xref: /freebsd/share/man/man9/SDT.9 (revision 9a0c3479e22feda1bdb2db4b97f9deb1b5fa6269)
1.\" Copyright (c) 2013 Mark Johnston <markj@freebsd.org>
2.\" All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\"    notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\"    notice, this list of conditions and the following disclaimer in the
11.\"    documentation and/or other materials provided with the distribution.
12.\"
13.\" THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
14.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
15.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
16.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
17.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
18.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
19.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
20.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
21.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
22.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
23.\" SUCH DAMAGE.
24.\"
25.\" $FreeBSD$
26.\"
27.Dd July 3, 2013
28.Dt SDT 9
29.Os
30.Sh NAME
31.Nm SDT
32.Nd a DTrace framework for adding statically-defined tracing probes
33.Sh SYNOPSIS
34.In sys/sdt.h
35.Fn SDT_PROVIDER_DECLARE prov
36.Fn SDT_PROVIDER_DEFINE prov
37.Fn SDT_PROBE_DECLARE prov mod func name
38.Fn SDT_PROBE_DEFINE prov mod func name sname
39.Fn SDT_PROBE_DEFINE0 prov mod func name sname
40.Fn SDT_PROBE_DEFINE1 prov mod func name sname arg0
41.Fn SDT_PROBE_DEFINE2 prov mod func name sname arg0 arg1
42.Fn SDT_PROBE_DEFINE3 prov mod func name sname arg0 arg1 arg2
43.Fn SDT_PROBE_DEFINE4 prov mod func name sname arg0 arg1 arg2 arg3
44.Fn SDT_PROBE_DEFINE5 prov mod func name sname arg0 arg1 arg2 arg3 arg4
45.Fn SDT_PROBE_DEFINE6 prov mod func name sname arg0 arg1 arg2 arg3 arg4 arg5
46.Fn SDT_PROBE_DEFINE7 prov mod func name sname arg0 arg1 arg2 arg3 arg4 arg5   \
47    arg6
48.Fn SDT_PROBE0 prov mod func name
49.Fn SDT_PROBE1 prov mod func name arg0
50.Fn SDT_PROBE2 prov mod func name arg0 arg1
51.Fn SDT_PROBE3 prov mod func name arg0 arg1 arg2
52.Fn SDT_PROBE4 prov mod func name arg0 arg1 arg2 arg3
53.Fn SDT_PROBE5 prov mod func name arg0 arg1 arg2 arg3 arg4
54.Fn SDT_PROBE6 prov mod func name arg0 arg1 arg2 arg3 arg4 arg5
55.Fn SDT_PROBE7 prov mod func name arg0 arg1 arg2 arg3 arg4 arg5 arg6
56.Sh DESCRIPTION
57.Pp
58The
59.Nm
60macros allow programmers to define static trace points in kernel code.
61These trace points are used by the
62.Nm
63framework to create DTrace probes, allowing the code to be instrumented
64using
65.Xr dtrace 1 .
66By default,
67.Nm
68trace points are disabled and have no effect on the surrounding code.
69When a DTrace probe corresponding to a given trace point is enabled, threads
70that execute the trace point will call a handler and cause the probe to fire.
71Moreover, trace points can take arguments, making it possible to pass data
72to the DTrace framework when an enabled probe fires.
73.Pp
74Multiple trace points may correspond to a single DTrace probe, allowing
75programmers to create DTrace probes that correspond to logical system events
76rather than tying probes to specific code execution paths.
77For instance, a DTrace probe corresponding to the arrival of an IP packet into
78the network stack may be defined using two
79.Nm
80trace points: one for IPv4 packets and one for IPv6 packets.
81.Pp
82In addition to defining DTrace probes, the
83.Nm
84macros allow programmers to define new DTrace providers, making it possible to
85namespace logically-related probes.
86An example is FreeBSD's sctp provider, which contains
87.Nm
88probes for FreeBSD's
89.Xr sctp 4
90implementation.
91.Pp
92The
93.Fn SDT_PROVIDER_DECLARE
94and
95.Fn SDT_PROVIDER_DEFINE
96macros are used respectively to declare and define a DTrace provider named
97.Ar prov
98with the
99.Nm
100framework.
101A provider need only be defined once; however, the provider must be declared
102before defining any
103.Nm
104probes belonging to that provider.
105.Pp
106Similarly, the
107.Fn SDT_PROBE_DECLARE
108and
109.Fn SDT_PROBE_DEFINE*
110macros are used to declare and define DTrace probes using the
111.Nm
112framework.
113Once a probe has been defined, trace points for that probe may be added to
114kernel code.
115DTrace probe identifiers consist of a provider, module, function and name, all
116of which may be specified in the
117.Nm
118probe definition.
119Note that probes should not specify a module name: the module name of a probe is
120used to determine whether or not it should be destroyed when a kernel module is
121unloaded.
122See the
123.Sx BUGS
124section.
125Note in particular that probes must not be defined across multiple kernel
126modules.
127The
128.Fn SDT_PROBE_DEFINE*
129macros also take an extra
130.Ar sname
131parameter.
132This is used to allow the creation of probes with names containing the
133.Ql -
134character.
135Specifically, the
136.Ar name
137argument should contain the probe name with all dashes converted to underscores,
138and the
139.Ar sname
140argument should be the probe name as it will be referenced by D scripts.
141.Pp
142The
143.Fn SDT_PROBE_DEFINE*
144macros also allow programmers to declare the types of the arguments that are
145passed to probes.
146This is optional; if the argument types are omitted (through use of the
147.Fn SDT_PROBE_DEFINE
148macro), users wishing to make use of the arguments will have to manually cast
149them to the correct types in their D scripts.
150It is strongly recommended that probe definitions include a declaration of their
151argument types.
152.Pp
153The
154.Fn SDT_PROBE*
155macros are used to create
156.Nm
157trace points.
158They are meant to be added to executable code and can be used to instrument the
159code in which they are called.
160.Sh EXAMPLES
161.Pp
162The following probe definition will create a DTrace probe called
163.Ql icmp::unreach:pkt-receive ,
164which would hypothetically be triggered when the kernel receives an ICMP packet
165of type Destination Unreachable:
166.Bd -literal -offset indent
167SDT_PROVIDER_DECLARE(icmp);
168
169SDT_PROBE_DEFINE2(icmp, , unreach, pkt_receive, pkt-receive,
170    "struct mbuf *", "struct icmp *");
171
172.Ed
173This particular probe would take two arguments: a pointer to the
174.Xr mbuf 9
175containing the incoming packet, and a pointer to the ICMP header for the packet.
176Note that the module name of this probe is not specified.
177.Pp
178Consider a DTrace probe which fires when the network stack receives an IP
179packet.
180Such a probe would be defined by multiple tracepoints:
181.Bd -literal -offset indent
182SDT_PROBE_DEFINE2(ip, , , receive, receive, "struct mbuf *",
183    "struct ifnet *", "struct ip *", "struct ip6_hdr *");
184
185int
186ip_input(struct mbuf *m)
187{
188	struct ip *ip;
189	...
190	ip = mtod(m, struct ip *);
191	SDT_PROBE4(ip, , , receive, m, m->m_pkthdr.rcvif, ip, NULL);
192	...
193}
194
195int
196ip6_input(struct mbuf *m)
197{
198	struct ip6_hdr *ip6;
199	...
200	ip6 = mtod(m, struct ip6_hdr *);
201	SDT_PROBE4(ip, , , receive, m, m->m_pkthdr.rcvif, NULL, ip6);
202	...
203}
204
205.Ed
206In particular, the probe should fire when the kernel receives either an IPv4
207packet or an IPv6 packet.
208.Sh SEE ALSO
209.Xr dtrace 1
210.Sh AUTHORS
211.An -nosplit
212DTrace and the
213.Nm
214framework were originally ported to FreeBSD from Solaris by
215.An John Birrell Aq jb@FreeBSD.org .
216This manual page was written by
217.An Mark Johnston Aq markj@FreeBSD.org .
218.Sh BUGS
219.Pp
220The
221.Nm
222macros allow the module name of a probe to be specified as part of a probe
223definition.
224However, the DTrace framework uses the module name of probes to determine
225which probes should be destroyed when a kernel module is unloaded, so the module
226name of a probe should match the name of the module in which its defined.
227.Nm
228will set the module name properly if it is left unspecified in the probe
229definition; see the
230.Sx EXAMPLES
231section.
232.Pp
233One of the goals of the original
234.Nm
235implementation (and by extension, of FreeBSD's port) is that inactive
236.Nm
237probes should have no performance impact.
238This is unfortunately not the case;
239.Nm
240trace points will add a small but non-zero amount of latency to the code
241in which they are defined.
242A more sophisticated implementation of the probes will help alleviate this
243problem.
244