xref: /freebsd/share/man/man9/netisr.9 (revision d9a42747950146bf03cda7f6e25d219253f8a57a)
1.\"
2.\" Copyright (c) 2009 Robert N. M. Watson
3.\" All rights reserved.
4.\"
5.\" Redistribution and use in source and binary forms, with or without
6.\" modification, are permitted provided that the following conditions
7.\" are met:
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.
15.\"
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
26.\" DAMAGE.
27.\"
28.\" $FreeBSD$
29.\"
30.Dd April 12, 2023
31.Dt NETISR 9
32.Os
33.Sh NAME
34.Nm netisr
35.Nd Kernel network dispatch service
36.Sh SYNOPSIS
37.In net/netisr.h
38.Ft void
39.Fn netisr_register "const struct netisr_handler *nhp"
40.Ft void
41.Fn netisr_unregister "const struct netisr_handler *nhp"
42.Ft int
43.Fn netisr_dispatch "u_int proto" "struct mbuf *m"
44.Ft int
45.Fn netisr_dispatch_src "u_int proto" "uintptr_t source" "struct mbuf *m"
46.Ft int
47.Fn netisr_queue "u_int proto" "struct mbuf *m"
48.Ft int
49.Fn netisr_queue_src "u_int proto" "uintptr_t source" "struct mbuf *m"
50.Ft void
51.Fn netisr_clearqdrops "const struct netisr_handler *nhp"
52.Ft void
53.Fn netisr_getqdrops "const struct netisr_handler *nhp" "uint64_t *qdropsp"
54.Ft void
55.Fn netisr_getqlimit "const struct netisr_handler *nhp" "u_int *qlimitp"
56.Ft int
57.Fn netisr_setqlimit "const struct netisr_handler *nhp" "u_int qlimit"
58.Ft u_int
59.Fn netisr_default_flow2cpu "u_int flowid"
60.Ft u_int
61.Fn netisr_get_cpucount "void"
62.Ft u_int
63.Fn netisr_get_cpuid "u_int cpunumber"
64.Pp
65With optional virtual network stack support enabled via the following kernel
66compile option:
67.Bd -ragged -offset indent
68.Cd "options VIMAGE"
69.Ed
70.Ft void
71.Fn netisr_register_vnet "const struct netisr_handler *nhp"
72.Ft void
73.Fn netisr_unregister_vnet "const struct netisr_handler *nhp"
74.Sh DESCRIPTION
75The
76.Nm
77kernel interface suite allows device drivers (and other packet sources) to
78direct packets to protocols for directly dispatched or deferred processing.
79Protocol registration and work stream statistics may be monitored using
80.Xr netstat 1 .
81.Ss Protocol registration
82Protocols register and unregister handlers using
83.Fn netisr_register
84and
85.Fn netisr_unregister ,
86and may also manage queue limits and statistics using the
87.Fn netisr_clearqdrops ,
88.Fn netisr_getqdrops ,
89.Fn netisr_getqlimit ,
90and
91.Fn netisr_setqlimit .
92.Pp
93In case of VIMAGE kernels each virtual network stack (vnet), that is not the
94default base system network stack, calls
95.Fn netisr_register_vnet
96and
97.Fn netisr_unregister_vnet
98to enable or disable packet processing by the
99.Nm
100for each protocol.
101Disabling will also purge any outstanding packet from the protocol queue.
102.Pp
103.Nm
104supports multi-processor execution of handlers, and relies on a combination
105of source ordering and protocol-specific ordering and work-placement
106policies to decide how to distribute work across one or more worker
107threads.
108Registering protocols will declare one of three policies:
109.Bl -tag -width NETISR_POLICY_SOURCE
110.It Dv NETISR_POLICY_SOURCE
111.Nm
112should maintain source ordering without advice from the protocol.
113.Nm
114will ignore any flow IDs present on
115.Vt mbuf
116headers for the purposes of work placement.
117.It Dv NETISR_POLICY_FLOW
118.Nm
119should maintain flow ordering as defined by the
120.Vt mbuf
121header flow ID field.
122If the protocol implements
123.Va nh_m2flow ,
124then
125.Nm
126will query the protocol in the event that the
127.Vt mbuf
128doesn't have a flow ID, falling back on source ordering.
129.It NETISR_POLICY_CPU
130.Nm
131will entirely delegate all work placement decisions to the protocol,
132querying
133.Va nh_m2cpuid
134for each packet.
135.El
136.Pp
137Registration is declared using
138.Vt "struct netisr_handler" ,
139whose fields are defined as follows:
140.Bl -tag -width "netisr_handler_t nh_handler"
141.It Vt "const char *" Va nh_name
142Unique character string name of the protocol, which may be included in
143.Xr sysctl 3
144MIB names, so should not contain whitespace.
145.It Vt netisr_handler_t Va nh_handler
146Protocol handler function that will be invoked on each packet received for
147the protocol.
148.It Vt netisr_m2flow_t Va nh_m2flow
149Optional protocol function to generate a flow ID and set a valid
150hashtype for packets that enter the
151.Nm
152with
153.Dv M_HASHTYPE_GET(m)
154equal to
155.Dv M_HASHTYPE_NONE .
156Will be used only with
157.Dv NETISR_POLICY_FLOW .
158.It Vt netisr_m2cpuid_t Va nh_m2cpuid
159Protocol function to determine what CPU a packet should be processed on.
160Will be used only with
161.Dv NETISR_POLICY_CPU .
162.It Vt netisr_drainedcpu_t Va nh_drainedcpu
163Optional callback function that will be invoked when a per-CPU queue
164was drained.
165It will never fire for directly dispatched packets.
166Unless fully understood, this special-purpose function should not be used.
167.\" In case you intend to use this please send 50 chocolate bars to each
168.\" of rwatson and bz and wait for an answer.
169.It Vt u_int Va nh_proto
170Protocol number used by both protocols to identify themselves to
171.Nm ,
172and by packet sources to select what handler will be used to process
173packets.
174A table of supported protocol numbers appears below.
175For implementation reasons, protocol numbers great than 15 are currently
176unsupported.
177.It Vt u_int Va nh_qlimit
178The maximum per-CPU queue depth for the protocol; due to internal
179implementation details, the effective queue depth may be as much as twice
180this number.
181.It Vt u_int Va nh_policy
182The ordering and work placement policy for the protocol, as described
183earlier.
184.El
185.Ss Packet source interface
186Packet sources, such as network interfaces, may request protocol processing
187using the
188.Fn netisr_dispatch
189and
190.Fn netisr_queue
191interfaces.
192Both accept a protocol number and
193.Vt mbuf
194argument, but while
195.Fn netisr_queue
196will always execute the protocol handler asynchronously in a deferred
197context,
198.Fn netisr_dispatch
199will optionally direct dispatch if permitted by global and per-protocol
200policy.
201.Pp
202In order to provide additional load balancing and flow information,
203packet sources may also specify an opaque source identifier, which in
204practice might be a network interface number or socket pointer, using
205the
206.Fn netisr_dispatch_src
207and
208.Fn netisr_queue_src
209variants.
210.Ss Protocol number constants
211The follow protocol numbers are currently defined:
212.Bl -tag -width NETISR_ROUTE
213.It Dv NETISR_IP
214IPv4
215.It Dv NETISR_IGMP
216IGMPv3 loopback
217.It Dv NETISR_ROUTE
218Routing socket loopback
219.It Dv NETISR_ARP
220ARP
221.It Dv NETISR_IPV6
222IPv6
223.El
224.Sh AUTHORS
225This manual page and the
226.Nm
227implementation were written by
228.An Robert N. M. Watson .
229