Penguin
Recent Changes
Annotated edit history of icmp(7) version 1, including all changes. View license author blame.
Rev Author # Line
1 perry 1 ICMP
2 !!!ICMP
3 NAME
4 DESCRIPTION
5 SYSCTLS
6 NOTES
7 VERSIONS
8 SEE ALSO
9 ----
10 !!NAME
11
12
13 icmp, IPPROTO_ICMP - Linux IPv4 ICMP kernel module.
14 !!DESCRIPTION
15
16
17 This kernel protocol module implements the Internet Control
18 Message Protocol defined in RFC792. It is used to signal
19 error conditions and for diagnosis. The user doesn't
20 interact directly with this module; instead it communicates
21 with the other protocols in the kernel and these pass the
22 ICMP errors to the application layers. The kernel ICMP
23 module also answers ICMP requests.
24
25
26 A user protocol may receive ICMP packets for all local
27 sockets by opening a raw socket with the protocol
28 __IPPROTO_ICMP__. See raw(7) for more information.
29 The types of ICMP packets passed to the socket can be
30 filtered using the __ICMP_FILTER__ socket option. ICMP
31 packets are always processed by the kernel too, even when
32 passed to a user socket.
33
34
35 Linux limits the rate of ICMP error packets to each
36 destination. __ICMP_REDIRECT__ and
37 __ICMP_DEST_UNREACH__ are also limited by the destination
38 route of the incoming packets.
39 !!SYSCTLS
40
41
42 ICMP supports a sysctl interface to configure some global IP
43 parameters. The sysctls can be accessed by reading or
44 writing the __/proc/sys/net/ipv4/*__ files or with the
45 sysctl(2) interface. Most of these sysctls are rate
46 limitations for specific ICMP types. Linux 2.2 uses a token
47 bucket filter to limit ICMPs. The value is the timeout in
48 jiffies until the token bucket filter is cleared after a
49 burst. A jiffy is a system dependent unit, usually 10ms on
50 x86 and about 1ms on alpha and IA64.
51
52
53 __icmp_destunreach_rate__
54
55
56 Maximum rate to send ICMP Destination Unreachable packets.
57 This limits the rate at which packets are sent to any
58 individual route or destination. The limit does not affect
59 sending of __ICMP_FRAG_NEEDED__ packets needed for path
60 MTU discovery.
61
62
63 __icmp_echo_ignore_all__
64
65
66 If this value is non-zero, Linux will ignore all
67 __ICMP_ECHO__ requests.
68
69
70 __icmp_echo_ignore_broadcasts__
71
72
73 If this value is non-zero, Linux will ignore all
74 __ICMP_ECHO__ packets sent to broadcast
75 addresses.
76
77
78 __icmp_echoreply_rate__
79
80
81 Maximum rate for sending __ICMP_ECHOREPLY__ packets in
82 response to __ICMP_ECHOREQUEST__ packets.
83
84
85 __icmp_paramprob_rate__
86
87
88 Maximum rate for sending __ICMP_PARAMETERPROB__ packets.
89 These packets are sent when a packet arrives with an invalid
90 IP header.
91
92
93 __icmp_timeexceed_rate__
94
95
96 Maximum rate for sending __ICMP_TIME_EXCEEDED__ packets.
97 These packets are sent to prevent loops when a packet has
98 crossed too many hops.
99 !!NOTES
100
101
102 As many other implementations don't support
103 __IPPROTO_ICMP__ raw sockets, this feature should not be
104 relied on in portable programs.
105
106
107 __ICMP_REDIRECT__ packets are not sent when Linux is not
108 acting as a router. They are also only accepted from the old
109 gateway defined in the routing table and the redirect routes
110 are expired after some time.
111
112
113 The 64-bit timestamp returned by __ICMP_TIMESTAMP__ is in
114 milliseconds since January 1, 1970.
115
116
117 Linux ICMP internally uses a raw socket to send ICMPs. This
118 raw socket may appear in netstat(8) output with a
119 zero inode.
120 !!VERSIONS
121
122
123 Support for the __ICMP_ADDRESS__ request was removed in
124 2.2.
125
126
127 Support for __ICMP_SOURCE_QUENCH__ was removed in Linux
128 2.2.
129 !!SEE ALSO
130
131
132 ip(7)
133
134
135 RFC792 for a description of the ICMP protocol.
136 ----
This page is a man page (or other imported legacy content). We are unable to automatically determine the license status of this page.