1 Documentation for /proc/sys/net/*
2 (c) 1999 Terrehon Bowden <terrehon@pacbell.net>
3 Bodo Bauer <bb@ricochet.net>
4 (c) 2000 Jorge Nerin <comandante@zaralinux.com>
5 (c) 2009 Shen Feng <shen@cn.fujitsu.com>
7 For general info and legal blurb, please look in README.
9 ==============================================================
11 This file contains the documentation for the sysctl files in
14 The interface to the networking parts of the kernel is located in
15 /proc/sys/net. The following table shows all possible subdirectories. You may
16 see only some of them, depending on your kernel's configuration.
19 Table : Subdirectories in /proc/sys/net
20 ..............................................................................
21 Directory Content Directory Content
22 core General parameter appletalk Appletalk protocol
23 unix Unix domain sockets netrom NET/ROM
24 802 E802 protocol ax25 AX25
25 ethernet Ethernet protocol rose X.25 PLP layer
26 ipv4 IP version 4 x25 X.25 protocol
27 ipx IPX token-ring IBM token ring
28 bridge Bridging decnet DEC net
29 ipv6 IP version 6 tipc TIPC
30 ..............................................................................
32 1. /proc/sys/net/core - Network core options
33 -------------------------------------------------------
38 This enables the BPF Just in Time (JIT) compiler. BPF is a flexible
39 and efficient infrastructure allowing to execute bytecode at various
40 hook points. It is used in a number of Linux kernel subsystems such
41 as networking (e.g. XDP, tc), tracing (e.g. kprobes, uprobes, tracepoints)
42 and security (e.g. seccomp). LLVM has a BPF back end that can compile
43 restricted C into a sequence of BPF instructions. After program load
44 through bpf(2) and passing a verifier in the kernel, a JIT will then
45 translate these BPF proglets into native CPU instructions. There are
46 two flavors of JITs, the newer eBPF JIT currently supported on:
53 And the older cBPF JIT supported on the following archs:
59 eBPF JITs are a superset of cBPF JITs, meaning the kernel will
60 migrate cBPF instructions into eBPF instructions and then JIT
61 compile them transparently. Older cBPF JITs can only translate
62 tcpdump filters, seccomp rules, etc, but not mentioned eBPF
63 programs loaded through bpf(2).
66 0 - disable the JIT (default value)
68 2 - enable the JIT and ask the compiler to emit traces on kernel log.
73 This enables hardening for the BPF JIT compiler. Supported are eBPF
74 JIT backends. Enabling hardening trades off performance, but can
75 mitigate JIT spraying.
77 0 - disable JIT hardening (default value)
78 1 - enable JIT hardening for unprivileged users only
79 2 - enable JIT hardening for all users
84 When BPF JIT compiler is enabled, then compiled images are unknown
85 addresses to the kernel, meaning they neither show up in traces nor
86 in /proc/kallsyms. This enables export of these addresses, which can
87 be used for debugging/tracing. If bpf_jit_harden is enabled, this
90 0 - disable JIT kallsyms export (default value)
91 1 - enable JIT kallsyms export for privileged users only
96 The maximum number of packets that kernel can handle on a NAPI interrupt,
97 it's a Per-CPU variable.
103 RPS (e.g. RFS, aRFS) processing is competing with the registered NAPI poll function
104 of the driver for the per softirq cycle netdev_budget. This parameter influences
105 the proportion of the configured netdev_budget that is spent on RPS based packet
106 processing during RX softirq cycles. It is further meant for making current
107 dev_weight adaptable for asymmetric CPU needs on RX/TX side of the network stack.
108 (see dev_weight_tx_bias) It is effective on a per CPU basis. Determination is based
109 on dev_weight and is calculated multiplicative (dev_weight * dev_weight_rx_bias).
115 Scales the maximum number of packets that can be processed during a TX softirq cycle.
116 Effective on a per CPU basis. Allows scaling of current dev_weight for asymmetric
117 net stack processing needs. Be careful to avoid making TX softirq processing a CPU hog.
118 Calculation is based on dev_weight (dev_weight * dev_weight_tx_bias).
124 The default queuing discipline to use for network devices. This allows
125 overriding the default of pfifo_fast with an alternative. Since the default
126 queuing discipline is created without additional parameters so is best suited
127 to queuing disciplines that work well without configuration like stochastic
128 fair queue (sfq), CoDel (codel) or fair queue CoDel (fq_codel). Don't use
129 queuing disciplines like Hierarchical Token Bucket or Deficit Round Robin
130 which require setting up classes and bandwidths. Note that physical multiqueue
131 interfaces still use mq as root qdisc, which in turn uses this default for its
132 leaves. Virtual devices (like e.g. lo or veth) ignore this setting and instead
138 Low latency busy poll timeout for socket reads. (needs CONFIG_NET_RX_BUSY_POLL)
139 Approximate time in us to busy loop waiting for packets on the device queue.
140 This sets the default value of the SO_BUSY_POLL socket option.
141 Can be set or overridden per socket by setting socket option SO_BUSY_POLL,
142 which is the preferred method of enabling. If you need to enable the feature
143 globally via sysctl, a value of 50 is recommended.
144 Will increase power usage.
149 Low latency busy poll timeout for poll and select. (needs CONFIG_NET_RX_BUSY_POLL)
150 Approximate time in us to busy loop waiting for events.
151 Recommended value depends on the number of sockets you poll on.
152 For several sockets 50, for several hundreds 100.
153 For more than that you probably want to use epoll.
154 Note that only sockets with SO_BUSY_POLL set will be busy polled,
155 so you want to either selectively set SO_BUSY_POLL on those sockets or set
156 sysctl.net.busy_read globally.
157 Will increase power usage.
163 The default setting of the socket receive buffer in bytes.
168 The maximum receive socket buffer size in bytes.
172 Allow processes to receive tx timestamps looped together with the original
173 packet contents. If disabled, transmit timestamp requests from unprivileged
174 processes are dropped unless socket option SOF_TIMESTAMPING_OPT_TSONLY is set.
181 The default setting (in bytes) of the socket send buffer.
186 The maximum send socket buffer size in bytes.
188 message_burst and message_cost
189 ------------------------------
191 These parameters are used to limit the warning messages written to the kernel
192 log from the networking code. They enforce a rate limit to make a
193 denial-of-service attack impossible. A higher message_cost factor, results in
194 fewer messages that will be written. Message_burst controls when messages will
195 be dropped. The default settings limit warning messages to one every five
201 This sysctl is now unused.
203 This was used to control console messages from the networking stack that
204 occur because of problems on the network like duplicate address or bad
207 These messages are now emitted at KERN_DEBUG and can generally be enabled
208 and controlled by the dynamic_debug facility.
213 Maximum number of packets taken from all interfaces in one polling cycle (NAPI
214 poll). In one polling cycle interfaces which are registered to polling are
215 probed in a round-robin manner. Also, a polling cycle may not exceed
216 netdev_budget_usecs microseconds, even if netdev_budget has not been
220 ---------------------
222 Maximum number of microseconds in one NAPI polling cycle. Polling
223 will exit when either netdev_budget_usecs have elapsed during the
224 poll cycle or the number of packets processed reaches netdev_budget.
229 Maximum number of packets, queued on the INPUT side, when the interface
230 receives packets faster than kernel can process them.
235 RSS (Receive Side Scaling) enabled drivers use a 40 bytes host key that is
237 Some user space might need to gather its content even if drivers do not
238 provide ethtool -x support yet.
240 myhost:~# cat /proc/sys/net/core/netdev_rss_key
241 84:50:f4:00:a8:15:d1:a7:e9:7f:1d:60:35:c7:47:25:42:97:74:ca:56:bb:b6:a1:d8: ... (52 bytes total)
243 File contains nul bytes if no driver ever called netdev_rss_key_fill() function.
245 /proc/sys/net/core/netdev_rss_key contains 52 bytes of key,
246 but most drivers only use 40 bytes of it.
248 myhost:~# ethtool -x eth0
249 RX flow hash indirection table for eth0 with 8 RX ring(s):
252 84:50:f4:00:a8:15:d1:a7:e9:7f:1d:60:35:c7:47:25:42:97:74:ca:56:bb:b6:a1:d8:43:e3:c9:0c:fd:17:55:c2:3a:4d:69:ed:f1:42:89
254 netdev_tstamp_prequeue
255 ----------------------
257 If set to 0, RX packet timestamps can be sampled after RPS processing, when
258 the target CPU processes packets. It might give some delay on timestamps, but
259 permit to distribute the load on several cpus.
261 If set to 1 (default), timestamps are sampled as soon as possible, before
267 Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
268 of struct cmsghdr structures with appended data.
270 2. /proc/sys/net/unix - Parameters for Unix domain sockets
271 -------------------------------------------------------
273 There is only one file in this directory.
274 unix_dgram_qlen limits the max number of datagrams queued in Unix domain
275 socket's buffer. It will not take effect unless PF_UNIX flag is specified.
278 3. /proc/sys/net/ipv4 - IPV4 settings
279 -------------------------------------------------------
280 Please see: Documentation/networking/ip-sysctl.txt and ipvs-sysctl.txt for
281 descriptions of these entries.
285 -------------------------------------------------------
287 The /proc/sys/net/appletalk directory holds the Appletalk configuration data
288 when Appletalk is loaded. The configurable parameters are:
293 The amount of time we keep an ARP entry before expiring it. Used to age out
299 The amount of time we will spend trying to resolve an Appletalk address.
301 aarp-retransmit-limit
302 ---------------------
304 The number of times we will retransmit a query before giving up.
309 Controls the rate at which expires are checked.
311 The directory /proc/net/appletalk holds the list of active Appletalk sockets
314 The fields indicate the DDP type, the local address (in network:node format)
315 the remote address, the size of the transmit pending queue, the size of the
316 received queue (bytes waiting for applications to read) the state and the uid
319 /proc/net/atalk_iface lists all the interfaces configured for appletalk.It
320 shows the name of the interface, its Appletalk address, the network range on
321 that address (or network number for phase 1 networks), and the status of the
324 /proc/net/atalk_route lists each known network route. It lists the target
325 (network) that the route leads to, the router (may be directly connected), the
326 route flags, and the device the route is using.
330 -------------------------------------------------------
332 The IPX protocol has no tunable values in proc/sys/net.
334 The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
335 socket giving the local and remote addresses in Novell format (that is
336 network:node:port). In accordance with the strange Novell tradition,
337 everything but the port is in hex. Not_Connected is displayed for sockets that
338 are not tied to a specific remote address. The Tx and Rx queue sizes indicate
339 the number of bytes pending for transmission and reception. The state
340 indicates the state the socket is in and the uid is the owning uid of the
343 The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
344 it gives the network number, the node number, and indicates if the network is
345 the primary network. It also indicates which device it is bound to (or
346 Internal for internal networks) and the Frame Type if appropriate. Linux
347 supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
350 The /proc/net/ipx_route table holds a list of IPX routes. For each route it
351 gives the destination network, the router node (or Directly) and the network
352 address of the router (or Connected) for internal networks.
355 -------------------------------------------------------
360 The TIPC protocol now has a tunable for the receive memory, similar to the
361 tcp_rmem - i.e. a vector of 3 INTEGERs: (min, default, max)
363 # cat /proc/sys/net/tipc/tipc_rmem
364 4252725 34021800 68043600
367 The max value is set to CONN_OVERLOAD_LIMIT, and the default and min values
368 are scaled (shifted) versions of that same value. Note that the min value
369 is not at this point in time used in any meaningful way, but the triplet is
370 preserved in order to be consistent with things like tcp_rmem.
375 TIPC name table updates are distributed asynchronously in a cluster, without
376 any form of transaction handling. This means that different race scenarios are
377 possible. One such is that a name withdrawal sent out by one node and received
378 by another node may arrive after a second, overlapping name publication already
379 has been accepted from a third node, although the conflicting updates
380 originally may have been issued in the correct sequential order.
381 If named_timeout is nonzero, failed topology updates will be placed on a defer
382 queue until another event arrives that clears the error, or until the timeout
383 expires. Value is in milliseconds.