1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 LTTng{nbsp}{revision} bears the name _KeKriek_. From
78 http://brasseriedunham.com/[Brasserie Dunham], the _**KeKriek**_ is a
79 sour mashed golden wheat ale fermented with local sour cherries from
80 Tougas orchards. Fresh sweet cherry notes with some tartness, lively
81 carbonation with a dry finish.
83 New features and changes in LTTng{nbsp}{revision}:
85 * **Tracing control**:
86 ** You can put more than one wildcard special character (`*`), and not
87 only at the end, when you <<enabling-disabling-events,create an event
88 rule>>, in both the instrumentation point name and the literal
90 link:/man/1/lttng-enable-event/v{revision}/#doc-filter-syntax[filter expressions]:
95 # lttng enable-event --kernel 'x86_*_local_timer_*' \
96 --filter='name == "*a*b*c*d*e" && count >= 23'
103 $ lttng enable-event --userspace '*_my_org:*msg*'
107 ** New trigger and notification API for
108 <<liblttng-ctl-lttng,`liblttng-ctl`>>. This new subsystem allows you
109 to register triggers which emit a notification when a given
110 condition is satisfied. As of LTTng{nbsp}{revision}, only
111 <<channel,channel>> buffer usage conditions are available.
112 Documentation is available in the
113 https://github.com/lttng/lttng-tools/tree/stable-{revision}/include/lttng[`liblttng-ctl`
115 <<notif-trigger-api,Get notified when a channel's buffer usage is too
118 ** You can now embed the whole textual LTTng-tools man pages into the
119 executables at build time with the `--enable-embedded-help`
120 configuration option. Thanks to this option, you don't need the
121 http://www.methods.co.nz/asciidoc/[AsciiDoc] and
122 https://directory.fsf.org/wiki/Xmlto[xmlto] tools at build time, and
123 a manual pager at run time, to get access to this documentation.
125 * **User space tracing**:
126 ** New blocking mode: an LTTng-UST tracepoint can now block until
127 <<channel,sub-buffer>> space is available instead of discarding event
128 records in <<channel-overwrite-mode-vs-discard-mode,discard mode>>.
129 With this feature, you can be sure that no event records are
130 discarded during your application's execution at the expense of
133 For example, the following command lines create a user space tracing
134 channel with an infinite blocking timeout and run an application
135 instrumented with LTTng-UST which is explicitly allowed to block:
141 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
142 $ lttng enable-event --userspace --channel=blocking-channel --all
144 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
148 See the complete <<blocking-timeout-example,blocking timeout example>>.
150 * **Linux kernel tracing**:
151 ** Linux 4.10, 4.11, and 4.12 support.
152 ** The thread state dump events recorded by LTTng-modules now contain
153 the task's CPU identifier. This improves the precision of the
154 scheduler model for analyses.
155 ** Extended man:socketpair(2) system call tracing data.
161 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
162 generation_ is a modern toolkit for tracing Linux systems and
163 applications. So your first question might be:
170 As the history of software engineering progressed and led to what
171 we now take for granted--complex, numerous and
172 interdependent software applications running in parallel on
173 sophisticated operating systems like Linux--the authors of such
174 components, software developers, began feeling a natural
175 urge to have tools that would ensure the robustness and good performance
176 of their masterpieces.
178 One major achievement in this field is, inarguably, the
179 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
180 an essential tool for developers to find and fix bugs. But even the best
181 debugger won't help make your software run faster, and nowadays, faster
182 software means either more work done by the same hardware, or cheaper
183 hardware for the same work.
185 A _profiler_ is often the tool of choice to identify performance
186 bottlenecks. Profiling is suitable to identify _where_ performance is
187 lost in a given software. The profiler outputs a profile, a statistical
188 summary of observed events, which you may use to discover which
189 functions took the most time to execute. However, a profiler won't
190 report _why_ some identified functions are the bottleneck. Bottlenecks
191 might only occur when specific conditions are met, conditions that are
192 sometimes impossible to capture by a statistical profiler, or impossible
193 to reproduce with an application altered by the overhead of an
194 event-based profiler. For a thorough investigation of software
195 performance issues, a history of execution is essential, with the
196 recorded values of variables and context fields you choose, and
197 with as little influence as possible on the instrumented software. This
198 is where tracing comes in handy.
200 _Tracing_ is a technique used to understand what goes on in a running
201 software system. The software used for tracing is called a _tracer_,
202 which is conceptually similar to a tape recorder. When recording,
203 specific instrumentation points placed in the software source code
204 generate events that are saved on a giant tape: a _trace_ file. You
205 can trace user applications and the operating system at the same time,
206 opening the possibility of resolving a wide range of problems that would
207 otherwise be extremely challenging.
209 Tracing is often compared to _logging_. However, tracers and loggers are
210 two different tools, serving two different purposes. Tracers are
211 designed to record much lower-level events that occur much more
212 frequently than log messages, often in the range of thousands per
213 second, with very little execution overhead. Logging is more appropriate
214 for a very high-level analysis of less frequent events: user accesses,
215 exceptional conditions (errors and warnings, for example), database
216 transactions, instant messaging communications, and such. Simply put,
217 logging is one of the many use cases that can be satisfied with tracing.
219 The list of recorded events inside a trace file can be read manually
220 like a log file for the maximum level of detail, but it is generally
221 much more interesting to perform application-specific analyses to
222 produce reduced statistics and graphs that are useful to resolve a
223 given problem. Trace viewers and analyzers are specialized tools
226 In the end, this is what LTTng is: a powerful, open source set of
227 tools to trace the Linux kernel and user applications at the same time.
228 LTTng is composed of several components actively maintained and
229 developed by its link:/community/#where[community].
232 [[lttng-alternatives]]
233 === Alternatives to noch:{LTTng}
235 Excluding proprietary solutions, a few competing software tracers
238 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
239 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
240 user scripts and is responsible for loading code into the
241 Linux kernel for further execution and collecting the outputted data.
242 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
243 subsystem in the Linux kernel in which a virtual machine can execute
244 programs passed from the user space to the kernel. You can attach
245 such programs to tracepoints and KProbes thanks to a system call, and
246 they can output data to the user space when executed thanks to
247 different mechanisms (pipe, VM register values, and eBPF maps, to name
249 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
250 is the de facto function tracer of the Linux kernel. Its user
251 interface is a set of special files in sysfs.
252 * https://perf.wiki.kernel.org/[perf] is
253 a performance analyzing tool for Linux which supports hardware
254 performance counters, tracepoints, as well as other counters and
255 types of probes. perf's controlling utility is the cmd:perf command
257 * http://linux.die.net/man/1/strace[strace]
258 is a command-line utility which records system calls made by a
259 user process, as well as signal deliveries and changes of process
260 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
261 to fulfill its function.
262 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
263 analyze Linux kernel events. You write scripts, or _chisels_ in
264 sysdig's jargon, in Lua and sysdig executes them while the system is
265 being traced or afterwards. sysdig's interface is the cmd:sysdig
266 command-line tool as well as the curses-based cmd:csysdig tool.
267 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
268 user space tracer which uses custom user scripts to produce plain text
269 traces. SystemTap converts the scripts to the C language, and then
270 compiles them as Linux kernel modules which are loaded to produce
271 trace data. SystemTap's primary user interface is the cmd:stap
274 The main distinctive features of LTTng is that it produces correlated
275 kernel and user space traces, as well as doing so with the lowest
276 overhead amongst other solutions. It produces trace files in the
277 http://diamon.org/ctf[CTF] format, a file format optimized
278 for the production and analyses of multi-gigabyte data.
280 LTTng is the result of more than 10 years of active open source
281 development by a community of passionate developers.
282 LTTng{nbsp}{revision} is currently available on major desktop and server
285 The main interface for tracing control is a single command-line tool
286 named cmd:lttng. The latter can create several tracing sessions, enable
287 and disable events on the fly, filter events efficiently with custom
288 user expressions, start and stop tracing, and much more. LTTng can
289 record the traces on the file system or send them over the network, and
290 keep them totally or partially. You can view the traces once tracing
291 becomes inactive or in real-time.
293 <<installing-lttng,Install LTTng now>> and
294 <<getting-started,start tracing>>!
300 **LTTng** is a set of software <<plumbing,components>> which interact to
301 <<instrumenting,instrument>> the Linux kernel and user applications, and
302 to <<controlling-tracing,control tracing>> (start and stop
303 tracing, enable and disable event rules, and the rest). Those
304 components are bundled into the following packages:
306 * **LTTng-tools**: Libraries and command-line interface to
308 * **LTTng-modules**: Linux kernel modules to instrument and
310 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
311 trace user applications.
313 Most distributions mark the LTTng-modules and LTTng-UST packages as
314 optional when installing LTTng-tools (which is always required). In the
315 following sections, we always provide the steps to install all three,
318 * You only need to install LTTng-modules if you intend to trace the
320 * You only need to install LTTng-UST if you intend to trace user
324 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 25 July 2017.
326 |Distribution |Available in releases |Alternatives
328 |https://www.ubuntu.com/[Ubuntu]
329 |Ubuntu{nbsp}14.04 _Trusty Tahr_ and Ubuntu{nbsp}16.04 _Xenial Xerus_:
330 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
331 |link:/docs/v2.9#doc-ubuntu[LTTng{nbsp}2.9 for Ubuntu{nbsp}17.04 _Zesty Zapus_].
333 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
334 other Ubuntu releases.
336 |https://getfedora.org/[Fedora]
338 |link:/docs/v2.9#doc-fedora[LTTng{nbsp}2.9 for Fedora 26].
340 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
342 |https://www.debian.org/[Debian]
344 |link:/docs/v2.9#doc-debian[LTTng{nbsp}2.9 for Debian "stretch"
345 (stable), Debian "buster" (testing), and Debian "sid" (unstable)].
347 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
349 |https://www.archlinux.org/[Arch Linux]
351 |link:/docs/v2.9#doc-arch-linux[LTTng{nbsp}2.9 in the latest AUR packages].
353 |https://alpinelinux.org/[Alpine Linux]
355 |link:/docs/v2.9#doc-alpine-linux[LTTng{nbsp}2.9 for Alpine Linux "edge"].
357 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
359 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
360 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
363 |https://buildroot.org/[Buildroot]
365 |link:/docs/v2.9#doc-buildroot[LTTng{nbsp}2.9 for Buildroot{nbsp}2017.02 and
366 Buildroot{nbsp}2017.05].
368 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
370 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
371 https://www.yoctoproject.org/[Yocto]
373 |link:/docs/v2.9#doc-oe-yocto[LTTng{nbsp}2.9 for Yocto Project{nbsp}2.3 _Pyro_]
374 (`openembedded-core` layer).
376 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
381 === [[ubuntu-official-repositories]]Ubuntu
384 ==== noch:{LTTng} Stable {revision} PPA
386 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
387 Stable{nbsp}{revision} PPA] offers the latest stable
388 LTTng{nbsp}{revision} packages for:
390 * Ubuntu{nbsp}14.04 _Trusty Tahr_
391 * Ubuntu{nbsp}16.04 _Xenial Xerus_
393 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
395 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
401 # apt-add-repository ppa:lttng/stable-2.10
406 . Install the main LTTng{nbsp}{revision} packages:
411 # apt-get install lttng-tools
412 # apt-get install lttng-modules-dkms
413 # apt-get install liblttng-ust-dev
417 . **If you need to instrument and trace
418 <<java-application,Java applications>>**, install the LTTng-UST
424 # apt-get install liblttng-ust-agent-java
428 . **If you need to instrument and trace
429 <<python-application,Python{nbsp}3 applications>>**, install the
430 LTTng-UST Python agent:
435 # apt-get install python3-lttngust
440 [[enterprise-distributions]]
441 === RHEL, SUSE, and other enterprise distributions
443 To install LTTng on enterprise Linux distributions, such as Red Hat
444 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
445 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
448 [[building-from-source]]
449 === Build from source
451 To build and install LTTng{nbsp}{revision} from source:
453 . Using your distribution's package manager, or from source, install
454 the following dependencies of LTTng-tools and LTTng-UST:
457 * https://sourceforge.net/projects/libuuid/[libuuid]
458 * http://directory.fsf.org/wiki/Popt[popt]
459 * http://liburcu.org/[Userspace RCU]
460 * http://www.xmlsoft.org/[libxml2]
463 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
469 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
470 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
471 cd lttng-modules-2.10.* &&
473 sudo make modules_install &&
478 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
484 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.10.tar.bz2 &&
485 tar -xf lttng-ust-latest-2.10.tar.bz2 &&
486 cd lttng-ust-2.10.* &&
496 .Java and Python application tracing
498 If you need to instrument and trace <<java-application,Java
499 applications>>, pass the `--enable-java-agent-jul`,
500 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
501 `configure` script, depending on which Java logging framework you use.
503 If you need to instrument and trace <<python-application,Python
504 applications>>, pass the `--enable-python-agent` option to the
505 `configure` script. You can set the `PYTHON` environment variable to the
506 path to the Python interpreter for which to install the LTTng-UST Python
514 By default, LTTng-UST libraries are installed to
515 dir:{/usr/local/lib}, which is the de facto directory in which to
516 keep self-compiled and third-party libraries.
518 When <<building-tracepoint-providers-and-user-application,linking an
519 instrumented user application with `liblttng-ust`>>:
521 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
523 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
524 man:gcc(1), man:g++(1), or man:clang(1).
528 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
534 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
535 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
536 cd lttng-tools-2.10.* &&
544 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
545 previous steps automatically for a given version of LTTng and confine
546 the installed files in a specific directory. This can be useful to test
547 LTTng without installing it on your system.
553 This is a short guide to get started quickly with LTTng kernel and user
556 Before you follow this guide, make sure to <<installing-lttng,install>>
559 This tutorial walks you through the steps to:
561 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
562 . <<tracing-your-own-user-application,Trace a user application>> written
564 . <<viewing-and-analyzing-your-traces,View and analyze the
568 [[tracing-the-linux-kernel]]
569 === Trace the Linux kernel
571 The following command lines start with the `#` prompt because you need
572 root privileges to trace the Linux kernel. You can also trace the kernel
573 as a regular user if your Unix user is a member of the
574 <<tracing-group,tracing group>>.
576 . Create a <<tracing-session,tracing session>> which writes its traces
577 to dir:{/tmp/my-kernel-trace}:
582 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
586 . List the available kernel tracepoints and system calls:
591 # lttng list --kernel
592 # lttng list --kernel --syscall
596 . Create <<event,event rules>> which match the desired instrumentation
597 point names, for example the `sched_switch` and `sched_process_fork`
598 tracepoints, and the man:open(2) and man:close(2) system calls:
603 # lttng enable-event --kernel sched_switch,sched_process_fork
604 # lttng enable-event --kernel --syscall open,close
608 You can also create an event rule which matches _all_ the Linux kernel
609 tracepoints (this will generate a lot of data when tracing):
614 # lttng enable-event --kernel --all
618 . <<basic-tracing-session-control,Start tracing>>:
627 . Do some operation on your system for a few seconds. For example,
628 load a website, or list the files of a directory.
629 . <<basic-tracing-session-control,Stop tracing>> and destroy the
640 The man:lttng-destroy(1) command does not destroy the trace data; it
641 only destroys the state of the tracing session.
643 . For the sake of this example, make the recorded trace accessible to
649 # chown -R $(whoami) /tmp/my-kernel-trace
653 See <<viewing-and-analyzing-your-traces,View and analyze the
654 recorded events>> to view the recorded events.
657 [[tracing-your-own-user-application]]
658 === Trace a user application
660 This section steps you through a simple example to trace a
661 _Hello world_ program written in C.
663 To create the traceable user application:
665 . Create the tracepoint provider header file, which defines the
666 tracepoints and the events they can generate:
672 #undef TRACEPOINT_PROVIDER
673 #define TRACEPOINT_PROVIDER hello_world
675 #undef TRACEPOINT_INCLUDE
676 #define TRACEPOINT_INCLUDE "./hello-tp.h"
678 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
681 #include <lttng/tracepoint.h>
691 ctf_string(my_string_field, my_string_arg)
692 ctf_integer(int, my_integer_field, my_integer_arg)
696 #endif /* _HELLO_TP_H */
698 #include <lttng/tracepoint-event.h>
702 . Create the tracepoint provider package source file:
708 #define TRACEPOINT_CREATE_PROBES
709 #define TRACEPOINT_DEFINE
711 #include "hello-tp.h"
715 . Build the tracepoint provider package:
720 $ gcc -c -I. hello-tp.c
724 . Create the _Hello World_ application source file:
731 #include "hello-tp.h"
733 int main(int argc, char *argv[])
737 puts("Hello, World!\nPress Enter to continue...");
740 * The following getchar() call is only placed here for the purpose
741 * of this demonstration, to pause the application in order for
742 * you to have time to list its tracepoints. It is not
748 * A tracepoint() call.
750 * Arguments, as defined in hello-tp.h:
752 * 1. Tracepoint provider name (required)
753 * 2. Tracepoint name (required)
754 * 3. my_integer_arg (first user-defined argument)
755 * 4. my_string_arg (second user-defined argument)
757 * Notice the tracepoint provider and tracepoint names are
758 * NOT strings: they are in fact parts of variables that the
759 * macros in hello-tp.h create.
761 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
763 for (x = 0; x < argc; ++x) {
764 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
767 puts("Quitting now!");
768 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
775 . Build the application:
784 . Link the application with the tracepoint provider package,
785 `liblttng-ust`, and `libdl`:
790 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
794 Here's the whole build process:
797 .User space tracing tutorial's build steps.
798 image::ust-flow.png[]
800 To trace the user application:
802 . Run the application with a few arguments:
807 $ ./hello world and beyond
816 Press Enter to continue...
820 . Start an LTTng <<lttng-sessiond,session daemon>>:
825 $ lttng-sessiond --daemonize
829 Note that a session daemon might already be running, for example as
830 a service that the distribution's service manager started.
832 . List the available user space tracepoints:
837 $ lttng list --userspace
841 You see the `hello_world:my_first_tracepoint` tracepoint listed
842 under the `./hello` process.
844 . Create a <<tracing-session,tracing session>>:
849 $ lttng create my-user-space-session
853 . Create an <<event,event rule>> which matches the
854 `hello_world:my_first_tracepoint` event name:
859 $ lttng enable-event --userspace hello_world:my_first_tracepoint
863 . <<basic-tracing-session-control,Start tracing>>:
872 . Go back to the running `hello` application and press Enter. The
873 program executes all `tracepoint()` instrumentation points and exits.
874 . <<basic-tracing-session-control,Stop tracing>> and destroy the
885 The man:lttng-destroy(1) command does not destroy the trace data; it
886 only destroys the state of the tracing session.
888 By default, LTTng saves the traces in
889 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
890 where +__name__+ is the tracing session name. The
891 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
893 See <<viewing-and-analyzing-your-traces,View and analyze the
894 recorded events>> to view the recorded events.
897 [[viewing-and-analyzing-your-traces]]
898 === View and analyze the recorded events
900 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
901 kernel>> and <<tracing-your-own-user-application,Trace a user
902 application>> tutorials, you can inspect the recorded events.
904 Many tools are available to read LTTng traces:
906 * **cmd:babeltrace** is a command-line utility which converts trace
907 formats; it supports the format that LTTng produces, CTF, as well as a
908 basic text output which can be ++grep++ed. The cmd:babeltrace command
909 is part of the http://diamon.org/babeltrace[Babeltrace] project.
910 * Babeltrace also includes
911 **https://www.python.org/[Python] bindings** so
912 that you can easily open and read an LTTng trace with your own script,
913 benefiting from the power of Python.
914 * http://tracecompass.org/[**Trace Compass**]
915 is a graphical user interface for viewing and analyzing any type of
916 logs or traces, including LTTng's.
917 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
918 project which includes many high-level analyses of LTTng kernel
919 traces, like scheduling statistics, interrupt frequency distribution,
920 top CPU usage, and more.
922 NOTE: This section assumes that the traces recorded during the previous
923 tutorials were saved to their default location, in the
924 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
925 environment variable defaults to `$HOME` if not set.
928 [[viewing-and-analyzing-your-traces-bt]]
929 ==== Use the cmd:babeltrace command-line tool
931 The simplest way to list all the recorded events of a trace is to pass
932 its path to cmd:babeltrace with no options:
936 $ babeltrace ~/lttng-traces/my-user-space-session*
939 cmd:babeltrace finds all traces recursively within the given path and
940 prints all their events, merging them in chronological order.
942 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
947 $ babeltrace /tmp/my-kernel-trace | grep _switch
950 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
951 count the recorded events:
955 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
959 [[viewing-and-analyzing-your-traces-bt-python]]
960 ==== Use the Babeltrace Python bindings
962 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
963 is useful to isolate events by simple matching using man:grep(1) and
964 similar utilities. However, more elaborate filters, such as keeping only
965 event records with a field value falling within a specific range, are
966 not trivial to write using a shell. Moreover, reductions and even the
967 most basic computations involving multiple event records are virtually
968 impossible to implement.
970 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
971 to read the event records of an LTTng trace sequentially and compute the
974 The following script accepts an LTTng Linux kernel trace path as its
975 first argument and prints the short names of the top 5 running processes
976 on CPU 0 during the whole trace:
981 from collections import Counter
987 if len(sys.argv) != 2:
988 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
989 print(msg, file=sys.stderr)
992 # A trace collection contains one or more traces
993 col = babeltrace.TraceCollection()
995 # Add the trace provided by the user (LTTng traces always have
997 if col.add_trace(sys.argv[1], 'ctf') is None:
998 raise RuntimeError('Cannot add trace')
1000 # This counter dict contains execution times:
1002 # task command name -> total execution time (ns)
1003 exec_times = Counter()
1005 # This contains the last `sched_switch` timestamp
1009 for event in col.events:
1010 # Keep only `sched_switch` events
1011 if event.name != 'sched_switch':
1014 # Keep only events which happened on CPU 0
1015 if event['cpu_id'] != 0:
1019 cur_ts = event.timestamp
1025 # Previous task command (short) name
1026 prev_comm = event['prev_comm']
1028 # Initialize entry in our dict if not yet done
1029 if prev_comm not in exec_times:
1030 exec_times[prev_comm] = 0
1032 # Compute previous command execution time
1033 diff = cur_ts - last_ts
1035 # Update execution time of this command
1036 exec_times[prev_comm] += diff
1038 # Update last timestamp
1042 for name, ns in exec_times.most_common(5):
1044 print('{:20}{} s'.format(name, s))
1049 if __name__ == '__main__':
1050 sys.exit(0 if top5proc() else 1)
1057 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1063 swapper/0 48.607245889 s
1064 chromium 7.192738188 s
1065 pavucontrol 0.709894415 s
1066 Compositor 0.660867933 s
1067 Xorg.bin 0.616753786 s
1070 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1071 weren't using the CPU that much when tracing, its first position in the
1076 == [[understanding-lttng]]Core concepts
1078 From a user's perspective, the LTTng system is built on a few concepts,
1079 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1080 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1081 Understanding how those objects relate to eachother is key in mastering
1084 The core concepts are:
1086 * <<tracing-session,Tracing session>>
1087 * <<domain,Tracing domain>>
1088 * <<channel,Channel and ring buffer>>
1089 * <<"event","Instrumentation point, event rule, event, and event record">>
1095 A _tracing session_ is a stateful dialogue between you and
1096 a <<lttng-sessiond,session daemon>>. You can
1097 <<creating-destroying-tracing-sessions,create a new tracing
1098 session>> with the `lttng create` command.
1100 Anything that you do when you control LTTng tracers happens within a
1101 tracing session. In particular, a tracing session:
1104 * Has its own set of trace files.
1105 * Has its own state of activity (started or stopped).
1106 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1108 * Has its own <<channel,channels>> which have their own
1109 <<event,event rules>>.
1112 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1113 image::concepts.png[]
1115 Those attributes and objects are completely isolated between different
1118 A tracing session is analogous to a cash machine session:
1119 the operations you do on the banking system through the cash machine do
1120 not alter the data of other users of the same system. In the case of
1121 the cash machine, a session lasts as long as your bank card is inside.
1122 In the case of LTTng, a tracing session lasts from the `lttng create`
1123 command to the `lttng destroy` command.
1126 .Each Unix user has its own set of tracing sessions.
1127 image::many-sessions.png[]
1130 [[tracing-session-mode]]
1131 ==== Tracing session mode
1133 LTTng can send the generated trace data to different locations. The
1134 _tracing session mode_ dictates where to send it. The following modes
1135 are available in LTTng{nbsp}{revision}:
1138 LTTng writes the traces to the file system of the machine being traced
1141 Network streaming mode::
1142 LTTng sends the traces over the network to a
1143 <<lttng-relayd,relay daemon>> running on a remote system.
1146 LTTng does not write the traces by default. Instead, you can request
1147 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1148 current tracing buffers, and to write it to the target's file system
1149 or to send it over the network to a <<lttng-relayd,relay daemon>>
1150 running on a remote system.
1153 This mode is similar to the network streaming mode, but a live
1154 trace viewer can connect to the distant relay daemon to
1155 <<lttng-live,view event records as LTTng generates them>> by
1162 A _tracing domain_ is a namespace for event sources. A tracing domain
1163 has its own properties and features.
1165 There are currently five available tracing domains:
1169 * `java.util.logging` (JUL)
1173 You must specify a tracing domain when using some commands to avoid
1174 ambiguity. For example, since all the domains support named tracepoints
1175 as event sources (instrumentation points that you manually insert in the
1176 source code), you need to specify a tracing domain when
1177 <<enabling-disabling-events,creating an event rule>> because all the
1178 tracing domains could have tracepoints with the same names.
1180 Some features are reserved to specific tracing domains. Dynamic function
1181 entry and return instrumentation points, for example, are currently only
1182 supported in the Linux kernel tracing domain, but support for other
1183 tracing domains could be added in the future.
1185 You can create <<channel,channels>> in the Linux kernel and user space
1186 tracing domains. The other tracing domains have a single default
1191 === Channel and ring buffer
1193 A _channel_ is an object which is responsible for a set of ring buffers.
1194 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1195 tracer emits an event, it can record it to one or more
1196 sub-buffers. The attributes of a channel determine what to do when
1197 there's no space left for a new event record because all sub-buffers
1198 are full, where to send a full sub-buffer, and other behaviours.
1200 A channel is always associated to a <<domain,tracing domain>>. The
1201 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1202 a default channel which you cannot configure.
1204 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1205 an event, it records it to the sub-buffers of all
1206 the enabled channels with a satisfied event rule, as long as those
1207 channels are part of active <<tracing-session,tracing sessions>>.
1210 [[channel-buffering-schemes]]
1211 ==== Per-user vs. per-process buffering schemes
1213 A channel has at least one ring buffer _per CPU_. LTTng always
1214 records an event to the ring buffer associated to the CPU on which it
1217 Two _buffering schemes_ are available when you
1218 <<enabling-disabling-channels,create a channel>> in the
1219 user space <<domain,tracing domain>>:
1221 Per-user buffering::
1222 Allocate one set of ring buffers--one per CPU--shared by all the
1223 instrumented processes of each Unix user.
1227 .Per-user buffering scheme.
1228 image::per-user-buffering.png[]
1231 Per-process buffering::
1232 Allocate one set of ring buffers--one per CPU--for each
1233 instrumented process.
1237 .Per-process buffering scheme.
1238 image::per-process-buffering.png[]
1241 The per-process buffering scheme tends to consume more memory than the
1242 per-user option because systems generally have more instrumented
1243 processes than Unix users running instrumented processes. However, the
1244 per-process buffering scheme ensures that one process having a high
1245 event throughput won't fill all the shared sub-buffers of the same
1248 The Linux kernel tracing domain has only one available buffering scheme
1249 which is to allocate a single set of ring buffers for the whole system.
1250 This scheme is similar to the per-user option, but with a single, global
1251 user "running" the kernel.
1254 [[channel-overwrite-mode-vs-discard-mode]]
1255 ==== Overwrite vs. discard event loss modes
1257 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1258 arc in the following animation) of a specific channel's ring buffer.
1259 When there's no space left in a sub-buffer, the tracer marks it as
1260 consumable (red) and another, empty sub-buffer starts receiving the
1261 following event records. A <<lttng-consumerd,consumer daemon>>
1262 eventually consumes the marked sub-buffer (returns to white).
1265 [role="docsvg-channel-subbuf-anim"]
1270 In an ideal world, sub-buffers are consumed faster than they are filled,
1271 as is the case in the previous animation. In the real world,
1272 however, all sub-buffers can be full at some point, leaving no space to
1273 record the following events.
1275 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1276 no empty sub-buffer is available, it is acceptable to lose event records
1277 when the alternative would be to cause substantial delays in the
1278 instrumented application's execution. LTTng privileges performance over
1279 integrity; it aims at perturbing the traced system as little as possible
1280 in order to make tracing of subtle race conditions and rare interrupt
1283 Starting from LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST,
1284 supports a _blocking mode_. See the <<blocking-timeout-example,blocking
1285 timeout example>> to learn how to use the blocking mode.
1287 When it comes to losing event records because no empty sub-buffer is
1288 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1289 reached, the channel's _event loss mode_ determines what to do. The
1290 available event loss modes are:
1293 Drop the newest event records until a the tracer releases a
1296 This is the only available mode when you specify a
1297 <<opt-blocking-timeout,blocking timeout>>.
1300 Clear the sub-buffer containing the oldest event records and start
1301 writing the newest event records there.
1303 This mode is sometimes called _flight recorder mode_ because it's
1305 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1306 always keep a fixed amount of the latest data.
1308 Which mechanism you should choose depends on your context: prioritize
1309 the newest or the oldest event records in the ring buffer?
1311 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1312 as soon as a there's no space left for a new event record, whereas in
1313 discard mode, the tracer only discards the event record that doesn't
1316 In discard mode, LTTng increments a count of lost event records when an
1317 event record is lost and saves this count to the trace. In overwrite
1318 mode, since LTTng 2.8, LTTng increments a count of lost sub-buffers when
1319 a sub-buffer is lost and saves this count to the trace. In this mode,
1320 the exact number of lost event records in those lost sub-buffers is not
1321 saved to the trace. Trace analyses can use the trace's saved discarded
1322 event record and sub-buffer counts to decide whether or not to perform
1323 the analyses even if trace data is known to be missing.
1325 There are a few ways to decrease your probability of losing event
1327 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1328 how you can fine-une the sub-buffer count and size of a channel to
1329 virtually stop losing event records, though at the cost of greater
1333 [[channel-subbuf-size-vs-subbuf-count]]
1334 ==== Sub-buffer count and size
1336 When you <<enabling-disabling-channels,create a channel>>, you can
1337 set its number of sub-buffers and their size.
1339 Note that there is noticeable CPU overhead introduced when
1340 switching sub-buffers (marking a full one as consumable and switching
1341 to an empty one for the following events to be recorded). Knowing this,
1342 the following list presents a few practical situations along with how
1343 to configure the sub-buffer count and size for them:
1345 * **High event throughput**: In general, prefer bigger sub-buffers to
1346 lower the risk of losing event records.
1348 Having bigger sub-buffers also ensures a lower
1349 <<channel-switch-timer,sub-buffer switching frequency>>.
1351 The number of sub-buffers is only meaningful if you create the channel
1352 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1353 other sub-buffers are left unaltered.
1355 * **Low event throughput**: In general, prefer smaller sub-buffers
1356 since the risk of losing event records is low.
1358 Because events occur less frequently, the sub-buffer switching frequency
1359 should remain low and thus the tracer's overhead should not be a
1362 * **Low memory system**: If your target system has a low memory
1363 limit, prefer fewer first, then smaller sub-buffers.
1365 Even if the system is limited in memory, you want to keep the
1366 sub-buffers as big as possible to avoid a high sub-buffer switching
1369 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1370 which means event data is very compact. For example, the average
1371 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1372 sub-buffer size of 1{nbsp}MiB is considered big.
1374 The previous situations highlight the major trade-off between a few big
1375 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1376 frequency vs. how much data is lost in overwrite mode. Assuming a
1377 constant event throughput and using the overwrite mode, the two
1378 following configurations have the same ring buffer total size:
1381 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1386 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1387 switching frequency, but if a sub-buffer overwrite happens, half of
1388 the event records so far (4{nbsp}MiB) are definitely lost.
1389 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1390 overhead as the previous configuration, but if a sub-buffer
1391 overwrite happens, only the eighth of event records so far are
1394 In discard mode, the sub-buffers count parameter is pointless: use two
1395 sub-buffers and set their size according to the requirements of your
1399 [[channel-switch-timer]]
1400 ==== Switch timer period
1402 The _switch timer period_ is an important configurable attribute of
1403 a channel to ensure periodic sub-buffer flushing.
1405 When the _switch timer_ expires, a sub-buffer switch happens. You can
1406 set the switch timer period attribute when you
1407 <<enabling-disabling-channels,create a channel>> to ensure that event
1408 data is consumed and committed to trace files or to a distant relay
1409 daemon periodically in case of a low event throughput.
1412 [role="docsvg-channel-switch-timer"]
1417 This attribute is also convenient when you use big sub-buffers to cope
1418 with a sporadic high event throughput, even if the throughput is
1422 [[channel-read-timer]]
1423 ==== Read timer period
1425 By default, the LTTng tracers use a notification mechanism to signal a
1426 full sub-buffer so that a consumer daemon can consume it. When such
1427 notifications must be avoided, for example in real-time applications,
1428 you can use the channel's _read timer_ instead. When the read timer
1429 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1430 consumable sub-buffers.
1433 [[tracefile-rotation]]
1434 ==== Trace file count and size
1436 By default, trace files can grow as large as needed. You can set the
1437 maximum size of each trace file that a channel writes when you
1438 <<enabling-disabling-channels,create a channel>>. When the size of
1439 a trace file reaches the channel's fixed maximum size, LTTng creates
1440 another file to contain the next event records. LTTng appends a file
1441 count to each trace file name in this case.
1443 If you set the trace file size attribute when you create a channel, the
1444 maximum number of trace files that LTTng creates is _unlimited_ by
1445 default. To limit them, you can also set a maximum number of trace
1446 files. When the number of trace files reaches the channel's fixed
1447 maximum count, the oldest trace file is overwritten. This mechanism is
1448 called _trace file rotation_.
1452 === Instrumentation point, event rule, event, and event record
1454 An _event rule_ is a set of conditions which must be **all** satisfied
1455 for LTTng to record an occuring event.
1457 You set the conditions when you <<enabling-disabling-events,create
1460 You always attach an event rule to <<channel,channel>> when you create
1463 When an event passes the conditions of an event rule, LTTng records it
1464 in one of the attached channel's sub-buffers.
1466 The available conditions, as of LTTng{nbsp}{revision}, are:
1468 * The event rule _is enabled_.
1469 * The instrumentation point's type _is{nbsp}T_.
1470 * The instrumentation point's name (sometimes called _event name_)
1471 _matches{nbsp}N_, but _is not{nbsp}E_.
1472 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1473 _is exactly{nbsp}L_.
1474 * The fields of the event's payload _satisfy_ a filter
1475 expression{nbsp}__F__.
1477 As you can see, all the conditions but the dynamic filter are related to
1478 the event rule's status or to the instrumentation point, not to the
1479 occurring events. This is why, without a filter, checking if an event
1480 passes an event rule is not a dynamic task: when you create or modify an
1481 event rule, all the tracers of its tracing domain enable or disable the
1482 instrumentation points themselves once. This is possible because the
1483 attributes of an instrumentation point (type, name, and log level) are
1484 defined statically. In other words, without a dynamic filter, the tracer
1485 _does not evaluate_ the arguments of an instrumentation point unless it
1486 matches an enabled event rule.
1488 Note that, for LTTng to record an event, the <<channel,channel>> to
1489 which a matching event rule is attached must also be enabled, and the
1490 tracing session owning this channel must be active.
1493 .Logical path from an instrumentation point to an event record.
1494 image::event-rule.png[]
1496 .Event, event record, or event rule?
1498 With so many similar terms, it's easy to get confused.
1500 An **event** is the consequence of the execution of an _instrumentation
1501 point_, like a tracepoint that you manually place in some source code,
1502 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1503 time. Different actions can be taken upon the occurrence of an event,
1504 like record the event's payload to a buffer.
1506 An **event record** is the representation of an event in a sub-buffer. A
1507 tracer is responsible for capturing the payload of an event, current
1508 context variables, the event's ID, and the event's timestamp. LTTng
1509 can append this sub-buffer to a trace file.
1511 An **event rule** is a set of conditions which must all be satisfied for
1512 LTTng to record an occuring event. Events still occur without
1513 satisfying event rules, but LTTng does not record them.
1518 == Components of noch:{LTTng}
1520 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1521 to call LTTng a simple _tool_ since it is composed of multiple
1522 interacting components. This section describes those components,
1523 explains their respective roles, and shows how they connect together to
1524 form the LTTng ecosystem.
1526 The following diagram shows how the most important components of LTTng
1527 interact with user applications, the Linux kernel, and you:
1530 .Control and trace data paths between LTTng components.
1531 image::plumbing.png[]
1533 The LTTng project incorporates:
1535 * **LTTng-tools**: Libraries and command-line interface to
1536 control tracing sessions.
1537 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1538 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1539 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1540 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1541 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1542 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1544 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1545 headers to instrument and trace any native user application.
1546 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1547 *** `liblttng-ust-libc-wrapper`
1548 *** `liblttng-ust-pthread-wrapper`
1549 *** `liblttng-ust-cyg-profile`
1550 *** `liblttng-ust-cyg-profile-fast`
1551 *** `liblttng-ust-dl`
1552 ** User space tracepoint provider source files generator command-line
1553 tool (man:lttng-gen-tp(1)).
1554 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1555 Java applications using `java.util.logging` or
1556 Apache log4j 1.2 logging.
1557 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1558 Python applications using the standard `logging` package.
1559 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1561 ** LTTng kernel tracer module.
1562 ** Tracing ring buffer kernel modules.
1563 ** Probe kernel modules.
1564 ** LTTng logger kernel module.
1568 === Tracing control command-line interface
1571 .The tracing control command-line interface.
1572 image::plumbing-lttng-cli.png[]
1574 The _man:lttng(1) command-line tool_ is the standard user interface to
1575 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1576 is part of LTTng-tools.
1578 The cmd:lttng tool is linked with
1579 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1580 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1582 The cmd:lttng tool has a Git-like interface:
1586 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1589 The <<controlling-tracing,Tracing control>> section explores the
1590 available features of LTTng using the cmd:lttng tool.
1593 [[liblttng-ctl-lttng]]
1594 === Tracing control library
1597 .The tracing control library.
1598 image::plumbing-liblttng-ctl.png[]
1600 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1601 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1602 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1604 The <<lttng-cli,cmd:lttng command-line tool>>
1605 is linked with `liblttng-ctl`.
1607 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1612 #include <lttng/lttng.h>
1615 Some objects are referenced by name (C string), such as tracing
1616 sessions, but most of them require to create a handle first using
1617 `lttng_create_handle()`.
1619 The best available developer documentation for `liblttng-ctl` is, as of
1620 LTTng{nbsp}{revision}, its installed header files. Every function and
1621 structure is thoroughly documented.
1625 === User space tracing library
1628 .The user space tracing library.
1629 image::plumbing-liblttng-ust.png[]
1631 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1632 is the LTTng user space tracer. It receives commands from a
1633 <<lttng-sessiond,session daemon>>, for example to
1634 enable and disable specific instrumentation points, and writes event
1635 records to ring buffers shared with a
1636 <<lttng-consumerd,consumer daemon>>.
1637 `liblttng-ust` is part of LTTng-UST.
1639 Public C header files are installed beside `liblttng-ust` to
1640 instrument any <<c-application,C or $$C++$$ application>>.
1642 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1643 packages, use their own library providing tracepoints which is
1644 linked with `liblttng-ust`.
1646 An application or library does not have to initialize `liblttng-ust`
1647 manually: its constructor does the necessary tasks to properly register
1648 to a session daemon. The initialization phase also enables the
1649 instrumentation points matching the <<event,event rules>> that you
1653 [[lttng-ust-agents]]
1654 === User space tracing agents
1657 .The user space tracing agents.
1658 image::plumbing-lttng-ust-agents.png[]
1660 The _LTTng-UST Java and Python agents_ are regular Java and Python
1661 packages which add LTTng tracing capabilities to the
1662 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1664 In the case of Java, the
1665 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1666 core logging facilities] and
1667 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1668 Note that Apache Log4{nbsp}2 is not supported.
1670 In the case of Python, the standard
1671 https://docs.python.org/3/library/logging.html[`logging`] package
1672 is supported. Both Python 2 and Python 3 modules can import the
1673 LTTng-UST Python agent package.
1675 The applications using the LTTng-UST agents are in the
1676 `java.util.logging` (JUL),
1677 log4j, and Python <<domain,tracing domains>>.
1679 Both agents use the same mechanism to trace the log statements. When an
1680 agent is initialized, it creates a log handler that attaches to the root
1681 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1682 When the application executes a log statement, it is passed to the
1683 agent's log handler by the root logger. The agent's log handler calls a
1684 native function in a tracepoint provider package shared library linked
1685 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1686 other fields, like its logger name and its log level. This native
1687 function contains a user space instrumentation point, hence tracing the
1690 The log level condition of an
1691 <<event,event rule>> is considered when tracing
1692 a Java or a Python application, and it's compatible with the standard
1693 JUL, log4j, and Python log levels.
1697 === LTTng kernel modules
1700 .The LTTng kernel modules.
1701 image::plumbing-lttng-modules.png[]
1703 The _LTTng kernel modules_ are a set of Linux kernel modules
1704 which implement the kernel tracer of the LTTng project. The LTTng
1705 kernel modules are part of LTTng-modules.
1707 The LTTng kernel modules include:
1709 * A set of _probe_ modules.
1711 Each module attaches to a specific subsystem
1712 of the Linux kernel using its tracepoint instrument points. There are
1713 also modules to attach to the entry and return points of the Linux
1714 system call functions.
1716 * _Ring buffer_ modules.
1718 A ring buffer implementation is provided as kernel modules. The LTTng
1719 kernel tracer writes to the ring buffer; a
1720 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1722 * The _LTTng kernel tracer_ module.
1723 * The _LTTng logger_ module.
1725 The LTTng logger module implements the special path:{/proc/lttng-logger}
1726 file so that any executable can generate LTTng events by opening and
1727 writing to this file.
1729 See <<proc-lttng-logger-abi,LTTng logger>>.
1731 Generally, you do not have to load the LTTng kernel modules manually
1732 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1733 daemon>> loads the necessary modules when starting. If you have extra
1734 probe modules, you can specify to load them to the session daemon on
1737 The LTTng kernel modules are installed in
1738 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1739 the kernel release (see `uname --kernel-release`).
1746 .The session daemon.
1747 image::plumbing-sessiond.png[]
1749 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1750 managing tracing sessions and for controlling the various components of
1751 LTTng. The session daemon is part of LTTng-tools.
1753 The session daemon sends control requests to and receives control
1756 * The <<lttng-ust,user space tracing library>>.
1758 Any instance of the user space tracing library first registers to
1759 a session daemon. Then, the session daemon can send requests to
1760 this instance, such as:
1763 ** Get the list of tracepoints.
1764 ** Share an <<event,event rule>> so that the user space tracing library
1765 can enable or disable tracepoints. Amongst the possible conditions
1766 of an event rule is a filter expression which `liblttng-ust` evalutes
1767 when an event occurs.
1768 ** Share <<channel,channel>> attributes and ring buffer locations.
1771 The session daemon and the user space tracing library use a Unix
1772 domain socket for their communication.
1774 * The <<lttng-ust-agents,user space tracing agents>>.
1776 Any instance of a user space tracing agent first registers to
1777 a session daemon. Then, the session daemon can send requests to
1778 this instance, such as:
1781 ** Get the list of loggers.
1782 ** Enable or disable a specific logger.
1785 The session daemon and the user space tracing agent use a TCP connection
1786 for their communication.
1788 * The <<lttng-modules,LTTng kernel tracer>>.
1789 * The <<lttng-consumerd,consumer daemon>>.
1791 The session daemon sends requests to the consumer daemon to instruct
1792 it where to send the trace data streams, amongst other information.
1794 * The <<lttng-relayd,relay daemon>>.
1796 The session daemon receives commands from the
1797 <<liblttng-ctl-lttng,tracing control library>>.
1799 The root session daemon loads the appropriate
1800 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1801 a <<lttng-consumerd,consumer daemon>> as soon as you create
1802 an <<event,event rule>>.
1804 The session daemon does not send and receive trace data: this is the
1805 role of the <<lttng-consumerd,consumer daemon>> and
1806 <<lttng-relayd,relay daemon>>. It does, however, generate the
1807 http://diamon.org/ctf/[CTF] metadata stream.
1809 Each Unix user can have its own session daemon instance. The
1810 tracing sessions managed by different session daemons are completely
1813 The root user's session daemon is the only one which is
1814 allowed to control the LTTng kernel tracer, and its spawned consumer
1815 daemon is the only one which is allowed to consume trace data from the
1816 LTTng kernel tracer. Note, however, that any Unix user which is a member
1817 of the <<tracing-group,tracing group>> is allowed
1818 to create <<channel,channels>> in the
1819 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1822 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1823 session daemon when using its `create` command if none is currently
1824 running. You can also start the session daemon manually.
1831 .The consumer daemon.
1832 image::plumbing-consumerd.png[]
1834 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
1835 ring buffers with user applications or with the LTTng kernel modules to
1836 collect trace data and send it to some location (on disk or to a
1837 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1838 is part of LTTng-tools.
1840 You do not start a consumer daemon manually: a consumer daemon is always
1841 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1842 <<event,event rule>>, that is, before you start tracing. When you kill
1843 its owner session daemon, the consumer daemon also exits because it is
1844 the session daemon's child process. Command-line options of
1845 man:lttng-sessiond(8) target the consumer daemon process.
1847 There are up to two running consumer daemons per Unix user, whereas only
1848 one session daemon can run per user. This is because each process can be
1849 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1850 and 64-bit processes, it is more efficient to have separate
1851 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1852 exception: it can have up to _three_ running consumer daemons: 32-bit
1853 and 64-bit instances for its user applications, and one more
1854 reserved for collecting kernel trace data.
1862 image::plumbing-relayd.png[]
1864 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1865 between remote session and consumer daemons, local trace files, and a
1866 remote live trace viewer. The relay daemon is part of LTTng-tools.
1868 The main purpose of the relay daemon is to implement a receiver of
1869 <<sending-trace-data-over-the-network,trace data over the network>>.
1870 This is useful when the target system does not have much file system
1871 space to record trace files locally.
1873 The relay daemon is also a server to which a
1874 <<lttng-live,live trace viewer>> can
1875 connect. The live trace viewer sends requests to the relay daemon to
1876 receive trace data as the target system emits events. The
1877 communication protocol is named _LTTng live_; it is used over TCP
1880 Note that you can start the relay daemon on the target system directly.
1881 This is the setup of choice when the use case is to view events as
1882 the target system emits them without the need of a remote system.
1886 == [[using-lttng]]Instrumentation
1888 There are many examples of tracing and monitoring in our everyday life:
1890 * You have access to real-time and historical weather reports and
1891 forecasts thanks to weather stations installed around the country.
1892 * You know your heart is safe thanks to an electrocardiogram.
1893 * You make sure not to drive your car too fast and to have enough fuel
1894 to reach your destination thanks to gauges visible on your dashboard.
1896 All the previous examples have something in common: they rely on
1897 **instruments**. Without the electrodes attached to the surface of your
1898 body's skin, cardiac monitoring is futile.
1900 LTTng, as a tracer, is no different from those real life examples. If
1901 you're about to trace a software system or, in other words, record its
1902 history of execution, you better have **instrumentation points** in the
1903 subject you're tracing, that is, the actual software.
1905 Various ways were developed to instrument a piece of software for LTTng
1906 tracing. The most straightforward one is to manually place
1907 instrumentation points, called _tracepoints_, in the software's source
1908 code. It is also possible to add instrumentation points dynamically in
1909 the Linux kernel <<domain,tracing domain>>.
1911 If you're only interested in tracing the Linux kernel, your
1912 instrumentation needs are probably already covered by LTTng's built-in
1913 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
1914 user application which is already instrumented for LTTng tracing.
1915 In such cases, you can skip this whole section and read the topics of
1916 the <<controlling-tracing,Tracing control>> section.
1918 Many methods are available to instrument a piece of software for LTTng
1921 * <<c-application,User space instrumentation for C and $$C++$$
1923 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
1924 * <<java-application,User space Java agent>>.
1925 * <<python-application,User space Python agent>>.
1926 * <<proc-lttng-logger-abi,LTTng logger>>.
1927 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
1931 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
1933 The procedure to instrument a C or $$C++$$ user application with
1934 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
1936 . <<tracepoint-provider,Create the source files of a tracepoint provider
1938 . <<probing-the-application-source-code,Add tracepoints to
1939 the application's source code>>.
1940 . <<building-tracepoint-providers-and-user-application,Build and link
1941 a tracepoint provider package and the user application>>.
1943 If you need quick, man:printf(3)-like instrumentation, you can skip
1944 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
1947 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
1948 instrument a user application with `liblttng-ust`.
1951 [[tracepoint-provider]]
1952 ==== Create the source files of a tracepoint provider package
1954 A _tracepoint provider_ is a set of compiled functions which provide
1955 **tracepoints** to an application, the type of instrumentation point
1956 supported by LTTng-UST. Those functions can emit events with
1957 user-defined fields and serialize those events as event records to one
1958 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
1959 macro, which you <<probing-the-application-source-code,insert in a user
1960 application's source code>>, calls those functions.
1962 A _tracepoint provider package_ is an object file (`.o`) or a shared
1963 library (`.so`) which contains one or more tracepoint providers.
1964 Its source files are:
1966 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
1967 * A <<tpp-source,tracepoint provider package source>> (`.c`).
1969 A tracepoint provider package is dynamically linked with `liblttng-ust`,
1970 the LTTng user space tracer, at run time.
1973 .User application linked with `liblttng-ust` and containing a tracepoint provider.
1974 image::ust-app.png[]
1976 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
1977 skip creating and using a tracepoint provider and use
1978 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
1982 ===== Create a tracepoint provider header file template
1984 A _tracepoint provider header file_ contains the tracepoint
1985 definitions of a tracepoint provider.
1987 To create a tracepoint provider header file:
1989 . Start from this template:
1993 .Tracepoint provider header file template (`.h` file extension).
1995 #undef TRACEPOINT_PROVIDER
1996 #define TRACEPOINT_PROVIDER provider_name
1998 #undef TRACEPOINT_INCLUDE
1999 #define TRACEPOINT_INCLUDE "./tp.h"
2001 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2004 #include <lttng/tracepoint.h>
2007 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2008 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2013 #include <lttng/tracepoint-event.h>
2019 * `provider_name` with the name of your tracepoint provider.
2020 * `"tp.h"` with the name of your tracepoint provider header file.
2022 . Below the `#include <lttng/tracepoint.h>` line, put your
2023 <<defining-tracepoints,tracepoint definitions>>.
2025 Your tracepoint provider name must be unique amongst all the possible
2026 tracepoint provider names used on the same target system. We
2027 suggest to include the name of your project or company in the name,
2028 for example, `org_lttng_my_project_tpp`.
2030 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2031 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2032 write are the <<defining-tracepoints,tracepoint definitions>>.
2035 [[defining-tracepoints]]
2036 ===== Create a tracepoint definition
2038 A _tracepoint definition_ defines, for a given tracepoint:
2040 * Its **input arguments**. They are the macro parameters that the
2041 `tracepoint()` macro accepts for this particular tracepoint
2042 in the user application's source code.
2043 * Its **output event fields**. They are the sources of event fields
2044 that form the payload of any event that the execution of the
2045 `tracepoint()` macro emits for this particular tracepoint.
2047 You can create a tracepoint definition by using the
2048 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2050 <<tpp-header,tracepoint provider header file template>>.
2052 The syntax of the `TRACEPOINT_EVENT()` macro is:
2055 .`TRACEPOINT_EVENT()` macro syntax.
2058 /* Tracepoint provider name */
2061 /* Tracepoint name */
2064 /* Input arguments */
2069 /* Output event fields */
2078 * `provider_name` with your tracepoint provider name.
2079 * `tracepoint_name` with your tracepoint name.
2080 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2081 * `fields` with the <<tpp-def-output-fields,output event field>>
2084 This tracepoint emits events named `provider_name:tracepoint_name`.
2087 .Event name's length limitation
2089 The concatenation of the tracepoint provider name and the
2090 tracepoint name must not exceed **254 characters**. If it does, the
2091 instrumented application compiles and runs, but LTTng throws multiple
2092 warnings and you could experience serious issues.
2095 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2098 .`TP_ARGS()` macro syntax.
2107 * `type` with the C type of the argument.
2108 * `arg_name` with the argument name.
2110 You can repeat `type` and `arg_name` up to 10 times to have
2111 more than one argument.
2113 .`TP_ARGS()` usage with three arguments.
2125 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2126 tracepoint definition with no input arguments.
2128 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2129 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2130 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2131 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2134 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2135 C expression that the tracer evalutes at the `tracepoint()` macro site
2136 in the application's source code. This expression provides a field's
2137 source of data. The argument expression can include input argument names
2138 listed in the `TP_ARGS()` macro.
2140 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2141 must be unique within a given tracepoint definition.
2143 Here's a complete tracepoint definition example:
2145 .Tracepoint definition.
2147 The following tracepoint definition defines a tracepoint which takes
2148 three input arguments and has four output event fields.
2152 #include "my-custom-structure.h"
2158 const struct my_custom_structure*, my_custom_structure,
2163 ctf_string(query_field, query)
2164 ctf_float(double, ratio_field, ratio)
2165 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2166 ctf_integer(int, send_size, my_custom_structure->send_size)
2171 You can refer to this tracepoint definition with the `tracepoint()`
2172 macro in your application's source code like this:
2176 tracepoint(my_provider, my_tracepoint,
2177 my_structure, some_ratio, the_query);
2181 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2182 if they satisfy an enabled <<event,event rule>>.
2185 [[using-tracepoint-classes]]
2186 ===== Use a tracepoint class
2188 A _tracepoint class_ is a class of tracepoints which share the same
2189 output event field definitions. A _tracepoint instance_ is one
2190 instance of such a defined tracepoint class, with its own tracepoint
2193 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2194 shorthand which defines both a tracepoint class and a tracepoint
2195 instance at the same time.
2197 When you build a tracepoint provider package, the C or $$C++$$ compiler
2198 creates one serialization function for each **tracepoint class**. A
2199 serialization function is responsible for serializing the event fields
2200 of a tracepoint to a sub-buffer when tracing.
2202 For various performance reasons, when your situation requires multiple
2203 tracepoint definitions with different names, but with the same event
2204 fields, we recommend that you manually create a tracepoint class
2205 and instantiate as many tracepoint instances as needed. One positive
2206 effect of such a design, amongst other advantages, is that all
2207 tracepoint instances of the same tracepoint class reuse the same
2208 serialization function, thus reducing
2209 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2211 .Use a tracepoint class and tracepoint instances.
2213 Consider the following three tracepoint definitions:
2225 ctf_integer(int, userid, userid)
2226 ctf_integer(size_t, len, len)
2238 ctf_integer(int, userid, userid)
2239 ctf_integer(size_t, len, len)
2251 ctf_integer(int, userid, userid)
2252 ctf_integer(size_t, len, len)
2257 In this case, we create three tracepoint classes, with one implicit
2258 tracepoint instance for each of them: `get_account`, `get_settings`, and
2259 `get_transaction`. However, they all share the same event field names
2260 and types. Hence three identical, yet independent serialization
2261 functions are created when you build the tracepoint provider package.
2263 A better design choice is to define a single tracepoint class and three
2264 tracepoint instances:
2268 /* The tracepoint class */
2269 TRACEPOINT_EVENT_CLASS(
2270 /* Tracepoint provider name */
2273 /* Tracepoint class name */
2276 /* Input arguments */
2282 /* Output event fields */
2284 ctf_integer(int, userid, userid)
2285 ctf_integer(size_t, len, len)
2289 /* The tracepoint instances */
2290 TRACEPOINT_EVENT_INSTANCE(
2291 /* Tracepoint provider name */
2294 /* Tracepoint class name */
2297 /* Tracepoint name */
2300 /* Input arguments */
2306 TRACEPOINT_EVENT_INSTANCE(
2315 TRACEPOINT_EVENT_INSTANCE(
2328 [[assigning-log-levels]]
2329 ===== Assign a log level to a tracepoint definition
2331 You can assign an optional _log level_ to a
2332 <<defining-tracepoints,tracepoint definition>>.
2334 Assigning different levels of severity to tracepoint definitions can
2335 be useful: when you <<enabling-disabling-events,create an event rule>>,
2336 you can target tracepoints having a log level as severe as a specific
2339 The concept of LTTng-UST log levels is similar to the levels found
2340 in typical logging frameworks:
2342 * In a logging framework, the log level is given by the function
2343 or method name you use at the log statement site: `debug()`,
2344 `info()`, `warn()`, `error()`, and so on.
2345 * In LTTng-UST, you statically assign the log level to a tracepoint
2346 definition; any `tracepoint()` macro invocation which refers to
2347 this definition has this log level.
2349 You can assign a log level to a tracepoint definition with the
2350 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2351 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2352 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2355 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2358 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2360 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2365 * `provider_name` with the tracepoint provider name.
2366 * `tracepoint_name` with the tracepoint name.
2367 * `log_level` with the log level to assign to the tracepoint
2368 definition named `tracepoint_name` in the `provider_name`
2369 tracepoint provider.
2371 See man:lttng-ust(3) for a list of available log level names.
2373 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2377 /* Tracepoint definition */
2386 ctf_integer(int, userid, userid)
2387 ctf_integer(size_t, len, len)
2391 /* Log level assignment */
2392 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2398 ===== Create a tracepoint provider package source file
2400 A _tracepoint provider package source file_ is a C source file which
2401 includes a <<tpp-header,tracepoint provider header file>> to expand its
2402 macros into event serialization and other functions.
2404 You can always use the following tracepoint provider package source
2408 .Tracepoint provider package source file template.
2410 #define TRACEPOINT_CREATE_PROBES
2415 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2416 header file>> name. You may also include more than one tracepoint
2417 provider header file here to create a tracepoint provider package
2418 holding more than one tracepoint providers.
2421 [[probing-the-application-source-code]]
2422 ==== Add tracepoints to an application's source code
2424 Once you <<tpp-header,create a tracepoint provider header file>>, you
2425 can use the `tracepoint()` macro in your application's
2426 source code to insert the tracepoints that this header
2427 <<defining-tracepoints,defines>>.
2429 The `tracepoint()` macro takes at least two parameters: the tracepoint
2430 provider name and the tracepoint name. The corresponding tracepoint
2431 definition defines the other parameters.
2433 .`tracepoint()` usage.
2435 The following <<defining-tracepoints,tracepoint definition>> defines a
2436 tracepoint which takes two input arguments and has two output event
2440 .Tracepoint provider header file.
2442 #include "my-custom-structure.h"
2449 const char*, cmd_name
2452 ctf_string(cmd_name, cmd_name)
2453 ctf_integer(int, number_of_args, argc)
2458 You can refer to this tracepoint definition with the `tracepoint()`
2459 macro in your application's source code like this:
2462 .Application's source file.
2466 int main(int argc, char* argv[])
2468 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2474 Note how the application's source code includes
2475 the tracepoint provider header file containing the tracepoint
2476 definitions to use, path:{tp.h}.
2479 .`tracepoint()` usage with a complex tracepoint definition.
2481 Consider this complex tracepoint definition, where multiple event
2482 fields refer to the same input arguments in their argument expression
2486 .Tracepoint provider header file.
2488 /* For `struct stat` */
2489 #include <sys/types.h>
2490 #include <sys/stat.h>
2502 ctf_integer(int, my_constant_field, 23 + 17)
2503 ctf_integer(int, my_int_arg_field, my_int_arg)
2504 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2505 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2506 my_str_arg[2] + my_str_arg[3])
2507 ctf_string(my_str_arg_field, my_str_arg)
2508 ctf_integer_hex(off_t, size_field, st->st_size)
2509 ctf_float(double, size_dbl_field, (double) st->st_size)
2510 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2511 size_t, strlen(my_str_arg) / 2)
2516 You can refer to this tracepoint definition with the `tracepoint()`
2517 macro in your application's source code like this:
2520 .Application's source file.
2522 #define TRACEPOINT_DEFINE
2529 stat("/etc/fstab", &s);
2530 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2536 If you look at the event record that LTTng writes when tracing this
2537 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2538 it should look like this:
2540 .Event record fields
2542 |Field's name |Field's value
2543 |`my_constant_field` |40
2544 |`my_int_arg_field` |23
2545 |`my_int_arg_field2` |529
2547 |`my_str_arg_field` |`Hello, World!`
2548 |`size_field` |0x12d
2549 |`size_dbl_field` |301.0
2550 |`half_my_str_arg_field` |`Hello,`
2554 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2555 compute--they use the call stack, for example. To avoid this
2556 computation when the tracepoint is disabled, you can use the
2557 `tracepoint_enabled()` and `do_tracepoint()` macros.
2559 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2563 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2565 tracepoint_enabled(provider_name, tracepoint_name)
2566 do_tracepoint(provider_name, tracepoint_name, ...)
2571 * `provider_name` with the tracepoint provider name.
2572 * `tracepoint_name` with the tracepoint name.
2574 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2575 `tracepoint_name` from the provider named `provider_name` is enabled
2578 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2579 if the tracepoint is enabled. Using `tracepoint()` with
2580 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2581 the `tracepoint_enabled()` check, thus a race condition is
2582 possible in this situation:
2585 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2587 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2588 stuff = prepare_stuff();
2591 tracepoint(my_provider, my_tracepoint, stuff);
2594 If the tracepoint is enabled after the condition, then `stuff` is not
2595 prepared: the emitted event will either contain wrong data, or the whole
2596 application could crash (segmentation fault, for example).
2598 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2599 `STAP_PROBEV()` call. If you need it, you must emit
2603 [[building-tracepoint-providers-and-user-application]]
2604 ==== Build and link a tracepoint provider package and an application
2606 Once you have one or more <<tpp-header,tracepoint provider header
2607 files>> and a <<tpp-source,tracepoint provider package source file>>,
2608 you can create the tracepoint provider package by compiling its source
2609 file. From here, multiple build and run scenarios are possible. The
2610 following table shows common application and library configurations
2611 along with the required command lines to achieve them.
2613 In the following diagrams, we use the following file names:
2616 Executable application.
2619 Application's object file.
2622 Tracepoint provider package object file.
2625 Tracepoint provider package archive file.
2628 Tracepoint provider package shared object file.
2631 User library object file.
2634 User library shared object file.
2636 We use the following symbols in the diagrams of table below:
2639 .Symbols used in the build scenario diagrams.
2640 image::ust-sit-symbols.png[]
2642 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2643 variable in the following instructions.
2645 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2646 .Common tracepoint provider package scenarios.
2648 |Scenario |Instructions
2651 The instrumented application is statically linked with
2652 the tracepoint provider package object.
2654 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2657 include::../common/ust-sit-step-tp-o.txt[]
2659 To build the instrumented application:
2661 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2666 #define TRACEPOINT_DEFINE
2670 . Compile the application source file:
2679 . Build the application:
2684 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2688 To run the instrumented application:
2690 * Start the application:
2700 The instrumented application is statically linked with the
2701 tracepoint provider package archive file.
2703 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2706 To create the tracepoint provider package archive file:
2708 . Compile the <<tpp-source,tracepoint provider package source file>>:
2717 . Create the tracepoint provider package archive file:
2722 $ ar rcs tpp.a tpp.o
2726 To build the instrumented application:
2728 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2733 #define TRACEPOINT_DEFINE
2737 . Compile the application source file:
2746 . Build the application:
2751 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2755 To run the instrumented application:
2757 * Start the application:
2767 The instrumented application is linked with the tracepoint provider
2768 package shared object.
2770 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2773 include::../common/ust-sit-step-tp-so.txt[]
2775 To build the instrumented application:
2777 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2782 #define TRACEPOINT_DEFINE
2786 . Compile the application source file:
2795 . Build the application:
2800 $ gcc -o app app.o -ldl -L. -ltpp
2804 To run the instrumented application:
2806 * Start the application:
2816 The tracepoint provider package shared object is preloaded before the
2817 instrumented application starts.
2819 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2822 include::../common/ust-sit-step-tp-so.txt[]
2824 To build the instrumented application:
2826 . In path:{app.c}, before including path:{tpp.h}, add the
2832 #define TRACEPOINT_DEFINE
2833 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2837 . Compile the application source file:
2846 . Build the application:
2851 $ gcc -o app app.o -ldl
2855 To run the instrumented application with tracing support:
2857 * Preload the tracepoint provider package shared object and
2858 start the application:
2863 $ LD_PRELOAD=./libtpp.so ./app
2867 To run the instrumented application without tracing support:
2869 * Start the application:
2879 The instrumented application dynamically loads the tracepoint provider
2880 package shared object.
2882 See the <<dlclose-warning,warning about `dlclose()`>>.
2884 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
2887 include::../common/ust-sit-step-tp-so.txt[]
2889 To build the instrumented application:
2891 . In path:{app.c}, before including path:{tpp.h}, add the
2897 #define TRACEPOINT_DEFINE
2898 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2902 . Compile the application source file:
2911 . Build the application:
2916 $ gcc -o app app.o -ldl
2920 To run the instrumented application:
2922 * Start the application:
2932 The application is linked with the instrumented user library.
2934 The instrumented user library is statically linked with the tracepoint
2935 provider package object file.
2937 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
2940 include::../common/ust-sit-step-tp-o-fpic.txt[]
2942 To build the instrumented user library:
2944 . In path:{emon.c}, before including path:{tpp.h}, add the
2950 #define TRACEPOINT_DEFINE
2954 . Compile the user library source file:
2959 $ gcc -I. -fpic -c emon.c
2963 . Build the user library shared object:
2968 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
2972 To build the application:
2974 . Compile the application source file:
2983 . Build the application:
2988 $ gcc -o app app.o -L. -lemon
2992 To run the application:
2994 * Start the application:
3004 The application is linked with the instrumented user library.
3006 The instrumented user library is linked with the tracepoint provider
3007 package shared object.
3009 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3012 include::../common/ust-sit-step-tp-so.txt[]
3014 To build the instrumented user library:
3016 . In path:{emon.c}, before including path:{tpp.h}, add the
3022 #define TRACEPOINT_DEFINE
3026 . Compile the user library source file:
3031 $ gcc -I. -fpic -c emon.c
3035 . Build the user library shared object:
3040 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3044 To build the application:
3046 . Compile the application source file:
3055 . Build the application:
3060 $ gcc -o app app.o -L. -lemon
3064 To run the application:
3066 * Start the application:
3076 The tracepoint provider package shared object is preloaded before the
3079 The application is linked with the instrumented user library.
3081 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3084 include::../common/ust-sit-step-tp-so.txt[]
3086 To build the instrumented user library:
3088 . In path:{emon.c}, before including path:{tpp.h}, add the
3094 #define TRACEPOINT_DEFINE
3095 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3099 . Compile the user library source file:
3104 $ gcc -I. -fpic -c emon.c
3108 . Build the user library shared object:
3113 $ gcc -shared -o libemon.so emon.o -ldl
3117 To build the application:
3119 . Compile the application source file:
3128 . Build the application:
3133 $ gcc -o app app.o -L. -lemon
3137 To run the application with tracing support:
3139 * Preload the tracepoint provider package shared object and
3140 start the application:
3145 $ LD_PRELOAD=./libtpp.so ./app
3149 To run the application without tracing support:
3151 * Start the application:
3161 The application is linked with the instrumented user library.
3163 The instrumented user library dynamically loads the tracepoint provider
3164 package shared object.
3166 See the <<dlclose-warning,warning about `dlclose()`>>.
3168 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3171 include::../common/ust-sit-step-tp-so.txt[]
3173 To build the instrumented user library:
3175 . In path:{emon.c}, before including path:{tpp.h}, add the
3181 #define TRACEPOINT_DEFINE
3182 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3186 . Compile the user library source file:
3191 $ gcc -I. -fpic -c emon.c
3195 . Build the user library shared object:
3200 $ gcc -shared -o libemon.so emon.o -ldl
3204 To build the application:
3206 . Compile the application source file:
3215 . Build the application:
3220 $ gcc -o app app.o -L. -lemon
3224 To run the application:
3226 * Start the application:
3236 The application dynamically loads the instrumented user library.
3238 The instrumented user library is linked with the tracepoint provider
3239 package shared object.
3241 See the <<dlclose-warning,warning about `dlclose()`>>.
3243 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3246 include::../common/ust-sit-step-tp-so.txt[]
3248 To build the instrumented user library:
3250 . In path:{emon.c}, before including path:{tpp.h}, add the
3256 #define TRACEPOINT_DEFINE
3260 . Compile the user library source file:
3265 $ gcc -I. -fpic -c emon.c
3269 . Build the user library shared object:
3274 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3278 To build the application:
3280 . Compile the application source file:
3289 . Build the application:
3294 $ gcc -o app app.o -ldl -L. -lemon
3298 To run the application:
3300 * Start the application:
3310 The application dynamically loads the instrumented user library.
3312 The instrumented user library dynamically loads the tracepoint provider
3313 package shared object.
3315 See the <<dlclose-warning,warning about `dlclose()`>>.
3317 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3320 include::../common/ust-sit-step-tp-so.txt[]
3322 To build the instrumented user library:
3324 . In path:{emon.c}, before including path:{tpp.h}, add the
3330 #define TRACEPOINT_DEFINE
3331 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3335 . Compile the user library source file:
3340 $ gcc -I. -fpic -c emon.c
3344 . Build the user library shared object:
3349 $ gcc -shared -o libemon.so emon.o -ldl
3353 To build the application:
3355 . Compile the application source file:
3364 . Build the application:
3369 $ gcc -o app app.o -ldl -L. -lemon
3373 To run the application:
3375 * Start the application:
3385 The tracepoint provider package shared object is preloaded before the
3388 The application dynamically loads the instrumented user library.
3390 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3393 include::../common/ust-sit-step-tp-so.txt[]
3395 To build the instrumented user library:
3397 . In path:{emon.c}, before including path:{tpp.h}, add the
3403 #define TRACEPOINT_DEFINE
3404 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3408 . Compile the user library source file:
3413 $ gcc -I. -fpic -c emon.c
3417 . Build the user library shared object:
3422 $ gcc -shared -o libemon.so emon.o -ldl
3426 To build the application:
3428 . Compile the application source file:
3437 . Build the application:
3442 $ gcc -o app app.o -L. -lemon
3446 To run the application with tracing support:
3448 * Preload the tracepoint provider package shared object and
3449 start the application:
3454 $ LD_PRELOAD=./libtpp.so ./app
3458 To run the application without tracing support:
3460 * Start the application:
3470 The application is statically linked with the tracepoint provider
3471 package object file.
3473 The application is linked with the instrumented user library.
3475 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3478 include::../common/ust-sit-step-tp-o.txt[]
3480 To build the instrumented user library:
3482 . In path:{emon.c}, before including path:{tpp.h}, add the
3488 #define TRACEPOINT_DEFINE
3492 . Compile the user library source file:
3497 $ gcc -I. -fpic -c emon.c
3501 . Build the user library shared object:
3506 $ gcc -shared -o libemon.so emon.o
3510 To build the application:
3512 . Compile the application source file:
3521 . Build the application:
3526 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3530 To run the instrumented application:
3532 * Start the application:
3542 The application is statically linked with the tracepoint provider
3543 package object file.
3545 The application dynamically loads the instrumented user library.
3547 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3550 include::../common/ust-sit-step-tp-o.txt[]
3552 To build the application:
3554 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3559 #define TRACEPOINT_DEFINE
3563 . Compile the application source file:
3572 . Build the application:
3577 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3582 The `--export-dynamic` option passed to the linker is necessary for the
3583 dynamically loaded library to ``see'' the tracepoint symbols defined in
3586 To build the instrumented user library:
3588 . Compile the user library source file:
3593 $ gcc -I. -fpic -c emon.c
3597 . Build the user library shared object:
3602 $ gcc -shared -o libemon.so emon.o
3606 To run the application:
3608 * Start the application:
3620 .Do not use man:dlclose(3) on a tracepoint provider package
3622 Never use man:dlclose(3) on any shared object which:
3624 * Is linked with, statically or dynamically, a tracepoint provider
3626 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3627 package shared object.
3629 This is currently considered **unsafe** due to a lack of reference
3630 counting from LTTng-UST to the shared object.
3632 A known workaround (available since glibc 2.2) is to use the
3633 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3634 effect of not unloading the loaded shared object, even if man:dlclose(3)
3637 You can also preload the tracepoint provider package shared object with
3638 the env:LD_PRELOAD environment variable to overcome this limitation.
3642 [[using-lttng-ust-with-daemons]]
3643 ===== Use noch:{LTTng-UST} with daemons
3645 If your instrumented application calls man:fork(2), man:clone(2),
3646 or BSD's man:rfork(2), without a following man:exec(3)-family
3647 system call, you must preload the path:{liblttng-ust-fork.so} shared
3648 object when you start the application.
3652 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3655 If your tracepoint provider package is
3656 a shared library which you also preload, you must put both
3657 shared objects in env:LD_PRELOAD:
3661 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3667 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3669 If your instrumented application closes one or more file descriptors
3670 which it did not open itself, you must preload the
3671 path:{liblttng-ust-fd.so} shared object when you start the application:
3675 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3678 Typical use cases include closing all the file descriptors after
3679 man:fork(2) or man:rfork(2) and buggy applications doing
3683 [[lttng-ust-pkg-config]]
3684 ===== Use noch:{pkg-config}
3686 On some distributions, LTTng-UST ships with a
3687 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3688 metadata file. If this is your case, then you can use cmd:pkg-config to
3689 build an application on the command line:
3693 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3697 [[instrumenting-32-bit-app-on-64-bit-system]]
3698 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3700 In order to trace a 32-bit application running on a 64-bit system,
3701 LTTng must use a dedicated 32-bit
3702 <<lttng-consumerd,consumer daemon>>.
3704 The following steps show how to build and install a 32-bit consumer
3705 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3706 build and install the 32-bit LTTng-UST libraries, and how to build and
3707 link an instrumented 32-bit application in that context.
3709 To build a 32-bit instrumented application for a 64-bit target system,
3710 assuming you have a fresh target system with no installed Userspace RCU
3713 . Download, build, and install a 32-bit version of Userspace RCU:
3718 $ cd $(mktemp -d) &&
3719 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3720 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3721 cd userspace-rcu-0.9.* &&
3722 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3724 sudo make install &&
3729 . Using your distribution's package manager, or from source, install
3730 the following 32-bit versions of the following dependencies of
3731 LTTng-tools and LTTng-UST:
3734 * https://sourceforge.net/projects/libuuid/[libuuid]
3735 * http://directory.fsf.org/wiki/Popt[popt]
3736 * http://www.xmlsoft.org/[libxml2]
3739 . Download, build, and install a 32-bit version of the latest
3740 LTTng-UST{nbsp}{revision}:
3745 $ cd $(mktemp -d) &&
3746 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.10.tar.bz2 &&
3747 tar -xf lttng-ust-latest-2.10.tar.bz2 &&
3748 cd lttng-ust-2.10.* &&
3749 ./configure --libdir=/usr/local/lib32 \
3750 CFLAGS=-m32 CXXFLAGS=-m32 \
3751 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3753 sudo make install &&
3760 Depending on your distribution,
3761 32-bit libraries could be installed at a different location than
3762 `/usr/lib32`. For example, Debian is known to install
3763 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3765 In this case, make sure to set `LDFLAGS` to all the
3766 relevant 32-bit library paths, for example:
3770 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3774 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3775 the 32-bit consumer daemon:
3780 $ cd $(mktemp -d) &&
3781 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
3782 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
3783 cd lttng-tools-2.10.* &&
3784 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3785 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3786 --disable-bin-lttng --disable-bin-lttng-crash \
3787 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3789 cd src/bin/lttng-consumerd &&
3790 sudo make install &&
3795 . From your distribution or from source,
3796 <<installing-lttng,install>> the 64-bit versions of
3797 LTTng-UST and Userspace RCU.
3798 . Download, build, and install the 64-bit version of the
3799 latest LTTng-tools{nbsp}{revision}:
3804 $ cd $(mktemp -d) &&
3805 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.10.tar.bz2 &&
3806 tar -xf lttng-tools-latest-2.10.tar.bz2 &&
3807 cd lttng-tools-2.10.* &&
3808 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3809 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3811 sudo make install &&
3816 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3817 when linking your 32-bit application:
3820 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3821 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3824 For example, let's rebuild the quick start example in
3825 <<tracing-your-own-user-application,Trace a user application>> as an
3826 instrumented 32-bit application:
3831 $ gcc -m32 -c -I. hello-tp.c
3832 $ gcc -m32 -c hello.c
3833 $ gcc -m32 -o hello hello.o hello-tp.o \
3834 -L/usr/lib32 -L/usr/local/lib32 \
3835 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3840 No special action is required to execute the 32-bit application and
3841 to trace it: use the command-line man:lttng(1) tool as usual.
3848 man:tracef(3) is a small LTTng-UST API designed for quick,
3849 man:printf(3)-like instrumentation without the burden of
3850 <<tracepoint-provider,creating>> and
3851 <<building-tracepoint-providers-and-user-application,building>>
3852 a tracepoint provider package.
3854 To use `tracef()` in your application:
3856 . In the C or C++ source files where you need to use `tracef()`,
3857 include `<lttng/tracef.h>`:
3862 #include <lttng/tracef.h>
3866 . In the application's source code, use `tracef()` like you would use
3874 tracef("my message: %d (%s)", my_integer, my_string);
3880 . Link your application with `liblttng-ust`:
3885 $ gcc -o app app.c -llttng-ust
3889 To trace the events that `tracef()` calls emit:
3891 * <<enabling-disabling-events,Create an event rule>> which matches the
3892 `lttng_ust_tracef:*` event name:
3897 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
3902 .Limitations of `tracef()`
3904 The `tracef()` utility function was developed to make user space tracing
3905 super simple, albeit with notable disadvantages compared to
3906 <<defining-tracepoints,user-defined tracepoints>>:
3908 * All the emitted events have the same tracepoint provider and
3909 tracepoint names, respectively `lttng_ust_tracef` and `event`.
3910 * There is no static type checking.
3911 * The only event record field you actually get, named `msg`, is a string
3912 potentially containing the values you passed to `tracef()`
3913 using your own format string. This also means that you cannot filter
3914 events with a custom expression at run time because there are no
3916 * Since `tracef()` uses the C standard library's man:vasprintf(3)
3917 function behind the scenes to format the strings at run time, its
3918 expected performance is lower than with user-defined tracepoints,
3919 which do not require a conversion to a string.
3921 Taking this into consideration, `tracef()` is useful for some quick
3922 prototyping and debugging, but you should not consider it for any
3923 permanent and serious applicative instrumentation.
3929 ==== Use `tracelog()`
3931 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
3932 the difference that it accepts an additional log level parameter.
3934 The goal of `tracelog()` is to ease the migration from logging to
3937 To use `tracelog()` in your application:
3939 . In the C or C++ source files where you need to use `tracelog()`,
3940 include `<lttng/tracelog.h>`:
3945 #include <lttng/tracelog.h>
3949 . In the application's source code, use `tracelog()` like you would use
3950 man:printf(3), except for the first parameter which is the log
3958 tracelog(TRACE_WARNING, "my message: %d (%s)",
3959 my_integer, my_string);
3965 See man:lttng-ust(3) for a list of available log level names.
3967 . Link your application with `liblttng-ust`:
3972 $ gcc -o app app.c -llttng-ust
3976 To trace the events that `tracelog()` calls emit with a log level
3977 _as severe as_ a specific log level:
3979 * <<enabling-disabling-events,Create an event rule>> which matches the
3980 `lttng_ust_tracelog:*` event name and a minimum level
3986 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
3987 --loglevel=TRACE_WARNING
3991 To trace the events that `tracelog()` calls emit with a
3992 _specific log level_:
3994 * Create an event rule which matches the `lttng_ust_tracelog:*`
3995 event name and a specific log level:
4000 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4001 --loglevel-only=TRACE_INFO
4006 [[prebuilt-ust-helpers]]
4007 === Prebuilt user space tracing helpers
4009 The LTTng-UST package provides a few helpers in the form or preloadable
4010 shared objects which automatically instrument system functions and
4013 The helper shared objects are normally found in dir:{/usr/lib}. If you
4014 built LTTng-UST <<building-from-source,from source>>, they are probably
4015 located in dir:{/usr/local/lib}.
4017 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4020 path:{liblttng-ust-libc-wrapper.so}::
4021 path:{liblttng-ust-pthread-wrapper.so}::
4022 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4023 memory and POSIX threads function tracing>>.
4025 path:{liblttng-ust-cyg-profile.so}::
4026 path:{liblttng-ust-cyg-profile-fast.so}::
4027 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4029 path:{liblttng-ust-dl.so}::
4030 <<liblttng-ust-dl,Dynamic linker tracing>>.
4032 To use a user space tracing helper with any user application:
4034 * Preload the helper shared object when you start the application:
4039 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4043 You can preload more than one helper:
4048 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4054 [[liblttng-ust-libc-pthread-wrapper]]
4055 ==== Instrument C standard library memory and POSIX threads functions
4057 The path:{liblttng-ust-libc-wrapper.so} and
4058 path:{liblttng-ust-pthread-wrapper.so} helpers
4059 add instrumentation to some C standard library and POSIX
4063 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4065 |TP provider name |TP name |Instrumented function
4067 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4068 |`calloc` |man:calloc(3)
4069 |`realloc` |man:realloc(3)
4070 |`free` |man:free(3)
4071 |`memalign` |man:memalign(3)
4072 |`posix_memalign` |man:posix_memalign(3)
4076 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4078 |TP provider name |TP name |Instrumented function
4080 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4081 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4082 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4083 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4086 When you preload the shared object, it replaces the functions listed
4087 in the previous tables by wrappers which contain tracepoints and call
4088 the replaced functions.
4091 [[liblttng-ust-cyg-profile]]
4092 ==== Instrument function entry and exit
4094 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4095 to the entry and exit points of functions.
4097 man:gcc(1) and man:clang(1) have an option named
4098 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4099 which generates instrumentation calls for entry and exit to functions.
4100 The LTTng-UST function tracing helpers,
4101 path:{liblttng-ust-cyg-profile.so} and
4102 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4103 to add tracepoints to the two generated functions (which contain
4104 `cyg_profile` in their names, hence the helper's name).
4106 To use the LTTng-UST function tracing helper, the source files to
4107 instrument must be built using the `-finstrument-functions` compiler
4110 There are two versions of the LTTng-UST function tracing helper:
4112 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4113 that you should only use when it can be _guaranteed_ that the
4114 complete event stream is recorded without any lost event record.
4115 Any kind of duplicate information is left out.
4117 Assuming no event record is lost, having only the function addresses on
4118 entry is enough to create a call graph, since an event record always
4119 contains the ID of the CPU that generated it.
4121 You can use a tool like man:addr2line(1) to convert function addresses
4122 back to source file names and line numbers.
4124 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4125 which also works in use cases where event records might get discarded or
4126 not recorded from application startup.
4127 In these cases, the trace analyzer needs more information to be
4128 able to reconstruct the program flow.
4130 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4131 points of this helper.
4133 All the tracepoints that this helper provides have the
4134 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4136 TIP: It's sometimes a good idea to limit the number of source files that
4137 you compile with the `-finstrument-functions` option to prevent LTTng
4138 from writing an excessive amount of trace data at run time. When using
4139 man:gcc(1), you can use the
4140 `-finstrument-functions-exclude-function-list` option to avoid
4141 instrument entries and exits of specific function names.
4146 ==== Instrument the dynamic linker
4148 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4149 man:dlopen(3) and man:dlclose(3) function calls.
4151 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4156 [[java-application]]
4157 === User space Java agent
4159 You can instrument any Java application which uses one of the following
4162 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4163 (JUL) core logging facilities.
4164 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4165 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4168 .LTTng-UST Java agent imported by a Java application.
4169 image::java-app.png[]
4171 Note that the methods described below are new in LTTng{nbsp}{revision}.
4172 Previous LTTng versions use another technique.
4174 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4175 and https://ci.lttng.org/[continuous integration], thus this version is
4176 directly supported. However, the LTTng-UST Java agent is also tested
4177 with OpenJDK{nbsp}7.
4182 ==== Use the LTTng-UST Java agent for `java.util.logging`
4184 To use the LTTng-UST Java agent in a Java application which uses
4185 `java.util.logging` (JUL):
4187 . In the Java application's source code, import the LTTng-UST
4188 log handler package for `java.util.logging`:
4193 import org.lttng.ust.agent.jul.LttngLogHandler;
4197 . Create an LTTng-UST JUL log handler:
4202 Handler lttngUstLogHandler = new LttngLogHandler();
4206 . Add this handler to the JUL loggers which should emit LTTng events:
4211 Logger myLogger = Logger.getLogger("some-logger");
4213 myLogger.addHandler(lttngUstLogHandler);
4217 . Use `java.util.logging` log statements and configuration as usual.
4218 The loggers with an attached LTTng-UST log handler can emit
4221 . Before exiting the application, remove the LTTng-UST log handler from
4222 the loggers attached to it and call its `close()` method:
4227 myLogger.removeHandler(lttngUstLogHandler);
4228 lttngUstLogHandler.close();
4232 This is not strictly necessary, but it is recommended for a clean
4233 disposal of the handler's resources.
4235 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4236 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4238 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4239 path] when you build the Java application.
4241 The JAR files are typically located in dir:{/usr/share/java}.
4243 IMPORTANT: The LTTng-UST Java agent must be
4244 <<installing-lttng,installed>> for the logging framework your
4247 .Use the LTTng-UST Java agent for `java.util.logging`.
4252 import java.io.IOException;
4253 import java.util.logging.Handler;
4254 import java.util.logging.Logger;
4255 import org.lttng.ust.agent.jul.LttngLogHandler;
4259 private static final int answer = 42;
4261 public static void main(String[] argv) throws Exception
4264 Logger logger = Logger.getLogger("jello");
4266 // Create an LTTng-UST log handler
4267 Handler lttngUstLogHandler = new LttngLogHandler();
4269 // Add the LTTng-UST log handler to our logger
4270 logger.addHandler(lttngUstLogHandler);
4273 logger.info("some info");
4274 logger.warning("some warning");
4276 logger.finer("finer information; the answer is " + answer);
4278 logger.severe("error!");
4280 // Not mandatory, but cleaner
4281 logger.removeHandler(lttngUstLogHandler);
4282 lttngUstLogHandler.close();
4291 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4294 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4295 <<enabling-disabling-events,create an event rule>> matching the
4296 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4301 $ lttng enable-event --jul jello
4305 Run the compiled class:
4309 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4312 <<basic-tracing-session-control,Stop tracing>> and inspect the
4322 In the resulting trace, an <<event,event record>> generated by a Java
4323 application using `java.util.logging` is named `lttng_jul:event` and
4324 has the following fields:
4327 Log record's message.
4333 Name of the class in which the log statement was executed.
4336 Name of the method in which the log statement was executed.
4339 Logging time (timestamp in milliseconds).
4342 Log level integer value.
4345 ID of the thread in which the log statement was executed.
4347 You can use the opt:lttng-enable-event(1):--loglevel or
4348 opt:lttng-enable-event(1):--loglevel-only option of the
4349 man:lttng-enable-event(1) command to target a range of JUL log levels
4350 or a specific JUL log level.
4355 ==== Use the LTTng-UST Java agent for Apache log4j
4357 To use the LTTng-UST Java agent in a Java application which uses
4360 . In the Java application's source code, import the LTTng-UST
4361 log appender package for Apache log4j:
4366 import org.lttng.ust.agent.log4j.LttngLogAppender;
4370 . Create an LTTng-UST log4j log appender:
4375 Appender lttngUstLogAppender = new LttngLogAppender();
4379 . Add this appender to the log4j loggers which should emit LTTng events:
4384 Logger myLogger = Logger.getLogger("some-logger");
4386 myLogger.addAppender(lttngUstLogAppender);
4390 . Use Apache log4j log statements and configuration as usual. The
4391 loggers with an attached LTTng-UST log appender can emit LTTng events.
4393 . Before exiting the application, remove the LTTng-UST log appender from
4394 the loggers attached to it and call its `close()` method:
4399 myLogger.removeAppender(lttngUstLogAppender);
4400 lttngUstLogAppender.close();
4404 This is not strictly necessary, but it is recommended for a clean
4405 disposal of the appender's resources.
4407 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4408 files, path:{lttng-ust-agent-common.jar} and
4409 path:{lttng-ust-agent-log4j.jar}, in the
4410 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4411 path] when you build the Java application.
4413 The JAR files are typically located in dir:{/usr/share/java}.
4415 IMPORTANT: The LTTng-UST Java agent must be
4416 <<installing-lttng,installed>> for the logging framework your
4419 .Use the LTTng-UST Java agent for Apache log4j.
4424 import org.apache.log4j.Appender;
4425 import org.apache.log4j.Logger;
4426 import org.lttng.ust.agent.log4j.LttngLogAppender;
4430 private static final int answer = 42;
4432 public static void main(String[] argv) throws Exception
4435 Logger logger = Logger.getLogger("jello");
4437 // Create an LTTng-UST log appender
4438 Appender lttngUstLogAppender = new LttngLogAppender();
4440 // Add the LTTng-UST log appender to our logger
4441 logger.addAppender(lttngUstLogAppender);
4444 logger.info("some info");
4445 logger.warn("some warning");
4447 logger.debug("debug information; the answer is " + answer);
4449 logger.fatal("error!");
4451 // Not mandatory, but cleaner
4452 logger.removeAppender(lttngUstLogAppender);
4453 lttngUstLogAppender.close();
4459 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4464 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4467 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4468 <<enabling-disabling-events,create an event rule>> matching the
4469 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4474 $ lttng enable-event --log4j jello
4478 Run the compiled class:
4482 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4485 <<basic-tracing-session-control,Stop tracing>> and inspect the
4495 In the resulting trace, an <<event,event record>> generated by a Java
4496 application using log4j is named `lttng_log4j:event` and
4497 has the following fields:
4500 Log record's message.
4506 Name of the class in which the log statement was executed.
4509 Name of the method in which the log statement was executed.
4512 Name of the file in which the executed log statement is located.
4515 Line number at which the log statement was executed.
4521 Log level integer value.
4524 Name of the Java thread in which the log statement was executed.
4526 You can use the opt:lttng-enable-event(1):--loglevel or
4527 opt:lttng-enable-event(1):--loglevel-only option of the
4528 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4529 or a specific log4j log level.
4533 [[java-application-context]]
4534 ==== Provide application-specific context fields in a Java application
4536 A Java application-specific context field is a piece of state provided
4537 by the application which <<adding-context,you can add>>, using the
4538 man:lttng-add-context(1) command, to each <<event,event record>>
4539 produced by the log statements of this application.
4541 For example, a given object might have a current request ID variable.
4542 You can create a context information retriever for this object and
4543 assign a name to this current request ID. You can then, using the
4544 man:lttng-add-context(1) command, add this context field by name to
4545 the JUL or log4j <<channel,channel>>.
4547 To provide application-specific context fields in a Java application:
4549 . In the Java application's source code, import the LTTng-UST
4550 Java agent context classes and interfaces:
4555 import org.lttng.ust.agent.context.ContextInfoManager;
4556 import org.lttng.ust.agent.context.IContextInfoRetriever;
4560 . Create a context information retriever class, that is, a class which
4561 implements the `IContextInfoRetriever` interface:
4566 class MyContextInfoRetriever implements IContextInfoRetriever
4569 public Object retrieveContextInfo(String key)
4571 if (key.equals("intCtx")) {
4573 } else if (key.equals("strContext")) {
4574 return "context value!";
4583 This `retrieveContextInfo()` method is the only member of the
4584 `IContextInfoRetriever` interface. Its role is to return the current
4585 value of a state by name to create a context field. The names of the
4586 context fields and which state variables they return depends on your
4589 All primitive types and objects are supported as context fields.
4590 When `retrieveContextInfo()` returns an object, the context field
4591 serializer calls its `toString()` method to add a string field to
4592 event records. The method can also return `null`, which means that
4593 no context field is available for the required name.
4595 . Register an instance of your context information retriever class to
4596 the context information manager singleton:
4601 IContextInfoRetriever cir = new MyContextInfoRetriever();
4602 ContextInfoManager cim = ContextInfoManager.getInstance();
4603 cim.registerContextInfoRetriever("retrieverName", cir);
4607 . Before exiting the application, remove your context information
4608 retriever from the context information manager singleton:
4613 ContextInfoManager cim = ContextInfoManager.getInstance();
4614 cim.unregisterContextInfoRetriever("retrieverName");
4618 This is not strictly necessary, but it is recommended for a clean
4619 disposal of some manager's resources.
4621 . Build your Java application with LTTng-UST Java agent support as
4622 usual, following the procedure for either the <<jul,JUL>> or
4623 <<log4j,Apache log4j>> framework.
4626 .Provide application-specific context fields in a Java application.
4631 import java.util.logging.Handler;
4632 import java.util.logging.Logger;
4633 import org.lttng.ust.agent.jul.LttngLogHandler;
4634 import org.lttng.ust.agent.context.ContextInfoManager;
4635 import org.lttng.ust.agent.context.IContextInfoRetriever;
4639 // Our context information retriever class
4640 private static class MyContextInfoRetriever
4641 implements IContextInfoRetriever
4644 public Object retrieveContextInfo(String key) {
4645 if (key.equals("intCtx")) {
4647 } else if (key.equals("strContext")) {
4648 return "context value!";
4655 private static final int answer = 42;
4657 public static void main(String args[]) throws Exception
4659 // Get the context information manager instance
4660 ContextInfoManager cim = ContextInfoManager.getInstance();
4662 // Create and register our context information retriever
4663 IContextInfoRetriever cir = new MyContextInfoRetriever();
4664 cim.registerContextInfoRetriever("myRetriever", cir);
4667 Logger logger = Logger.getLogger("jello");
4669 // Create an LTTng-UST log handler
4670 Handler lttngUstLogHandler = new LttngLogHandler();
4672 // Add the LTTng-UST log handler to our logger
4673 logger.addHandler(lttngUstLogHandler);
4676 logger.info("some info");
4677 logger.warning("some warning");
4679 logger.finer("finer information; the answer is " + answer);
4681 logger.severe("error!");
4683 // Not mandatory, but cleaner
4684 logger.removeHandler(lttngUstLogHandler);
4685 lttngUstLogHandler.close();
4686 cim.unregisterContextInfoRetriever("myRetriever");
4695 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4698 <<creating-destroying-tracing-sessions,Create a tracing session>>
4699 and <<enabling-disabling-events,create an event rule>> matching the
4705 $ lttng enable-event --jul jello
4708 <<adding-context,Add the application-specific context fields>> to the
4713 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4714 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4717 <<basic-tracing-session-control,Start tracing>>:
4724 Run the compiled class:
4728 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4731 <<basic-tracing-session-control,Stop tracing>> and inspect the
4743 [[python-application]]
4744 === User space Python agent
4746 You can instrument a Python 2 or Python 3 application which uses the
4747 standard https://docs.python.org/3/library/logging.html[`logging`]
4750 Each log statement emits an LTTng event once the
4751 application module imports the
4752 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4755 .A Python application importing the LTTng-UST Python agent.
4756 image::python-app.png[]
4758 To use the LTTng-UST Python agent:
4760 . In the Python application's source code, import the LTTng-UST Python
4770 The LTTng-UST Python agent automatically adds its logging handler to the
4771 root logger at import time.
4773 Any log statement that the application executes before this import does
4774 not emit an LTTng event.
4776 IMPORTANT: The LTTng-UST Python agent must be
4777 <<installing-lttng,installed>>.
4779 . Use log statements and logging configuration as usual.
4780 Since the LTTng-UST Python agent adds a handler to the _root_
4781 logger, you can trace any log statement from any logger.
4783 .Use the LTTng-UST Python agent.
4794 logging.basicConfig()
4795 logger = logging.getLogger('my-logger')
4798 logger.debug('debug message')
4799 logger.info('info message')
4800 logger.warn('warn message')
4801 logger.error('error message')
4802 logger.critical('critical message')
4806 if __name__ == '__main__':
4810 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4811 logging handler which prints to the standard error stream, is not
4812 strictly required for LTTng-UST tracing to work, but in versions of
4813 Python preceding 3.2, you could see a warning message which indicates
4814 that no handler exists for the logger `my-logger`.
4816 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4817 <<enabling-disabling-events,create an event rule>> matching the
4818 `my-logger` Python logger, and <<basic-tracing-session-control,start
4824 $ lttng enable-event --python my-logger
4828 Run the Python script:
4835 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4845 In the resulting trace, an <<event,event record>> generated by a Python
4846 application is named `lttng_python:event` and has the following fields:
4849 Logging time (string).
4852 Log record's message.
4858 Name of the function in which the log statement was executed.
4861 Line number at which the log statement was executed.
4864 Log level integer value.
4867 ID of the Python thread in which the log statement was executed.
4870 Name of the Python thread in which the log statement was executed.
4872 You can use the opt:lttng-enable-event(1):--loglevel or
4873 opt:lttng-enable-event(1):--loglevel-only option of the
4874 man:lttng-enable-event(1) command to target a range of Python log levels
4875 or a specific Python log level.
4877 When an application imports the LTTng-UST Python agent, the agent tries
4878 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4879 <<start-sessiond,start the session daemon>> _before_ you run the Python
4880 application. If a session daemon is found, the agent tries to register
4881 to it during 5{nbsp}seconds, after which the application continues
4882 without LTTng tracing support. You can override this timeout value with
4883 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4886 If the session daemon stops while a Python application with an imported
4887 LTTng-UST Python agent runs, the agent retries to connect and to
4888 register to a session daemon every 3{nbsp}seconds. You can override this
4889 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4894 [[proc-lttng-logger-abi]]
4897 The `lttng-tracer` Linux kernel module, part of
4898 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4899 path:{/proc/lttng-logger} when it's loaded. Any application can write
4900 text data to this file to emit an LTTng event.
4903 .An application writes to the LTTng logger file to emit an LTTng event.
4904 image::lttng-logger.png[]
4906 The LTTng logger is the quickest method--not the most efficient,
4907 however--to add instrumentation to an application. It is designed
4908 mostly to instrument shell scripts:
4912 $ echo "Some message, some $variable" > /proc/lttng-logger
4915 Any event that the LTTng logger emits is named `lttng_logger` and
4916 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4917 other instrumentation points in the kernel tracing domain, **any Unix
4918 user** can <<enabling-disabling-events,create an event rule>> which
4919 matches its event name, not only the root user or users in the
4920 <<tracing-group,tracing group>>.
4922 To use the LTTng logger:
4924 * From any application, write text data to the path:{/proc/lttng-logger}
4927 The `msg` field of `lttng_logger` event records contains the
4930 NOTE: The maximum message length of an LTTng logger event is
4931 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4932 than one event to contain the remaining data.
4934 You should not use the LTTng logger to trace a user application which
4935 can be instrumented in a more efficient way, namely:
4937 * <<c-application,C and $$C++$$ applications>>.
4938 * <<java-application,Java applications>>.
4939 * <<python-application,Python applications>>.
4941 .Use the LTTng logger.
4946 echo 'Hello, World!' > /proc/lttng-logger
4948 df --human-readable --print-type / > /proc/lttng-logger
4951 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4952 <<enabling-disabling-events,create an event rule>> matching the
4953 `lttng_logger` Linux kernel tracepoint, and
4954 <<basic-tracing-session-control,start tracing>>:
4959 $ lttng enable-event --kernel lttng_logger
4963 Run the Bash script:
4970 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4981 [[instrumenting-linux-kernel]]
4982 === LTTng kernel tracepoints
4984 NOTE: This section shows how to _add_ instrumentation points to the
4985 Linux kernel. The kernel's subsystems are already thoroughly
4986 instrumented at strategic places for LTTng when you
4987 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
4991 There are two methods to instrument the Linux kernel:
4993 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
4994 tracepoint which uses the `TRACE_EVENT()` API.
4996 Choose this if you want to instrumentation a Linux kernel tree with an
4997 instrumentation point compatible with ftrace, perf, and SystemTap.
4999 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5000 instrument an out-of-tree kernel module.
5002 Choose this if you don't need ftrace, perf, or SystemTap support.
5006 [[linux-add-lttng-layer]]
5007 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5009 This section shows how to add an LTTng layer to existing ftrace
5010 instrumentation using the `TRACE_EVENT()` API.
5012 This section does not document the `TRACE_EVENT()` macro. You can
5013 read the following articles to learn more about this API:
5015 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5016 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5017 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5019 The following procedure assumes that your ftrace tracepoints are
5020 correctly defined in their own header and that they are created in
5021 one source file using the `CREATE_TRACE_POINTS` definition.
5023 To add an LTTng layer over an existing ftrace tracepoint:
5025 . Make sure the following kernel configuration options are
5031 * `CONFIG_HIGH_RES_TIMERS`
5032 * `CONFIG_TRACEPOINTS`
5035 . Build the Linux source tree with your custom ftrace tracepoints.
5036 . Boot the resulting Linux image on your target system.
5038 Confirm that the tracepoints exist by looking for their names in the
5039 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5040 is your subsystem's name.
5042 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5047 $ cd $(mktemp -d) &&
5048 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.10.tar.bz2 &&
5049 tar -xf lttng-modules-latest-2.10.tar.bz2 &&
5050 cd lttng-modules-2.10.*
5054 . In dir:{instrumentation/events/lttng-module}, relative to the root
5055 of the LTTng-modules source tree, create a header file named
5056 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5057 LTTng-modules tracepoint definitions using the LTTng-modules
5060 Start with this template:
5064 .path:{instrumentation/events/lttng-module/my_subsys.h}
5067 #define TRACE_SYSTEM my_subsys
5069 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5070 #define _LTTNG_MY_SUBSYS_H
5072 #include "../../../probes/lttng-tracepoint-event.h"
5073 #include <linux/tracepoint.h>
5075 LTTNG_TRACEPOINT_EVENT(
5077 * Format is identical to TRACE_EVENT()'s version for the three
5078 * following macro parameters:
5081 TP_PROTO(int my_int, const char *my_string),
5082 TP_ARGS(my_int, my_string),
5084 /* LTTng-modules specific macros */
5086 ctf_integer(int, my_int_field, my_int)
5087 ctf_string(my_bar_field, my_bar)
5091 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5093 #include "../../../probes/define_trace.h"
5097 The entries in the `TP_FIELDS()` section are the list of fields for the
5098 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5099 ftrace's `TRACE_EVENT()` macro.
5101 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5102 complete description of the available `ctf_*()` macros.
5104 . Create the LTTng-modules probe's kernel module C source file,
5105 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5110 .path:{probes/lttng-probe-my-subsys.c}
5112 #include <linux/module.h>
5113 #include "../lttng-tracer.h"
5116 * Build-time verification of mismatch between mainline
5117 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5118 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5120 #include <trace/events/my_subsys.h>
5122 /* Create LTTng tracepoint probes */
5123 #define LTTNG_PACKAGE_BUILD
5124 #define CREATE_TRACE_POINTS
5125 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5127 #include "../instrumentation/events/lttng-module/my_subsys.h"
5129 MODULE_LICENSE("GPL and additional rights");
5130 MODULE_AUTHOR("Your name <your-email>");
5131 MODULE_DESCRIPTION("LTTng my_subsys probes");
5132 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5133 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5134 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5135 LTTNG_MODULES_EXTRAVERSION);
5139 . Edit path:{probes/KBuild} and add your new kernel module object
5140 next to the existing ones:
5144 .path:{probes/KBuild}
5148 obj-m += lttng-probe-module.o
5149 obj-m += lttng-probe-power.o
5151 obj-m += lttng-probe-my-subsys.o
5157 . Build and install the LTTng kernel modules:
5162 $ make KERNELDIR=/path/to/linux
5163 # make modules_install && depmod -a
5167 Replace `/path/to/linux` with the path to the Linux source tree where
5168 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5170 Note that you can also use the
5171 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5172 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5173 C code that need to be executed before the event fields are recorded.
5175 The best way to learn how to use the previous LTTng-modules macros is to
5176 inspect the existing LTTng-modules tracepoint definitions in the
5177 dir:{instrumentation/events/lttng-module} header files. Compare them
5178 with the Linux kernel mainline versions in the
5179 dir:{include/trace/events} directory of the Linux source tree.
5183 [[lttng-tracepoint-event-code]]
5184 ===== Use custom C code to access the data for tracepoint fields
5186 Although we recommended to always use the
5187 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5188 the arguments and fields of an LTTng-modules tracepoint when possible,
5189 sometimes you need a more complex process to access the data that the
5190 tracer records as event record fields. In other words, you need local
5191 variables and multiple C{nbsp}statements instead of simple
5192 argument-based expressions that you pass to the
5193 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5195 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5196 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5197 a block of C{nbsp}code to be executed before LTTng records the fields.
5198 The structure of this macro is:
5201 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5203 LTTNG_TRACEPOINT_EVENT_CODE(
5205 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5206 * version for the following three macro parameters:
5209 TP_PROTO(int my_int, const char *my_string),
5210 TP_ARGS(my_int, my_string),
5212 /* Declarations of custom local variables */
5215 unsigned long b = 0;
5216 const char *name = "(undefined)";
5217 struct my_struct *my_struct;
5221 * Custom code which uses both tracepoint arguments
5222 * (in TP_ARGS()) and local variables (in TP_locvar()).
5224 * Local variables are actually members of a structure pointed
5225 * to by the special variable tp_locvar.
5229 tp_locvar->a = my_int + 17;
5230 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5231 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5232 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5233 put_my_struct(tp_locvar->my_struct);
5242 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5243 * version for this, except that tp_locvar members can be
5244 * used in the argument expression parameters of
5245 * the ctf_*() macros.
5248 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5249 ctf_integer(int, my_struct_a, tp_locvar->a)
5250 ctf_string(my_string_field, my_string)
5251 ctf_string(my_struct_name, tp_locvar->name)
5256 IMPORTANT: The C code defined in `TP_code()` must not have any side
5257 effects when executed. In particular, the code must not allocate
5258 memory or get resources without deallocating this memory or putting
5259 those resources afterwards.
5262 [[instrumenting-linux-kernel-tracing]]
5263 ==== Load and unload a custom probe kernel module
5265 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5266 kernel module>> in the kernel before it can emit LTTng events.
5268 To load the default probe kernel modules and a custom probe kernel
5271 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5272 probe modules to load when starting a root <<lttng-sessiond,session
5276 .Load the `my_subsys`, `usb`, and the default probe modules.
5280 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5285 You only need to pass the subsystem name, not the whole kernel module
5288 To load _only_ a given custom probe kernel module:
5290 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5291 modules to load when starting a root session daemon:
5294 .Load only the `my_subsys` and `usb` probe modules.
5298 # lttng-sessiond --kmod-probes=my_subsys,usb
5303 To confirm that a probe module is loaded:
5310 $ lsmod | grep lttng_probe_usb
5314 To unload the loaded probe modules:
5316 * Kill the session daemon with `SIGTERM`:
5321 # pkill lttng-sessiond
5325 You can also use man:modprobe(8)'s `--remove` option if the session
5326 daemon terminates abnormally.
5329 [[controlling-tracing]]
5332 Once an application or a Linux kernel is
5333 <<instrumenting,instrumented>> for LTTng tracing,
5336 This section is divided in topics on how to use the various
5337 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5338 command-line tool>>, to _control_ the LTTng daemons and tracers.
5340 NOTE: In the following subsections, we refer to an man:lttng(1) command
5341 using its man page name. For example, instead of _Run the `create`
5342 command to..._, we use _Run the man:lttng-create(1) command to..._.
5346 === Start a session daemon
5348 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5349 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5352 You will see the following error when you run a command while no session
5356 Error: No session daemon is available
5359 The only command that automatically runs a session daemon is
5360 man:lttng-create(1), which you use to
5361 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5362 this is most of the time the first operation that you do, sometimes it's
5363 not. Some examples are:
5365 * <<list-instrumentation-points,List the available instrumentation points>>.
5366 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5368 [[tracing-group]] Each Unix user must have its own running session
5369 daemon to trace user applications. The session daemon that the root user
5370 starts is the only one allowed to control the LTTng kernel tracer. Users
5371 that are part of the _tracing group_ can control the root session
5372 daemon. The default tracing group name is `tracing`; you can set it to
5373 something else with the opt:lttng-sessiond(8):--group option when you
5374 start the root session daemon.
5376 To start a user session daemon:
5378 * Run man:lttng-sessiond(8):
5383 $ lttng-sessiond --daemonize
5387 To start the root session daemon:
5389 * Run man:lttng-sessiond(8) as the root user:
5394 # lttng-sessiond --daemonize
5398 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5399 start the session daemon in foreground.
5401 To stop a session daemon, use man:kill(1) on its process ID (standard
5404 Note that some Linux distributions could manage the LTTng session daemon
5405 as a service. In this case, you should use the service manager to
5406 start, restart, and stop session daemons.
5409 [[creating-destroying-tracing-sessions]]
5410 === Create and destroy a tracing session
5412 Almost all the LTTng control operations happen in the scope of
5413 a <<tracing-session,tracing session>>, which is the dialogue between the
5414 <<lttng-sessiond,session daemon>> and you.
5416 To create a tracing session with a generated name:
5418 * Use the man:lttng-create(1) command:
5427 The created tracing session's name is `auto` followed by the
5430 To create a tracing session with a specific name:
5432 * Use the optional argument of the man:lttng-create(1) command:
5437 $ lttng create my-session
5441 Replace `my-session` with the specific tracing session name.
5443 LTTng appends the creation date to the created tracing session's name.
5445 LTTng writes the traces of a tracing session in
5446 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5447 name of the tracing session. Note that the env:LTTNG_HOME environment
5448 variable defaults to `$HOME` if not set.
5450 To output LTTng traces to a non-default location:
5452 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5457 $ lttng create my-session --output=/tmp/some-directory
5461 You may create as many tracing sessions as you wish.
5463 To list all the existing tracing sessions for your Unix user:
5465 * Use the man:lttng-list(1) command:
5474 When you create a tracing session, it is set as the _current tracing
5475 session_. The following man:lttng(1) commands operate on the current
5476 tracing session when you don't specify one:
5478 [role="list-3-cols"]
5495 To change the current tracing session:
5497 * Use the man:lttng-set-session(1) command:
5502 $ lttng set-session new-session
5506 Replace `new-session` by the name of the new current tracing session.
5508 When you are done tracing in a given tracing session, you can destroy
5509 it. This operation frees the resources taken by the tracing session
5510 to destroy; it does not destroy the trace data that LTTng wrote for
5511 this tracing session.
5513 To destroy the current tracing session:
5515 * Use the man:lttng-destroy(1) command:
5525 [[list-instrumentation-points]]
5526 === List the available instrumentation points
5528 The <<lttng-sessiond,session daemon>> can query the running instrumented
5529 user applications and the Linux kernel to get a list of available
5530 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5531 they are tracepoints and system calls. For the user space tracing
5532 domain, they are tracepoints. For the other tracing domains, they are
5535 To list the available instrumentation points:
5537 * Use the man:lttng-list(1) command with the requested tracing domain's
5541 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5542 must be a root user, or it must be a member of the
5543 <<tracing-group,tracing group>>).
5544 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5545 kernel system calls (your Unix user must be a root user, or it must be
5546 a member of the tracing group).
5547 * opt:lttng-list(1):--userspace: user space tracepoints.
5548 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5549 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5550 * opt:lttng-list(1):--python: Python loggers.
5553 .List the available user space tracepoints.
5557 $ lttng list --userspace
5561 .List the available Linux kernel system call tracepoints.
5565 $ lttng list --kernel --syscall
5570 [[enabling-disabling-events]]
5571 === Create and enable an event rule
5573 Once you <<creating-destroying-tracing-sessions,create a tracing
5574 session>>, you can create <<event,event rules>> with the
5575 man:lttng-enable-event(1) command.
5577 You specify each condition with a command-line option. The available
5578 condition options are shown in the following table.
5580 [role="growable",cols="asciidoc,asciidoc,default"]
5581 .Condition command-line options for the man:lttng-enable-event(1) command.
5583 |Option |Description |Applicable tracing domains
5589 . +--probe=__ADDR__+
5590 . +--function=__ADDR__+
5593 Instead of using the default _tracepoint_ instrumentation type, use:
5595 . A Linux system call.
5596 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5597 . The entry and return points of a Linux function (symbol or address).
5601 |First positional argument.
5604 Tracepoint or system call name. In the case of a Linux KProbe or
5605 function, this is a custom name given to the event rule. With the
5606 JUL, log4j, and Python domains, this is a logger name.
5608 With a tracepoint, logger, or system call name, the last character
5609 can be `*` to match anything that remains.
5616 . +--loglevel=__LEVEL__+
5617 . +--loglevel-only=__LEVEL__+
5620 . Match only tracepoints or log statements with a logging level at
5621 least as severe as +__LEVEL__+.
5622 . Match only tracepoints or log statements with a logging level
5623 equal to +__LEVEL__+.
5625 See man:lttng-enable-event(1) for the list of available logging level
5628 |User space, JUL, log4j, and Python.
5630 |+--exclude=__EXCLUSIONS__+
5633 When you use a `*` character at the end of the tracepoint or logger
5634 name (first positional argument), exclude the specific names in the
5635 comma-delimited list +__EXCLUSIONS__+.
5638 User space, JUL, log4j, and Python.
5640 |+--filter=__EXPR__+
5643 Match only events which satisfy the expression +__EXPR__+.
5645 See man:lttng-enable-event(1) to learn more about the syntax of a
5652 You attach an event rule to a <<channel,channel>> on creation. If you do
5653 not specify the channel with the opt:lttng-enable-event(1):--channel
5654 option, and if the event rule to create is the first in its
5655 <<domain,tracing domain>> for a given tracing session, then LTTng
5656 creates a _default channel_ for you. This default channel is reused in
5657 subsequent invocations of the man:lttng-enable-event(1) command for the
5658 same tracing domain.
5660 An event rule is always enabled at creation time.
5662 The following examples show how you can combine the previous
5663 command-line options to create simple to more complex event rules.
5665 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5669 $ lttng enable-event --kernel sched_switch
5673 .Create an event rule matching four Linux kernel system calls (default channel).
5677 $ lttng enable-event --kernel --syscall open,write,read,close
5681 .Create event rules matching tracepoints with filter expressions (default channel).
5685 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5690 $ lttng enable-event --kernel --all \
5691 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5696 $ lttng enable-event --jul my_logger \
5697 --filter='$app.retriever:cur_msg_id > 3'
5700 IMPORTANT: Make sure to always quote the filter string when you
5701 use man:lttng(1) from a shell.
5704 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5708 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5711 IMPORTANT: Make sure to always quote the wildcard character when you
5712 use man:lttng(1) from a shell.
5715 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5719 $ lttng enable-event --python my-app.'*' \
5720 --exclude='my-app.module,my-app.hello'
5724 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5728 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5732 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5736 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5740 The event rules of a given channel form a whitelist: as soon as an
5741 emitted event passes one of them, LTTng can record the event. For
5742 example, an event named `my_app:my_tracepoint` emitted from a user space
5743 tracepoint with a `TRACE_ERROR` log level passes both of the following
5748 $ lttng enable-event --userspace my_app:my_tracepoint
5749 $ lttng enable-event --userspace my_app:my_tracepoint \
5750 --loglevel=TRACE_INFO
5753 The second event rule is redundant: the first one includes
5757 [[disable-event-rule]]
5758 === Disable an event rule
5760 To disable an event rule that you <<enabling-disabling-events,created>>
5761 previously, use the man:lttng-disable-event(1) command. This command
5762 disables _all_ the event rules (of a given tracing domain and channel)
5763 which match an instrumentation point. The other conditions are not
5764 supported as of LTTng{nbsp}{revision}.
5766 The LTTng tracer does not record an emitted event which passes
5767 a _disabled_ event rule.
5769 .Disable an event rule matching a Python logger (default channel).
5773 $ lttng disable-event --python my-logger
5777 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5781 $ lttng disable-event --jul '*'
5785 .Disable _all_ the event rules of the default channel.
5787 The opt:lttng-disable-event(1):--all-events option is not, like the
5788 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5789 equivalent of the event name `*` (wildcard): it disables _all_ the event
5790 rules of a given channel.
5794 $ lttng disable-event --jul --all-events
5798 NOTE: You cannot delete an event rule once you create it.
5802 === Get the status of a tracing session
5804 To get the status of the current tracing session, that is, its
5805 parameters, its channels, event rules, and their attributes:
5807 * Use the man:lttng-status(1) command:
5817 To get the status of any tracing session:
5819 * Use the man:lttng-list(1) command with the tracing session's name:
5824 $ lttng list my-session
5828 Replace `my-session` with the desired tracing session's name.
5831 [[basic-tracing-session-control]]
5832 === Start and stop a tracing session
5834 Once you <<creating-destroying-tracing-sessions,create a tracing
5836 <<enabling-disabling-events,create one or more event rules>>,
5837 you can start and stop the tracers for this tracing session.
5839 To start tracing in the current tracing session:
5841 * Use the man:lttng-start(1) command:
5850 LTTng is very flexible: you can launch user applications before
5851 or after the you start the tracers. The tracers only record the events
5852 if they pass enabled event rules and if they occur while the tracers are
5855 To stop tracing in the current tracing session:
5857 * Use the man:lttng-stop(1) command:
5866 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5867 records>> or lost sub-buffers since the last time you ran
5868 man:lttng-start(1), warnings are printed when you run the
5869 man:lttng-stop(1) command.
5872 [[enabling-disabling-channels]]
5873 === Create a channel
5875 Once you create a tracing session, you can create a <<channel,channel>>
5876 with the man:lttng-enable-channel(1) command.
5878 Note that LTTng automatically creates a default channel when, for a
5879 given <<domain,tracing domain>>, no channels exist and you
5880 <<enabling-disabling-events,create>> the first event rule. This default
5881 channel is named `channel0` and its attributes are set to reasonable
5882 values. Therefore, you only need to create a channel when you need
5883 non-default attributes.
5885 You specify each non-default channel attribute with a command-line
5886 option when you use the man:lttng-enable-channel(1) command. The
5887 available command-line options are:
5889 [role="growable",cols="asciidoc,asciidoc"]
5890 .Command-line options for the man:lttng-enable-channel(1) command.
5892 |Option |Description
5898 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5899 the default _discard_ mode.
5901 |`--buffers-pid` (user space tracing domain only)
5904 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5905 instead of the default per-user buffering scheme.
5907 |+--subbuf-size=__SIZE__+
5910 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5911 either for each Unix user (default), or for each instrumented process.
5913 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5915 |+--num-subbuf=__COUNT__+
5918 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5919 for each Unix user (default), or for each instrumented process.
5921 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5923 |+--tracefile-size=__SIZE__+
5926 Set the maximum size of each trace file that this channel writes within
5927 a stream to +__SIZE__+ bytes instead of no maximum.
5929 See <<tracefile-rotation,Trace file count and size>>.
5931 |+--tracefile-count=__COUNT__+
5934 Limit the number of trace files that this channel creates to
5935 +__COUNT__+ channels instead of no limit.
5937 See <<tracefile-rotation,Trace file count and size>>.
5939 |+--switch-timer=__PERIODUS__+
5942 Set the <<channel-switch-timer,switch timer period>>
5943 to +__PERIODUS__+{nbsp}µs.
5945 |+--read-timer=__PERIODUS__+
5948 Set the <<channel-read-timer,read timer period>>
5949 to +__PERIODUS__+{nbsp}µs.
5951 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
5954 Set the timeout of user space applications which load LTTng-UST
5955 in blocking mode to +__TIMEOUTUS__+:
5958 Never block (non-blocking mode).
5961 Block forever until space is available in a sub-buffer to record
5964 __n__, a positive value::
5965 Wait for at most __n__ µs when trying to write into a sub-buffer.
5967 Note that, for this option to have any effect on an instrumented
5968 user space application, you need to run the application with a set
5969 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
5971 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5974 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5978 You can only create a channel in the Linux kernel and user space
5979 <<domain,tracing domains>>: other tracing domains have their own channel
5980 created on the fly when <<enabling-disabling-events,creating event
5985 Because of a current LTTng limitation, you must create all channels
5986 _before_ you <<basic-tracing-session-control,start tracing>> in a given
5987 tracing session, that is, before the first time you run
5990 Since LTTng automatically creates a default channel when you use the
5991 man:lttng-enable-event(1) command with a specific tracing domain, you
5992 cannot, for example, create a Linux kernel event rule, start tracing,
5993 and then create a user space event rule, because no user space channel
5994 exists yet and it's too late to create one.
5996 For this reason, make sure to configure your channels properly
5997 before starting the tracers for the first time!
6000 The following examples show how you can combine the previous
6001 command-line options to create simple to more complex channels.
6003 .Create a Linux kernel channel with default attributes.
6007 $ lttng enable-channel --kernel my-channel
6011 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6015 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6016 --buffers-pid my-channel
6020 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6022 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6023 create the channel, <<enabling-disabling-events,create an event rule>>,
6024 and <<basic-tracing-session-control,start tracing>>:
6029 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6030 $ lttng enable-event --userspace --channel=blocking-channel --all
6034 Run an application instrumented with LTTng-UST and allow it to block:
6038 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6042 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6046 $ lttng enable-channel --kernel --tracefile-count=8 \
6047 --tracefile-size=4194304 my-channel
6051 .Create a user space channel in overwrite (or _flight recorder_) mode.
6055 $ lttng enable-channel --userspace --overwrite my-channel
6059 You can <<enabling-disabling-events,create>> the same event rule in
6060 two different channels:
6064 $ lttng enable-event --userspace --channel=my-channel app:tp
6065 $ lttng enable-event --userspace --channel=other-channel app:tp
6068 If both channels are enabled, when a tracepoint named `app:tp` is
6069 reached, LTTng records two events, one for each channel.
6073 === Disable a channel
6075 To disable a specific channel that you <<enabling-disabling-channels,created>>
6076 previously, use the man:lttng-disable-channel(1) command.
6078 .Disable a specific Linux kernel channel.
6082 $ lttng disable-channel --kernel my-channel
6086 The state of a channel precedes the individual states of event rules
6087 attached to it: event rules which belong to a disabled channel, even if
6088 they are enabled, are also considered disabled.
6092 === Add context fields to a channel
6094 Event record fields in trace files provide important information about
6095 events that occured previously, but sometimes some external context may
6096 help you solve a problem faster. Examples of context fields are:
6098 * The **process ID**, **thread ID**, **process name**, and
6099 **process priority** of the thread in which the event occurs.
6100 * The **hostname** of the system on which the event occurs.
6101 * The current values of many possible **performance counters** using
6103 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6105 ** Branch instructions, misses, and loads.
6107 * Any context defined at the application level (supported for the
6108 JUL and log4j <<domain,tracing domains>>).
6110 To get the full list of available context fields, see
6111 `lttng add-context --list`. Some context fields are reserved for a
6112 specific <<domain,tracing domain>> (Linux kernel or user space).
6114 You add context fields to <<channel,channels>>. All the events
6115 that a channel with added context fields records contain those fields.
6117 To add context fields to one or all the channels of a given tracing
6120 * Use the man:lttng-add-context(1) command.
6122 .Add context fields to all the channels of the current tracing session.
6124 The following command line adds the virtual process identifier and
6125 the per-thread CPU cycles count fields to all the user space channels
6126 of the current tracing session.
6130 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6134 .Add performance counter context fields by raw ID
6136 See man:lttng-add-context(1) for the exact format of the context field
6137 type, which is partly compatible with the format used in
6142 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6143 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6147 .Add a context field to a specific channel.
6149 The following command line adds the thread identifier context field
6150 to the Linux kernel channel named `my-channel` in the current
6155 $ lttng add-context --kernel --channel=my-channel --type=tid
6159 .Add an application-specific context field to a specific channel.
6161 The following command line adds the `cur_msg_id` context field of the
6162 `retriever` context retriever for all the instrumented
6163 <<java-application,Java applications>> recording <<event,event records>>
6164 in the channel named `my-channel`:
6168 $ lttng add-context --kernel --channel=my-channel \
6169 --type='$app:retriever:cur_msg_id'
6172 IMPORTANT: Make sure to always quote the `$` character when you
6173 use man:lttng-add-context(1) from a shell.
6176 NOTE: You cannot remove context fields from a channel once you add it.
6181 === Track process IDs
6183 It's often useful to allow only specific process IDs (PIDs) to emit
6184 events. For example, you may wish to record all the system calls made by
6185 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6187 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6188 purpose. Both commands operate on a whitelist of process IDs. You _add_
6189 entries to this whitelist with the man:lttng-track(1) command and remove
6190 entries with the man:lttng-untrack(1) command. Any process which has one
6191 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6192 an enabled <<event,event rule>>.
6194 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6195 process with a given tracked ID exit and another process be given this
6196 ID, then the latter would also be allowed to emit events.
6198 .Track and untrack process IDs.
6200 For the sake of the following example, assume the target system has 16
6204 <<creating-destroying-tracing-sessions,create a tracing session>>,
6205 the whitelist contains all the possible PIDs:
6208 .All PIDs are tracked.
6209 image::track-all.png[]
6211 When the whitelist is full and you use the man:lttng-track(1) command to
6212 specify some PIDs to track, LTTng first clears the whitelist, then it
6213 tracks the specific PIDs. After:
6217 $ lttng track --pid=3,4,7,10,13
6223 .PIDs 3, 4, 7, 10, and 13 are tracked.
6224 image::track-3-4-7-10-13.png[]
6226 You can add more PIDs to the whitelist afterwards:
6230 $ lttng track --pid=1,15,16
6236 .PIDs 1, 15, and 16 are added to the whitelist.
6237 image::track-1-3-4-7-10-13-15-16.png[]
6239 The man:lttng-untrack(1) command removes entries from the PID tracker's
6240 whitelist. Given the previous example, the following command:
6244 $ lttng untrack --pid=3,7,10,13
6247 leads to this whitelist:
6250 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6251 image::track-1-4-15-16.png[]
6253 LTTng can track all possible PIDs again using the opt:track(1):--all
6258 $ lttng track --pid --all
6261 The result is, again:
6264 .All PIDs are tracked.
6265 image::track-all.png[]
6268 .Track only specific PIDs
6270 A very typical use case with PID tracking is to start with an empty
6271 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6272 then add PIDs manually while tracers are active. You can accomplish this
6273 by using the opt:lttng-untrack(1):--all option of the
6274 man:lttng-untrack(1) command to clear the whitelist after you
6275 <<creating-destroying-tracing-sessions,create a tracing session>>:
6279 $ lttng untrack --pid --all
6285 .No PIDs are tracked.
6286 image::untrack-all.png[]
6288 If you trace with this whitelist configuration, the tracer records no
6289 events for this <<domain,tracing domain>> because no processes are
6290 tracked. You can use the man:lttng-track(1) command as usual to track
6291 specific PIDs, for example:
6295 $ lttng track --pid=6,11
6301 .PIDs 6 and 11 are tracked.
6302 image::track-6-11.png[]
6307 [[saving-loading-tracing-session]]
6308 === Save and load tracing session configurations
6310 Configuring a <<tracing-session,tracing session>> can be long. Some of
6311 the tasks involved are:
6313 * <<enabling-disabling-channels,Create channels>> with
6314 specific attributes.
6315 * <<adding-context,Add context fields>> to specific channels.
6316 * <<enabling-disabling-events,Create event rules>> with specific log
6317 level and filter conditions.
6319 If you use LTTng to solve real world problems, chances are you have to
6320 record events using the same tracing session setup over and over,
6321 modifying a few variables each time in your instrumented program
6322 or environment. To avoid constant tracing session reconfiguration,
6323 the man:lttng(1) command-line tool can save and load tracing session
6324 configurations to/from XML files.
6326 To save a given tracing session configuration:
6328 * Use the man:lttng-save(1) command:
6333 $ lttng save my-session
6337 Replace `my-session` with the name of the tracing session to save.
6339 LTTng saves tracing session configurations to
6340 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6341 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6342 the opt:lttng-save(1):--output-path option to change this destination
6345 LTTng saves all configuration parameters, for example:
6347 * The tracing session name.
6348 * The trace data output path.
6349 * The channels with their state and all their attributes.
6350 * The context fields you added to channels.
6351 * The event rules with their state, log level and filter conditions.
6353 To load a tracing session:
6355 * Use the man:lttng-load(1) command:
6360 $ lttng load my-session
6364 Replace `my-session` with the name of the tracing session to load.
6366 When LTTng loads a configuration, it restores your saved tracing session
6367 as if you just configured it manually.
6369 See man:lttng(1) for the complete list of command-line options. You
6370 can also save and load all many sessions at a time, and decide in which
6371 directory to output the XML files.
6374 [[sending-trace-data-over-the-network]]
6375 === Send trace data over the network
6377 LTTng can send the recorded trace data to a remote system over the
6378 network instead of writing it to the local file system.
6380 To send the trace data over the network:
6382 . On the _remote_ system (which can also be the target system),
6383 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6392 . On the _target_ system, create a tracing session configured to
6393 send trace data over the network:
6398 $ lttng create my-session --set-url=net://remote-system
6402 Replace `remote-system` by the host name or IP address of the
6403 remote system. See man:lttng-create(1) for the exact URL format.
6405 . On the target system, use the man:lttng(1) command-line tool as usual.
6406 When tracing is active, the target's consumer daemon sends sub-buffers
6407 to the relay daemon running on the remote system instead of flushing
6408 them to the local file system. The relay daemon writes the received
6409 packets to the local file system.
6411 The relay daemon writes trace files to
6412 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6413 +__hostname__+ is the host name of the target system and +__session__+
6414 is the tracing session name. Note that the env:LTTNG_HOME environment
6415 variable defaults to `$HOME` if not set. Use the
6416 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6417 trace files to another base directory.
6422 === View events as LTTng emits them (noch:{LTTng} live)
6424 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6425 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6426 display events as LTTng emits them on the target system while tracing is
6429 The relay daemon creates a _tee_: it forwards the trace data to both
6430 the local file system and to connected live viewers:
6433 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6438 . On the _target system_, create a <<tracing-session,tracing session>>
6444 $ lttng create my-session --live
6448 This spawns a local relay daemon.
6450 . Start the live viewer and configure it to connect to the relay
6451 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6456 $ babeltrace --input-format=lttng-live \
6457 net://localhost/host/hostname/my-session
6464 * `hostname` with the host name of the target system.
6465 * `my-session` with the name of the tracing session to view.
6468 . Configure the tracing session as usual with the man:lttng(1)
6469 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6471 You can list the available live tracing sessions with Babeltrace:
6475 $ babeltrace --input-format=lttng-live net://localhost
6478 You can start the relay daemon on another system. In this case, you need
6479 to specify the relay daemon's URL when you create the tracing session
6480 with the opt:lttng-create(1):--set-url option. You also need to replace
6481 `localhost` in the procedure above with the host name of the system on
6482 which the relay daemon is running.
6484 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6485 command-line options.
6489 [[taking-a-snapshot]]
6490 === Take a snapshot of the current sub-buffers of a tracing session
6492 The normal behavior of LTTng is to append full sub-buffers to growing
6493 trace data files. This is ideal to keep a full history of the events
6494 that occurred on the target system, but it can
6495 represent too much data in some situations. For example, you may wish
6496 to trace your application continuously until some critical situation
6497 happens, in which case you only need the latest few recorded
6498 events to perform the desired analysis, not multi-gigabyte trace files.
6500 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6501 current sub-buffers of a given <<tracing-session,tracing session>>.
6502 LTTng can write the snapshot to the local file system or send it over
6507 . Create a tracing session in _snapshot mode_:
6512 $ lttng create my-session --snapshot
6516 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6517 <<channel,channels>> created in this mode is automatically set to
6518 _overwrite_ (flight recorder mode).
6520 . Configure the tracing session as usual with the man:lttng(1)
6521 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6523 . **Optional**: When you need to take a snapshot,
6524 <<basic-tracing-session-control,stop tracing>>.
6526 You can take a snapshot when the tracers are active, but if you stop
6527 them first, you are sure that the data in the sub-buffers does not
6528 change before you actually take the snapshot.
6535 $ lttng snapshot record --name=my-first-snapshot
6539 LTTng writes the current sub-buffers of all the current tracing
6540 session's channels to trace files on the local file system. Those trace
6541 files have `my-first-snapshot` in their name.
6543 There is no difference between the format of a normal trace file and the
6544 format of a snapshot: viewers of LTTng traces also support LTTng
6547 By default, LTTng writes snapshot files to the path shown by
6548 `lttng snapshot list-output`. You can change this path or decide to send
6549 snapshots over the network using either:
6551 . An output path or URL that you specify when you create the
6553 . An snapshot output path or URL that you add using
6554 `lttng snapshot add-output`
6555 . An output path or URL that you provide directly to the
6556 `lttng snapshot record` command.
6558 Method 3 overrides method 2, which overrides method 1. When you
6559 specify a URL, a relay daemon must listen on a remote system (see
6560 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6565 === Use the machine interface
6567 With any command of the man:lttng(1) command-line tool, you can set the
6568 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6569 XML machine interface output, for example:
6573 $ lttng --mi=xml enable-event --kernel --syscall open
6576 A schema definition (XSD) is
6577 https://github.com/lttng/lttng-tools/blob/stable-2.10/src/common/mi-lttng-3.0.xsd[available]
6578 to ease the integration with external tools as much as possible.
6582 [[metadata-regenerate]]
6583 === Regenerate the metadata of an LTTng trace
6585 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6586 data stream files and a metadata file. This metadata file contains,
6587 amongst other things, information about the offset of the clock sources
6588 used to timestamp <<event,event records>> when tracing.
6590 If, once a <<tracing-session,tracing session>> is
6591 <<basic-tracing-session-control,started>>, a major
6592 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6593 happens, the trace's clock offset also needs to be updated. You
6594 can use the `metadata` item of the man:lttng-regenerate(1) command
6597 The main use case of this command is to allow a system to boot with
6598 an incorrect wall time and trace it with LTTng before its wall time
6599 is corrected. Once the system is known to be in a state where its
6600 wall time is correct, it can run `lttng regenerate metadata`.
6602 To regenerate the metadata of an LTTng trace:
6604 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6609 $ lttng regenerate metadata
6615 `lttng regenerate metadata` has the following limitations:
6617 * Tracing session <<creating-destroying-tracing-sessions,created>>
6619 * User space <<channel,channels>>, if any, are using
6620 <<channel-buffering-schemes,per-user buffering>>.
6625 [[regenerate-statedump]]
6626 === Regenerate the state dump of a tracing session
6628 The LTTng kernel and user space tracers generate state dump
6629 <<event,event records>> when the application starts or when you
6630 <<basic-tracing-session-control,start a tracing session>>. An analysis
6631 can use the state dump event records to set an initial state before it
6632 builds the rest of the state from the following event records.
6633 http://tracecompass.org/[Trace Compass] is a notable example of an
6634 application which uses the state dump of an LTTng trace.
6636 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6637 state dump event records are not included in the snapshot because they
6638 were recorded to a sub-buffer that has been consumed or overwritten
6641 You can use the `lttng regenerate statedump` command to emit the state
6642 dump event records again.
6644 To regenerate the state dump of the current tracing session, provided
6645 create it in snapshot mode, before you take a snapshot:
6647 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6652 $ lttng regenerate statedump
6656 . <<basic-tracing-session-control,Stop the tracing session>>:
6665 . <<taking-a-snapshot,Take a snapshot>>:
6670 $ lttng snapshot record --name=my-snapshot
6674 Depending on the event throughput, you should run steps 1 and 2
6675 as closely as possible.
6677 NOTE: To record the state dump events, you need to
6678 <<enabling-disabling-events,create event rules>> which enable them.
6679 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6680 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6684 [[persistent-memory-file-systems]]
6685 === Record trace data on persistent memory file systems
6687 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6688 (NVRAM) is random-access memory that retains its information when power
6689 is turned off (non-volatile). Systems with such memory can store data
6690 structures in RAM and retrieve them after a reboot, without flushing
6691 to typical _storage_.
6693 Linux supports NVRAM file systems thanks to either
6694 http://pramfs.sourceforge.net/[PRAMFS] or
6695 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6696 (requires Linux 4.1+).
6698 This section does not describe how to operate such file systems;
6699 we assume that you have a working persistent memory file system.
6701 When you create a <<tracing-session,tracing session>>, you can specify
6702 the path of the shared memory holding the sub-buffers. If you specify a
6703 location on an NVRAM file system, then you can retrieve the latest
6704 recorded trace data when the system reboots after a crash.
6706 To record trace data on a persistent memory file system and retrieve the
6707 trace data after a system crash:
6709 . Create a tracing session with a sub-buffer shared memory path located
6710 on an NVRAM file system:
6715 $ lttng create my-session --shm-path=/path/to/shm
6719 . Configure the tracing session as usual with the man:lttng(1)
6720 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6722 . After a system crash, use the man:lttng-crash(1) command-line tool to
6723 view the trace data recorded on the NVRAM file system:
6728 $ lttng-crash /path/to/shm
6732 The binary layout of the ring buffer files is not exactly the same as
6733 the trace files layout. This is why you need to use man:lttng-crash(1)
6734 instead of your preferred trace viewer directly.
6736 To convert the ring buffer files to LTTng trace files:
6738 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6743 $ lttng-crash --extract=/path/to/trace /path/to/shm
6749 [[notif-trigger-api]]
6750 === Get notified when a channel's buffer usage is too high or too low
6752 With LTTng's $$C/C++$$ notification and trigger API, your user
6753 application can get notified when the buffer usage of one or more
6754 <<channel,channels>> becomes too low or too high. You can use this API
6755 and enable or disable <<event,event rules>> during tracing to avoid
6756 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
6758 .Have a user application get notified when an LTTng channel's buffer usage is too high.
6760 In this example, we create and build an application which gets notified
6761 when the buffer usage of a specific LTTng channel is higher than
6762 75{nbsp}%. We only print that it is the case in the example, but we
6763 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
6764 disable event rules when this happens.
6766 . Create the application's C source file:
6774 #include <lttng/domain.h>
6775 #include <lttng/action/action.h>
6776 #include <lttng/action/notify.h>
6777 #include <lttng/condition/condition.h>
6778 #include <lttng/condition/buffer-usage.h>
6779 #include <lttng/condition/evaluation.h>
6780 #include <lttng/notification/channel.h>
6781 #include <lttng/notification/notification.h>
6782 #include <lttng/trigger/trigger.h>
6783 #include <lttng/endpoint.h>
6785 int main(int argc, char *argv[])
6787 int exit_status = 0;
6788 struct lttng_notification_channel *notification_channel;
6789 struct lttng_condition *condition;
6790 struct lttng_action *action;
6791 struct lttng_trigger *trigger;
6792 const char *tracing_session_name;
6793 const char *channel_name;
6796 tracing_session_name = argv[1];
6797 channel_name = argv[2];
6800 * Create a notification channel. A notification channel
6801 * connects the user application to the LTTng session daemon.
6802 * This notification channel can be used to listen to various
6803 * types of notifications.
6805 notification_channel = lttng_notification_channel_create(
6806 lttng_session_daemon_notification_endpoint);
6809 * Create a "high buffer usage" condition. In this case, the
6810 * condition is reached when the buffer usage is greater than or
6811 * equal to 75 %. We create the condition for a specific tracing
6812 * session name, channel name, and for the user space tracing
6815 * The "low buffer usage" condition type also exists.
6817 condition = lttng_condition_buffer_usage_high_create();
6818 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
6819 lttng_condition_buffer_usage_set_session_name(
6820 condition, tracing_session_name);
6821 lttng_condition_buffer_usage_set_channel_name(condition,
6823 lttng_condition_buffer_usage_set_domain_type(condition,
6827 * Create an action (get a notification) to take when the
6828 * condition created above is reached.
6830 action = lttng_action_notify_create();
6833 * Create a trigger. A trigger associates a condition to an
6834 * action: the action is executed when the condition is reached.
6836 trigger = lttng_trigger_create(condition, action);
6838 /* Register the trigger to LTTng. */
6839 lttng_register_trigger(trigger);
6842 * Now that we have registered a trigger, a notification will be
6843 * emitted everytime its condition is met. To receive this
6844 * notification, we must subscribe to notifications that match
6845 * the same condition.
6847 lttng_notification_channel_subscribe(notification_channel,
6851 * Notification loop. You can put this in a dedicated thread to
6852 * avoid blocking the main thread.
6855 struct lttng_notification *notification;
6856 enum lttng_notification_channel_status status;
6857 const struct lttng_evaluation *notification_evaluation;
6858 const struct lttng_condition *notification_condition;
6859 double buffer_usage;
6861 /* Receive the next notification. */
6862 status = lttng_notification_channel_get_next_notification(
6863 notification_channel, ¬ification);
6866 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
6868 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
6870 * The session daemon can drop notifications if
6871 * a monitoring application is not consuming the
6872 * notifications fast enough.
6875 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
6877 * The notification channel has been closed by the
6878 * session daemon. This is typically caused by a session
6879 * daemon shutting down.
6883 /* Unhandled conditions or errors. */
6889 * A notification provides, amongst other things:
6891 * * The condition that caused this notification to be
6893 * * The condition evaluation, which provides more
6894 * specific information on the evaluation of the
6897 * The condition evaluation provides the buffer usage
6898 * value at the moment the condition was reached.
6900 notification_condition = lttng_notification_get_condition(
6902 notification_evaluation = lttng_notification_get_evaluation(
6905 /* We're subscribed to only one condition. */
6906 assert(lttng_condition_get_type(notification_condition) ==
6907 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
6910 * Get the exact sampled buffer usage from the
6911 * condition evaluation.
6913 lttng_evaluation_buffer_usage_get_usage_ratio(
6914 notification_evaluation, &buffer_usage);
6917 * At this point, instead of printing a message, we
6918 * could do something to reduce the channel's buffer
6919 * usage, like disable specific events.
6921 printf("Buffer usage is %f %% in tracing session \"%s\", "
6922 "user space channel \"%s\".\n", buffer_usage * 100,
6923 tracing_session_name, channel_name);
6924 lttng_notification_destroy(notification);
6928 lttng_action_destroy(action);
6929 lttng_condition_destroy(condition);
6930 lttng_trigger_destroy(trigger);
6931 lttng_notification_channel_destroy(notification_channel);
6937 . Build the `notif-app` application, linking it to `liblttng-ctl`:
6942 $ gcc -o notif-app notif-app.c -llttng-ctl
6946 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
6947 <<enabling-disabling-events,create an event rule>> matching all the
6948 user space tracepoints, and
6949 <<basic-tracing-session-control,start tracing>>:
6954 $ lttng create my-session
6955 $ lttng enable-event --userspace --all
6960 If you create the channel manually with the man:lttng-enable-channel(1)
6961 command, you can control how frequently are the current values of the
6962 channel's properties sampled to evaluate user conditions with the
6963 opt:lttng-enable-channel(1):--monitor-timer option.
6965 . Run the `notif-app` application. This program accepts the
6966 <<tracing-session,tracing session>> name and the user space channel
6967 name as its two first arguments. The channel which LTTng automatically
6968 creates with the man:lttng-enable-event(1) command above is named
6974 $ ./notif-app my-session channel0
6978 . In another terminal, run an application with a very high event
6979 throughput so that the 75{nbsp}% buffer usage condition is reached.
6981 In the first terminal, the application should print lines like this:
6984 Buffer usage is 81.45197 % in tracing session "my-session", user space
6988 If you don't see anything, try modifying the condition in
6989 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
6990 (step 2) and running it again (step 4).
6997 [[lttng-modules-ref]]
6998 === noch:{LTTng-modules}
7002 [[lttng-tracepoint-enum]]
7003 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7005 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7009 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7014 * `name` with the name of the enumeration (C identifier, unique
7015 amongst all the defined enumerations).
7016 * `entries` with a list of enumeration entries.
7018 The available enumeration entry macros are:
7020 +ctf_enum_value(__name__, __value__)+::
7021 Entry named +__name__+ mapped to the integral value +__value__+.
7023 +ctf_enum_range(__name__, __begin__, __end__)+::
7024 Entry named +__name__+ mapped to the range of integral values between
7025 +__begin__+ (included) and +__end__+ (included).
7027 +ctf_enum_auto(__name__)+::
7028 Entry named +__name__+ mapped to the integral value following the
7029 last mapping's value.
7031 The last value of a `ctf_enum_value()` entry is its +__value__+
7034 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7036 If `ctf_enum_auto()` is the first entry in the list, its integral
7039 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7040 to use a defined enumeration as a tracepoint field.
7042 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7046 LTTNG_TRACEPOINT_ENUM(
7049 ctf_enum_auto("AUTO: EXPECT 0")
7050 ctf_enum_value("VALUE: 23", 23)
7051 ctf_enum_value("VALUE: 27", 27)
7052 ctf_enum_auto("AUTO: EXPECT 28")
7053 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7054 ctf_enum_auto("AUTO: EXPECT 304")
7062 [[lttng-modules-tp-fields]]
7063 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7065 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7066 tracepoint fields, which must be listed within `TP_FIELDS()` in
7067 `LTTNG_TRACEPOINT_EVENT()`, are:
7069 [role="func-desc growable",cols="asciidoc,asciidoc"]
7070 .Available macros to define LTTng-modules tracepoint fields
7072 |Macro |Description and parameters
7075 +ctf_integer(__t__, __n__, __e__)+
7077 +ctf_integer_nowrite(__t__, __n__, __e__)+
7079 +ctf_user_integer(__t__, __n__, __e__)+
7081 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7083 Standard integer, displayed in base 10.
7086 Integer C type (`int`, `long`, `size_t`, ...).
7092 Argument expression.
7095 +ctf_integer_hex(__t__, __n__, __e__)+
7097 +ctf_user_integer_hex(__t__, __n__, __e__)+
7099 Standard integer, displayed in base 16.
7108 Argument expression.
7110 |+ctf_integer_oct(__t__, __n__, __e__)+
7112 Standard integer, displayed in base 8.
7121 Argument expression.
7124 +ctf_integer_network(__t__, __n__, __e__)+
7126 +ctf_user_integer_network(__t__, __n__, __e__)+
7128 Integer in network byte order (big-endian), displayed in base 10.
7137 Argument expression.
7140 +ctf_integer_network_hex(__t__, __n__, __e__)+
7142 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7144 Integer in network byte order, displayed in base 16.
7153 Argument expression.
7156 +ctf_enum(__N__, __t__, __n__, __e__)+
7158 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7160 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7162 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7167 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7170 Integer C type (`int`, `long`, `size_t`, ...).
7176 Argument expression.
7179 +ctf_string(__n__, __e__)+
7181 +ctf_string_nowrite(__n__, __e__)+
7183 +ctf_user_string(__n__, __e__)+
7185 +ctf_user_string_nowrite(__n__, __e__)+
7187 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7193 Argument expression.
7196 +ctf_array(__t__, __n__, __e__, __s__)+
7198 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7200 +ctf_user_array(__t__, __n__, __e__, __s__)+
7202 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7204 Statically-sized array of integers.
7207 Array element C type.
7213 Argument expression.
7219 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7221 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7223 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7225 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7227 Statically-sized array of bits.
7229 The type of +__e__+ must be an integer type. +__s__+ is the number
7230 of elements of such type in +__e__+, not the number of bits.
7233 Array element C type.
7239 Argument expression.
7245 +ctf_array_text(__t__, __n__, __e__, __s__)+
7247 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7249 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7251 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7253 Statically-sized array, printed as text.
7255 The string does not need to be null-terminated.
7258 Array element C type (always `char`).
7264 Argument expression.
7270 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7272 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7274 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7276 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7278 Dynamically-sized array of integers.
7280 The type of +__E__+ must be unsigned.
7283 Array element C type.
7289 Argument expression.
7292 Length expression C type.
7298 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7300 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7302 Dynamically-sized array of integers, displayed in base 16.
7304 The type of +__E__+ must be unsigned.
7307 Array element C type.
7313 Argument expression.
7316 Length expression C type.
7321 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7323 Dynamically-sized array of integers in network byte order (big-endian),
7324 displayed in base 10.
7326 The type of +__E__+ must be unsigned.
7329 Array element C type.
7335 Argument expression.
7338 Length expression C type.
7344 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7346 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7348 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7350 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7352 Dynamically-sized array of bits.
7354 The type of +__e__+ must be an integer type. +__s__+ is the number
7355 of elements of such type in +__e__+, not the number of bits.
7357 The type of +__E__+ must be unsigned.
7360 Array element C type.
7366 Argument expression.
7369 Length expression C type.
7375 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7377 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7379 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7381 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7383 Dynamically-sized array, displayed as text.
7385 The string does not need to be null-terminated.
7387 The type of +__E__+ must be unsigned.
7389 The behaviour is undefined if +__e__+ is `NULL`.
7392 Sequence element C type (always `char`).
7398 Argument expression.
7401 Length expression C type.
7407 Use the `_user` versions when the argument expression, `e`, is
7408 a user space address. In the cases of `ctf_user_integer*()` and
7409 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7412 The `_nowrite` versions omit themselves from the session trace, but are
7413 otherwise identical. This means the `_nowrite` fields won't be written
7414 in the recorded trace. Their primary purpose is to make some
7415 of the event context available to the
7416 <<enabling-disabling-events,event filters>> without having to
7417 commit the data to sub-buffers.
7423 Terms related to LTTng and to tracing in general:
7426 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7427 the cmd:babeltrace command, some libraries, and Python bindings.
7429 <<channel-buffering-schemes,buffering scheme>>::
7430 A layout of sub-buffers applied to a given channel.
7432 <<channel,channel>>::
7433 An entity which is responsible for a set of ring buffers.
7435 <<event,Event rules>> are always attached to a specific channel.
7438 A reference of time for a tracer.
7440 <<lttng-consumerd,consumer daemon>>::
7441 A process which is responsible for consuming the full sub-buffers
7442 and write them to a file system or send them over the network.
7444 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7445 mode in which the tracer _discards_ new event records when there's no
7446 sub-buffer space left to store them.
7449 The consequence of the execution of an instrumentation
7450 point, like a tracepoint that you manually place in some source code,
7451 or a Linux kernel KProbe.
7453 An event is said to _occur_ at a specific time. Different actions can
7454 be taken upon the occurrence of an event, like record the event's payload
7457 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7458 The mechanism by which event records of a given channel are lost
7459 (not recorded) when there is no sub-buffer space left to store them.
7461 [[def-event-name]]event name::
7462 The name of an event, which is also the name of the event record.
7463 This is also called the _instrumentation point name_.
7466 A record, in a trace, of the payload of an event which occured.
7468 <<event,event rule>>::
7469 Set of conditions which must be satisfied for one or more occuring
7470 events to be recorded.
7472 `java.util.logging`::
7474 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7476 <<instrumenting,instrumentation>>::
7477 The use of LTTng probes to make a piece of software traceable.
7479 instrumentation point::
7480 A point in the execution path of a piece of software that, when
7481 reached by this execution, can emit an event.
7483 instrumentation point name::
7484 See _<<def-event-name,event name>>_.
7487 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7488 developed by the Apache Software Foundation.
7491 Level of severity of a log statement or user space
7492 instrumentation point.
7495 The _Linux Trace Toolkit: next generation_ project.
7497 <<lttng-cli,cmd:lttng>>::
7498 A command-line tool provided by the LTTng-tools project which you
7499 can use to send and receive control messages to and from a
7503 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7504 which is a set of analyzing programs that are used to obtain a
7505 higher level view of an LTTng trace.
7507 cmd:lttng-consumerd::
7508 The name of the consumer daemon program.
7511 A utility provided by the LTTng-tools project which can convert
7512 ring buffer files (usually
7513 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7516 LTTng Documentation::
7519 <<lttng-live,LTTng live>>::
7520 A communication protocol between the relay daemon and live viewers
7521 which makes it possible to see events "live", as they are received by
7524 <<lttng-modules,LTTng-modules>>::
7525 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7526 which contains the Linux kernel modules to make the Linux kernel
7527 instrumentation points available for LTTng tracing.
7530 The name of the relay daemon program.
7532 cmd:lttng-sessiond::
7533 The name of the session daemon program.
7536 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7537 contains the various programs and libraries used to
7538 <<controlling-tracing,control tracing>>.
7540 <<lttng-ust,LTTng-UST>>::
7541 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7542 contains libraries to instrument user applications.
7544 <<lttng-ust-agents,LTTng-UST Java agent>>::
7545 A Java package provided by the LTTng-UST project to allow the
7546 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7549 <<lttng-ust-agents,LTTng-UST Python agent>>::
7550 A Python package provided by the LTTng-UST project to allow the
7551 LTTng instrumentation of Python logging statements.
7553 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7554 The event loss mode in which new event records overwrite older
7555 event records when there's no sub-buffer space left to store them.
7557 <<channel-buffering-schemes,per-process buffering>>::
7558 A buffering scheme in which each instrumented process has its own
7559 sub-buffers for a given user space channel.
7561 <<channel-buffering-schemes,per-user buffering>>::
7562 A buffering scheme in which all the processes of a Unix user share the
7563 same sub-buffer for a given user space channel.
7565 <<lttng-relayd,relay daemon>>::
7566 A process which is responsible for receiving the trace data sent by
7567 a distant consumer daemon.
7570 A set of sub-buffers.
7572 <<lttng-sessiond,session daemon>>::
7573 A process which receives control commands from you and orchestrates
7574 the tracers and various LTTng daemons.
7576 <<taking-a-snapshot,snapshot>>::
7577 A copy of the current data of all the sub-buffers of a given tracing
7578 session, saved as trace files.
7581 One part of an LTTng ring buffer which contains event records.
7584 The time information attached to an event when it is emitted.
7587 A set of files which are the concatenations of one or more
7588 flushed sub-buffers.
7591 The action of recording the events emitted by an application
7592 or by a system, or to initiate such recording by controlling
7596 The http://tracecompass.org[Trace Compass] project and application.
7599 An instrumentation point using the tracepoint mechanism of the Linux
7600 kernel or of LTTng-UST.
7602 tracepoint definition::
7603 The definition of a single tracepoint.
7606 The name of a tracepoint.
7608 tracepoint provider::
7609 A set of functions providing tracepoints to an instrumented user
7612 Not to be confused with a _tracepoint provider package_: many tracepoint
7613 providers can exist within a tracepoint provider package.
7615 tracepoint provider package::
7616 One or more tracepoint providers compiled as an object file or as
7620 A software which records emitted events.
7622 <<domain,tracing domain>>::
7623 A namespace for event sources.
7625 <<tracing-group,tracing group>>::
7626 The Unix group in which a Unix user can be to be allowed to trace the
7629 <<tracing-session,tracing session>>::
7630 A stateful dialogue between you and a <<lttng-sessiond,session
7634 An application running in user space, as opposed to a Linux kernel
7635 module, for example.