1NFT(8) NFT(8)
2
3
4
6 nft - Administration tool of the nftables framework for packet
7 filtering and classification
8
10 nft [ -nNscaeSupyjt ] [ -I directory ] [ -f filename | -i | cmd ...]
11 nft -h
12 nft -v
13
15 nft is the command line tool used to set up, maintain and inspect
16 packet filtering and classification rules in the Linux kernel, in the
17 nftables framework. The Linux kernel subsystem is known as nf_tables,
18 and ‘nf’ stands for Netfilter.
19
21 The command accepts several different options which are documented here
22 in groups for better understanding of their meaning. You can get
23 information about options by running nft --help.
24
25 General options:
26
27 -h, --help
28 Show help message and all options.
29
30 -v, --version
31 Show version.
32
33 -V
34 Show long version information, including compile-time
35 configuration.
36
37 Ruleset input handling options that specify to how to load rulesets:
38
39 -f, --file filename
40 Read input from filename. If filename is -, read from stdin.
41
42 -i, --interactive
43 Read input from an interactive readline CLI. You can use quit to
44 exit, or use the EOF marker, normally this is CTRL-D.
45
46 -I, --includepath directory
47 Add the directory directory to the list of directories to be
48 searched for included files. This option may be specified multiple
49 times.
50
51 -c, --check
52 Check commands validity without actually applying the changes.
53
54 Ruleset list output formatting that modify the output of the list
55 ruleset command:
56
57 -a, --handle
58 Show object handles in output.
59
60 -s, --stateless
61 Omit stateful information of rules and stateful objects.
62
63 -t, --terse
64 Omit contents of sets from output.
65
66 -S, --service
67 Translate ports to service names as defined by /etc/services.
68
69 -N, --reversedns
70 Translate IP address to names via reverse DNS lookup. This may slow
71 down your listing since it generates network traffic.
72
73 -u, --guid
74 Translate numeric UID/GID to names as defined by /etc/passwd and
75 /etc/group.
76
77 -n, --numeric
78 Print fully numerical output.
79
80 -y, --numeric-priority
81 Display base chain priority numerically.
82
83 -p, --numeric-protocol
84 Display layer 4 protocol numerically.
85
86 -T, --numeric-time
87 Show time, day and hour values in numeric format.
88
89 Command output formatting:
90
91 -e, --echo
92 When inserting items into the ruleset using add, insert or replace
93 commands, print notifications just like nft monitor.
94
95 -j, --json
96 Format output in JSON. See libnftables-json(5) for a schema
97 description.
98
99 -d, --debug level
100 Enable debugging output. The debug level can be any of scanner,
101 parser, eval, netlink, mnl, proto-ctx, segtree, all. You can
102 combine more than one by separating by the , symbol, for example -d
103 eval,mnl.
104
106 LEXICAL CONVENTIONS
107 Input is parsed line-wise. When the last character of a line, just
108 before the newline character, is a non-quoted backslash (\), the next
109 line is treated as a continuation. Multiple commands on the same line
110 can be separated using a semicolon (;).
111
112 A hash sign (#) begins a comment. All following characters on the same
113 line are ignored.
114
115 Identifiers begin with an alphabetic character (a-z,A-Z), followed zero
116 or more alphanumeric characters (a-z,A-Z,0-9) and the characters slash
117 (/), backslash (\), underscore (_) and dot (.). Identifiers using
118 different characters or clashing with a keyword need to be enclosed in
119 double quotes (").
120
121 INCLUDE FILES
122 include filename
123
124 Other files can be included by using the include statement. The
125 directories to be searched for include files can be specified using the
126 -I/--includepath option. You can override this behaviour either by
127 prepending ‘./’ to your path to force inclusion of files located in the
128 current working directory (i.e. relative path) or / for file location
129 expressed as an absolute path.
130
131 If -I/--includepath is not specified, then nft relies on the default
132 directory that is specified at compile time. You can retrieve this
133 default directory via -h/--help option.
134
135 Include statements support the usual shell wildcard symbols (\*,?,[]).
136 Having no matches for an include statement is not an error, if wildcard
137 symbols are used in the include statement. This allows having
138 potentially empty include directories for statements like include
139 "/etc/firewall/rules/". The wildcard matches are loaded in alphabetical
140 order. Files beginning with dot (.) are not matched by include
141 statements.
142
143 SYMBOLIC VARIABLES
144 define variable = expr
145 $variable
146
147 Symbolic variables can be defined using the define statement. Variable
148 references are expressions and can be used initialize other variables.
149 The scope of a definition is the current block and all blocks contained
150 within.
151
152 Using symbolic variables.
153
154 define int_if1 = eth0
155 define int_if2 = eth1
156 define int_ifs = { $int_if1, $int_if2 }
157
158 filter input iif $int_ifs accept
159
160
162 Address families determine the type of packets which are processed. For
163 each address family, the kernel contains so called hooks at specific
164 stages of the packet processing paths, which invoke nftables if rules
165 for these hooks exist.
166
167
168 ip IPv4 address family.
169
170 ip6 IPv6 address family.
171
172 inet Internet (IPv4/IPv6)
173 address family.
174
175 arp ARP address family,
176 handling IPv4 ARP packets.
177
178 bridge Bridge address family,
179 handling packets which
180 traverse a bridge device.
181
182 netdev Netdev address family,
183 handling packets from
184 ingress.
185
186
187 All nftables objects exist in address family specific namespaces,
188 therefore all identifiers include an address family. If an identifier
189 is specified without an address family, the ip family is used by
190 default.
191
192 IPV4/IPV6/INET ADDRESS FAMILIES
193 The IPv4/IPv6/Inet address families handle IPv4, IPv6 or both types of
194 packets. They contain five hooks at different packet processing stages
195 in the network stack.
196
197 Table 1. IPv4/IPv6/Inet address family hooks
198 ┌────────────┬────────────────────────────┐
199 │Hook │ Description │
200 ├────────────┼────────────────────────────┤
201 │ │ │
202 │prerouting │ All packets entering the │
203 │ │ system are processed by │
204 │ │ the prerouting hook. It is │
205 │ │ invoked before the routing │
206 │ │ process and is used for │
207 │ │ early filtering or │
208 │ │ changing packet attributes │
209 │ │ that affect routing. │
210 ├────────────┼────────────────────────────┤
211 │ │ │
212 │input │ Packets delivered to the │
213 │ │ local system are processed │
214 │ │ by the input hook. │
215 ├────────────┼────────────────────────────┤
216 │ │ │
217 │forward │ Packets forwarded to a │
218 │ │ different host are │
219 │ │ processed by the forward │
220 │ │ hook. │
221 ├────────────┼────────────────────────────┤
222 │ │ │
223 │output │ Packets sent by local │
224 │ │ processes are processed by │
225 │ │ the output hook. │
226 ├────────────┼────────────────────────────┤
227 │ │ │
228 │postrouting │ All packets leaving the │
229 │ │ system are processed by │
230 │ │ the postrouting hook. │
231 ├────────────┼────────────────────────────┤
232 │ │ │
233 │ingress │ All packets entering the │
234 │ │ system are processed by │
235 │ │ this hook. It is invoked │
236 │ │ before layer 3 protocol │
237 │ │ handlers, hence before the │
238 │ │ prerouting hook, and it │
239 │ │ can be used for filtering │
240 │ │ and policing. Ingress is │
241 │ │ only available for Inet │
242 │ │ family (since Linux kernel │
243 │ │ 5.10). │
244 └────────────┴────────────────────────────┘
245
246 ARP ADDRESS FAMILY
247 The ARP address family handles ARP packets received and sent by the
248 system. It is commonly used to mangle ARP packets for clustering.
249
250 Table 2. ARP address family hooks
251 ┌───────┬────────────────────────────┐
252 │Hook │ Description │
253 ├───────┼────────────────────────────┤
254 │ │ │
255 │input │ Packets delivered to the │
256 │ │ local system are processed │
257 │ │ by the input hook. │
258 ├───────┼────────────────────────────┤
259 │ │ │
260 │output │ Packets send by the local │
261 │ │ system are processed by │
262 │ │ the output hook. │
263 └───────┴────────────────────────────┘
264
265 BRIDGE ADDRESS FAMILY
266 The bridge address family handles Ethernet packets traversing bridge
267 devices.
268
269 The list of supported hooks is identical to IPv4/IPv6/Inet address
270 families above.
271
272 NETDEV ADDRESS FAMILY
273 The Netdev address family handles packets from the device ingress path.
274 This family allows you to filter packets of any ethertype such as ARP,
275 VLAN 802.1q, VLAN 802.1ad (Q-in-Q) as well as IPv4 and IPv6 packets.
276
277 Table 3. Netdev address family hooks
278 ┌────────┬────────────────────────────┐
279 │Hook │ Description │
280 ├────────┼────────────────────────────┤
281 │ │ │
282 │ingress │ All packets entering the │
283 │ │ system are processed by │
284 │ │ this hook. It is invoked │
285 │ │ after the network taps │
286 │ │ (ie. tcpdump), right after │
287 │ │ tc ingress and before │
288 │ │ layer 3 protocol handlers, │
289 │ │ it can be used for early │
290 │ │ filtering and policing. │
291 └────────┴────────────────────────────┘
292
294 {list | flush} ruleset [family]
295
296 The ruleset keyword is used to identify the whole set of tables,
297 chains, etc. currently in place in kernel. The following ruleset
298 commands exist:
299
300
301 list Print the ruleset in
302 human-readable format.
303
304 flush Clear the whole ruleset.
305 Note that, unlike
306 iptables, this will remove
307 all tables and whatever
308 they contain, effectively
309 leading to an empty
310 ruleset - no packet
311 filtering will happen
312 anymore, so the kernel
313 accepts any valid packet
314 it receives.
315
316
317 It is possible to limit list and flush to a specific address family
318 only. For a list of valid family names, see the section called “ADDRESS
319 FAMILIES” above.
320
321 By design, list ruleset command output may be used as input to nft -f.
322 Effectively, this is the nft-equivalent of iptables-save and
323 iptables-restore.
324
326 {add | create} table [family] table [{ flags flags ; }]
327 {delete | list | flush} table [family] table
328 list tables [family]
329 delete table [family] handle handle
330
331 Tables are containers for chains, sets and stateful objects. They are
332 identified by their address family and their name. The address family
333 must be one of ip, ip6, inet, arp, bridge, netdev. The inet address
334 family is a dummy family which is used to create hybrid IPv4/IPv6
335 tables. The meta expression nfproto keyword can be used to test which
336 family (ipv4 or ipv6) context the packet is being processed in. When no
337 address family is specified, ip is used by default. The only difference
338 between add and create is that the former will not return an error if
339 the specified table already exists while create will return an error.
340
341 Table 4. Table flags
342 ┌────────┬────────────────────────────┐
343 │Flag │ Description │
344 ├────────┼────────────────────────────┤
345 │ │ │
346 │dormant │ table is not evaluated any │
347 │ │ more (base chains are │
348 │ │ unregistered). │
349 └────────┴────────────────────────────┘
350
351 Add, change, delete a table.
352
353 # start nft in interactive mode
354 nft --interactive
355
356 # create a new table.
357 create table inet mytable
358
359 # add a new base chain: get input packets
360 add chain inet mytable myin { type filter hook input priority 0; }
361
362 # add a single counter to the chain
363 add rule inet mytable myin counter
364
365 # disable the table temporarily -- rules are not evaluated anymore
366 add table inet mytable { flags dormant; }
367
368 # make table active again:
369 add table inet mytable
370
371
372
373 add Add a new table for the
374 given family with the
375 given name.
376
377 delete Delete the specified
378 table.
379
380 list List all chains and rules
381 of the specified table.
382
383 flush Flush all chains and rules
384 of the specified table.
385
386
388 {add | create} chain [family] table chain [{ type type hook hook [device device] priority priority ; [policy policy ;] }]
389 {delete | list | flush} chain [family] table chain
390 list chains [family]
391 delete chain [family] table handle handle
392 rename chain [family] table chain newname
393
394 Chains are containers for rules. They exist in two kinds, base chains
395 and regular chains. A base chain is an entry point for packets from the
396 networking stack, a regular chain may be used as jump target and is
397 used for better rule organization.
398
399
400 add Add a new chain in the
401 specified table. When a
402 hook and priority value
403 are specified, the chain
404 is created as a base chain
405 and hooked up to the
406 networking stack.
407
408 create Similar to the add
409 command, but returns an
410 error if the chain already
411 exists.
412
413 delete Delete the specified
414 chain. The chain must not
415 contain any rules or be
416 used as jump target.
417
418 rename Rename the specified
419 chain.
420
421 list List all rules of the
422 specified chain.
423
424 flush Flush all rules of the
425 specified chain.
426
427
428 For base chains, type, hook and priority parameters are mandatory.
429
430 Table 5. Supported chain types
431 ┌───────┬───────────────┬────────────────┬──────────────────┐
432 │Type │ Families │ Hooks │ Description │
433 ├───────┼───────────────┼────────────────┼──────────────────┤
434 │ │ │ │ │
435 │filter │ all │ all │ Standard chain │
436 │ │ │ │ type to use in │
437 │ │ │ │ doubt. │
438 ├───────┼───────────────┼────────────────┼──────────────────┤
439 │ │ │ │ │
440 │nat │ ip, ip6, inet │ prerouting, │ Chains of this │
441 │ │ │ input, output, │ type perform │
442 │ │ │ postrouting │ Native Address │
443 │ │ │ │ Translation │
444 │ │ │ │ based on │
445 │ │ │ │ conntrack │
446 │ │ │ │ entries. Only │
447 │ │ │ │ the first packet │
448 │ │ │ │ of a connection │
449 │ │ │ │ actually │
450 │ │ │ │ traverses this │
451 │ │ │ │ chain - its │
452 │ │ │ │ rules usually │
453 │ │ │ │ define details │
454 │ │ │ │ of the created │
455 │ │ │ │ conntrack entry │
456 │ │ │ │ (NAT statements │
457 │ │ │ │ for instance). │
458 ├───────┼───────────────┼────────────────┼──────────────────┤
459 │ │ │ │ │
460 │route │ ip, ip6 │ output │ If a packet has │
461 │ │ │ │ traversed a │
462 │ │ │ │ chain of this │
463 │ │ │ │ type and is │
464 │ │ │ │ about to be │
465 │ │ │ │ accepted, a new │
466 │ │ │ │ route lookup is │
467 │ │ │ │ performed if │
468 │ │ │ │ relevant parts │
469 │ │ │ │ of the IP header │
470 │ │ │ │ have changed. │
471 │ │ │ │ This allows to │
472 │ │ │ │ e.g. implement │
473 │ │ │ │ policy routing │
474 │ │ │ │ selectors in │
475 │ │ │ │ nftables. │
476 └───────┴───────────────┴────────────────┴──────────────────┘
477
478 Apart from the special cases illustrated above (e.g. nat type not
479 supporting forward hook or route type only supporting output hook),
480 there are three further quirks worth noticing:
481
482 • The netdev family supports merely a single combination, namely
483 filter type and ingress hook. Base chains in this family also
484 require the device parameter to be present since they exist per
485 incoming interface only.
486
487 • The arp family supports only the input and output hooks, both in
488 chains of type filter.
489
490 • The inet family also supports the ingress hook (since Linux kernel
491 5.10), to filter IPv4 and IPv6 packet at the same location as the
492 netdev ingress hook. This inet hook allows you to share sets and
493 maps between the usual prerouting, input, forward, output,
494 postrouting and this ingress hook.
495
496 The priority parameter accepts a signed integer value or a standard
497 priority name which specifies the order in which chains with same hook
498 value are traversed. The ordering is ascending, i.e. lower priority
499 values have precedence over higher ones.
500
501 Standard priority values can be replaced with easily memorizable names.
502 Not all names make sense in every family with every hook (see the
503 compatibility matrices below) but their numerical value can still be
504 used for prioritizing chains.
505
506 These names and values are defined and made available based on what
507 priorities are used by xtables when registering their default chains.
508
509 Most of the families use the same values, but bridge uses different
510 ones from the others. See the following tables that describe the values
511 and compatibility.
512
513 Table 6. Standard priority names, family and hook compatibility matrix
514 ┌─────────┬───────┬────────────────┬─────────────┐
515 │Name │ Value │ Families │ Hooks │
516 ├─────────┼───────┼────────────────┼─────────────┤
517 │ │ │ │ │
518 │raw │ -300 │ ip, ip6, inet │ all │
519 ├─────────┼───────┼────────────────┼─────────────┤
520 │ │ │ │ │
521 │mangle │ -150 │ ip, ip6, inet │ all │
522 ├─────────┼───────┼────────────────┼─────────────┤
523 │ │ │ │ │
524 │dstnat │ -100 │ ip, ip6, inet │ prerouting │
525 ├─────────┼───────┼────────────────┼─────────────┤
526 │ │ │ │ │
527 │filter │ 0 │ ip, ip6, inet, │ all │
528 │ │ │ arp, netdev │ │
529 ├─────────┼───────┼────────────────┼─────────────┤
530 │ │ │ │ │
531 │security │ 50 │ ip, ip6, inet │ all │
532 ├─────────┼───────┼────────────────┼─────────────┤
533 │ │ │ │ │
534 │srcnat │ 100 │ ip, ip6, inet │ postrouting │
535 └─────────┴───────┴────────────────┴─────────────┘
536
537 Table 7. Standard priority names and hook compatibility for the bridge
538 family
539 ┌───────┬───────┬─────────────┐
540 │ │ │ │
541 │Name │ Value │ Hooks │
542 ├───────┼───────┼─────────────┤
543 │ │ │ │
544 │dstnat │ -300 │ prerouting │
545 ├───────┼───────┼─────────────┤
546 │ │ │ │
547 │filter │ -200 │ all │
548 ├───────┼───────┼─────────────┤
549 │ │ │ │
550 │out │ 100 │ output │
551 ├───────┼───────┼─────────────┤
552 │ │ │ │
553 │srcnat │ 300 │ postrouting │
554 └───────┴───────┴─────────────┘
555
556 Basic arithmetic expressions (addition and subtraction) can also be
557 achieved with these standard names to ease relative prioritizing, e.g.
558 mangle - 5 stands for -155. Values will also be printed like this until
559 the value is not further than 10 form the standard value.
560
561 Base chains also allow to set the chain’s policy, i.e. what happens to
562 packets not explicitly accepted or refused in contained rules.
563 Supported policy values are accept (which is the default) or drop.
564
566 {add | insert} rule [family] table chain [handle handle | index index] statement ... [comment comment]
567 replace rule [family] table chain handle handle statement ... [comment comment]
568 delete rule [family] table chain handle handle
569
570 Rules are added to chains in the given table. If the family is not
571 specified, the ip family is used. Rules are constructed from two kinds
572 of components according to a set of grammatical rules: expressions and
573 statements.
574
575 The add and insert commands support an optional location specifier,
576 which is either a handle or the index (starting at zero) of an existing
577 rule. Internally, rule locations are always identified by handle and
578 the translation from index happens in userspace. This has two potential
579 implications in case a concurrent ruleset change happens after the
580 translation was done: The effective rule index might change if a rule
581 was inserted or deleted before the referred one. If the referred rule
582 was deleted, the command is rejected by the kernel just as if an
583 invalid handle was given.
584
585 A comment is a single word or a double-quoted (") multi-word string
586 which can be used to make notes regarding the actual rule. Note: If you
587 use bash for adding rules, you have to escape the quotation marks, e.g.
588 \"enable ssh for servers\".
589
590
591
592
593
594
595
596
597
598 add Add a new rule described
599 by the list of statements.
600 The rule is appended to
601 the given chain unless a
602 location is specified, in
603 which case the rule is
604 inserted after the
605 specified rule.
606
607 insert Same as add except the
608 rule is inserted at the
609 beginning of the chain or
610 before the specified rule.
611
612 replace Similar to add, but the
613 rule replaces the
614 specified rule.
615
616 delete Delete the specified rule.
617
618
619 add a rule to ip table output chain.
620
621 nft add rule filter output ip daddr 192.168.0.0/24 accept # 'ip filter' is assumed
622 # same command, slightly more verbose
623 nft add rule ip filter output ip daddr 192.168.0.0/24 accept
624
625 delete rule from inet table.
626
627 # nft -a list ruleset
628 table inet filter {
629 chain input {
630 type filter hook input priority 0; policy accept;
631 ct state established,related accept # handle 4
632 ip saddr 10.1.1.1 tcp dport ssh accept # handle 5
633 ...
634 # delete the rule with handle 5
635 # nft delete rule inet filter input handle 5
636
637
639 nftables offers two kinds of set concepts. Anonymous sets are sets that
640 have no specific name. The set members are enclosed in curly braces,
641 with commas to separate elements when creating the rule the set is used
642 in. Once that rule is removed, the set is removed as well. They cannot
643 be updated, i.e. once an anonymous set is declared it cannot be changed
644 anymore except by removing/altering the rule that uses the anonymous
645 set.
646
647 Using anonymous sets to accept particular subnets and ports.
648
649 nft add rule filter input ip saddr { 10.0.0.0/8, 192.168.0.0/16 } tcp dport { 22, 443 } accept
650
651 Named sets are sets that need to be defined first before they can be
652 referenced in rules. Unlike anonymous sets, elements can be added to or
653 removed from a named set at any time. Sets are referenced from rules
654 using an @ prefixed to the sets name.
655
656 Using named sets to accept addresses and ports.
657
658 nft add rule filter input ip saddr @allowed_hosts tcp dport @allowed_ports accept
659
660 The sets allowed_hosts and allowed_ports need to be created first. The
661 next section describes nft set syntax in more detail.
662
663 add set [family] table set { type type | typeof expression ; [flags flags ;] [timeout timeout ;] [gc-interval gc-interval ;] [elements = { element[, ...] } ;] [size size ;] [policy policy ;] [auto-merge ;] }
664 {delete | list | flush} set [family] table set
665 list sets [family]
666 delete set [family] table handle handle
667 {add | delete} element [family] table set { element[, ...] }
668
669 Sets are element containers of a user-defined data type, they are
670 uniquely identified by a user-defined name and attached to tables.
671 Their behaviour can be tuned with the flags that can be specified at
672 set creation time.
673
674
675 add Add a new set in the
676 specified table. See the
677 Set specification table
678 below for more information
679 about how to specify a
680 sets properties.
681
682 delete Delete the specified set.
683
684 list Display the elements in
685 the specified set.
686
687 flush Remove all elements from
688 the specified set.
689
690
691 Table 8. Set specifications
692 ┌────────────┬──────────────────────┬─────────────────────┐
693 │Keyword │ Description │ Type │
694 ├────────────┼──────────────────────┼─────────────────────┤
695 │ │ │ │
696 │type │ data type of set │ string: ipv4_addr, │
697 │ │ elements │ ipv6_addr, │
698 │ │ │ ether_addr, │
699 │ │ │ inet_proto, │
700 │ │ │ inet_service, mark │
701 ├────────────┼──────────────────────┼─────────────────────┤
702 │ │ │ │
703 │typeof │ data type of set │ expression to │
704 │ │ element │ derive the data │
705 │ │ │ type from │
706 ├────────────┼──────────────────────┼─────────────────────┤
707 │ │ │ │
708 │flags │ set flags │ string: constant, │
709 │ │ │ dynamic, interval, │
710 │ │ │ timeout │
711 ├────────────┼──────────────────────┼─────────────────────┤
712 │ │ │ │
713 │timeout │ time an element │ string, decimal │
714 │ │ stays in the set, │ followed by unit. │
715 │ │ mandatory if set is │ Units are: d, h, m, │
716 │ │ added to from the │ s │
717 │ │ packet path │ │
718 │ │ (ruleset). │ │
719 ├────────────┼──────────────────────┼─────────────────────┤
720 │ │ │ │
721 │gc-interval │ garbage collection │ string, decimal │
722 │ │ interval, only │ followed by unit. │
723 │ │ available when │ Units are: d, h, m, │
724 │ │ timeout or flag │ s │
725 │ │ timeout are active │ │
726 ├────────────┼──────────────────────┼─────────────────────┤
727 │ │ │ │
728 │elements │ elements contained │ set data type │
729 │ │ by the set │ │
730 ├────────────┼──────────────────────┼─────────────────────┤
731 │ │ │ │
732 │size │ maximum number of │ unsigned integer │
733 │ │ elements in the │ (64 bit) │
734 │ │ set, mandatory if │ │
735 │ │ set is added to │ │
736 │ │ from the packet │ │
737 │ │ path (ruleset). │ │
738 ├────────────┼──────────────────────┼─────────────────────┤
739 │ │ │ │
740 │policy │ set policy │ string: performance │
741 │ │ │ [default], memory │
742 ├────────────┼──────────────────────┼─────────────────────┤
743 │ │ │ │
744 │auto-merge │ automatic merge of │ │
745 │ │ adjacent/overlapping │ │
746 │ │ set elements (only │ │
747 │ │ for interval sets) │ │
748 └────────────┴──────────────────────┴─────────────────────┘
749
751 add map [family] table map { type type | typeof expression [flags flags ;] [elements = { element[, ...] } ;] [size size ;] [policy policy ;] }
752 {delete | list | flush} map [family] table map
753 list maps [family]
754
755 Maps store data based on some specific key used as input. They are
756 uniquely identified by a user-defined name and attached to tables.
757
758
759 add Add a new map in the
760 specified table.
761
762 delete Delete the specified map.
763
764 list Display the elements in
765 the specified map.
766
767 flush Remove all elements from
768 the specified map.
769
770 add element Comma-separated list of
771 elements to add into the
772 specified map.
773
774 delete element Comma-separated list of
775 element keys to delete
776 from the specified map.
777
778
779 Table 9. Map specifications
780 ┌─────────┬─────────────────────┬─────────────────────┐
781 │Keyword │ Description │ Type │
782 ├─────────┼─────────────────────┼─────────────────────┤
783 │ │ │ │
784 │type │ data type of map │ string: ipv4_addr, │
785 │ │ elements │ ipv6_addr, │
786 │ │ │ ether_addr, │
787 │ │ │ inet_proto, │
788 │ │ │ inet_service, mark, │
789 │ │ │ counter, quota. │
790 │ │ │ Counter and quota │
791 │ │ │ can’t be used as │
792 │ │ │ keys │
793 ├─────────┼─────────────────────┼─────────────────────┤
794 │ │ │ │
795 │typeof │ data type of set │ expression to │
796 │ │ element │ derive the data │
797 │ │ │ type from │
798 ├─────────┼─────────────────────┼─────────────────────┤
799 │ │ │ │
800 │flags │ map flags │ string: constant, │
801 │ │ │ interval │
802 ├─────────┼─────────────────────┼─────────────────────┤
803 │ │ │ │
804 │elements │ elements contained │ map data type │
805 │ │ by the map │ │
806 ├─────────┼─────────────────────┼─────────────────────┤
807 │ │ │ │
808 │size │ maximum number of │ unsigned integer │
809 │ │ elements in the map │ (64 bit) │
810 ├─────────┼─────────────────────┼─────────────────────┤
811 │ │ │ │
812 │policy │ map policy │ string: performance │
813 │ │ │ [default], memory │
814 └─────────┴─────────────────────┴─────────────────────┘
815
817 {add | create | delete | get } element [family] table set { ELEMENT[, ...] }
818
819 ELEMENT := key_expression OPTIONS [: value_expression]
820 OPTIONS := [timeout TIMESPEC] [expires TIMESPEC] [comment string]
821 TIMESPEC := [numd][numh][numm][num[s]]
822
823 Element-related commands allow to change contents of named sets and
824 maps. key_expression is typically a value matching the set type.
825 value_expression is not allowed in sets but mandatory when adding to
826 maps, where it matches the data part in it’s type definition. When
827 deleting from maps, it may be specified but is optional as
828 key_expression uniquely identifies the element.
829
830 create command is similar to add with the exception that none of the
831 listed elements may already exist.
832
833 get command is useful to check if an element is contained in a set
834 which may be non-trivial in very large and/or interval sets. In the
835 latter case, the containing interval is returned instead of just the
836 element itself.
837
838 Table 10. Element options
839 ┌────────┬───────────────────────────┐
840 │Option │ Description │
841 ├────────┼───────────────────────────┤
842 │ │ │
843 │timeout │ timeout value for │
844 │ │ sets/maps with flag │
845 │ │ timeout │
846 ├────────┼───────────────────────────┤
847 │ │ │
848 │expires │ the time until given │
849 │ │ element expires, useful │
850 │ │ for ruleset replication │
851 │ │ only │
852 ├────────┼───────────────────────────┤
853 │ │ │
854 │comment │ per element comment field │
855 └────────┴───────────────────────────┘
856
858 {add | create} flowtable [family] table flowtable { hook hook priority priority ; devices = { device[, ...] } ; }
859 list flowtables [family]
860 {delete | list} flowtable [family] table flowtable
861 delete flowtable [family] table handle handle
862
863 Flowtables allow you to accelerate packet forwarding in software.
864 Flowtables entries are represented through a tuple that is composed of
865 the input interface, source and destination address, source and
866 destination port; and layer 3/4 protocols. Each entry also caches the
867 destination interface and the gateway address - to update the
868 destination link-layer address - to forward packets. The ttl and
869 hoplimit fields are also decremented. Hence, flowtables provides an
870 alternative path that allow packets to bypass the classic forwarding
871 path. Flowtables reside in the ingress hook that is located before the
872 prerouting hook. You can select which flows you want to offload through
873 the flow expression from the forward chain. Flowtables are identified
874 by their address family and their name. The address family must be one
875 of ip, ip6, or inet. The inet address family is a dummy family which is
876 used to create hybrid IPv4/IPv6 tables. When no address family is
877 specified, ip is used by default.
878
879 The priority can be a signed integer or filter which stands for 0.
880 Addition and subtraction can be used to set relative priority, e.g.
881 filter + 5 equals to 5.
882
883
884 add Add a new flowtable for
885 the given family with the
886 given name.
887
888 delete Delete the specified
889 flowtable.
890
891 list List all flowtables.
892
893
895 {add | delete | list | reset} type [family] table object
896 delete type [family] table handle handle
897 list counters [family]
898 list quotas [family]
899
900 Stateful objects are attached to tables and are identified by an unique
901 name. They group stateful information from rules, to reference them in
902 rules the keywords "type name" are used e.g. "counter name".
903
904
905 add Add a new stateful object
906 in the specified table.
907
908 delete Delete the specified
909 object.
910
911 list Display stateful
912 information the object
913 holds.
914
915 reset List-and-reset stateful
916 object.
917
918
919 CT HELPER
920 ct helper helper { type type protocol protocol ; [l3proto family ;] }
921
922 Ct helper is used to define connection tracking helpers that can then
923 be used in combination with the ct helper set statement. type and
924 protocol are mandatory, l3proto is derived from the table family by
925 default, i.e. in the inet table the kernel will try to load both the
926 ipv4 and ipv6 helper backends, if they are supported by the kernel.
927
928 Table 11. conntrack helper specifications
929 ┌─────────┬─────────────────────┬─────────────────────┐
930 │Keyword │ Description │ Type │
931 ├─────────┼─────────────────────┼─────────────────────┤
932 │ │ │ │
933 │type │ name of helper type │ quoted string (e.g. │
934 │ │ │ "ftp") │
935 ├─────────┼─────────────────────┼─────────────────────┤
936 │ │ │ │
937 │protocol │ layer 4 protocol of │ string (e.g. ip) │
938 │ │ the helper │ │
939 ├─────────┼─────────────────────┼─────────────────────┤
940 │ │ │ │
941 │l3proto │ layer 3 protocol of │ address family │
942 │ │ the helper │ (e.g. ip) │
943 └─────────┴─────────────────────┴─────────────────────┘
944
945 defining and assigning ftp helper.
946
947 Unlike iptables, helper assignment needs to be performed after the conntrack
948 lookup has completed, for example with the default 0 hook priority.
949
950 table inet myhelpers {
951 ct helper ftp-standard {
952 type "ftp" protocol tcp
953 }
954 chain prerouting {
955 type filter hook prerouting priority 0;
956 tcp dport 21 ct helper set "ftp-standard"
957 }
958 }
959
960
961 CT TIMEOUT
962 ct timeout name { protocol protocol ; policy = { state: value [, ...] } ; [l3proto family ;] }
963
964 Ct timeout is used to update connection tracking timeout values.Timeout
965 policies are assigned with the ct timeout set statement. protocol and
966 policy are mandatory, l3proto is derived from the table family by
967 default.
968
969 Table 12. conntrack timeout specifications
970 ┌─────────┬─────────────────────┬──────────────────┐
971 │Keyword │ Description │ Type │
972 ├─────────┼─────────────────────┼──────────────────┤
973 │ │ │ │
974 │protocol │ layer 4 protocol of │ string (e.g. ip) │
975 │ │ the timeout object │ │
976 ├─────────┼─────────────────────┼──────────────────┤
977 │ │ │ │
978 │state │ connection state │ string (e.g. │
979 │ │ name │ "established") │
980 ├─────────┼─────────────────────┼──────────────────┤
981 │ │ │ │
982 │value │ timeout value for │ unsigned integer │
983 │ │ connection state │ │
984 ├─────────┼─────────────────────┼──────────────────┤
985 │ │ │ │
986 │l3proto │ layer 3 protocol of │ address family │
987 │ │ the timeout object │ (e.g. ip) │
988 └─────────┴─────────────────────┴──────────────────┘
989
990 defining and assigning ct timeout policy.
991
992 table ip filter {
993 ct timeout customtimeout {
994 protocol tcp;
995 l3proto ip
996 policy = { established: 120, close: 20 }
997 }
998
999 chain output {
1000 type filter hook output priority filter; policy accept;
1001 ct timeout set "customtimeout"
1002 }
1003 }
1004
1005 testing the updated timeout policy.
1006
1007 % conntrack -E
1008
1009 It should display:
1010
1011 [UPDATE] tcp 6 120 ESTABLISHED src=172.16.19.128 dst=172.16.19.1
1012 sport=22 dport=41360 [UNREPLIED] src=172.16.19.1 dst=172.16.19.128
1013 sport=41360 dport=22
1014
1015
1016 CT EXPECTATION
1017 ct expectation name { protocol protocol ; dport dport ; timeout timeout ; size size ; [*l3proto family ;] }
1018
1019 Ct expectation is used to create connection expectations. Expectations
1020 are assigned with the ct expectation set statement. protocol, dport,
1021 timeout and size are mandatory, l3proto is derived from the table
1022 family by default.
1023
1024 Table 13. conntrack expectation specifications
1025 ┌─────────┬─────────────────────┬──────────────────┐
1026 │Keyword │ Description │ Type │
1027 ├─────────┼─────────────────────┼──────────────────┤
1028 │ │ │ │
1029 │protocol │ layer 4 protocol of │ string (e.g. ip) │
1030 │ │ the expectation │ │
1031 │ │ object │ │
1032 ├─────────┼─────────────────────┼──────────────────┤
1033 │ │ │ │
1034 │dport │ destination port of │ unsigned integer │
1035 │ │ expected connection │ │
1036 ├─────────┼─────────────────────┼──────────────────┤
1037 │ │ │ │
1038 │timeout │ timeout value for │ unsigned integer │
1039 │ │ expectation │ │
1040 ├─────────┼─────────────────────┼──────────────────┤
1041 │ │ │ │
1042 │size │ size value for │ unsigned integer │
1043 │ │ expectation │ │
1044 ├─────────┼─────────────────────┼──────────────────┤
1045 │ │ │ │
1046 │l3proto │ layer 3 protocol of │ address family │
1047 │ │ the expectation │ (e.g. ip) │
1048 │ │ object │ │
1049 └─────────┴─────────────────────┴──────────────────┘
1050
1051 defining and assigning ct expectation policy.
1052
1053 table ip filter {
1054 ct expectation expect {
1055 protocol udp
1056 dport 9876
1057 timeout 2m
1058 size 8
1059 l3proto ip
1060 }
1061
1062 chain input {
1063 type filter hook input priority filter; policy accept;
1064 ct expectation set "expect"
1065 }
1066 }
1067
1068
1069 COUNTER
1070 counter [packets bytes]
1071
1072 Table 14. Counter specifications
1073 ┌────────┬──────────────────┬──────────────────┐
1074 │Keyword │ Description │ Type │
1075 ├────────┼──────────────────┼──────────────────┤
1076 │ │ │ │
1077 │packets │ initial count of │ unsigned integer │
1078 │ │ packets │ (64 bit) │
1079 ├────────┼──────────────────┼──────────────────┤
1080 │ │ │ │
1081 │bytes │ initial count of │ unsigned integer │
1082 │ │ bytes │ (64 bit) │
1083 └────────┴──────────────────┴──────────────────┘
1084
1085 QUOTA
1086 quota [over | until] [used]
1087
1088 Table 15. Quota specifications
1089 ┌────────┬───────────────────┬────────────────────┐
1090 │Keyword │ Description │ Type │
1091 ├────────┼───────────────────┼────────────────────┤
1092 │ │ │ │
1093 │quota │ quota limit, used │ Two arguments, │
1094 │ │ as the quota name │ unsigned integer │
1095 │ │ │ (64 bit) and │
1096 │ │ │ string: bytes, │
1097 │ │ │ kbytes, mbytes. │
1098 │ │ │ "over" and "until" │
1099 │ │ │ go before these │
1100 │ │ │ arguments │
1101 ├────────┼───────────────────┼────────────────────┤
1102 │ │ │ │
1103 │used │ initial value of │ Two arguments, │
1104 │ │ used quota │ unsigned integer │
1105 │ │ │ (64 bit) and │
1106 │ │ │ string: bytes, │
1107 │ │ │ kbytes, mbytes │
1108 └────────┴───────────────────┴────────────────────┘
1109
1111 Expressions represent values, either constants like network addresses,
1112 port numbers, etc., or data gathered from the packet during ruleset
1113 evaluation. Expressions can be combined using binary, logical,
1114 relational and other types of expressions to form complex or relational
1115 (match) expressions. They are also used as arguments to certain types
1116 of operations, like NAT, packet marking etc.
1117
1118 Each expression has a data type, which determines the size, parsing and
1119 representation of symbolic values and type compatibility with other
1120 expressions.
1121
1122 DESCRIBE COMMAND
1123 describe expression | data type
1124
1125 The describe command shows information about the type of an expression
1126 and its data type. A data type may also be given, in which nft will
1127 display more information about the type.
1128
1129 The describe command.
1130
1131 $ nft describe tcp flags
1132 payload expression, datatype tcp_flag (TCP flag) (basetype bitmask, integer), 8 bits
1133
1134 predefined symbolic constants:
1135 fin 0x01
1136 syn 0x02
1137 rst 0x04
1138 psh 0x08
1139 ack 0x10
1140 urg 0x20
1141 ecn 0x40
1142 cwr 0x80
1143
1144
1146 Data types determine the size, parsing and representation of symbolic
1147 values and type compatibility of expressions. A number of global data
1148 types exist, in addition some expression types define further data
1149 types specific to the expression type. Most data types have a fixed
1150 size, some however may have a dynamic size, f.i. the string type. Some
1151 types also have predefined symbolic constants. Those can be listed
1152 using the nft describe command:
1153
1154 $ nft describe ct_state
1155 datatype ct_state (conntrack state) (basetype bitmask, integer), 32 bits
1156
1157 pre-defined symbolic constants (in hexadecimal):
1158 invalid 0x00000001
1159 new ...
1160
1161 Types may be derived from lower order types, f.i. the IPv4 address type
1162 is derived from the integer type, meaning an IPv4 address can also be
1163 specified as an integer value.
1164
1165 In certain contexts (set and map definitions), it is necessary to
1166 explicitly specify a data type. Each type has a name which is used for
1167 this.
1168
1169 INTEGER TYPE
1170 ┌────────┬─────────┬──────────┬───────────┐
1171 │Name │ Keyword │ Size │ Base type │
1172 ├────────┼─────────┼──────────┼───────────┤
1173 │ │ │ │ │
1174 │Integer │ integer │ variable │ - │
1175 └────────┴─────────┴──────────┴───────────┘
1176
1177 The integer type is used for numeric values. It may be specified as a
1178 decimal, hexadecimal or octal number. The integer type does not have a
1179 fixed size, its size is determined by the expression for which it is
1180 used.
1181
1182 BITMASK TYPE
1183 ┌────────┬─────────┬──────────┬───────────┐
1184 │Name │ Keyword │ Size │ Base type │
1185 ├────────┼─────────┼──────────┼───────────┤
1186 │ │ │ │ │
1187 │Bitmask │ bitmask │ variable │ integer │
1188 └────────┴─────────┴──────────┴───────────┘
1189
1190 The bitmask type (bitmask) is used for bitmasks.
1191
1192 STRING TYPE
1193 ┌───────┬─────────┬──────────┬───────────┐
1194 │Name │ Keyword │ Size │ Base type │
1195 ├───────┼─────────┼──────────┼───────────┤
1196 │ │ │ │ │
1197 │String │ string │ variable │ - │
1198 └───────┴─────────┴──────────┴───────────┘
1199
1200 The string type is used for character strings. A string begins with an
1201 alphabetic character (a-zA-Z) followed by zero or more alphanumeric
1202 characters or the characters /, -, _ and .. In addition, anything
1203 enclosed in double quotes (") is recognized as a string.
1204
1205 String specification.
1206
1207 # Interface name
1208 filter input iifname eth0
1209
1210 # Weird interface name
1211 filter input iifname "(eth0)"
1212
1213
1214 LINK LAYER ADDRESS TYPE
1215 ┌───────────┬─────────┬──────────┬───────────┐
1216 │Name │ Keyword │ Size │ Base type │
1217 ├───────────┼─────────┼──────────┼───────────┤
1218 │ │ │ │ │
1219 │Link layer │ lladdr │ variable │ integer │
1220 │address │ │ │ │
1221 └───────────┴─────────┴──────────┴───────────┘
1222
1223 The link layer address type is used for link layer addresses. Link
1224 layer addresses are specified as a variable amount of groups of two
1225 hexadecimal digits separated using colons (:).
1226
1227 Link layer address specification.
1228
1229 # Ethernet destination MAC address
1230 filter input ether daddr 20:c9:d0:43:12:d9
1231
1232
1233 IPV4 ADDRESS TYPE
1234 ┌─────────────┬───────────┬────────┬───────────┐
1235 │Name │ Keyword │ Size │ Base type │
1236 ├─────────────┼───────────┼────────┼───────────┤
1237 │ │ │ │ │
1238 │IPV4 address │ ipv4_addr │ 32 bit │ integer │
1239 └─────────────┴───────────┴────────┴───────────┘
1240
1241 The IPv4 address type is used for IPv4 addresses. Addresses are
1242 specified in either dotted decimal, dotted hexadecimal, dotted octal,
1243 decimal, hexadecimal, octal notation or as a host name. A host name
1244 will be resolved using the standard system resolver.
1245
1246 IPv4 address specification.
1247
1248 # dotted decimal notation
1249 filter output ip daddr 127.0.0.1
1250
1251 # host name
1252 filter output ip daddr localhost
1253
1254
1255 IPV6 ADDRESS TYPE
1256 ┌─────────────┬───────────┬─────────┬───────────┐
1257 │Name │ Keyword │ Size │ Base type │
1258 ├─────────────┼───────────┼─────────┼───────────┤
1259 │ │ │ │ │
1260 │IPv6 address │ ipv6_addr │ 128 bit │ integer │
1261 └─────────────┴───────────┴─────────┴───────────┘
1262
1263 The IPv6 address type is used for IPv6 addresses. Addresses are
1264 specified as a host name or as hexadecimal halfwords separated by
1265 colons. Addresses might be enclosed in square brackets ("[]") to
1266 differentiate them from port numbers.
1267
1268 IPv6 address specification.
1269
1270 # abbreviated loopback address
1271 filter output ip6 daddr ::1
1272
1273 IPv6 address specification with bracket notation.
1274
1275 # without [] the port number (22) would be parsed as part of the
1276 # ipv6 address
1277 ip6 nat prerouting tcp dport 2222 dnat to [1ce::d0]:22
1278
1279
1280 BOOLEAN TYPE
1281 ┌────────┬─────────┬───────┬───────────┐
1282 │Name │ Keyword │ Size │ Base type │
1283 ├────────┼─────────┼───────┼───────────┤
1284 │ │ │ │ │
1285 │Boolean │ boolean │ 1 bit │ integer │
1286 └────────┴─────────┴───────┴───────────┘
1287
1288 The boolean type is a syntactical helper type in userspace. Its use is
1289 in the right-hand side of a (typically implicit) relational expression
1290 to change the expression on the left-hand side into a boolean check
1291 (usually for existence).
1292
1293 Table 16. The following keywords will automatically resolve into a
1294 boolean type with given value
1295 ┌────────┬───────┐
1296 │Keyword │ Value │
1297 ├────────┼───────┤
1298 │ │ │
1299 │exists │ 1 │
1300 ├────────┼───────┤
1301 │ │ │
1302 │missing │ 0 │
1303 └────────┴───────┘
1304
1305 Table 17. expressions support a boolean comparison
1306 ┌───────────┬─────────────────────────┐
1307 │Expression │ Behaviour │
1308 ├───────────┼─────────────────────────┤
1309 │ │ │
1310 │fib │ Check route existence. │
1311 ├───────────┼─────────────────────────┤
1312 │ │ │
1313 │exthdr │ Check IPv6 extension │
1314 │ │ header existence. │
1315 ├───────────┼─────────────────────────┤
1316 │ │ │
1317 │tcp option │ Check TCP option header │
1318 │ │ existence. │
1319 └───────────┴─────────────────────────┘
1320
1321 Boolean specification.
1322
1323 # match if route exists
1324 filter input fib daddr . iif oif exists
1325
1326 # match only non-fragmented packets in IPv6 traffic
1327 filter input exthdr frag missing
1328
1329 # match if TCP timestamp option is present
1330 filter input tcp option timestamp exists
1331
1332
1333 ICMP TYPE TYPE
1334 ┌──────────┬───────────┬───────┬───────────┐
1335 │Name │ Keyword │ Size │ Base type │
1336 ├──────────┼───────────┼───────┼───────────┤
1337 │ │ │ │ │
1338 │ICMP Type │ icmp_type │ 8 bit │ integer │
1339 └──────────┴───────────┴───────┴───────────┘
1340
1341 The ICMP Type type is used to conveniently specify the ICMP header’s
1342 type field.
1343
1344 Table 18. Keywords may be used when specifying the ICMP type
1345 ┌────────────────────────┬───────┐
1346 │Keyword │ Value │
1347 ├────────────────────────┼───────┤
1348 │ │ │
1349 │echo-reply │ 0 │
1350 ├────────────────────────┼───────┤
1351 │ │ │
1352 │destination-unreachable │ 3 │
1353 ├────────────────────────┼───────┤
1354 │ │ │
1355 │source-quench │ 4 │
1356 ├────────────────────────┼───────┤
1357 │ │ │
1358 │redirect │ 5 │
1359 ├────────────────────────┼───────┤
1360 │ │ │
1361 │echo-request │ 8 │
1362 ├────────────────────────┼───────┤
1363 │ │ │
1364 │router-advertisement │ 9 │
1365 ├────────────────────────┼───────┤
1366 │ │ │
1367 │router-solicitation │ 10 │
1368 ├────────────────────────┼───────┤
1369 │ │ │
1370 │time-exceeded │ 11 │
1371 ├────────────────────────┼───────┤
1372 │ │ │
1373 │parameter-problem │ 12 │
1374 ├────────────────────────┼───────┤
1375 │ │ │
1376 │timestamp-request │ 13 │
1377 ├────────────────────────┼───────┤
1378 │ │ │
1379 │timestamp-reply │ 14 │
1380 ├────────────────────────┼───────┤
1381 │ │ │
1382 │info-request │ 15 │
1383 ├────────────────────────┼───────┤
1384 │ │ │
1385 │info-reply │ 16 │
1386 ├────────────────────────┼───────┤
1387 │ │ │
1388 │address-mask-request │ 17 │
1389 ├────────────────────────┼───────┤
1390 │ │ │
1391 │address-mask-reply │ 18 │
1392 └────────────────────────┴───────┘
1393
1394 ICMP Type specification.
1395
1396 # match ping packets
1397 filter output icmp type { echo-request, echo-reply }
1398
1399
1400 ICMP CODE TYPE
1401 ┌──────────┬───────────┬───────┬───────────┐
1402 │Name │ Keyword │ Size │ Base type │
1403 ├──────────┼───────────┼───────┼───────────┤
1404 │ │ │ │ │
1405 │ICMP Code │ icmp_code │ 8 bit │ integer │
1406 └──────────┴───────────┴───────┴───────────┘
1407
1408 The ICMP Code type is used to conveniently specify the ICMP header’s
1409 code field.
1410
1411 Table 19. Keywords may be used when specifying the ICMP code
1412 ┌─────────────────┬───────┐
1413 │Keyword │ Value │
1414 ├─────────────────┼───────┤
1415 │ │ │
1416 │net-unreachable │ 0 │
1417 ├─────────────────┼───────┤
1418 │ │ │
1419 │host-unreachable │ 1 │
1420 ├─────────────────┼───────┤
1421 │ │ │
1422 │prot-unreachable │ 2 │
1423 ├─────────────────┼───────┤
1424 │ │ │
1425 │port-unreachable │ 3 │
1426 ├─────────────────┼───────┤
1427 │ │ │
1428 │frag-needed │ 4 │
1429 ├─────────────────┼───────┤
1430 │ │ │
1431 │net-prohibited │ 9 │
1432 ├─────────────────┼───────┤
1433 │ │ │
1434 │host-prohibited │ 10 │
1435 ├─────────────────┼───────┤
1436 │ │ │
1437 │admin-prohibited │ 13 │
1438 └─────────────────┴───────┘
1439
1440 ICMPV6 TYPE TYPE
1441 ┌────────────┬────────────┬───────┬───────────┐
1442 │Name │ Keyword │ Size │ Base type │
1443 ├────────────┼────────────┼───────┼───────────┤
1444 │ │ │ │ │
1445 │ICMPv6 Type │ icmpx_code │ 8 bit │ integer │
1446 └────────────┴────────────┴───────┴───────────┘
1447
1448 The ICMPv6 Type type is used to conveniently specify the ICMPv6
1449 header’s type field.
1450
1451 Table 20. keywords may be used when specifying the ICMPv6 type:
1452 ┌────────────────────────┬───────┐
1453 │Keyword │ Value │
1454 ├────────────────────────┼───────┤
1455 │ │ │
1456 │destination-unreachable │ 1 │
1457 ├────────────────────────┼───────┤
1458 │ │ │
1459 │packet-too-big │ 2 │
1460 ├────────────────────────┼───────┤
1461 │ │ │
1462 │time-exceeded │ 3 │
1463 ├────────────────────────┼───────┤
1464 │ │ │
1465 │parameter-problem │ 4 │
1466 ├────────────────────────┼───────┤
1467 │ │ │
1468 │echo-request │ 128 │
1469 ├────────────────────────┼───────┤
1470 │ │ │
1471 │echo-reply │ 129 │
1472 ├────────────────────────┼───────┤
1473 │ │ │
1474 │mld-listener-query │ 130 │
1475 ├────────────────────────┼───────┤
1476 │ │ │
1477 │mld-listener-report │ 131 │
1478 ├────────────────────────┼───────┤
1479 │ │ │
1480 │mld-listener-done │ 132 │
1481 ├────────────────────────┼───────┤
1482 │ │ │
1483 │mld-listener-reduction │ 132 │
1484 ├────────────────────────┼───────┤
1485 │ │ │
1486 │nd-router-solicit │ 133 │
1487 ├────────────────────────┼───────┤
1488 │ │ │
1489 │nd-router-advert │ 134 │
1490 ├────────────────────────┼───────┤
1491 │ │ │
1492 │nd-neighbor-solicit │ 135 │
1493 ├────────────────────────┼───────┤
1494 │ │ │
1495 │nd-neighbor-advert │ 136 │
1496 ├────────────────────────┼───────┤
1497 │ │ │
1498 │nd-redirect │ 137 │
1499 ├────────────────────────┼───────┤
1500 │ │ │
1501 │router-renumbering │ 138 │
1502 ├────────────────────────┼───────┤
1503 │ │ │
1504 │ind-neighbor-solicit │ 141 │
1505 ├────────────────────────┼───────┤
1506 │ │ │
1507 │ind-neighbor-advert │ 142 │
1508 ├────────────────────────┼───────┤
1509 │ │ │
1510 │mld2-listener-report │ 143 │
1511 └────────────────────────┴───────┘
1512
1513 ICMPv6 Type specification.
1514
1515 # match ICMPv6 ping packets
1516 filter output icmpv6 type { echo-request, echo-reply }
1517
1518
1519 ICMPV6 CODE TYPE
1520 ┌────────────┬─────────────┬───────┬───────────┐
1521 │Name │ Keyword │ Size │ Base type │
1522 ├────────────┼─────────────┼───────┼───────────┤
1523 │ │ │ │ │
1524 │ICMPv6 Code │ icmpv6_code │ 8 bit │ integer │
1525 └────────────┴─────────────┴───────┴───────────┘
1526
1527 The ICMPv6 Code type is used to conveniently specify the ICMPv6
1528 header’s code field.
1529
1530 Table 21. keywords may be used when specifying the ICMPv6 code
1531 ┌─────────────────┬───────┐
1532 │Keyword │ Value │
1533 ├─────────────────┼───────┤
1534 │ │ │
1535 │no-route │ 0 │
1536 ├─────────────────┼───────┤
1537 │ │ │
1538 │admin-prohibited │ 1 │
1539 ├─────────────────┼───────┤
1540 │ │ │
1541 │addr-unreachable │ 3 │
1542 ├─────────────────┼───────┤
1543 │ │ │
1544 │port-unreachable │ 4 │
1545 ├─────────────────┼───────┤
1546 │ │ │
1547 │policy-fail │ 5 │
1548 ├─────────────────┼───────┤
1549 │ │ │
1550 │reject-route │ 6 │
1551 └─────────────────┴───────┘
1552
1553 ICMPVX CODE TYPE
1554 ┌────────────┬─────────────┬───────┬───────────┐
1555 │Name │ Keyword │ Size │ Base type │
1556 ├────────────┼─────────────┼───────┼───────────┤
1557 │ │ │ │ │
1558 │ICMPvX Code │ icmpv6_type │ 8 bit │ integer │
1559 └────────────┴─────────────┴───────┴───────────┘
1560
1561 The ICMPvX Code type abstraction is a set of values which overlap
1562 between ICMP and ICMPv6 Code types to be used from the inet family.
1563
1564 Table 22. keywords may be used when specifying the ICMPvX code
1565 ┌─────────────────┬───────┐
1566 │Keyword │ Value │
1567 ├─────────────────┼───────┤
1568 │ │ │
1569 │no-route │ 0 │
1570 ├─────────────────┼───────┤
1571 │ │ │
1572 │port-unreachable │ 1 │
1573 ├─────────────────┼───────┤
1574 │ │ │
1575 │host-unreachable │ 2 │
1576 ├─────────────────┼───────┤
1577 │ │ │
1578 │admin-prohibited │ 3 │
1579 └─────────────────┴───────┘
1580
1581 CONNTRACK TYPES
1582 Table 23. overview of types used in ct expression and statement
1583 ┌─────────────────┬───────────┬─────────┬───────────┐
1584 │Name │ Keyword │ Size │ Base type │
1585 ├─────────────────┼───────────┼─────────┼───────────┤
1586 │ │ │ │ │
1587 │conntrack state │ ct_state │ 4 byte │ bitmask │
1588 ├─────────────────┼───────────┼─────────┼───────────┤
1589 │ │ │ │ │
1590 │conntrack │ ct_dir │ 8 bit │ integer │
1591 │direction │ │ │ │
1592 ├─────────────────┼───────────┼─────────┼───────────┤
1593 │ │ │ │ │
1594 │conntrack status │ ct_status │ 4 byte │ bitmask │
1595 ├─────────────────┼───────────┼─────────┼───────────┤
1596 │ │ │ │ │
1597 │conntrack event │ ct_event │ 4 byte │ bitmask │
1598 │bits │ │ │ │
1599 ├─────────────────┼───────────┼─────────┼───────────┤
1600 │ │ │ │ │
1601 │conntrack label │ ct_label │ 128 bit │ bitmask │
1602 └─────────────────┴───────────┴─────────┴───────────┘
1603
1604 For each of the types above, keywords are available for convenience:
1605
1606 Table 24. conntrack state (ct_state)
1607 ┌────────────┬───────┐
1608 │Keyword │ Value │
1609 ├────────────┼───────┤
1610 │ │ │
1611 │invalid │ 1 │
1612 ├────────────┼───────┤
1613 │ │ │
1614 │established │ 2 │
1615 ├────────────┼───────┤
1616 │ │ │
1617 │related │ 4 │
1618 ├────────────┼───────┤
1619 │ │ │
1620 │new │ 8 │
1621 ├────────────┼───────┤
1622 │ │ │
1623 │untracked │ 64 │
1624 └────────────┴───────┘
1625
1626 Table 25. conntrack direction (ct_dir)
1627 ┌─────────┬───────┐
1628 │Keyword │ Value │
1629 ├─────────┼───────┤
1630 │ │ │
1631 │original │ 0 │
1632 ├─────────┼───────┤
1633 │ │ │
1634 │reply │ 1 │
1635 └─────────┴───────┘
1636
1637 Table 26. conntrack status (ct_status)
1638 ┌───────────┬───────┐
1639 │Keyword │ Value │
1640 ├───────────┼───────┤
1641 │ │ │
1642 │expected │ 1 │
1643 ├───────────┼───────┤
1644 │ │ │
1645 │seen-reply │ 2 │
1646 ├───────────┼───────┤
1647 │ │ │
1648 │assured │ 4 │
1649 ├───────────┼───────┤
1650 │ │ │
1651 │confirmed │ 8 │
1652 ├───────────┼───────┤
1653 │ │ │
1654 │snat │ 16 │
1655 ├───────────┼───────┤
1656 │ │ │
1657 │dnat │ 32 │
1658 ├───────────┼───────┤
1659 │ │ │
1660 │dying │ 512 │
1661 └───────────┴───────┘
1662
1663 Table 27. conntrack event bits (ct_event)
1664 ┌──────────┬───────┐
1665 │Keyword │ Value │
1666 ├──────────┼───────┤
1667 │ │ │
1668 │new │ 1 │
1669 ├──────────┼───────┤
1670 │ │ │
1671 │related │ 2 │
1672 ├──────────┼───────┤
1673 │ │ │
1674 │destroy │ 4 │
1675 ├──────────┼───────┤
1676 │ │ │
1677 │reply │ 8 │
1678 ├──────────┼───────┤
1679 │ │ │
1680 │assured │ 16 │
1681 ├──────────┼───────┤
1682 │ │ │
1683 │protoinfo │ 32 │
1684 ├──────────┼───────┤
1685 │ │ │
1686 │helper │ 64 │
1687 ├──────────┼───────┤
1688 │ │ │
1689 │mark │ 128 │
1690 ├──────────┼───────┤
1691 │ │ │
1692 │seqadj │ 256 │
1693 ├──────────┼───────┤
1694 │ │ │
1695 │secmark │ 512 │
1696 ├──────────┼───────┤
1697 │ │ │
1698 │label │ 1024 │
1699 └──────────┴───────┘
1700
1701 Possible keywords for conntrack label type (ct_label) are read at
1702 runtime from /etc/connlabel.conf.
1703
1704 DCCP PKTTYPE TYPE
1705 ┌─────────────────┬──────────────┬───────┬───────────┐
1706 │Name │ Keyword │ Size │ Base type │
1707 ├─────────────────┼──────────────┼───────┼───────────┤
1708 │ │ │ │ │
1709 │DCCP packet type │ dccp_pkttype │ 4 bit │ integer │
1710 └─────────────────┴──────────────┴───────┴───────────┘
1711
1712 The DCCP packet type abstracts the different legal values of the
1713 respective four bit field in the DCCP header, as stated by RFC4340.
1714 Note that possible values 10-15 are considered reserved and therefore
1715 not allowed to be used. In iptables' dccp match, these values are
1716 aliased INVALID. With nftables, one may simply match on the numeric
1717 value range, i.e. 10-15.
1718
1719 Table 28. keywords may be used when specifying the DCCP packet type
1720 ┌─────────┬───────┐
1721 │Keyword │ Value │
1722 ├─────────┼───────┤
1723 │ │ │
1724 │request │ 0 │
1725 ├─────────┼───────┤
1726 │ │ │
1727 │response │ 1 │
1728 ├─────────┼───────┤
1729 │ │ │
1730 │data │ 2 │
1731 ├─────────┼───────┤
1732 │ │ │
1733 │ack │ 3 │
1734 ├─────────┼───────┤
1735 │ │ │
1736 │dataack │ 4 │
1737 ├─────────┼───────┤
1738 │ │ │
1739 │closereq │ 5 │
1740 ├─────────┼───────┤
1741 │ │ │
1742 │close │ 6 │
1743 ├─────────┼───────┤
1744 │ │ │
1745 │reset │ 7 │
1746 ├─────────┼───────┤
1747 │ │ │
1748 │sync │ 8 │
1749 ├─────────┼───────┤
1750 │ │ │
1751 │syncack │ 9 │
1752 └─────────┴───────┘
1753
1755 The lowest order expression is a primary expression, representing
1756 either a constant or a single datum from a packet’s payload, meta data
1757 or a stateful module.
1758
1759 META EXPRESSIONS
1760 meta {length | nfproto | l4proto | protocol | priority}
1761 [meta] {mark | iif | iifname | iiftype | oif | oifname | oiftype | skuid | skgid | nftrace | rtclassid | ibrname | obrname | pkttype | cpu | iifgroup | oifgroup | cgroup | random | ipsec | iifkind | oifkind | time | hour | day }
1762
1763 A meta expression refers to meta data associated with a packet.
1764
1765 There are two types of meta expressions: unqualified and qualified meta
1766 expressions. Qualified meta expressions require the meta keyword before
1767 the meta key, unqualified meta expressions can be specified by using
1768 the meta key directly or as qualified meta expressions. Meta l4proto is
1769 useful to match a particular transport protocol that is part of either
1770 an IPv4 or IPv6 packet. It will also skip any IPv6 extension headers
1771 present in an IPv6 packet.
1772
1773 meta iif, oif, iifname and oifname are used to match the interface a
1774 packet arrived on or is about to be sent out on.
1775
1776 iif and oif are used to match on the interface index, whereas iifname
1777 and oifname are used to match on the interface name. This is not the
1778 same — assuming the rule
1779
1780 filter input meta iif "foo"
1781
1782 Then this rule can only be added if the interface "foo" exists. Also,
1783 the rule will continue to match even if the interface "foo" is renamed
1784 to "bar".
1785
1786 This is because internally the interface index is used. In case of
1787 dynamically created interfaces, such as tun/tap or dialup interfaces
1788 (ppp for example), it might be better to use iifname or oifname
1789 instead.
1790
1791 In these cases, the name is used so the interface doesn’t have to exist
1792 to add such a rule, it will stop matching if the interface gets renamed
1793 and it will match again in case interface gets deleted and later a new
1794 interface with the same name is created.
1795
1796 Like with iptables, wildcard matching on interface name prefixes is
1797 available for iifname and oifname matches by appending an asterisk (*)
1798 character. Note however that unlike iptables, nftables does not accept
1799 interface names consisting of the wildcard character only - users are
1800 supposed to just skip those always matching expressions. In order to
1801 match on literal asterisk character, one may escape it using backslash
1802 (\).
1803
1804 Table 29. Meta expression types
1805 ┌──────────┬─────────────────────┬─────────────────────┐
1806 │Keyword │ Description │ Type │
1807 ├──────────┼─────────────────────┼─────────────────────┤
1808 │ │ │ │
1809 │length │ Length of the │ integer (32-bit) │
1810 │ │ packet in bytes │ │
1811 ├──────────┼─────────────────────┼─────────────────────┤
1812 │ │ │ │
1813 │nfproto │ real hook protocol │ integer (32 bit) │
1814 │ │ family, useful only │ │
1815 │ │ in inet table │ │
1816 ├──────────┼─────────────────────┼─────────────────────┤
1817 │ │ │ │
1818 │l4proto │ layer 4 protocol, │ integer (8 bit) │
1819 │ │ skips ipv6 │ │
1820 │ │ extension headers │ │
1821 ├──────────┼─────────────────────┼─────────────────────┤
1822 │ │ │ │
1823 │protocol │ EtherType protocol │ ether_type │
1824 │ │ value │ │
1825 ├──────────┼─────────────────────┼─────────────────────┤
1826 │ │ │ │
1827 │priority │ TC packet priority │ tc_handle │
1828 ├──────────┼─────────────────────┼─────────────────────┤
1829 │ │ │ │
1830 │mark │ Packet mark │ mark │
1831 ├──────────┼─────────────────────┼─────────────────────┤
1832 │ │ │ │
1833 │iif │ Input interface │ iface_index │
1834 │ │ index │ │
1835 ├──────────┼─────────────────────┼─────────────────────┤
1836 │ │ │ │
1837 │iifname │ Input interface │ ifname │
1838 │ │ name │ │
1839 ├──────────┼─────────────────────┼─────────────────────┤
1840 │ │ │ │
1841 │iiftype │ Input interface │ iface_type │
1842 │ │ type │ │
1843 ├──────────┼─────────────────────┼─────────────────────┤
1844 │ │ │ │
1845 │oif │ Output interface │ iface_index │
1846 │ │ index │ │
1847 ├──────────┼─────────────────────┼─────────────────────┤
1848 │ │ │ │
1849 │oifname │ Output interface │ ifname │
1850 │ │ name │ │
1851 ├──────────┼─────────────────────┼─────────────────────┤
1852 │ │ │ │
1853 │oiftype │ Output interface │ iface_type │
1854 │ │ hardware type │ │
1855 ├──────────┼─────────────────────┼─────────────────────┤
1856 │ │ │ │
1857 │sdif │ Slave device input │ iface_index │
1858 │ │ interface index │ │
1859 ├──────────┼─────────────────────┼─────────────────────┤
1860 │ │ │ │
1861 │sdifname │ Slave device │ ifname │
1862 │ │ interface name │ │
1863 ├──────────┼─────────────────────┼─────────────────────┤
1864 │ │ │ │
1865 │skuid │ UID associated with │ uid │
1866 │ │ originating socket │ │
1867 ├──────────┼─────────────────────┼─────────────────────┤
1868 │ │ │ │
1869 │skgid │ GID associated with │ gid │
1870 │ │ originating socket │ │
1871 ├──────────┼─────────────────────┼─────────────────────┤
1872 │ │ │ │
1873 │rtclassid │ Routing realm │ realm │
1874 ├──────────┼─────────────────────┼─────────────────────┤
1875 │ │ │ │
1876 │ibrname │ Input bridge │ ifname │
1877 │ │ interface name │ │
1878 ├──────────┼─────────────────────┼─────────────────────┤
1879 │ │ │ │
1880 │obrname │ Output bridge │ ifname │
1881 │ │ interface name │ │
1882 ├──────────┼─────────────────────┼─────────────────────┤
1883 │ │ │ │
1884 │pkttype │ packet type │ pkt_type │
1885 ├──────────┼─────────────────────┼─────────────────────┤
1886 │ │ │ │
1887 │cpu │ cpu number │ integer (32 bit) │
1888 │ │ processing the │ │
1889 │ │ packet │ │
1890 ├──────────┼─────────────────────┼─────────────────────┤
1891 │ │ │ │
1892 │iifgroup │ incoming device │ devgroup │
1893 │ │ group │ │
1894 ├──────────┼─────────────────────┼─────────────────────┤
1895 │ │ │ │
1896 │oifgroup │ outgoing device │ devgroup │
1897 │ │ group │ │
1898 ├──────────┼─────────────────────┼─────────────────────┤
1899 │ │ │ │
1900 │cgroup │ control group id │ integer (32 bit) │
1901 ├──────────┼─────────────────────┼─────────────────────┤
1902 │ │ │ │
1903 │random │ pseudo-random │ integer (32 bit) │
1904 │ │ number │ │
1905 ├──────────┼─────────────────────┼─────────────────────┤
1906 │ │ │ │
1907 │ipsec │ true if packet was │ boolean (1 bit) │
1908 │ │ ipsec encrypted │ │
1909 ├──────────┼─────────────────────┼─────────────────────┤
1910 │ │ │ │
1911 │iifkind │ Input interface │ │
1912 │ │ kind │ │
1913 ├──────────┼─────────────────────┼─────────────────────┤
1914 │ │ │ │
1915 │oifkind │ Output interface │ │
1916 │ │ kind │ │
1917 ├──────────┼─────────────────────┼─────────────────────┤
1918 │ │ │ │
1919 │time │ Absolute time of │ Integer (32 bit) or │
1920 │ │ packet reception │ string │
1921 ├──────────┼─────────────────────┼─────────────────────┤
1922 │ │ │ │
1923 │day │ Day of week │ Integer (8 bit) or │
1924 │ │ │ string │
1925 ├──────────┼─────────────────────┼─────────────────────┤
1926 │ │ │ │
1927 │hour │ Hour of day │ String │
1928 └──────────┴─────────────────────┴─────────────────────┘
1929
1930 Table 30. Meta expression specific types
1931 ┌──────────────┬────────────────────────────┐
1932 │Type │ Description │
1933 ├──────────────┼────────────────────────────┤
1934 │ │ │
1935 │iface_index │ Interface index (32 bit │
1936 │ │ number). Can be specified │
1937 │ │ numerically or as name of │
1938 │ │ an existing interface. │
1939 ├──────────────┼────────────────────────────┤
1940 │ │ │
1941 │ifname │ Interface name (16 byte │
1942 │ │ string). Does not have to │
1943 │ │ exist. │
1944 ├──────────────┼────────────────────────────┤
1945 │ │ │
1946 │iface_type │ Interface type (16 bit │
1947 │ │ number). │
1948 ├──────────────┼────────────────────────────┤
1949 │ │ │
1950 │uid │ User ID (32 bit number). │
1951 │ │ Can be specified │
1952 │ │ numerically or as user │
1953 │ │ name. │
1954 ├──────────────┼────────────────────────────┤
1955 │ │ │
1956 │gid │ Group ID (32 bit number). │
1957 │ │ Can be specified │
1958 │ │ numerically or as group │
1959 │ │ name. │
1960 ├──────────────┼────────────────────────────┤
1961 │ │ │
1962 │realm │ Routing Realm (32 bit │
1963 │ │ number). Can be specified │
1964 │ │ numerically or as symbolic │
1965 │ │ name defined in │
1966 │ │ /etc/iproute2/rt_realms. │
1967 ├──────────────┼────────────────────────────┤
1968 │ │ │
1969 │devgroup_type │ Device group (32 bit │
1970 │ │ number). Can be specified │
1971 │ │ numerically or as symbolic │
1972 │ │ name defined in │
1973 │ │ /etc/iproute2/group. │
1974 ├──────────────┼────────────────────────────┤
1975 │ │ │
1976 │pkt_type │ Packet type: host │
1977 │ │ (addressed to local host), │
1978 │ │ broadcast (to all), │
1979 │ │ multicast (to group), │
1980 │ │ other (addressed to │
1981 │ │ another host). │
1982 ├──────────────┼────────────────────────────┤
1983 │ │ │
1984 │ifkind │ Interface kind (16 byte │
1985 │ │ string). See TYPES in │
1986 │ │ ip-link(8) for a list. │
1987 ├──────────────┼────────────────────────────┤
1988 │ │ │
1989 │time │ Either an integer or a │
1990 │ │ date in ISO format. For │
1991 │ │ example: "2019-06-06 │
1992 │ │ 17:00". Hour and seconds │
1993 │ │ are optional and can be │
1994 │ │ omitted if desired. If │
1995 │ │ omitted, midnight will be │
1996 │ │ assumed. The following │
1997 │ │ three would be equivalent: │
1998 │ │ "2019-06-06", "2019-06-06 │
1999 │ │ 00:00" and "2019-06-06 │
2000 │ │ 00:00:00". When an integer │
2001 │ │ is given, it is assumed to │
2002 │ │ be a UNIX timestamp. │
2003 ├──────────────┼────────────────────────────┤
2004 │ │ │
2005 │day │ Either a day of week │
2006 │ │ ("Monday", "Tuesday", │
2007 │ │ etc.), or an integer │
2008 │ │ between 0 and 6. Strings │
2009 │ │ are matched │
2010 │ │ case-insensitively, and a │
2011 │ │ full match is not expected │
2012 │ │ (e.g. "Mon" would match │
2013 │ │ "Monday"). When an integer │
2014 │ │ is given, 0 is Sunday and │
2015 │ │ 6 is Saturday. │
2016 ├──────────────┼────────────────────────────┤
2017 │ │ │
2018 │hour │ A string representing an │
2019 │ │ hour in 24-hour format. │
2020 │ │ Seconds can optionally be │
2021 │ │ specified. For example, │
2022 │ │ 17:00 and 17:00:00 would │
2023 │ │ be equivalent. │
2024 └──────────────┴────────────────────────────┘
2025
2026 Using meta expressions.
2027
2028 # qualified meta expression
2029 filter output meta oif eth0
2030 filter forward meta iifkind { "tun", "veth" }
2031
2032 # unqualified meta expression
2033 filter output oif eth0
2034
2035 # incoming packet was subject to ipsec processing
2036 raw prerouting meta ipsec exists accept
2037
2038
2039 SOCKET EXPRESSION
2040 socket {transparent | mark | wildcard}
2041
2042 Socket expression can be used to search for an existing open TCP/UDP
2043 socket and its attributes that can be associated with a packet. It
2044 looks for an established or non-zero bound listening socket (possibly
2045 with a non-local address).
2046
2047 Table 31. Available socket attributes
2048 ┌────────────┬─────────────────────┬─────────────────┐
2049 │Name │ Description │ Type │
2050 ├────────────┼─────────────────────┼─────────────────┤
2051 │ │ │ │
2052 │transparent │ Value of the │ boolean (1 bit) │
2053 │ │ IP_TRANSPARENT │ │
2054 │ │ socket option in │ │
2055 │ │ the found socket. │ │
2056 │ │ It can be 0 or 1. │ │
2057 ├────────────┼─────────────────────┼─────────────────┤
2058 │ │ │ │
2059 │mark │ Value of the socket │ mark │
2060 │ │ mark (SOL_SOCKET, │ │
2061 │ │ SO_MARK). │ │
2062 ├────────────┼─────────────────────┼─────────────────┤
2063 │ │ │ │
2064 │wildcard │ Indicates whether │ boolean (1 bit) │
2065 │ │ the socket is │ │
2066 │ │ wildcard-bound │ │
2067 │ │ (e.g. 0.0.0.0 or │ │
2068 │ │ ::0). │ │
2069 └────────────┴─────────────────────┴─────────────────┘
2070
2071 Using socket expression.
2072
2073 # Mark packets that correspond to a transparent socket. "socket wildcard 0"
2074 # means that zero-bound listener sockets are NOT matched (which is usually
2075 # exactly what you want).
2076 table inet x {
2077 chain y {
2078 type filter hook prerouting priority -150; policy accept;
2079 socket transparent 1 socket wildcard 0 mark set 0x00000001 accept
2080 }
2081 }
2082
2083 # Trace packets that corresponds to a socket with a mark value of 15
2084 table inet x {
2085 chain y {
2086 type filter hook prerouting priority -150; policy accept;
2087 socket mark 0x0000000f nftrace set 1
2088 }
2089 }
2090
2091 # Set packet mark to socket mark
2092 table inet x {
2093 chain y {
2094 type filter hook prerouting priority -150; policy accept;
2095 tcp dport 8080 mark set socket mark
2096 }
2097 }
2098
2099
2100 OSF EXPRESSION
2101 osf [ttl {loose | skip}] {name | version}
2102
2103 The osf expression does passive operating system fingerprinting. This
2104 expression compares some data (Window Size, MSS, options and their
2105 order, DF, and others) from packets with the SYN bit set.
2106
2107 Table 32. Available osf attributes
2108 ┌────────┬─────────────────────┬────────┐
2109 │Name │ Description │ Type │
2110 ├────────┼─────────────────────┼────────┤
2111 │ │ │ │
2112 │ttl │ Do TTL checks on │ string │
2113 │ │ the packet to │ │
2114 │ │ determine the │ │
2115 │ │ operating system. │ │
2116 ├────────┼─────────────────────┼────────┤
2117 │ │ │ │
2118 │version │ Do OS version │ │
2119 │ │ checks on the │ │
2120 │ │ packet. │ │
2121 ├────────┼─────────────────────┼────────┤
2122 │ │ │ │
2123 │name │ Name of the OS │ string │
2124 │ │ signature to match. │ │
2125 │ │ All signatures can │ │
2126 │ │ be found at pf.os │ │
2127 │ │ file. Use "unknown" │ │
2128 │ │ for OS signatures │ │
2129 │ │ that the expression │ │
2130 │ │ could not detect. │ │
2131 └────────┴─────────────────────┴────────┘
2132
2133 Available ttl values.
2134
2135 If no TTL attribute is passed, make a true IP header and fingerprint TTL true comparison. This generally works for LANs.
2136
2137 * loose: Check if the IP header's TTL is less than the fingerprint one. Works for globally-routable addresses.
2138 * skip: Do not compare the TTL at all.
2139
2140 Using osf expression.
2141
2142 # Accept packets that match the "Linux" OS genre signature without comparing TTL.
2143 table inet x {
2144 chain y {
2145 type filter hook input priority 0; policy accept;
2146 osf ttl skip name "Linux"
2147 }
2148 }
2149
2150
2151 FIB EXPRESSIONS
2152 fib {saddr | daddr | mark | iif | oif} [. ...] {oif | oifname | type}
2153
2154 A fib expression queries the fib (forwarding information base) to
2155 obtain information such as the output interface index a particular
2156 address would use. The input is a tuple of elements that is used as
2157 input to the fib lookup functions.
2158
2159 Table 33. fib expression specific types
2160 ┌────────┬──────────────────┬──────────────────┐
2161 │Keyword │ Description │ Type │
2162 ├────────┼──────────────────┼──────────────────┤
2163 │ │ │ │
2164 │oif │ Output interface │ integer (32 bit) │
2165 │ │ index │ │
2166 ├────────┼──────────────────┼──────────────────┤
2167 │ │ │ │
2168 │oifname │ Output interface │ string │
2169 │ │ name │ │
2170 ├────────┼──────────────────┼──────────────────┤
2171 │ │ │ │
2172 │type │ Address type │ fib_addrtype │
2173 └────────┴──────────────────┴──────────────────┘
2174
2175 Use nft describe fib_addrtype to get a list of all address types.
2176
2177 Using fib expressions.
2178
2179 # drop packets without a reverse path
2180 filter prerouting fib saddr . iif oif missing drop
2181
2182 In this example, 'saddr . iif' looks up routing information based on the source address and the input interface.
2183 oif picks the output interface index from the routing information.
2184 If no route was found for the source address/input interface combination, the output interface index is zero.
2185 In case the input interface is specified as part of the input key, the output interface index is always the same as the input interface index or zero.
2186 If only 'saddr oif' is given, then oif can be any interface index or zero.
2187
2188 # drop packets to address not configured on incoming interface
2189 filter prerouting fib daddr . iif type != { local, broadcast, multicast } drop
2190
2191 # perform lookup in a specific 'blackhole' table (0xdead, needs ip appropriate ip rule)
2192 filter prerouting meta mark set 0xdead fib daddr . mark type vmap { blackhole : drop, prohibit : jump prohibited, unreachable : drop }
2193
2194
2195 ROUTING EXPRESSIONS
2196 rt [ip | ip6] {classid | nexthop | mtu | ipsec}
2197
2198 A routing expression refers to routing data associated with a packet.
2199
2200 Table 34. Routing expression types
2201 ┌────────┬─────────────────────┬─────────────────────┐
2202 │Keyword │ Description │ Type │
2203 ├────────┼─────────────────────┼─────────────────────┤
2204 │ │ │ │
2205 │classid │ Routing realm │ realm │
2206 ├────────┼─────────────────────┼─────────────────────┤
2207 │ │ │ │
2208 │nexthop │ Routing nexthop │ ipv4_addr/ipv6_addr │
2209 ├────────┼─────────────────────┼─────────────────────┤
2210 │ │ │ │
2211 │mtu │ TCP maximum segment │ integer (16 bit) │
2212 │ │ size of route │ │
2213 ├────────┼─────────────────────┼─────────────────────┤
2214 │ │ │ │
2215 │ipsec │ route via ipsec │ boolean │
2216 │ │ tunnel or transport │ │
2217 └────────┴─────────────────────┴─────────────────────┘
2218
2219 Table 35. Routing expression specific types
2220 ┌──────┬────────────────────────────┐
2221 │Type │ Description │
2222 ├──────┼────────────────────────────┤
2223 │ │ │
2224 │realm │ Routing Realm (32 bit │
2225 │ │ number). Can be specified │
2226 │ │ numerically or as symbolic │
2227 │ │ name defined in │
2228 │ │ /etc/iproute2/rt_realms. │
2229 └──────┴────────────────────────────┘
2230
2231 Using routing expressions.
2232
2233 # IP family independent rt expression
2234 filter output rt classid 10
2235
2236 # IP family dependent rt expressions
2237 ip filter output rt nexthop 192.168.0.1
2238 ip6 filter output rt nexthop fd00::1
2239 inet filter output rt ip nexthop 192.168.0.1
2240 inet filter output rt ip6 nexthop fd00::1
2241
2242 # outgoing packet will be encapsulated/encrypted by ipsec
2243 filter output rt ipsec exists
2244
2245
2246 IPSEC EXPRESSIONS
2247 ipsec {in | out} [ spnum NUM ] {reqid | spi}
2248 ipsec {in | out} [ spnum NUM ] {ip | ip6} {saddr | daddr}
2249
2250 An ipsec expression refers to ipsec data associated with a packet.
2251
2252 The in or out keyword needs to be used to specify if the expression
2253 should examine inbound or outbound policies. The in keyword can be used
2254 in the prerouting, input and forward hooks. The out keyword applies to
2255 forward, output and postrouting hooks. The optional keyword spnum can
2256 be used to match a specific state in a chain, it defaults to 0.
2257
2258 Table 36. Ipsec expression types
2259 ┌────────┬─────────────────────┬─────────────────────┐
2260 │Keyword │ Description │ Type │
2261 ├────────┼─────────────────────┼─────────────────────┤
2262 │ │ │ │
2263 │reqid │ Request ID │ integer (32 bit) │
2264 ├────────┼─────────────────────┼─────────────────────┤
2265 │ │ │ │
2266 │spi │ Security Parameter │ integer (32 bit) │
2267 │ │ Index │ │
2268 ├────────┼─────────────────────┼─────────────────────┤
2269 │ │ │ │
2270 │saddr │ Source address of │ ipv4_addr/ipv6_addr │
2271 │ │ the tunnel │ │
2272 ├────────┼─────────────────────┼─────────────────────┤
2273 │ │ │ │
2274 │daddr │ Destination address │ ipv4_addr/ipv6_addr │
2275 │ │ of the tunnel │ │
2276 └────────┴─────────────────────┴─────────────────────┘
2277
2278 NUMGEN EXPRESSION
2279 numgen {inc | random} mod NUM [ offset NUM ]
2280
2281 Create a number generator. The inc or random keywords control its
2282 operation mode: In inc mode, the last returned value is simply
2283 incremented. In random mode, a new random number is returned. The value
2284 after mod keyword specifies an upper boundary (read: modulus) which is
2285 not reached by returned numbers. The optional offset allows to
2286 increment the returned value by a fixed offset.
2287
2288 A typical use-case for numgen is load-balancing:
2289
2290 Using numgen expression.
2291
2292 # round-robin between 192.168.10.100 and 192.168.20.200:
2293 add rule nat prerouting dnat to numgen inc mod 2 map \
2294 { 0 : 192.168.10.100, 1 : 192.168.20.200 }
2295
2296 # probability-based with odd bias using intervals:
2297 add rule nat prerouting dnat to numgen random mod 10 map \
2298 { 0-2 : 192.168.10.100, 3-9 : 192.168.20.200 }
2299
2300
2301 HASH EXPRESSIONS
2302 jhash {ip saddr | ip6 daddr | tcp dport | udp sport | ether saddr} [. ...] mod NUM [ seed NUM ] [ offset NUM ]
2303 symhash mod NUM [ offset NUM ]
2304
2305 Use a hashing function to generate a number. The functions available
2306 are jhash, known as Jenkins Hash, and symhash, for Symmetric Hash. The
2307 jhash requires an expression to determine the parameters of the packet
2308 header to apply the hashing, concatenations are possible as well. The
2309 value after mod keyword specifies an upper boundary (read: modulus)
2310 which is not reached by returned numbers. The optional seed is used to
2311 specify an init value used as seed in the hashing function. The
2312 optional offset allows to increment the returned value by a fixed
2313 offset.
2314
2315 A typical use-case for jhash and symhash is load-balancing:
2316
2317 Using hash expressions.
2318
2319 # load balance based on source ip between 2 ip addresses:
2320 add rule nat prerouting dnat to jhash ip saddr mod 2 map \
2321 { 0 : 192.168.10.100, 1 : 192.168.20.200 }
2322
2323 # symmetric load balancing between 2 ip addresses:
2324 add rule nat prerouting dnat to symhash mod 2 map \
2325 { 0 : 192.168.10.100, 1 : 192.168.20.200 }
2326
2327
2329 Payload expressions refer to data from the packet’s payload.
2330
2331 ETHERNET HEADER EXPRESSION
2332 ether {daddr | saddr | type}
2333
2334 Table 37. Ethernet header expression types
2335 ┌────────┬────────────────────┬────────────┐
2336 │Keyword │ Description │ Type │
2337 ├────────┼────────────────────┼────────────┤
2338 │ │ │ │
2339 │daddr │ Destination MAC │ ether_addr │
2340 │ │ address │ │
2341 ├────────┼────────────────────┼────────────┤
2342 │ │ │ │
2343 │saddr │ Source MAC address │ ether_addr │
2344 ├────────┼────────────────────┼────────────┤
2345 │ │ │ │
2346 │type │ EtherType │ ether_type │
2347 └────────┴────────────────────┴────────────┘
2348
2349 VLAN HEADER EXPRESSION
2350 vlan {id | cfi | pcp | type}
2351
2352 Table 38. VLAN header expression
2353 ┌────────┬─────────────────────┬──────────────────┐
2354 │Keyword │ Description │ Type │
2355 ├────────┼─────────────────────┼──────────────────┤
2356 │ │ │ │
2357 │id │ VLAN ID (VID) │ integer (12 bit) │
2358 ├────────┼─────────────────────┼──────────────────┤
2359 │ │ │ │
2360 │cfi │ Canonical Format │ integer (1 bit) │
2361 │ │ Indicator │ │
2362 ├────────┼─────────────────────┼──────────────────┤
2363 │ │ │ │
2364 │pcp │ Priority code point │ integer (3 bit) │
2365 ├────────┼─────────────────────┼──────────────────┤
2366 │ │ │ │
2367 │type │ EtherType │ ether_type │
2368 └────────┴─────────────────────┴──────────────────┘
2369
2370 ARP HEADER EXPRESSION
2371 arp {htype | ptype | hlen | plen | operation | saddr { ip | ether } | daddr { ip | ether }
2372
2373 Table 39. ARP header expression
2374 ┌────────────┬─────────────────────┬──────────────────┐
2375 │Keyword │ Description │ Type │
2376 ├────────────┼─────────────────────┼──────────────────┤
2377 │ │ │ │
2378 │htype │ ARP hardware type │ integer (16 bit) │
2379 ├────────────┼─────────────────────┼──────────────────┤
2380 │ │ │ │
2381 │ptype │ EtherType │ ether_type │
2382 ├────────────┼─────────────────────┼──────────────────┤
2383 │ │ │ │
2384 │hlen │ Hardware address │ integer (8 bit) │
2385 │ │ len │ │
2386 ├────────────┼─────────────────────┼──────────────────┤
2387 │ │ │ │
2388 │plen │ Protocol address │ integer (8 bit) │
2389 │ │ len │ │
2390 ├────────────┼─────────────────────┼──────────────────┤
2391 │ │ │ │
2392 │operation │ Operation │ arp_op │
2393 ├────────────┼─────────────────────┼──────────────────┤
2394 │ │ │ │
2395 │saddr ether │ Ethernet sender │ ether_addr │
2396 │ │ address │ │
2397 ├────────────┼─────────────────────┼──────────────────┤
2398 │ │ │ │
2399 │daddr ether │ Ethernet target │ ether_addr │
2400 │ │ address │ │
2401 ├────────────┼─────────────────────┼──────────────────┤
2402 │ │ │ │
2403 │saddr ip │ IPv4 sender address │ ipv4_addr │
2404 ├────────────┼─────────────────────┼──────────────────┤
2405 │ │ │ │
2406 │daddr ip │ IPv4 target address │ ipv4_addr │
2407 └────────────┴─────────────────────┴──────────────────┘
2408
2409 IPV4 HEADER EXPRESSION
2410 ip {version | hdrlength | dscp | ecn | length | id | frag-off | ttl | protocol | checksum | saddr | daddr }
2411
2412 Table 40. IPv4 header expression
2413 ┌──────────┬─────────────────────┬──────────────────┐
2414 │Keyword │ Description │ Type │
2415 ├──────────┼─────────────────────┼──────────────────┤
2416 │ │ │ │
2417 │version │ IP header version │ integer (4 bit) │
2418 │ │ (4) │ │
2419 ├──────────┼─────────────────────┼──────────────────┤
2420 │ │ │ │
2421 │hdrlength │ IP header length │ integer (4 bit) │
2422 │ │ including options │ FIXME scaling │
2423 ├──────────┼─────────────────────┼──────────────────┤
2424 │ │ │ │
2425 │dscp │ Differentiated │ dscp │
2426 │ │ Services Code Point │ │
2427 ├──────────┼─────────────────────┼──────────────────┤
2428 │ │ │ │
2429 │ecn │ Explicit Congestion │ ecn │
2430 │ │ Notification │ │
2431 ├──────────┼─────────────────────┼──────────────────┤
2432 │ │ │ │
2433 │length │ Total packet length │ integer (16 bit) │
2434 ├──────────┼─────────────────────┼──────────────────┤
2435 │ │ │ │
2436 │id │ IP ID │ integer (16 bit) │
2437 ├──────────┼─────────────────────┼──────────────────┤
2438 │ │ │ │
2439 │frag-off │ Fragment offset │ integer (16 bit) │
2440 ├──────────┼─────────────────────┼──────────────────┤
2441 │ │ │ │
2442 │ttl │ Time to live │ integer (8 bit) │
2443 ├──────────┼─────────────────────┼──────────────────┤
2444 │ │ │ │
2445 │protocol │ Upper layer │ inet_proto │
2446 │ │ protocol │ │
2447 ├──────────┼─────────────────────┼──────────────────┤
2448 │ │ │ │
2449 │checksum │ IP header checksum │ integer (16 bit) │
2450 ├──────────┼─────────────────────┼──────────────────┤
2451 │ │ │ │
2452 │saddr │ Source address │ ipv4_addr │
2453 ├──────────┼─────────────────────┼──────────────────┤
2454 │ │ │ │
2455 │daddr │ Destination address │ ipv4_addr │
2456 └──────────┴─────────────────────┴──────────────────┘
2457
2458 ICMP HEADER EXPRESSION
2459 icmp {type | code | checksum | id | sequence | gateway | mtu}
2460
2461 This expression refers to ICMP header fields. When using it in inet,
2462 bridge or netdev families, it will cause an implicit dependency on IPv4
2463 to be created. To match on unusual cases like ICMP over IPv6, one has
2464 to add an explicit meta protocol ip6 match to the rule.
2465
2466 Table 41. ICMP header expression
2467 ┌─────────┬─────────────────────┬──────────────────┐
2468 │Keyword │ Description │ Type │
2469 ├─────────┼─────────────────────┼──────────────────┤
2470 │ │ │ │
2471 │type │ ICMP type field │ icmp_type │
2472 ├─────────┼─────────────────────┼──────────────────┤
2473 │ │ │ │
2474 │code │ ICMP code field │ integer (8 bit) │
2475 ├─────────┼─────────────────────┼──────────────────┤
2476 │ │ │ │
2477 │checksum │ ICMP checksum field │ integer (16 bit) │
2478 ├─────────┼─────────────────────┼──────────────────┤
2479 │ │ │ │
2480 │id │ ID of echo │ integer (16 bit) │
2481 │ │ request/response │ │
2482 ├─────────┼─────────────────────┼──────────────────┤
2483 │ │ │ │
2484 │sequence │ sequence number of │ integer (16 bit) │
2485 │ │ echo │ │
2486 │ │ request/response │ │
2487 ├─────────┼─────────────────────┼──────────────────┤
2488 │ │ │ │
2489 │gateway │ gateway of │ integer (32 bit) │
2490 │ │ redirects │ │
2491 ├─────────┼─────────────────────┼──────────────────┤
2492 │ │ │ │
2493 │mtu │ MTU of path MTU │ integer (16 bit) │
2494 │ │ discovery │ │
2495 └─────────┴─────────────────────┴──────────────────┘
2496
2497 IGMP HEADER EXPRESSION
2498 igmp {type | mrt | checksum | group}
2499
2500 This expression refers to IGMP header fields. When using it in inet,
2501 bridge or netdev families, it will cause an implicit dependency on IPv4
2502 to be created. To match on unusual cases like IGMP over IPv6, one has
2503 to add an explicit meta protocol ip6 match to the rule.
2504
2505 Table 42. IGMP header expression
2506 ┌─────────┬─────────────────────┬──────────────────┐
2507 │Keyword │ Description │ Type │
2508 ├─────────┼─────────────────────┼──────────────────┤
2509 │ │ │ │
2510 │type │ IGMP type field │ igmp_type │
2511 ├─────────┼─────────────────────┼──────────────────┤
2512 │ │ │ │
2513 │mrt │ IGMP maximum │ integer (8 bit) │
2514 │ │ response time field │ │
2515 ├─────────┼─────────────────────┼──────────────────┤
2516 │ │ │ │
2517 │checksum │ IGMP checksum field │ integer (16 bit) │
2518 ├─────────┼─────────────────────┼──────────────────┤
2519 │ │ │ │
2520 │group │ Group address │ integer (32 bit) │
2521 └─────────┴─────────────────────┴──────────────────┘
2522
2523 IPV6 HEADER EXPRESSION
2524 ip6 {version | dscp | ecn | flowlabel | length | nexthdr | hoplimit | saddr | daddr}
2525
2526 This expression refers to the ipv6 header fields. Caution when using
2527 ip6 nexthdr, the value only refers to the next header, i.e. ip6 nexthdr
2528 tcp will only match if the ipv6 packet does not contain any extension
2529 headers. Packets that are fragmented or e.g. contain a routing
2530 extension headers will not be matched. Please use meta l4proto if you
2531 wish to match the real transport header and ignore any additional
2532 extension headers instead.
2533
2534 Table 43. IPv6 header expression
2535 ┌──────────┬─────────────────────┬──────────────────┐
2536 │Keyword │ Description │ Type │
2537 ├──────────┼─────────────────────┼──────────────────┤
2538 │ │ │ │
2539 │version │ IP header version │ integer (4 bit) │
2540 │ │ (6) │ │
2541 ├──────────┼─────────────────────┼──────────────────┤
2542 │ │ │ │
2543 │dscp │ Differentiated │ dscp │
2544 │ │ Services Code Point │ │
2545 ├──────────┼─────────────────────┼──────────────────┤
2546 │ │ │ │
2547 │ecn │ Explicit Congestion │ ecn │
2548 │ │ Notification │ │
2549 ├──────────┼─────────────────────┼──────────────────┤
2550 │ │ │ │
2551 │flowlabel │ Flow label │ integer (20 bit) │
2552 ├──────────┼─────────────────────┼──────────────────┤
2553 │ │ │ │
2554 │length │ Payload length │ integer (16 bit) │
2555 ├──────────┼─────────────────────┼──────────────────┤
2556 │ │ │ │
2557 │nexthdr │ Nexthdr protocol │ inet_proto │
2558 ├──────────┼─────────────────────┼──────────────────┤
2559 │ │ │ │
2560 │hoplimit │ Hop limit │ integer (8 bit) │
2561 ├──────────┼─────────────────────┼──────────────────┤
2562 │ │ │ │
2563 │saddr │ Source address │ ipv6_addr │
2564 ├──────────┼─────────────────────┼──────────────────┤
2565 │ │ │ │
2566 │daddr │ Destination address │ ipv6_addr │
2567 └──────────┴─────────────────────┴──────────────────┘
2568
2569 Using ip6 header expressions.
2570
2571 # matching if first extension header indicates a fragment
2572 ip6 nexthdr ipv6-frag
2573
2574
2575 ICMPV6 HEADER EXPRESSION
2576 icmpv6 {type | code | checksum | parameter-problem | packet-too-big | id | sequence | max-delay}
2577
2578 This expression refers to ICMPv6 header fields. When using it in inet,
2579 bridge or netdev families, it will cause an implicit dependency on IPv6
2580 to be created. To match on unusual cases like ICMPv6 over IPv4, one has
2581 to add an explicit meta protocol ip match to the rule.
2582
2583 Table 44. ICMPv6 header expression
2584 ┌──────────────────┬────────────────────┬──────────────────┐
2585 │Keyword │ Description │ Type │
2586 ├──────────────────┼────────────────────┼──────────────────┤
2587 │ │ │ │
2588 │type │ ICMPv6 type field │ icmpv6_type │
2589 ├──────────────────┼────────────────────┼──────────────────┤
2590 │ │ │ │
2591 │code │ ICMPv6 code field │ integer (8 bit) │
2592 ├──────────────────┼────────────────────┼──────────────────┤
2593 │ │ │ │
2594 │checksum │ ICMPv6 checksum │ integer (16 bit) │
2595 │ │ field │ │
2596 ├──────────────────┼────────────────────┼──────────────────┤
2597 │ │ │ │
2598 │parameter-problem │ pointer to problem │ integer (32 bit) │
2599 ├──────────────────┼────────────────────┼──────────────────┤
2600 │ │ │ │
2601 │packet-too-big │ oversized MTU │ integer (32 bit) │
2602 ├──────────────────┼────────────────────┼──────────────────┤
2603 │ │ │ │
2604 │id │ ID of echo │ integer (16 bit) │
2605 │ │ request/response │ │
2606 ├──────────────────┼────────────────────┼──────────────────┤
2607 │ │ │ │
2608 │sequence │ sequence number of │ integer (16 bit) │
2609 │ │ echo │ │
2610 │ │ request/response │ │
2611 ├──────────────────┼────────────────────┼──────────────────┤
2612 │ │ │ │
2613 │max-delay │ maximum response │ integer (16 bit) │
2614 │ │ delay of MLD │ │
2615 │ │ queries │ │
2616 └──────────────────┴────────────────────┴──────────────────┘
2617
2618 TCP HEADER EXPRESSION
2619 tcp {sport | dport | sequence | ackseq | doff | reserved | flags | window | checksum | urgptr}
2620
2621 Table 45. TCP header expression
2622 ┌─────────┬──────────────────┬──────────────────┐
2623 │Keyword │ Description │ Type │
2624 ├─────────┼──────────────────┼──────────────────┤
2625 │ │ │ │
2626 │sport │ Source port │ inet_service │
2627 ├─────────┼──────────────────┼──────────────────┤
2628 │ │ │ │
2629 │dport │ Destination port │ inet_service │
2630 ├─────────┼──────────────────┼──────────────────┤
2631 │ │ │ │
2632 │sequence │ Sequence number │ integer (32 bit) │
2633 ├─────────┼──────────────────┼──────────────────┤
2634 │ │ │ │
2635 │ackseq │ Acknowledgement │ integer (32 bit) │
2636 │ │ number │ │
2637 ├─────────┼──────────────────┼──────────────────┤
2638 │ │ │ │
2639 │doff │ Data offset │ integer (4 bit) │
2640 │ │ │ FIXME scaling │
2641 ├─────────┼──────────────────┼──────────────────┤
2642 │ │ │ │
2643 │reserved │ Reserved area │ integer (4 bit) │
2644 ├─────────┼──────────────────┼──────────────────┤
2645 │ │ │ │
2646 │flags │ TCP flags │ tcp_flag │
2647 ├─────────┼──────────────────┼──────────────────┤
2648 │ │ │ │
2649 │window │ Window │ integer (16 bit) │
2650 ├─────────┼──────────────────┼──────────────────┤
2651 │ │ │ │
2652 │checksum │ Checksum │ integer (16 bit) │
2653 ├─────────┼──────────────────┼──────────────────┤
2654 │ │ │ │
2655 │urgptr │ Urgent pointer │ integer (16 bit) │
2656 └─────────┴──────────────────┴──────────────────┘
2657
2658 UDP HEADER EXPRESSION
2659 udp {sport | dport | length | checksum}
2660
2661 Table 46. UDP header expression
2662 ┌─────────┬─────────────────────┬──────────────────┐
2663 │Keyword │ Description │ Type │
2664 ├─────────┼─────────────────────┼──────────────────┤
2665 │ │ │ │
2666 │sport │ Source port │ inet_service │
2667 ├─────────┼─────────────────────┼──────────────────┤
2668 │ │ │ │
2669 │dport │ Destination port │ inet_service │
2670 ├─────────┼─────────────────────┼──────────────────┤
2671 │ │ │ │
2672 │length │ Total packet length │ integer (16 bit) │
2673 ├─────────┼─────────────────────┼──────────────────┤
2674 │ │ │ │
2675 │checksum │ Checksum │ integer (16 bit) │
2676 └─────────┴─────────────────────┴──────────────────┘
2677
2678 UDP-LITE HEADER EXPRESSION
2679 udplite {sport | dport | checksum}
2680
2681 Table 47. UDP-Lite header expression
2682 ┌─────────┬──────────────────┬──────────────────┐
2683 │Keyword │ Description │ Type │
2684 ├─────────┼──────────────────┼──────────────────┤
2685 │ │ │ │
2686 │sport │ Source port │ inet_service │
2687 ├─────────┼──────────────────┼──────────────────┤
2688 │ │ │ │
2689 │dport │ Destination port │ inet_service │
2690 ├─────────┼──────────────────┼──────────────────┤
2691 │ │ │ │
2692 │checksum │ Checksum │ integer (16 bit) │
2693 └─────────┴──────────────────┴──────────────────┘
2694
2695 SCTP HEADER EXPRESSION
2696 sctp {sport | dport | vtag | checksum}
2697
2698 Table 48. SCTP header expression
2699 ┌─────────┬──────────────────┬──────────────────┐
2700 │Keyword │ Description │ Type │
2701 ├─────────┼──────────────────┼──────────────────┤
2702 │ │ │ │
2703 │sport │ Source port │ inet_service │
2704 ├─────────┼──────────────────┼──────────────────┤
2705 │ │ │ │
2706 │dport │ Destination port │ inet_service │
2707 ├─────────┼──────────────────┼──────────────────┤
2708 │ │ │ │
2709 │vtag │ Verification Tag │ integer (32 bit) │
2710 ├─────────┼──────────────────┼──────────────────┤
2711 │ │ │ │
2712 │checksum │ Checksum │ integer (32 bit) │
2713 └─────────┴──────────────────┴──────────────────┘
2714
2715 DCCP HEADER EXPRESSION
2716 dccp {sport | dport | type}
2717
2718 Table 49. DCCP header expression
2719 ┌────────┬──────────────────┬──────────────┐
2720 │Keyword │ Description │ Type │
2721 ├────────┼──────────────────┼──────────────┤
2722 │ │ │ │
2723 │sport │ Source port │ inet_service │
2724 ├────────┼──────────────────┼──────────────┤
2725 │ │ │ │
2726 │dport │ Destination port │ inet_service │
2727 ├────────┼──────────────────┼──────────────┤
2728 │ │ │ │
2729 │type │ Packet type │ dccp_pkttype │
2730 └────────┴──────────────────┴──────────────┘
2731
2732 AUTHENTICATION HEADER EXPRESSION
2733 ah {nexthdr | hdrlength | reserved | spi | sequence}
2734
2735 Table 50. AH header expression
2736 ┌──────────┬────────────────────┬──────────────────┐
2737 │Keyword │ Description │ Type │
2738 ├──────────┼────────────────────┼──────────────────┤
2739 │ │ │ │
2740 │nexthdr │ Next header │ inet_proto │
2741 │ │ protocol │ │
2742 ├──────────┼────────────────────┼──────────────────┤
2743 │ │ │ │
2744 │hdrlength │ AH Header length │ integer (8 bit) │
2745 ├──────────┼────────────────────┼──────────────────┤
2746 │ │ │ │
2747 │reserved │ Reserved area │ integer (16 bit) │
2748 ├──────────┼────────────────────┼──────────────────┤
2749 │ │ │ │
2750 │spi │ Security Parameter │ integer (32 bit) │
2751 │ │ Index │ │
2752 ├──────────┼────────────────────┼──────────────────┤
2753 │ │ │ │
2754 │sequence │ Sequence number │ integer (32 bit) │
2755 └──────────┴────────────────────┴──────────────────┘
2756
2757 ENCRYPTED SECURITY PAYLOAD HEADER EXPRESSION
2758 esp {spi | sequence}
2759
2760 Table 51. ESP header expression
2761 ┌─────────┬────────────────────┬──────────────────┐
2762 │Keyword │ Description │ Type │
2763 ├─────────┼────────────────────┼──────────────────┤
2764 │ │ │ │
2765 │spi │ Security Parameter │ integer (32 bit) │
2766 │ │ Index │ │
2767 ├─────────┼────────────────────┼──────────────────┤
2768 │ │ │ │
2769 │sequence │ Sequence number │ integer (32 bit) │
2770 └─────────┴────────────────────┴──────────────────┘
2771
2772 IPCOMP HEADER EXPRESSION
2773 comp {nexthdr | flags | cpi}
2774
2775 Table 52. IPComp header expression
2776 ┌────────┬─────────────────┬──────────────────┐
2777 │Keyword │ Description │ Type │
2778 ├────────┼─────────────────┼──────────────────┤
2779 │ │ │ │
2780 │nexthdr │ Next header │ inet_proto │
2781 │ │ protocol │ │
2782 ├────────┼─────────────────┼──────────────────┤
2783 │ │ │ │
2784 │flags │ Flags │ bitmask │
2785 ├────────┼─────────────────┼──────────────────┤
2786 │ │ │ │
2787 │cpi │ compression │ integer (16 bit) │
2788 │ │ Parameter Index │ │
2789 └────────┴─────────────────┴──────────────────┘
2790
2791 RAW PAYLOAD EXPRESSION
2792 @base,offset,length
2793
2794 The raw payload expression instructs to load length bits starting at
2795 offset bits. Bit 0 refers to the very first bit — in the C programming
2796 language, this corresponds to the topmost bit, i.e. 0x80 in case of an
2797 octet. They are useful to match headers that do not have a
2798 human-readable template expression yet. Note that nft will not add
2799 dependencies for Raw payload expressions. If you e.g. want to match
2800 protocol fields of a transport header with protocol number 5, you need
2801 to manually exclude packets that have a different transport header, for
2802 instance by using meta l4proto 5 before the raw expression.
2803
2804 Table 53. Supported payload protocol bases
2805 ┌─────┬─────────────────────────┐
2806 │Base │ Description │
2807 ├─────┼─────────────────────────┤
2808 │ │ │
2809 │ll │ Link layer, for example │
2810 │ │ the Ethernet header │
2811 ├─────┼─────────────────────────┤
2812 │ │ │
2813 │nh │ Network header, for │
2814 │ │ example IPv4 or IPv6 │
2815 ├─────┼─────────────────────────┤
2816 │ │ │
2817 │th │ Transport Header, for │
2818 │ │ example TCP │
2819 └─────┴─────────────────────────┘
2820
2821 Matching destination port of both UDP and TCP.
2822
2823 inet filter input meta l4proto {tcp, udp} @th,16,16 { 53, 80 }
2824
2825 The above can also be written as
2826
2827 inet filter input meta l4proto {tcp, udp} th dport { 53, 80 }
2828
2829 it is more convenient, but like the raw expression notation no
2830 dependencies are created or checked. It is the users responsibility to
2831 restrict matching to those header types that have a notion of ports.
2832 Otherwise, rules using raw expressions will errnously match unrelated
2833 packets, e.g. mis-interpreting ESP packets SPI field as a port.
2834
2835 Rewrite arp packet target hardware address if target protocol address
2836 matches a given address.
2837
2838 input meta iifname enp2s0 arp ptype 0x0800 arp htype 1 arp hlen 6 arp plen 4 @nh,192,32 0xc0a88f10 @nh,144,48 set 0x112233445566 accept
2839
2840
2841 EXTENSION HEADER EXPRESSIONS
2842 Extension header expressions refer to data from variable-sized protocol
2843 headers, such as IPv6 extension headers, TCP options and IPv4 options.
2844
2845 nftables currently supports matching (finding) a given ipv6 extension
2846 header, TCP option or IPv4 option.
2847
2848 hbh {nexthdr | hdrlength}
2849 frag {nexthdr | frag-off | more-fragments | id}
2850 rt {nexthdr | hdrlength | type | seg-left}
2851 dst {nexthdr | hdrlength}
2852 mh {nexthdr | hdrlength | checksum | type}
2853 srh {flags | tag | sid | seg-left}
2854 tcp option {eol | nop | maxseg | window | sack-perm | sack | sack0 | sack1 | sack2 | sack3 | timestamp} tcp_option_field
2855 ip option { lsrr | ra | rr | ssrr } ip_option_field
2856
2857 The following syntaxes are valid only in a relational expression with
2858 boolean type on right-hand side for checking header existence only:
2859
2860 exthdr {hbh | frag | rt | dst | mh}
2861 tcp option {eol | nop | maxseg | window | sack-perm | sack | sack0 | sack1 | sack2 | sack3 | timestamp}
2862 ip option { lsrr | ra | rr | ssrr }
2863
2864 Table 54. IPv6 extension headers
2865 ┌────────┬────────────────────────┐
2866 │Keyword │ Description │
2867 ├────────┼────────────────────────┤
2868 │ │ │
2869 │hbh │ Hop by Hop │
2870 ├────────┼────────────────────────┤
2871 │ │ │
2872 │rt │ Routing Header │
2873 ├────────┼────────────────────────┤
2874 │ │ │
2875 │frag │ Fragmentation header │
2876 ├────────┼────────────────────────┤
2877 │ │ │
2878 │dst │ dst options │
2879 ├────────┼────────────────────────┤
2880 │ │ │
2881 │mh │ Mobility Header │
2882 ├────────┼────────────────────────┤
2883 │ │ │
2884 │srh │ Segment Routing Header │
2885 └────────┴────────────────────────┘
2886
2887 Table 55. TCP Options
2888 ┌──────────┬─────────────────────┬─────────────────────┐
2889 │Keyword │ Description │ TCP option fields │
2890 ├──────────┼─────────────────────┼─────────────────────┤
2891 │ │ │ │
2892 │eol │ End if option list │ kind │
2893 ├──────────┼─────────────────────┼─────────────────────┤
2894 │ │ │ │
2895 │nop │ 1 Byte TCP Nop │ kind │
2896 │ │ padding option │ │
2897 ├──────────┼─────────────────────┼─────────────────────┤
2898 │ │ │ │
2899 │maxseg │ TCP Maximum Segment │ kind, length, size │
2900 │ │ Size │ │
2901 ├──────────┼─────────────────────┼─────────────────────┤
2902 │ │ │ │
2903 │window │ TCP Window Scaling │ kind, length, count │
2904 ├──────────┼─────────────────────┼─────────────────────┤
2905 │ │ │ │
2906 │sack-perm │ TCP SACK permitted │ kind, length │
2907 ├──────────┼─────────────────────┼─────────────────────┤
2908 │ │ │ │
2909 │sack │ TCP Selective │ kind, length, left, │
2910 │ │ Acknowledgement │ right │
2911 │ │ (alias of block 0) │ │
2912 ├──────────┼─────────────────────┼─────────────────────┤
2913 │ │ │ │
2914 │sack0 │ TCP Selective │ kind, length, left, │
2915 │ │ Acknowledgement │ right │
2916 │ │ (block 0) │ │
2917 ├──────────┼─────────────────────┼─────────────────────┤
2918 │ │ │ │
2919 │sack1 │ TCP Selective │ kind, length, left, │
2920 │ │ Acknowledgement │ right │
2921 │ │ (block 1) │ │
2922 ├──────────┼─────────────────────┼─────────────────────┤
2923 │ │ │ │
2924 │sack2 │ TCP Selective │ kind, length, left, │
2925 │ │ Acknowledgement │ right │
2926 │ │ (block 2) │ │
2927 ├──────────┼─────────────────────┼─────────────────────┤
2928 │ │ │ │
2929 │sack3 │ TCP Selective │ kind, length, left, │
2930 │ │ Acknowledgement │ right │
2931 │ │ (block 3) │ │
2932 ├──────────┼─────────────────────┼─────────────────────┤
2933 │ │ │ │
2934 │timestamp │ TCP Timestamps │ kind, length, │
2935 │ │ │ tsval, tsecr │
2936 └──────────┴─────────────────────┴─────────────────────┘
2937
2938 TCP option matching also supports raw expression syntax to access
2939 arbitrary options:
2940
2941 tcp option
2942
2943 tcp option @number,offset,length
2944
2945 Table 56. IP Options
2946 ┌────────┬─────────────────────┬─────────────────────┐
2947 │Keyword │ Description │ IP option fields │
2948 ├────────┼─────────────────────┼─────────────────────┤
2949 │ │ │ │
2950 │lsrr │ Loose Source Route │ type, length, ptr, │
2951 │ │ │ addr │
2952 ├────────┼─────────────────────┼─────────────────────┤
2953 │ │ │ │
2954 │ra │ Router Alert │ type, length, value │
2955 ├────────┼─────────────────────┼─────────────────────┤
2956 │ │ │ │
2957 │rr │ Record Route │ type, length, ptr, │
2958 │ │ │ addr │
2959 ├────────┼─────────────────────┼─────────────────────┤
2960 │ │ │ │
2961 │ssrr │ Strict Source Route │ type, length, ptr, │
2962 │ │ │ addr │
2963 └────────┴─────────────────────┴─────────────────────┘
2964
2965 finding TCP options.
2966
2967 filter input tcp option sack-perm kind 1 counter
2968
2969 matching IPv6 exthdr.
2970
2971 ip6 filter input frag more-fragments 1 counter
2972
2973 finding IP option.
2974
2975 filter input ip option lsrr exists counter
2976
2977
2978 CONNTRACK EXPRESSIONS
2979 Conntrack expressions refer to meta data of the connection tracking
2980 entry associated with a packet.
2981
2982 There are three types of conntrack expressions. Some conntrack
2983 expressions require the flow direction before the conntrack key, others
2984 must be used directly because they are direction agnostic. The packets,
2985 bytes and avgpkt keywords can be used with or without a direction. If
2986 the direction is omitted, the sum of the original and the reply
2987 direction is returned. The same is true for the zone, if a direction is
2988 given, the zone is only matched if the zone id is tied to the given
2989 direction.
2990
2991 ct {state | direction | status | mark | expiration | helper | label}
2992 ct [original | reply] {l3proto | protocol | bytes | packets | avgpkt | zone | id}
2993 ct {original | reply} {proto-src | proto-dst}
2994 ct {original | reply} {ip | ip6} {saddr | daddr}
2995
2996 The conntrack-specific types in this table are described in the
2997 sub-section CONNTRACK TYPES above.
2998
2999 Table 57. Conntrack expressions
3000 ┌───────────┬─────────────────────┬─────────────────────┐
3001 │Keyword │ Description │ Type │
3002 ├───────────┼─────────────────────┼─────────────────────┤
3003 │ │ │ │
3004 │state │ State of the │ ct_state │
3005 │ │ connection │ │
3006 ├───────────┼─────────────────────┼─────────────────────┤
3007 │ │ │ │
3008 │direction │ Direction of the │ ct_dir │
3009 │ │ packet relative to │ │
3010 │ │ the connection │ │
3011 ├───────────┼─────────────────────┼─────────────────────┤
3012 │ │ │ │
3013 │status │ Status of the │ ct_status │
3014 │ │ connection │ │
3015 ├───────────┼─────────────────────┼─────────────────────┤
3016 │ │ │ │
3017 │mark │ Connection mark │ mark │
3018 ├───────────┼─────────────────────┼─────────────────────┤
3019 │ │ │ │
3020 │expiration │ Connection │ time │
3021 │ │ expiration time │ │
3022 ├───────────┼─────────────────────┼─────────────────────┤
3023 │ │ │ │
3024 │helper │ Helper associated │ string │
3025 │ │ with the connection │ │
3026 ├───────────┼─────────────────────┼─────────────────────┤
3027 │ │ │ │
3028 │label │ Connection tracking │ ct_label │
3029 │ │ label bit or │ │
3030 │ │ symbolic name │ │
3031 │ │ defined in │ │
3032 │ │ connlabel.conf in │ │
3033 │ │ the nftables │ │
3034 │ │ include path │ │
3035 ├───────────┼─────────────────────┼─────────────────────┤
3036 │ │ │ │
3037 │l3proto │ Layer 3 protocol of │ nf_proto │
3038 │ │ the connection │ │
3039 ├───────────┼─────────────────────┼─────────────────────┤
3040 │ │ │ │
3041 │saddr │ Source address of │ ipv4_addr/ipv6_addr │
3042 │ │ the connection for │ │
3043 │ │ the given direction │ │
3044 ├───────────┼─────────────────────┼─────────────────────┤
3045 │ │ │ │
3046 │daddr │ Destination address │ ipv4_addr/ipv6_addr │
3047 │ │ of the connection │ │
3048 │ │ for the given │ │
3049 │ │ direction │ │
3050 ├───────────┼─────────────────────┼─────────────────────┤
3051 │ │ │ │
3052 │protocol │ Layer 4 protocol of │ inet_proto │
3053 │ │ the connection for │ │
3054 │ │ the given direction │ │
3055 ├───────────┼─────────────────────┼─────────────────────┤
3056 │ │ │ │
3057 │proto-src │ Layer 4 protocol │ integer (16 bit) │
3058 │ │ source for the │ │
3059 │ │ given direction │ │
3060 ├───────────┼─────────────────────┼─────────────────────┤
3061 │ │ │ │
3062 │proto-dst │ Layer 4 protocol │ integer (16 bit) │
3063 │ │ destination for the │ │
3064 │ │ given direction │ │
3065 ├───────────┼─────────────────────┼─────────────────────┤
3066 │ │ │ │
3067 │packets │ packet count seen │ integer (64 bit) │
3068 │ │ in the given │ │
3069 │ │ direction or sum of │ │
3070 │ │ original and reply │ │
3071 ├───────────┼─────────────────────┼─────────────────────┤
3072 │ │ │ │
3073 │bytes │ byte count seen, │ integer (64 bit) │
3074 │ │ see description for │ │
3075 │ │ packets keyword │ │
3076 ├───────────┼─────────────────────┼─────────────────────┤
3077 │ │ │ │
3078 │avgpkt │ average bytes per │ integer (64 bit) │
3079 │ │ packet, see │ │
3080 │ │ description for │ │
3081 │ │ packets keyword │ │
3082 ├───────────┼─────────────────────┼─────────────────────┤
3083 │ │ │ │
3084 │zone │ conntrack zone │ integer (16 bit) │
3085 ├───────────┼─────────────────────┼─────────────────────┤
3086 │ │ │ │
3087 │count │ number of current │ integer (32 bit) │
3088 │ │ connections │ │
3089 ├───────────┼─────────────────────┼─────────────────────┤
3090 │ │ │ │
3091 │id │ Connection id │ ct_id │
3092 └───────────┴─────────────────────┴─────────────────────┘
3093
3094 restrict the number of parallel connections to a server.
3095
3096 nft add set filter ssh_flood '{ type ipv4_addr; flags dynamic; }'
3097 nft add rule filter input tcp dport 22 add @ssh_flood '{ ip saddr ct count over 2 }' reject
3098
3099
3101 Statements represent actions to be performed. They can alter control
3102 flow (return, jump to a different chain, accept or drop the packet) or
3103 can perform actions, such as logging, rejecting a packet, etc.
3104
3105 Statements exist in two kinds. Terminal statements unconditionally
3106 terminate evaluation of the current rule, non-terminal statements
3107 either only conditionally or never terminate evaluation of the current
3108 rule, in other words, they are passive from the ruleset evaluation
3109 perspective. There can be an arbitrary amount of non-terminal
3110 statements in a rule, but only a single terminal statement as the final
3111 statement.
3112
3113 VERDICT STATEMENT
3114 The verdict statement alters control flow in the ruleset and issues
3115 policy decisions for packets.
3116
3117 {accept | drop | queue | continue | return}
3118 {jump | goto} chain
3119
3120 accept and drop are absolute verdicts — they terminate ruleset
3121 evaluation immediately.
3122
3123
3124 accept Terminate ruleset
3125 evaluation and accept the
3126 packet. The packet can
3127 still be dropped later by
3128 another hook, for instance
3129 accept in the forward hook
3130 still allows to drop the
3131 packet later in the
3132 postrouting hook, or
3133 another forward base chain
3134 that has a higher priority
3135 number and is evaluated
3136 afterwards in the
3137 processing pipeline.
3138
3139
3140 drop Terminate ruleset
3141 evaluation and drop the
3142 packet. The drop occurs
3143 instantly, no further
3144 chains or hooks are
3145 evaluated. It is not
3146 possible to accept the
3147 packet in a later chain
3148 again, as those are not
3149 evaluated anymore for the
3150 packet.
3151
3152 queue Terminate ruleset
3153 evaluation and queue the
3154 packet to userspace.
3155 Userspace must provide a
3156 drop or accept verdict. In
3157 case of accept, processing
3158 resumes with the next base
3159 chain hook, not the rule
3160 following the queue
3161 verdict.
3162
3163 continue Continue ruleset
3164 evaluation with the next
3165 rule. This is the default
3166 behaviour in case a rule
3167 issues no verdict.
3168
3169 return Return from the current
3170 chain and continue
3171 evaluation at the next
3172 rule in the last chain. If
3173 issued in a base chain, it
3174 is equivalent to the base
3175 chain policy.
3176
3177 jump chain Continue evaluation at the
3178 first rule in chain. The
3179 current position in the
3180 ruleset is pushed to a
3181 call stack and evaluation
3182 will continue there when
3183 the new chain is entirely
3184 evaluated or a return
3185 verdict is issued. In case
3186 an absolute verdict is
3187 issued by a rule in the
3188 chain, ruleset evaluation
3189 terminates immediately and
3190 the specific action is
3191 taken.
3192
3193 goto chain Similar to jump, but the
3194 current position is not
3195 pushed to the call stack,
3196 meaning that after the new
3197 chain evaluation will
3198 continue at the last chain
3199 instead of the one
3200 containing the goto
3201 statement.
3202
3203
3204 Using verdict statements.
3205
3206 # process packets from eth0 and the internal network in from_lan
3207 # chain, drop all packets from eth0 with different source addresses.
3208
3209 filter input iif eth0 ip saddr 192.168.0.0/24 jump from_lan
3210 filter input iif eth0 drop
3211
3212
3213 PAYLOAD STATEMENT
3214 payload_expression set value
3215
3216 The payload statement alters packet content. It can be used for example
3217 to set ip DSCP (diffserv) header field or ipv6 flow labels.
3218
3219 route some packets instead of bridging.
3220
3221 # redirect tcp:http from 192.160.0.0/16 to local machine for routing instead of bridging
3222 # assumes 00:11:22:33:44:55 is local MAC address.
3223 bridge input meta iif eth0 ip saddr 192.168.0.0/16 tcp dport 80 meta pkttype set unicast ether daddr set 00:11:22:33:44:55
3224
3225 Set IPv4 DSCP header field.
3226
3227 ip forward ip dscp set 42
3228
3229
3230 EXTENSION HEADER STATEMENT
3231 extension_header_expression set value
3232
3233 The extension header statement alters packet content in variable-sized
3234 headers. This can currently be used to alter the TCP Maximum segment
3235 size of packets, similar to TCPMSS.
3236
3237 change tcp mss.
3238
3239 tcp flags syn tcp option maxseg size set 1360
3240 # set a size based on route information:
3241 tcp flags syn tcp option maxseg size set rt mtu
3242
3243
3244 LOG STATEMENT
3245 log [prefix quoted_string] [level syslog-level] [flags log-flags]
3246 log group nflog_group [prefix quoted_string] [queue-threshold value] [snaplen size]
3247 log level audit
3248
3249 The log statement enables logging of matching packets. When this
3250 statement is used from a rule, the Linux kernel will print some
3251 information on all matching packets, such as header fields, via the
3252 kernel log (where it can be read with dmesg(1) or read in the syslog).
3253
3254 In the second form of invocation (if nflog_group is specified), the
3255 Linux kernel will pass the packet to nfnetlink_log which will multicast
3256 the packet through a netlink socket to the specified multicast group.
3257 One or more userspace processes may subscribe to the group to receive
3258 the packets, see libnetfilter_queue documentation for details.
3259
3260 In the third form of invocation (if level audit is specified), the
3261 Linux kernel writes a message into the audit buffer suitably formatted
3262 for reading with auditd. Therefore no further formatting options (such
3263 as prefix or flags) are allowed in this mode.
3264
3265 This is a non-terminating statement, so the rule evaluation continues
3266 after the packet is logged.
3267
3268 Table 58. log statement options
3269 ┌────────────────┬─────────────────────┬───────────────────┐
3270 │Keyword │ Description │ Type │
3271 ├────────────────┼─────────────────────┼───────────────────┤
3272 │ │ │ │
3273 │prefix │ Log message prefix │ quoted string │
3274 ├────────────────┼─────────────────────┼───────────────────┤
3275 │ │ │ │
3276 │level │ Syslog level of │ string: emerg, │
3277 │ │ logging │ alert, crit, err, │
3278 │ │ │ warn [default], │
3279 │ │ │ notice, info, │
3280 │ │ │ debug, audit │
3281 ├────────────────┼─────────────────────┼───────────────────┤
3282 │ │ │ │
3283 │group │ NFLOG group to send │ unsigned integer │
3284 │ │ messages to │ (16 bit) │
3285 ├────────────────┼─────────────────────┼───────────────────┤
3286 │ │ │ │
3287 │snaplen │ Length of packet │ unsigned integer │
3288 │ │ payload to include │ (32 bit) │
3289 │ │ in netlink message │ │
3290 ├────────────────┼─────────────────────┼───────────────────┤
3291 │ │ │ │
3292 │queue-threshold │ Number of packets │ unsigned integer │
3293 │ │ to queue inside the │ (32 bit) │
3294 │ │ kernel before │ │
3295 │ │ sending them to │ │
3296 │ │ userspace │ │
3297 └────────────────┴─────────────────────┴───────────────────┘
3298
3299 Table 59. log-flags
3300 ┌─────────────┬───────────────────────────┐
3301 │Flag │ Description │
3302 ├─────────────┼───────────────────────────┤
3303 │ │ │
3304 │tcp sequence │ Log TCP sequence numbers. │
3305 ├─────────────┼───────────────────────────┤
3306 │ │ │
3307 │tcp options │ Log options from the TCP │
3308 │ │ packet header. │
3309 ├─────────────┼───────────────────────────┤
3310 │ │ │
3311 │ip options │ Log options from the │
3312 │ │ IP/IPv6 packet header. │
3313 ├─────────────┼───────────────────────────┤
3314 │ │ │
3315 │skuid │ Log the userid of the │
3316 │ │ process which generated │
3317 │ │ the packet. │
3318 ├─────────────┼───────────────────────────┤
3319 │ │ │
3320 │ether │ Decode MAC addresses and │
3321 │ │ protocol. │
3322 ├─────────────┼───────────────────────────┤
3323 │ │ │
3324 │all │ Enable all log flags │
3325 │ │ listed above. │
3326 └─────────────┴───────────────────────────┘
3327
3328 Using log statement.
3329
3330 # log the UID which generated the packet and ip options
3331 ip filter output log flags skuid flags ip options
3332
3333 # log the tcp sequence numbers and tcp options from the TCP packet
3334 ip filter output log flags tcp sequence,options
3335
3336 # enable all supported log flags
3337 ip6 filter output log flags all
3338
3339
3340 REJECT STATEMENT
3341 reject [ with REJECT_WITH ]
3342
3343 REJECT_WITH := icmp type icmp_code |
3344 icmpv6 type icmpv6_code |
3345 icmpx type icmpx_code |
3346 tcp reset
3347
3348 A reject statement is used to send back an error packet in response to
3349 the matched packet otherwise it is equivalent to drop so it is a
3350 terminating statement, ending rule traversal. This statement is only
3351 valid in base chains using the input, forward or output hooks, and
3352 user-defined chains which are only called from those chains.
3353
3354 Table 60. different ICMP reject variants are meant for use in different
3355 table families
3356 ┌────────┬────────┬─────────────┐
3357 │Variant │ Family │ Type │
3358 ├────────┼────────┼─────────────┤
3359 │ │ │ │
3360 │icmp │ ip │ icmp_code │
3361 ├────────┼────────┼─────────────┤
3362 │ │ │ │
3363 │icmpv6 │ ip6 │ icmpv6_code │
3364 ├────────┼────────┼─────────────┤
3365 │ │ │ │
3366 │icmpx │ inet │ icmpx_code │
3367 └────────┴────────┴─────────────┘
3368
3369 For a description of the different types and a list of supported
3370 keywords refer to DATA TYPES section above. The common default reject
3371 value is port-unreachable.
3372
3373 Note that in bridge family, reject statement is only allowed in base
3374 chains which hook into input or prerouting.
3375
3376 COUNTER STATEMENT
3377 A counter statement sets the hit count of packets along with the number
3378 of bytes.
3379
3380 counter packets number bytes number
3381 counter { packets number | bytes number }
3382
3383 CONNTRACK STATEMENT
3384 The conntrack statement can be used to set the conntrack mark and
3385 conntrack labels.
3386
3387 ct {mark | event | label | zone} set value
3388
3389 The ct statement sets meta data associated with a connection. The zone
3390 id has to be assigned before a conntrack lookup takes place, i.e. this
3391 has to be done in prerouting and possibly output (if locally generated
3392 packets need to be placed in a distinct zone), with a hook priority of
3393 -300.
3394
3395 Unlike iptables, where the helper assignment happens in the raw table,
3396 the helper needs to be assigned after a conntrack entry has been found,
3397 i.e. it will not work when used with hook priorities equal or before
3398 -200.
3399
3400 Table 61. Conntrack statement types
3401 ┌────────┬─────────────────────┬──────────────────┐
3402 │Keyword │ Description │ Value │
3403 ├────────┼─────────────────────┼──────────────────┤
3404 │ │ │ │
3405 │event │ conntrack event │ bitmask, integer │
3406 │ │ bits │ (32 bit) │
3407 ├────────┼─────────────────────┼──────────────────┤
3408 │ │ │ │
3409 │helper │ name of ct helper │ quoted string │
3410 │ │ object to assign to │ │
3411 │ │ the connection │ │
3412 ├────────┼─────────────────────┼──────────────────┤
3413 │ │ │ │
3414 │mark │ Connection tracking │ mark │
3415 │ │ mark │ │
3416 ├────────┼─────────────────────┼──────────────────┤
3417 │ │ │ │
3418 │label │ Connection tracking │ label │
3419 │ │ label │ │
3420 ├────────┼─────────────────────┼──────────────────┤
3421 │ │ │ │
3422 │zone │ conntrack zone │ integer (16 bit) │
3423 └────────┴─────────────────────┴──────────────────┘
3424
3425 save packet nfmark in conntrack.
3426
3427 ct mark set meta mark
3428
3429 set zone mapped via interface.
3430
3431 table inet raw {
3432 chain prerouting {
3433 type filter hook prerouting priority -300;
3434 ct zone set iif map { "eth1" : 1, "veth1" : 2 }
3435 }
3436 chain output {
3437 type filter hook output priority -300;
3438 ct zone set oif map { "eth1" : 1, "veth1" : 2 }
3439 }
3440 }
3441
3442 restrict events reported by ctnetlink.
3443
3444 ct event set new,related,destroy
3445
3446
3447 NOTRACK STATEMENT
3448 The notrack statement allows to disable connection tracking for certain
3449 packets.
3450
3451 notrack
3452
3453 Note that for this statement to be effective, it has to be applied to
3454 packets before a conntrack lookup happens. Therefore, it needs to sit
3455 in a chain with either prerouting or output hook and a hook priority of
3456 -300 or less.
3457
3458 See SYNPROXY STATEMENT for an example usage.
3459
3460 META STATEMENT
3461 A meta statement sets the value of a meta expression. The existing meta
3462 fields are: priority, mark, pkttype, nftrace.
3463
3464 meta {mark | priority | pkttype | nftrace} set value
3465
3466 A meta statement sets meta data associated with a packet.
3467
3468 Table 62. Meta statement types
3469 ┌─────────┬─────────────────────┬───────────┐
3470 │Keyword │ Description │ Value │
3471 ├─────────┼─────────────────────┼───────────┤
3472 │ │ │ │
3473 │priority │ TC packet priority │ tc_handle │
3474 ├─────────┼─────────────────────┼───────────┤
3475 │ │ │ │
3476 │mark │ Packet mark │ mark │
3477 ├─────────┼─────────────────────┼───────────┤
3478 │ │ │ │
3479 │pkttype │ packet type │ pkt_type │
3480 ├─────────┼─────────────────────┼───────────┤
3481 │ │ │ │
3482 │nftrace │ ruleset packet │ 0, 1 │
3483 │ │ tracing on/off. Use │ │
3484 │ │ monitor trace │ │
3485 │ │ command to watch │ │
3486 │ │ traces │ │
3487 └─────────┴─────────────────────┴───────────┘
3488
3489 LIMIT STATEMENT
3490 limit rate [over] packet_number / TIME_UNIT [burst packet_number packets]
3491 limit rate [over] byte_number BYTE_UNIT / TIME_UNIT [burst byte_number BYTE_UNIT]
3492
3493 TIME_UNIT := second | minute | hour | day
3494 BYTE_UNIT := bytes | kbytes | mbytes
3495
3496 A limit statement matches at a limited rate using a token bucket
3497 filter. A rule using this statement will match until this limit is
3498 reached. It can be used in combination with the log statement to give
3499 limited logging. The optional over keyword makes it match over the
3500 specified rate. Default burst is 5. if you specify burst, it must be
3501 non-zero value.
3502
3503 Table 63. limit statement values
3504 ┌──────────────┬───────────────────┬──────────────────┐
3505 │Value │ Description │ Type │
3506 ├──────────────┼───────────────────┼──────────────────┤
3507 │ │ │ │
3508 │packet_number │ Number of packets │ unsigned integer │
3509 │ │ │ (32 bit) │
3510 ├──────────────┼───────────────────┼──────────────────┤
3511 │ │ │ │
3512 │byte_number │ Number of bytes │ unsigned integer │
3513 │ │ │ (32 bit) │
3514 └──────────────┴───────────────────┴──────────────────┘
3515
3516 NAT STATEMENTS
3517 snat to address [:port] [PRF_FLAGS]
3518 snat to address - address [:port - port] [PRF_FLAGS]
3519 snat { ip | ip6 } to address - address [:port - port] [PR_FLAGS]
3520 dnat to address [:port] [PRF_FLAGS]
3521 dnat to address [:port - port] [PR_FLAGS]
3522 dnat { ip | ip6 } to address [:port - port] [PR_FLAGS]
3523 masquerade to [:port] [PRF_FLAGS]
3524 masquerade to [:port - port] [PRF_FLAGS]
3525 redirect to [:port] [PRF_FLAGS]
3526 redirect to [:port - port] [PRF_FLAGS]
3527
3528 PRF_FLAGS := PRF_FLAG [, PRF_FLAGS]
3529 PR_FLAGS := PR_FLAG [, PR_FLAGS]
3530 PRF_FLAG := PR_FLAG | fully-random
3531 PR_FLAG := persistent | random
3532
3533 The nat statements are only valid from nat chain types.
3534
3535 The snat and masquerade statements specify that the source address of
3536 the packet should be modified. While snat is only valid in the
3537 postrouting and input chains, masquerade makes sense only in
3538 postrouting. The dnat and redirect statements are only valid in the
3539 prerouting and output chains, they specify that the destination address
3540 of the packet should be modified. You can use non-base chains which are
3541 called from base chains of nat chain type too. All future packets in
3542 this connection will also be mangled, and rules should cease being
3543 examined.
3544
3545 The masquerade statement is a special form of snat which always uses
3546 the outgoing interface’s IP address to translate to. It is particularly
3547 useful on gateways with dynamic (public) IP addresses.
3548
3549 The redirect statement is a special form of dnat which always
3550 translates the destination address to the local host’s one. It comes in
3551 handy if one only wants to alter the destination port of incoming
3552 traffic on different interfaces.
3553
3554 When used in the inet family (available with kernel 5.2), the dnat and
3555 snat statements require the use of the ip and ip6 keyword in case an
3556 address is provided, see the examples below.
3557
3558 Before kernel 4.18 nat statements require both prerouting and
3559 postrouting base chains to be present since otherwise packets on the
3560 return path won’t be seen by netfilter and therefore no reverse
3561 translation will take place.
3562
3563 Table 64. NAT statement values
3564 ┌───────────┬─────────────────────┬─────────────────────┐
3565 │Expression │ Description │ Type │
3566 ├───────────┼─────────────────────┼─────────────────────┤
3567 │ │ │ │
3568 │address │ Specifies that the │ ipv4_addr, │
3569 │ │ source/destination │ ipv6_addr, e.g. │
3570 │ │ address of the │ abcd::1234, or you │
3571 │ │ packet should be │ can use a mapping, │
3572 │ │ modified. You may │ e.g. meta mark map │
3573 │ │ specify a mapping │ { 10 : 192.168.1.2, │
3574 │ │ to relate a list of │ 20 : 192.168.1.3 } │
3575 │ │ tuples composed of │ │
3576 │ │ arbitrary │ │
3577 │ │ expression key with │ │
3578 │ │ address value. │ │
3579 ├───────────┼─────────────────────┼─────────────────────┤
3580 │ │ │ │
3581 │port │ Specifies that the │ port number (16 │
3582 │ │ source/destination │ bit) │
3583 │ │ address of the │ │
3584 │ │ packet should be │ │
3585 │ │ modified. │ │
3586 └───────────┴─────────────────────┴─────────────────────┘
3587
3588 Table 65. NAT statement flags
3589 ┌─────────────┬─────────────────────────────┐
3590 │Flag │ Description │
3591 ├─────────────┼─────────────────────────────┤
3592 │ │ │
3593 │persistent │ Gives a client the same │
3594 │ │ source-/destination-address │
3595 │ │ for each connection. │
3596 ├─────────────┼─────────────────────────────┤
3597 │ │ │
3598 │random │ In kernel 5.0 and newer │
3599 │ │ this is the same as │
3600 │ │ fully-random. In earlier │
3601 │ │ kernels the port mapping │
3602 │ │ will be randomized using a │
3603 │ │ seeded MD5 hash mix using │
3604 │ │ source and destination │
3605 │ │ address and destination │
3606 │ │ port. │
3607 ├─────────────┼─────────────────────────────┤
3608 │ │ │
3609 │fully-random │ If used then port mapping │
3610 │ │ is generated based on a │
3611 │ │ 32-bit pseudo-random │
3612 │ │ algorithm. │
3613 └─────────────┴─────────────────────────────┘
3614
3615 Using NAT statements.
3616
3617 # create a suitable table/chain setup for all further examples
3618 add table nat
3619 add chain nat prerouting { type nat hook prerouting priority 0; }
3620 add chain nat postrouting { type nat hook postrouting priority 100; }
3621
3622 # translate source addresses of all packets leaving via eth0 to address 1.2.3.4
3623 add rule nat postrouting oif eth0 snat to 1.2.3.4
3624
3625 # redirect all traffic entering via eth0 to destination address 192.168.1.120
3626 add rule nat prerouting iif eth0 dnat to 192.168.1.120
3627
3628 # translate source addresses of all packets leaving via eth0 to whatever
3629 # locally generated packets would use as source to reach the same destination
3630 add rule nat postrouting oif eth0 masquerade
3631
3632 # redirect incoming TCP traffic for port 22 to port 2222
3633 add rule nat prerouting tcp dport 22 redirect to :2222
3634
3635 # inet family:
3636 # handle ip dnat:
3637 add rule inet nat prerouting dnat ip to 10.0.2.99
3638 # handle ip6 dnat:
3639 add rule inet nat prerouting dnat ip6 to fe80::dead
3640 # this masquerades both ipv4 and ipv6:
3641 add rule inet nat postrouting meta oif ppp0 masquerade
3642
3643
3644 TPROXY STATEMENT
3645 Tproxy redirects the packet to a local socket without changing the
3646 packet header in any way. If any of the arguments is missing the data
3647 of the incoming packet is used as parameter. Tproxy matching requires
3648 another rule that ensures the presence of transport protocol header is
3649 specified.
3650
3651 tproxy to address:port
3652 tproxy to {address | :port}
3653
3654 This syntax can be used in ip/ip6 tables where network layer protocol
3655 is obvious. Either IP address or port can be specified, but at least
3656 one of them is necessary.
3657
3658 tproxy {ip | ip6} to address[:port]
3659 tproxy to :port
3660
3661 This syntax can be used in inet tables. The ip/ip6 parameter defines
3662 the family the rule will match. The address parameter must be of this
3663 family. When only port is defined, the address family should not be
3664 specified. In this case the rule will match for both families.
3665
3666 Table 66. tproxy attributes
3667 ┌────────┬────────────────────────────┐
3668 │Name │ Description │
3669 ├────────┼────────────────────────────┤
3670 │ │ │
3671 │address │ IP address the listening │
3672 │ │ socket with IP_TRANSPARENT │
3673 │ │ option is bound to. │
3674 ├────────┼────────────────────────────┤
3675 │ │ │
3676 │port │ Port the listening socket │
3677 │ │ with IP_TRANSPARENT option │
3678 │ │ is bound to. │
3679 └────────┴────────────────────────────┘
3680
3681 Example ruleset for tproxy statement.
3682
3683 table ip x {
3684 chain y {
3685 type filter hook prerouting priority -150; policy accept;
3686 tcp dport ntp tproxy to 1.1.1.1
3687 udp dport ssh tproxy to :2222
3688 }
3689 }
3690 table ip6 x {
3691 chain y {
3692 type filter hook prerouting priority -150; policy accept;
3693 tcp dport ntp tproxy to [dead::beef]
3694 udp dport ssh tproxy to :2222
3695 }
3696 }
3697 table inet x {
3698 chain y {
3699 type filter hook prerouting priority -150; policy accept;
3700 tcp dport 321 tproxy to :ssh
3701 tcp dport 99 tproxy ip to 1.1.1.1:999
3702 udp dport 155 tproxy ip6 to [dead::beef]:smux
3703 }
3704 }
3705
3706
3707 SYNPROXY STATEMENT
3708 This statement will process TCP three-way-handshake parallel in
3709 netfilter context to protect either local or backend system. This
3710 statement requires connection tracking because sequence numbers need to
3711 be translated.
3712
3713 synproxy [mss mss_value] [wscale wscale_value] [SYNPROXY_FLAGS]
3714
3715 Table 67. synproxy statement attributes
3716 ┌───────┬────────────────────────────┐
3717 │Name │ Description │
3718 ├───────┼────────────────────────────┤
3719 │ │ │
3720 │mss │ Maximum segment size │
3721 │ │ announced to clients. This │
3722 │ │ must match the backend. │
3723 ├───────┼────────────────────────────┤
3724 │ │ │
3725 │wscale │ Window scale announced to │
3726 │ │ clients. This must match │
3727 │ │ the backend. │
3728 └───────┴────────────────────────────┘
3729
3730 Table 68. synproxy statement flags
3731 ┌──────────┬────────────────────────────┐
3732 │Flag │ Description │
3733 ├──────────┼────────────────────────────┤
3734 │ │ │
3735 │sack-perm │ Pass client selective │
3736 │ │ acknowledgement option to │
3737 │ │ backend (will be disabled │
3738 │ │ if not present). │
3739 ├──────────┼────────────────────────────┤
3740 │ │ │
3741 │timestamp │ Pass client timestamp │
3742 │ │ option to backend (will be │
3743 │ │ disabled if not present, │
3744 │ │ also needed for selective │
3745 │ │ acknowledgement and window │
3746 │ │ scaling). │
3747 └──────────┴────────────────────────────┘
3748
3749 Example ruleset for synproxy statement.
3750
3751 Determine tcp options used by backend, from an external system
3752
3753 tcpdump -pni eth0 -c 1 'tcp[tcpflags] == (tcp-syn|tcp-ack)'
3754 port 80 &
3755 telnet 192.0.2.42 80
3756 18:57:24.693307 IP 192.0.2.42.80 > 192.0.2.43.48757:
3757 Flags [S.], seq 360414582, ack 788841994, win 14480,
3758 options [mss 1460,sackOK,
3759 TS val 1409056151 ecr 9690221,
3760 nop,wscale 9],
3761 length 0
3762
3763 Switch tcp_loose mode off, so conntrack will mark out-of-flow packets as state INVALID.
3764
3765 echo 0 > /proc/sys/net/netfilter/nf_conntrack_tcp_loose
3766
3767 Make SYN packets untracked.
3768
3769 table ip x {
3770 chain y {
3771 type filter hook prerouting priority raw; policy accept;
3772 tcp flags syn notrack
3773 }
3774 }
3775
3776 Catch UNTRACKED (SYN packets) and INVALID (3WHS ACK packets) states and send
3777 them to SYNPROXY. This rule will respond to SYN packets with SYN+ACK
3778 syncookies, create ESTABLISHED for valid client response (3WHS ACK packets) and
3779 drop incorrect cookies. Flags combinations not expected during 3WHS will not
3780 match and continue (e.g. SYN+FIN, SYN+ACK). Finally, drop invalid packets, this
3781 will be out-of-flow packets that were not matched by SYNPROXY.
3782
3783 table ip foo {
3784 chain z {
3785 type filter hook input priority filter; policy accept;
3786 ct state { invalid, untracked } synproxy mss 1460 wscale 9 timestamp sack-perm
3787 ct state invalid drop
3788 }
3789 }
3790
3791 The outcome ruleset of the steps above should be similar to the one below.
3792
3793 table ip x {
3794 chain y {
3795 type filter hook prerouting priority raw; policy accept;
3796 tcp flags syn notrack
3797 }
3798
3799 chain z {
3800 type filter hook input priority filter; policy accept;
3801 ct state { invalid, untracked } synproxy mss 1460 wscale 9 timestamp sack-perm
3802 ct state invalid drop
3803 }
3804 }
3805
3806
3807 FLOW STATEMENT
3808 A flow statement allows us to select what flows you want to accelerate
3809 forwarding through layer 3 network stack bypass. You have to specify
3810 the flowtable name where you want to offload this flow.
3811
3812 flow add @flowtable
3813
3814 QUEUE STATEMENT
3815 This statement passes the packet to userspace using the nfnetlink_queue
3816 handler. The packet is put into the queue identified by its 16-bit
3817 queue number. Userspace can inspect and modify the packet if desired.
3818 Userspace must then drop or re-inject the packet into the kernel. See
3819 libnetfilter_queue documentation for details.
3820
3821 queue [num queue_number] [bypass]
3822 queue [num queue_number_from - queue_number_to] [QUEUE_FLAGS]
3823
3824 QUEUE_FLAGS := QUEUE_FLAG [, QUEUE_FLAGS]
3825 QUEUE_FLAG := bypass | fanout
3826
3827 Table 69. queue statement values
3828 ┌──────────────────┬────────────────────┬──────────────────┐
3829 │Value │ Description │ Type │
3830 ├──────────────────┼────────────────────┼──────────────────┤
3831 │ │ │ │
3832 │queue_number │ Sets queue number, │ unsigned integer │
3833 │ │ default is 0. │ (16 bit) │
3834 ├──────────────────┼────────────────────┼──────────────────┤
3835 │ │ │ │
3836 │queue_number_from │ Sets initial queue │ unsigned integer │
3837 │ │ in the range, if │ (16 bit) │
3838 │ │ fanout is used. │ │
3839 ├──────────────────┼────────────────────┼──────────────────┤
3840 │ │ │ │
3841 │queue_number_to │ Sets closing queue │ unsigned integer │
3842 │ │ in the range, if │ (16 bit) │
3843 │ │ fanout is used. │ │
3844 └──────────────────┴────────────────────┴──────────────────┘
3845
3846 Table 70. queue statement flags
3847 ┌───────┬────────────────────────────┐
3848 │Flag │ Description │
3849 ├───────┼────────────────────────────┤
3850 │ │ │
3851 │bypass │ Let packets go through if │
3852 │ │ userspace application │
3853 │ │ cannot back off. Before │
3854 │ │ using this flag, read │
3855 │ │ libnetfilter_queue │
3856 │ │ documentation for │
3857 │ │ performance tuning │
3858 │ │ recommendations. │
3859 ├───────┼────────────────────────────┤
3860 │ │ │
3861 │fanout │ Distribute packets between │
3862 │ │ several queues. │
3863 └───────┴────────────────────────────┘
3864
3865 DUP STATEMENT
3866 The dup statement is used to duplicate a packet and send the copy to a
3867 different destination.
3868
3869 dup to device
3870 dup to address device device
3871
3872 Table 71. Dup statement values
3873 ┌───────────┬─────────────────────┬─────────────────────┐
3874 │Expression │ Description │ Type │
3875 ├───────────┼─────────────────────┼─────────────────────┤
3876 │ │ │ │
3877 │address │ Specifies that the │ ipv4_addr, │
3878 │ │ copy of the packet │ ipv6_addr, e.g. │
3879 │ │ should be sent to a │ abcd::1234, or you │
3880 │ │ new gateway. │ can use a mapping, │
3881 │ │ │ e.g. ip saddr map { │
3882 │ │ │ 192.168.1.2 : │
3883 │ │ │ 10.1.1.1 } │
3884 ├───────────┼─────────────────────┼─────────────────────┤
3885 │ │ │ │
3886 │device │ Specifies that the │ string │
3887 │ │ copy should be │ │
3888 │ │ transmitted via │ │
3889 │ │ device. │ │
3890 └───────────┴─────────────────────┴─────────────────────┘
3891
3892 Using the dup statement.
3893
3894 # send to machine with ip address 10.2.3.4 on eth0
3895 ip filter forward dup to 10.2.3.4 device "eth0"
3896
3897 # copy raw frame to another interface
3898 netdetv ingress dup to "eth0"
3899 dup to "eth0"
3900
3901 # combine with map dst addr to gateways
3902 dup to ip daddr map { 192.168.7.1 : "eth0", 192.168.7.2 : "eth1" }
3903
3904
3905 FWD STATEMENT
3906 The fwd statement is used to redirect a raw packet to another
3907 interface. It is only available in the netdev family ingress hook. It
3908 is similar to the dup statement except that no copy is made.
3909
3910 fwd to device
3911
3912 SET STATEMENT
3913 The set statement is used to dynamically add or update elements in a
3914 set from the packet path. The set setname must already exist in the
3915 given table and must have been created with one or both of the dynamic
3916 and the timeout flags. The dynamic flag is required if the set
3917 statement expression includes a stateful object. The timeout flag is
3918 implied if the set is created with a timeout, and is required if the
3919 set statement updates elements, rather than adding them. Furthermore,
3920 these sets should specify both a maximum set size (to prevent memory
3921 exhaustion), and their elements should have a timeout (so their number
3922 will not grow indefinitely) either from the set definition or from the
3923 statement that adds or updates them. The set statement can be used to
3924 e.g. create dynamic blacklists.
3925
3926 {add | update} @setname { expression [timeout timeout] [comment string] }
3927
3928 Example for simple blacklist.
3929
3930 # declare a set, bound to table "filter", in family "ip".
3931 # Timeout and size are mandatory because we will add elements from packet path.
3932 # Entries will timeout after one minute, after which they might be
3933 # re-added if limit condition persists.
3934 nft add set ip filter blackhole \
3935 "{ type ipv4_addr; flags dynamic; timeout 1m; size 65536; }"
3936
3937 # declare a set to store the limit per saddr.
3938 # This must be separate from blackhole since the timeout is different
3939 nft add set ip filter flood \
3940 "{ type ipv4_addr; flags dynamic; timeout 10s; size 128000; }"
3941
3942 # whitelist internal interface.
3943 nft add rule ip filter input meta iifname "internal" accept
3944
3945 # drop packets coming from blacklisted ip addresses.
3946 nft add rule ip filter input ip saddr @blackhole counter drop
3947
3948 # add source ip addresses to the blacklist if more than 10 tcp connection
3949 # requests occurred per second and ip address.
3950 nft add rule ip filter input tcp flags syn tcp dport ssh \
3951 add @flood { ip saddr limit rate over 10/second } \
3952 add @blackhole { ip saddr } drop
3953
3954 # inspect state of the sets.
3955 nft list set ip filter flood
3956 nft list set ip filter blackhole
3957
3958 # manually add two addresses to the blackhole.
3959 nft add element filter blackhole { 10.2.3.4, 10.23.1.42 }
3960
3961
3962 MAP STATEMENT
3963 The map statement is used to lookup data based on some specific input
3964 key.
3965
3966 expression map { MAP_ELEMENTS }
3967
3968 MAP_ELEMENTS := MAP_ELEMENT [, MAP_ELEMENTS]
3969 MAP_ELEMENT := key : value
3970
3971 The key is a value returned by expression.
3972
3973 Using the map statement.
3974
3975 # select DNAT target based on TCP dport:
3976 # connections to port 80 are redirected to 192.168.1.100,
3977 # connections to port 8888 are redirected to 192.168.1.101
3978 nft add rule ip nat prerouting dnat tcp dport map { 80 : 192.168.1.100, 8888 : 192.168.1.101 }
3979
3980 # source address based SNAT:
3981 # packets from net 192.168.1.0/24 will appear as originating from 10.0.0.1,
3982 # packets from net 192.168.2.0/24 will appear as originating from 10.0.0.2
3983 nft add rule ip nat postrouting snat to ip saddr map { 192.168.1.0/24 : 10.0.0.1, 192.168.2.0/24 : 10.0.0.2 }
3984
3985
3986 VMAP STATEMENT
3987 The verdict map (vmap) statement works analogous to the map statement,
3988 but contains verdicts as values.
3989
3990 expression vmap { VMAP_ELEMENTS }
3991
3992 VMAP_ELEMENTS := VMAP_ELEMENT [, VMAP_ELEMENTS]
3993 VMAP_ELEMENT := key : verdict
3994
3995 Using the vmap statement.
3996
3997 # jump to different chains depending on layer 4 protocol type:
3998 nft add rule ip filter input ip protocol vmap { tcp : jump tcp-chain, udp : jump udp-chain , icmp : jump icmp-chain }
3999
4000
4002 These are some additional commands included in nft.
4003
4004 MONITOR
4005 The monitor command allows you to listen to Netlink events produced by
4006 the nf_tables subsystem, related to creation and deletion of objects.
4007 When they occur, nft will print to stdout the monitored events in
4008 either JSON or native nft format.
4009
4010 To filter events related to a concrete object, use one of the keywords
4011 tables, chains, sets, rules, elements, ruleset.
4012
4013 To filter events related to a concrete action, use keyword new or
4014 destroy.
4015
4016 Hit ^C to finish the monitor operation.
4017
4018 Listen to all events, report in native nft format.
4019
4020 % nft monitor
4021
4022 Listen to deleted rules, report in JSON format.
4023
4024 % nft -j monitor destroy rules
4025
4026 Listen to both new and destroyed chains, in native nft format.
4027
4028 % nft monitor chains
4029
4030 Listen to ruleset events such as table, chain, rule, set, counters and
4031 quotas, in native nft format.
4032
4033 % nft monitor ruleset
4034
4035
4037 When an error is detected, nft shows the line(s) containing the error,
4038 the position of the erroneous parts in the input stream and marks up
4039 the erroneous parts using carets (^). If the error results from the
4040 combination of two expressions or statements, the part imposing the
4041 constraints which are violated is marked using tildes (~).
4042
4043 For errors returned by the kernel, nft cannot detect which parts of the
4044 input caused the error and the entire command is marked.
4045
4046 Error caused by single incorrect expression.
4047
4048 <cmdline>:1:19-22: Error: Interface does not exist
4049 filter output oif eth0
4050 ^^^^
4051
4052 Error caused by invalid combination of two expressions.
4053
4054 <cmdline>:1:28-36: Error: Right hand side of relational expression (==) must be constant
4055 filter output tcp dport == tcp dport
4056 ~~ ^^^^^^^^^
4057
4058 Error returned by the kernel.
4059
4060 <cmdline>:0:0-23: Error: Could not process rule: Operation not permitted
4061 filter output oif wlan0
4062 ^^^^^^^^^^^^^^^^^^^^^^^
4063
4064
4066 On success, nft exits with a status of 0. Unspecified errors cause it
4067 to exit with a status of 1, memory allocation errors with a status of
4068 2, unable to open Netlink socket with 3.
4069
4071 libnftables(3), libnftables-json(5), iptables(8), ip6tables(8), arptables(8), ebtables(8), ip(8), tc(8)
4072
4073 There is an official wiki at: https://wiki.nftables.org
4074
4076 nftables was written by Patrick McHardy and Pablo Neira Ayuso, among
4077 many other contributors from the Netfilter community.
4078
4080 Copyright © 2008-2014 Patrick McHardy <kaber@trash.net> Copyright ©
4081 2013-2018 Pablo Neira Ayuso <pablo@netfilter.org>
4082
4083 nftables is free software; you can redistribute it and/or modify it
4084 under the terms of the GNU General Public License version 2 as
4085 published by the Free Software Foundation.
4086
4087 This documentation is licensed under the terms of the Creative Commons
4088 Attribution-ShareAlike 4.0 license, CC BY-SA 4.0
4089 http://creativecommons.org/licenses/by-sa/4.0/.
4090
4091
4092
4093 01/15/2021 NFT(8)