2 .\" Copyright (c) 2009 Robert N. M. Watson
3 .\" All rights reserved.
5 .\" Redistribution and use in source and binary forms, with or without
6 .\" modification, are permitted provided that the following conditions
8 .\" 1. Redistributions of source code must retain the above copyright
9 .\" notice(s), this list of conditions and the following disclaimer as
10 .\" the first lines of this file unmodified other than the possible
11 .\" addition of one or more copyright notices.
12 .\" 2. Redistributions in binary form must reproduce the above copyright
13 .\" notice(s), this list of conditions and the following disclaimer in the
14 .\" documentation and/or other materials provided with the distribution.
16 .\" THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDER(S) ``AS IS'' AND ANY
17 .\" EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
18 .\" WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE
19 .\" DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER(S) BE LIABLE FOR ANY
20 .\" DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES
21 .\" (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR
22 .\" SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER
23 .\" CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24 .\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
35 .Nd Kernel network dispatch service
39 .Fn netisr_register "const struct netisr_handler *nhp"
41 .Fn netisr_unregister "const struct netisr_handler *nhp"
43 .Fn netisr_dispatch "u_int proto" "struct mbuf *m"
45 .Fn netisr_dispatch_src "u_int proto" "uintptr_t source" "struct mbuf *m"
47 .Fn netisr_queue "u_int proto" "struct mbuf *m"
49 .Fn netisr_queue_src "u_int proto" "uintptr_t source" "struct mbuf *m"
51 .Fn netisr_clearqdrops "const struct netisr_handler *nhp"
53 .Fn netisr_getqdrops "const struct netisr_handler *nhp" "uint64_t *qdropsp"
55 .Fn netisr_getqlimit "const struct netisr_handler *nhp" "u_int *qlimitp"
57 .Fn netisr_setqlimit "const struct netisr_handler *nhp" "u_int qlimit"
59 .Fn netisr_default_flow2cpu "u_int flowid"
61 .Fn netisr_get_cpucount "void"
63 .Fn netisr_get_cpuid "u_int cpunumber"
67 kernel interface suite allows device drivers (and other packet sources) to
68 direct packets to protocols for directly dispatched or deferred processing.
69 Protocol registration and work stream statistics may be monitored using
71 .Ss Protocol registration
72 Protocols register and unregister handlers using
75 .Fn netisr_unregister ,
76 and may also manage queue limits and statistics using the
77 .Fn netisr_clearqdrops ,
78 .Fn netisr_getqdrops ,
79 .Fn netisr_getqlimit ,
81 .Fn netisr_setqlimit .
84 supports multi-processor execution of handlers, and relies on a combination
85 of source ordering and protocol-specific ordering and work-placement
86 policies to decide how to distribute work across one or more worker
88 Registering protocols will declare one of three policies:
89 .Bl -tag -width NETISR_POLICY_SOURCE
90 .It Dv NETISR_POLICY_SOURCE
92 should maintain source ordering without advice from the protocol.
94 will ignore any flow IDs present on
96 headers for the purposes of work placement.
97 .It Dv NETISR_POLICY_FLOW
99 should maintain flow ordering as defined by the
101 header flow ID field.
102 If the protocol implements
106 will query the protocol in the event that the
108 doesn't have a flow ID, falling back on source ordering.
109 .It NETISR_POLICY_CPU
111 will entirely delegate all work placement decisions to the protocol,
117 Registration is declared using
118 .Vt "struct netisr_handler" ,
119 whose fields are defined as follows:
120 .Bl -tag -width "netisr_handler_t nh_handler"
121 .It Vt "const char *" Va nh_name
122 Unique character string name of the protocol, which may be included in
124 MIB names, so should not contain whitespace.
125 .It Vt netisr_handler_t Va nh_handler
126 Protocol handler function that will be invoked on each packet received for
128 .It Vt netisr_m2flow_t Va nh_m2flow
129 Optional protocol function to generate a flow ID and set a valid
130 hashtype for packets that enter the
133 .Dv M_HASHTYPE_GET(m)
135 .Dv M_HASHTYPE_NONE .
136 Will be used only with
137 .Dv NETISR_POLICY_FLOW .
138 .It Vt netisr_m2cpuid_t Va nh_m2cpuid
139 Protocol function to determine what CPU a packet should be processed on.
140 Will be used only with
141 .Dv NETISR_POLICY_CPU .
142 .It Vt netisr_drainedcpu_t Va nh_drainedcpu
143 Optional callback function that will be invoked when a per-CPU queue
145 It will never fire for directly dispatched packets.
146 Unless fully understood, this special-purpose function should not be used.
147 .\" In case you intend to use this please send 50 chocolate bars to each
148 .\" of rwatson and bz and wait for an answer.
149 .It Vt u_int Va nh_proto
150 Protocol number used by both protocols to identify themselves to
152 and by packet sources to select what handler will be used to process
154 A table of supported protocol numbers appears below.
155 For implementation reasons, protocol numbers great than 15 are currently
157 .It Vt u_int Va nh_qlimit
158 The maximum per-CPU queue depth for the protocol; due to internal
159 implementation details, the effective queue depth may be as much as twice
161 .It Vt u_int Va nh_policy
162 The ordering and work placement policy for the protocol, as described
165 .Ss Packet source interface
166 Packet sources, such as network interfaces, may request protocol processing
172 Both accept a protocol number and
176 will always execute the protocol handler asynchronously in a deferred
179 will optionally direct dispatch if permitted by global and per-protocol
182 In order to provide additional load balancing and flow information,
183 packet sources may also specify an opaque source identifier, which in
184 practice might be a network interface number or socket pointer, using
186 .Fn netisr_dispatch_src
190 .Ss Protocol number constants
191 The follow protocol numbers are currently defined:
192 .Bl -tag -width NETISR_ROUTE
198 Routing socket loopback
210 This manual page and the
212 implementation were written by
213 .An Robert N. M. Watson .