_DROP; log = 1; goto done; } if (pd.proto == IPPROTO_AH) off += (opt6.ip6e_len + 2) * 4; else off += (opt6.ip6e_len + 1) * 8; pd.proto = opt6.ip6e_nxt; /* goto the next header */ break; } default: terminal++; break; } } while (!terminal); /* if there's no routing header, use unmodified mbuf for checksumming */ if (!n) n = m; switch (pd.proto) { case IPPROTO_TCP: { struct tcphdr th; pd.hdr.tcp = &th; if (!pf_pull_hdr(m, off, &th, sizeof(th), &action, &reason, AF_INET6)) { log = action != PF_PASS; goto done; } pd.p_len = pd.tot_len - off - (th.th_off << 2); action = pf_normalize_tcp(dir, kif, m, 0, off, h, &pd); if (action == PF_DROP) goto done; action = pf_test_state_tcp(&s, dir, kif, m, off, h, &pd, &reason); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } case IPPROTO_UDP: { struct udphdr uh; pd.hdr.udp = &uh; if (!pf_pull_hdr(m, off, &uh, sizeof(uh), &action, &reason, AF_INET6)) { log = action != PF_PASS; goto done; } if (uh.uh_dport == 0 || ntohs(uh.uh_ulen) > m->m_pkthdr.len - off || ntohs(uh.uh_ulen) < sizeof(struct udphdr)) { action = PF_DROP; REASON_SET(&reason, PFRES_SHORT); goto done; } action = pf_test_state_udp(&s, dir, kif, m, off, h, &pd); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } case IPPROTO_ICMP: { action = PF_DROP; DPFPRINTF(PF_DEBUG_MISC, ("pf: dropping IPv6 packet with ICMPv4 payload\n")); goto done; } case IPPROTO_ICMPV6: { struct icmp6_hdr ih; pd.hdr.icmp6 = &ih; if (!pf_pull_hdr(m, off, &ih, sizeof(ih), &action, &reason, AF_INET6)) { log = action != PF_PASS; goto done; } action = pf_test_state_icmp(&s, dir, kif, m, off, h, &pd, &reason); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } default: action = pf_test_state_other(&s, dir, kif, m, &pd); if (action == PF_PASS) { if (pfsync_update_state_ptr != NULL) pfsync_update_state_ptr(s); r = s->rule.ptr; a = s->anchor.ptr; log = s->log; } else if (s == NULL) action = pf_test_rule(&r, &s, dir, kif, m, off, &pd, &a, &ruleset, inp); break; } done: PF_RULES_RUNLOCK(); if (n != m) { m_freem(n); n = NULL; } /* handle dangerous IPv6 extension headers. */ if (action == PF_PASS && rh_cnt && !((s && s->state_flags & PFSTATE_ALLOWOPTS) || r->allow_opts)) { action = PF_DROP; REASON_SET(&reason, PFRES_IPOPTIONS); log = r->log; DPFPRINTF(PF_DEBUG_MISC, ("pf: dropping packet with dangerous v6 headers\n")); } if (s && s->tag > 0 && pf_tag_packet(m, &pd, s->tag)) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); } if (r->rtableid >= 0) M_SETFIB(m, r->rtableid); if (r->scrub_flags & PFSTATE_SETPRIO) { if (pd.tos & IPTOS_LOWDELAY) pqid = 1; if (pf_ieee8021q_setpcp(m, r->set_prio[pqid])) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); log = 1; DPFPRINTF(PF_DEBUG_MISC, ("pf: failed to allocate 802.1q mtag\n")); } } #ifdef ALTQ if (action == PF_PASS && r->qid) { if (pd.pf_mtag == NULL && ((pd.pf_mtag = pf_get_mtag(m)) == NULL)) { action = PF_DROP; REASON_SET(&reason, PFRES_MEMORY); } else { if (s != NULL) pd.pf_mtag->qid_hash = pf_state_hash(s); if (pd.tos & IPTOS_LOWDELAY) pd.pf_mtag->qid = r->pqid; else pd.pf_mtag->qid = r->qid; /* Add hints for ecn. */ pd.pf_mtag->hdr = h; } } #endif /* ALTQ */ if (dir == PF_IN && action == PF_PASS && (pd.proto == IPPROTO_TCP || pd.proto == IPPROTO_UDP) && s != NULL && s->nat_rule.ptr != NULL && (s->nat_rule.ptr->action == PF_RDR || s->nat_rule.ptr->action == PF_BINAT) && IN6_IS_ADDR_LOOPBACK(&pd.dst->v6)) m->m_flags |= M_SKIP_FIREWALL; /* XXX: Anybody working on it?! */ if (r->divert.port) printf("pf: divert(9) is not supported for IPv6\n"); if (log) { struct pf_rule *lr; if (s != NULL && s->nat_rule.ptr != NULL && s->nat_rule.ptr->log & PF_LOG_ALL) lr = s->nat_rule.ptr; else lr = r; PFLOG_PACKET(kif, m, AF_INET6, dir, reason, lr, a, ruleset, &pd, (s == NULL)); } kif->pfik_bytes[1][dir == PF_OUT][action != PF_PASS] += pd.tot_len; kif->pfik_packets[1][dir == PF_OUT][action != PF_PASS]++; if (action == PF_PASS || r->action == PF_DROP) { dirndx = (dir == PF_OUT); r->packets[dirndx]++; r->bytes[dirndx] += pd.tot_len; if (a != NULL) { a->packets[dirndx]++; a->bytes[dirndx] += pd.tot_len; } if (s != NULL) { if (s->nat_rule.ptr != NULL) { s->nat_rule.ptr->packets[dirndx]++; s->nat_rule.ptr->bytes[dirndx] += pd.tot_len; } if (s->src_node != NULL) { s->src_node->packets[dirndx]++; s->src_node->bytes[dirndx] += pd.tot_len; } if (s->nat_src_node != NULL) { s->nat_src_node->packets[dirndx]++; s->nat_src_node->bytes[dirndx] += pd.tot_len; } dirndx = (dir == s->direction) ? 0 : 1; s->packets[dirndx]++; s->bytes[dirndx] += pd.tot_len; } tr = r; nr = (s != NULL) ? s->nat_rule.ptr : pd.nat_rule; if (nr != NULL && r == &V_pf_default_rule) tr = nr; if (tr->src.addr.type == PF_ADDR_TABLE) pfr_update_stats(tr->src.addr.p.tbl, (s == NULL) ? pd.src : &s->key[(s->direction == PF_IN)]->addr[0], pd.af, pd.tot_len, dir == PF_OUT, r->action == PF_PASS, tr->src.neg); if (tr->dst.addr.type == PF_ADDR_TABLE) pfr_update_stats(tr->dst.addr.p.tbl, (s == NULL) ? pd.dst : &s->key[(s->direction == PF_IN)]->addr[1], pd.af, pd.tot_len, dir == PF_OUT, r->action == PF_PASS, tr->dst.neg); } switch (action) { case PF_SYNPROXY_DROP: m_freem(*m0); case PF_DEFER: *m0 = NULL; action = PF_PASS; break; case PF_DROP: m_freem(*m0); *m0 = NULL; break; default: /* pf_route6() returns unlocked. */ if (r->rt) { pf_route6(m0, r, dir, kif->pfik_ifp, s, &pd); return (action); } break; } if (s) PF_STATE_UNLOCK(s); /* If reassembled packet passed, create new fragments. */ if (action == PF_PASS && *m0 && fwdir == PF_FWD && (mtag = m_tag_find(m, PF_REASSEMBLED, NULL)) != NULL) action = pf_refragment6(ifp, m0, mtag); return (action); } #endif /* INET6 */ Index: head/sys/netpfil/pf/pf_ioctl.c =================================================================== --- head/sys/netpfil/pf/pf_ioctl.c (revision 306683) +++ head/sys/netpfil/pf/pf_ioctl.c (revision 306684) @@ -1,3860 +1,3860 @@ /*- * Copyright (c) 2001 Daniel Hartmeier * Copyright (c) 2002,2003 Henning Brauer * Copyright (c) 2012 Gleb Smirnoff * All rights reserved. * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions * are met: * * - Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * - 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. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS 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 * COPYRIGHT HOLDERS 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. * * Effort sponsored in part by the Defense Advanced Research Projects * Agency (DARPA) and Air Force Research Laboratory, Air Force * Materiel Command, USAF, under agreement number F30602-01-2-0537. * * $OpenBSD: pf_ioctl.c,v 1.213 2009/02/15 21:46:12 mbalmer Exp $ */ #include __FBSDID("$FreeBSD$"); #include "opt_inet.h" #include "opt_inet6.h" #include "opt_bpf.h" #include "opt_pf.h" #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #ifdef INET6 #include #endif /* INET6 */ #ifdef ALTQ #include #endif static struct pf_pool *pf_get_pool(char *, u_int32_t, u_int8_t, u_int32_t, u_int8_t, u_int8_t, u_int8_t); static void pf_mv_pool(struct pf_palist *, struct pf_palist *); static void pf_empty_pool(struct pf_palist *); static int pfioctl(struct cdev *, u_long, caddr_t, int, struct thread *); #ifdef ALTQ static int pf_begin_altq(u_int32_t *); static int pf_rollback_altq(u_int32_t); static int pf_commit_altq(u_int32_t); static int pf_enable_altq(struct pf_altq *); static int pf_disable_altq(struct pf_altq *); static u_int32_t pf_qname2qid(char *); static void pf_qid_unref(u_int32_t); #endif /* ALTQ */ static int pf_begin_rules(u_int32_t *, int, const char *); static int pf_rollback_rules(u_int32_t, int, char *); static int pf_setup_pfsync_matching(struct pf_ruleset *); static void pf_hash_rule(MD5_CTX *, struct pf_rule *); static void pf_hash_rule_addr(MD5_CTX *, struct pf_rule_addr *); static int pf_commit_rules(u_int32_t, int, char *); static int pf_addr_setup(struct pf_ruleset *, struct pf_addr_wrap *, sa_family_t); static void pf_addr_copyout(struct pf_addr_wrap *); VNET_DEFINE(struct pf_rule, pf_default_rule); #ifdef ALTQ static VNET_DEFINE(int, pf_altq_running); #define V_pf_altq_running VNET(pf_altq_running) #endif #define TAGID_MAX 50000 struct pf_tagname { TAILQ_ENTRY(pf_tagname) entries; char name[PF_TAG_NAME_SIZE]; uint16_t tag; int ref; }; TAILQ_HEAD(pf_tags, pf_tagname); #define V_pf_tags VNET(pf_tags) VNET_DEFINE(struct pf_tags, pf_tags); #define V_pf_qids VNET(pf_qids) VNET_DEFINE(struct pf_tags, pf_qids); static MALLOC_DEFINE(M_PFTAG, "pf_tag", "pf(4) tag names"); static MALLOC_DEFINE(M_PFALTQ, "pf_altq", "pf(4) altq configuration db"); static MALLOC_DEFINE(M_PFRULE, "pf_rule", "pf(4) rules"); #if (PF_QNAME_SIZE != PF_TAG_NAME_SIZE) #error PF_QNAME_SIZE must be equal to PF_TAG_NAME_SIZE #endif static u_int16_t tagname2tag(struct pf_tags *, char *); static u_int16_t pf_tagname2tag(char *); static void tag_unref(struct pf_tags *, u_int16_t); #define DPFPRINTF(n, x) if (V_pf_status.debug >= (n)) printf x struct cdev *pf_dev; /* * XXX - These are new and need to be checked when moveing to a new version */ static void pf_clear_states(void); static int pf_clear_tables(void); static void pf_clear_srcnodes(struct pf_src_node *); static void pf_kill_srcnodes(struct pfioc_src_node_kill *); static void pf_tbladdr_copyout(struct pf_addr_wrap *); /* * Wrapper functions for pfil(9) hooks */ #ifdef INET static int pf_check_in(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp); static int pf_check_out(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp); #endif #ifdef INET6 static int pf_check6_in(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp); static int pf_check6_out(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp); #endif static int hook_pf(void); static int dehook_pf(void); static int shutdown_pf(void); static int pf_load(void); static int pf_unload(void); static struct cdevsw pf_cdevsw = { .d_ioctl = pfioctl, .d_name = PF_NAME, .d_version = D_VERSION, }; static volatile VNET_DEFINE(int, pf_pfil_hooked); #define V_pf_pfil_hooked VNET(pf_pfil_hooked) /* * We need a flag that is neither hooked nor running to know when * the VNET is "valid". We primarily need this to control (global) * external event, e.g., eventhandlers. */ VNET_DEFINE(int, pf_vnet_active); #define V_pf_vnet_active VNET(pf_vnet_active) int pf_end_threads; struct rwlock pf_rules_lock; struct sx pf_ioctl_lock; /* pfsync */ pfsync_state_import_t *pfsync_state_import_ptr = NULL; pfsync_insert_state_t *pfsync_insert_state_ptr = NULL; pfsync_update_state_t *pfsync_update_state_ptr = NULL; pfsync_delete_state_t *pfsync_delete_state_ptr = NULL; pfsync_clear_states_t *pfsync_clear_states_ptr = NULL; pfsync_defer_t *pfsync_defer_ptr = NULL; /* pflog */ pflog_packet_t *pflog_packet_ptr = NULL; static void pfattach_vnet(void) { u_int32_t *my_timeout = V_pf_default_rule.timeout; pf_initialize(); pfr_initialize(); pfi_initialize_vnet(); pf_normalize_init(); V_pf_limits[PF_LIMIT_STATES].limit = PFSTATE_HIWAT; V_pf_limits[PF_LIMIT_SRC_NODES].limit = PFSNODE_HIWAT; RB_INIT(&V_pf_anchors); pf_init_ruleset(&pf_main_ruleset); /* default rule should never be garbage collected */ V_pf_default_rule.entries.tqe_prev = &V_pf_default_rule.entries.tqe_next; #ifdef PF_DEFAULT_TO_DROP V_pf_default_rule.action = PF_DROP; #else V_pf_default_rule.action = PF_PASS; #endif V_pf_default_rule.nr = -1; V_pf_default_rule.rtableid = -1; V_pf_default_rule.states_cur = counter_u64_alloc(M_WAITOK); V_pf_default_rule.states_tot = counter_u64_alloc(M_WAITOK); V_pf_default_rule.src_nodes = counter_u64_alloc(M_WAITOK); /* initialize default timeouts */ my_timeout[PFTM_TCP_FIRST_PACKET] = PFTM_TCP_FIRST_PACKET_VAL; my_timeout[PFTM_TCP_OPENING] = PFTM_TCP_OPENING_VAL; my_timeout[PFTM_TCP_ESTABLISHED] = PFTM_TCP_ESTABLISHED_VAL; my_timeout[PFTM_TCP_CLOSING] = PFTM_TCP_CLOSING_VAL; my_timeout[PFTM_TCP_FIN_WAIT] = PFTM_TCP_FIN_WAIT_VAL; my_timeout[PFTM_TCP_CLOSED] = PFTM_TCP_CLOSED_VAL; my_timeout[PFTM_UDP_FIRST_PACKET] = PFTM_UDP_FIRST_PACKET_VAL; my_timeout[PFTM_UDP_SINGLE] = PFTM_UDP_SINGLE_VAL; my_timeout[PFTM_UDP_MULTIPLE] = PFTM_UDP_MULTIPLE_VAL; my_timeout[PFTM_ICMP_FIRST_PACKET] = PFTM_ICMP_FIRST_PACKET_VAL; my_timeout[PFTM_ICMP_ERROR_REPLY] = PFTM_ICMP_ERROR_REPLY_VAL; my_timeout[PFTM_OTHER_FIRST_PACKET] = PFTM_OTHER_FIRST_PACKET_VAL; my_timeout[PFTM_OTHER_SINGLE] = PFTM_OTHER_SINGLE_VAL; my_timeout[PFTM_OTHER_MULTIPLE] = PFTM_OTHER_MULTIPLE_VAL; my_timeout[PFTM_FRAG] = PFTM_FRAG_VAL; my_timeout[PFTM_INTERVAL] = PFTM_INTERVAL_VAL; my_timeout[PFTM_SRC_NODE] = PFTM_SRC_NODE_VAL; my_timeout[PFTM_TS_DIFF] = PFTM_TS_DIFF_VAL; my_timeout[PFTM_ADAPTIVE_START] = PFSTATE_ADAPT_START; my_timeout[PFTM_ADAPTIVE_END] = PFSTATE_ADAPT_END; bzero(&V_pf_status, sizeof(V_pf_status)); V_pf_status.debug = PF_DEBUG_URGENT; V_pf_pfil_hooked = 0; /* XXX do our best to avoid a conflict */ V_pf_status.hostid = arc4random(); for (int i = 0; i < PFRES_MAX; i++) V_pf_status.counters[i] = counter_u64_alloc(M_WAITOK); for (int i = 0; i < LCNT_MAX; i++) V_pf_status.lcounters[i] = counter_u64_alloc(M_WAITOK); for (int i = 0; i < FCNT_MAX; i++) V_pf_status.fcounters[i] = counter_u64_alloc(M_WAITOK); for (int i = 0; i < SCNT_MAX; i++) V_pf_status.scounters[i] = counter_u64_alloc(M_WAITOK); if (swi_add(NULL, "pf send", pf_intr, curvnet, SWI_NET, INTR_MPSAFE, &V_pf_swi_cookie) != 0) /* XXXGL: leaked all above. */ return; } static struct pf_pool * pf_get_pool(char *anchor, u_int32_t ticket, u_int8_t rule_action, u_int32_t rule_number, u_int8_t r_last, u_int8_t active, u_int8_t check_ticket) { struct pf_ruleset *ruleset; struct pf_rule *rule; int rs_num; ruleset = pf_find_ruleset(anchor); if (ruleset == NULL) return (NULL); rs_num = pf_get_ruleset_number(rule_action); if (rs_num >= PF_RULESET_MAX) return (NULL); if (active) { if (check_ticket && ticket != ruleset->rules[rs_num].active.ticket) return (NULL); if (r_last) rule = TAILQ_LAST(ruleset->rules[rs_num].active.ptr, pf_rulequeue); else rule = TAILQ_FIRST(ruleset->rules[rs_num].active.ptr); } else { if (check_ticket && ticket != ruleset->rules[rs_num].inactive.ticket) return (NULL); if (r_last) rule = TAILQ_LAST(ruleset->rules[rs_num].inactive.ptr, pf_rulequeue); else rule = TAILQ_FIRST(ruleset->rules[rs_num].inactive.ptr); } if (!r_last) { while ((rule != NULL) && (rule->nr != rule_number)) rule = TAILQ_NEXT(rule, entries); } if (rule == NULL) return (NULL); return (&rule->rpool); } static void pf_mv_pool(struct pf_palist *poola, struct pf_palist *poolb) { struct pf_pooladdr *mv_pool_pa; while ((mv_pool_pa = TAILQ_FIRST(poola)) != NULL) { TAILQ_REMOVE(poola, mv_pool_pa, entries); TAILQ_INSERT_TAIL(poolb, mv_pool_pa, entries); } } static void pf_empty_pool(struct pf_palist *poola) { struct pf_pooladdr *pa; while ((pa = TAILQ_FIRST(poola)) != NULL) { switch (pa->addr.type) { case PF_ADDR_DYNIFTL: pfi_dynaddr_remove(pa->addr.p.dyn); break; case PF_ADDR_TABLE: /* XXX: this could be unfinished pooladdr on pabuf */ if (pa->addr.p.tbl != NULL) pfr_detach_table(pa->addr.p.tbl); break; } if (pa->kif) pfi_kif_unref(pa->kif); TAILQ_REMOVE(poola, pa, entries); free(pa, M_PFRULE); } } static void pf_unlink_rule(struct pf_rulequeue *rulequeue, struct pf_rule *rule) { PF_RULES_WASSERT(); TAILQ_REMOVE(rulequeue, rule, entries); PF_UNLNKDRULES_LOCK(); rule->rule_flag |= PFRULE_REFS; TAILQ_INSERT_TAIL(&V_pf_unlinked_rules, rule, entries); PF_UNLNKDRULES_UNLOCK(); } void pf_free_rule(struct pf_rule *rule) { PF_RULES_WASSERT(); if (rule->tag) tag_unref(&V_pf_tags, rule->tag); if (rule->match_tag) tag_unref(&V_pf_tags, rule->match_tag); #ifdef ALTQ if (rule->pqid != rule->qid) pf_qid_unref(rule->pqid); pf_qid_unref(rule->qid); #endif switch (rule->src.addr.type) { case PF_ADDR_DYNIFTL: pfi_dynaddr_remove(rule->src.addr.p.dyn); break; case PF_ADDR_TABLE: pfr_detach_table(rule->src.addr.p.tbl); break; } switch (rule->dst.addr.type) { case PF_ADDR_DYNIFTL: pfi_dynaddr_remove(rule->dst.addr.p.dyn); break; case PF_ADDR_TABLE: pfr_detach_table(rule->dst.addr.p.tbl); break; } if (rule->overload_tbl) pfr_detach_table(rule->overload_tbl); if (rule->kif) pfi_kif_unref(rule->kif); pf_anchor_remove(rule); pf_empty_pool(&rule->rpool.list); counter_u64_free(rule->states_cur); counter_u64_free(rule->states_tot); counter_u64_free(rule->src_nodes); free(rule, M_PFRULE); } static u_int16_t tagname2tag(struct pf_tags *head, char *tagname) { struct pf_tagname *tag, *p = NULL; u_int16_t new_tagid = 1; PF_RULES_WASSERT(); TAILQ_FOREACH(tag, head, entries) if (strcmp(tagname, tag->name) == 0) { tag->ref++; return (tag->tag); } /* * to avoid fragmentation, we do a linear search from the beginning * and take the first free slot we find. if there is none or the list * is empty, append a new entry at the end. */ /* new entry */ if (!TAILQ_EMPTY(head)) for (p = TAILQ_FIRST(head); p != NULL && p->tag == new_tagid; p = TAILQ_NEXT(p, entries)) new_tagid = p->tag + 1; if (new_tagid > TAGID_MAX) return (0); /* allocate and fill new struct pf_tagname */ tag = malloc(sizeof(*tag), M_PFTAG, M_NOWAIT|M_ZERO); if (tag == NULL) return (0); strlcpy(tag->name, tagname, sizeof(tag->name)); tag->tag = new_tagid; tag->ref++; if (p != NULL) /* insert new entry before p */ TAILQ_INSERT_BEFORE(p, tag, entries); else /* either list empty or no free slot in between */ TAILQ_INSERT_TAIL(head, tag, entries); return (tag->tag); } static void tag_unref(struct pf_tags *head, u_int16_t tag) { struct pf_tagname *p, *next; PF_RULES_WASSERT(); for (p = TAILQ_FIRST(head); p != NULL; p = next) { next = TAILQ_NEXT(p, entries); if (tag == p->tag) { if (--p->ref == 0) { TAILQ_REMOVE(head, p, entries); free(p, M_PFTAG); } break; } } } static u_int16_t pf_tagname2tag(char *tagname) { return (tagname2tag(&V_pf_tags, tagname)); } #ifdef ALTQ static u_int32_t pf_qname2qid(char *qname) { return ((u_int32_t)tagname2tag(&V_pf_qids, qname)); } static void pf_qid_unref(u_int32_t qid) { tag_unref(&V_pf_qids, (u_int16_t)qid); } static int pf_begin_altq(u_int32_t *ticket) { struct pf_altq *altq; int error = 0; PF_RULES_WASSERT(); /* Purge the old altq list */ while ((altq = TAILQ_FIRST(V_pf_altqs_inactive)) != NULL) { TAILQ_REMOVE(V_pf_altqs_inactive, altq, entries); if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { /* detach and destroy the discipline */ error = altq_remove(altq); } else pf_qid_unref(altq->qid); free(altq, M_PFALTQ); } if (error) return (error); *ticket = ++V_ticket_altqs_inactive; V_altqs_inactive_open = 1; return (0); } static int pf_rollback_altq(u_int32_t ticket) { struct pf_altq *altq; int error = 0; PF_RULES_WASSERT(); if (!V_altqs_inactive_open || ticket != V_ticket_altqs_inactive) return (0); /* Purge the old altq list */ while ((altq = TAILQ_FIRST(V_pf_altqs_inactive)) != NULL) { TAILQ_REMOVE(V_pf_altqs_inactive, altq, entries); if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { /* detach and destroy the discipline */ error = altq_remove(altq); } else pf_qid_unref(altq->qid); free(altq, M_PFALTQ); } V_altqs_inactive_open = 0; return (error); } static int pf_commit_altq(u_int32_t ticket) { struct pf_altqqueue *old_altqs; struct pf_altq *altq; int err, error = 0; PF_RULES_WASSERT(); if (!V_altqs_inactive_open || ticket != V_ticket_altqs_inactive) return (EBUSY); /* swap altqs, keep the old. */ old_altqs = V_pf_altqs_active; V_pf_altqs_active = V_pf_altqs_inactive; V_pf_altqs_inactive = old_altqs; V_ticket_altqs_active = V_ticket_altqs_inactive; /* Attach new disciplines */ TAILQ_FOREACH(altq, V_pf_altqs_active, entries) { if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { /* attach the discipline */ error = altq_pfattach(altq); if (error == 0 && V_pf_altq_running) error = pf_enable_altq(altq); if (error != 0) return (error); } } /* Purge the old altq list */ while ((altq = TAILQ_FIRST(V_pf_altqs_inactive)) != NULL) { TAILQ_REMOVE(V_pf_altqs_inactive, altq, entries); if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { /* detach and destroy the discipline */ if (V_pf_altq_running) error = pf_disable_altq(altq); err = altq_pfdetach(altq); if (err != 0 && error == 0) error = err; err = altq_remove(altq); if (err != 0 && error == 0) error = err; } else pf_qid_unref(altq->qid); free(altq, M_PFALTQ); } V_altqs_inactive_open = 0; return (error); } static int pf_enable_altq(struct pf_altq *altq) { struct ifnet *ifp; struct tb_profile tb; int error = 0; if ((ifp = ifunit(altq->ifname)) == NULL) return (EINVAL); if (ifp->if_snd.altq_type != ALTQT_NONE) error = altq_enable(&ifp->if_snd); /* set tokenbucket regulator */ if (error == 0 && ifp != NULL && ALTQ_IS_ENABLED(&ifp->if_snd)) { tb.rate = altq->ifbandwidth; tb.depth = altq->tbrsize; error = tbr_set(&ifp->if_snd, &tb); } return (error); } static int pf_disable_altq(struct pf_altq *altq) { struct ifnet *ifp; struct tb_profile tb; int error; if ((ifp = ifunit(altq->ifname)) == NULL) return (EINVAL); /* * when the discipline is no longer referenced, it was overridden * by a new one. if so, just return. */ if (altq->altq_disc != ifp->if_snd.altq_disc) return (0); error = altq_disable(&ifp->if_snd); if (error == 0) { /* clear tokenbucket regulator */ tb.rate = 0; error = tbr_set(&ifp->if_snd, &tb); } return (error); } void pf_altq_ifnet_event(struct ifnet *ifp, int remove) { struct ifnet *ifp1; struct pf_altq *a1, *a2, *a3; u_int32_t ticket; int error = 0; /* Interrupt userland queue modifications */ if (V_altqs_inactive_open) pf_rollback_altq(V_ticket_altqs_inactive); /* Start new altq ruleset */ if (pf_begin_altq(&ticket)) return; /* Copy the current active set */ TAILQ_FOREACH(a1, V_pf_altqs_active, entries) { a2 = malloc(sizeof(*a2), M_PFALTQ, M_NOWAIT); if (a2 == NULL) { error = ENOMEM; break; } bcopy(a1, a2, sizeof(struct pf_altq)); if (a2->qname[0] != 0) { if ((a2->qid = pf_qname2qid(a2->qname)) == 0) { error = EBUSY; free(a2, M_PFALTQ); break; } a2->altq_disc = NULL; TAILQ_FOREACH(a3, V_pf_altqs_inactive, entries) { if (strncmp(a3->ifname, a2->ifname, IFNAMSIZ) == 0 && a3->qname[0] == 0) { a2->altq_disc = a3->altq_disc; break; } } } /* Deactivate the interface in question */ a2->local_flags &= ~PFALTQ_FLAG_IF_REMOVED; if ((ifp1 = ifunit(a2->ifname)) == NULL || (remove && ifp1 == ifp)) { a2->local_flags |= PFALTQ_FLAG_IF_REMOVED; } else { error = altq_add(a2); if (ticket != V_ticket_altqs_inactive) error = EBUSY; if (error) { free(a2, M_PFALTQ); break; } } TAILQ_INSERT_TAIL(V_pf_altqs_inactive, a2, entries); } if (error != 0) pf_rollback_altq(ticket); else pf_commit_altq(ticket); } #endif /* ALTQ */ static int pf_begin_rules(u_int32_t *ticket, int rs_num, const char *anchor) { struct pf_ruleset *rs; struct pf_rule *rule; PF_RULES_WASSERT(); if (rs_num < 0 || rs_num >= PF_RULESET_MAX) return (EINVAL); rs = pf_find_or_create_ruleset(anchor); if (rs == NULL) return (EINVAL); while ((rule = TAILQ_FIRST(rs->rules[rs_num].inactive.ptr)) != NULL) { pf_unlink_rule(rs->rules[rs_num].inactive.ptr, rule); rs->rules[rs_num].inactive.rcount--; } *ticket = ++rs->rules[rs_num].inactive.ticket; rs->rules[rs_num].inactive.open = 1; return (0); } static int pf_rollback_rules(u_int32_t ticket, int rs_num, char *anchor) { struct pf_ruleset *rs; struct pf_rule *rule; PF_RULES_WASSERT(); if (rs_num < 0 || rs_num >= PF_RULESET_MAX) return (EINVAL); rs = pf_find_ruleset(anchor); if (rs == NULL || !rs->rules[rs_num].inactive.open || rs->rules[rs_num].inactive.ticket != ticket) return (0); while ((rule = TAILQ_FIRST(rs->rules[rs_num].inactive.ptr)) != NULL) { pf_unlink_rule(rs->rules[rs_num].inactive.ptr, rule); rs->rules[rs_num].inactive.rcount--; } rs->rules[rs_num].inactive.open = 0; return (0); } #define PF_MD5_UPD(st, elm) \ MD5Update(ctx, (u_int8_t *) &(st)->elm, sizeof((st)->elm)) #define PF_MD5_UPD_STR(st, elm) \ MD5Update(ctx, (u_int8_t *) (st)->elm, strlen((st)->elm)) #define PF_MD5_UPD_HTONL(st, elm, stor) do { \ (stor) = htonl((st)->elm); \ MD5Update(ctx, (u_int8_t *) &(stor), sizeof(u_int32_t));\ } while (0) #define PF_MD5_UPD_HTONS(st, elm, stor) do { \ (stor) = htons((st)->elm); \ MD5Update(ctx, (u_int8_t *) &(stor), sizeof(u_int16_t));\ } while (0) static void pf_hash_rule_addr(MD5_CTX *ctx, struct pf_rule_addr *pfr) { PF_MD5_UPD(pfr, addr.type); switch (pfr->addr.type) { case PF_ADDR_DYNIFTL: PF_MD5_UPD(pfr, addr.v.ifname); PF_MD5_UPD(pfr, addr.iflags); break; case PF_ADDR_TABLE: PF_MD5_UPD(pfr, addr.v.tblname); break; case PF_ADDR_ADDRMASK: /* XXX ignore af? */ PF_MD5_UPD(pfr, addr.v.a.addr.addr32); PF_MD5_UPD(pfr, addr.v.a.mask.addr32); break; } PF_MD5_UPD(pfr, port[0]); PF_MD5_UPD(pfr, port[1]); PF_MD5_UPD(pfr, neg); PF_MD5_UPD(pfr, port_op); } static void pf_hash_rule(MD5_CTX *ctx, struct pf_rule *rule) { u_int16_t x; u_int32_t y; pf_hash_rule_addr(ctx, &rule->src); pf_hash_rule_addr(ctx, &rule->dst); PF_MD5_UPD_STR(rule, label); PF_MD5_UPD_STR(rule, ifname); PF_MD5_UPD_STR(rule, match_tagname); PF_MD5_UPD_HTONS(rule, match_tag, x); /* dup? */ PF_MD5_UPD_HTONL(rule, os_fingerprint, y); PF_MD5_UPD_HTONL(rule, prob, y); PF_MD5_UPD_HTONL(rule, uid.uid[0], y); PF_MD5_UPD_HTONL(rule, uid.uid[1], y); PF_MD5_UPD(rule, uid.op); PF_MD5_UPD_HTONL(rule, gid.gid[0], y); PF_MD5_UPD_HTONL(rule, gid.gid[1], y); PF_MD5_UPD(rule, gid.op); PF_MD5_UPD_HTONL(rule, rule_flag, y); PF_MD5_UPD(rule, action); PF_MD5_UPD(rule, direction); PF_MD5_UPD(rule, af); PF_MD5_UPD(rule, quick); PF_MD5_UPD(rule, ifnot); PF_MD5_UPD(rule, match_tag_not); PF_MD5_UPD(rule, natpass); PF_MD5_UPD(rule, keep_state); PF_MD5_UPD(rule, proto); PF_MD5_UPD(rule, type); PF_MD5_UPD(rule, code); PF_MD5_UPD(rule, flags); PF_MD5_UPD(rule, flagset); PF_MD5_UPD(rule, allow_opts); PF_MD5_UPD(rule, rt); PF_MD5_UPD(rule, tos); } static int pf_commit_rules(u_int32_t ticket, int rs_num, char *anchor) { struct pf_ruleset *rs; struct pf_rule *rule, **old_array; struct pf_rulequeue *old_rules; int error; u_int32_t old_rcount; PF_RULES_WASSERT(); if (rs_num < 0 || rs_num >= PF_RULESET_MAX) return (EINVAL); rs = pf_find_ruleset(anchor); if (rs == NULL || !rs->rules[rs_num].inactive.open || ticket != rs->rules[rs_num].inactive.ticket) return (EBUSY); /* Calculate checksum for the main ruleset */ if (rs == &pf_main_ruleset) { error = pf_setup_pfsync_matching(rs); if (error != 0) return (error); } /* Swap rules, keep the old. */ old_rules = rs->rules[rs_num].active.ptr; old_rcount = rs->rules[rs_num].active.rcount; old_array = rs->rules[rs_num].active.ptr_array; rs->rules[rs_num].active.ptr = rs->rules[rs_num].inactive.ptr; rs->rules[rs_num].active.ptr_array = rs->rules[rs_num].inactive.ptr_array; rs->rules[rs_num].active.rcount = rs->rules[rs_num].inactive.rcount; rs->rules[rs_num].inactive.ptr = old_rules; rs->rules[rs_num].inactive.ptr_array = old_array; rs->rules[rs_num].inactive.rcount = old_rcount; rs->rules[rs_num].active.ticket = rs->rules[rs_num].inactive.ticket; pf_calc_skip_steps(rs->rules[rs_num].active.ptr); /* Purge the old rule list. */ while ((rule = TAILQ_FIRST(old_rules)) != NULL) pf_unlink_rule(old_rules, rule); if (rs->rules[rs_num].inactive.ptr_array) free(rs->rules[rs_num].inactive.ptr_array, M_TEMP); rs->rules[rs_num].inactive.ptr_array = NULL; rs->rules[rs_num].inactive.rcount = 0; rs->rules[rs_num].inactive.open = 0; pf_remove_if_empty_ruleset(rs); return (0); } static int pf_setup_pfsync_matching(struct pf_ruleset *rs) { MD5_CTX ctx; struct pf_rule *rule; int rs_cnt; u_int8_t digest[PF_MD5_DIGEST_LENGTH]; MD5Init(&ctx); for (rs_cnt = 0; rs_cnt < PF_RULESET_MAX; rs_cnt++) { /* XXX PF_RULESET_SCRUB as well? */ if (rs_cnt == PF_RULESET_SCRUB) continue; if (rs->rules[rs_cnt].inactive.ptr_array) free(rs->rules[rs_cnt].inactive.ptr_array, M_TEMP); rs->rules[rs_cnt].inactive.ptr_array = NULL; if (rs->rules[rs_cnt].inactive.rcount) { rs->rules[rs_cnt].inactive.ptr_array = malloc(sizeof(caddr_t) * rs->rules[rs_cnt].inactive.rcount, M_TEMP, M_NOWAIT); if (!rs->rules[rs_cnt].inactive.ptr_array) return (ENOMEM); } TAILQ_FOREACH(rule, rs->rules[rs_cnt].inactive.ptr, entries) { pf_hash_rule(&ctx, rule); (rs->rules[rs_cnt].inactive.ptr_array)[rule->nr] = rule; } } MD5Final(digest, &ctx); memcpy(V_pf_status.pf_chksum, digest, sizeof(V_pf_status.pf_chksum)); return (0); } static int pf_addr_setup(struct pf_ruleset *ruleset, struct pf_addr_wrap *addr, sa_family_t af) { int error = 0; switch (addr->type) { case PF_ADDR_TABLE: addr->p.tbl = pfr_attach_table(ruleset, addr->v.tblname); if (addr->p.tbl == NULL) error = ENOMEM; break; case PF_ADDR_DYNIFTL: error = pfi_dynaddr_setup(addr, af); break; } return (error); } static void pf_addr_copyout(struct pf_addr_wrap *addr) { switch (addr->type) { case PF_ADDR_DYNIFTL: pfi_dynaddr_copyout(addr); break; case PF_ADDR_TABLE: pf_tbladdr_copyout(addr); break; } } static int pfioctl(struct cdev *dev, u_long cmd, caddr_t addr, int flags, struct thread *td) { int error = 0; /* XXX keep in sync with switch() below */ if (securelevel_gt(td->td_ucred, 2)) switch (cmd) { case DIOCGETRULES: case DIOCGETRULE: case DIOCGETADDRS: case DIOCGETADDR: case DIOCGETSTATE: case DIOCSETSTATUSIF: case DIOCGETSTATUS: case DIOCCLRSTATUS: case DIOCNATLOOK: case DIOCSETDEBUG: case DIOCGETSTATES: case DIOCGETTIMEOUT: case DIOCCLRRULECTRS: case DIOCGETLIMIT: case DIOCGETALTQS: case DIOCGETALTQ: case DIOCGETQSTATS: case DIOCGETRULESETS: case DIOCGETRULESET: case DIOCRGETTABLES: case DIOCRGETTSTATS: case DIOCRCLRTSTATS: case DIOCRCLRADDRS: case DIOCRADDADDRS: case DIOCRDELADDRS: case DIOCRSETADDRS: case DIOCRGETADDRS: case DIOCRGETASTATS: case DIOCRCLRASTATS: case DIOCRTSTADDRS: case DIOCOSFPGET: case DIOCGETSRCNODES: case DIOCCLRSRCNODES: case DIOCIGETIFACES: case DIOCGIFSPEED: case DIOCSETIFFLAG: case DIOCCLRIFFLAG: break; case DIOCRCLRTABLES: case DIOCRADDTABLES: case DIOCRDELTABLES: case DIOCRSETTFLAGS: if (((struct pfioc_table *)addr)->pfrio_flags & PFR_FLAG_DUMMY) break; /* dummy operation ok */ return (EPERM); default: return (EPERM); } if (!(flags & FWRITE)) switch (cmd) { case DIOCGETRULES: case DIOCGETADDRS: case DIOCGETADDR: case DIOCGETSTATE: case DIOCGETSTATUS: case DIOCGETSTATES: case DIOCGETTIMEOUT: case DIOCGETLIMIT: case DIOCGETALTQS: case DIOCGETALTQ: case DIOCGETQSTATS: case DIOCGETRULESETS: case DIOCGETRULESET: case DIOCNATLOOK: case DIOCRGETTABLES: case DIOCRGETTSTATS: case DIOCRGETADDRS: case DIOCRGETASTATS: case DIOCRTSTADDRS: case DIOCOSFPGET: case DIOCGETSRCNODES: case DIOCIGETIFACES: case DIOCGIFSPEED: break; case DIOCRCLRTABLES: case DIOCRADDTABLES: case DIOCRDELTABLES: case DIOCRCLRTSTATS: case DIOCRCLRADDRS: case DIOCRADDADDRS: case DIOCRDELADDRS: case DIOCRSETADDRS: case DIOCRSETTFLAGS: if (((struct pfioc_table *)addr)->pfrio_flags & PFR_FLAG_DUMMY) { flags |= FWRITE; /* need write lock for dummy */ break; /* dummy operation ok */ } return (EACCES); case DIOCGETRULE: if (((struct pfioc_rule *)addr)->action == PF_GET_CLR_CNTR) return (EACCES); break; default: return (EACCES); } CURVNET_SET(TD_TO_VNET(td)); switch (cmd) { case DIOCSTART: sx_xlock(&pf_ioctl_lock); if (V_pf_status.running) error = EEXIST; else { int cpu; error = hook_pf(); if (error) { DPFPRINTF(PF_DEBUG_MISC, ("pf: pfil registration failed\n")); break; } V_pf_status.running = 1; V_pf_status.since = time_second; CPU_FOREACH(cpu) V_pf_stateid[cpu] = time_second; DPFPRINTF(PF_DEBUG_MISC, ("pf: started\n")); } break; case DIOCSTOP: sx_xlock(&pf_ioctl_lock); if (!V_pf_status.running) error = ENOENT; else { V_pf_status.running = 0; error = dehook_pf(); if (error) { V_pf_status.running = 1; DPFPRINTF(PF_DEBUG_MISC, ("pf: pfil unregistration failed\n")); } V_pf_status.since = time_second; DPFPRINTF(PF_DEBUG_MISC, ("pf: stopped\n")); } break; case DIOCADDRULE: { struct pfioc_rule *pr = (struct pfioc_rule *)addr; struct pf_ruleset *ruleset; struct pf_rule *rule, *tail; struct pf_pooladdr *pa; struct pfi_kif *kif = NULL; int rs_num; if (pr->rule.return_icmp >> 8 > ICMP_MAXTYPE) { error = EINVAL; break; } #ifndef INET if (pr->rule.af == AF_INET) { error = EAFNOSUPPORT; break; } #endif /* INET */ #ifndef INET6 if (pr->rule.af == AF_INET6) { error = EAFNOSUPPORT; break; } #endif /* INET6 */ rule = malloc(sizeof(*rule), M_PFRULE, M_WAITOK); bcopy(&pr->rule, rule, sizeof(struct pf_rule)); if (rule->ifname[0]) kif = malloc(sizeof(*kif), PFI_MTYPE, M_WAITOK); rule->states_cur = counter_u64_alloc(M_WAITOK); rule->states_tot = counter_u64_alloc(M_WAITOK); rule->src_nodes = counter_u64_alloc(M_WAITOK); rule->cuid = td->td_ucred->cr_ruid; rule->cpid = td->td_proc ? td->td_proc->p_pid : 0; TAILQ_INIT(&rule->rpool.list); #define ERROUT(x) { error = (x); goto DIOCADDRULE_error; } PF_RULES_WLOCK(); pr->anchor[sizeof(pr->anchor) - 1] = 0; ruleset = pf_find_ruleset(pr->anchor); if (ruleset == NULL) ERROUT(EINVAL); rs_num = pf_get_ruleset_number(pr->rule.action); if (rs_num >= PF_RULESET_MAX) ERROUT(EINVAL); if (pr->ticket != ruleset->rules[rs_num].inactive.ticket) { DPFPRINTF(PF_DEBUG_MISC, ("ticket: %d != [%d]%d\n", pr->ticket, rs_num, ruleset->rules[rs_num].inactive.ticket)); ERROUT(EBUSY); } if (pr->pool_ticket != V_ticket_pabuf) { DPFPRINTF(PF_DEBUG_MISC, ("pool_ticket: %d != %d\n", pr->pool_ticket, V_ticket_pabuf)); ERROUT(EBUSY); } tail = TAILQ_LAST(ruleset->rules[rs_num].inactive.ptr, pf_rulequeue); if (tail) rule->nr = tail->nr + 1; else rule->nr = 0; if (rule->ifname[0]) { rule->kif = pfi_kif_attach(kif, rule->ifname); pfi_kif_ref(rule->kif); } else rule->kif = NULL; if (rule->rtableid > 0 && rule->rtableid >= rt_numfibs) error = EBUSY; #ifdef ALTQ /* set queue IDs */ if (rule->qname[0] != 0) { if ((rule->qid = pf_qname2qid(rule->qname)) == 0) error = EBUSY; else if (rule->pqname[0] != 0) { if ((rule->pqid = pf_qname2qid(rule->pqname)) == 0) error = EBUSY; } else rule->pqid = rule->qid; } #endif if (rule->tagname[0]) if ((rule->tag = pf_tagname2tag(rule->tagname)) == 0) error = EBUSY; if (rule->match_tagname[0]) if ((rule->match_tag = pf_tagname2tag(rule->match_tagname)) == 0) error = EBUSY; if (rule->rt && !rule->direction) error = EINVAL; if (!rule->log) rule->logif = 0; if (rule->logif >= PFLOGIFS_MAX) error = EINVAL; if (pf_addr_setup(ruleset, &rule->src.addr, rule->af)) error = ENOMEM; if (pf_addr_setup(ruleset, &rule->dst.addr, rule->af)) error = ENOMEM; if (pf_anchor_setup(rule, ruleset, pr->anchor_call)) error = EINVAL; if (rule->scrub_flags & PFSTATE_SETPRIO && (rule->set_prio[0] > PF_PRIO_MAX || rule->set_prio[1] > PF_PRIO_MAX)) error = EINVAL; TAILQ_FOREACH(pa, &V_pf_pabuf, entries) if (pa->addr.type == PF_ADDR_TABLE) { pa->addr.p.tbl = pfr_attach_table(ruleset, pa->addr.v.tblname); if (pa->addr.p.tbl == NULL) error = ENOMEM; } rule->overload_tbl = NULL; if (rule->overload_tblname[0]) { if ((rule->overload_tbl = pfr_attach_table(ruleset, rule->overload_tblname)) == NULL) error = EINVAL; else rule->overload_tbl->pfrkt_flags |= PFR_TFLAG_ACTIVE; } pf_mv_pool(&V_pf_pabuf, &rule->rpool.list); if (((((rule->action == PF_NAT) || (rule->action == PF_RDR) || (rule->action == PF_BINAT)) && rule->anchor == NULL) || - (rule->rt > PF_FASTROUTE)) && + (rule->rt > PF_NOPFROUTE)) && (TAILQ_FIRST(&rule->rpool.list) == NULL)) error = EINVAL; if (error) { pf_free_rule(rule); PF_RULES_WUNLOCK(); break; } rule->rpool.cur = TAILQ_FIRST(&rule->rpool.list); rule->evaluations = rule->packets[0] = rule->packets[1] = rule->bytes[0] = rule->bytes[1] = 0; TAILQ_INSERT_TAIL(ruleset->rules[rs_num].inactive.ptr, rule, entries); ruleset->rules[rs_num].inactive.rcount++; PF_RULES_WUNLOCK(); break; #undef ERROUT DIOCADDRULE_error: PF_RULES_WUNLOCK(); counter_u64_free(rule->states_cur); counter_u64_free(rule->states_tot); counter_u64_free(rule->src_nodes); free(rule, M_PFRULE); if (kif) free(kif, PFI_MTYPE); break; } case DIOCGETRULES: { struct pfioc_rule *pr = (struct pfioc_rule *)addr; struct pf_ruleset *ruleset; struct pf_rule *tail; int rs_num; PF_RULES_WLOCK(); pr->anchor[sizeof(pr->anchor) - 1] = 0; ruleset = pf_find_ruleset(pr->anchor); if (ruleset == NULL) { PF_RULES_WUNLOCK(); error = EINVAL; break; } rs_num = pf_get_ruleset_number(pr->rule.action); if (rs_num >= PF_RULESET_MAX) { PF_RULES_WUNLOCK(); error = EINVAL; break; } tail = TAILQ_LAST(ruleset->rules[rs_num].active.ptr, pf_rulequeue); if (tail) pr->nr = tail->nr + 1; else pr->nr = 0; pr->ticket = ruleset->rules[rs_num].active.ticket; PF_RULES_WUNLOCK(); break; } case DIOCGETRULE: { struct pfioc_rule *pr = (struct pfioc_rule *)addr; struct pf_ruleset *ruleset; struct pf_rule *rule; int rs_num, i; PF_RULES_WLOCK(); pr->anchor[sizeof(pr->anchor) - 1] = 0; ruleset = pf_find_ruleset(pr->anchor); if (ruleset == NULL) { PF_RULES_WUNLOCK(); error = EINVAL; break; } rs_num = pf_get_ruleset_number(pr->rule.action); if (rs_num >= PF_RULESET_MAX) { PF_RULES_WUNLOCK(); error = EINVAL; break; } if (pr->ticket != ruleset->rules[rs_num].active.ticket) { PF_RULES_WUNLOCK(); error = EBUSY; break; } rule = TAILQ_FIRST(ruleset->rules[rs_num].active.ptr); while ((rule != NULL) && (rule->nr != pr->nr)) rule = TAILQ_NEXT(rule, entries); if (rule == NULL) { PF_RULES_WUNLOCK(); error = EBUSY; break; } bcopy(rule, &pr->rule, sizeof(struct pf_rule)); pr->rule.u_states_cur = counter_u64_fetch(rule->states_cur); pr->rule.u_states_tot = counter_u64_fetch(rule->states_tot); pr->rule.u_src_nodes = counter_u64_fetch(rule->src_nodes); if (pf_anchor_copyout(ruleset, rule, pr)) { PF_RULES_WUNLOCK(); error = EBUSY; break; } pf_addr_copyout(&pr->rule.src.addr); pf_addr_copyout(&pr->rule.dst.addr); for (i = 0; i < PF_SKIP_COUNT; ++i) if (rule->skip[i].ptr == NULL) pr->rule.skip[i].nr = -1; else pr->rule.skip[i].nr = rule->skip[i].ptr->nr; if (pr->action == PF_GET_CLR_CNTR) { rule->evaluations = 0; rule->packets[0] = rule->packets[1] = 0; rule->bytes[0] = rule->bytes[1] = 0; counter_u64_zero(rule->states_tot); } PF_RULES_WUNLOCK(); break; } case DIOCCHANGERULE: { struct pfioc_rule *pcr = (struct pfioc_rule *)addr; struct pf_ruleset *ruleset; struct pf_rule *oldrule = NULL, *newrule = NULL; struct pfi_kif *kif = NULL; struct pf_pooladdr *pa; u_int32_t nr = 0; int rs_num; if (pcr->action < PF_CHANGE_ADD_HEAD || pcr->action > PF_CHANGE_GET_TICKET) { error = EINVAL; break; } if (pcr->rule.return_icmp >> 8 > ICMP_MAXTYPE) { error = EINVAL; break; } if (pcr->action != PF_CHANGE_REMOVE) { #ifndef INET if (pcr->rule.af == AF_INET) { error = EAFNOSUPPORT; break; } #endif /* INET */ #ifndef INET6 if (pcr->rule.af == AF_INET6) { error = EAFNOSUPPORT; break; } #endif /* INET6 */ newrule = malloc(sizeof(*newrule), M_PFRULE, M_WAITOK); bcopy(&pcr->rule, newrule, sizeof(struct pf_rule)); if (newrule->ifname[0]) kif = malloc(sizeof(*kif), PFI_MTYPE, M_WAITOK); newrule->states_cur = counter_u64_alloc(M_WAITOK); newrule->states_tot = counter_u64_alloc(M_WAITOK); newrule->src_nodes = counter_u64_alloc(M_WAITOK); newrule->cuid = td->td_ucred->cr_ruid; newrule->cpid = td->td_proc ? td->td_proc->p_pid : 0; TAILQ_INIT(&newrule->rpool.list); } #define ERROUT(x) { error = (x); goto DIOCCHANGERULE_error; } PF_RULES_WLOCK(); if (!(pcr->action == PF_CHANGE_REMOVE || pcr->action == PF_CHANGE_GET_TICKET) && pcr->pool_ticket != V_ticket_pabuf) ERROUT(EBUSY); ruleset = pf_find_ruleset(pcr->anchor); if (ruleset == NULL) ERROUT(EINVAL); rs_num = pf_get_ruleset_number(pcr->rule.action); if (rs_num >= PF_RULESET_MAX) ERROUT(EINVAL); if (pcr->action == PF_CHANGE_GET_TICKET) { pcr->ticket = ++ruleset->rules[rs_num].active.ticket; ERROUT(0); } else if (pcr->ticket != ruleset->rules[rs_num].active.ticket) ERROUT(EINVAL); if (pcr->action != PF_CHANGE_REMOVE) { if (newrule->ifname[0]) { newrule->kif = pfi_kif_attach(kif, newrule->ifname); pfi_kif_ref(newrule->kif); } else newrule->kif = NULL; if (newrule->rtableid > 0 && newrule->rtableid >= rt_numfibs) error = EBUSY; #ifdef ALTQ /* set queue IDs */ if (newrule->qname[0] != 0) { if ((newrule->qid = pf_qname2qid(newrule->qname)) == 0) error = EBUSY; else if (newrule->pqname[0] != 0) { if ((newrule->pqid = pf_qname2qid(newrule->pqname)) == 0) error = EBUSY; } else newrule->pqid = newrule->qid; } #endif /* ALTQ */ if (newrule->tagname[0]) if ((newrule->tag = pf_tagname2tag(newrule->tagname)) == 0) error = EBUSY; if (newrule->match_tagname[0]) if ((newrule->match_tag = pf_tagname2tag( newrule->match_tagname)) == 0) error = EBUSY; if (newrule->rt && !newrule->direction) error = EINVAL; if (!newrule->log) newrule->logif = 0; if (newrule->logif >= PFLOGIFS_MAX) error = EINVAL; if (pf_addr_setup(ruleset, &newrule->src.addr, newrule->af)) error = ENOMEM; if (pf_addr_setup(ruleset, &newrule->dst.addr, newrule->af)) error = ENOMEM; if (pf_anchor_setup(newrule, ruleset, pcr->anchor_call)) error = EINVAL; TAILQ_FOREACH(pa, &V_pf_pabuf, entries) if (pa->addr.type == PF_ADDR_TABLE) { pa->addr.p.tbl = pfr_attach_table(ruleset, pa->addr.v.tblname); if (pa->addr.p.tbl == NULL) error = ENOMEM; } newrule->overload_tbl = NULL; if (newrule->overload_tblname[0]) { if ((newrule->overload_tbl = pfr_attach_table( ruleset, newrule->overload_tblname)) == NULL) error = EINVAL; else newrule->overload_tbl->pfrkt_flags |= PFR_TFLAG_ACTIVE; } pf_mv_pool(&V_pf_pabuf, &newrule->rpool.list); if (((((newrule->action == PF_NAT) || (newrule->action == PF_RDR) || (newrule->action == PF_BINAT) || - (newrule->rt > PF_FASTROUTE)) && + (newrule->rt > PF_NOPFROUTE)) && !newrule->anchor)) && (TAILQ_FIRST(&newrule->rpool.list) == NULL)) error = EINVAL; if (error) { pf_free_rule(newrule); PF_RULES_WUNLOCK(); break; } newrule->rpool.cur = TAILQ_FIRST(&newrule->rpool.list); newrule->evaluations = 0; newrule->packets[0] = newrule->packets[1] = 0; newrule->bytes[0] = newrule->bytes[1] = 0; } pf_empty_pool(&V_pf_pabuf); if (pcr->action == PF_CHANGE_ADD_HEAD) oldrule = TAILQ_FIRST( ruleset->rules[rs_num].active.ptr); else if (pcr->action == PF_CHANGE_ADD_TAIL) oldrule = TAILQ_LAST( ruleset->rules[rs_num].active.ptr, pf_rulequeue); else { oldrule = TAILQ_FIRST( ruleset->rules[rs_num].active.ptr); while ((oldrule != NULL) && (oldrule->nr != pcr->nr)) oldrule = TAILQ_NEXT(oldrule, entries); if (oldrule == NULL) { if (newrule != NULL) pf_free_rule(newrule); PF_RULES_WUNLOCK(); error = EINVAL; break; } } if (pcr->action == PF_CHANGE_REMOVE) { pf_unlink_rule(ruleset->rules[rs_num].active.ptr, oldrule); ruleset->rules[rs_num].active.rcount--; } else { if (oldrule == NULL) TAILQ_INSERT_TAIL( ruleset->rules[rs_num].active.ptr, newrule, entries); else if (pcr->action == PF_CHANGE_ADD_HEAD || pcr->action == PF_CHANGE_ADD_BEFORE) TAILQ_INSERT_BEFORE(oldrule, newrule, entries); else TAILQ_INSERT_AFTER( ruleset->rules[rs_num].active.ptr, oldrule, newrule, entries); ruleset->rules[rs_num].active.rcount++; } nr = 0; TAILQ_FOREACH(oldrule, ruleset->rules[rs_num].active.ptr, entries) oldrule->nr = nr++; ruleset->rules[rs_num].active.ticket++; pf_calc_skip_steps(ruleset->rules[rs_num].active.ptr); pf_remove_if_empty_ruleset(ruleset); PF_RULES_WUNLOCK(); break; #undef ERROUT DIOCCHANGERULE_error: PF_RULES_WUNLOCK(); if (newrule != NULL) { counter_u64_free(newrule->states_cur); counter_u64_free(newrule->states_tot); counter_u64_free(newrule->src_nodes); free(newrule, M_PFRULE); } if (kif != NULL) free(kif, PFI_MTYPE); break; } case DIOCCLRSTATES: { struct pf_state *s; struct pfioc_state_kill *psk = (struct pfioc_state_kill *)addr; u_int i, killed = 0; for (i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; relock_DIOCCLRSTATES: PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) if (!psk->psk_ifname[0] || !strcmp(psk->psk_ifname, s->kif->pfik_name)) { /* * Don't send out individual * delete messages. */ s->state_flags |= PFSTATE_NOSYNC; pf_unlink_state(s, PF_ENTER_LOCKED); killed++; goto relock_DIOCCLRSTATES; } PF_HASHROW_UNLOCK(ih); } psk->psk_killed = killed; if (pfsync_clear_states_ptr != NULL) pfsync_clear_states_ptr(V_pf_status.hostid, psk->psk_ifname); break; } case DIOCKILLSTATES: { struct pf_state *s; struct pf_state_key *sk; struct pf_addr *srcaddr, *dstaddr; u_int16_t srcport, dstport; struct pfioc_state_kill *psk = (struct pfioc_state_kill *)addr; u_int i, killed = 0; if (psk->psk_pfcmp.id) { if (psk->psk_pfcmp.creatorid == 0) psk->psk_pfcmp.creatorid = V_pf_status.hostid; if ((s = pf_find_state_byid(psk->psk_pfcmp.id, psk->psk_pfcmp.creatorid))) { pf_unlink_state(s, PF_ENTER_LOCKED); psk->psk_killed = 1; } break; } for (i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; relock_DIOCKILLSTATES: PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { sk = s->key[PF_SK_WIRE]; if (s->direction == PF_OUT) { srcaddr = &sk->addr[1]; dstaddr = &sk->addr[0]; srcport = sk->port[1]; dstport = sk->port[0]; } else { srcaddr = &sk->addr[0]; dstaddr = &sk->addr[1]; srcport = sk->port[0]; dstport = sk->port[1]; } if ((!psk->psk_af || sk->af == psk->psk_af) && (!psk->psk_proto || psk->psk_proto == sk->proto) && PF_MATCHA(psk->psk_src.neg, &psk->psk_src.addr.v.a.addr, &psk->psk_src.addr.v.a.mask, srcaddr, sk->af) && PF_MATCHA(psk->psk_dst.neg, &psk->psk_dst.addr.v.a.addr, &psk->psk_dst.addr.v.a.mask, dstaddr, sk->af) && (psk->psk_src.port_op == 0 || pf_match_port(psk->psk_src.port_op, psk->psk_src.port[0], psk->psk_src.port[1], srcport)) && (psk->psk_dst.port_op == 0 || pf_match_port(psk->psk_dst.port_op, psk->psk_dst.port[0], psk->psk_dst.port[1], dstport)) && (!psk->psk_label[0] || (s->rule.ptr->label[0] && !strcmp(psk->psk_label, s->rule.ptr->label))) && (!psk->psk_ifname[0] || !strcmp(psk->psk_ifname, s->kif->pfik_name))) { pf_unlink_state(s, PF_ENTER_LOCKED); killed++; goto relock_DIOCKILLSTATES; } } PF_HASHROW_UNLOCK(ih); } psk->psk_killed = killed; break; } case DIOCADDSTATE: { struct pfioc_state *ps = (struct pfioc_state *)addr; struct pfsync_state *sp = &ps->state; if (sp->timeout >= PFTM_MAX) { error = EINVAL; break; } if (pfsync_state_import_ptr != NULL) { PF_RULES_RLOCK(); error = pfsync_state_import_ptr(sp, PFSYNC_SI_IOCTL); PF_RULES_RUNLOCK(); } else error = EOPNOTSUPP; break; } case DIOCGETSTATE: { struct pfioc_state *ps = (struct pfioc_state *)addr; struct pf_state *s; s = pf_find_state_byid(ps->state.id, ps->state.creatorid); if (s == NULL) { error = ENOENT; break; } pfsync_state_export(&ps->state, s); PF_STATE_UNLOCK(s); break; } case DIOCGETSTATES: { struct pfioc_states *ps = (struct pfioc_states *)addr; struct pf_state *s; struct pfsync_state *pstore, *p; int i, nr; if (ps->ps_len == 0) { nr = uma_zone_get_cur(V_pf_state_z); ps->ps_len = sizeof(struct pfsync_state) * nr; break; } p = pstore = malloc(ps->ps_len, M_TEMP, M_WAITOK); nr = 0; for (i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { if (s->timeout == PFTM_UNLINKED) continue; if ((nr+1) * sizeof(*p) > ps->ps_len) { PF_HASHROW_UNLOCK(ih); goto DIOCGETSTATES_full; } pfsync_state_export(p, s); p++; nr++; } PF_HASHROW_UNLOCK(ih); } DIOCGETSTATES_full: error = copyout(pstore, ps->ps_states, sizeof(struct pfsync_state) * nr); if (error) { free(pstore, M_TEMP); break; } ps->ps_len = sizeof(struct pfsync_state) * nr; free(pstore, M_TEMP); break; } case DIOCGETSTATUS: { struct pf_status *s = (struct pf_status *)addr; PF_RULES_RLOCK(); s->running = V_pf_status.running; s->since = V_pf_status.since; s->debug = V_pf_status.debug; s->hostid = V_pf_status.hostid; s->states = V_pf_status.states; s->src_nodes = V_pf_status.src_nodes; for (int i = 0; i < PFRES_MAX; i++) s->counters[i] = counter_u64_fetch(V_pf_status.counters[i]); for (int i = 0; i < LCNT_MAX; i++) s->lcounters[i] = counter_u64_fetch(V_pf_status.lcounters[i]); for (int i = 0; i < FCNT_MAX; i++) s->fcounters[i] = counter_u64_fetch(V_pf_status.fcounters[i]); for (int i = 0; i < SCNT_MAX; i++) s->scounters[i] = counter_u64_fetch(V_pf_status.scounters[i]); bcopy(V_pf_status.ifname, s->ifname, IFNAMSIZ); bcopy(V_pf_status.pf_chksum, s->pf_chksum, PF_MD5_DIGEST_LENGTH); pfi_update_status(s->ifname, s); PF_RULES_RUNLOCK(); break; } case DIOCSETSTATUSIF: { struct pfioc_if *pi = (struct pfioc_if *)addr; if (pi->ifname[0] == 0) { bzero(V_pf_status.ifname, IFNAMSIZ); break; } PF_RULES_WLOCK(); strlcpy(V_pf_status.ifname, pi->ifname, IFNAMSIZ); PF_RULES_WUNLOCK(); break; } case DIOCCLRSTATUS: { PF_RULES_WLOCK(); for (int i = 0; i < PFRES_MAX; i++) counter_u64_zero(V_pf_status.counters[i]); for (int i = 0; i < FCNT_MAX; i++) counter_u64_zero(V_pf_status.fcounters[i]); for (int i = 0; i < SCNT_MAX; i++) counter_u64_zero(V_pf_status.scounters[i]); V_pf_status.since = time_second; if (*V_pf_status.ifname) pfi_update_status(V_pf_status.ifname, NULL); PF_RULES_WUNLOCK(); break; } case DIOCNATLOOK: { struct pfioc_natlook *pnl = (struct pfioc_natlook *)addr; struct pf_state_key *sk; struct pf_state *state; struct pf_state_key_cmp key; int m = 0, direction = pnl->direction; int sidx, didx; /* NATLOOK src and dst are reversed, so reverse sidx/didx */ sidx = (direction == PF_IN) ? 1 : 0; didx = (direction == PF_IN) ? 0 : 1; if (!pnl->proto || PF_AZERO(&pnl->saddr, pnl->af) || PF_AZERO(&pnl->daddr, pnl->af) || ((pnl->proto == IPPROTO_TCP || pnl->proto == IPPROTO_UDP) && (!pnl->dport || !pnl->sport))) error = EINVAL; else { bzero(&key, sizeof(key)); key.af = pnl->af; key.proto = pnl->proto; PF_ACPY(&key.addr[sidx], &pnl->saddr, pnl->af); key.port[sidx] = pnl->sport; PF_ACPY(&key.addr[didx], &pnl->daddr, pnl->af); key.port[didx] = pnl->dport; state = pf_find_state_all(&key, direction, &m); if (m > 1) error = E2BIG; /* more than one state */ else if (state != NULL) { /* XXXGL: not locked read */ sk = state->key[sidx]; PF_ACPY(&pnl->rsaddr, &sk->addr[sidx], sk->af); pnl->rsport = sk->port[sidx]; PF_ACPY(&pnl->rdaddr, &sk->addr[didx], sk->af); pnl->rdport = sk->port[didx]; } else error = ENOENT; } break; } case DIOCSETTIMEOUT: { struct pfioc_tm *pt = (struct pfioc_tm *)addr; int old; if (pt->timeout < 0 || pt->timeout >= PFTM_MAX || pt->seconds < 0) { error = EINVAL; break; } PF_RULES_WLOCK(); old = V_pf_default_rule.timeout[pt->timeout]; if (pt->timeout == PFTM_INTERVAL && pt->seconds == 0) pt->seconds = 1; V_pf_default_rule.timeout[pt->timeout] = pt->seconds; if (pt->timeout == PFTM_INTERVAL && pt->seconds < old) wakeup(pf_purge_thread); pt->seconds = old; PF_RULES_WUNLOCK(); break; } case DIOCGETTIMEOUT: { struct pfioc_tm *pt = (struct pfioc_tm *)addr; if (pt->timeout < 0 || pt->timeout >= PFTM_MAX) { error = EINVAL; break; } PF_RULES_RLOCK(); pt->seconds = V_pf_default_rule.timeout[pt->timeout]; PF_RULES_RUNLOCK(); break; } case DIOCGETLIMIT: { struct pfioc_limit *pl = (struct pfioc_limit *)addr; if (pl->index < 0 || pl->index >= PF_LIMIT_MAX) { error = EINVAL; break; } PF_RULES_RLOCK(); pl->limit = V_pf_limits[pl->index].limit; PF_RULES_RUNLOCK(); break; } case DIOCSETLIMIT: { struct pfioc_limit *pl = (struct pfioc_limit *)addr; int old_limit; PF_RULES_WLOCK(); if (pl->index < 0 || pl->index >= PF_LIMIT_MAX || V_pf_limits[pl->index].zone == NULL) { PF_RULES_WUNLOCK(); error = EINVAL; break; } uma_zone_set_max(V_pf_limits[pl->index].zone, pl->limit); old_limit = V_pf_limits[pl->index].limit; V_pf_limits[pl->index].limit = pl->limit; pl->limit = old_limit; PF_RULES_WUNLOCK(); break; } case DIOCSETDEBUG: { u_int32_t *level = (u_int32_t *)addr; PF_RULES_WLOCK(); V_pf_status.debug = *level; PF_RULES_WUNLOCK(); break; } case DIOCCLRRULECTRS: { /* obsoleted by DIOCGETRULE with action=PF_GET_CLR_CNTR */ struct pf_ruleset *ruleset = &pf_main_ruleset; struct pf_rule *rule; PF_RULES_WLOCK(); TAILQ_FOREACH(rule, ruleset->rules[PF_RULESET_FILTER].active.ptr, entries) { rule->evaluations = 0; rule->packets[0] = rule->packets[1] = 0; rule->bytes[0] = rule->bytes[1] = 0; } PF_RULES_WUNLOCK(); break; } case DIOCGIFSPEED: { struct pf_ifspeed *psp = (struct pf_ifspeed *)addr; struct pf_ifspeed ps; struct ifnet *ifp; if (psp->ifname[0] != 0) { /* Can we completely trust user-land? */ strlcpy(ps.ifname, psp->ifname, IFNAMSIZ); ifp = ifunit(ps.ifname); if (ifp != NULL) psp->baudrate = ifp->if_baudrate; else error = EINVAL; } else error = EINVAL; break; } #ifdef ALTQ case DIOCSTARTALTQ: { struct pf_altq *altq; PF_RULES_WLOCK(); /* enable all altq interfaces on active list */ TAILQ_FOREACH(altq, V_pf_altqs_active, entries) { if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { error = pf_enable_altq(altq); if (error != 0) break; } } if (error == 0) V_pf_altq_running = 1; PF_RULES_WUNLOCK(); DPFPRINTF(PF_DEBUG_MISC, ("altq: started\n")); break; } case DIOCSTOPALTQ: { struct pf_altq *altq; PF_RULES_WLOCK(); /* disable all altq interfaces on active list */ TAILQ_FOREACH(altq, V_pf_altqs_active, entries) { if (altq->qname[0] == 0 && (altq->local_flags & PFALTQ_FLAG_IF_REMOVED) == 0) { error = pf_disable_altq(altq); if (error != 0) break; } } if (error == 0) V_pf_altq_running = 0; PF_RULES_WUNLOCK(); DPFPRINTF(PF_DEBUG_MISC, ("altq: stopped\n")); break; } case DIOCADDALTQ: { struct pfioc_altq *pa = (struct pfioc_altq *)addr; struct pf_altq *altq, *a; struct ifnet *ifp; altq = malloc(sizeof(*altq), M_PFALTQ, M_WAITOK); bcopy(&pa->altq, altq, sizeof(struct pf_altq)); altq->local_flags = 0; PF_RULES_WLOCK(); if (pa->ticket != V_ticket_altqs_inactive) { PF_RULES_WUNLOCK(); free(altq, M_PFALTQ); error = EBUSY; break; } /* * if this is for a queue, find the discipline and * copy the necessary fields */ if (altq->qname[0] != 0) { if ((altq->qid = pf_qname2qid(altq->qname)) == 0) { PF_RULES_WUNLOCK(); error = EBUSY; free(altq, M_PFALTQ); break; } altq->altq_disc = NULL; TAILQ_FOREACH(a, V_pf_altqs_inactive, entries) { if (strncmp(a->ifname, altq->ifname, IFNAMSIZ) == 0 && a->qname[0] == 0) { altq->altq_disc = a->altq_disc; break; } } } if ((ifp = ifunit(altq->ifname)) == NULL) altq->local_flags |= PFALTQ_FLAG_IF_REMOVED; else error = altq_add(altq); if (error) { PF_RULES_WUNLOCK(); free(altq, M_PFALTQ); break; } TAILQ_INSERT_TAIL(V_pf_altqs_inactive, altq, entries); bcopy(altq, &pa->altq, sizeof(struct pf_altq)); PF_RULES_WUNLOCK(); break; } case DIOCGETALTQS: { struct pfioc_altq *pa = (struct pfioc_altq *)addr; struct pf_altq *altq; PF_RULES_RLOCK(); pa->nr = 0; TAILQ_FOREACH(altq, V_pf_altqs_active, entries) pa->nr++; pa->ticket = V_ticket_altqs_active; PF_RULES_RUNLOCK(); break; } case DIOCGETALTQ: { struct pfioc_altq *pa = (struct pfioc_altq *)addr; struct pf_altq *altq; u_int32_t nr; PF_RULES_RLOCK(); if (pa->ticket != V_ticket_altqs_active) { PF_RULES_RUNLOCK(); error = EBUSY; break; } nr = 0; altq = TAILQ_FIRST(V_pf_altqs_active); while ((altq != NULL) && (nr < pa->nr)) { altq = TAILQ_NEXT(altq, entries); nr++; } if (altq == NULL) { PF_RULES_RUNLOCK(); error = EBUSY; break; } bcopy(altq, &pa->altq, sizeof(struct pf_altq)); PF_RULES_RUNLOCK(); break; } case DIOCCHANGEALTQ: /* CHANGEALTQ not supported yet! */ error = ENODEV; break; case DIOCGETQSTATS: { struct pfioc_qstats *pq = (struct pfioc_qstats *)addr; struct pf_altq *altq; u_int32_t nr; int nbytes; PF_RULES_RLOCK(); if (pq->ticket != V_ticket_altqs_active) { PF_RULES_RUNLOCK(); error = EBUSY; break; } nbytes = pq->nbytes; nr = 0; altq = TAILQ_FIRST(V_pf_altqs_active); while ((altq != NULL) && (nr < pq->nr)) { altq = TAILQ_NEXT(altq, entries); nr++; } if (altq == NULL) { PF_RULES_RUNLOCK(); error = EBUSY; break; } if ((altq->local_flags & PFALTQ_FLAG_IF_REMOVED) != 0) { PF_RULES_RUNLOCK(); error = ENXIO; break; } PF_RULES_RUNLOCK(); error = altq_getqstats(altq, pq->buf, &nbytes); if (error == 0) { pq->scheduler = altq->scheduler; pq->nbytes = nbytes; } break; } #endif /* ALTQ */ case DIOCBEGINADDRS: { struct pfioc_pooladdr *pp = (struct pfioc_pooladdr *)addr; PF_RULES_WLOCK(); pf_empty_pool(&V_pf_pabuf); pp->ticket = ++V_ticket_pabuf; PF_RULES_WUNLOCK(); break; } case DIOCADDADDR: { struct pfioc_pooladdr *pp = (struct pfioc_pooladdr *)addr; struct pf_pooladdr *pa; struct pfi_kif *kif = NULL; #ifndef INET if (pp->af == AF_INET) { error = EAFNOSUPPORT; break; } #endif /* INET */ #ifndef INET6 if (pp->af == AF_INET6) { error = EAFNOSUPPORT; break; } #endif /* INET6 */ if (pp->addr.addr.type != PF_ADDR_ADDRMASK && pp->addr.addr.type != PF_ADDR_DYNIFTL && pp->addr.addr.type != PF_ADDR_TABLE) { error = EINVAL; break; } pa = malloc(sizeof(*pa), M_PFRULE, M_WAITOK); bcopy(&pp->addr, pa, sizeof(struct pf_pooladdr)); if (pa->ifname[0]) kif = malloc(sizeof(*kif), PFI_MTYPE, M_WAITOK); PF_RULES_WLOCK(); if (pp->ticket != V_ticket_pabuf) { PF_RULES_WUNLOCK(); if (pa->ifname[0]) free(kif, PFI_MTYPE); free(pa, M_PFRULE); error = EBUSY; break; } if (pa->ifname[0]) { pa->kif = pfi_kif_attach(kif, pa->ifname); pfi_kif_ref(pa->kif); } else pa->kif = NULL; if (pa->addr.type == PF_ADDR_DYNIFTL && ((error = pfi_dynaddr_setup(&pa->addr, pp->af)) != 0)) { if (pa->ifname[0]) pfi_kif_unref(pa->kif); PF_RULES_WUNLOCK(); free(pa, M_PFRULE); break; } TAILQ_INSERT_TAIL(&V_pf_pabuf, pa, entries); PF_RULES_WUNLOCK(); break; } case DIOCGETADDRS: { struct pfioc_pooladdr *pp = (struct pfioc_pooladdr *)addr; struct pf_pool *pool; struct pf_pooladdr *pa; PF_RULES_RLOCK(); pp->nr = 0; pool = pf_get_pool(pp->anchor, pp->ticket, pp->r_action, pp->r_num, 0, 1, 0); if (pool == NULL) { PF_RULES_RUNLOCK(); error = EBUSY; break; } TAILQ_FOREACH(pa, &pool->list, entries) pp->nr++; PF_RULES_RUNLOCK(); break; } case DIOCGETADDR: { struct pfioc_pooladdr *pp = (struct pfioc_pooladdr *)addr; struct pf_pool *pool; struct pf_pooladdr *pa; u_int32_t nr = 0; PF_RULES_RLOCK(); pool = pf_get_pool(pp->anchor, pp->ticket, pp->r_action, pp->r_num, 0, 1, 1); if (pool == NULL) { PF_RULES_RUNLOCK(); error = EBUSY; break; } pa = TAILQ_FIRST(&pool->list); while ((pa != NULL) && (nr < pp->nr)) { pa = TAILQ_NEXT(pa, entries); nr++; } if (pa == NULL) { PF_RULES_RUNLOCK(); error = EBUSY; break; } bcopy(pa, &pp->addr, sizeof(struct pf_pooladdr)); pf_addr_copyout(&pp->addr.addr); PF_RULES_RUNLOCK(); break; } case DIOCCHANGEADDR: { struct pfioc_pooladdr *pca = (struct pfioc_pooladdr *)addr; struct pf_pool *pool; struct pf_pooladdr *oldpa = NULL, *newpa = NULL; struct pf_ruleset *ruleset; struct pfi_kif *kif = NULL; if (pca->action < PF_CHANGE_ADD_HEAD || pca->action > PF_CHANGE_REMOVE) { error = EINVAL; break; } if (pca->addr.addr.type != PF_ADDR_ADDRMASK && pca->addr.addr.type != PF_ADDR_DYNIFTL && pca->addr.addr.type != PF_ADDR_TABLE) { error = EINVAL; break; } if (pca->action != PF_CHANGE_REMOVE) { #ifndef INET if (pca->af == AF_INET) { error = EAFNOSUPPORT; break; } #endif /* INET */ #ifndef INET6 if (pca->af == AF_INET6) { error = EAFNOSUPPORT; break; } #endif /* INET6 */ newpa = malloc(sizeof(*newpa), M_PFRULE, M_WAITOK); bcopy(&pca->addr, newpa, sizeof(struct pf_pooladdr)); if (newpa->ifname[0]) kif = malloc(sizeof(*kif), PFI_MTYPE, M_WAITOK); newpa->kif = NULL; } #define ERROUT(x) { error = (x); goto DIOCCHANGEADDR_error; } PF_RULES_WLOCK(); ruleset = pf_find_ruleset(pca->anchor); if (ruleset == NULL) ERROUT(EBUSY); pool = pf_get_pool(pca->anchor, pca->ticket, pca->r_action, pca->r_num, pca->r_last, 1, 1); if (pool == NULL) ERROUT(EBUSY); if (pca->action != PF_CHANGE_REMOVE) { if (newpa->ifname[0]) { newpa->kif = pfi_kif_attach(kif, newpa->ifname); pfi_kif_ref(newpa->kif); kif = NULL; } switch (newpa->addr.type) { case PF_ADDR_DYNIFTL: error = pfi_dynaddr_setup(&newpa->addr, pca->af); break; case PF_ADDR_TABLE: newpa->addr.p.tbl = pfr_attach_table(ruleset, newpa->addr.v.tblname); if (newpa->addr.p.tbl == NULL) error = ENOMEM; break; } if (error) goto DIOCCHANGEADDR_error; } switch (pca->action) { case PF_CHANGE_ADD_HEAD: oldpa = TAILQ_FIRST(&pool->list); break; case PF_CHANGE_ADD_TAIL: oldpa = TAILQ_LAST(&pool->list, pf_palist); break; default: oldpa = TAILQ_FIRST(&pool->list); for (int i = 0; oldpa && i < pca->nr; i++) oldpa = TAILQ_NEXT(oldpa, entries); if (oldpa == NULL) ERROUT(EINVAL); } if (pca->action == PF_CHANGE_REMOVE) { TAILQ_REMOVE(&pool->list, oldpa, entries); switch (oldpa->addr.type) { case PF_ADDR_DYNIFTL: pfi_dynaddr_remove(oldpa->addr.p.dyn); break; case PF_ADDR_TABLE: pfr_detach_table(oldpa->addr.p.tbl); break; } if (oldpa->kif) pfi_kif_unref(oldpa->kif); free(oldpa, M_PFRULE); } else { if (oldpa == NULL) TAILQ_INSERT_TAIL(&pool->list, newpa, entries); else if (pca->action == PF_CHANGE_ADD_HEAD || pca->action == PF_CHANGE_ADD_BEFORE) TAILQ_INSERT_BEFORE(oldpa, newpa, entries); else TAILQ_INSERT_AFTER(&pool->list, oldpa, newpa, entries); } pool->cur = TAILQ_FIRST(&pool->list); PF_ACPY(&pool->counter, &pool->cur->addr.v.a.addr, pca->af); PF_RULES_WUNLOCK(); break; #undef ERROUT DIOCCHANGEADDR_error: if (newpa->kif) pfi_kif_unref(newpa->kif); PF_RULES_WUNLOCK(); if (newpa != NULL) free(newpa, M_PFRULE); if (kif != NULL) free(kif, PFI_MTYPE); break; } case DIOCGETRULESETS: { struct pfioc_ruleset *pr = (struct pfioc_ruleset *)addr; struct pf_ruleset *ruleset; struct pf_anchor *anchor; PF_RULES_RLOCK(); pr->path[sizeof(pr->path) - 1] = 0; if ((ruleset = pf_find_ruleset(pr->path)) == NULL) { PF_RULES_RUNLOCK(); error = ENOENT; break; } pr->nr = 0; if (ruleset->anchor == NULL) { /* XXX kludge for pf_main_ruleset */ RB_FOREACH(anchor, pf_anchor_global, &V_pf_anchors) if (anchor->parent == NULL) pr->nr++; } else { RB_FOREACH(anchor, pf_anchor_node, &ruleset->anchor->children) pr->nr++; } PF_RULES_RUNLOCK(); break; } case DIOCGETRULESET: { struct pfioc_ruleset *pr = (struct pfioc_ruleset *)addr; struct pf_ruleset *ruleset; struct pf_anchor *anchor; u_int32_t nr = 0; PF_RULES_RLOCK(); pr->path[sizeof(pr->path) - 1] = 0; if ((ruleset = pf_find_ruleset(pr->path)) == NULL) { PF_RULES_RUNLOCK(); error = ENOENT; break; } pr->name[0] = 0; if (ruleset->anchor == NULL) { /* XXX kludge for pf_main_ruleset */ RB_FOREACH(anchor, pf_anchor_global, &V_pf_anchors) if (anchor->parent == NULL && nr++ == pr->nr) { strlcpy(pr->name, anchor->name, sizeof(pr->name)); break; } } else { RB_FOREACH(anchor, pf_anchor_node, &ruleset->anchor->children) if (nr++ == pr->nr) { strlcpy(pr->name, anchor->name, sizeof(pr->name)); break; } } if (!pr->name[0]) error = EBUSY; PF_RULES_RUNLOCK(); break; } case DIOCRCLRTABLES: { struct pfioc_table *io = (struct pfioc_table *)addr; if (io->pfrio_esize != 0) { error = ENODEV; break; } PF_RULES_WLOCK(); error = pfr_clr_tables(&io->pfrio_table, &io->pfrio_ndel, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); break; } case DIOCRADDTABLES: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_table *pfrts; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_table)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_table); pfrts = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfrts, totlen); if (error) { free(pfrts, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_add_tables(pfrts, io->pfrio_size, &io->pfrio_nadd, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); free(pfrts, M_TEMP); break; } case DIOCRDELTABLES: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_table *pfrts; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_table)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_table); pfrts = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfrts, totlen); if (error) { free(pfrts, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_del_tables(pfrts, io->pfrio_size, &io->pfrio_ndel, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); free(pfrts, M_TEMP); break; } case DIOCRGETTABLES: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_table *pfrts; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_table)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_table); pfrts = malloc(totlen, M_TEMP, M_WAITOK); PF_RULES_RLOCK(); error = pfr_get_tables(&io->pfrio_table, pfrts, &io->pfrio_size, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_RUNLOCK(); if (error == 0) error = copyout(pfrts, io->pfrio_buffer, totlen); free(pfrts, M_TEMP); break; } case DIOCRGETTSTATS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_tstats *pfrtstats; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_tstats)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_tstats); pfrtstats = malloc(totlen, M_TEMP, M_WAITOK); PF_RULES_WLOCK(); error = pfr_get_tstats(&io->pfrio_table, pfrtstats, &io->pfrio_size, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); if (error == 0) error = copyout(pfrtstats, io->pfrio_buffer, totlen); free(pfrtstats, M_TEMP); break; } case DIOCRCLRTSTATS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_table *pfrts; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_table)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_table); pfrts = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfrts, totlen); if (error) { free(pfrts, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_clr_tstats(pfrts, io->pfrio_size, &io->pfrio_nzero, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); free(pfrts, M_TEMP); break; } case DIOCRSETTFLAGS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_table *pfrts; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_table)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_table); pfrts = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfrts, totlen); if (error) { free(pfrts, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_set_tflags(pfrts, io->pfrio_size, io->pfrio_setflag, io->pfrio_clrflag, &io->pfrio_nchange, &io->pfrio_ndel, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); free(pfrts, M_TEMP); break; } case DIOCRCLRADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; if (io->pfrio_esize != 0) { error = ENODEV; break; } PF_RULES_WLOCK(); error = pfr_clr_addrs(&io->pfrio_table, &io->pfrio_ndel, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); break; } case DIOCRADDADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_add_addrs(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_nadd, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); if (error == 0 && io->pfrio_flags & PFR_FLAG_FEEDBACK) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRDELADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_del_addrs(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_ndel, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); if (error == 0 && io->pfrio_flags & PFR_FLAG_FEEDBACK) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRSETADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen, count; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } count = max(io->pfrio_size, io->pfrio_size2); totlen = count * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_set_addrs(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_size2, &io->pfrio_nadd, &io->pfrio_ndel, &io->pfrio_nchange, io->pfrio_flags | PFR_FLAG_USERIOCTL, 0); PF_RULES_WUNLOCK(); if (error == 0 && io->pfrio_flags & PFR_FLAG_FEEDBACK) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRGETADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); PF_RULES_RLOCK(); error = pfr_get_addrs(&io->pfrio_table, pfras, &io->pfrio_size, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_RUNLOCK(); if (error == 0) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRGETASTATS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_astats *pfrastats; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_astats)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_astats); pfrastats = malloc(totlen, M_TEMP, M_WAITOK); PF_RULES_RLOCK(); error = pfr_get_astats(&io->pfrio_table, pfrastats, &io->pfrio_size, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_RUNLOCK(); if (error == 0) error = copyout(pfrastats, io->pfrio_buffer, totlen); free(pfrastats, M_TEMP); break; } case DIOCRCLRASTATS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_clr_astats(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_nzero, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); if (error == 0 && io->pfrio_flags & PFR_FLAG_FEEDBACK) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRTSTADDRS: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_RLOCK(); error = pfr_tst_addrs(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_nmatch, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_RUNLOCK(); if (error == 0) error = copyout(pfras, io->pfrio_buffer, totlen); free(pfras, M_TEMP); break; } case DIOCRINADEFINE: { struct pfioc_table *io = (struct pfioc_table *)addr; struct pfr_addr *pfras; size_t totlen; if (io->pfrio_esize != sizeof(struct pfr_addr)) { error = ENODEV; break; } totlen = io->pfrio_size * sizeof(struct pfr_addr); pfras = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->pfrio_buffer, pfras, totlen); if (error) { free(pfras, M_TEMP); break; } PF_RULES_WLOCK(); error = pfr_ina_define(&io->pfrio_table, pfras, io->pfrio_size, &io->pfrio_nadd, &io->pfrio_naddr, io->pfrio_ticket, io->pfrio_flags | PFR_FLAG_USERIOCTL); PF_RULES_WUNLOCK(); free(pfras, M_TEMP); break; } case DIOCOSFPADD: { struct pf_osfp_ioctl *io = (struct pf_osfp_ioctl *)addr; PF_RULES_WLOCK(); error = pf_osfp_add(io); PF_RULES_WUNLOCK(); break; } case DIOCOSFPGET: { struct pf_osfp_ioctl *io = (struct pf_osfp_ioctl *)addr; PF_RULES_RLOCK(); error = pf_osfp_get(io); PF_RULES_RUNLOCK(); break; } case DIOCXBEGIN: { struct pfioc_trans *io = (struct pfioc_trans *)addr; struct pfioc_trans_e *ioes, *ioe; size_t totlen; int i; if (io->esize != sizeof(*ioe)) { error = ENODEV; break; } totlen = sizeof(struct pfioc_trans_e) * io->size; ioes = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->array, ioes, totlen); if (error) { free(ioes, M_TEMP); break; } PF_RULES_WLOCK(); for (i = 0, ioe = ioes; i < io->size; i++, ioe++) { switch (ioe->rs_num) { #ifdef ALTQ case PF_RULESET_ALTQ: if (ioe->anchor[0]) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EINVAL; goto fail; } if ((error = pf_begin_altq(&ioe->ticket))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; } break; #endif /* ALTQ */ case PF_RULESET_TABLE: { struct pfr_table table; bzero(&table, sizeof(table)); strlcpy(table.pfrt_anchor, ioe->anchor, sizeof(table.pfrt_anchor)); if ((error = pfr_ina_begin(&table, &ioe->ticket, NULL, 0))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; } break; } default: if ((error = pf_begin_rules(&ioe->ticket, ioe->rs_num, ioe->anchor))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; } break; } } PF_RULES_WUNLOCK(); error = copyout(ioes, io->array, totlen); free(ioes, M_TEMP); break; } case DIOCXROLLBACK: { struct pfioc_trans *io = (struct pfioc_trans *)addr; struct pfioc_trans_e *ioe, *ioes; size_t totlen; int i; if (io->esize != sizeof(*ioe)) { error = ENODEV; break; } totlen = sizeof(struct pfioc_trans_e) * io->size; ioes = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->array, ioes, totlen); if (error) { free(ioes, M_TEMP); break; } PF_RULES_WLOCK(); for (i = 0, ioe = ioes; i < io->size; i++, ioe++) { switch (ioe->rs_num) { #ifdef ALTQ case PF_RULESET_ALTQ: if (ioe->anchor[0]) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EINVAL; goto fail; } if ((error = pf_rollback_altq(ioe->ticket))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; #endif /* ALTQ */ case PF_RULESET_TABLE: { struct pfr_table table; bzero(&table, sizeof(table)); strlcpy(table.pfrt_anchor, ioe->anchor, sizeof(table.pfrt_anchor)); if ((error = pfr_ina_rollback(&table, ioe->ticket, NULL, 0))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; } default: if ((error = pf_rollback_rules(ioe->ticket, ioe->rs_num, ioe->anchor))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; } } PF_RULES_WUNLOCK(); free(ioes, M_TEMP); break; } case DIOCXCOMMIT: { struct pfioc_trans *io = (struct pfioc_trans *)addr; struct pfioc_trans_e *ioe, *ioes; struct pf_ruleset *rs; size_t totlen; int i; if (io->esize != sizeof(*ioe)) { error = ENODEV; break; } totlen = sizeof(struct pfioc_trans_e) * io->size; ioes = malloc(totlen, M_TEMP, M_WAITOK); error = copyin(io->array, ioes, totlen); if (error) { free(ioes, M_TEMP); break; } PF_RULES_WLOCK(); /* First makes sure everything will succeed. */ for (i = 0, ioe = ioes; i < io->size; i++, ioe++) { switch (ioe->rs_num) { #ifdef ALTQ case PF_RULESET_ALTQ: if (ioe->anchor[0]) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EINVAL; goto fail; } if (!V_altqs_inactive_open || ioe->ticket != V_ticket_altqs_inactive) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EBUSY; goto fail; } break; #endif /* ALTQ */ case PF_RULESET_TABLE: rs = pf_find_ruleset(ioe->anchor); if (rs == NULL || !rs->topen || ioe->ticket != rs->tticket) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EBUSY; goto fail; } break; default: if (ioe->rs_num < 0 || ioe->rs_num >= PF_RULESET_MAX) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EINVAL; goto fail; } rs = pf_find_ruleset(ioe->anchor); if (rs == NULL || !rs->rules[ioe->rs_num].inactive.open || rs->rules[ioe->rs_num].inactive.ticket != ioe->ticket) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); error = EBUSY; goto fail; } break; } } /* Now do the commit - no errors should happen here. */ for (i = 0, ioe = ioes; i < io->size; i++, ioe++) { switch (ioe->rs_num) { #ifdef ALTQ case PF_RULESET_ALTQ: if ((error = pf_commit_altq(ioe->ticket))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; #endif /* ALTQ */ case PF_RULESET_TABLE: { struct pfr_table table; bzero(&table, sizeof(table)); strlcpy(table.pfrt_anchor, ioe->anchor, sizeof(table.pfrt_anchor)); if ((error = pfr_ina_commit(&table, ioe->ticket, NULL, NULL, 0))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; } default: if ((error = pf_commit_rules(ioe->ticket, ioe->rs_num, ioe->anchor))) { PF_RULES_WUNLOCK(); free(ioes, M_TEMP); goto fail; /* really bad */ } break; } } PF_RULES_WUNLOCK(); free(ioes, M_TEMP); break; } case DIOCGETSRCNODES: { struct pfioc_src_nodes *psn = (struct pfioc_src_nodes *)addr; struct pf_srchash *sh; struct pf_src_node *n, *p, *pstore; uint32_t i, nr = 0; if (psn->psn_len == 0) { for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) { PF_HASHROW_LOCK(sh); LIST_FOREACH(n, &sh->nodes, entry) nr++; PF_HASHROW_UNLOCK(sh); } psn->psn_len = sizeof(struct pf_src_node) * nr; break; } p = pstore = malloc(psn->psn_len, M_TEMP, M_WAITOK); for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) { PF_HASHROW_LOCK(sh); LIST_FOREACH(n, &sh->nodes, entry) { int secs = time_uptime, diff; if ((nr + 1) * sizeof(*p) > (unsigned)psn->psn_len) break; bcopy(n, p, sizeof(struct pf_src_node)); if (n->rule.ptr != NULL) p->rule.nr = n->rule.ptr->nr; p->creation = secs - p->creation; if (p->expire > secs) p->expire -= secs; else p->expire = 0; /* Adjust the connection rate estimate. */ diff = secs - n->conn_rate.last; if (diff >= n->conn_rate.seconds) p->conn_rate.count = 0; else p->conn_rate.count -= n->conn_rate.count * diff / n->conn_rate.seconds; p++; nr++; } PF_HASHROW_UNLOCK(sh); } error = copyout(pstore, psn->psn_src_nodes, sizeof(struct pf_src_node) * nr); if (error) { free(pstore, M_TEMP); break; } psn->psn_len = sizeof(struct pf_src_node) * nr; free(pstore, M_TEMP); break; } case DIOCCLRSRCNODES: { pf_clear_srcnodes(NULL); pf_purge_expired_src_nodes(); break; } case DIOCKILLSRCNODES: pf_kill_srcnodes((struct pfioc_src_node_kill *)addr); break; case DIOCSETHOSTID: { u_int32_t *hostid = (u_int32_t *)addr; PF_RULES_WLOCK(); if (*hostid == 0) V_pf_status.hostid = arc4random(); else V_pf_status.hostid = *hostid; PF_RULES_WUNLOCK(); break; } case DIOCOSFPFLUSH: PF_RULES_WLOCK(); pf_osfp_flush(); PF_RULES_WUNLOCK(); break; case DIOCIGETIFACES: { struct pfioc_iface *io = (struct pfioc_iface *)addr; struct pfi_kif *ifstore; size_t bufsiz; if (io->pfiio_esize != sizeof(struct pfi_kif)) { error = ENODEV; break; } bufsiz = io->pfiio_size * sizeof(struct pfi_kif); ifstore = malloc(bufsiz, M_TEMP, M_WAITOK); PF_RULES_RLOCK(); pfi_get_ifaces(io->pfiio_name, ifstore, &io->pfiio_size); PF_RULES_RUNLOCK(); error = copyout(ifstore, io->pfiio_buffer, bufsiz); free(ifstore, M_TEMP); break; } case DIOCSETIFFLAG: { struct pfioc_iface *io = (struct pfioc_iface *)addr; PF_RULES_WLOCK(); error = pfi_set_flags(io->pfiio_name, io->pfiio_flags); PF_RULES_WUNLOCK(); break; } case DIOCCLRIFFLAG: { struct pfioc_iface *io = (struct pfioc_iface *)addr; PF_RULES_WLOCK(); error = pfi_clear_flags(io->pfiio_name, io->pfiio_flags); PF_RULES_WUNLOCK(); break; } default: error = ENODEV; break; } fail: if (sx_xlocked(&pf_ioctl_lock)) sx_xunlock(&pf_ioctl_lock); CURVNET_RESTORE(); return (error); } void pfsync_state_export(struct pfsync_state *sp, struct pf_state *st) { bzero(sp, sizeof(struct pfsync_state)); /* copy from state key */ sp->key[PF_SK_WIRE].addr[0] = st->key[PF_SK_WIRE]->addr[0]; sp->key[PF_SK_WIRE].addr[1] = st->key[PF_SK_WIRE]->addr[1]; sp->key[PF_SK_WIRE].port[0] = st->key[PF_SK_WIRE]->port[0]; sp->key[PF_SK_WIRE].port[1] = st->key[PF_SK_WIRE]->port[1]; sp->key[PF_SK_STACK].addr[0] = st->key[PF_SK_STACK]->addr[0]; sp->key[PF_SK_STACK].addr[1] = st->key[PF_SK_STACK]->addr[1]; sp->key[PF_SK_STACK].port[0] = st->key[PF_SK_STACK]->port[0]; sp->key[PF_SK_STACK].port[1] = st->key[PF_SK_STACK]->port[1]; sp->proto = st->key[PF_SK_WIRE]->proto; sp->af = st->key[PF_SK_WIRE]->af; /* copy from state */ strlcpy(sp->ifname, st->kif->pfik_name, sizeof(sp->ifname)); bcopy(&st->rt_addr, &sp->rt_addr, sizeof(sp->rt_addr)); sp->creation = htonl(time_uptime - st->creation); sp->expire = pf_state_expires(st); if (sp->expire <= time_uptime) sp->expire = htonl(0); else sp->expire = htonl(sp->expire - time_uptime); sp->direction = st->direction; sp->log = st->log; sp->timeout = st->timeout; sp->state_flags = st->state_flags; if (st->src_node) sp->sync_flags |= PFSYNC_FLAG_SRCNODE; if (st->nat_src_node) sp->sync_flags |= PFSYNC_FLAG_NATSRCNODE; sp->id = st->id; sp->creatorid = st->creatorid; pf_state_peer_hton(&st->src, &sp->src); pf_state_peer_hton(&st->dst, &sp->dst); if (st->rule.ptr == NULL) sp->rule = htonl(-1); else sp->rule = htonl(st->rule.ptr->nr); if (st->anchor.ptr == NULL) sp->anchor = htonl(-1); else sp->anchor = htonl(st->anchor.ptr->nr); if (st->nat_rule.ptr == NULL) sp->nat_rule = htonl(-1); else sp->nat_rule = htonl(st->nat_rule.ptr->nr); pf_state_counter_hton(st->packets[0], sp->packets[0]); pf_state_counter_hton(st->packets[1], sp->packets[1]); pf_state_counter_hton(st->bytes[0], sp->bytes[0]); pf_state_counter_hton(st->bytes[1], sp->bytes[1]); } static void pf_tbladdr_copyout(struct pf_addr_wrap *aw) { struct pfr_ktable *kt; KASSERT(aw->type == PF_ADDR_TABLE, ("%s: type %u", __func__, aw->type)); kt = aw->p.tbl; if (!(kt->pfrkt_flags & PFR_TFLAG_ACTIVE) && kt->pfrkt_root != NULL) kt = kt->pfrkt_root; aw->p.tbl = NULL; aw->p.tblcnt = (kt->pfrkt_flags & PFR_TFLAG_ACTIVE) ? kt->pfrkt_cnt : -1; } /* * XXX - Check for version missmatch!!! */ static void pf_clear_states(void) { struct pf_state *s; u_int i; for (i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; relock: PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { s->timeout = PFTM_PURGE; /* Don't send out individual delete messages. */ s->state_flags |= PFSTATE_NOSYNC; pf_unlink_state(s, PF_ENTER_LOCKED); goto relock; } PF_HASHROW_UNLOCK(ih); } } static int pf_clear_tables(void) { struct pfioc_table io; int error; bzero(&io, sizeof(io)); error = pfr_clr_tables(&io.pfrio_table, &io.pfrio_ndel, io.pfrio_flags); return (error); } static void pf_clear_srcnodes(struct pf_src_node *n) { struct pf_state *s; int i; for (i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { if (n == NULL || n == s->src_node) s->src_node = NULL; if (n == NULL || n == s->nat_src_node) s->nat_src_node = NULL; } PF_HASHROW_UNLOCK(ih); } if (n == NULL) { struct pf_srchash *sh; for (i = 0, sh = V_pf_srchash; i <= pf_srchashmask; i++, sh++) { PF_HASHROW_LOCK(sh); LIST_FOREACH(n, &sh->nodes, entry) { n->expire = 1; n->states = 0; } PF_HASHROW_UNLOCK(sh); } } else { /* XXX: hash slot should already be locked here. */ n->expire = 1; n->states = 0; } } static void pf_kill_srcnodes(struct pfioc_src_node_kill *psnk) { struct pf_src_node_list kill; LIST_INIT(&kill); for (int i = 0; i <= pf_srchashmask; i++) { struct pf_srchash *sh = &V_pf_srchash[i]; struct pf_src_node *sn, *tmp; PF_HASHROW_LOCK(sh); LIST_FOREACH_SAFE(sn, &sh->nodes, entry, tmp) if (PF_MATCHA(psnk->psnk_src.neg, &psnk->psnk_src.addr.v.a.addr, &psnk->psnk_src.addr.v.a.mask, &sn->addr, sn->af) && PF_MATCHA(psnk->psnk_dst.neg, &psnk->psnk_dst.addr.v.a.addr, &psnk->psnk_dst.addr.v.a.mask, &sn->raddr, sn->af)) { pf_unlink_src_node(sn); LIST_INSERT_HEAD(&kill, sn, entry); sn->expire = 1; } PF_HASHROW_UNLOCK(sh); } for (int i = 0; i <= pf_hashmask; i++) { struct pf_idhash *ih = &V_pf_idhash[i]; struct pf_state *s; PF_HASHROW_LOCK(ih); LIST_FOREACH(s, &ih->states, entry) { if (s->src_node && s->src_node->expire == 1) s->src_node = NULL; if (s->nat_src_node && s->nat_src_node->expire == 1) s->nat_src_node = NULL; } PF_HASHROW_UNLOCK(ih); } psnk->psnk_killed = pf_free_src_nodes(&kill); } /* * XXX - Check for version missmatch!!! */ /* * Duplicate pfctl -Fa operation to get rid of as much as we can. */ static int shutdown_pf(void) { int error = 0; u_int32_t t[5]; char nn = '\0'; do { if ((error = pf_begin_rules(&t[0], PF_RULESET_SCRUB, &nn)) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: SCRUB\n")); break; } if ((error = pf_begin_rules(&t[1], PF_RULESET_FILTER, &nn)) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: FILTER\n")); break; /* XXX: rollback? */ } if ((error = pf_begin_rules(&t[2], PF_RULESET_NAT, &nn)) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: NAT\n")); break; /* XXX: rollback? */ } if ((error = pf_begin_rules(&t[3], PF_RULESET_BINAT, &nn)) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: BINAT\n")); break; /* XXX: rollback? */ } if ((error = pf_begin_rules(&t[4], PF_RULESET_RDR, &nn)) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: RDR\n")); break; /* XXX: rollback? */ } /* XXX: these should always succeed here */ pf_commit_rules(t[0], PF_RULESET_SCRUB, &nn); pf_commit_rules(t[1], PF_RULESET_FILTER, &nn); pf_commit_rules(t[2], PF_RULESET_NAT, &nn); pf_commit_rules(t[3], PF_RULESET_BINAT, &nn); pf_commit_rules(t[4], PF_RULESET_RDR, &nn); if ((error = pf_clear_tables()) != 0) break; #ifdef ALTQ if ((error = pf_begin_altq(&t[0])) != 0) { DPFPRINTF(PF_DEBUG_MISC, ("shutdown_pf: ALTQ\n")); break; } pf_commit_altq(t[0]); #endif pf_clear_states(); pf_clear_srcnodes(NULL); /* status does not use malloced mem so no need to cleanup */ /* fingerprints and interfaces have their own cleanup code */ /* Free counters last as we updated them during shutdown. */ counter_u64_free(V_pf_default_rule.states_cur); counter_u64_free(V_pf_default_rule.states_tot); counter_u64_free(V_pf_default_rule.src_nodes); for (int i = 0; i < PFRES_MAX; i++) counter_u64_free(V_pf_status.counters[i]); for (int i = 0; i < LCNT_MAX; i++) counter_u64_free(V_pf_status.lcounters[i]); for (int i = 0; i < FCNT_MAX; i++) counter_u64_free(V_pf_status.fcounters[i]); for (int i = 0; i < SCNT_MAX; i++) counter_u64_free(V_pf_status.scounters[i]); } while(0); return (error); } #ifdef INET static int pf_check_in(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp) { int chk; chk = pf_test(PF_IN, ifp, m, inp); if (chk && *m) { m_freem(*m); *m = NULL; } if (chk != PF_PASS) return (EACCES); return (0); } static int pf_check_out(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp) { int chk; chk = pf_test(PF_OUT, ifp, m, inp); if (chk && *m) { m_freem(*m); *m = NULL; } if (chk != PF_PASS) return (EACCES); return (0); } #endif #ifdef INET6 static int pf_check6_in(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp) { int chk; /* * In case of loopback traffic IPv6 uses the real interface in * order to support scoped addresses. In order to support stateful * filtering we have change this to lo0 as it is the case in IPv4. */ CURVNET_SET(ifp->if_vnet); chk = pf_test6(PF_IN, (*m)->m_flags & M_LOOP ? V_loif : ifp, m, inp); CURVNET_RESTORE(); if (chk && *m) { m_freem(*m); *m = NULL; } if (chk != PF_PASS) return (EACCES); return (0); } static int pf_check6_out(void *arg, struct mbuf **m, struct ifnet *ifp, int dir, struct inpcb *inp) { int chk; CURVNET_SET(ifp->if_vnet); chk = pf_test6(PF_OUT, ifp, m, inp); CURVNET_RESTORE(); if (chk && *m) { m_freem(*m); *m = NULL; } if (chk != PF_PASS) return (EACCES); return (0); } #endif /* INET6 */ static int hook_pf(void) { #ifdef INET struct pfil_head *pfh_inet; #endif #ifdef INET6 struct pfil_head *pfh_inet6; #endif if (V_pf_pfil_hooked) return (0); #ifdef INET pfh_inet = pfil_head_get(PFIL_TYPE_AF, AF_INET); if (pfh_inet == NULL) return (ESRCH); /* XXX */ pfil_add_hook(pf_check_in, NULL, PFIL_IN | PFIL_WAITOK, pfh_inet); pfil_add_hook(pf_check_out, NULL, PFIL_OUT | PFIL_WAITOK, pfh_inet); #endif #ifdef INET6 pfh_inet6 = pfil_head_get(PFIL_TYPE_AF, AF_INET6); if (pfh_inet6 == NULL) { #ifdef INET pfil_remove_hook(pf_check_in, NULL, PFIL_IN | PFIL_WAITOK, pfh_inet); pfil_remove_hook(pf_check_out, NULL, PFIL_OUT | PFIL_WAITOK, pfh_inet); #endif return (ESRCH); /* XXX */ } pfil_add_hook(pf_check6_in, NULL, PFIL_IN | PFIL_WAITOK, pfh_inet6); pfil_add_hook(pf_check6_out, NULL, PFIL_OUT | PFIL_WAITOK, pfh_inet6); #endif V_pf_pfil_hooked = 1; return (0); } static int dehook_pf(void) { #ifdef INET struct pfil_head *pfh_inet; #endif #ifdef INET6 struct pfil_head *pfh_inet6; #endif if (V_pf_pfil_hooked == 0) return (0); #ifdef INET pfh_inet = pfil_head_get(PFIL_TYPE_AF, AF_INET); if (pfh_inet == NULL) return (ESRCH); /* XXX */ pfil_remove_hook(pf_check_in, NULL, PFIL_IN | PFIL_WAITOK, pfh_inet); pfil_remove_hook(pf_check_out, NULL, PFIL_OUT | PFIL_WAITOK, pfh_inet); #endif #ifdef INET6 pfh_inet6 = pfil_head_get(PFIL_TYPE_AF, AF_INET6); if (pfh_inet6 == NULL) return (ESRCH); /* XXX */ pfil_remove_hook(pf_check6_in, NULL, PFIL_IN | PFIL_WAITOK, pfh_inet6); pfil_remove_hook(pf_check6_out, NULL, PFIL_OUT | PFIL_WAITOK, pfh_inet6); #endif V_pf_pfil_hooked = 0; return (0); } static void pf_load_vnet(void) { VNET_ITERATOR_DECL(vnet_iter); VNET_LIST_RLOCK(); VNET_FOREACH(vnet_iter) { CURVNET_SET(vnet_iter); V_pf_pfil_hooked = 0; TAILQ_INIT(&V_pf_tags); TAILQ_INIT(&V_pf_qids); CURVNET_RESTORE(); } VNET_LIST_RUNLOCK(); pfattach_vnet(); V_pf_vnet_active = 1; } static int pf_load(void) { int error; rw_init(&pf_rules_lock, "pf rulesets"); sx_init(&pf_ioctl_lock, "pf ioctl"); pf_mtag_initialize(); pf_dev = make_dev(&pf_cdevsw, 0, 0, 0, 0600, PF_NAME); if (pf_dev == NULL) return (ENOMEM); pf_end_threads = 0; error = kproc_create(pf_purge_thread, NULL, NULL, 0, 0, "pf purge"); if (error != 0) return (error); pfi_initialize(); return (0); } static void pf_unload_vnet(void) { int error; V_pf_vnet_active = 0; V_pf_status.running = 0; swi_remove(V_pf_swi_cookie); error = dehook_pf(); if (error) { /* * Should not happen! * XXX Due to error code ESRCH, kldunload will show * a message like 'No such process'. */ printf("%s : pfil unregisteration fail\n", __FUNCTION__); return; } pf_unload_vnet_purge(); PF_RULES_WLOCK(); shutdown_pf(); PF_RULES_WUNLOCK(); pf_normalize_cleanup(); PF_RULES_WLOCK(); pfi_cleanup_vnet(); PF_RULES_WUNLOCK(); pfr_cleanup(); pf_osfp_flush(); pf_cleanup(); if (IS_DEFAULT_VNET(curvnet)) pf_mtag_cleanup(); } static int pf_unload(void) { int error = 0; pf_end_threads = 1; while (pf_end_threads < 2) { wakeup_one(pf_purge_thread); rw_sleep(pf_purge_thread, &pf_rules_lock, 0, "pftmo", 0); } if (pf_dev != NULL) destroy_dev(pf_dev); pfi_cleanup(); rw_destroy(&pf_rules_lock); sx_destroy(&pf_ioctl_lock); return (error); } static void vnet_pf_init(void *unused __unused) { pf_load_vnet(); } VNET_SYSINIT(vnet_pf_init, SI_SUB_PROTO_FIREWALL, SI_ORDER_THIRD, vnet_pf_init, NULL); static void vnet_pf_uninit(const void *unused __unused) { pf_unload_vnet(); } VNET_SYSUNINIT(vnet_pf_uninit, SI_SUB_PROTO_FIREWALL, SI_ORDER_THIRD, vnet_pf_uninit, NULL); static int pf_modevent(module_t mod, int type, void *data) { int error = 0; switch(type) { case MOD_LOAD: error = pf_load(); break; case MOD_QUIESCE: /* * Module should not be unloaded due to race conditions. */ error = EBUSY; break; case MOD_UNLOAD: error = pf_unload(); break; default: error = EINVAL; break; } return (error); } static moduledata_t pf_mod = { "pf", pf_modevent, 0 }; DECLARE_MODULE(pf, pf_mod, SI_SUB_PROTO_FIREWALL, SI_ORDER_SECOND); MODULE_VERSION(pf, PF_MODVER);