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 * **Tracing control**:
78 ** You can attach <<java-application-context,Java application-specific
79 context fields>> to a <<channel,channel>> with the
80 man:lttng-add-context(1) command:
85 lttng add-context --jul --type='$app.retriever:cur_msg_id'
89 Here, `$app` is the prefix of all application-specific context fields,
90 `retriever` names a _context information retriever_ defined at the
91 application level, and `cur_msg_id` names a context field read from this
94 Both the `java.util.logging` and Apache log4j <<domain,tracing domains>>
97 ** You can use Java application-specific <<adding-context,context>>
98 fields in the <<enabling-disabling-events,filter expression>> of an
104 lttng enable-event --log4j my_logger \
105 --filter='$app.retriever:cur_msg_id == 23'
109 ** New `lttng status` command which is the equivalent of +lttng list
110 __CUR__+, where +__CUR__+ is the name of the current
111 <<tracing-session,tracing session>>.
113 See man:lttng-status(1).
115 ** New `lttng metadata regenerate` command to
116 <<metadata-regenerate,regenerate the metadata file of an LTTng
117 trace>> at any moment. This command is meant to be used to resample
118 the wall time following a major
119 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
120 so that a system which boots with an incorrect wall time can be
121 traced before its wall time is NTP-corrected.
123 See man:lttng-metadata(1).
125 ** New command-line interface warnings when <<event,event records>> or
126 whole sub-buffers are
127 <<channel-overwrite-mode-vs-discard-mode,lost>>. The warning messages
128 are printed when a <<tracing-session,tracing session>> is
129 <<basic-tracing-session-control,stopped>> (man:lttng-stop(1)
132 * **User space tracing**:
133 ** Shared object base address dump in order to map <<event,event
134 records>> to original source location (file and line number).
136 If you attach the `ip` and `vpid` <<adding-context,context fields>> to a
137 user space <<channel,channel>> and if you use the
138 <<liblttng-ust-dl,path:{liblttng-ust-dl.so} helper>>, you can retrieve
139 the source location where a given event record was generated.
141 The http://diamon.org/babeltrace/[Babeltrace] trace viewer supports this
142 state dump and those context fields since version 1.4 to print the
143 source location of a given event record. http://tracecompass.org/[Trace
144 Compass] also supports this since version 2.0.
146 ** A <<java-application,Java application>> which uses
147 `java.util.logging` now adds an LTTng-UST log handler to the desired
150 The previous workflow was to initialize the LTTng-UST Java agent
151 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
152 an LTTng-UST log handler to the root loggers.
154 ** A <<java-application,Java application>> which uses Apache log4j now
155 adds an LTTng-UST log appender to the desired log4j loggers.
157 The previous workflow was to initialize the LTTng-UST Java agent
158 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
159 an LTTng-UST appender to the root loggers.
161 ** Any <<java-application,Java application>> can provide
162 <<java-application-context,dynamic context fields>> while running
163 thanks to a new API provided by the <<lttng-ust-agents,LTTng-UST Java
164 agent>>. You can require LTTng to record specific context fields in
165 event records, and you can use them in the filter expression of
166 <<event,event rules>>.
168 * **Linux kernel tracing**:
169 ** The LTTng kernel modules can now be built into a Linux kernel image,
170 that is, not as loadable modules.
173 https://github.com/lttng/lttng-modules/blob/stable-{revision}/README.md#kernel-built-in-support[`README.md`]
176 ** New instrumentation:
177 *** ARM64 architecture support.
179 *** x86 `irq_vectors`.
180 ** New <<adding-context,context fields>>:
183 *** `need_reschedule`
184 *** `migratable` (specific to RT-Preempt)
185 ** Clock source plugin support for advanced cases where a custom source
186 of time is needed to timestamp LTTng event records.
188 See https://github.com/lttng/lttng-modules/blob/stable-{revision}/lttng-clock.h[`lttng-clock.h`]
189 for an overview of the small API.
192 ** The link:/man[man pages] of the man:lttng(1) command-line tool are
193 split into one man page per command (à la Git), for example:
198 man lttng-enable-event
202 You can also use the `--help` option of any man:lttng(1) command to
205 The content and formatting of all the LTTng man pages has improved
212 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
213 generation_ is a modern toolkit for tracing Linux systems and
214 applications. So your first question might be:
221 As the history of software engineering progressed and led to what
222 we now take for granted--complex, numerous and
223 interdependent software applications running in parallel on
224 sophisticated operating systems like Linux--the authors of such
225 components, software developers, began feeling a natural
226 urge to have tools that would ensure the robustness and good performance
227 of their masterpieces.
229 One major achievement in this field is, inarguably, the
230 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
231 an essential tool for developers to find and fix bugs. But even the best
232 debugger won't help make your software run faster, and nowadays, faster
233 software means either more work done by the same hardware, or cheaper
234 hardware for the same work.
236 A _profiler_ is often the tool of choice to identify performance
237 bottlenecks. Profiling is suitable to identify _where_ performance is
238 lost in a given software. The profiler outputs a profile, a statistical
239 summary of observed events, which you may use to discover which
240 functions took the most time to execute. However, a profiler won't
241 report _why_ some identified functions are the bottleneck. Bottlenecks
242 might only occur when specific conditions are met, conditions that are
243 sometimes impossible to capture by a statistical profiler, or impossible
244 to reproduce with an application altered by the overhead of an
245 event-based profiler. For a thorough investigation of software
246 performance issues, a history of execution is essential, with the
247 recorded values of variables and context fields you choose, and
248 with as little influence as possible on the instrumented software. This
249 is where tracing comes in handy.
251 _Tracing_ is a technique used to understand what goes on in a running
252 software system. The software used for tracing is called a _tracer_,
253 which is conceptually similar to a tape recorder. When recording,
254 specific instrumentation points placed in the software source code
255 generate events that are saved on a giant tape: a _trace_ file. You
256 can trace user applications and the operating system at the same time,
257 opening the possibility of resolving a wide range of problems that would
258 otherwise be extremely challenging.
260 Tracing is often compared to _logging_. However, tracers and loggers are
261 two different tools, serving two different purposes. Tracers are
262 designed to record much lower-level events that occur much more
263 frequently than log messages, often in the range of thousands per
264 second, with very little execution overhead. Logging is more appropriate
265 for a very high-level analysis of less frequent events: user accesses,
266 exceptional conditions (errors and warnings, for example), database
267 transactions, instant messaging communications, and such. Simply put,
268 logging is one of the many use cases that can be satisfied with tracing.
270 The list of recorded events inside a trace file can be read manually
271 like a log file for the maximum level of detail, but it is generally
272 much more interesting to perform application-specific analyses to
273 produce reduced statistics and graphs that are useful to resolve a
274 given problem. Trace viewers and analyzers are specialized tools
277 In the end, this is what LTTng is: a powerful, open source set of
278 tools to trace the Linux kernel and user applications at the same time.
279 LTTng is composed of several components actively maintained and
280 developed by its link:/community/#where[community].
283 [[lttng-alternatives]]
284 === Alternatives to noch:{LTTng}
286 Excluding proprietary solutions, a few competing software tracers
289 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
290 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
291 user scripts and is responsible for loading code into the
292 Linux kernel for further execution and collecting the outputted data.
293 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
294 subsystem in the Linux kernel in which a virtual machine can execute
295 programs passed from the user space to the kernel. You can attach
296 such programs to tracepoints and KProbes thanks to a system call, and
297 they can output data to the user space when executed thanks to
298 different mechanisms (pipe, VM register values, and eBPF maps, to name
300 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
301 is the de facto function tracer of the Linux kernel. Its user
302 interface is a set of special files in sysfs.
303 * https://perf.wiki.kernel.org/[perf] is
304 a performance analyzing tool for Linux which supports hardware
305 performance counters, tracepoints, as well as other counters and
306 types of probes. perf's controlling utility is the cmd:perf command
308 * http://linux.die.net/man/1/strace[strace]
309 is a command-line utility which records system calls made by a
310 user process, as well as signal deliveries and changes of process
311 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
312 to fulfill its function.
313 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
314 analyze Linux kernel events. You write scripts, or _chisels_ in
315 sysdig's jargon, in Lua and sysdig executes them while the system is
316 being traced or afterwards. sysdig's interface is the cmd:sysdig
317 command-line tool as well as the curses-based cmd:csysdig tool.
318 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
319 user space tracer which uses custom user scripts to produce plain text
320 traces. SystemTap converts the scripts to the C language, and then
321 compiles them as Linux kernel modules which are loaded to produce
322 trace data. SystemTap's primary user interface is the cmd:stap
325 The main distinctive features of LTTng is that it produces correlated
326 kernel and user space traces, as well as doing so with the lowest
327 overhead amongst other solutions. It produces trace files in the
328 http://diamon.org/ctf[CTF] format, a file format optimized
329 for the production and analyses of multi-gigabyte data.
331 LTTng is the result of more than 10 years of active open source
332 development by a community of passionate developers.
333 LTTng{nbsp}{revision} is currently available on major desktop and server
336 The main interface for tracing control is a single command-line tool
337 named cmd:lttng. The latter can create several tracing sessions, enable
338 and disable events on the fly, filter events efficiently with custom
339 user expressions, start and stop tracing, and much more. LTTng can
340 record the traces on the file system or send them over the network, and
341 keep them totally or partially. You can view the traces once tracing
342 becomes inactive or in real-time.
344 <<installing-lttng,Install LTTng now>> and
345 <<getting-started,start tracing>>!
351 **LTTng** is a set of software <<plumbing,components>> which interact to
352 <<instrumenting,instrument>> the Linux kernel and user applications, and
353 to <<controlling-tracing,control tracing>> (start and stop
354 tracing, enable and disable event rules, and the rest). Those
355 components are bundled into the following packages:
357 * **LTTng-tools**: Libraries and command-line interface to
359 * **LTTng-modules**: Linux kernel modules to instrument and
361 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
362 trace user applications.
364 Most distributions mark the LTTng-modules and LTTng-UST packages as
365 optional when installing LTTng-tools (which is always required). In the
366 following sections, we always provide the steps to install all three,
369 * You only need to install LTTng-modules if you intend to trace the
371 * You only need to install LTTng-UST if you intend to trace user
375 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 28 November 2016.
377 |Distribution |Available in releases |Alternatives
380 |<<ubuntu,Ubuntu{nbsp}16.10 _Yakkety Yak_>>.
381 |LTTng{nbsp}{revision} for Ubuntu{nbsp}14.04 _Trusty Tahr_
382 and Ubuntu{nbsp}16.04 _Xenial Xerus_:
383 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
385 LTTng{nbsp}2.9 for Ubuntu{nbsp}14.04 _Trusty Tahr_
386 and Ubuntu{nbsp}16.04 _Xenial Xerus_:
387 link:/docs/v2.9#doc-ubuntu-ppa[use the LTTng Stable{nbsp}2.9 PPA].
389 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
390 other Ubuntu releases.
393 |<<fedora,Fedora{nbsp}25>>.
394 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
395 other Fedora releases.
398 |<<debian,Debian "stretch" (testing)>>.
399 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
400 previous Debian releases.
404 |<<building-from-source,Build LTTng{nbsp}{revision} from source>>.
408 |link:/docs/v2.9#doc-arch-linux[LTTng{nbsp}2.9 from the AUR].
411 |<<alpine-linux,Alpine Linux "edge">>.
412 |LTTng{nbsp}{revision} for Alpine Linux{nbsp}3.5 (not released yet).
414 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
415 other Alpine Linux releases.
418 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
423 |LTTng{nbsp}{revision} for Buildroot{nbsp}2016.11 (not released yet).
425 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
426 other Buildroot releases.
428 |OpenEmbedded and Yocto
429 |<<oe-yocto,Yocto Project{nbsp}2.2 _Morty_>> (`openembedded-core` layer).
430 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
431 other OpenEmbedded releases.
436 === [[ubuntu-official-repositories]]Ubuntu
438 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.10 _Yakkety Yak_.
439 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
440 Stable{nbsp}{revision} PPA>>.
442 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}16.10 _Yakkety Yak_:
444 . Install the main LTTng{nbsp}{revision} packages:
449 sudo apt-get install lttng-tools
450 sudo apt-get install lttng-modules-dkms
451 sudo apt-get install liblttng-ust-dev
455 . **If you need to instrument and trace
456 <<java-application,Java applications>>**, install the LTTng-UST
462 sudo apt-get install liblttng-ust-agent-java
466 . **If you need to instrument and trace
467 <<python-application,Python{nbsp}3 applications>>**, install the
468 LTTng-UST Python agent:
473 sudo apt-get install python3-lttngust
479 ==== noch:{LTTng} Stable {revision} PPA
481 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
482 Stable{nbsp}{revision} PPA] offers the latest stable
483 LTTng{nbsp}{revision} packages for:
485 * Ubuntu{nbsp}14.04 _Trusty Tahr_
486 * Ubuntu{nbsp}16.04 _Xenial Xerus_
488 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
490 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
496 sudo apt-add-repository ppa:lttng/stable-2.8
501 . Install the main LTTng{nbsp}{revision} packages:
506 sudo apt-get install lttng-tools
507 sudo apt-get install lttng-modules-dkms
508 sudo apt-get install liblttng-ust-dev
512 . **If you need to instrument and trace
513 <<java-application,Java applications>>**, install the LTTng-UST
519 sudo apt-get install liblttng-ust-agent-java
523 . **If you need to instrument and trace
524 <<python-application,Python{nbsp}3 applications>>**, install the
525 LTTng-UST Python agent:
530 sudo apt-get install python3-lttngust
538 To install LTTng{nbsp}{revision} on Fedora{nbsp}25:
540 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
546 sudo yum install lttng-tools
547 sudo yum install lttng-ust
551 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
557 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
558 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
559 cd lttng-modules-2.8.* &&
561 sudo make modules_install &&
567 .Java and Python application instrumentation and tracing
569 If you need to instrument and trace <<java-application,Java
570 applications>> on openSUSE, you need to build and install
571 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
572 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
573 `--enable-java-agent-all` options to the `configure` script, depending
574 on which Java logging framework you use.
576 If you need to instrument and trace <<python-application,Python
577 applications>> on openSUSE, you need to build and install
578 LTTng-UST{nbsp}{revision} from source and pass the
579 `--enable-python-agent` option to the `configure` script.
586 To install LTTng{nbsp}{revision} on Debian "stretch" (testing):
588 . Install the main LTTng{nbsp}{revision} packages:
593 sudo apt-get install lttng-modules-dkms
594 sudo apt-get install liblttng-ust-dev
595 sudo apt-get install lttng-tools
599 . **If you need to instrument and trace <<java-application,Java
600 applications>>**, install the LTTng-UST Java agent:
605 sudo apt-get install liblttng-ust-agent-java
609 . **If you need to instrument and trace <<python-application,Python
610 applications>>**, install the LTTng-UST Python agent:
615 sudo apt-get install python3-lttngust
623 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
626 . Make sure your system is
627 https://wiki.alpinelinux.org/wiki/Edge[configured for "edge"].
628 . Enable the _testing_ repository by uncommenting the corresponding
629 line in path:{/etc/apk/repositories}.
630 . Add the LTTng packages:
635 sudo apk add lttng-tools
636 sudo apk add lttng-ust-dev
640 To install LTTng-modules{nbsp}{revision} (Linux kernel tracing support)
641 on Alpine Linux "edge":
643 . Add the vanilla Linux kernel:
648 apk add linux-vanilla linux-vanilla-dev
652 . Reboot with the vanilla Linux kernel.
653 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
659 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
660 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
661 cd lttng-modules-2.8.* &&
663 sudo make modules_install &&
670 === OpenEmbedded and Yocto
672 LTTng{nbsp}{revision} recipes are available in the
673 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
674 layer for Yocto Project{nbsp}2.2 _Morty_ under the following names:
680 With BitBake, the simplest way to include LTTng recipes in your target
681 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
684 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
689 . Select a machine and an image recipe.
690 . Click **Edit image recipe**.
691 . Under the **All recipes** tab, search for **lttng**.
692 . Check the desired LTTng recipes.
695 .Java and Python application instrumentation and tracing
697 If you need to instrument and trace <<java-application,Java
698 applications>> on openSUSE, you need to build and install
699 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
700 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
701 `--enable-java-agent-all` options to the `configure` script, depending
702 on which Java logging framework you use.
704 If you need to instrument and trace <<python-application,Python
705 applications>> on openSUSE, you need to build and install
706 LTTng-UST{nbsp}{revision} from source and pass the
707 `--enable-python-agent` option to the `configure` script.
711 [[enterprise-distributions]]
712 === RHEL, SUSE, and other enterprise distributions
714 To install LTTng on enterprise Linux distributions, such as Red Hat
715 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
716 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
719 [[building-from-source]]
720 === Build from source
722 To build and install LTTng{nbsp}{revision} from source:
724 . Using your distribution's package manager, or from source, install
725 the following dependencies of LTTng-tools and LTTng-UST:
728 * https://sourceforge.net/projects/libuuid/[libuuid]
729 * http://directory.fsf.org/wiki/Popt[popt]
730 * http://liburcu.org/[Userspace RCU]
731 * http://www.xmlsoft.org/[libxml2]
734 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
740 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
741 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
742 cd lttng-modules-2.8.* &&
744 sudo make modules_install &&
749 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
755 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
756 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
757 cd lttng-ust-2.8.* &&
767 .Java and Python application tracing
769 If you need to instrument and trace <<java-application,Java
770 applications>>, pass the `--enable-java-agent-jul`,
771 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
772 `configure` script, depending on which Java logging framework you use.
774 If you need to instrument and trace <<python-application,Python
775 applications>>, pass the `--enable-python-agent` option to the
776 `configure` script. You can set the `PYTHON` environment variable to the
777 path to the Python interpreter for which to install the LTTng-UST Python
785 By default, LTTng-UST libraries are installed to
786 dir:{/usr/local/lib}, which is the de facto directory in which to
787 keep self-compiled and third-party libraries.
789 When <<building-tracepoint-providers-and-user-application,linking an
790 instrumented user application with `liblttng-ust`>>:
792 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
794 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
795 man:gcc(1), man:g++(1), or man:clang(1).
799 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
805 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
806 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
807 cd lttng-tools-2.8.* &&
815 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
816 previous steps automatically for a given version of LTTng and confine
817 the installed files in a specific directory. This can be useful to test
818 LTTng without installing it on your system.
824 This is a short guide to get started quickly with LTTng kernel and user
827 Before you follow this guide, make sure to <<installing-lttng,install>>
830 This tutorial walks you through the steps to:
832 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
833 . <<tracing-your-own-user-application,Trace a user application>> written
835 . <<viewing-and-analyzing-your-traces,View and analyze the
839 [[tracing-the-linux-kernel]]
840 === Trace the Linux kernel
842 The following command lines start with cmd:sudo because you need root
843 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
844 your Unix user is a member of the <<lttng-sessiond,tracing group>>.
846 . Create a <<tracing-session,tracing session>>:
851 sudo lttng create my-kernel-session
855 . List the available kernel tracepoints and system calls:
864 . Create an <<event,event rule>> which matches the desired event names,
865 for example `sched_switch` and `sched_process_fork`:
870 sudo lttng enable-event --kernel sched_switch,sched_process_fork
874 You can also create an event rule which _matches_ all the Linux kernel
875 tracepoints (this will generate a lot of data when tracing):
880 sudo lttng enable-event --kernel --all
893 . Do some operation on your system for a few seconds. For example,
894 load a website, or list the files of a directory.
895 . Stop tracing and destroy the tracing session:
905 The man:lttng-destroy(1) command does not destroy the trace data; it
906 only destroys the state of the tracing session.
908 By default, LTTng saves the traces in
909 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
910 where +__name__+ is the tracing session name. Note that the
911 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
913 See <<viewing-and-analyzing-your-traces,View and analyze the
914 recorded events>> to view the recorded events.
917 [[tracing-your-own-user-application]]
918 === Trace a user application
920 This section steps you through a simple example to trace a
921 _Hello world_ program written in C.
923 To create the traceable user application:
925 . Create the tracepoint provider header file, which defines the
926 tracepoints and the events they can generate:
932 #undef TRACEPOINT_PROVIDER
933 #define TRACEPOINT_PROVIDER hello_world
935 #undef TRACEPOINT_INCLUDE
936 #define TRACEPOINT_INCLUDE "./hello-tp.h"
938 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
941 #include <lttng/tracepoint.h>
951 ctf_string(my_string_field, my_string_arg)
952 ctf_integer(int, my_integer_field, my_integer_arg)
956 #endif /* _HELLO_TP_H */
958 #include <lttng/tracepoint-event.h>
962 . Create the tracepoint provider package source file:
968 #define TRACEPOINT_CREATE_PROBES
969 #define TRACEPOINT_DEFINE
971 #include "hello-tp.h"
975 . Build the tracepoint provider package:
980 gcc -c -I. hello-tp.c
984 . Create the _Hello World_ application source file:
991 #include "hello-tp.h"
993 int main(int argc, char *argv[])
997 puts("Hello, World!\nPress Enter to continue...");
1000 * The following getchar() call is only placed here for the purpose
1001 * of this demonstration, to pause the application in order for
1002 * you to have time to list its tracepoints. It is not
1008 * A tracepoint() call.
1010 * Arguments, as defined in hello-tp.h:
1012 * 1. Tracepoint provider name (required)
1013 * 2. Tracepoint name (required)
1014 * 3. my_integer_arg (first user-defined argument)
1015 * 4. my_string_arg (second user-defined argument)
1017 * Notice the tracepoint provider and tracepoint names are
1018 * NOT strings: they are in fact parts of variables that the
1019 * macros in hello-tp.h create.
1021 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
1023 for (x = 0; x < argc; ++x) {
1024 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
1027 puts("Quitting now!");
1028 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
1035 . Build the application:
1044 . Link the application with the tracepoint provider package,
1045 `liblttng-ust`, and `libdl`:
1050 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1054 Here's the whole build process:
1057 .User space tracing tutorial's build steps.
1058 image::ust-flow.png[]
1060 To trace the user application:
1062 . Run the application with a few arguments:
1067 ./hello world and beyond
1076 Press Enter to continue...
1080 . Start an LTTng <<lttng-sessiond,session daemon>>:
1085 lttng-sessiond --daemonize
1089 Note that a session daemon might already be running, for example as
1090 a service that the distribution's service manager started.
1092 . List the available user space tracepoints:
1097 lttng list --userspace
1101 You see the `hello_world:my_first_tracepoint` tracepoint listed
1102 under the `./hello` process.
1104 . Create a <<tracing-session,tracing session>>:
1109 lttng create my-user-space-session
1113 . Create an <<event,event rule>> which matches the
1114 `hello_world:my_first_tracepoint` event name:
1119 lttng enable-event --userspace hello_world:my_first_tracepoint
1132 . Go back to the running `hello` application and press Enter. The
1133 program executes all `tracepoint()` instrumentation points and exits.
1134 . Stop tracing and destroy the tracing session:
1144 The man:lttng-destroy(1) command does not destroy the trace data; it
1145 only destroys the state of the tracing session.
1147 By default, LTTng saves the traces in
1148 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1149 where +__name__+ is the tracing session name. Note that the
1150 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1152 See <<viewing-and-analyzing-your-traces,View and analyze the
1153 recorded events>> to view the recorded events.
1156 [[viewing-and-analyzing-your-traces]]
1157 === View and analyze the recorded events
1159 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1160 kernel>> and <<tracing-your-own-user-application,Trace a user
1161 application>> tutorials, you can inspect the recorded events.
1163 Many tools are available to read LTTng traces:
1165 * **cmd:babeltrace** is a command-line utility which converts trace
1166 formats; it supports the format that LTTng produces, CTF, as well as a
1167 basic text output which can be ++grep++ed. The cmd:babeltrace command
1168 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1169 * Babeltrace also includes
1170 **https://www.python.org/[Python] bindings** so
1171 that you can easily open and read an LTTng trace with your own script,
1172 benefiting from the power of Python.
1173 * http://tracecompass.org/[**Trace Compass**]
1174 is a graphical user interface for viewing and analyzing any type of
1175 logs or traces, including LTTng's.
1176 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1177 project which includes many high-level analyses of LTTng kernel
1178 traces, like scheduling statistics, interrupt frequency distribution,
1179 top CPU usage, and more.
1181 NOTE: This section assumes that the traces recorded during the previous
1182 tutorials were saved to their default location, in the
1183 dir:{$LTTNG_HOME/lttng-traces} directory. Note that the env:LTTNG_HOME
1184 environment variable defaults to `$HOME` if not set.
1187 [[viewing-and-analyzing-your-traces-bt]]
1188 ==== Use the cmd:babeltrace command-line tool
1190 The simplest way to list all the recorded events of a trace is to pass
1191 its path to cmd:babeltrace with no options:
1195 babeltrace ~/lttng-traces/my-user-space-session*
1198 cmd:babeltrace finds all traces recursively within the given path and
1199 prints all their events, merging them in chronological order.
1201 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1206 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_
1209 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1210 count the recorded events:
1214 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_read | wc --lines
1218 [[viewing-and-analyzing-your-traces-bt-python]]
1219 ==== Use the Babeltrace Python bindings
1221 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1222 is useful to isolate events by simple matching using man:grep(1) and
1223 similar utilities. However, more elaborate filters, such as keeping only
1224 event records with a field value falling within a specific range, are
1225 not trivial to write using a shell. Moreover, reductions and even the
1226 most basic computations involving multiple event records are virtually
1227 impossible to implement.
1229 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1230 to read the event records of an LTTng trace sequentially and compute the
1231 desired information.
1233 The following script accepts an LTTng Linux kernel trace path as its
1234 first argument and prints the short names of the top 5 running processes
1235 on CPU 0 during the whole trace:
1240 from collections import Counter
1246 if len(sys.argv) != 2:
1247 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1248 print(msg, file=sys.stderr)
1251 # A trace collection contains one or more traces
1252 col = babeltrace.TraceCollection()
1254 # Add the trace provided by the user (LTTng traces always have
1256 if col.add_trace(sys.argv[1], 'ctf') is None:
1257 raise RuntimeError('Cannot add trace')
1259 # This counter dict contains execution times:
1261 # task command name -> total execution time (ns)
1262 exec_times = Counter()
1264 # This contains the last `sched_switch` timestamp
1268 for event in col.events:
1269 # Keep only `sched_switch` events
1270 if event.name != 'sched_switch':
1273 # Keep only events which happened on CPU 0
1274 if event['cpu_id'] != 0:
1278 cur_ts = event.timestamp
1284 # Previous task command (short) name
1285 prev_comm = event['prev_comm']
1287 # Initialize entry in our dict if not yet done
1288 if prev_comm not in exec_times:
1289 exec_times[prev_comm] = 0
1291 # Compute previous command execution time
1292 diff = cur_ts - last_ts
1294 # Update execution time of this command
1295 exec_times[prev_comm] += diff
1297 # Update last timestamp
1301 for name, ns in exec_times.most_common(5):
1303 print('{:20}{} s'.format(name, s))
1308 if __name__ == '__main__':
1309 sys.exit(0 if top5proc() else 1)
1316 python3 top5proc.py ~/lttng-traces/my-kernel-session*/kernel
1322 swapper/0 48.607245889 s
1323 chromium 7.192738188 s
1324 pavucontrol 0.709894415 s
1325 Compositor 0.660867933 s
1326 Xorg.bin 0.616753786 s
1329 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1330 weren't using the CPU that much when tracing, its first position in the
1335 == [[understanding-lttng]]Core concepts
1337 From a user's perspective, the LTTng system is built on a few concepts,
1338 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1339 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1340 Understanding how those objects relate to eachother is key in mastering
1343 The core concepts are:
1345 * <<tracing-session,Tracing session>>
1346 * <<domain,Tracing domain>>
1347 * <<channel,Channel and ring buffer>>
1348 * <<"event","Instrumentation point, event rule, event, and event record">>
1354 A _tracing session_ is a stateful dialogue between you and
1355 a <<lttng-sessiond,session daemon>>. You can
1356 <<creating-destroying-tracing-sessions,create a new tracing
1357 session>> with the `lttng create` command.
1359 Anything that you do when you control LTTng tracers happens within a
1360 tracing session. In particular, a tracing session:
1363 * Has its own set of trace files.
1364 * Has its own state of activity (started or stopped).
1365 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1367 * Has its own <<channel,channels>> which have their own
1368 <<event,event rules>>.
1371 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1372 image::concepts.png[]
1374 Those attributes and objects are completely isolated between different
1377 A tracing session is analogous to a cash machine session:
1378 the operations you do on the banking system through the cash machine do
1379 not alter the data of other users of the same system. In the case of
1380 the cash machine, a session lasts as long as your bank card is inside.
1381 In the case of LTTng, a tracing session lasts from the `lttng create`
1382 command to the `lttng destroy` command.
1385 .Each Unix user has its own set of tracing sessions.
1386 image::many-sessions.png[]
1389 [[tracing-session-mode]]
1390 ==== Tracing session mode
1392 LTTng can send the generated trace data to different locations. The
1393 _tracing session mode_ dictates where to send it. The following modes
1394 are available in LTTng{nbsp}{revision}:
1397 LTTng writes the traces to the file system of the machine being traced
1400 Network streaming mode::
1401 LTTng sends the traces over the network to a
1402 <<lttng-relayd,relay daemon>> running on a remote system.
1405 LTTng does not write the traces by default. Instead, you can request
1406 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1407 current tracing buffers, and to write it to the target's file system
1408 or to send it over the network to a <<lttng-relayd,relay daemon>>
1409 running on a remote system.
1412 This mode is similar to the network streaming mode, but a live
1413 trace viewer can connect to the distant relay daemon to
1414 <<lttng-live,view event records as LTTng generates them>> by
1421 A _tracing domain_ is a namespace for event sources. A tracing domain
1422 has its own properties and features.
1424 There are currently five available tracing domains:
1428 * `java.util.logging` (JUL)
1432 You must specify a tracing domain when using some commands to avoid
1433 ambiguity. For example, since all the domains support named tracepoints
1434 as event sources (instrumentation points that you manually insert in the
1435 source code), you need to specify a tracing domain when
1436 <<enabling-disabling-events,creating an event rule>> because all the
1437 tracing domains could have tracepoints with the same names.
1439 Some features are reserved to specific tracing domains. Dynamic function
1440 entry and return instrumentation points, for example, are currently only
1441 supported in the Linux kernel tracing domain, but support for other
1442 tracing domains could be added in the future.
1444 You can create <<channel,channels>> in the Linux kernel and user space
1445 tracing domains. The other tracing domains have a single default
1450 === Channel and ring buffer
1452 A _channel_ is an object which is responsible for a set of ring buffers.
1453 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1454 tracer emits an event, it can record it to one or more
1455 sub-buffers. The attributes of a channel determine what to do when
1456 there's no space left for a new event record because all sub-buffers
1457 are full, where to send a full sub-buffer, and other behaviours.
1459 A channel is always associated to a <<domain,tracing domain>>. The
1460 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1461 a default channel which you cannot configure.
1463 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1464 an event, it records it to the sub-buffers of all
1465 the enabled channels with a satisfied event rule, as long as those
1466 channels are part of active <<tracing-session,tracing sessions>>.
1469 [[channel-buffering-schemes]]
1470 ==== Per-user vs. per-process buffering schemes
1472 A channel has at least one ring buffer _per CPU_. LTTng always
1473 records an event to the ring buffer associated to the CPU on which it
1476 Two _buffering schemes_ are available when you
1477 <<enabling-disabling-channels,create a channel>> in the
1478 user space <<domain,tracing domain>>:
1480 Per-user buffering::
1481 Allocate one set of ring buffers--one per CPU--shared by all the
1482 instrumented processes of each Unix user.
1486 .Per-user buffering scheme.
1487 image::per-user-buffering.png[]
1490 Per-process buffering::
1491 Allocate one set of ring buffers--one per CPU--for each
1492 instrumented process.
1496 .Per-process buffering scheme.
1497 image::per-process-buffering.png[]
1500 The per-process buffering scheme tends to consume more memory than the
1501 per-user option because systems generally have more instrumented
1502 processes than Unix users running instrumented processes. However, the
1503 per-process buffering scheme ensures that one process having a high
1504 event throughput won't fill all the shared sub-buffers of the same
1507 The Linux kernel tracing domain has only one available buffering scheme
1508 which is to allocate a single set of ring buffers for the whole system.
1509 This scheme is similar to the per-user option, but with a single, global
1510 user "running" the kernel.
1513 [[channel-overwrite-mode-vs-discard-mode]]
1514 ==== Overwrite vs. discard event loss modes
1516 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1517 arc in the following animation) of a specific channel's ring buffer.
1518 When there's no space left in a sub-buffer, the tracer marks it as
1519 consumable (red) and another, empty sub-buffer starts receiving the
1520 following event records. A <<lttng-consumerd,consumer daemon>>
1521 eventually consumes the marked sub-buffer (returns to white).
1524 [role="docsvg-channel-subbuf-anim"]
1529 In an ideal world, sub-buffers are consumed faster than they are filled,
1530 as is the case in the previous animation. In the real world,
1531 however, all sub-buffers can be full at some point, leaving no space to
1532 record the following events.
1534 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1535 available, it is acceptable to lose event records when the alternative
1536 would be to cause substantial delays in the instrumented application's
1537 execution. LTTng privileges performance over integrity; it aims at
1538 perturbing the traced system as little as possible in order to make
1539 tracing of subtle race conditions and rare interrupt cascades possible.
1541 When it comes to losing event records because no empty sub-buffer is
1542 available, the channel's _event loss mode_ determines what to do. The
1543 available event loss modes are:
1546 Drop the newest event records until a the tracer
1547 releases a sub-buffer.
1550 Clear the sub-buffer containing the oldest event records and start
1551 writing the newest event records there.
1553 This mode is sometimes called _flight recorder mode_ because it's
1555 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1556 always keep a fixed amount of the latest data.
1558 Which mechanism you should choose depends on your context: prioritize
1559 the newest or the oldest event records in the ring buffer?
1561 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1562 as soon as a there's no space left for a new event record, whereas in
1563 discard mode, the tracer only discards the event record that doesn't
1566 In discard mode, LTTng increments a count of lost event records when
1567 an event record is lost and saves this count to the trace. In
1568 overwrite mode, LTTng keeps no information when it overwrites a
1569 sub-buffer before consuming it.
1571 There are a few ways to decrease your probability of losing event
1573 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1574 how you can fine-une the sub-buffer count and size of a channel to
1575 virtually stop losing event records, though at the cost of greater
1579 [[channel-subbuf-size-vs-subbuf-count]]
1580 ==== Sub-buffer count and size
1582 When you <<enabling-disabling-channels,create a channel>>, you can
1583 set its number of sub-buffers and their size.
1585 Note that there is noticeable CPU overhead introduced when
1586 switching sub-buffers (marking a full one as consumable and switching
1587 to an empty one for the following events to be recorded). Knowing this,
1588 the following list presents a few practical situations along with how
1589 to configure the sub-buffer count and size for them:
1591 * **High event throughput**: In general, prefer bigger sub-buffers to
1592 lower the risk of losing event records.
1594 Having bigger sub-buffers also ensures a lower sub-buffer switching
1597 The number of sub-buffers is only meaningful if you create the channel
1598 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1599 other sub-buffers are left unaltered.
1601 * **Low event throughput**: In general, prefer smaller sub-buffers
1602 since the risk of losing event records is low.
1604 Because events occur less frequently, the sub-buffer switching frequency
1605 should remain low and thus the tracer's overhead should not be a
1608 * **Low memory system**: If your target system has a low memory
1609 limit, prefer fewer first, then smaller sub-buffers.
1611 Even if the system is limited in memory, you want to keep the
1612 sub-buffers as big as possible to avoid a high sub-buffer switching
1615 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1616 which means event data is very compact. For example, the average
1617 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1618 sub-buffer size of 1{nbsp}MiB is considered big.
1620 The previous situations highlight the major trade-off between a few big
1621 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1622 frequency vs. how much data is lost in overwrite mode. Assuming a
1623 constant event throughput and using the overwrite mode, the two
1624 following configurations have the same ring buffer total size:
1627 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1632 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1633 switching frequency, but if a sub-buffer overwrite happens, half of
1634 the event records so far (4{nbsp}MiB) are definitely lost.
1635 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1636 overhead as the previous configuration, but if a sub-buffer
1637 overwrite happens, only the eighth of event records so far are
1640 In discard mode, the sub-buffers count parameter is pointless: use two
1641 sub-buffers and set their size according to the requirements of your
1645 [[channel-switch-timer]]
1646 ==== Switch timer period
1648 The _switch timer period_ is an important configurable attribute of
1649 a channel to ensure periodic sub-buffer flushing.
1651 When the _switch timer_ expires, a sub-buffer switch happens. You can
1652 set the switch timer period attribute when you
1653 <<enabling-disabling-channels,create a channel>> to ensure that event
1654 data is consumed and committed to trace files or to a distant relay
1655 daemon periodically in case of a low event throughput.
1658 [role="docsvg-channel-switch-timer"]
1663 This attribute is also convenient when you use big sub-buffers to cope
1664 with a sporadic high event throughput, even if the throughput is
1668 [[channel-read-timer]]
1669 ==== Read timer period
1671 By default, the LTTng tracers use a notification mechanism to signal a
1672 full sub-buffer so that a consumer daemon can consume it. When such
1673 notifications must be avoided, for example in real-time applications,
1674 you can use the channel's _read timer_ instead. When the read timer
1675 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1676 consumable sub-buffers.
1679 [[tracefile-rotation]]
1680 ==== Trace file count and size
1682 By default, trace files can grow as large as needed. You can set the
1683 maximum size of each trace file that a channel writes when you
1684 <<enabling-disabling-channels,create a channel>>. When the size of
1685 a trace file reaches the channel's fixed maximum size, LTTng creates
1686 another file to contain the next event records. LTTng appends a file
1687 count to each trace file name in this case.
1689 If you set the trace file size attribute when you create a channel, the
1690 maximum number of trace files that LTTng creates is _unlimited_ by
1691 default. To limit them, you can also set a maximum number of trace
1692 files. When the number of trace files reaches the channel's fixed
1693 maximum count, the oldest trace file is overwritten. This mechanism is
1694 called _trace file rotation_.
1698 === Instrumentation point, event rule, event, and event record
1700 An _event rule_ is a set of conditions which must be **all** satisfied
1701 for LTTng to record an occuring event.
1703 You set the conditions when you <<enabling-disabling-events,create
1706 You always attach an event rule to <<channel,channel>> when you create
1709 When an event passes the conditions of an event rule, LTTng records it
1710 in one of the attached channel's sub-buffers.
1712 The available conditions, as of LTTng{nbsp}{revision}, are:
1714 * The event rule _is enabled_.
1715 * The instrumentation point's type _is{nbsp}T_.
1716 * The instrumentation point's name (sometimes called _event name_)
1717 _matches{nbsp}N_, but _is not{nbsp}E_.
1718 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1719 _is exactly{nbsp}L_.
1720 * The fields of the event's payload _satisfy_ a filter
1721 expression{nbsp}__F__.
1723 As you can see, all the conditions but the dynamic filter are related to
1724 the event rule's status or to the instrumentation point, not to the
1725 occurring events. This is why, without a filter, checking if an event
1726 passes an event rule is not a dynamic task: when you create or modify an
1727 event rule, all the tracers of its tracing domain enable or disable the
1728 instrumentation points themselves once. This is possible because the
1729 attributes of an instrumentation point (type, name, and log level) are
1730 defined statically. In other words, without a dynamic filter, the tracer
1731 _does not evaluate_ the arguments of an instrumentation point unless it
1732 matches an enabled event rule.
1734 Note that, for LTTng to record an event, the <<channel,channel>> to
1735 which a matching event rule is attached must also be enabled, and the
1736 tracing session owning this channel must be active.
1739 .Logical path from an instrumentation point to an event record.
1740 image::event-rule.png[]
1742 .Event, event record, or event rule?
1744 With so many similar terms, it's easy to get confused.
1746 An **event** is the consequence of the execution of an _instrumentation
1747 point_, like a tracepoint that you manually place in some source code,
1748 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1749 time. Different actions can be taken upon the occurance of an event,
1750 like record the event's payload to a buffer.
1752 An **event record** is the representation of an event in a sub-buffer. A
1753 tracer is responsible for capturing the payload of an event, current
1754 context variables, the event's ID, and the event's timestamp. LTTng
1755 can append this sub-buffer to a trace file.
1757 An **event rule** is a set of conditions which must all be satisfied for
1758 LTTng to record an occuring event. Events still occur without
1759 satisfying event rules, but LTTng does not record them.
1764 == Components of noch:{LTTng}
1766 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1767 to call LTTng a simple _tool_ since it is composed of multiple
1768 interacting components. This section describes those components,
1769 explains their respective roles, and shows how they connect together to
1770 form the LTTng ecosystem.
1772 The following diagram shows how the most important components of LTTng
1773 interact with user applications, the Linux kernel, and you:
1776 .Control and trace data paths between LTTng components.
1777 image::plumbing.png[]
1779 The LTTng project incorporates:
1781 * **LTTng-tools**: Libraries and command-line interface to
1782 control tracing sessions.
1783 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1784 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1785 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1786 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1787 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1788 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1790 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1791 headers to instrument and trace any native user application.
1792 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1793 *** `liblttng-ust-libc-wrapper`
1794 *** `liblttng-ust-pthread-wrapper`
1795 *** `liblttng-ust-cyg-profile`
1796 *** `liblttng-ust-cyg-profile-fast`
1797 *** `liblttng-ust-dl`
1798 ** User space tracepoint provider source files generator command-line
1799 tool (man:lttng-gen-tp(1)).
1800 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1801 Java applications using `java.util.logging` or
1802 Apache log4j 1.2 logging.
1803 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1804 Python applications using the standard `logging` package.
1805 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1807 ** LTTng kernel tracer module.
1808 ** Tracing ring buffer kernel modules.
1809 ** Probe kernel modules.
1810 ** LTTng logger kernel module.
1814 === Tracing control command-line interface
1817 .The tracing control command-line interface.
1818 image::plumbing-lttng-cli.png[]
1820 The _man:lttng(1) command-line tool_ is the standard user interface to
1821 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1822 is part of LTTng-tools.
1824 The cmd:lttng tool is linked with
1825 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1826 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1828 The cmd:lttng tool has a Git-like interface:
1832 lttng <general options> <command> <command options>
1835 The <<controlling-tracing,Tracing control>> section explores the
1836 available features of LTTng using the cmd:lttng tool.
1839 [[liblttng-ctl-lttng]]
1840 === Tracing control library
1843 .The tracing control library.
1844 image::plumbing-liblttng-ctl.png[]
1846 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1847 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1848 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1850 The <<lttng-cli,cmd:lttng command-line tool>>
1851 is linked with `liblttng-ctl`.
1853 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1858 #include <lttng/lttng.h>
1861 Some objects are referenced by name (C string), such as tracing
1862 sessions, but most of them require to create a handle first using
1863 `lttng_create_handle()`.
1865 The best available developer documentation for `liblttng-ctl` is, as of
1866 LTTng{nbsp}{revision}, its installed header files. Every function and
1867 structure is thoroughly documented.
1871 === User space tracing library
1874 .The user space tracing library.
1875 image::plumbing-liblttng-ust.png[]
1877 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1878 is the LTTng user space tracer. It receives commands from a
1879 <<lttng-sessiond,session daemon>>, for example to
1880 enable and disable specific instrumentation points, and writes event
1881 records to ring buffers shared with a
1882 <<lttng-consumerd,consumer daemon>>.
1883 `liblttng-ust` is part of LTTng-UST.
1885 Public C header files are installed beside `liblttng-ust` to
1886 instrument any <<c-application,C or $$C++$$ application>>.
1888 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1889 packages, use their own library providing tracepoints which is
1890 linked with `liblttng-ust`.
1892 An application or library does not have to initialize `liblttng-ust`
1893 manually: its constructor does the necessary tasks to properly register
1894 to a session daemon. The initialization phase also enables the
1895 instrumentation points matching the <<event,event rules>> that you
1899 [[lttng-ust-agents]]
1900 === User space tracing agents
1903 .The user space tracing agents.
1904 image::plumbing-lttng-ust-agents.png[]
1906 The _LTTng-UST Java and Python agents_ are regular Java and Python
1907 packages which add LTTng tracing capabilities to the
1908 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1910 In the case of Java, the
1911 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1912 core logging facilities] and
1913 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1914 Note that Apache Log4{nbsp}2 is not supported.
1916 In the case of Python, the standard
1917 https://docs.python.org/3/library/logging.html[`logging`] package
1918 is supported. Both Python 2 and Python 3 modules can import the
1919 LTTng-UST Python agent package.
1921 The applications using the LTTng-UST agents are in the
1922 `java.util.logging` (JUL),
1923 log4j, and Python <<domain,tracing domains>>.
1925 Both agents use the same mechanism to trace the log statements. When an
1926 agent is initialized, it creates a log handler that attaches to the root
1927 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1928 When the application executes a log statement, it is passed to the
1929 agent's log handler by the root logger. The agent's log handler calls a
1930 native function in a tracepoint provider package shared library linked
1931 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1932 other fields, like its logger name and its log level. This native
1933 function contains a user space instrumentation point, hence tracing the
1936 The log level condition of an
1937 <<event,event rule>> is considered when tracing
1938 a Java or a Python application, and it's compatible with the standard
1939 JUL, log4j, and Python log levels.
1943 === LTTng kernel modules
1946 .The LTTng kernel modules.
1947 image::plumbing-lttng-modules.png[]
1949 The _LTTng kernel modules_ are a set of Linux kernel modules
1950 which implement the kernel tracer of the LTTng project. The LTTng
1951 kernel modules are part of LTTng-modules.
1953 The LTTng kernel modules include:
1955 * A set of _probe_ modules.
1957 Each module attaches to a specific subsystem
1958 of the Linux kernel using its tracepoint instrument points. There are
1959 also modules to attach to the entry and return points of the Linux
1960 system call functions.
1962 * _Ring buffer_ modules.
1964 A ring buffer implementation is provided as kernel modules. The LTTng
1965 kernel tracer writes to the ring buffer; a
1966 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1968 * The _LTTng kernel tracer_ module.
1969 * The _LTTng logger_ module.
1971 The LTTng logger module implements the special path:{/proc/lttng-logger}
1972 file so that any executable can generate LTTng events by opening and
1973 writing to this file.
1975 See <<proc-lttng-logger-abi,LTTng logger>>.
1977 Generally, you do not have to load the LTTng kernel modules manually
1978 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1979 daemon>> loads the necessary modules when starting. If you have extra
1980 probe modules, you can specify to load them to the session daemon on
1983 The LTTng kernel modules are installed in
1984 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1985 the kernel release (see `uname --kernel-release`).
1992 .The session daemon.
1993 image::plumbing-sessiond.png[]
1995 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1996 managing tracing sessions and for controlling the various components of
1997 LTTng. The session daemon is part of LTTng-tools.
1999 The session daemon sends control requests to and receives control
2002 * The <<lttng-ust,user space tracing library>>.
2004 Any instance of the user space tracing library first registers to
2005 a session daemon. Then, the session daemon can send requests to
2006 this instance, such as:
2009 ** Get the list of tracepoints.
2010 ** Share an <<event,event rule>> so that the user space tracing library
2011 can enable or disable tracepoints. Amongst the possible conditions
2012 of an event rule is a filter expression which `liblttng-ust` evalutes
2013 when an event occurs.
2014 ** Share <<channel,channel>> attributes and ring buffer locations.
2017 The session daemon and the user space tracing library use a Unix
2018 domain socket for their communication.
2020 * The <<lttng-ust-agents,user space tracing agents>>.
2022 Any instance of a user space tracing agent first registers to
2023 a session daemon. Then, the session daemon can send requests to
2024 this instance, such as:
2027 ** Get the list of loggers.
2028 ** Enable or disable a specific logger.
2031 The session daemon and the user space tracing agent use a TCP connection
2032 for their communication.
2034 * The <<lttng-modules,LTTng kernel tracer>>.
2035 * The <<lttng-consumerd,consumer daemon>>.
2037 The session daemon sends requests to the consumer daemon to instruct
2038 it where to send the trace data streams, amongst other information.
2040 * The <<lttng-relayd,relay daemon>>.
2042 The session daemon receives commands from the
2043 <<liblttng-ctl-lttng,tracing control library>>.
2045 The root session daemon loads the appropriate
2046 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2047 a <<lttng-consumerd,consumer daemon>> as soon as you create
2048 an <<event,event rule>>.
2050 The session daemon does not send and receive trace data: this is the
2051 role of the <<lttng-consumerd,consumer daemon>> and
2052 <<lttng-relayd,relay daemon>>. It does, however, generate the
2053 http://diamon.org/ctf/[CTF] metadata stream.
2055 Each Unix user can have its own session daemon instance. The
2056 tracing sessions managed by different session daemons are completely
2059 The root user's session daemon is the only one which is
2060 allowed to control the LTTng kernel tracer, and its spawned consumer
2061 daemon is the only one which is allowed to consume trace data from the
2062 LTTng kernel tracer. Note, however, that any Unix user which is a member
2063 of the <<tracing-group,tracing group>> is allowed
2064 to create <<channel,channels>> in the
2065 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2068 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2069 session daemon when using its `create` command if none is currently
2070 running. You can also start the session daemon manually.
2077 .The consumer daemon.
2078 image::plumbing-consumerd.png[]
2080 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
2081 ring buffers with user applications or with the LTTng kernel modules to
2082 collect trace data and send it to some location (on disk or to a
2083 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2084 is part of LTTng-tools.
2086 You do not start a consumer daemon manually: a consumer daemon is always
2087 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2088 <<event,event rule>>, that is, before you start tracing. When you kill
2089 its owner session daemon, the consumer daemon also exits because it is
2090 the session daemon's child process. Command-line options of
2091 man:lttng-sessiond(8) target the consumer daemon process.
2093 There are up to two running consumer daemons per Unix user, whereas only
2094 one session daemon can run per user. This is because each process can be
2095 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2096 and 64-bit processes, it is more efficient to have separate
2097 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2098 exception: it can have up to _three_ running consumer daemons: 32-bit
2099 and 64-bit instances for its user applications, and one more
2100 reserved for collecting kernel trace data.
2108 image::plumbing-relayd.png[]
2110 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2111 between remote session and consumer daemons, local trace files, and a
2112 remote live trace viewer. The relay daemon is part of LTTng-tools.
2114 The main purpose of the relay daemon is to implement a receiver of
2115 <<sending-trace-data-over-the-network,trace data over the network>>.
2116 This is useful when the target system does not have much file system
2117 space to record trace files locally.
2119 The relay daemon is also a server to which a
2120 <<lttng-live,live trace viewer>> can
2121 connect. The live trace viewer sends requests to the relay daemon to
2122 receive trace data as the target system emits events. The
2123 communication protocol is named _LTTng live_; it is used over TCP
2126 Note that you can start the relay daemon on the target system directly.
2127 This is the setup of choice when the use case is to view events as
2128 the target system emits them without the need of a remote system.
2132 == [[using-lttng]]Instrumentation
2134 There are many examples of tracing and monitoring in our everyday life:
2136 * You have access to real-time and historical weather reports and
2137 forecasts thanks to weather stations installed around the country.
2138 * You know your heart is safe thanks to an electrocardiogram.
2139 * You make sure not to drive your car too fast and to have enough fuel
2140 to reach your destination thanks to gauges visible on your dashboard.
2142 All the previous examples have something in common: they rely on
2143 **instruments**. Without the electrodes attached to the surface of your
2144 body's skin, cardiac monitoring is futile.
2146 LTTng, as a tracer, is no different from those real life examples. If
2147 you're about to trace a software system or, in other words, record its
2148 history of execution, you better have **instrumentation points** in the
2149 subject you're tracing, that is, the actual software.
2151 Various ways were developed to instrument a piece of software for LTTng
2152 tracing. The most straightforward one is to manually place
2153 instrumentation points, called _tracepoints_, in the software's source
2154 code. It is also possible to add instrumentation points dynamically in
2155 the Linux kernel <<domain,tracing domain>>.
2157 If you're only interested in tracing the Linux kernel, your
2158 instrumentation needs are probably already covered by LTTng's built-in
2159 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2160 user application which is already instrumented for LTTng tracing.
2161 In such cases, you can skip this whole section and read the topics of
2162 the <<controlling-tracing,Tracing control>> section.
2164 Many methods are available to instrument a piece of software for LTTng
2167 * <<c-application,User space instrumentation for C and $$C++$$
2169 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2170 * <<java-application,User space Java agent>>.
2171 * <<python-application,User space Python agent>>.
2172 * <<proc-lttng-logger-abi,LTTng logger>>.
2173 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2177 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2179 The procedure to instrument a C or $$C++$$ user application with
2180 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2182 . <<tracepoint-provider,Create the source files of a tracepoint provider
2184 . <<probing-the-application-source-code,Add tracepoints to
2185 the application's source code>>.
2186 . <<building-tracepoint-providers-and-user-application,Build and link
2187 a tracepoint provider package and the user application>>.
2189 If you need quick, man:printf(3)-like instrumentation, you can skip
2190 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2193 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2194 instrument a user application with `liblttng-ust`.
2197 [[tracepoint-provider]]
2198 ==== Create the source files of a tracepoint provider package
2200 A _tracepoint provider_ is a set of compiled functions which provide
2201 **tracepoints** to an application, the type of instrumentation point
2202 supported by LTTng-UST. Those functions can emit events with
2203 user-defined fields and serialize those events as event records to one
2204 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2205 macro, which you <<probing-the-application-source-code,insert in a user
2206 application's source code>>, calls those functions.
2208 A _tracepoint provider package_ is an object file (`.o`) or a shared
2209 library (`.so`) which contains one or more tracepoint providers.
2210 Its source files are:
2212 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2213 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2215 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2216 the LTTng user space tracer, at run time.
2219 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2220 image::ust-app.png[]
2222 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2223 skip creating and using a tracepoint provider and use
2224 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2228 ===== Create a tracepoint provider header file template
2230 A _tracepoint provider header file_ contains the tracepoint
2231 definitions of a tracepoint provider.
2233 To create a tracepoint provider header file:
2235 . Start from this template:
2239 .Tracepoint provider header file template (`.h` file extension).
2241 #undef TRACEPOINT_PROVIDER
2242 #define TRACEPOINT_PROVIDER provider_name
2244 #undef TRACEPOINT_INCLUDE
2245 #define TRACEPOINT_INCLUDE "./tp.h"
2247 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2250 #include <lttng/tracepoint.h>
2253 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2254 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2259 #include <lttng/tracepoint-event.h>
2265 * `provider_name` with the name of your tracepoint provider.
2266 * `"tp.h"` with the name of your tracepoint provider header file.
2268 . Below the `#include <lttng/tracepoint.h>` line, put your
2269 <<defining-tracepoints,tracepoint definitions>>.
2271 Your tracepoint provider name must be unique amongst all the possible
2272 tracepoint provider names used on the same target system. We
2273 suggest to include the name of your project or company in the name,
2274 for example, `org_lttng_my_project_tpp`.
2276 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2277 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2278 write are the <<defining-tracepoints,tracepoint definitions>>.
2281 [[defining-tracepoints]]
2282 ===== Create a tracepoint definition
2284 A _tracepoint definition_ defines, for a given tracepoint:
2286 * Its **input arguments**. They are the macro parameters that the
2287 `tracepoint()` macro accepts for this particular tracepoint
2288 in the user application's source code.
2289 * Its **output event fields**. They are the sources of event fields
2290 that form the payload of any event that the execution of the
2291 `tracepoint()` macro emits for this particular tracepoint.
2293 You can create a tracepoint definition by using the
2294 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2296 <<tpp-header,tracepoint provider header file template>>.
2298 The syntax of the `TRACEPOINT_EVENT()` macro is:
2301 .`TRACEPOINT_EVENT()` macro syntax.
2304 /* Tracepoint provider name */
2307 /* Tracepoint name */
2310 /* Input arguments */
2315 /* Output event fields */
2324 * `provider_name` with your tracepoint provider name.
2325 * `tracepoint_name` with your tracepoint name.
2326 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2327 * `fields` with the <<tpp-def-output-fields,output event field>>
2330 This tracepoint emits events named `provider_name:tracepoint_name`.
2333 .Event name's length limitation
2335 The concatenation of the tracepoint provider name and the
2336 tracepoint name must not exceed **254 characters**. If it does, the
2337 instrumented application compiles and runs, but LTTng throws multiple
2338 warnings and you could experience serious issues.
2341 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2344 .`TP_ARGS()` macro syntax.
2353 * `type` with the C type of the argument.
2354 * `arg_name` with the argument name.
2356 You can repeat `type` and `arg_name` up to 10 times to have
2357 more than one argument.
2359 .`TP_ARGS()` usage with three arguments.
2371 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2372 tracepoint definition with no input arguments.
2374 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2375 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2376 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2377 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2380 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2381 C expression that the tracer evalutes at the `tracepoint()` macro site
2382 in the application's source code. This expression provides a field's
2383 source of data. The argument expression can include input argument names
2384 listed in the `TP_ARGS()` macro.
2386 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2387 must be unique within a given tracepoint definition.
2389 Here's a complete tracepoint definition example:
2391 .Tracepoint definition.
2393 The following tracepoint definition defines a tracepoint which takes
2394 three input arguments and has four output event fields.
2398 #include "my-custom-structure.h"
2404 const struct my_custom_structure*, my_custom_structure,
2409 ctf_string(query_field, query)
2410 ctf_float(double, ratio_field, ratio)
2411 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2412 ctf_integer(int, send_size, my_custom_structure->send_size)
2417 You can refer to this tracepoint definition with the `tracepoint()`
2418 macro in your application's source code like this:
2422 tracepoint(my_provider, my_tracepoint,
2423 my_structure, some_ratio, the_query);
2427 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2428 if they satisfy an enabled <<event,event rule>>.
2431 [[using-tracepoint-classes]]
2432 ===== Use a tracepoint class
2434 A _tracepoint class_ is a class of tracepoints which share the same
2435 output event field definitions. A _tracepoint instance_ is one
2436 instance of such a defined tracepoint class, with its own tracepoint
2439 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2440 shorthand which defines both a tracepoint class and a tracepoint
2441 instance at the same time.
2443 When you build a tracepoint provider package, the C or $$C++$$ compiler
2444 creates one serialization function for each **tracepoint class**. A
2445 serialization function is responsible for serializing the event fields
2446 of a tracepoint to a sub-buffer when tracing.
2448 For various performance reasons, when your situation requires multiple
2449 tracepoint definitions with different names, but with the same event
2450 fields, we recommend that you manually create a tracepoint class
2451 and instantiate as many tracepoint instances as needed. One positive
2452 effect of such a design, amongst other advantages, is that all
2453 tracepoint instances of the same tracepoint class reuse the same
2454 serialization function, thus reducing
2455 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2457 .Use a tracepoint class and tracepoint instances.
2459 Consider the following three tracepoint definitions:
2471 ctf_integer(int, userid, userid)
2472 ctf_integer(size_t, len, len)
2484 ctf_integer(int, userid, userid)
2485 ctf_integer(size_t, len, len)
2497 ctf_integer(int, userid, userid)
2498 ctf_integer(size_t, len, len)
2503 In this case, we create three tracepoint classes, with one implicit
2504 tracepoint instance for each of them: `get_account`, `get_settings`, and
2505 `get_transaction`. However, they all share the same event field names
2506 and types. Hence three identical, yet independent serialization
2507 functions are created when you build the tracepoint provider package.
2509 A better design choice is to define a single tracepoint class and three
2510 tracepoint instances:
2514 /* The tracepoint class */
2515 TRACEPOINT_EVENT_CLASS(
2516 /* Tracepoint provider name */
2519 /* Tracepoint class name */
2522 /* Input arguments */
2528 /* Output event fields */
2530 ctf_integer(int, userid, userid)
2531 ctf_integer(size_t, len, len)
2535 /* The tracepoint instances */
2536 TRACEPOINT_EVENT_INSTANCE(
2537 /* Tracepoint provider name */
2540 /* Tracepoint class name */
2543 /* Tracepoint name */
2546 /* Input arguments */
2552 TRACEPOINT_EVENT_INSTANCE(
2561 TRACEPOINT_EVENT_INSTANCE(
2574 [[assigning-log-levels]]
2575 ===== Assign a log level to a tracepoint definition
2577 You can assign an optional _log level_ to a
2578 <<defining-tracepoints,tracepoint definition>>.
2580 Assigning different levels of severity to tracepoint definitions can
2581 be useful: when you <<enabling-disabling-events,create an event rule>>,
2582 you can target tracepoints having a log level as severe as a specific
2585 The concept of LTTng-UST log levels is similar to the levels found
2586 in typical logging frameworks:
2588 * In a logging framework, the log level is given by the function
2589 or method name you use at the log statement site: `debug()`,
2590 `info()`, `warn()`, `error()`, and so on.
2591 * In LTTng-UST, you statically assign the log level to a tracepoint
2592 definition; any `tracepoint()` macro invocation which refers to
2593 this definition has this log level.
2595 You can assign a log level to a tracepoint definition with the
2596 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2597 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2598 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2601 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2604 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2606 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2611 * `provider_name` with the tracepoint provider name.
2612 * `tracepoint_name` with the tracepoint name.
2613 * `log_level` with the log level to assign to the tracepoint
2614 definition named `tracepoint_name` in the `provider_name`
2615 tracepoint provider.
2617 See man:lttng-ust(3) for a list of available log level names.
2619 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2623 /* Tracepoint definition */
2632 ctf_integer(int, userid, userid)
2633 ctf_integer(size_t, len, len)
2637 /* Log level assignment */
2638 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2644 ===== Create a tracepoint provider package source file
2646 A _tracepoint provider package source file_ is a C source file which
2647 includes a <<tpp-header,tracepoint provider header file>> to expand its
2648 macros into event serialization and other functions.
2650 You can always use the following tracepoint provider package source
2654 .Tracepoint provider package source file template.
2656 #define TRACEPOINT_CREATE_PROBES
2661 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2662 header file>> name. You may also include more than one tracepoint
2663 provider header file here to create a tracepoint provider package
2664 holding more than one tracepoint providers.
2667 [[probing-the-application-source-code]]
2668 ==== Add tracepoints to an application's source code
2670 Once you <<tpp-header,create a tracepoint provider header file>>, you
2671 can use the `tracepoint()` macro in your application's
2672 source code to insert the tracepoints that this header
2673 <<defining-tracepoints,defines>>.
2675 The `tracepoint()` macro takes at least two parameters: the tracepoint
2676 provider name and the tracepoint name. The corresponding tracepoint
2677 definition defines the other parameters.
2679 .`tracepoint()` usage.
2681 The following <<defining-tracepoints,tracepoint definition>> defines a
2682 tracepoint which takes two input arguments and has two output event
2686 .Tracepoint provider header file.
2688 #include "my-custom-structure.h"
2695 const char*, cmd_name
2698 ctf_string(cmd_name, cmd_name)
2699 ctf_integer(int, number_of_args, argc)
2704 You can refer to this tracepoint definition with the `tracepoint()`
2705 macro in your application's source code like this:
2708 .Application's source file.
2712 int main(int argc, char* argv[])
2714 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2720 Note how the application's source code includes
2721 the tracepoint provider header file containing the tracepoint
2722 definitions to use, path:{tp.h}.
2725 .`tracepoint()` usage with a complex tracepoint definition.
2727 Consider this complex tracepoint definition, where multiple event
2728 fields refer to the same input arguments in their argument expression
2732 .Tracepoint provider header file.
2734 /* For `struct stat` */
2735 #include <sys/types.h>
2736 #include <sys/stat.h>
2748 ctf_integer(int, my_constant_field, 23 + 17)
2749 ctf_integer(int, my_int_arg_field, my_int_arg)
2750 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2751 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2752 my_str_arg[2] + my_str_arg[3])
2753 ctf_string(my_str_arg_field, my_str_arg)
2754 ctf_integer_hex(off_t, size_field, st->st_size)
2755 ctf_float(double, size_dbl_field, (double) st->st_size)
2756 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2757 size_t, strlen(my_str_arg) / 2)
2762 You can refer to this tracepoint definition with the `tracepoint()`
2763 macro in your application's source code like this:
2766 .Application's source file.
2768 #define TRACEPOINT_DEFINE
2775 stat("/etc/fstab", &s);
2776 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2782 If you look at the event record that LTTng writes when tracing this
2783 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2784 it should look like this:
2786 .Event record fields
2788 |Field's name |Field's value
2789 |`my_constant_field` |40
2790 |`my_int_arg_field` |23
2791 |`my_int_arg_field2` |529
2793 |`my_str_arg_field` |`Hello, World!`
2794 |`size_field` |0x12d
2795 |`size_dbl_field` |301.0
2796 |`half_my_str_arg_field` |`Hello,`
2800 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2801 compute--they use the call stack, for example. To avoid this
2802 computation when the tracepoint is disabled, you can use the
2803 `tracepoint_enabled()` and `do_tracepoint()` macros.
2805 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2809 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2811 tracepoint_enabled(provider_name, tracepoint_name)
2812 do_tracepoint(provider_name, tracepoint_name, ...)
2817 * `provider_name` with the tracepoint provider name.
2818 * `tracepoint_name` with the tracepoint name.
2820 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2821 `tracepoint_name` from the provider named `provider_name` is enabled
2824 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2825 if the tracepoint is enabled. Using `tracepoint()` with
2826 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2827 the `tracepoint_enabled()` check, thus a race condition is
2828 possible in this situation:
2831 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2833 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2834 stuff = prepare_stuff();
2837 tracepoint(my_provider, my_tracepoint, stuff);
2840 If the tracepoint is enabled after the condition, then `stuff` is not
2841 prepared: the emitted event will either contain wrong data, or the whole
2842 application could crash (segmentation fault, for example).
2844 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2845 `STAP_PROBEV()` call. If you need it, you must emit
2849 [[building-tracepoint-providers-and-user-application]]
2850 ==== Build and link a tracepoint provider package and an application
2852 Once you have one or more <<tpp-header,tracepoint provider header
2853 files>> and a <<tpp-source,tracepoint provider package source file>>,
2854 you can create the tracepoint provider package by compiling its source
2855 file. From here, multiple build and run scenarios are possible. The
2856 following table shows common application and library configurations
2857 along with the required command lines to achieve them.
2859 In the following diagrams, we use the following file names:
2862 Executable application.
2865 Application's object file.
2868 Tracepoint provider package object file.
2871 Tracepoint provider package archive file.
2874 Tracepoint provider package shared object file.
2877 User library object file.
2880 User library shared object file.
2882 We use the following symbols in the diagrams of table below:
2885 .Symbols used in the build scenario diagrams.
2886 image::ust-sit-symbols.png[]
2888 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2889 variable in the following instructions.
2891 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2892 .Common tracepoint provider package scenarios.
2894 |Scenario |Instructions
2897 The instrumented application is statically linked with
2898 the tracepoint provider package object.
2900 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2903 include::../common/ust-sit-step-tp-o.txt[]
2905 To build the instrumented application:
2907 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2912 #define TRACEPOINT_DEFINE
2916 . Compile the application source file:
2925 . Build the application:
2930 gcc -o app app.o tpp.o -llttng-ust -ldl
2934 To run the instrumented application:
2936 * Start the application:
2946 The instrumented application is statically linked with the
2947 tracepoint provider package archive file.
2949 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2952 To create the tracepoint provider package archive file:
2954 . Compile the <<tpp-source,tracepoint provider package source file>>:
2963 . Create the tracepoint provider package archive file:
2972 To build the instrumented application:
2974 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2979 #define TRACEPOINT_DEFINE
2983 . Compile the application source file:
2992 . Build the application:
2997 gcc -o app app.o tpp.a -llttng-ust -ldl
3001 To run the instrumented application:
3003 * Start the application:
3013 The instrumented application is linked with the tracepoint provider
3014 package shared object.
3016 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
3019 include::../common/ust-sit-step-tp-so.txt[]
3021 To build the instrumented application:
3023 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3028 #define TRACEPOINT_DEFINE
3032 . Compile the application source file:
3041 . Build the application:
3046 gcc -o app app.o -ldl -L. -ltpp
3050 To run the instrumented application:
3052 * Start the application:
3062 The tracepoint provider package shared object is preloaded before the
3063 instrumented application starts.
3065 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3068 include::../common/ust-sit-step-tp-so.txt[]
3070 To build the instrumented application:
3072 . In path:{app.c}, before including path:{tpp.h}, add the
3078 #define TRACEPOINT_DEFINE
3079 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3083 . Compile the application source file:
3092 . Build the application:
3097 gcc -o app app.o -ldl
3101 To run the instrumented application with tracing support:
3103 * Preload the tracepoint provider package shared object and
3104 start the application:
3109 LD_PRELOAD=./libtpp.so ./app
3113 To run the instrumented application without tracing support:
3115 * Start the application:
3125 The instrumented application dynamically loads the tracepoint provider
3126 package shared object.
3128 See the <<dlclose-warning,warning about `dlclose()`>>.
3130 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3133 include::../common/ust-sit-step-tp-so.txt[]
3135 To build the instrumented application:
3137 . In path:{app.c}, before including path:{tpp.h}, add the
3143 #define TRACEPOINT_DEFINE
3144 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3148 . Compile the application source file:
3157 . Build the application:
3162 gcc -o app app.o -ldl
3166 To run the instrumented application:
3168 * Start the application:
3178 The application is linked with the instrumented user library.
3180 The instrumented user library is statically linked with the tracepoint
3181 provider package object file.
3183 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3186 include::../common/ust-sit-step-tp-o-fpic.txt[]
3188 To build the instrumented user library:
3190 . In path:{emon.c}, before including path:{tpp.h}, add the
3196 #define TRACEPOINT_DEFINE
3200 . Compile the user library source file:
3205 gcc -I. -fpic -c emon.c
3209 . Build the user library shared object:
3214 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3218 To build the application:
3220 . Compile the application source file:
3229 . Build the application:
3234 gcc -o app app.o -L. -lemon
3238 To run the application:
3240 * Start the application:
3250 The application is linked with the instrumented user library.
3252 The instrumented user library is linked with the tracepoint provider
3253 package shared object.
3255 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3258 include::../common/ust-sit-step-tp-so.txt[]
3260 To build the instrumented user library:
3262 . In path:{emon.c}, before including path:{tpp.h}, add the
3268 #define TRACEPOINT_DEFINE
3272 . Compile the user library source file:
3277 gcc -I. -fpic -c emon.c
3281 . Build the user library shared object:
3286 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3290 To build the application:
3292 . Compile the application source file:
3301 . Build the application:
3306 gcc -o app app.o -L. -lemon
3310 To run the application:
3312 * Start the application:
3322 The tracepoint provider package shared object is preloaded before the
3325 The application is linked with the instrumented user library.
3327 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3330 include::../common/ust-sit-step-tp-so.txt[]
3332 To build the instrumented user library:
3334 . In path:{emon.c}, before including path:{tpp.h}, add the
3340 #define TRACEPOINT_DEFINE
3341 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3345 . Compile the user library source file:
3350 gcc -I. -fpic -c emon.c
3354 . Build the user library shared object:
3359 gcc -shared -o libemon.so emon.o -ldl
3363 To build the application:
3365 . Compile the application source file:
3374 . Build the application:
3379 gcc -o app app.o -L. -lemon
3383 To run the application with tracing support:
3385 * Preload the tracepoint provider package shared object and
3386 start the application:
3391 LD_PRELOAD=./libtpp.so ./app
3395 To run the application without tracing support:
3397 * Start the application:
3407 The application is linked with the instrumented user library.
3409 The instrumented user library dynamically loads the tracepoint provider
3410 package shared object.
3412 See the <<dlclose-warning,warning about `dlclose()`>>.
3414 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3417 include::../common/ust-sit-step-tp-so.txt[]
3419 To build the instrumented user library:
3421 . In path:{emon.c}, before including path:{tpp.h}, add the
3427 #define TRACEPOINT_DEFINE
3428 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3432 . Compile the user library source file:
3437 gcc -I. -fpic -c emon.c
3441 . Build the user library shared object:
3446 gcc -shared -o libemon.so emon.o -ldl
3450 To build the application:
3452 . Compile the application source file:
3461 . Build the application:
3466 gcc -o app app.o -L. -lemon
3470 To run the application:
3472 * Start the application:
3482 The application dynamically loads the instrumented user library.
3484 The instrumented user library is linked with the tracepoint provider
3485 package shared object.
3487 See the <<dlclose-warning,warning about `dlclose()`>>.
3489 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3492 include::../common/ust-sit-step-tp-so.txt[]
3494 To build the instrumented user library:
3496 . In path:{emon.c}, before including path:{tpp.h}, add the
3502 #define TRACEPOINT_DEFINE
3506 . Compile the user library source file:
3511 gcc -I. -fpic -c emon.c
3515 . Build the user library shared object:
3520 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3524 To build the application:
3526 . Compile the application source file:
3535 . Build the application:
3540 gcc -o app app.o -ldl -L. -lemon
3544 To run the application:
3546 * Start the application:
3556 The application dynamically loads the instrumented user library.
3558 The instrumented user library dynamically loads the tracepoint provider
3559 package shared object.
3561 See the <<dlclose-warning,warning about `dlclose()`>>.
3563 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3566 include::../common/ust-sit-step-tp-so.txt[]
3568 To build the instrumented user library:
3570 . In path:{emon.c}, before including path:{tpp.h}, add the
3576 #define TRACEPOINT_DEFINE
3577 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3581 . Compile the user library source file:
3586 gcc -I. -fpic -c emon.c
3590 . Build the user library shared object:
3595 gcc -shared -o libemon.so emon.o -ldl
3599 To build the application:
3601 . Compile the application source file:
3610 . Build the application:
3615 gcc -o app app.o -ldl -L. -lemon
3619 To run the application:
3621 * Start the application:
3631 The tracepoint provider package shared object is preloaded before the
3634 The application dynamically loads the instrumented user library.
3636 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3639 include::../common/ust-sit-step-tp-so.txt[]
3641 To build the instrumented user library:
3643 . In path:{emon.c}, before including path:{tpp.h}, add the
3649 #define TRACEPOINT_DEFINE
3650 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3654 . Compile the user library source file:
3659 gcc -I. -fpic -c emon.c
3663 . Build the user library shared object:
3668 gcc -shared -o libemon.so emon.o -ldl
3672 To build the application:
3674 . Compile the application source file:
3683 . Build the application:
3688 gcc -o app app.o -L. -lemon
3692 To run the application with tracing support:
3694 * Preload the tracepoint provider package shared object and
3695 start the application:
3700 LD_PRELOAD=./libtpp.so ./app
3704 To run the application without tracing support:
3706 * Start the application:
3716 The application is statically linked with the tracepoint provider
3717 package object file.
3719 The application is linked with the instrumented user library.
3721 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3724 include::../common/ust-sit-step-tp-o.txt[]
3726 To build the instrumented user library:
3728 . In path:{emon.c}, before including path:{tpp.h}, add the
3734 #define TRACEPOINT_DEFINE
3738 . Compile the user library source file:
3743 gcc -I. -fpic -c emon.c
3747 . Build the user library shared object:
3752 gcc -shared -o libemon.so emon.o
3756 To build the application:
3758 . Compile the application source file:
3767 . Build the application:
3772 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3776 To run the instrumented application:
3778 * Start the application:
3788 The application is statically linked with the tracepoint provider
3789 package object file.
3791 The application dynamically loads the instrumented user library.
3793 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3796 include::../common/ust-sit-step-tp-o.txt[]
3798 To build the application:
3800 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3805 #define TRACEPOINT_DEFINE
3809 . Compile the application source file:
3818 . Build the application:
3823 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3828 The `--export-dynamic` option passed to the linker is necessary for the
3829 dynamically loaded library to ``see'' the tracepoint symbols defined in
3832 To build the instrumented user library:
3834 . Compile the user library source file:
3839 gcc -I. -fpic -c emon.c
3843 . Build the user library shared object:
3848 gcc -shared -o libemon.so emon.o
3852 To run the application:
3854 * Start the application:
3866 .Do not use man:dlclose(3) on a tracepoint provider package
3868 Never use man:dlclose(3) on any shared object which:
3870 * Is linked with, statically or dynamically, a tracepoint provider
3872 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3873 package shared object.
3875 This is currently considered **unsafe** due to a lack of reference
3876 counting from LTTng-UST to the shared object.
3878 A known workaround (available since glibc 2.2) is to use the
3879 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3880 effect of not unloading the loaded shared object, even if man:dlclose(3)
3883 You can also preload the tracepoint provider package shared object with
3884 the env:LD_PRELOAD environment variable to overcome this limitation.
3888 [[using-lttng-ust-with-daemons]]
3889 ===== Use noch:{LTTng-UST} with daemons
3891 If your instrumented application calls man:fork(2), man:clone(2),
3892 or BSD's man:rfork(2), without a following man:exec(3)-family
3893 system call, you must preload the path:{liblttng-ust-fork.so} shared
3894 object when starting the application.
3898 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3901 If your tracepoint provider package is
3902 a shared library which you also preload, you must put both
3903 shared objects in env:LD_PRELOAD:
3907 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3911 [[lttng-ust-pkg-config]]
3912 ===== Use noch:{pkg-config}
3914 On some distributions, LTTng-UST ships with a
3915 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3916 metadata file. If this is your case, then you can use cmd:pkg-config to
3917 build an application on the command line:
3921 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3925 [[instrumenting-32-bit-app-on-64-bit-system]]
3926 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3928 In order to trace a 32-bit application running on a 64-bit system,
3929 LTTng must use a dedicated 32-bit
3930 <<lttng-consumerd,consumer daemon>>.
3932 The following steps show how to build and install a 32-bit consumer
3933 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3934 build and install the 32-bit LTTng-UST libraries, and how to build and
3935 link an instrumented 32-bit application in that context.
3937 To build a 32-bit instrumented application for a 64-bit target system,
3938 assuming you have a fresh target system with no installed Userspace RCU
3941 . Download, build, and install a 32-bit version of Userspace RCU:
3947 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3948 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3949 cd userspace-rcu-0.9.* &&
3950 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3952 sudo make install &&
3957 . Using your distribution's package manager, or from source, install
3958 the following 32-bit versions of the following dependencies of
3959 LTTng-tools and LTTng-UST:
3962 * https://sourceforge.net/projects/libuuid/[libuuid]
3963 * http://directory.fsf.org/wiki/Popt[popt]
3964 * http://www.xmlsoft.org/[libxml2]
3967 . Download, build, and install a 32-bit version of the latest
3968 LTTng-UST{nbsp}{revision}:
3974 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
3975 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
3976 cd lttng-ust-2.8.* &&
3977 ./configure --libdir=/usr/local/lib32 \
3978 CFLAGS=-m32 CXXFLAGS=-m32 \
3979 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3981 sudo make install &&
3988 Depending on your distribution,
3989 32-bit libraries could be installed at a different location than
3990 `/usr/lib32`. For example, Debian is known to install
3991 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3993 In this case, make sure to set `LDFLAGS` to all the
3994 relevant 32-bit library paths, for example:
3998 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
4002 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
4003 the 32-bit consumer daemon:
4009 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4010 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4011 cd lttng-tools-2.8.* &&
4012 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
4013 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4015 cd src/bin/lttng-consumerd &&
4016 sudo make install &&
4021 . From your distribution or from source,
4022 <<installing-lttng,install>> the 64-bit versions of
4023 LTTng-UST and Userspace RCU.
4024 . Download, build, and install the 64-bit version of the
4025 latest LTTng-tools{nbsp}{revision}:
4031 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4032 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4033 cd lttng-tools-2.8.* &&
4034 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
4035 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
4037 sudo make install &&
4042 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
4043 when linking your 32-bit application:
4046 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4047 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4050 For example, let's rebuild the quick start example in
4051 <<tracing-your-own-user-application,Trace a user application>> as an
4052 instrumented 32-bit application:
4057 gcc -m32 -c -I. hello-tp.c
4059 gcc -m32 -o hello hello.o hello-tp.o \
4060 -L/usr/lib32 -L/usr/local/lib32 \
4061 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4066 No special action is required to execute the 32-bit application and
4067 to trace it: use the command-line man:lttng(1) tool as usual.
4074 man:tracef(3) is a small LTTng-UST API designed for quick,
4075 man:printf(3)-like instrumentation without the burden of
4076 <<tracepoint-provider,creating>> and
4077 <<building-tracepoint-providers-and-user-application,building>>
4078 a tracepoint provider package.
4080 To use `tracef()` in your application:
4082 . In the C or C++ source files where you need to use `tracef()`,
4083 include `<lttng/tracef.h>`:
4088 #include <lttng/tracef.h>
4092 . In the application's source code, use `tracef()` like you would use
4100 tracef("my message: %d (%s)", my_integer, my_string);
4106 . Link your application with `liblttng-ust`:
4111 gcc -o app app.c -llttng-ust
4115 To trace the events that `tracef()` calls emit:
4117 * <<enabling-disabling-events,Create an event rule>> which matches the
4118 `lttng_ust_tracef:*` event name:
4123 lttng enable-event --userspace 'lttng_ust_tracef:*'
4128 .Limitations of `tracef()`
4130 The `tracef()` utility function was developed to make user space tracing
4131 super simple, albeit with notable disadvantages compared to
4132 <<defining-tracepoints,user-defined tracepoints>>:
4134 * All the emitted events have the same tracepoint provider and
4135 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4136 * There is no static type checking.
4137 * The only event record field you actually get, named `msg`, is a string
4138 potentially containing the values you passed to `tracef()`
4139 using your own format string. This also means that you cannot filter
4140 events with a custom expression at run time because there are no
4142 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4143 function behind the scenes to format the strings at run time, its
4144 expected performance is lower than with user-defined tracepoints,
4145 which do not require a conversion to a string.
4147 Taking this into consideration, `tracef()` is useful for some quick
4148 prototyping and debugging, but you should not consider it for any
4149 permanent and serious applicative instrumentation.
4155 ==== Use `tracelog()`
4157 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4158 the difference that it accepts an additional log level parameter.
4160 The goal of `tracelog()` is to ease the migration from logging to
4163 To use `tracelog()` in your application:
4165 . In the C or C++ source files where you need to use `tracelog()`,
4166 include `<lttng/tracelog.h>`:
4171 #include <lttng/tracelog.h>
4175 . In the application's source code, use `tracelog()` like you would use
4176 man:printf(3), except for the first parameter which is the log
4184 tracelog(TRACE_WARNING, "my message: %d (%s)",
4185 my_integer, my_string);
4191 See man:lttng-ust(3) for a list of available log level names.
4193 . Link your application with `liblttng-ust`:
4198 gcc -o app app.c -llttng-ust
4202 To trace the events that `tracelog()` calls emit with a log level
4203 _as severe as_ a specific log level:
4205 * <<enabling-disabling-events,Create an event rule>> which matches the
4206 `lttng_ust_tracelog:*` event name and a minimum level
4212 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4213 --loglevel=TRACE_WARNING
4217 To trace the events that `tracelog()` calls emit with a
4218 _specific log level_:
4220 * Create an event rule which matches the `lttng_ust_tracelog:*`
4221 event name and a specific log level:
4226 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4227 --loglevel-only=TRACE_INFO
4232 [[prebuilt-ust-helpers]]
4233 === Prebuilt user space tracing helpers
4235 The LTTng-UST package provides a few helpers in the form or preloadable
4236 shared objects which automatically instrument system functions and
4239 The helper shared objects are normally found in dir:{/usr/lib}. If you
4240 built LTTng-UST <<building-from-source,from source>>, they are probably
4241 located in dir:{/usr/local/lib}.
4243 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4246 path:{liblttng-ust-libc-wrapper.so}::
4247 path:{liblttng-ust-pthread-wrapper.so}::
4248 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4249 memory and POSIX threads function tracing>>.
4251 path:{liblttng-ust-cyg-profile.so}::
4252 path:{liblttng-ust-cyg-profile-fast.so}::
4253 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4255 path:{liblttng-ust-dl.so}::
4256 <<liblttng-ust-dl,Dynamic linker tracing>>.
4258 To use a user space tracing helper with any user application:
4260 * Preload the helper shared object when you start the application:
4265 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4269 You can preload more than one helper:
4274 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4280 [[liblttng-ust-libc-pthread-wrapper]]
4281 ==== Instrument C standard library memory and POSIX threads functions
4283 The path:{liblttng-ust-libc-wrapper.so} and
4284 path:{liblttng-ust-pthread-wrapper.so} helpers
4285 add instrumentation to some C standard library and POSIX
4289 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4291 |TP provider name |TP name |Instrumented function
4293 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4294 |`calloc` |man:calloc(3)
4295 |`realloc` |man:realloc(3)
4296 |`free` |man:free(3)
4297 |`memalign` |man:memalign(3)
4298 |`posix_memalign` |man:posix_memalign(3)
4302 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4304 |TP provider name |TP name |Instrumented function
4306 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4307 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4308 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4309 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4312 When you preload the shared object, it replaces the functions listed
4313 in the previous tables by wrappers which contain tracepoints and call
4314 the replaced functions.
4317 [[liblttng-ust-cyg-profile]]
4318 ==== Instrument function entry and exit
4320 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4321 to the entry and exit points of functions.
4323 man:gcc(1) and man:clang(1) have an option named
4324 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4325 which generates instrumentation calls for entry and exit to functions.
4326 The LTTng-UST function tracing helpers,
4327 path:{liblttng-ust-cyg-profile.so} and
4328 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4329 to add tracepoints to the two generated functions (which contain
4330 `cyg_profile` in their names, hence the helper's name).
4332 To use the LTTng-UST function tracing helper, the source files to
4333 instrument must be built using the `-finstrument-functions` compiler
4336 There are two versions of the LTTng-UST function tracing helper:
4338 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4339 that you should only use when it can be _guaranteed_ that the
4340 complete event stream is recorded without any lost event record.
4341 Any kind of duplicate information is left out.
4343 Assuming no event record is lost, having only the function addresses on
4344 entry is enough to create a call graph, since an event record always
4345 contains the ID of the CPU that generated it.
4347 You can use a tool like man:addr2line(1) to convert function addresses
4348 back to source file names and line numbers.
4350 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4351 which also works in use cases where event records might get discarded or
4352 not recorded from application startup.
4353 In these cases, the trace analyzer needs more information to be
4354 able to reconstruct the program flow.
4356 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4357 points of this helper.
4359 All the tracepoints that this helper provides have the
4360 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4362 TIP: It's sometimes a good idea to limit the number of source files that
4363 you compile with the `-finstrument-functions` option to prevent LTTng
4364 from writing an excessive amount of trace data at run time. When using
4365 man:gcc(1), you can use the
4366 `-finstrument-functions-exclude-function-list` option to avoid
4367 instrument entries and exits of specific function names.
4372 ==== Instrument the dynamic linker
4374 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4375 man:dlopen(3) and man:dlclose(3) function calls.
4377 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4382 [[java-application]]
4383 === User space Java agent
4385 You can instrument any Java application which uses one of the following
4388 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4389 (JUL) core logging facilities.
4390 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4391 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4394 .LTTng-UST Java agent imported by a Java application.
4395 image::java-app.png[]
4397 Note that the methods described below are new in LTTng{nbsp}{revision}.
4398 Previous LTTng versions use another technique.
4400 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4401 and https://ci.lttng.org/[continuous integration], thus this version is
4402 directly supported. However, the LTTng-UST Java agent is also tested
4403 with OpenJDK{nbsp}7.
4408 ==== Use the LTTng-UST Java agent for `java.util.logging`
4410 To use the LTTng-UST Java agent in a Java application which uses
4411 `java.util.logging` (JUL):
4413 . In the Java application's source code, import the LTTng-UST
4414 log handler package for `java.util.logging`:
4419 import org.lttng.ust.agent.jul.LttngLogHandler;
4423 . Create an LTTng-UST JUL log handler:
4428 Handler lttngUstLogHandler = new LttngLogHandler();
4432 . Add this handler to the JUL loggers which should emit LTTng events:
4437 Logger myLogger = Logger.getLogger("some-logger");
4439 myLogger.addHandler(lttngUstLogHandler);
4443 . Use `java.util.logging` log statements and configuration as usual.
4444 The loggers with an attached LTTng-UST log handler can emit
4447 . Before exiting the application, remove the LTTng-UST log handler from
4448 the loggers attached to it and call its `close()` method:
4453 myLogger.removeHandler(lttngUstLogHandler);
4454 lttngUstLogHandler.close();
4458 This is not strictly necessary, but it is recommended for a clean
4459 disposal of the handler's resources.
4461 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4462 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4464 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4465 path] when you build the Java application.
4467 The JAR files are typically located in dir:{/usr/share/java}.
4469 IMPORTANT: The LTTng-UST Java agent must be
4470 <<installing-lttng,installed>> for the logging framework your
4473 .Use the LTTng-UST Java agent for `java.util.logging`.
4478 import java.io.IOException;
4479 import java.util.logging.Handler;
4480 import java.util.logging.Logger;
4481 import org.lttng.ust.agent.jul.LttngLogHandler;
4485 private static final int answer = 42;
4487 public static void main(String[] argv) throws Exception
4490 Logger logger = Logger.getLogger("jello");
4492 // Create an LTTng-UST log handler
4493 Handler lttngUstLogHandler = new LttngLogHandler();
4495 // Add the LTTng-UST log handler to our logger
4496 logger.addHandler(lttngUstLogHandler);
4499 logger.info("some info");
4500 logger.warning("some warning");
4502 logger.finer("finer information; the answer is " + answer);
4504 logger.severe("error!");
4506 // Not mandatory, but cleaner
4507 logger.removeHandler(lttngUstLogHandler);
4508 lttngUstLogHandler.close();
4517 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4520 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4521 <<enabling-disabling-events,create an event rule>> matching the
4522 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4527 lttng enable-event --jul jello
4531 Run the compiled class:
4535 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4538 <<basic-tracing-session-control,Stop tracing>> and inspect the
4548 You can use the opt:lttng-enable-event(1):--loglevel or
4549 opt:lttng-enable-event(1):--loglevel-only option of the
4550 man:lttng-enable-event(1) command to target a range of JUL log levels
4551 or a specific JUL log level.
4556 ==== Use the LTTng-UST Java agent for Apache log4j
4558 To use the LTTng-UST Java agent in a Java application which uses
4561 . In the Java application's source code, import the LTTng-UST
4562 log appender package for Apache log4j:
4567 import org.lttng.ust.agent.log4j.LttngLogAppender;
4571 . Create an LTTng-UST log4j log appender:
4576 Appender lttngUstLogAppender = new LttngLogAppender();
4580 . Add this appender to the log4j loggers which should emit LTTng events:
4585 Logger myLogger = Logger.getLogger("some-logger");
4587 myLogger.addAppender(lttngUstLogAppender);
4591 . Use Apache log4j log statements and configuration as usual. The
4592 loggers with an attached LTTng-UST log appender can emit LTTng events.
4594 . Before exiting the application, remove the LTTng-UST log appender from
4595 the loggers attached to it and call its `close()` method:
4600 myLogger.removeAppender(lttngUstLogAppender);
4601 lttngUstLogAppender.close();
4605 This is not strictly necessary, but it is recommended for a clean
4606 disposal of the appender's resources.
4608 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4609 files, path:{lttng-ust-agent-common.jar} and
4610 path:{lttng-ust-agent-log4j.jar}, in the
4611 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4612 path] when you build the Java application.
4614 The JAR files are typically located in dir:{/usr/share/java}.
4616 IMPORTANT: The LTTng-UST Java agent must be
4617 <<installing-lttng,installed>> for the logging framework your
4620 .Use the LTTng-UST Java agent for Apache log4j.
4625 import org.apache.log4j.Appender;
4626 import org.apache.log4j.Logger;
4627 import org.lttng.ust.agent.log4j.LttngLogAppender;
4631 private static final int answer = 42;
4633 public static void main(String[] argv) throws Exception
4636 Logger logger = Logger.getLogger("jello");
4638 // Create an LTTng-UST log appender
4639 Appender lttngUstLogAppender = new LttngLogAppender();
4641 // Add the LTTng-UST log appender to our logger
4642 logger.addAppender(lttngUstLogAppender);
4645 logger.info("some info");
4646 logger.warn("some warning");
4648 logger.debug("debug information; the answer is " + answer);
4650 logger.fatal("error!");
4652 // Not mandatory, but cleaner
4653 logger.removeAppender(lttngUstLogAppender);
4654 lttngUstLogAppender.close();
4660 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4665 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4668 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4669 <<enabling-disabling-events,create an event rule>> matching the
4670 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4675 lttng enable-event --log4j jello
4679 Run the compiled class:
4683 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4686 <<basic-tracing-session-control,Stop tracing>> and inspect the
4696 You can use the opt:lttng-enable-event(1):--loglevel or
4697 opt:lttng-enable-event(1):--loglevel-only option of the
4698 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4699 or a specific log4j log level.
4703 [[java-application-context]]
4704 ==== Provide application-specific context fields in a Java application
4706 A Java application-specific context field is a piece of state provided
4707 by the application which <<adding-context,you can add>>, using the
4708 man:lttng-add-context(1) command, to each <<event,event record>>
4709 produced by the log statements of this application.
4711 For example, a given object might have a current request ID variable.
4712 You can create a context information retriever for this object and
4713 assign a name to this current request ID. You can then, using the
4714 man:lttng-add-context(1) command, add this context field by name to
4715 the JUL or log4j <<channel,channel>>.
4717 To provide application-specific context fields in a Java application:
4719 . In the Java application's source code, import the LTTng-UST
4720 Java agent context classes and interfaces:
4725 import org.lttng.ust.agent.context.ContextInfoManager;
4726 import org.lttng.ust.agent.context.IContextInfoRetriever;
4730 . Create a context information retriever class, that is, a class which
4731 implements the `IContextInfoRetriever` interface:
4736 class MyContextInfoRetriever implements IContextInfoRetriever
4739 public Object retrieveContextInfo(String key)
4741 if (key.equals("intCtx")) {
4743 } else if (key.equals("strContext")) {
4744 return "context value!";
4753 This `retrieveContextInfo()` method is the only member of the
4754 `IContextInfoRetriever` interface. Its role is to return the current
4755 value of a state by name to create a context field. The names of the
4756 context fields and which state variables they return depends on your
4759 All primitive types and objects are supported as context fields.
4760 When `retrieveContextInfo()` returns an object, the context field
4761 serializer calls its `toString()` method to add a string field to
4762 event records. The method can also return `null`, which means that
4763 no context field is available for the required name.
4765 . Register an instance of your context information retriever class to
4766 the context information manager singleton:
4771 IContextInfoRetriever cir = new MyContextInfoRetriever();
4772 ContextInfoManager cim = ContextInfoManager.getInstance();
4773 cim.registerContextInfoRetriever("retrieverName", cir);
4777 . Before exiting the application, remove your context information
4778 retriever from the context information manager singleton:
4783 ContextInfoManager cim = ContextInfoManager.getInstance();
4784 cim.unregisterContextInfoRetriever("retrieverName");
4788 This is not strictly necessary, but it is recommended for a clean
4789 disposal of some manager's resources.
4791 . Build your Java application with LTTng-UST Java agent support as
4792 usual, following the procedure for either the <<jul,JUL>> or
4793 <<log4j,Apache log4j>> framework.
4796 .Provide application-specific context fields in a Java application.
4801 import java.util.logging.Handler;
4802 import java.util.logging.Logger;
4803 import org.lttng.ust.agent.jul.LttngLogHandler;
4804 import org.lttng.ust.agent.context.ContextInfoManager;
4805 import org.lttng.ust.agent.context.IContextInfoRetriever;
4809 // Our context information retriever class
4810 private static class MyContextInfoRetriever
4811 implements IContextInfoRetriever
4814 public Object retrieveContextInfo(String key) {
4815 if (key.equals("intCtx")) {
4817 } else if (key.equals("strContext")) {
4818 return "context value!";
4825 private static final int answer = 42;
4827 public static void main(String args[]) throws Exception
4829 // Get the context information manager instance
4830 ContextInfoManager cim = ContextInfoManager.getInstance();
4832 // Create and register our context information retriever
4833 IContextInfoRetriever cir = new MyContextInfoRetriever();
4834 cim.registerContextInfoRetriever("myRetriever", cir);
4837 Logger logger = Logger.getLogger("jello");
4839 // Create an LTTng-UST log handler
4840 Handler lttngUstLogHandler = new LttngLogHandler();
4842 // Add the LTTng-UST log handler to our logger
4843 logger.addHandler(lttngUstLogHandler);
4846 logger.info("some info");
4847 logger.warning("some warning");
4849 logger.finer("finer information; the answer is " + answer);
4851 logger.severe("error!");
4853 // Not mandatory, but cleaner
4854 logger.removeHandler(lttngUstLogHandler);
4855 lttngUstLogHandler.close();
4856 cim.unregisterContextInfoRetriever("myRetriever");
4865 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4868 <<creating-destroying-tracing-sessions,Create a tracing session>>
4869 and <<enabling-disabling-events,create an event rule>> matching the
4875 lttng enable-event --jul jello
4878 <<adding-context,Add the application-specific context fields>> to the
4883 lttng add-context --jul --type='$app.myRetriever:intCtx'
4884 lttng add-context --jul --type='$app.myRetriever:strContext'
4887 <<basic-tracing-session-control,Start tracing>>:
4894 Run the compiled class:
4898 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4901 <<basic-tracing-session-control,Stop tracing>> and inspect the
4913 [[python-application]]
4914 === User space Python agent
4916 You can instrument a Python 2 or Python 3 application which uses the
4917 standard https://docs.python.org/3/library/logging.html[`logging`]
4920 Each log statement emits an LTTng event once the
4921 application module imports the
4922 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4925 .A Python application importing the LTTng-UST Python agent.
4926 image::python-app.png[]
4928 To use the LTTng-UST Python agent:
4930 . In the Python application's source code, import the LTTng-UST Python
4940 The LTTng-UST Python agent automatically adds its logging handler to the
4941 root logger at import time.
4943 Any log statement that the application executes before this import does
4944 not emit an LTTng event.
4946 IMPORTANT: The LTTng-UST Python agent must be
4947 <<installing-lttng,installed>>.
4949 . Use log statements and logging configuration as usual.
4950 Since the LTTng-UST Python agent adds a handler to the _root_
4951 logger, you can trace any log statement from any logger.
4953 .Use the LTTng-UST Python agent.
4964 logging.basicConfig()
4965 logger = logging.getLogger('my-logger')
4968 logger.debug('debug message')
4969 logger.info('info message')
4970 logger.warn('warn message')
4971 logger.error('error message')
4972 logger.critical('critical message')
4976 if __name__ == '__main__':
4980 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4981 logging handler which prints to the standard error stream, is not
4982 strictly required for LTTng-UST tracing to work, but in versions of
4983 Python preceding 3.2, you could see a warning message which indicates
4984 that no handler exists for the logger `my-logger`.
4986 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4987 <<enabling-disabling-events,create an event rule>> matching the
4988 `my-logger` Python logger, and <<basic-tracing-session-control,start
4994 lttng enable-event --python my-logger
4998 Run the Python script:
5005 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5015 You can use the opt:lttng-enable-event(1):--loglevel or
5016 opt:lttng-enable-event(1):--loglevel-only option of the
5017 man:lttng-enable-event(1) command to target a range of Python log levels
5018 or a specific Python log level.
5020 When an application imports the LTTng-UST Python agent, the agent tries
5021 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5022 <<start-sessiond,start the session daemon>> _before_ you run the Python
5023 application. If a session daemon is found, the agent tries to register
5024 to it during 5{nbsp}seconds, after which the application continues
5025 without LTTng tracing support. You can override this timeout value with
5026 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5029 If the session daemon stops while a Python application with an imported
5030 LTTng-UST Python agent runs, the agent retries to connect and to
5031 register to a session daemon every 3{nbsp}seconds. You can override this
5032 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5037 [[proc-lttng-logger-abi]]
5040 The `lttng-tracer` Linux kernel module, part of
5041 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5042 path:{/proc/lttng-logger} when it's loaded. Any application can write
5043 text data to this file to emit an LTTng event.
5046 .An application writes to the LTTng logger file to emit an LTTng event.
5047 image::lttng-logger.png[]
5049 The LTTng logger is the quickest method--not the most efficient,
5050 however--to add instrumentation to an application. It is designed
5051 mostly to instrument shell scripts:
5055 echo "Some message, some $variable" > /proc/lttng-logger
5058 Any event that the LTTng logger emits is named `lttng_logger` and
5059 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5060 other instrumentation points in the kernel tracing domain, **any Unix
5061 user** can <<enabling-disabling-events,create an event rule>> which
5062 matches its event name, not only the root user or users in the tracing
5065 To use the LTTng logger:
5067 * From any application, write text data to the path:{/proc/lttng-logger}
5070 The `msg` field of `lttng_logger` event records contains the
5073 NOTE: The maximum message length of an LTTng logger event is
5074 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5075 than one event to contain the remaining data.
5077 You should not use the LTTng logger to trace a user application which
5078 can be instrumented in a more efficient way, namely:
5080 * <<c-application,C and $$C++$$ applications>>.
5081 * <<java-application,Java applications>>.
5082 * <<python-application,Python applications>>.
5084 .Use the LTTng logger.
5089 echo 'Hello, World!' > /proc/lttng-logger
5091 df --human-readable --print-type / > /proc/lttng-logger
5094 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5095 <<enabling-disabling-events,create an event rule>> matching the
5096 `lttng_logger` Linux kernel tracepoint, and
5097 <<basic-tracing-session-control,start tracing>>:
5102 lttng enable-event --kernel lttng_logger
5106 Run the Bash script:
5113 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5124 [[instrumenting-linux-kernel]]
5125 === LTTng kernel tracepoints
5127 NOTE: This section shows how to _add_ instrumentation points to the
5128 Linux kernel. The kernel's subsystems are already thoroughly
5129 instrumented at strategic places for LTTng when you
5130 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5134 There are two methods to instrument the Linux kernel:
5136 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5137 tracepoint which uses the `TRACE_EVENT()` API.
5139 Choose this if you want to instrumentation a Linux kernel tree with an
5140 instrumentation point compatible with ftrace, perf, and SystemTap.
5142 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5143 instrument an out-of-tree kernel module.
5145 Choose this if you don't need ftrace, perf, or SystemTap support.
5149 [[linux-add-lttng-layer]]
5150 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5152 This section shows how to add an LTTng layer to existing ftrace
5153 instrumentation using the `TRACE_EVENT()` API.
5155 This section does not document the `TRACE_EVENT()` macro. You can
5156 read the following articles to learn more about this API:
5158 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5159 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5160 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5162 The following procedure assumes that your ftrace tracepoints are
5163 correctly defined in their own header and that they are created in
5164 one source file using the `CREATE_TRACE_POINTS` definition.
5166 To add an LTTng layer over an existing ftrace tracepoint:
5168 . Make sure the following kernel configuration options are
5174 * `CONFIG_HIGH_RES_TIMERS`
5175 * `CONFIG_TRACEPOINTS`
5178 . Build the Linux source tree with your custom ftrace tracepoints.
5179 . Boot the resulting Linux image on your target system.
5181 Confirm that the tracepoints exist by looking for their names in the
5182 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5183 is your subsystem's name.
5185 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5191 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
5192 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
5193 cd lttng-modules-2.8.*
5197 . In dir:{instrumentation/events/lttng-module}, relative to the root
5198 of the LTTng-modules source tree, create a header file named
5199 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5200 LTTng-modules tracepoint definitions using the LTTng-modules
5203 Start with this template:
5207 .path:{instrumentation/events/lttng-module/my_subsys.h}
5210 #define TRACE_SYSTEM my_subsys
5212 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5213 #define _LTTNG_MY_SUBSYS_H
5215 #include "../../../probes/lttng-tracepoint-event.h"
5216 #include <linux/tracepoint.h>
5218 LTTNG_TRACEPOINT_EVENT(
5220 * Format is identical to TRACE_EVENT()'s version for the three
5221 * following macro parameters:
5224 TP_PROTO(int my_int, const char *my_string),
5225 TP_ARGS(my_int, my_string),
5227 /* LTTng-modules specific macros */
5229 ctf_integer(int, my_int_field, my_int)
5230 ctf_string(my_bar_field, my_bar)
5234 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5236 #include "../../../probes/define_trace.h"
5240 The entries in the `TP_FIELDS()` section are the list of fields for the
5241 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5242 ftrace's `TRACE_EVENT()` macro.
5244 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5245 complete description of the available `ctf_*()` macros.
5247 . Create the LTTng-modules probe's kernel module C source file,
5248 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5253 .path:{probes/lttng-probe-my-subsys.c}
5255 #include <linux/module.h>
5256 #include "../lttng-tracer.h"
5259 * Build-time verification of mismatch between mainline
5260 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5261 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5263 #include <trace/events/my_subsys.h>
5265 /* Create LTTng tracepoint probes */
5266 #define LTTNG_PACKAGE_BUILD
5267 #define CREATE_TRACE_POINTS
5268 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5270 #include "../instrumentation/events/lttng-module/my_subsys.h"
5272 MODULE_LICENSE("GPL and additional rights");
5273 MODULE_AUTHOR("Your name <your-email>");
5274 MODULE_DESCRIPTION("LTTng my_subsys probes");
5275 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5276 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5277 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5278 LTTNG_MODULES_EXTRAVERSION);
5282 . Edit path:{probes/Makefile} and add your new kernel module object
5283 next to the existing ones:
5287 .path:{probes/Makefile}
5291 obj-m += lttng-probe-module.o
5292 obj-m += lttng-probe-power.o
5294 obj-m += lttng-probe-my-subsys.o
5300 . Build and install the LTTng kernel modules:
5305 make KERNELDIR=/path/to/linux
5306 sudo make modules_install
5310 Replace `/path/to/linux` with the path to the Linux source tree where
5311 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5313 Note that you can also use the
5314 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5315 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5316 C code that need to be executed before the event fields are recorded.
5318 The best way to learn how to use the previous LTTng-modules macros is to
5319 inspect the existing LTTng-modules tracepoint definitions in the
5320 dir:{instrumentation/events/lttng-module} header files. Compare them
5321 with the Linux kernel mainline versions in the
5322 dir:{include/trace/events} directory of the Linux source tree.
5326 [[lttng-tracepoint-event-code]]
5327 ===== Use custom C code to access the data for tracepoint fields
5329 Although we recommended to always use the
5330 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5331 the arguments and fields of an LTTng-modules tracepoint when possible,
5332 sometimes you need a more complex process to access the data that the
5333 tracer records as event record fields. In other words, you need local
5334 variables and multiple C{nbsp}statements instead of simple
5335 argument-based expressions that you pass to the
5336 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5338 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5339 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5340 a block of C{nbsp}code to be executed before LTTng records the fields.
5341 The structure of this macro is:
5344 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5346 LTTNG_TRACEPOINT_EVENT_CODE(
5348 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5349 * version for the following three macro parameters:
5352 TP_PROTO(int my_int, const char *my_string),
5353 TP_ARGS(my_int, my_string),
5355 /* Declarations of custom local variables */
5358 unsigned long b = 0;
5359 const char *name = "(undefined)";
5360 struct my_struct *my_struct;
5364 * Custom code which uses both tracepoint arguments
5365 * (in TP_ARGS()) and local variables (in TP_locvar()).
5367 * Local variables are actually members of a structure pointed
5368 * to by the special variable tp_locvar.
5372 tp_locvar->a = my_int + 17;
5373 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5374 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5375 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5376 put_my_struct(tp_locvar->my_struct);
5385 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5386 * version for this, except that tp_locvar members can be
5387 * used in the argument expression parameters of
5388 * the ctf_*() macros.
5391 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5392 ctf_integer(int, my_struct_a, tp_locvar->a)
5393 ctf_string(my_string_field, my_string)
5394 ctf_string(my_struct_name, tp_locvar->name)
5399 IMPORTANT: The C code defined in `TP_code()` must not have any side
5400 effects when executed. In particular, the code must not allocate
5401 memory or get resources without deallocating this memory or putting
5402 those resources afterwards.
5405 [[instrumenting-linux-kernel-tracing]]
5406 ==== Load and unload a custom probe kernel module
5408 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5409 kernel module>> in the kernel before it can emit LTTng events.
5411 To load the default probe kernel modules and a custom probe kernel
5414 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5415 probe modules to load when starting a root <<lttng-sessiond,session
5419 .Load the `my_subsys`, `usb`, and the default probe modules.
5423 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5428 You only need to pass the subsystem name, not the whole kernel module
5431 To load _only_ a given custom probe kernel module:
5433 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5434 modules to load when starting a root session daemon:
5437 .Load only the `my_subsys` and `usb` probe modules.
5441 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5446 To confirm that a probe module is loaded:
5453 lsmod | grep lttng_probe_usb
5457 To unload the loaded probe modules:
5459 * Kill the session daemon with `SIGTERM`:
5464 sudo pkill lttng-sessiond
5468 You can also use man:modprobe(8)'s `--remove` option if the session
5469 daemon terminates abnormally.
5472 [[controlling-tracing]]
5475 Once an application or a Linux kernel is
5476 <<instrumenting,instrumented>> for LTTng tracing,
5479 This section is divided in topics on how to use the various
5480 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5481 command-line tool>>, to _control_ the LTTng daemons and tracers.
5483 NOTE: In the following subsections, we refer to an man:lttng(1) command
5484 using its man page name. For example, instead of _Run the `create`
5485 command to..._, we use _Run the man:lttng-create(1) command to..._.
5489 === Start a session daemon
5491 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5492 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5495 You will see the following error when you run a command while no session
5499 Error: No session daemon is available
5502 The only command that automatically runs a session daemon is
5503 man:lttng-create(1), which you use to
5504 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5505 this is most of the time the first operation that you do, sometimes it's
5506 not. Some examples are:
5508 * <<list-instrumentation-points,List the available instrumentation points>>.
5509 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5511 [[tracing-group]] Each Unix user must have its own running session
5512 daemon to trace user applications. The session daemon that the root user
5513 starts is the only one allowed to control the LTTng kernel tracer. Users
5514 that are part of the _tracing group_ can control the root session
5515 daemon. The default tracing group name is `tracing`; you can set it to
5516 something else with the opt:lttng-sessiond(8):--group option when you
5517 start the root session daemon.
5519 To start a user session daemon:
5521 * Run man:lttng-sessiond(8):
5526 lttng-sessiond --daemonize
5530 To start the root session daemon:
5532 * Run man:lttng-sessiond(8) as the root user:
5537 sudo lttng-sessiond --daemonize
5541 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5542 start the session daemon in foreground.
5544 To stop a session daemon, use man:kill(1) on its process ID (standard
5547 Note that some Linux distributions could manage the LTTng session daemon
5548 as a service. In this case, you should use the service manager to
5549 start, restart, and stop session daemons.
5552 [[creating-destroying-tracing-sessions]]
5553 === Create and destroy a tracing session
5555 Almost all the LTTng control operations happen in the scope of
5556 a <<tracing-session,tracing session>>, which is the dialogue between the
5557 <<lttng-sessiond,session daemon>> and you.
5559 To create a tracing session with a generated name:
5561 * Use the man:lttng-create(1) command:
5570 The created tracing session's name is `auto` followed by the
5573 To create a tracing session with a specific name:
5575 * Use the optional argument of the man:lttng-create(1) command:
5580 lttng create my-session
5584 Replace `my-session` with the specific tracing session name.
5586 LTTng appends the creation date to the created tracing session's name.
5588 LTTng writes the traces of a tracing session in
5589 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5590 name of the tracing session. Note that the env:LTTNG_HOME environment
5591 variable defaults to `$HOME` if not set.
5593 To output LTTng traces to a non-default location:
5595 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5600 lttng create --output=/tmp/some-directory my-session
5604 You may create as many tracing sessions as you wish.
5606 To list all the existing tracing sessions for your Unix user:
5608 * Use the man:lttng-list(1) command:
5617 When you create a tracing session, it is set as the _current tracing
5618 session_. The following man:lttng(1) commands operate on the current
5619 tracing session when you don't specify one:
5621 [role="list-3-cols"]
5637 To change the current tracing session:
5639 * Use the man:lttng-set-session(1) command:
5644 lttng set-session new-session
5648 Replace `new-session` by the name of the new current tracing session.
5650 When you are done tracing in a given tracing session, you can destroy
5651 it. This operation frees the resources taken by the tracing session
5652 to destroy; it does not destroy the trace data that LTTng wrote for
5653 this tracing session.
5655 To destroy the current tracing session:
5657 * Use the man:lttng-destroy(1) command:
5667 [[list-instrumentation-points]]
5668 === List the available instrumentation points
5670 The <<lttng-sessiond,session daemon>> can query the running instrumented
5671 user applications and the Linux kernel to get a list of available
5672 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5673 they are tracepoints and system calls. For the user space tracing
5674 domain, they are tracepoints. For the other tracing domains, they are
5677 To list the available instrumentation points:
5679 * Use the man:lttng-list(1) command with the requested tracing domain's
5683 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5684 must be a root user, or it must be a member of the
5685 <<tracing-group,tracing group>>).
5686 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5687 kernel system calls (your Unix user must be a root user, or it must be
5688 a member of the tracing group).
5689 * opt:lttng-list(1):--userspace: user space tracepoints.
5690 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5691 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5692 * opt:lttng-list(1):--python: Python loggers.
5695 .List the available user space tracepoints.
5699 lttng list --userspace
5703 .List the available Linux kernel system call tracepoints.
5707 lttng list --kernel --syscall
5712 [[enabling-disabling-events]]
5713 === Create and enable an event rule
5715 Once you <<creating-destroying-tracing-sessions,create a tracing
5716 session>>, you can create <<event,event rules>> with the
5717 man:lttng-enable-event(1) command.
5719 You specify each condition with a command-line option. The available
5720 condition options are shown in the following table.
5722 [role="growable",cols="asciidoc,asciidoc,default"]
5723 .Condition command-line options for the man:lttng-enable-event(1) command.
5725 |Option |Description |Applicable tracing domains
5731 . +--probe=__ADDR__+
5732 . +--function=__ADDR__+
5735 Instead of using the default _tracepoint_ instrumentation type, use:
5737 . A Linux system call.
5738 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5739 . The entry and return points of a Linux function (symbol or address).
5743 |First positional argument.
5746 Tracepoint or system call name. In the case of a Linux KProbe or
5747 function, this is a custom name given to the event rule. With the
5748 JUL, log4j, and Python domains, this is a logger name.
5750 With a tracepoint, logger, or system call name, the last character
5751 can be `*` to match anything that remains.
5758 . +--loglevel=__LEVEL__+
5759 . +--loglevel-only=__LEVEL__+
5762 . Match only tracepoints or log statements with a logging level at
5763 least as severe as +__LEVEL__+.
5764 . Match only tracepoints or log statements with a logging level
5765 equal to +__LEVEL__+.
5767 See man:lttng-enable-event(1) for the list of available logging level
5770 |User space, JUL, log4j, and Python.
5772 |+--exclude=__EXCLUSIONS__+
5775 When you use a `*` character at the end of the tracepoint or logger
5776 name (first positional argument), exclude the specific names in the
5777 comma-delimited list +__EXCLUSIONS__+.
5780 User space, JUL, log4j, and Python.
5782 |+--filter=__EXPR__+
5785 Match only events which satisfy the expression +__EXPR__+.
5787 See man:lttng-enable-event(1) to learn more about the syntax of a
5794 You attach an event rule to a <<channel,channel>> on creation. If you do
5795 not specify the channel with the opt:lttng-enable-event(1):--channel
5796 option, and if the event rule to create is the first in its
5797 <<domain,tracing domain>> for a given tracing session, then LTTng
5798 creates a _default channel_ for you. This default channel is reused in
5799 subsequent invocations of the man:lttng-enable-event(1) command for the
5800 same tracing domain.
5802 An event rule is always enabled at creation time.
5804 The following examples show how you can combine the previous
5805 command-line options to create simple to more complex event rules.
5807 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5811 lttng enable-event --kernel sched_switch
5815 .Create an event rule matching four Linux kernel system calls (default channel).
5819 lttng enable-event --kernel --syscall open,write,read,close
5823 .Create event rules matching tracepoints with filter expressions (default channel).
5827 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5832 lttng enable-event --kernel --all \
5833 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5838 lttng enable-event --jul my_logger \
5839 --filter='$app.retriever:cur_msg_id > 3'
5842 IMPORTANT: Make sure to always quote the filter string when you
5843 use man:lttng(1) from a shell.
5846 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5850 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5853 IMPORTANT: Make sure to always quote the wildcard character when you
5854 use man:lttng(1) from a shell.
5857 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5861 lttng enable-event --python my-app.'*' \
5862 --exclude='my-app.module,my-app.hello'
5866 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5870 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5874 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5878 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5882 The event rules of a given channel form a whitelist: as soon as an
5883 emitted event passes one of them, LTTng can record the event. For
5884 example, an event named `my_app:my_tracepoint` emitted from a user space
5885 tracepoint with a `TRACE_ERROR` log level passes both of the following
5890 lttng enable-event --userspace my_app:my_tracepoint
5891 lttng enable-event --userspace my_app:my_tracepoint \
5892 --loglevel=TRACE_INFO
5895 The second event rule is redundant: the first one includes
5899 [[disable-event-rule]]
5900 === Disable an event rule
5902 To disable an event rule that you <<enabling-disabling-events,created>>
5903 previously, use the man:lttng-disable-event(1) command. This command
5904 disables _all_ the event rules (of a given tracing domain and channel)
5905 which match an instrumentation point. The other conditions are not
5906 supported as of LTTng{nbsp}{revision}.
5908 The LTTng tracer does not record an emitted event which passes
5909 a _disabled_ event rule.
5911 .Disable an event rule matching a Python logger (default channel).
5915 lttng disable-event --python my-logger
5919 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5923 lttng disable-event --jul '*'
5927 .Disable _all_ the event rules of the default channel.
5929 The opt:lttng-disable-event(1):--all-events option is not, like the
5930 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5931 equivalent of the event name `*` (wildcard): it disables _all_ the event
5932 rules of a given channel.
5936 lttng disable-event --jul --all-events
5940 NOTE: You cannot delete an event rule once you create it.
5944 === Get the status of a tracing session
5946 To get the status of the current tracing session, that is, its
5947 parameters, its channels, event rules, and their attributes:
5949 * Use the man:lttng-status(1) command:
5959 To get the status of any tracing session:
5961 * Use the man:lttng-list(1) command with the tracing session's name:
5966 lttng list my-session
5970 Replace `my-session` with the desired tracing session's name.
5973 [[basic-tracing-session-control]]
5974 === Start and stop a tracing session
5976 Once you <<creating-destroying-tracing-sessions,create a tracing
5978 <<enabling-disabling-events,create one or more event rules>>,
5979 you can start and stop the tracers for this tracing session.
5981 To start tracing in the current tracing session:
5983 * Use the man:lttng-start(1) command:
5992 LTTng is very flexible: you can launch user applications before
5993 or after the you start the tracers. The tracers only record the events
5994 if they pass enabled event rules and if they occur while the tracers are
5997 To stop tracing in the current tracing session:
5999 * Use the man:lttng-stop(1) command:
6008 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6009 records>> or lost sub-buffers since the last time you ran
6010 man:lttng-start(1), warnings are printed when you run the
6011 man:lttng-stop(1) command.
6014 [[enabling-disabling-channels]]
6015 === Create a channel
6017 Once you create a tracing session, you can create a <<channel,channel>>
6018 with the man:lttng-enable-channel(1) command.
6020 Note that LTTng automatically creates a default channel when, for a
6021 given <<domain,tracing domain>>, no channels exist and you
6022 <<enabling-disabling-events,create>> the first event rule. This default
6023 channel is named `channel0` and its attributes are set to reasonable
6024 values. Therefore, you only need to create a channel when you need
6025 non-default attributes.
6027 You specify each non-default channel attribute with a command-line
6028 option when you use the man:lttng-enable-channel(1) command. The
6029 available command-line options are:
6031 [role="growable",cols="asciidoc,asciidoc"]
6032 .Command-line options for the man:lttng-enable-channel(1) command.
6034 |Option |Description
6040 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
6041 the default _discard_ mode.
6043 |`--buffers-pid` (user space tracing domain only)
6046 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6047 instead of the default per-user buffering scheme.
6049 |+--subbuf-size=__SIZE__+
6052 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6053 either for each Unix user (default), or for each instrumented process.
6055 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6057 |+--num-subbuf=__COUNT__+
6060 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6061 for each Unix user (default), or for each instrumented process.
6063 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6065 |+--tracefile-size=__SIZE__+
6068 Set the maximum size of each trace file that this channel writes within
6069 a stream to +__SIZE__+ bytes instead of no maximum.
6071 See <<tracefile-rotation,Trace file count and size>>.
6073 |+--tracefile-count=__COUNT__+
6076 Limit the number of trace files that this channel creates to
6077 +__COUNT__+ channels instead of no limit.
6079 See <<tracefile-rotation,Trace file count and size>>.
6081 |+--switch-timer=__PERIODUS__+
6084 Set the <<channel-switch-timer,switch timer period>>
6085 to +__PERIODUS__+{nbsp}µs.
6087 |+--read-timer=__PERIODUS__+
6090 Set the <<channel-read-timer,read timer period>>
6091 to +__PERIODUS__+{nbsp}µs.
6093 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6096 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6100 You can only create a channel in the Linux kernel and user space
6101 <<domain,tracing domains>>: other tracing domains have their own channel
6102 created on the fly when <<enabling-disabling-events,creating event
6107 Because of a current LTTng limitation, you must create all channels
6108 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6109 tracing session, that is, before the first time you run
6112 Since LTTng automatically creates a default channel when you use the
6113 man:lttng-enable-event(1) command with a specific tracing domain, you
6114 cannot, for example, create a Linux kernel event rule, start tracing,
6115 and then create a user space event rule, because no user space channel
6116 exists yet and it's too late to create one.
6118 For this reason, make sure to configure your channels properly
6119 before starting the tracers for the first time!
6122 The following examples show how you can combine the previous
6123 command-line options to create simple to more complex channels.
6125 .Create a Linux kernel channel with default attributes.
6129 lttng enable-channel --kernel my-channel
6133 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6137 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6138 --buffers-pid my-channel
6142 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6146 lttng enable-channel --kernel --tracefile-count=8 \
6147 --tracefile-size=4194304 my-channel
6151 .Create a user space channel in overwrite (or _flight recorder_) mode.
6155 lttng enable-channel --userspace --overwrite my-channel
6159 You can <<enabling-disabling-events,create>> the same event rule in
6160 two different channels:
6164 lttng enable-event --userspace --channel=my-channel app:tp
6165 lttng enable-event --userspace --channel=other-channel app:tp
6168 If both channels are enabled, when a tracepoint named `app:tp` is
6169 reached, LTTng records two events, one for each channel.
6173 === Disable a channel
6175 To disable a specific channel that you <<enabling-disabling-channels,created>>
6176 previously, use the man:lttng-disable-channel(1) command.
6178 .Disable a specific Linux kernel channel.
6182 lttng disable-channel --kernel my-channel
6186 The state of a channel precedes the individual states of event rules
6187 attached to it: event rules which belong to a disabled channel, even if
6188 they are enabled, are also considered disabled.
6192 === Add context fields to a channel
6194 Event record fields in trace files provide important information about
6195 events that occured previously, but sometimes some external context may
6196 help you solve a problem faster. Examples of context fields are:
6198 * The **process ID**, **thread ID**, **process name**, and
6199 **process priority** of the thread in which the event occurs.
6200 * The **hostname** of the system on which the event occurs.
6201 * The current values of many possible **performance counters** using
6203 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6205 ** Branch instructions, misses, and loads.
6207 * Any context defined at the application level (supported for the
6208 JUL and log4j <<domain,tracing domains>>).
6210 To get the full list of available context fields, see
6211 `lttng add-context --list`. Some context fields are reserved for a
6212 specific <<domain,tracing domain>> (Linux kernel or user space).
6214 You add context fields to <<channel,channels>>. All the events
6215 that a channel with added context fields records contain those fields.
6217 To add context fields to one or all the channels of a given tracing
6220 * Use the man:lttng-add-context(1) command.
6222 .Add context fields to all the channels of the current tracing session.
6224 The following command line adds the virtual process identifier and
6225 the per-thread CPU cycles count fields to all the user space channels
6226 of the current tracing session.
6230 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6234 .Add a context field to a specific channel.
6236 The following command line adds the thread identifier context field
6237 to the Linux kernel channel named `my-channel` in the current
6242 lttng add-context --kernel --channel=my-channel --type=tid
6246 .Add an application-specific context field to a specific channel.
6248 The following command line adds the `cur_msg_id` context field of the
6249 `retriever` context retriever for all the instrumented
6250 <<java-application,Java applications>> recording <<event,event records>>
6251 in the channel named `my-channel`:
6255 lttng add-context --kernel --channel=my-channel \
6256 --type='$app:retriever:cur_msg_id'
6259 IMPORTANT: Make sure to always quote the `$` character when you
6260 use man:lttng-add-context(1) from a shell.
6263 NOTE: You cannot remove context fields from a channel once you add it.
6268 === Track process IDs
6270 It's often useful to allow only specific process IDs (PIDs) to emit
6271 events. For example, you may wish to record all the system calls made by
6272 a given process (à la http://linux.die.net/man/1/strace[strace]).
6274 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6275 purpose. Both commands operate on a whitelist of process IDs. You _add_
6276 entries to this whitelist with the man:lttng-track(1) command and remove
6277 entries with the man:lttng-untrack(1) command. Any process which has one
6278 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6279 an enabled <<event,event rule>>.
6281 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6282 process with a given tracked ID exit and another process be given this
6283 ID, then the latter would also be allowed to emit events.
6285 .Track and untrack process IDs.
6287 For the sake of the following example, assume the target system has 16
6291 <<creating-destroying-tracing-sessions,create a tracing session>>,
6292 the whitelist contains all the possible PIDs:
6295 .All PIDs are tracked.
6296 image::track-all.png[]
6298 When the whitelist is full and you use the man:lttng-track(1) command to
6299 specify some PIDs to track, LTTng first clears the whitelist, then it
6300 tracks the specific PIDs. After:
6304 lttng track --pid=3,4,7,10,13
6310 .PIDs 3, 4, 7, 10, and 13 are tracked.
6311 image::track-3-4-7-10-13.png[]
6313 You can add more PIDs to the whitelist afterwards:
6317 lttng track --pid=1,15,16
6323 .PIDs 1, 15, and 16 are added to the whitelist.
6324 image::track-1-3-4-7-10-13-15-16.png[]
6326 The man:lttng-untrack(1) command removes entries from the PID tracker's
6327 whitelist. Given the previous example, the following command:
6331 lttng untrack --pid=3,7,10,13
6334 leads to this whitelist:
6337 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6338 image::track-1-4-15-16.png[]
6340 LTTng can track all possible PIDs again using the opt:track(1):--all
6345 lttng track --pid --all
6348 The result is, again:
6351 .All PIDs are tracked.
6352 image::track-all.png[]
6355 .Track only specific PIDs
6357 A very typical use case with PID tracking is to start with an empty
6358 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6359 then add PIDs manually while tracers are active. You can accomplish this
6360 by using the opt:lttng-untrack(1):--all option of the
6361 man:lttng-untrack(1) command to clear the whitelist after you
6362 <<creating-destroying-tracing-sessions,create a tracing session>>:
6366 lttng untrack --pid --all
6372 .No PIDs are tracked.
6373 image::untrack-all.png[]
6375 If you trace with this whitelist configuration, the tracer records no
6376 events for this <<domain,tracing domain>> because no processes are
6377 tracked. You can use the man:lttng-track(1) command as usual to track
6378 specific PIDs, for example:
6382 lttng track --pid=6,11
6388 .PIDs 6 and 11 are tracked.
6389 image::track-6-11.png[]
6394 [[saving-loading-tracing-session]]
6395 === Save and load tracing session configurations
6397 Configuring a <<tracing-session,tracing session>> can be long. Some of
6398 the tasks involved are:
6400 * <<enabling-disabling-channels,Create channels>> with
6401 specific attributes.
6402 * <<adding-context,Add context fields>> to specific channels.
6403 * <<enabling-disabling-events,Create event rules>> with specific log
6404 level and filter conditions.
6406 If you use LTTng to solve real world problems, chances are you have to
6407 record events using the same tracing session setup over and over,
6408 modifying a few variables each time in your instrumented program
6409 or environment. To avoid constant tracing session reconfiguration,
6410 the man:lttng(1) command-line tool can save and load tracing session
6411 configurations to/from XML files.
6413 To save a given tracing session configuration:
6415 * Use the man:lttng-save(1) command:
6420 lttng save my-session
6424 Replace `my-session` with the name of the tracing session to save.
6426 LTTng saves tracing session configurations to
6427 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6428 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6429 the opt:lttng-save(1):--output-path option to change this destination
6432 LTTng saves all configuration parameters, for example:
6434 * The tracing session name.
6435 * The trace data output path.
6436 * The channels with their state and all their attributes.
6437 * The context fields you added to channels.
6438 * The event rules with their state, log level and filter conditions.
6440 To load a tracing session:
6442 * Use the man:lttng-load(1) command:
6447 lttng load my-session
6451 Replace `my-session` with the name of the tracing session to load.
6453 When LTTng loads a configuration, it restores your saved tracing session
6454 as if you just configured it manually.
6456 See man:lttng(1) for the complete list of command-line options. You
6457 can also save and load all many sessions at a time, and decide in which
6458 directory to output the XML files.
6461 [[sending-trace-data-over-the-network]]
6462 === Send trace data over the network
6464 LTTng can send the recorded trace data to a remote system over the
6465 network instead of writing it to the local file system.
6467 To send the trace data over the network:
6469 . On the _remote_ system (which can also be the target system),
6470 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6479 . On the _target_ system, create a tracing session configured to
6480 send trace data over the network:
6485 lttng create my-session --set-url=net://remote-system
6489 Replace `remote-system` by the host name or IP address of the
6490 remote system. See man:lttng-create(1) for the exact URL format.
6492 . On the target system, use the man:lttng(1) command-line tool as usual.
6493 When tracing is active, the target's consumer daemon sends sub-buffers
6494 to the relay daemon running on the remote system intead of flushing
6495 them to the local file system. The relay daemon writes the received
6496 packets to the local file system.
6498 The relay daemon writes trace files to
6499 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6500 +__hostname__+ is the host name of the target system and +__session__+
6501 is the tracing session name. Note that the env:LTTNG_HOME environment
6502 variable defaults to `$HOME` if not set. Use the
6503 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6504 trace files to another base directory.
6509 === View events as LTTng emits them (noch:{LTTng} live)
6511 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6512 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6513 display events as LTTng emits them on the target system while tracing is
6516 The relay daemon creates a _tee_: it forwards the trace data to both
6517 the local file system and to connected live viewers:
6520 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6525 . On the _target system_, create a <<tracing-session,tracing session>>
6531 lttng create --live my-session
6535 This spawns a local relay daemon.
6537 . Start the live viewer and configure it to connect to the relay
6538 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6543 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6550 * `hostname` with the host name of the target system.
6551 * `my-session` with the name of the tracing session to view.
6554 . Configure the tracing session as usual with the man:lttng(1)
6555 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6557 You can list the available live tracing sessions with Babeltrace:
6561 babeltrace --input-format=lttng-live net://localhost
6564 You can start the relay daemon on another system. In this case, you need
6565 to specify the relay daemon's URL when you create the tracing session
6566 with the opt:lttng-create(1):--set-url option. You also need to replace
6567 `localhost` in the procedure above with the host name of the system on
6568 which the relay daemon is running.
6570 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6571 command-line options.
6575 [[taking-a-snapshot]]
6576 === Take a snapshot of the current sub-buffers of a tracing session
6578 The normal behavior of LTTng is to append full sub-buffers to growing
6579 trace data files. This is ideal to keep a full history of the events
6580 that occurred on the target system, but it can
6581 represent too much data in some situations. For example, you may wish
6582 to trace your application continuously until some critical situation
6583 happens, in which case you only need the latest few recorded
6584 events to perform the desired analysis, not multi-gigabyte trace files.
6586 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6587 current sub-buffers of a given <<tracing-session,tracing session>>.
6588 LTTng can write the snapshot to the local file system or send it over
6593 . Create a tracing session in _snapshot mode_:
6598 lttng create --snapshot my-session
6602 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6603 <<channel,channels>> created in this mode is automatically set to
6604 _overwrite_ (flight recorder mode).
6606 . Configure the tracing session as usual with the man:lttng(1)
6607 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6609 . **Optional**: When you need to take a snapshot,
6610 <<basic-tracing-session-control,stop tracing>>.
6612 You can take a snapshot when the tracers are active, but if you stop
6613 them first, you are sure that the data in the sub-buffers does not
6614 change before you actually take the snapshot.
6621 lttng snapshot record --name=my-first-snapshot
6625 LTTng writes the current sub-buffers of all the current tracing
6626 session's channels to trace files on the local file system. Those trace
6627 files have `my-first-snapshot` in their name.
6629 There is no difference between the format of a normal trace file and the
6630 format of a snapshot: viewers of LTTng traces also support LTTng
6633 By default, LTTng writes snapshot files to the path shown by
6634 `lttng snapshot list-output`. You can change this path or decide to send
6635 snapshots over the network using either:
6637 . An output path or URL that you specify when you create the
6639 . An snapshot output path or URL that you add using
6640 `lttng snapshot add-output`
6641 . An output path or URL that you provide directly to the
6642 `lttng snapshot record` command.
6644 Method 3 overrides method 2, which overrides method 1. When you
6645 specify a URL, a relay daemon must listen on a remote system (see
6646 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6651 === Use the machine interface
6653 With any command of the man:lttng(1) command-line tool, you can set the
6654 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6655 XML machine interface output, for example:
6659 lttng --mi=xml enable-event --kernel --syscall open
6662 A schema definition (XSD) is
6663 https://github.com/lttng/lttng-tools/blob/stable-2.8/src/common/mi-lttng-3.0.xsd[available]
6664 to ease the integration with external tools as much as possible.
6668 [[metadata-regenerate]]
6669 === Regenerate the metadata of an LTTng trace
6671 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6672 data stream files and a metadata file. This metadata file contains,
6673 amongst other things, information about the offset of the clock sources
6674 used to timestamp <<event,event records>> when tracing.
6676 If, once a <<tracing-session,tracing session>> is
6677 <<basic-tracing-session-control,started>>, a major
6678 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6679 happens, the trace's clock offset also needs to be updated. You
6680 can use the man:lttng-metadata(1) command to do so.
6682 The main use case of this command is to allow a system to boot with
6683 an incorrect wall time and trace it with LTTng before its wall time
6684 is corrected. Once the system is known to be in a state where its
6685 wall time is correct, it can run `lttng metadata regenerate`.
6687 To regenerate the metadata of an LTTng trace:
6689 * Use the `regenerate` action of the man:lttng-metadata(1) command:
6694 lttng metadata regenerate
6700 `lttng metadata regenerate` has the following limitations:
6702 * Tracing session <<creating-destroying-tracing-sessions,created>>
6704 * User space <<channel,channels>>, if any, using
6705 <<channel-buffering-schemes,per-user buffering>>.
6710 [[persistent-memory-file-systems]]
6711 === Record trace data on persistent memory file systems
6713 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6714 (NVRAM) is random-access memory that retains its information when power
6715 is turned off (non-volatile). Systems with such memory can store data
6716 structures in RAM and retrieve them after a reboot, without flushing
6717 to typical _storage_.
6719 Linux supports NVRAM file systems thanks to either
6720 http://pramfs.sourceforge.net/[PRAMFS] or
6721 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6722 (requires Linux 4.1+).
6724 This section does not describe how to operate such file systems;
6725 we assume that you have a working persistent memory file system.
6727 When you create a <<tracing-session,tracing session>>, you can specify
6728 the path of the shared memory holding the sub-buffers. If you specify a
6729 location on an NVRAM file system, then you can retrieve the latest
6730 recorded trace data when the system reboots after a crash.
6732 To record trace data on a persistent memory file system and retrieve the
6733 trace data after a system crash:
6735 . Create a tracing session with a sub-buffer shared memory path located
6736 on an NVRAM file system:
6741 lttng create --shm-path=/path/to/shm
6745 . Configure the tracing session as usual with the man:lttng(1)
6746 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6748 . After a system crash, use the man:lttng-crash(1) command-line tool to
6749 view the trace data recorded on the NVRAM file system:
6754 lttng-crash /path/to/shm
6758 The binary layout of the ring buffer files is not exactly the same as
6759 the trace files layout. This is why you need to use man:lttng-crash(1)
6760 instead of your preferred trace viewer directly.
6762 To convert the ring buffer files to LTTng trace files:
6764 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6769 lttng-crash --extract=/path/to/trace /path/to/shm
6777 [[lttng-modules-ref]]
6778 === noch:{LTTng-modules}
6781 [[lttng-modules-tp-fields]]
6782 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6784 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6785 tracepoint fields, which must be listed within `TP_FIELDS()` in
6786 `LTTNG_TRACEPOINT_EVENT()`, are:
6788 [role="func-desc growable",cols="asciidoc,asciidoc"]
6789 .Available macros to define LTTng-modules tracepoint fields
6791 |Macro |Description and parameters
6794 +ctf_integer(__t__, __n__, __e__)+
6796 +ctf_integer_nowrite(__t__, __n__, __e__)+
6798 +ctf_user_integer(__t__, __n__, __e__)+
6800 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6802 Standard integer, displayed in base 10.
6805 Integer C type (`int`, `long`, `size_t`, ...).
6811 Argument expression.
6814 +ctf_integer_hex(__t__, __n__, __e__)+
6816 +ctf_user_integer_hex(__t__, __n__, __e__)+
6818 Standard integer, displayed in base 16.
6827 Argument expression.
6829 |+ctf_integer_oct(__t__, __n__, __e__)+
6831 Standard integer, displayed in base 8.
6840 Argument expression.
6843 +ctf_integer_network(__t__, __n__, __e__)+
6845 +ctf_user_integer_network(__t__, __n__, __e__)+
6847 Integer in network byte order (big-endian), displayed in base 10.
6856 Argument expression.
6859 +ctf_integer_network_hex(__t__, __n__, __e__)+
6861 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6863 Integer in network byte order, displayed in base 16.
6872 Argument expression.
6875 +ctf_string(__n__, __e__)+
6877 +ctf_string_nowrite(__n__, __e__)+
6879 +ctf_user_string(__n__, __e__)+
6881 +ctf_user_string_nowrite(__n__, __e__)+
6883 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6889 Argument expression.
6892 +ctf_array(__t__, __n__, __e__, __s__)+
6894 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6896 +ctf_user_array(__t__, __n__, __e__, __s__)+
6898 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6900 Statically-sized array of integers.
6903 Array element C type.
6909 Argument expression.
6915 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
6917 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6919 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
6921 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6923 Statically-sized array of bits.
6925 The type of +__e__+ must be an integer type. +__s__+ is the number
6926 of elements of such type in +__e__+, not the number of bits.
6929 Array element C type.
6935 Argument expression.
6941 +ctf_array_text(__t__, __n__, __e__, __s__)+
6943 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6945 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6947 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6949 Statically-sized array, printed as text.
6951 The string does not need to be null-terminated.
6954 Array element C type (always `char`).
6960 Argument expression.
6966 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6968 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6970 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6972 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6974 Dynamically-sized array of integers.
6976 The type of +__E__+ must be unsigned.
6979 Array element C type.
6985 Argument expression.
6988 Length expression C type.
6994 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6996 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6998 Dynamically-sized array of integers, displayed in base 16.
7000 The type of +__E__+ must be unsigned.
7003 Array element C type.
7009 Argument expression.
7012 Length expression C type.
7017 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7019 Dynamically-sized array of integers in network byte order (big-endian),
7020 displayed in base 10.
7022 The type of +__E__+ must be unsigned.
7025 Array element C type.
7031 Argument expression.
7034 Length expression C type.
7040 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7042 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7044 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7046 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7048 Dynamically-sized array of bits.
7050 The type of +__e__+ must be an integer type. +__s__+ is the number
7051 of elements of such type in +__e__+, not the number of bits.
7053 The type of +__E__+ must be unsigned.
7056 Array element C type.
7062 Argument expression.
7065 Length expression C type.
7071 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7073 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7075 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7077 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7079 Dynamically-sized array, displayed as text.
7081 The string does not need to be null-terminated.
7083 The type of +__E__+ must be unsigned.
7085 The behaviour is undefined if +__e__+ is `NULL`.
7088 Sequence element C type (always `char`).
7094 Argument expression.
7097 Length expression C type.
7103 Use the `_user` versions when the argument expression, `e`, is
7104 a user space address. In the cases of `ctf_user_integer*()` and
7105 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7108 The `_nowrite` versions omit themselves from the session trace, but are
7109 otherwise identical. This means the `_nowrite` fields won't be written
7110 in the recorded trace. Their primary purpose is to make some
7111 of the event context available to the
7112 <<enabling-disabling-events,event filters>> without having to
7113 commit the data to sub-buffers.
7119 Terms related to LTTng and to tracing in general:
7122 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7123 the cmd:babeltrace command, some libraries, and Python bindings.
7125 <<channel-buffering-schemes,buffering scheme>>::
7126 A layout of sub-buffers applied to a given channel.
7128 <<channel,channel>>::
7129 An entity which is responsible for a set of ring buffers.
7131 <<event,Event rules>> are always attached to a specific channel.
7134 A reference of time for a tracer.
7136 <<lttng-consumerd,consumer daemon>>::
7137 A process which is responsible for consuming the full sub-buffers
7138 and write them to a file system or send them over the network.
7140 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7141 mode in which the tracer _discards_ new event records when there's no
7142 sub-buffer space left to store them.
7145 The consequence of the execution of an instrumentation
7146 point, like a tracepoint that you manually place in some source code,
7147 or a Linux kernel KProbe.
7149 An event is said to _occur_ at a specific time. Different actions can
7150 be taken upon the occurance of an event, like record the event's payload
7153 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7154 The mechanism by which event records of a given channel are lost
7155 (not recorded) when there is no sub-buffer space left to store them.
7157 [[def-event-name]]event name::
7158 The name of an event, which is also the name of the event record.
7159 This is also called the _instrumentation point name_.
7162 A record, in a trace, of the payload of an event which occured.
7164 <<event,event rule>>::
7165 Set of conditions which must be satisfied for one or more occuring
7166 events to be recorded.
7168 `java.util.logging`::
7170 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7172 <<instrumenting,instrumentation>>::
7173 The use of LTTng probes to make a piece of software traceable.
7175 instrumentation point::
7176 A point in the execution path of a piece of software that, when
7177 reached by this execution, can emit an event.
7179 instrumentation point name::
7180 See _<<def-event-name,event name>>_.
7183 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7184 developed by the Apache Software Foundation.
7187 Level of severity of a log statement or user space
7188 instrumentation point.
7191 The _Linux Trace Toolkit: next generation_ project.
7193 <<lttng-cli,cmd:lttng>>::
7194 A command-line tool provided by the LTTng-tools project which you
7195 can use to send and receive control messages to and from a
7199 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7200 which is a set of analyzing programs that are used to obtain a
7201 higher level view of an LTTng trace.
7203 cmd:lttng-consumerd::
7204 The name of the consumer daemon program.
7207 A utility provided by the LTTng-tools project which can convert
7208 ring buffer files (usually
7209 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7212 LTTng Documentation::
7215 <<lttng-live,LTTng live>>::
7216 A communication protocol between the relay daemon and live viewers
7217 which makes it possible to see events "live", as they are received by
7220 <<lttng-modules,LTTng-modules>>::
7221 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7222 which contains the Linux kernel modules to make the Linux kernel
7223 instrumentation points available for LTTng tracing.
7226 The name of the relay daemon program.
7228 cmd:lttng-sessiond::
7229 The name of the session daemon program.
7232 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7233 contains the various programs and libraries used to
7234 <<controlling-tracing,control tracing>>.
7236 <<lttng-ust,LTTng-UST>>::
7237 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7238 contains libraries to instrument user applications.
7240 <<lttng-ust-agents,LTTng-UST Java agent>>::
7241 A Java package provided by the LTTng-UST project to allow the
7242 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7245 <<lttng-ust-agents,LTTng-UST Python agent>>::
7246 A Python package provided by the LTTng-UST project to allow the
7247 LTTng instrumentation of Python logging statements.
7249 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7250 The event loss mode in which new event records overwrite older
7251 event records when there's no sub-buffer space left to store them.
7253 <<channel-buffering-schemes,per-process buffering>>::
7254 A buffering scheme in which each instrumented process has its own
7255 sub-buffers for a given user space channel.
7257 <<channel-buffering-schemes,per-user buffering>>::
7258 A buffering scheme in which all the processes of a Unix user share the
7259 same sub-buffer for a given user space channel.
7261 <<lttng-relayd,relay daemon>>::
7262 A process which is responsible for receiving the trace data sent by
7263 a distant consumer daemon.
7266 A set of sub-buffers.
7268 <<lttng-sessiond,session daemon>>::
7269 A process which receives control commands from you and orchestrates
7270 the tracers and various LTTng daemons.
7272 <<taking-a-snapshot,snapshot>>::
7273 A copy of the current data of all the sub-buffers of a given tracing
7274 session, saved as trace files.
7277 One part of an LTTng ring buffer which contains event records.
7280 The time information attached to an event when it is emitted.
7283 A set of files which are the concatenations of one or more
7284 flushed sub-buffers.
7287 The action of recording the events emitted by an application
7288 or by a system, or to initiate such recording by controlling
7292 The http://tracecompass.org[Trace Compass] project and application.
7295 An instrumentation point using the tracepoint mechanism of the Linux
7296 kernel or of LTTng-UST.
7298 tracepoint definition::
7299 The definition of a single tracepoint.
7302 The name of a tracepoint.
7304 tracepoint provider::
7305 A set of functions providing tracepoints to an instrumented user
7308 Not to be confused with a _tracepoint provider package_: many tracepoint
7309 providers can exist within a tracepoint provider package.
7311 tracepoint provider package::
7312 One or more tracepoint providers compiled as an object file or as
7316 A software which records emitted events.
7318 <<domain,tracing domain>>::
7319 A namespace for event sources.
7322 The Unix group in which a Unix user can be to be allowed to trace the
7325 <<tracing-session,tracing session>>::
7326 A stateful dialogue between you and a <<lttng-sessiond,session
7330 An application running in user space, as opposed to a Linux kernel
7331 module, for example.