tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
1# SPDX-License-Identifier: GPL-2.0-only
2#
3# Network configuration
4#
5
6menuconfig NET
7 bool "Networking support"
8 select NLATTR
9 select GENERIC_NET_UTILS
10 select BPF
11 help
12 Unless you really know what you are doing, you should say Y here.
13 The reason is that some programs need kernel networking support even
14 when running on a stand-alone machine that isn't connected to any
15 other computer.
16
17 If you are upgrading from an older kernel, you
18 should consider updating your networking tools too because changes
19 in the kernel and the tools often go hand in hand. The tools are
20 contained in the package net-tools, the location and version number
21 of which are given in <file:Documentation/Changes>.
22
23 For a general introduction to Linux networking, it is highly
24 recommended to read the NET-HOWTO, available from
25 <http://www.tldp.org/docs.html#howto>.
26
27if NET
28
29config WANT_COMPAT_NETLINK_MESSAGES
30 bool
31 help
32 This option can be selected by other options that need compat
33 netlink messages.
34
35config COMPAT_NETLINK_MESSAGES
36 def_bool y
37 depends on COMPAT
38 depends on WEXT_CORE || WANT_COMPAT_NETLINK_MESSAGES
39 help
40 This option makes it possible to send different netlink messages
41 to tasks depending on whether the task is a compat task or not. To
42 achieve this, you need to set skb_shinfo(skb)->frag_list to the
43 compat skb before sending the skb, the netlink code will sort out
44 which message to actually pass to the task.
45
46 Newly written code should NEVER need this option but do
47 compat-independent messages instead!
48
49config NET_INGRESS
50 bool
51
52config NET_EGRESS
53 bool
54
55config NET_XGRESS
56 select NET_INGRESS
57 select NET_EGRESS
58 bool
59
60config NET_REDIRECT
61 bool
62
63config SKB_EXTENSIONS
64 bool
65
66menu "Networking options"
67
68source "net/packet/Kconfig"
69source "net/unix/Kconfig"
70source "net/tls/Kconfig"
71source "net/xfrm/Kconfig"
72source "net/iucv/Kconfig"
73source "net/smc/Kconfig"
74source "net/xdp/Kconfig"
75
76config NET_HANDSHAKE
77 bool
78 depends on SUNRPC || NVME_TARGET_TCP || NVME_TCP
79 default y
80
81config NET_HANDSHAKE_KUNIT_TEST
82 tristate "KUnit tests for the handshake upcall mechanism" if !KUNIT_ALL_TESTS
83 default KUNIT_ALL_TESTS
84 depends on KUNIT
85 help
86 This builds the KUnit tests for the handshake upcall mechanism.
87
88 KUnit tests run during boot and output the results to the debug
89 log in TAP format (https://testanything.org/). Only useful for
90 kernel devs running KUnit test harness and are not for inclusion
91 into a production build.
92
93 For more information on KUnit and unit tests in general, refer
94 to the KUnit documentation in Documentation/dev-tools/kunit/.
95
96config INET
97 bool "TCP/IP networking"
98 help
99 These are the protocols used on the Internet and on most local
100 Ethernets. It is highly recommended to say Y here (this will enlarge
101 your kernel by about 400 KB), since some programs (e.g. the X window
102 system) use TCP/IP even if your machine is not connected to any
103 other computer. You will get the so-called loopback device which
104 allows you to ping yourself (great fun, that!).
105
106 For an excellent introduction to Linux networking, please read the
107 Linux Networking HOWTO, available from
108 <http://www.tldp.org/docs.html#howto>.
109
110 If you say Y here and also to "/proc file system support" and
111 "Sysctl support" below, you can change various aspects of the
112 behavior of the TCP/IP code by writing to the (virtual) files in
113 /proc/sys/net/ipv4/*; the options are explained in the file
114 <file:Documentation/networking/ip-sysctl.rst>.
115
116 Short answer: say Y.
117
118if INET
119source "net/ipv4/Kconfig"
120source "net/ipv6/Kconfig"
121source "net/netlabel/Kconfig"
122source "net/mptcp/Kconfig"
123
124endif # if INET
125
126config NETWORK_SECMARK
127 bool "Security Marking"
128 help
129 This enables security marking of network packets, similar
130 to nfmark, but designated for security purposes.
131 If you are unsure how to answer this question, answer N.
132
133config NET_PTP_CLASSIFY
134 def_bool n
135
136config NETWORK_PHY_TIMESTAMPING
137 bool "Timestamping in PHY devices"
138 select NET_PTP_CLASSIFY
139 help
140 This allows timestamping of network packets by PHYs (or
141 other MII bus snooping devices) with hardware timestamping
142 capabilities. This option adds some overhead in the transmit
143 and receive paths.
144
145 If you are unsure how to answer this question, answer N.
146
147menuconfig NETFILTER
148 bool "Network packet filtering framework (Netfilter)"
149 help
150 Netfilter is a framework for filtering and mangling network packets
151 that pass through your Linux box.
152
153 The most common use of packet filtering is to run your Linux box as
154 a firewall protecting a local network from the Internet. The type of
155 firewall provided by this kernel support is called a "packet
156 filter", which means that it can reject individual network packets
157 based on type, source, destination etc. The other kind of firewall,
158 a "proxy-based" one, is more secure but more intrusive and more
159 bothersome to set up; it inspects the network traffic much more
160 closely, modifies it and has knowledge about the higher level
161 protocols, which a packet filter lacks. Moreover, proxy-based
162 firewalls often require changes to the programs running on the local
163 clients. Proxy-based firewalls don't need support by the kernel, but
164 they are often combined with a packet filter, which only works if
165 you say Y here.
166
167 You should also say Y here if you intend to use your Linux box as
168 the gateway to the Internet for a local network of machines without
169 globally valid IP addresses. This is called "masquerading": if one
170 of the computers on your local network wants to send something to
171 the outside, your box can "masquerade" as that computer, i.e. it
172 forwards the traffic to the intended outside destination, but
173 modifies the packets to make it look like they came from the
174 firewall box itself. It works both ways: if the outside host
175 replies, the Linux box will silently forward the traffic to the
176 correct local computer. This way, the computers on your local net
177 are completely invisible to the outside world, even though they can
178 reach the outside and can receive replies. It is even possible to
179 run globally visible servers from within a masqueraded local network
180 using a mechanism called portforwarding. Masquerading is also often
181 called NAT (Network Address Translation).
182
183 Another use of Netfilter is in transparent proxying: if a machine on
184 the local network tries to connect to an outside host, your Linux
185 box can transparently forward the traffic to a local server,
186 typically a caching proxy server.
187
188 Yet another use of Netfilter is building a bridging firewall. Using
189 a bridge with Network packet filtering enabled makes iptables "see"
190 the bridged traffic. For filtering on the lower network and Ethernet
191 protocols over the bridge, use ebtables (under bridge netfilter
192 configuration).
193
194 Various modules exist for netfilter which replace the previous
195 masquerading (ipmasqadm), packet filtering (ipchains), transparent
196 proxying, and portforwarding mechanisms. Please see
197 <file:Documentation/Changes> under "iptables" for the location of
198 these packages.
199
200if NETFILTER
201
202config NETFILTER_ADVANCED
203 bool "Advanced netfilter configuration"
204 depends on NETFILTER
205 default y
206 help
207 If you say Y here you can select between all the netfilter modules.
208 If you say N the more unusual ones will not be shown and the
209 basic ones needed by most people will default to 'M'.
210
211 If unsure, say Y.
212
213config BRIDGE_NETFILTER
214 tristate "Bridged IP/ARP packets filtering"
215 depends on BRIDGE
216 depends on NETFILTER && INET
217 depends on NETFILTER_ADVANCED
218 select NETFILTER_FAMILY_BRIDGE
219 select SKB_EXTENSIONS
220 help
221 Enabling this option will let arptables resp. iptables see bridged
222 ARP resp. IP traffic. If you want a bridging firewall, you probably
223 want this option enabled.
224 Enabling or disabling this option doesn't enable or disable
225 ebtables.
226
227 If unsure, say N.
228
229source "net/netfilter/Kconfig"
230source "net/ipv4/netfilter/Kconfig"
231source "net/ipv6/netfilter/Kconfig"
232source "net/bridge/netfilter/Kconfig"
233
234endif
235
236source "net/bpfilter/Kconfig"
237
238source "net/dccp/Kconfig"
239source "net/sctp/Kconfig"
240source "net/rds/Kconfig"
241source "net/tipc/Kconfig"
242source "net/atm/Kconfig"
243source "net/l2tp/Kconfig"
244source "net/802/Kconfig"
245source "net/bridge/Kconfig"
246source "net/dsa/Kconfig"
247source "net/8021q/Kconfig"
248source "net/llc/Kconfig"
249source "drivers/net/appletalk/Kconfig"
250source "net/x25/Kconfig"
251source "net/lapb/Kconfig"
252source "net/phonet/Kconfig"
253source "net/6lowpan/Kconfig"
254source "net/ieee802154/Kconfig"
255source "net/mac802154/Kconfig"
256source "net/sched/Kconfig"
257source "net/dcb/Kconfig"
258source "net/dns_resolver/Kconfig"
259source "net/batman-adv/Kconfig"
260source "net/openvswitch/Kconfig"
261source "net/vmw_vsock/Kconfig"
262source "net/netlink/Kconfig"
263source "net/mpls/Kconfig"
264source "net/nsh/Kconfig"
265source "net/hsr/Kconfig"
266source "net/switchdev/Kconfig"
267source "net/l3mdev/Kconfig"
268source "net/qrtr/Kconfig"
269source "net/ncsi/Kconfig"
270
271config PCPU_DEV_REFCNT
272 bool "Use percpu variables to maintain network device refcount"
273 depends on SMP
274 default y
275 help
276 network device refcount are using per cpu variables if this option is set.
277 This can be forced to N to detect underflows (with a performance drop).
278
279config MAX_SKB_FRAGS
280 int "Maximum number of fragments per skb_shared_info"
281 range 17 45
282 default 17
283 help
284 Having more fragments per skb_shared_info can help GRO efficiency.
285 This helps BIG TCP workloads, but might expose bugs in some
286 legacy drivers.
287 This also increases memory overhead of small packets,
288 and in drivers using build_skb().
289 If unsure, say 17.
290
291config RPS
292 bool
293 depends on SMP && SYSFS
294 default y
295
296config RFS_ACCEL
297 bool
298 depends on RPS
299 select CPU_RMAP
300 default y
301
302config SOCK_RX_QUEUE_MAPPING
303 bool
304
305config XPS
306 bool
307 depends on SMP
308 select SOCK_RX_QUEUE_MAPPING
309 default y
310
311config HWBM
312 bool
313
314config CGROUP_NET_PRIO
315 bool "Network priority cgroup"
316 depends on CGROUPS
317 select SOCK_CGROUP_DATA
318 help
319 Cgroup subsystem for use in assigning processes to network priorities on
320 a per-interface basis.
321
322config CGROUP_NET_CLASSID
323 bool "Network classid cgroup"
324 depends on CGROUPS
325 select SOCK_CGROUP_DATA
326 help
327 Cgroup subsystem for use as general purpose socket classid marker that is
328 being used in cls_cgroup and for netfilter matching.
329
330config NET_RX_BUSY_POLL
331 bool
332 default y if !PREEMPT_RT || (PREEMPT_RT && !NETCONSOLE)
333
334config BQL
335 bool
336 depends on SYSFS
337 select DQL
338 default y
339
340config BPF_STREAM_PARSER
341 bool "enable BPF STREAM_PARSER"
342 depends on INET
343 depends on BPF_SYSCALL
344 depends on CGROUP_BPF
345 select STREAM_PARSER
346 select NET_SOCK_MSG
347 help
348 Enabling this allows a TCP stream parser to be used with
349 BPF_MAP_TYPE_SOCKMAP.
350
351config NET_FLOW_LIMIT
352 bool
353 depends on RPS
354 default y
355 help
356 The network stack has to drop packets when a receive processing CPU's
357 backlog reaches netdev_max_backlog. If a few out of many active flows
358 generate the vast majority of load, drop their traffic earlier to
359 maintain capacity for the other flows. This feature provides servers
360 with many clients some protection against DoS by a single (spoofed)
361 flow that greatly exceeds average workload.
362
363menu "Network testing"
364
365config NET_PKTGEN
366 tristate "Packet Generator (USE WITH CAUTION)"
367 depends on INET && PROC_FS
368 help
369 This module will inject preconfigured packets, at a configurable
370 rate, out of a given interface. It is used for network interface
371 stress testing and performance analysis. If you don't understand
372 what was just said, you don't need it: say N.
373
374 Documentation on how to use the packet generator can be found
375 at <file:Documentation/networking/pktgen.rst>.
376
377 To compile this code as a module, choose M here: the
378 module will be called pktgen.
379
380config NET_DROP_MONITOR
381 tristate "Network packet drop alerting service"
382 depends on INET && TRACEPOINTS
383 help
384 This feature provides an alerting service to userspace in the
385 event that packets are discarded in the network stack. Alerts
386 are broadcast via netlink socket to any listening user space
387 process. If you don't need network drop alerts, or if you are ok
388 just checking the various proc files and other utilities for
389 drop statistics, say N here.
390
391endmenu
392
393endmenu
394
395source "net/ax25/Kconfig"
396source "net/can/Kconfig"
397source "net/bluetooth/Kconfig"
398source "net/rxrpc/Kconfig"
399source "net/kcm/Kconfig"
400source "net/strparser/Kconfig"
401source "net/mctp/Kconfig"
402
403config FIB_RULES
404 bool
405
406menuconfig WIRELESS
407 bool "Wireless"
408 depends on !S390
409 default y
410
411if WIRELESS
412
413source "net/wireless/Kconfig"
414source "net/mac80211/Kconfig"
415
416endif # WIRELESS
417
418source "net/rfkill/Kconfig"
419source "net/9p/Kconfig"
420source "net/caif/Kconfig"
421source "net/ceph/Kconfig"
422source "net/nfc/Kconfig"
423source "net/psample/Kconfig"
424source "net/ife/Kconfig"
425
426config LWTUNNEL
427 bool "Network light weight tunnels"
428 help
429 This feature provides an infrastructure to support light weight
430 tunnels like mpls. There is no netdevice associated with a light
431 weight tunnel endpoint. Tunnel encapsulation parameters are stored
432 with light weight tunnel state associated with fib routes.
433
434config LWTUNNEL_BPF
435 bool "Execute BPF program as route nexthop action"
436 depends on LWTUNNEL && INET
437 default y if LWTUNNEL=y
438 help
439 Allows to run BPF programs as a nexthop action following a route
440 lookup for incoming and outgoing packets.
441
442config DST_CACHE
443 bool
444 default n
445
446config GRO_CELLS
447 bool
448 default n
449
450config SOCK_VALIDATE_XMIT
451 bool
452
453config NET_SELFTESTS
454 def_tristate PHYLIB
455 depends on PHYLIB && INET
456
457config NET_SOCK_MSG
458 bool
459 default n
460 help
461 The NET_SOCK_MSG provides a framework for plain sockets (e.g. TCP) or
462 ULPs (upper layer modules, e.g. TLS) to process L7 application data
463 with the help of BPF programs.
464
465config NET_DEVLINK
466 bool
467 default n
468
469config PAGE_POOL
470 bool
471
472config PAGE_POOL_STATS
473 default n
474 bool "Page pool stats"
475 depends on PAGE_POOL
476 help
477 Enable page pool statistics to track page allocation and recycling
478 in page pools. This option incurs additional CPU cost in allocation
479 and recycle paths and additional memory cost to store the statistics.
480 These statistics are only available if this option is enabled and if
481 the driver using the page pool supports exporting this data.
482
483 If unsure, say N.
484
485config FAILOVER
486 tristate "Generic failover module"
487 help
488 The failover module provides a generic interface for paravirtual
489 drivers to register a netdev and a set of ops with a failover
490 instance. The ops are used as event handlers that get called to
491 handle netdev register/unregister/link change/name change events
492 on slave pci ethernet devices with the same mac address as the
493 failover netdev. This enables paravirtual drivers to use a
494 VF as an accelerated low latency datapath. It also allows live
495 migration of VMs with direct attached VFs by failing over to the
496 paravirtual datapath when the VF is unplugged.
497
498config ETHTOOL_NETLINK
499 bool "Netlink interface for ethtool"
500 default y
501 help
502 An alternative userspace interface for ethtool based on generic
503 netlink. It provides better extensibility and some new features,
504 e.g. notification messages.
505
506config NETDEV_ADDR_LIST_TEST
507 tristate "Unit tests for device address list"
508 default KUNIT_ALL_TESTS
509 depends on KUNIT
510
511endif # if NET