tjh.dev
Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
1Documentation 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>
6
7For general info and legal blurb, please look in README.
8
9==============================================================
10
11This file contains the documentation for the sysctl files in
12/proc/sys/net
13
14The interface to the networking parts of the kernel is located in
15/proc/sys/net. The following table shows all possible subdirectories. You may
16see only some of them, depending on your kernel's configuration.
17
18
19Table : 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..............................................................................
31
321. /proc/sys/net/core - Network core options
33-------------------------------------------------------
34
35bpf_jit_enable
36--------------
37
38This enables Berkeley Packet Filter Just in Time compiler.
39Currently supported on x86_64 architecture, bpf_jit provides a framework
40to speed packet filtering, the one used by tcpdump/libpcap for example.
41Values :
42 0 - disable the JIT (default value)
43 1 - enable the JIT
44 2 - enable the JIT and ask the compiler to emit traces on kernel log.
45
46dev_weight
47--------------
48
49The maximum number of packets that kernel can handle on a NAPI interrupt,
50it's a Per-CPU variable.
51Default: 64
52
53default_qdisc
54--------------
55
56The default queuing discipline to use for network devices. This allows
57overriding the default queue discipline of pfifo_fast with an
58alternative. Since the default queuing discipline is created with the
59no additional parameters so is best suited to queuing disciplines that
60work well without configuration like stochastic fair queue (sfq),
61CoDel (codel) or fair queue CoDel (fq_codel). Don't use queuing disciplines
62like Hierarchical Token Bucket or Deficit Round Robin which require setting
63up classes and bandwidths.
64Default: pfifo_fast
65
66busy_read
67----------------
68Low latency busy poll timeout for socket reads. (needs CONFIG_NET_RX_BUSY_POLL)
69Approximate time in us to busy loop waiting for packets on the device queue.
70This sets the default value of the SO_BUSY_POLL socket option.
71Can be set or overridden per socket by setting socket option SO_BUSY_POLL,
72which is the preferred method of enabling. If you need to enable the feature
73globally via sysctl, a value of 50 is recommended.
74Will increase power usage.
75Default: 0 (off)
76
77busy_poll
78----------------
79Low latency busy poll timeout for poll and select. (needs CONFIG_NET_RX_BUSY_POLL)
80Approximate time in us to busy loop waiting for events.
81Recommended value depends on the number of sockets you poll on.
82For several sockets 50, for several hundreds 100.
83For more than that you probably want to use epoll.
84Note that only sockets with SO_BUSY_POLL set will be busy polled,
85so you want to either selectively set SO_BUSY_POLL on those sockets or set
86sysctl.net.busy_read globally.
87Will increase power usage.
88Default: 0 (off)
89
90rmem_default
91------------
92
93The default setting of the socket receive buffer in bytes.
94
95rmem_max
96--------
97
98The maximum receive socket buffer size in bytes.
99
100wmem_default
101------------
102
103The default setting (in bytes) of the socket send buffer.
104
105wmem_max
106--------
107
108The maximum send socket buffer size in bytes.
109
110message_burst and message_cost
111------------------------------
112
113These parameters are used to limit the warning messages written to the kernel
114log from the networking code. They enforce a rate limit to make a
115denial-of-service attack impossible. A higher message_cost factor, results in
116fewer messages that will be written. Message_burst controls when messages will
117be dropped. The default settings limit warning messages to one every five
118seconds.
119
120warnings
121--------
122
123This controls console messages from the networking stack that can occur because
124of problems on the network like duplicate address or bad checksums. Normally,
125this should be enabled, but if the problem persists the messages can be
126disabled.
127
128netdev_budget
129-------------
130
131Maximum number of packets taken from all interfaces in one polling cycle (NAPI
132poll). In one polling cycle interfaces which are registered to polling are
133probed in a round-robin manner.
134
135netdev_max_backlog
136------------------
137
138Maximum number of packets, queued on the INPUT side, when the interface
139receives packets faster than kernel can process them.
140
141netdev_tstamp_prequeue
142----------------------
143
144If set to 0, RX packet timestamps can be sampled after RPS processing, when
145the target CPU processes packets. It might give some delay on timestamps, but
146permit to distribute the load on several cpus.
147
148If set to 1 (default), timestamps are sampled as soon as possible, before
149queueing.
150
151optmem_max
152----------
153
154Maximum ancillary buffer size allowed per socket. Ancillary data is a sequence
155of struct cmsghdr structures with appended data.
156
1572. /proc/sys/net/unix - Parameters for Unix domain sockets
158-------------------------------------------------------
159
160There is only one file in this directory.
161unix_dgram_qlen limits the max number of datagrams queued in Unix domain
162socket's buffer. It will not take effect unless PF_UNIX flag is specified.
163
164
1653. /proc/sys/net/ipv4 - IPV4 settings
166-------------------------------------------------------
167Please see: Documentation/networking/ip-sysctl.txt and ipvs-sysctl.txt for
168descriptions of these entries.
169
170
1714. Appletalk
172-------------------------------------------------------
173
174The /proc/sys/net/appletalk directory holds the Appletalk configuration data
175when Appletalk is loaded. The configurable parameters are:
176
177aarp-expiry-time
178----------------
179
180The amount of time we keep an ARP entry before expiring it. Used to age out
181old hosts.
182
183aarp-resolve-time
184-----------------
185
186The amount of time we will spend trying to resolve an Appletalk address.
187
188aarp-retransmit-limit
189---------------------
190
191The number of times we will retransmit a query before giving up.
192
193aarp-tick-time
194--------------
195
196Controls the rate at which expires are checked.
197
198The directory /proc/net/appletalk holds the list of active Appletalk sockets
199on a machine.
200
201The fields indicate the DDP type, the local address (in network:node format)
202the remote address, the size of the transmit pending queue, the size of the
203received queue (bytes waiting for applications to read) the state and the uid
204owning the socket.
205
206/proc/net/atalk_iface lists all the interfaces configured for appletalk.It
207shows the name of the interface, its Appletalk address, the network range on
208that address (or network number for phase 1 networks), and the status of the
209interface.
210
211/proc/net/atalk_route lists each known network route. It lists the target
212(network) that the route leads to, the router (may be directly connected), the
213route flags, and the device the route is using.
214
215
2165. IPX
217-------------------------------------------------------
218
219The IPX protocol has no tunable values in proc/sys/net.
220
221The IPX protocol does, however, provide proc/net/ipx. This lists each IPX
222socket giving the local and remote addresses in Novell format (that is
223network:node:port). In accordance with the strange Novell tradition,
224everything but the port is in hex. Not_Connected is displayed for sockets that
225are not tied to a specific remote address. The Tx and Rx queue sizes indicate
226the number of bytes pending for transmission and reception. The state
227indicates the state the socket is in and the uid is the owning uid of the
228socket.
229
230The /proc/net/ipx_interface file lists all IPX interfaces. For each interface
231it gives the network number, the node number, and indicates if the network is
232the primary network. It also indicates which device it is bound to (or
233Internal for internal networks) and the Frame Type if appropriate. Linux
234supports 802.3, 802.2, 802.2 SNAP and DIX (Blue Book) ethernet framing for
235IPX.
236
237The /proc/net/ipx_route table holds a list of IPX routes. For each route it
238gives the destination network, the router node (or Directly) and the network
239address of the router (or Connected) for internal networks.
240
2416. TIPC
242-------------------------------------------------------
243
244The TIPC protocol now has a tunable for the receive memory, similar to the
245tcp_rmem - i.e. a vector of 3 INTEGERs: (min, default, max)
246
247 # cat /proc/sys/net/tipc/tipc_rmem
248 4252725 34021800 68043600
249 #
250
251The max value is set to CONN_OVERLOAD_LIMIT, and the default and min values
252are scaled (shifted) versions of that same value. Note that the min value
253is not at this point in time used in any meaningful way, but the triplet is
254preserved in order to be consistent with things like tcp_rmem.