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.
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 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
386 other Ubuntu releases.
390 |LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} for
391 Fedora{nbsp}25 and Fedora{nbsp}26 (both are not released yet).
393 <<building-from-source,Build LTTng-modules{nbsp}{revision} from
396 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
397 other Fedora releases.
400 |<<debian,Debian "stretch" (testing)>>
402 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
403 previous Debian releases.
407 |LTTng{nbsp}2.7 for openSUSE Leap{nbsp}42.1.
409 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
410 other openSUSE releases.
413 |<<arch-linux,Latest AUR packages>>
417 |<<alpine-linux,Alpine Linux "edge">>
418 |LTTng{nbsp}{revision} for Alpine Linux{nbsp}3.5 (not released yet).
420 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
421 other Alpine Linux releases.
424 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
429 |LTTng{nbsp}{revision} for Buildroot{nbsp}2016.11 (not released yet).
431 LTTng{nbsp}2.7 for Buildroot{nbsp}2016.02, Buildroot{nbsp}2016.05,
432 and Buildroot{nbsp}2016.08.
434 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
435 other Buildroot releases.
437 |OpenEmbedded and Yocto
438 |<<oe-yocto,Yocto Project{nbsp}2.2 _Morty_>> (`openembedded-core` layer)
439 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
440 other OpenEmbedded releases.
445 === [[ubuntu-official-repositories]]Ubuntu
447 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.10 _Yakkety Yak_.
448 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
449 Stable{nbsp}{revision} PPA>>.
451 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}16.10 _Yakkety Yak_:
453 . Install the main LTTng{nbsp}{revision} packages:
458 sudo apt-get install lttng-tools
459 sudo apt-get install lttng-modules-dkms
460 sudo apt-get install liblttng-ust-dev
464 . **If you need to instrument and trace
465 <<java-application,Java applications>>**, install the LTTng-UST
471 sudo apt-get install liblttng-ust-agent-java
475 . **If you need to instrument and trace
476 <<python-application,Python applications>>**, install the
477 LTTng-UST Python agent:
482 sudo apt-get install python3-lttngust
488 ==== noch:{LTTng} Stable {revision} PPA
490 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
491 Stable{nbsp}{revision} PPA] offers the latest stable
492 LTTng{nbsp}{revision} packages for:
494 * Ubuntu{nbsp}14.04 _Trusty Tahr_
495 * Ubuntu{nbsp}16.04 _Xenial Xerus_
497 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
499 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
505 sudo apt-add-repository ppa:lttng/stable-2.8
510 . Install the main LTTng{nbsp}{revision} packages:
515 sudo apt-get install lttng-tools
516 sudo apt-get install lttng-modules-dkms
517 sudo apt-get install liblttng-ust-dev
521 . **If you need to instrument and trace
522 <<java-application,Java applications>>**, install the LTTng-UST
528 sudo apt-get install liblttng-ust-agent-java
532 . **If you need to instrument and trace
533 <<python-application,Python applications>>**, install the
534 LTTng-UST Python agent:
539 sudo apt-get install python3-lttngust
547 To install LTTng{nbsp}{revision} on Debian "stretch" (testing):
549 . Install the main LTTng{nbsp}{revision} packages:
554 sudo apt-get install lttng-modules-dkms
555 sudo apt-get install liblttng-ust-dev
556 sudo apt-get install lttng-tools
560 . **If you need to instrument and trace <<java-application,Java
561 applications>>**, install the LTTng-UST Java agent:
566 sudo apt-get install liblttng-ust-agent-java
570 . **If you need to instrument and trace <<python-application,Python
571 applications>>**, install the LTTng-UST Python agent:
576 sudo apt-get install python3-lttngust
584 To install LTTng{nbsp}{revision} on Arch Linux using
585 https://archlinux.fr/yaourt-en[Yaourt]:
587 . Install the main LTTng{nbsp}{revision} packages:
592 yaourt -S lttng-tools
594 yaourt -S lttng-modules
598 . **If you need to instrument and trace <<python-application,Python
599 applications>>**, install the LTTng-UST Python agent:
604 yaourt -S python-lttngust
605 yaourt -S python2-lttngust
613 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
616 . Make sure your system is
617 https://wiki.alpinelinux.org/wiki/Edge[configured for "edge"].
618 . Enable the _testing_ repository by uncommenting the corresponding
619 line in path:{/etc/apk/repositories}.
620 . Add the LTTng packages:
625 sudo apk add lttng-tools
626 sudo apk add lttng-ust-dev
630 To install LTTng-modules{nbsp}{revision} (Linux kernel tracing support)
631 on Alpine Linux "edge":
633 . Add the vanilla Linux kernel:
638 apk add linux-vanilla linux-vanilla-dev
642 . Reboot with the vanilla Linux kernel.
643 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
649 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
650 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
651 cd lttng-modules-2.8.* &&
653 sudo make modules_install &&
660 === OpenEmbedded and Yocto
662 LTTng{nbsp}{revision} recipes are available in the
663 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
664 layer for Yocto Project{nbsp}2.2 _Morty_ under the following names:
670 With BitBake, the simplest way to include LTTng recipes in your target
671 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
674 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
679 . Select a machine and an image recipe.
680 . Click **Edit image recipe**.
681 . Under the **All recipes** tab, search for **lttng**.
682 . Check the desired LTTng recipes.
685 .Java and Python application instrumentation and tracing
687 If you need to instrument and trace <<java-application,Java
688 applications>> on openSUSE, you need to build and install
689 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
690 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
691 `--enable-java-agent-all` options to the `configure` script, depending
692 on which Java logging framework you use.
694 If you need to instrument and trace <<python-application,Python
695 applications>> on openSUSE, you need to build and install
696 LTTng-UST{nbsp}{revision} from source and pass the
697 `--enable-python-agent` option to the `configure` script.
701 [[enterprise-distributions]]
702 === RHEL, SUSE, and other enterprise distributions
704 To install LTTng on enterprise Linux distributions, such as Red Hat
705 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
706 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
709 [[building-from-source]]
710 === Build from source
712 To build and install LTTng{nbsp}{revision} from source:
714 . Using your distribution's package manager, or from source, install
715 the following dependencies of LTTng-tools and LTTng-UST:
718 * https://sourceforge.net/projects/libuuid/[libuuid]
719 * http://directory.fsf.org/wiki/Popt[popt]
720 * http://liburcu.org/[Userspace RCU]
721 * http://www.xmlsoft.org/[libxml2]
724 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
730 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
731 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
732 cd lttng-modules-2.8.* &&
734 sudo make modules_install &&
739 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
745 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
746 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
747 cd lttng-ust-2.8.* &&
757 .Java and Python application tracing
759 If you need to instrument and trace <<java-application,Java
760 applications>>, pass the `--enable-java-agent-jul`,
761 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
762 `configure` script, depending on which Java logging framework you use.
764 If you need to instrument and trace <<python-application,Python
765 applications>>, pass the `--enable-python-agent` option to the
766 `configure` script. You can set the `PYTHON` environment variable to the
767 path to the Python interpreter for which to install the LTTng-UST Python
775 By default, LTTng-UST libraries are installed to
776 dir:{/usr/local/lib}, which is the de facto directory in which to
777 keep self-compiled and third-party libraries.
779 When <<building-tracepoint-providers-and-user-application,linking an
780 instrumented user application with `liblttng-ust`>>:
782 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
784 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
785 man:gcc(1), man:g++(1), or man:clang(1).
789 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
795 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
796 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
797 cd lttng-tools-2.8.* &&
805 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
806 previous steps automatically for a given version of LTTng and confine
807 the installed files in a specific directory. This can be useful to test
808 LTTng without installing it on your system.
814 This is a short guide to get started quickly with LTTng kernel and user
817 Before you follow this guide, make sure to <<installing-lttng,install>>
820 This tutorial walks you through the steps to:
822 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
823 . <<tracing-your-own-user-application,Trace a user application>> written
825 . <<viewing-and-analyzing-your-traces,View and analyze the
829 [[tracing-the-linux-kernel]]
830 === Trace the Linux kernel
832 The following command lines start with cmd:sudo because you need root
833 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
834 your Unix user is a member of the <<lttng-sessiond,tracing group>>.
836 . Create a <<tracing-session,tracing session>>:
841 sudo lttng create my-kernel-session
845 . List the available kernel tracepoints and system calls:
854 . Create an <<event,event rule>> which matches the desired event names,
855 for example `sched_switch` and `sched_process_fork`:
860 sudo lttng enable-event --kernel sched_switch,sched_process_fork
864 You can also create an event rule which _matches_ all the Linux kernel
865 tracepoints (this will generate a lot of data when tracing):
870 sudo lttng enable-event --kernel --all
883 . Do some operation on your system for a few seconds. For example,
884 load a website, or list the files of a directory.
885 . Stop tracing and destroy the tracing session:
895 The man:lttng-destroy(1) command does not destroy the trace data; it
896 only destroys the state of the tracing session.
898 By default, LTTng saves the traces in
899 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
900 where +__name__+ is the tracing session name. Note that the
901 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
903 See <<viewing-and-analyzing-your-traces,View and analyze the
904 recorded events>> to view the recorded events.
907 [[tracing-your-own-user-application]]
908 === Trace a user application
910 This section steps you through a simple example to trace a
911 _Hello world_ program written in C.
913 To create the traceable user application:
915 . Create the tracepoint provider header file, which defines the
916 tracepoints and the events they can generate:
922 #undef TRACEPOINT_PROVIDER
923 #define TRACEPOINT_PROVIDER hello_world
925 #undef TRACEPOINT_INCLUDE
926 #define TRACEPOINT_INCLUDE "./hello-tp.h"
928 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
931 #include <lttng/tracepoint.h>
941 ctf_string(my_string_field, my_string_arg)
942 ctf_integer(int, my_integer_field, my_integer_arg)
946 #endif /* _HELLO_TP_H */
948 #include <lttng/tracepoint-event.h>
952 . Create the tracepoint provider package source file:
958 #define TRACEPOINT_CREATE_PROBES
959 #define TRACEPOINT_DEFINE
961 #include "hello-tp.h"
965 . Build the tracepoint provider package:
970 gcc -c -I. hello-tp.c
974 . Create the _Hello World_ application source file:
981 #include "hello-tp.h"
983 int main(int argc, char *argv[])
987 puts("Hello, World!\nPress Enter to continue...");
990 * The following getchar() call is only placed here for the purpose
991 * of this demonstration, to pause the application in order for
992 * you to have time to list its tracepoints. It is not
998 * A tracepoint() call.
1000 * Arguments, as defined in hello-tp.h:
1002 * 1. Tracepoint provider name (required)
1003 * 2. Tracepoint name (required)
1004 * 3. my_integer_arg (first user-defined argument)
1005 * 4. my_string_arg (second user-defined argument)
1007 * Notice the tracepoint provider and tracepoint names are
1008 * NOT strings: they are in fact parts of variables that the
1009 * macros in hello-tp.h create.
1011 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
1013 for (x = 0; x < argc; ++x) {
1014 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
1017 puts("Quitting now!");
1018 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
1025 . Build the application:
1034 . Link the application with the tracepoint provider package,
1035 `liblttng-ust`, and `libdl`:
1040 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1044 Here's the whole build process:
1047 .User space tracing tutorial's build steps.
1048 image::ust-flow.png[]
1050 To trace the user application:
1052 . Run the application with a few arguments:
1057 ./hello world and beyond
1066 Press Enter to continue...
1070 . Start an LTTng <<lttng-sessiond,session daemon>>:
1075 lttng-sessiond --daemonize
1079 Note that a session daemon might already be running, for example as
1080 a service that the distribution's service manager started.
1082 . List the available user space tracepoints:
1087 lttng list --userspace
1091 You see the `hello_world:my_first_tracepoint` tracepoint listed
1092 under the `./hello` process.
1094 . Create a <<tracing-session,tracing session>>:
1099 lttng create my-user-space-session
1103 . Create an <<event,event rule>> which matches the
1104 `hello_world:my_first_tracepoint` event name:
1109 lttng enable-event --userspace hello_world:my_first_tracepoint
1122 . Go back to the running `hello` application and press Enter. The
1123 program executes all `tracepoint()` instrumentation points and exits.
1124 . Stop tracing and destroy the tracing session:
1134 The man:lttng-destroy(1) command does not destroy the trace data; it
1135 only destroys the state of the tracing session.
1137 By default, LTTng saves the traces in
1138 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1139 where +__name__+ is the tracing session name. Note that the
1140 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1142 See <<viewing-and-analyzing-your-traces,View and analyze the
1143 recorded events>> to view the recorded events.
1146 [[viewing-and-analyzing-your-traces]]
1147 === View and analyze the recorded events
1149 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1150 kernel>> and <<tracing-your-own-user-application,Trace a user
1151 application>> tutorials, you can inspect the recorded events.
1153 Many tools are available to read LTTng traces:
1155 * **cmd:babeltrace** is a command-line utility which converts trace
1156 formats; it supports the format that LTTng produces, CTF, as well as a
1157 basic text output which can be ++grep++ed. The cmd:babeltrace command
1158 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1159 * Babeltrace also includes
1160 **https://www.python.org/[Python] bindings** so
1161 that you can easily open and read an LTTng trace with your own script,
1162 benefiting from the power of Python.
1163 * http://tracecompass.org/[**Trace Compass**]
1164 is a graphical user interface for viewing and analyzing any type of
1165 logs or traces, including LTTng's.
1166 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1167 project which includes many high-level analyses of LTTng kernel
1168 traces, like scheduling statistics, interrupt frequency distribution,
1169 top CPU usage, and more.
1171 NOTE: This section assumes that the traces recorded during the previous
1172 tutorials were saved to their default location, in the
1173 dir:{$LTTNG_HOME/lttng-traces} directory. Note that the env:LTTNG_HOME
1174 environment variable defaults to `$HOME` if not set.
1177 [[viewing-and-analyzing-your-traces-bt]]
1178 ==== Use the cmd:babeltrace command-line tool
1180 The simplest way to list all the recorded events of a trace is to pass
1181 its path to cmd:babeltrace with no options:
1185 babeltrace ~/lttng-traces/my-user-space-session*
1188 cmd:babeltrace finds all traces recursively within the given path and
1189 prints all their events, merging them in chronological order.
1191 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1196 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_
1199 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1200 count the recorded events:
1204 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_read | wc --lines
1208 [[viewing-and-analyzing-your-traces-bt-python]]
1209 ==== Use the Babeltrace Python bindings
1211 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1212 is useful to isolate events by simple matching using man:grep(1) and
1213 similar utilities. However, more elaborate filters, such as keeping only
1214 event records with a field value falling within a specific range, are
1215 not trivial to write using a shell. Moreover, reductions and even the
1216 most basic computations involving multiple event records are virtually
1217 impossible to implement.
1219 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1220 to read the event records of an LTTng trace sequentially and compute the
1221 desired information.
1223 The following script accepts an LTTng Linux kernel trace path as its
1224 first argument and prints the short names of the top 5 running processes
1225 on CPU 0 during the whole trace:
1230 from collections import Counter
1236 if len(sys.argv) != 2:
1237 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1238 print(msg, file=sys.stderr)
1241 # A trace collection contains one or more traces
1242 col = babeltrace.TraceCollection()
1244 # Add the trace provided by the user (LTTng traces always have
1246 if col.add_trace(sys.argv[1], 'ctf') is None:
1247 raise RuntimeError('Cannot add trace')
1249 # This counter dict contains execution times:
1251 # task command name -> total execution time (ns)
1252 exec_times = Counter()
1254 # This contains the last `sched_switch` timestamp
1258 for event in col.events:
1259 # Keep only `sched_switch` events
1260 if event.name != 'sched_switch':
1263 # Keep only events which happened on CPU 0
1264 if event['cpu_id'] != 0:
1268 cur_ts = event.timestamp
1274 # Previous task command (short) name
1275 prev_comm = event['prev_comm']
1277 # Initialize entry in our dict if not yet done
1278 if prev_comm not in exec_times:
1279 exec_times[prev_comm] = 0
1281 # Compute previous command execution time
1282 diff = cur_ts - last_ts
1284 # Update execution time of this command
1285 exec_times[prev_comm] += diff
1287 # Update last timestamp
1291 for name, ns in exec_times.most_common(5):
1293 print('{:20}{} s'.format(name, s))
1298 if __name__ == '__main__':
1299 sys.exit(0 if top5proc() else 1)
1306 python3 top5proc.py ~/lttng-traces/my-kernel-session*/kernel
1312 swapper/0 48.607245889 s
1313 chromium 7.192738188 s
1314 pavucontrol 0.709894415 s
1315 Compositor 0.660867933 s
1316 Xorg.bin 0.616753786 s
1319 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1320 weren't using the CPU that much when tracing, its first position in the
1325 == [[understanding-lttng]]Core concepts
1327 From a user's perspective, the LTTng system is built on a few concepts,
1328 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1329 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1330 Understanding how those objects relate to eachother is key in mastering
1333 The core concepts are:
1335 * <<tracing-session,Tracing session>>
1336 * <<domain,Tracing domain>>
1337 * <<channel,Channel and ring buffer>>
1338 * <<"event","Instrumentation point, event rule, event, and event record">>
1344 A _tracing session_ is a stateful dialogue between you and
1345 a <<lttng-sessiond,session daemon>>. You can
1346 <<creating-destroying-tracing-sessions,create a new tracing
1347 session>> with the `lttng create` command.
1349 Anything that you do when you control LTTng tracers happens within a
1350 tracing session. In particular, a tracing session:
1353 * Has its own set of trace files.
1354 * Has its own state of activity (started or stopped).
1355 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1357 * Has its own <<channel,channels>> which have their own
1358 <<event,event rules>>.
1361 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1362 image::concepts.png[]
1364 Those attributes and objects are completely isolated between different
1367 A tracing session is analogous to a cash machine session:
1368 the operations you do on the banking system through the cash machine do
1369 not alter the data of other users of the same system. In the case of
1370 the cash machine, a session lasts as long as your bank card is inside.
1371 In the case of LTTng, a tracing session lasts from the `lttng create`
1372 command to the `lttng destroy` command.
1375 .Each Unix user has its own set of tracing sessions.
1376 image::many-sessions.png[]
1379 [[tracing-session-mode]]
1380 ==== Tracing session mode
1382 LTTng can send the generated trace data to different locations. The
1383 _tracing session mode_ dictates where to send it. The following modes
1384 are available in LTTng{nbsp}{revision}:
1387 LTTng writes the traces to the file system of the machine being traced
1390 Network streaming mode::
1391 LTTng sends the traces over the network to a
1392 <<lttng-relayd,relay daemon>> running on a remote system.
1395 LTTng does not write the traces by default. Instead, you can request
1396 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1397 current tracing buffers, and to write it to the target's file system
1398 or to send it over the network to a <<lttng-relayd,relay daemon>>
1399 running on a remote system.
1402 This mode is similar to the network streaming mode, but a live
1403 trace viewer can connect to the distant relay daemon to
1404 <<lttng-live,view event records as LTTng generates them>> by
1411 A _tracing domain_ is a namespace for event sources. A tracing domain
1412 has its own properties and features.
1414 There are currently five available tracing domains:
1418 * `java.util.logging` (JUL)
1422 You must specify a tracing domain when using some commands to avoid
1423 ambiguity. For example, since all the domains support named tracepoints
1424 as event sources (instrumentation points that you manually insert in the
1425 source code), you need to specify a tracing domain when
1426 <<enabling-disabling-events,creating an event rule>> because all the
1427 tracing domains could have tracepoints with the same names.
1429 Some features are reserved to specific tracing domains. Dynamic function
1430 entry and return instrumentation points, for example, are currently only
1431 supported in the Linux kernel tracing domain, but support for other
1432 tracing domains could be added in the future.
1434 You can create <<channel,channels>> in the Linux kernel and user space
1435 tracing domains. The other tracing domains have a single default
1440 === Channel and ring buffer
1442 A _channel_ is an object which is responsible for a set of ring buffers.
1443 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1444 tracer emits an event, it can record it to one or more
1445 sub-buffers. The attributes of a channel determine what to do when
1446 there's no space left for a new event record because all sub-buffers
1447 are full, where to send a full sub-buffer, and other behaviours.
1449 A channel is always associated to a <<domain,tracing domain>>. The
1450 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1451 a default channel which you cannot configure.
1453 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1454 an event, it records it to the sub-buffers of all
1455 the enabled channels with a satisfied event rule, as long as those
1456 channels are part of active <<tracing-session,tracing sessions>>.
1459 [[channel-buffering-schemes]]
1460 ==== Per-user vs. per-process buffering schemes
1462 A channel has at least one ring buffer _per CPU_. LTTng always
1463 records an event to the ring buffer associated to the CPU on which it
1466 Two _buffering schemes_ are available when you
1467 <<enabling-disabling-channels,create a channel>> in the
1468 user space <<domain,tracing domain>>:
1470 Per-user buffering::
1471 Allocate one set of ring buffers--one per CPU--shared by all the
1472 instrumented processes of each Unix user.
1476 .Per-user buffering scheme.
1477 image::per-user-buffering.png[]
1480 Per-process buffering::
1481 Allocate one set of ring buffers--one per CPU--for each
1482 instrumented process.
1486 .Per-process buffering scheme.
1487 image::per-process-buffering.png[]
1490 The per-process buffering scheme tends to consume more memory than the
1491 per-user option because systems generally have more instrumented
1492 processes than Unix users running instrumented processes. However, the
1493 per-process buffering scheme ensures that one process having a high
1494 event throughput won't fill all the shared sub-buffers of the same
1497 The Linux kernel tracing domain has only one available buffering scheme
1498 which is to allocate a single set of ring buffers for the whole system.
1499 This scheme is similar to the per-user option, but with a single, global
1500 user "running" the kernel.
1503 [[channel-overwrite-mode-vs-discard-mode]]
1504 ==== Overwrite vs. discard event loss modes
1506 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1507 arc in the following animation) of a specific channel's ring buffer.
1508 When there's no space left in a sub-buffer, the tracer marks it as
1509 consumable (red) and another, empty sub-buffer starts receiving the
1510 following event records. A <<lttng-consumerd,consumer daemon>>
1511 eventually consumes the marked sub-buffer (returns to white).
1514 [role="docsvg-channel-subbuf-anim"]
1519 In an ideal world, sub-buffers are consumed faster than they are filled,
1520 as is the case in the previous animation. In the real world,
1521 however, all sub-buffers can be full at some point, leaving no space to
1522 record the following events.
1524 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1525 available, it is acceptable to lose event records when the alternative
1526 would be to cause substantial delays in the instrumented application's
1527 execution. LTTng privileges performance over integrity; it aims at
1528 perturbing the traced system as little as possible in order to make
1529 tracing of subtle race conditions and rare interrupt cascades possible.
1531 When it comes to losing event records because no empty sub-buffer is
1532 available, the channel's _event loss mode_ determines what to do. The
1533 available event loss modes are:
1536 Drop the newest event records until a the tracer
1537 releases a sub-buffer.
1540 Clear the sub-buffer containing the oldest event records and start
1541 writing the newest event records there.
1543 This mode is sometimes called _flight recorder mode_ because it's
1545 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1546 always keep a fixed amount of the latest data.
1548 Which mechanism you should choose depends on your context: prioritize
1549 the newest or the oldest event records in the ring buffer?
1551 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1552 as soon as a there's no space left for a new event record, whereas in
1553 discard mode, the tracer only discards the event record that doesn't
1556 In discard mode, LTTng increments a count of lost event records when
1557 an event record is lost and saves this count to the trace. In
1558 overwrite mode, LTTng keeps no information when it overwrites a
1559 sub-buffer before consuming it.
1561 There are a few ways to decrease your probability of losing event
1563 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1564 how you can fine-une the sub-buffer count and size of a channel to
1565 virtually stop losing event records, though at the cost of greater
1569 [[channel-subbuf-size-vs-subbuf-count]]
1570 ==== Sub-buffer count and size
1572 When you <<enabling-disabling-channels,create a channel>>, you can
1573 set its number of sub-buffers and their size.
1575 Note that there is noticeable CPU overhead introduced when
1576 switching sub-buffers (marking a full one as consumable and switching
1577 to an empty one for the following events to be recorded). Knowing this,
1578 the following list presents a few practical situations along with how
1579 to configure the sub-buffer count and size for them:
1581 * **High event throughput**: In general, prefer bigger sub-buffers to
1582 lower the risk of losing event records.
1584 Having bigger sub-buffers also ensures a lower sub-buffer switching
1587 The number of sub-buffers is only meaningful if you create the channel
1588 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1589 other sub-buffers are left unaltered.
1591 * **Low event throughput**: In general, prefer smaller sub-buffers
1592 since the risk of losing event records is low.
1594 Because events occur less frequently, the sub-buffer switching frequency
1595 should remain low and thus the tracer's overhead should not be a
1598 * **Low memory system**: If your target system has a low memory
1599 limit, prefer fewer first, then smaller sub-buffers.
1601 Even if the system is limited in memory, you want to keep the
1602 sub-buffers as big as possible to avoid a high sub-buffer switching
1605 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1606 which means event data is very compact. For example, the average
1607 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1608 sub-buffer size of 1{nbsp}MiB is considered big.
1610 The previous situations highlight the major trade-off between a few big
1611 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1612 frequency vs. how much data is lost in overwrite mode. Assuming a
1613 constant event throughput and using the overwrite mode, the two
1614 following configurations have the same ring buffer total size:
1617 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1622 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1623 switching frequency, but if a sub-buffer overwrite happens, half of
1624 the event records so far (4{nbsp}MiB) are definitely lost.
1625 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1626 overhead as the previous configuration, but if a sub-buffer
1627 overwrite happens, only the eighth of event records so far are
1630 In discard mode, the sub-buffers count parameter is pointless: use two
1631 sub-buffers and set their size according to the requirements of your
1635 [[channel-switch-timer]]
1636 ==== Switch timer period
1638 The _switch timer period_ is an important configurable attribute of
1639 a channel to ensure periodic sub-buffer flushing.
1641 When the _switch timer_ expires, a sub-buffer switch happens. You can
1642 set the switch timer period attribute when you
1643 <<enabling-disabling-channels,create a channel>> to ensure that event
1644 data is consumed and committed to trace files or to a distant relay
1645 daemon periodically in case of a low event throughput.
1648 [role="docsvg-channel-switch-timer"]
1653 This attribute is also convenient when you use big sub-buffers to cope
1654 with a sporadic high event throughput, even if the throughput is
1658 [[channel-read-timer]]
1659 ==== Read timer period
1661 By default, the LTTng tracers use a notification mechanism to signal a
1662 full sub-buffer so that a consumer daemon can consume it. When such
1663 notifications must be avoided, for example in real-time applications,
1664 you can use the channel's _read timer_ instead. When the read timer
1665 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1666 consumable sub-buffers.
1669 [[tracefile-rotation]]
1670 ==== Trace file count and size
1672 By default, trace files can grow as large as needed. You can set the
1673 maximum size of each trace file that a channel writes when you
1674 <<enabling-disabling-channels,create a channel>>. When the size of
1675 a trace file reaches the channel's fixed maximum size, LTTng creates
1676 another file to contain the next event records. LTTng appends a file
1677 count to each trace file name in this case.
1679 If you set the trace file size attribute when you create a channel, the
1680 maximum number of trace files that LTTng creates is _unlimited_ by
1681 default. To limit them, you can also set a maximum number of trace
1682 files. When the number of trace files reaches the channel's fixed
1683 maximum count, the oldest trace file is overwritten. This mechanism is
1684 called _trace file rotation_.
1688 === Instrumentation point, event rule, event, and event record
1690 An _event rule_ is a set of conditions which must be **all** satisfied
1691 for LTTng to record an occuring event.
1693 You set the conditions when you <<enabling-disabling-events,create
1696 You always attach an event rule to <<channel,channel>> when you create
1699 When an event passes the conditions of an event rule, LTTng records it
1700 in one of the attached channel's sub-buffers.
1702 The available conditions, as of LTTng{nbsp}{revision}, are:
1704 * The event rule _is enabled_.
1705 * The instrumentation point's type _is{nbsp}T_.
1706 * The instrumentation point's name (sometimes called _event name_)
1707 _matches{nbsp}N_, but _is not{nbsp}E_.
1708 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1709 _is exactly{nbsp}L_.
1710 * The fields of the event's payload _satisfy_ a filter
1711 expression{nbsp}__F__.
1713 As you can see, all the conditions but the dynamic filter are related to
1714 the event rule's status or to the instrumentation point, not to the
1715 occurring events. This is why, without a filter, checking if an event
1716 passes an event rule is not a dynamic task: when you create or modify an
1717 event rule, all the tracers of its tracing domain enable or disable the
1718 instrumentation points themselves once. This is possible because the
1719 attributes of an instrumentation point (type, name, and log level) are
1720 defined statically. In other words, without a dynamic filter, the tracer
1721 _does not evaluate_ the arguments of an instrumentation point unless it
1722 matches an enabled event rule.
1724 Note that, for LTTng to record an event, the <<channel,channel>> to
1725 which a matching event rule is attached must also be enabled, and the
1726 tracing session owning this channel must be active.
1729 .Logical path from an instrumentation point to an event record.
1730 image::event-rule.png[]
1732 .Event, event record, or event rule?
1734 With so many similar terms, it's easy to get confused.
1736 An **event** is the consequence of the execution of an _instrumentation
1737 point_, like a tracepoint that you manually place in some source code,
1738 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1739 time. Different actions can be taken upon the occurance of an event,
1740 like record the event's payload to a buffer.
1742 An **event record** is the representation of an event in a sub-buffer. A
1743 tracer is responsible for capturing the payload of an event, current
1744 context variables, the event's ID, and the event's timestamp. LTTng
1745 can append this sub-buffer to a trace file.
1747 An **event rule** is a set of conditions which must all be satisfied for
1748 LTTng to record an occuring event. Events still occur without
1749 satisfying event rules, but LTTng does not record them.
1754 == Components of noch:{LTTng}
1756 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1757 to call LTTng a simple _tool_ since it is composed of multiple
1758 interacting components. This section describes those components,
1759 explains their respective roles, and shows how they connect together to
1760 form the LTTng ecosystem.
1762 The following diagram shows how the most important components of LTTng
1763 interact with user applications, the Linux kernel, and you:
1766 .Control and trace data paths between LTTng components.
1767 image::plumbing.png[]
1769 The LTTng project incorporates:
1771 * **LTTng-tools**: Libraries and command-line interface to
1772 control tracing sessions.
1773 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1774 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1775 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1776 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1777 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1778 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1780 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1781 headers to instrument and trace any native user application.
1782 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1783 *** `liblttng-ust-libc-wrapper`
1784 *** `liblttng-ust-pthread-wrapper`
1785 *** `liblttng-ust-cyg-profile`
1786 *** `liblttng-ust-cyg-profile-fast`
1787 *** `liblttng-ust-dl`
1788 ** User space tracepoint provider source files generator command-line
1789 tool (man:lttng-gen-tp(1)).
1790 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1791 Java applications using `java.util.logging` or
1792 Apache log4j 1.2 logging.
1793 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1794 Python applications using the standard `logging` package.
1795 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1797 ** LTTng kernel tracer module.
1798 ** Tracing ring buffer kernel modules.
1799 ** Probe kernel modules.
1800 ** LTTng logger kernel module.
1804 === Tracing control command-line interface
1807 .The tracing control command-line interface.
1808 image::plumbing-lttng-cli.png[]
1810 The _man:lttng(1) command-line tool_ is the standard user interface to
1811 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1812 is part of LTTng-tools.
1814 The cmd:lttng tool is linked with
1815 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1816 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1818 The cmd:lttng tool has a Git-like interface:
1822 lttng <general options> <command> <command options>
1825 The <<controlling-tracing,Tracing control>> section explores the
1826 available features of LTTng using the cmd:lttng tool.
1829 [[liblttng-ctl-lttng]]
1830 === Tracing control library
1833 .The tracing control library.
1834 image::plumbing-liblttng-ctl.png[]
1836 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1837 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1838 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1840 The <<lttng-cli,cmd:lttng command-line tool>>
1841 is linked with `liblttng-ctl`.
1843 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1848 #include <lttng/lttng.h>
1851 Some objects are referenced by name (C string), such as tracing
1852 sessions, but most of them require to create a handle first using
1853 `lttng_create_handle()`.
1855 The best available developer documentation for `liblttng-ctl` is, as of
1856 LTTng{nbsp}{revision}, its installed header files. Every function and
1857 structure is thoroughly documented.
1861 === User space tracing library
1864 .The user space tracing library.
1865 image::plumbing-liblttng-ust.png[]
1867 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1868 is the LTTng user space tracer. It receives commands from a
1869 <<lttng-sessiond,session daemon>>, for example to
1870 enable and disable specific instrumentation points, and writes event
1871 records to ring buffers shared with a
1872 <<lttng-consumerd,consumer daemon>>.
1873 `liblttng-ust` is part of LTTng-UST.
1875 Public C header files are installed beside `liblttng-ust` to
1876 instrument any <<c-application,C or $$C++$$ application>>.
1878 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1879 packages, use their own library providing tracepoints which is
1880 linked with `liblttng-ust`.
1882 An application or library does not have to initialize `liblttng-ust`
1883 manually: its constructor does the necessary tasks to properly register
1884 to a session daemon. The initialization phase also enables the
1885 instrumentation points matching the <<event,event rules>> that you
1889 [[lttng-ust-agents]]
1890 === User space tracing agents
1893 .The user space tracing agents.
1894 image::plumbing-lttng-ust-agents.png[]
1896 The _LTTng-UST Java and Python agents_ are regular Java and Python
1897 packages which add LTTng tracing capabilities to the
1898 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1900 In the case of Java, the
1901 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1902 core logging facilities] and
1903 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1904 Note that Apache Log4{nbsp}2 is not supported.
1906 In the case of Python, the standard
1907 https://docs.python.org/3/library/logging.html[`logging`] package
1908 is supported. Both Python 2 and Python 3 modules can import the
1909 LTTng-UST Python agent package.
1911 The applications using the LTTng-UST agents are in the
1912 `java.util.logging` (JUL),
1913 log4j, and Python <<domain,tracing domains>>.
1915 Both agents use the same mechanism to trace the log statements. When an
1916 agent is initialized, it creates a log handler that attaches to the root
1917 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1918 When the application executes a log statement, it is passed to the
1919 agent's log handler by the root logger. The agent's log handler calls a
1920 native function in a tracepoint provider package shared library linked
1921 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1922 other fields, like its logger name and its log level. This native
1923 function contains a user space instrumentation point, hence tracing the
1926 The log level condition of an
1927 <<event,event rule>> is considered when tracing
1928 a Java or a Python application, and it's compatible with the standard
1929 JUL, log4j, and Python log levels.
1933 === LTTng kernel modules
1936 .The LTTng kernel modules.
1937 image::plumbing-lttng-modules.png[]
1939 The _LTTng kernel modules_ are a set of Linux kernel modules
1940 which implement the kernel tracer of the LTTng project. The LTTng
1941 kernel modules are part of LTTng-modules.
1943 The LTTng kernel modules include:
1945 * A set of _probe_ modules.
1947 Each module attaches to a specific subsystem
1948 of the Linux kernel using its tracepoint instrument points. There are
1949 also modules to attach to the entry and return points of the Linux
1950 system call functions.
1952 * _Ring buffer_ modules.
1954 A ring buffer implementation is provided as kernel modules. The LTTng
1955 kernel tracer writes to the ring buffer; a
1956 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1958 * The _LTTng kernel tracer_ module.
1959 * The _LTTng logger_ module.
1961 The LTTng logger module implements the special path:{/proc/lttng-logger}
1962 file so that any executable can generate LTTng events by opening and
1963 writing to this file.
1965 See <<proc-lttng-logger-abi,LTTng logger>>.
1967 Generally, you do not have to load the LTTng kernel modules manually
1968 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1969 daemon>> loads the necessary modules when starting. If you have extra
1970 probe modules, you can specify to load them to the session daemon on
1973 The LTTng kernel modules are installed in
1974 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1975 the kernel release (see `uname --kernel-release`).
1982 .The session daemon.
1983 image::plumbing-sessiond.png[]
1985 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1986 managing tracing sessions and for controlling the various components of
1987 LTTng. The session daemon is part of LTTng-tools.
1989 The session daemon sends control requests to and receives control
1992 * The <<lttng-ust,user space tracing library>>.
1994 Any instance of the user space tracing library first registers to
1995 a session daemon. Then, the session daemon can send requests to
1996 this instance, such as:
1999 ** Get the list of tracepoints.
2000 ** Share an <<event,event rule>> so that the user space tracing library
2001 can enable or disable tracepoints. Amongst the possible conditions
2002 of an event rule is a filter expression which `liblttng-ust` evalutes
2003 when an event occurs.
2004 ** Share <<channel,channel>> attributes and ring buffer locations.
2007 The session daemon and the user space tracing library use a Unix
2008 domain socket for their communication.
2010 * The <<lttng-ust-agents,user space tracing agents>>.
2012 Any instance of a user space tracing agent first registers to
2013 a session daemon. Then, the session daemon can send requests to
2014 this instance, such as:
2017 ** Get the list of loggers.
2018 ** Enable or disable a specific logger.
2021 The session daemon and the user space tracing agent use a TCP connection
2022 for their communication.
2024 * The <<lttng-modules,LTTng kernel tracer>>.
2025 * The <<lttng-consumerd,consumer daemon>>.
2027 The session daemon sends requests to the consumer daemon to instruct
2028 it where to send the trace data streams, amongst other information.
2030 * The <<lttng-relayd,relay daemon>>.
2032 The session daemon receives commands from the
2033 <<liblttng-ctl-lttng,tracing control library>>.
2035 The root session daemon loads the appropriate
2036 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2037 a <<lttng-consumerd,consumer daemon>> as soon as you create
2038 an <<event,event rule>>.
2040 The session daemon does not send and receive trace data: this is the
2041 role of the <<lttng-consumerd,consumer daemon>> and
2042 <<lttng-relayd,relay daemon>>. It does, however, generate the
2043 http://diamon.org/ctf/[CTF] metadata stream.
2045 Each Unix user can have its own session daemon instance. The
2046 tracing sessions managed by different session daemons are completely
2049 The root user's session daemon is the only one which is
2050 allowed to control the LTTng kernel tracer, and its spawned consumer
2051 daemon is the only one which is allowed to consume trace data from the
2052 LTTng kernel tracer. Note, however, that any Unix user which is a member
2053 of the <<tracing-group,tracing group>> is allowed
2054 to create <<channel,channels>> in the
2055 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2058 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2059 session daemon when using its `create` command if none is currently
2060 running. You can also start the session daemon manually.
2067 .The consumer daemon.
2068 image::plumbing-consumerd.png[]
2070 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
2071 ring buffers with user applications or with the LTTng kernel modules to
2072 collect trace data and send it to some location (on disk or to a
2073 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2074 is part of LTTng-tools.
2076 You do not start a consumer daemon manually: a consumer daemon is always
2077 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2078 <<event,event rule>>, that is, before you start tracing. When you kill
2079 its owner session daemon, the consumer daemon also exits because it is
2080 the session daemon's child process. Command-line options of
2081 man:lttng-sessiond(8) target the consumer daemon process.
2083 There are up to two running consumer daemons per Unix user, whereas only
2084 one session daemon can run per user. This is because each process can be
2085 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2086 and 64-bit processes, it is more efficient to have separate
2087 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2088 exception: it can have up to _three_ running consumer daemons: 32-bit
2089 and 64-bit instances for its user applications, and one more
2090 reserved for collecting kernel trace data.
2098 image::plumbing-relayd.png[]
2100 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2101 between remote session and consumer daemons, local trace files, and a
2102 remote live trace viewer. The relay daemon is part of LTTng-tools.
2104 The main purpose of the relay daemon is to implement a receiver of
2105 <<sending-trace-data-over-the-network,trace data over the network>>.
2106 This is useful when the target system does not have much file system
2107 space to record trace files locally.
2109 The relay daemon is also a server to which a
2110 <<lttng-live,live trace viewer>> can
2111 connect. The live trace viewer sends requests to the relay daemon to
2112 receive trace data as the target system emits events. The
2113 communication protocol is named _LTTng live_; it is used over TCP
2116 Note that you can start the relay daemon on the target system directly.
2117 This is the setup of choice when the use case is to view events as
2118 the target system emits them without the need of a remote system.
2122 == [[using-lttng]]Instrumentation
2124 There are many examples of tracing and monitoring in our everyday life:
2126 * You have access to real-time and historical weather reports and
2127 forecasts thanks to weather stations installed around the country.
2128 * You know your heart is safe thanks to an electrocardiogram.
2129 * You make sure not to drive your car too fast and to have enough fuel
2130 to reach your destination thanks to gauges visible on your dashboard.
2132 All the previous examples have something in common: they rely on
2133 **instruments**. Without the electrodes attached to the surface of your
2134 body's skin, cardiac monitoring is futile.
2136 LTTng, as a tracer, is no different from those real life examples. If
2137 you're about to trace a software system or, in other words, record its
2138 history of execution, you better have **instrumentation points** in the
2139 subject you're tracing, that is, the actual software.
2141 Various ways were developed to instrument a piece of software for LTTng
2142 tracing. The most straightforward one is to manually place
2143 instrumentation points, called _tracepoints_, in the software's source
2144 code. It is also possible to add instrumentation points dynamically in
2145 the Linux kernel <<domain,tracing domain>>.
2147 If you're only interested in tracing the Linux kernel, your
2148 instrumentation needs are probably already covered by LTTng's built-in
2149 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2150 user application which is already instrumented for LTTng tracing.
2151 In such cases, you can skip this whole section and read the topics of
2152 the <<controlling-tracing,Tracing control>> section.
2154 Many methods are available to instrument a piece of software for LTTng
2157 * <<c-application,User space instrumentation for C and $$C++$$
2159 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2160 * <<java-application,User space Java agent>>.
2161 * <<python-application,User space Python agent>>.
2162 * <<proc-lttng-logger-abi,LTTng logger>>.
2163 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2167 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2169 The procedure to instrument a C or $$C++$$ user application with
2170 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2172 . <<tracepoint-provider,Create the source files of a tracepoint provider
2174 . <<probing-the-application-source-code,Add tracepoints to
2175 the application's source code>>.
2176 . <<building-tracepoint-providers-and-user-application,Build and link
2177 a tracepoint provider package and the user application>>.
2179 If you need quick, man:printf(3)-like instrumentation, you can skip
2180 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2183 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2184 instrument a user application with `liblttng-ust`.
2187 [[tracepoint-provider]]
2188 ==== Create the source files of a tracepoint provider package
2190 A _tracepoint provider_ is a set of compiled functions which provide
2191 **tracepoints** to an application, the type of instrumentation point
2192 supported by LTTng-UST. Those functions can emit events with
2193 user-defined fields and serialize those events as event records to one
2194 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2195 macro, which you <<probing-the-application-source-code,insert in a user
2196 application's source code>>, calls those functions.
2198 A _tracepoint provider package_ is an object file (`.o`) or a shared
2199 library (`.so`) which contains one or more tracepoint providers.
2200 Its source files are:
2202 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2203 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2205 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2206 the LTTng user space tracer, at run time.
2209 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2210 image::ust-app.png[]
2212 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2213 skip creating and using a tracepoint provider and use
2214 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2218 ===== Create a tracepoint provider header file template
2220 A _tracepoint provider header file_ contains the tracepoint
2221 definitions of a tracepoint provider.
2223 To create a tracepoint provider header file:
2225 . Start from this template:
2229 .Tracepoint provider header file template (`.h` file extension).
2231 #undef TRACEPOINT_PROVIDER
2232 #define TRACEPOINT_PROVIDER provider_name
2234 #undef TRACEPOINT_INCLUDE
2235 #define TRACEPOINT_INCLUDE "./tp.h"
2237 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2240 #include <lttng/tracepoint.h>
2243 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2244 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2249 #include <lttng/tracepoint-event.h>
2255 * `provider_name` with the name of your tracepoint provider.
2256 * `"tp.h"` with the name of your tracepoint provider header file.
2258 . Below the `#include <lttng/tracepoint.h>` line, put your
2259 <<defining-tracepoints,tracepoint definitions>>.
2261 Your tracepoint provider name must be unique amongst all the possible
2262 tracepoint provider names used on the same target system. We
2263 suggest to include the name of your project or company in the name,
2264 for example, `org_lttng_my_project_tpp`.
2266 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2267 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2268 write are the <<defining-tracepoints,tracepoint definitions>>.
2271 [[defining-tracepoints]]
2272 ===== Create a tracepoint definition
2274 A _tracepoint definition_ defines, for a given tracepoint:
2276 * Its **input arguments**. They are the macro parameters that the
2277 `tracepoint()` macro accepts for this particular tracepoint
2278 in the user application's source code.
2279 * Its **output event fields**. They are the sources of event fields
2280 that form the payload of any event that the execution of the
2281 `tracepoint()` macro emits for this particular tracepoint.
2283 You can create a tracepoint definition by using the
2284 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2286 <<tpp-header,tracepoint provider header file template>>.
2288 The syntax of the `TRACEPOINT_EVENT()` macro is:
2291 .`TRACEPOINT_EVENT()` macro syntax.
2294 /* Tracepoint provider name */
2297 /* Tracepoint name */
2300 /* Input arguments */
2305 /* Output event fields */
2314 * `provider_name` with your tracepoint provider name.
2315 * `tracepoint_name` with your tracepoint name.
2316 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2317 * `fields` with the <<tpp-def-output-fields,output event field>>
2320 This tracepoint emits events named `provider_name:tracepoint_name`.
2323 .Event name's length limitation
2325 The concatenation of the tracepoint provider name and the
2326 tracepoint name must not exceed **254 characters**. If it does, the
2327 instrumented application compiles and runs, but LTTng throws multiple
2328 warnings and you could experience serious issues.
2331 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2334 .`TP_ARGS()` macro syntax.
2343 * `type` with the C type of the argument.
2344 * `arg_name` with the argument name.
2346 You can repeat `type` and `arg_name` up to 10 times to have
2347 more than one argument.
2349 .`TP_ARGS()` usage with three arguments.
2361 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2362 tracepoint definition with no input arguments.
2364 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2365 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2366 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2367 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2370 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2371 C expression that the tracer evalutes at the `tracepoint()` macro site
2372 in the application's source code. This expression provides a field's
2373 source of data. The argument expression can include input argument names
2374 listed in the `TP_ARGS()` macro.
2376 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2377 must be unique within a given tracepoint definition.
2379 Here's a complete tracepoint definition example:
2381 .Tracepoint definition.
2383 The following tracepoint definition defines a tracepoint which takes
2384 three input arguments and has four output event fields.
2388 #include "my-custom-structure.h"
2394 const struct my_custom_structure*, my_custom_structure,
2399 ctf_string(query_field, query)
2400 ctf_float(double, ratio_field, ratio)
2401 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2402 ctf_integer(int, send_size, my_custom_structure->send_size)
2407 You can refer to this tracepoint definition with the `tracepoint()`
2408 macro in your application's source code like this:
2412 tracepoint(my_provider, my_tracepoint,
2413 my_structure, some_ratio, the_query);
2417 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2418 if they satisfy an enabled <<event,event rule>>.
2421 [[using-tracepoint-classes]]
2422 ===== Use a tracepoint class
2424 A _tracepoint class_ is a class of tracepoints which share the same
2425 output event field definitions. A _tracepoint instance_ is one
2426 instance of such a defined tracepoint class, with its own tracepoint
2429 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2430 shorthand which defines both a tracepoint class and a tracepoint
2431 instance at the same time.
2433 When you build a tracepoint provider package, the C or $$C++$$ compiler
2434 creates one serialization function for each **tracepoint class**. A
2435 serialization function is responsible for serializing the event fields
2436 of a tracepoint to a sub-buffer when tracing.
2438 For various performance reasons, when your situation requires multiple
2439 tracepoint definitions with different names, but with the same event
2440 fields, we recommend that you manually create a tracepoint class
2441 and instantiate as many tracepoint instances as needed. One positive
2442 effect of such a design, amongst other advantages, is that all
2443 tracepoint instances of the same tracepoint class reuse the same
2444 serialization function, thus reducing
2445 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2447 .Use a tracepoint class and tracepoint instances.
2449 Consider the following three tracepoint definitions:
2461 ctf_integer(int, userid, userid)
2462 ctf_integer(size_t, len, len)
2474 ctf_integer(int, userid, userid)
2475 ctf_integer(size_t, len, len)
2487 ctf_integer(int, userid, userid)
2488 ctf_integer(size_t, len, len)
2493 In this case, we create three tracepoint classes, with one implicit
2494 tracepoint instance for each of them: `get_account`, `get_settings`, and
2495 `get_transaction`. However, they all share the same event field names
2496 and types. Hence three identical, yet independent serialization
2497 functions are created when you build the tracepoint provider package.
2499 A better design choice is to define a single tracepoint class and three
2500 tracepoint instances:
2504 /* The tracepoint class */
2505 TRACEPOINT_EVENT_CLASS(
2506 /* Tracepoint provider name */
2509 /* Tracepoint class name */
2512 /* Input arguments */
2518 /* Output event fields */
2520 ctf_integer(int, userid, userid)
2521 ctf_integer(size_t, len, len)
2525 /* The tracepoint instances */
2526 TRACEPOINT_EVENT_INSTANCE(
2527 /* Tracepoint provider name */
2530 /* Tracepoint class name */
2533 /* Tracepoint name */
2536 /* Input arguments */
2542 TRACEPOINT_EVENT_INSTANCE(
2551 TRACEPOINT_EVENT_INSTANCE(
2564 [[assigning-log-levels]]
2565 ===== Assign a log level to a tracepoint definition
2567 You can assign an optional _log level_ to a
2568 <<defining-tracepoints,tracepoint definition>>.
2570 Assigning different levels of severity to tracepoint definitions can
2571 be useful: when you <<enabling-disabling-events,create an event rule>>,
2572 you can target tracepoints having a log level as severe as a specific
2575 The concept of LTTng-UST log levels is similar to the levels found
2576 in typical logging frameworks:
2578 * In a logging framework, the log level is given by the function
2579 or method name you use at the log statement site: `debug()`,
2580 `info()`, `warn()`, `error()`, and so on.
2581 * In LTTng-UST, you statically assign the log level to a tracepoint
2582 definition; any `tracepoint()` macro invocation which refers to
2583 this definition has this log level.
2585 You can assign a log level to a tracepoint definition with the
2586 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2587 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2588 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2591 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2594 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2596 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2601 * `provider_name` with the tracepoint provider name.
2602 * `tracepoint_name` with the tracepoint name.
2603 * `log_level` with the log level to assign to the tracepoint
2604 definition named `tracepoint_name` in the `provider_name`
2605 tracepoint provider.
2607 See man:lttng-ust(3) for a list of available log level names.
2609 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2613 /* Tracepoint definition */
2622 ctf_integer(int, userid, userid)
2623 ctf_integer(size_t, len, len)
2627 /* Log level assignment */
2628 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2634 ===== Create a tracepoint provider package source file
2636 A _tracepoint provider package source file_ is a C source file which
2637 includes a <<tpp-header,tracepoint provider header file>> to expand its
2638 macros into event serialization and other functions.
2640 You can always use the following tracepoint provider package source
2644 .Tracepoint provider package source file template.
2646 #define TRACEPOINT_CREATE_PROBES
2651 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2652 header file>> name. You may also include more than one tracepoint
2653 provider header file here to create a tracepoint provider package
2654 holding more than one tracepoint providers.
2657 [[probing-the-application-source-code]]
2658 ==== Add tracepoints to an application's source code
2660 Once you <<tpp-header,create a tracepoint provider header file>>, you
2661 can use the `tracepoint()` macro in your application's
2662 source code to insert the tracepoints that this header
2663 <<defining-tracepoints,defines>>.
2665 The `tracepoint()` macro takes at least two parameters: the tracepoint
2666 provider name and the tracepoint name. The corresponding tracepoint
2667 definition defines the other parameters.
2669 .`tracepoint()` usage.
2671 The following <<defining-tracepoints,tracepoint definition>> defines a
2672 tracepoint which takes two input arguments and has two output event
2676 .Tracepoint provider header file.
2678 #include "my-custom-structure.h"
2685 const char*, cmd_name
2688 ctf_string(cmd_name, cmd_name)
2689 ctf_integer(int, number_of_args, argc)
2694 You can refer to this tracepoint definition with the `tracepoint()`
2695 macro in your application's source code like this:
2698 .Application's source file.
2702 int main(int argc, char* argv[])
2704 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2710 Note how the application's source code includes
2711 the tracepoint provider header file containing the tracepoint
2712 definitions to use, path:{tp.h}.
2715 .`tracepoint()` usage with a complex tracepoint definition.
2717 Consider this complex tracepoint definition, where multiple event
2718 fields refer to the same input arguments in their argument expression
2722 .Tracepoint provider header file.
2724 /* For `struct stat` */
2725 #include <sys/types.h>
2726 #include <sys/stat.h>
2738 ctf_integer(int, my_constant_field, 23 + 17)
2739 ctf_integer(int, my_int_arg_field, my_int_arg)
2740 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2741 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2742 my_str_arg[2] + my_str_arg[3])
2743 ctf_string(my_str_arg_field, my_str_arg)
2744 ctf_integer_hex(off_t, size_field, st->st_size)
2745 ctf_float(double, size_dbl_field, (double) st->st_size)
2746 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2747 size_t, strlen(my_str_arg) / 2)
2752 You can refer to this tracepoint definition with the `tracepoint()`
2753 macro in your application's source code like this:
2756 .Application's source file.
2758 #define TRACEPOINT_DEFINE
2765 stat("/etc/fstab", &s);
2766 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2772 If you look at the event record that LTTng writes when tracing this
2773 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2774 it should look like this:
2776 .Event record fields
2778 |Field's name |Field's value
2779 |`my_constant_field` |40
2780 |`my_int_arg_field` |23
2781 |`my_int_arg_field2` |529
2783 |`my_str_arg_field` |`Hello, World!`
2784 |`size_field` |0x12d
2785 |`size_dbl_field` |301.0
2786 |`half_my_str_arg_field` |`Hello,`
2790 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2791 compute--they use the call stack, for example. To avoid this
2792 computation when the tracepoint is disabled, you can use the
2793 `tracepoint_enabled()` and `do_tracepoint()` macros.
2795 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2799 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2801 tracepoint_enabled(provider_name, tracepoint_name)
2802 do_tracepoint(provider_name, tracepoint_name, ...)
2807 * `provider_name` with the tracepoint provider name.
2808 * `tracepoint_name` with the tracepoint name.
2810 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2811 `tracepoint_name` from the provider named `provider_name` is enabled
2814 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2815 if the tracepoint is enabled. Using `tracepoint()` with
2816 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2817 the `tracepoint_enabled()` check, thus a race condition is
2818 possible in this situation:
2821 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2823 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2824 stuff = prepare_stuff();
2827 tracepoint(my_provider, my_tracepoint, stuff);
2830 If the tracepoint is enabled after the condition, then `stuff` is not
2831 prepared: the emitted event will either contain wrong data, or the whole
2832 application could crash (segmentation fault, for example).
2834 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2835 `STAP_PROBEV()` call. If you need it, you must emit
2839 [[building-tracepoint-providers-and-user-application]]
2840 ==== Build and link a tracepoint provider package and an application
2842 Once you have one or more <<tpp-header,tracepoint provider header
2843 files>> and a <<tpp-source,tracepoint provider package source file>>,
2844 you can create the tracepoint provider package by compiling its source
2845 file. From here, multiple build and run scenarios are possible. The
2846 following table shows common application and library configurations
2847 along with the required command lines to achieve them.
2849 In the following diagrams, we use the following file names:
2852 Executable application.
2855 Application's object file.
2858 Tracepoint provider package object file.
2861 Tracepoint provider package archive file.
2864 Tracepoint provider package shared object file.
2867 User library object file.
2870 User library shared object file.
2872 We use the following symbols in the diagrams of table below:
2875 .Symbols used in the build scenario diagrams.
2876 image::ust-sit-symbols.png[]
2878 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2879 variable in the following instructions.
2881 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2882 .Common tracepoint provider package scenarios.
2884 |Scenario |Instructions
2887 The instrumented application is statically linked with
2888 the tracepoint provider package object.
2890 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2893 include::../common/ust-sit-step-tp-o.txt[]
2895 To build the instrumented application:
2897 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2902 #define TRACEPOINT_DEFINE
2906 . Compile the application source file:
2915 . Build the application:
2920 gcc -o app app.o tpp.o -llttng-ust -ldl
2924 To run the instrumented application:
2926 * Start the application:
2936 The instrumented application is statically linked with the
2937 tracepoint provider package archive file.
2939 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2942 To create the tracepoint provider package archive file:
2944 . Compile the <<tpp-source,tracepoint provider package source file>>:
2953 . Create the tracepoint provider package archive file:
2962 To build the instrumented application:
2964 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2969 #define TRACEPOINT_DEFINE
2973 . Compile the application source file:
2982 . Build the application:
2987 gcc -o app app.o tpp.a -llttng-ust -ldl
2991 To run the instrumented application:
2993 * Start the application:
3003 The instrumented application is linked with the tracepoint provider
3004 package shared object.
3006 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
3009 include::../common/ust-sit-step-tp-so.txt[]
3011 To build the instrumented application:
3013 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3018 #define TRACEPOINT_DEFINE
3022 . Compile the application source file:
3031 . Build the application:
3036 gcc -o app app.o -ldl -L. -ltpp
3040 To run the instrumented application:
3042 * Start the application:
3052 The tracepoint provider package shared object is preloaded before the
3053 instrumented application starts.
3055 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3058 include::../common/ust-sit-step-tp-so.txt[]
3060 To build the instrumented application:
3062 . In path:{app.c}, before including path:{tpp.h}, add the
3068 #define TRACEPOINT_DEFINE
3069 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3073 . Compile the application source file:
3082 . Build the application:
3087 gcc -o app app.o -ldl
3091 To run the instrumented application with tracing support:
3093 * Preload the tracepoint provider package shared object and
3094 start the application:
3099 LD_PRELOAD=./libtpp.so ./app
3103 To run the instrumented application without tracing support:
3105 * Start the application:
3115 The instrumented application dynamically loads the tracepoint provider
3116 package shared object.
3118 See the <<dlclose-warning,warning about `dlclose()`>>.
3120 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3123 include::../common/ust-sit-step-tp-so.txt[]
3125 To build the instrumented application:
3127 . In path:{app.c}, before including path:{tpp.h}, add the
3133 #define TRACEPOINT_DEFINE
3134 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3138 . Compile the application source file:
3147 . Build the application:
3152 gcc -o app app.o -ldl
3156 To run the instrumented application:
3158 * Start the application:
3168 The application is linked with the instrumented user library.
3170 The instrumented user library is statically linked with the tracepoint
3171 provider package object file.
3173 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3176 include::../common/ust-sit-step-tp-o-fpic.txt[]
3178 To build the instrumented user library:
3180 . In path:{emon.c}, before including path:{tpp.h}, add the
3186 #define TRACEPOINT_DEFINE
3190 . Compile the user library source file:
3195 gcc -I. -fpic -c emon.c
3199 . Build the user library shared object:
3204 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3208 To build the application:
3210 . Compile the application source file:
3219 . Build the application:
3224 gcc -o app app.o -L. -lemon
3228 To run the application:
3230 * Start the application:
3240 The application is linked with the instrumented user library.
3242 The instrumented user library is linked with the tracepoint provider
3243 package shared object.
3245 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3248 include::../common/ust-sit-step-tp-so.txt[]
3250 To build the instrumented user library:
3252 . In path:{emon.c}, before including path:{tpp.h}, add the
3258 #define TRACEPOINT_DEFINE
3262 . Compile the user library source file:
3267 gcc -I. -fpic -c emon.c
3271 . Build the user library shared object:
3276 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3280 To build the application:
3282 . Compile the application source file:
3291 . Build the application:
3296 gcc -o app app.o -L. -lemon
3300 To run the application:
3302 * Start the application:
3312 The tracepoint provider package shared object is preloaded before the
3315 The application is linked with the instrumented user library.
3317 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3320 include::../common/ust-sit-step-tp-so.txt[]
3322 To build the instrumented user library:
3324 . In path:{emon.c}, before including path:{tpp.h}, add the
3330 #define TRACEPOINT_DEFINE
3331 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3335 . Compile the user library source file:
3340 gcc -I. -fpic -c emon.c
3344 . Build the user library shared object:
3349 gcc -shared -o libemon.so emon.o -ldl
3353 To build the application:
3355 . Compile the application source file:
3364 . Build the application:
3369 gcc -o app app.o -L. -lemon
3373 To run the application with tracing support:
3375 * Preload the tracepoint provider package shared object and
3376 start the application:
3381 LD_PRELOAD=./libtpp.so ./app
3385 To run the application without tracing support:
3387 * Start the application:
3397 The application is linked with the instrumented user library.
3399 The instrumented user library dynamically loads the tracepoint provider
3400 package shared object.
3402 See the <<dlclose-warning,warning about `dlclose()`>>.
3404 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3407 include::../common/ust-sit-step-tp-so.txt[]
3409 To build the instrumented user library:
3411 . In path:{emon.c}, before including path:{tpp.h}, add the
3417 #define TRACEPOINT_DEFINE
3418 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3422 . Compile the user library source file:
3427 gcc -I. -fpic -c emon.c
3431 . Build the user library shared object:
3436 gcc -shared -o libemon.so emon.o -ldl
3440 To build the application:
3442 . Compile the application source file:
3451 . Build the application:
3456 gcc -o app app.o -L. -lemon
3460 To run the application:
3462 * Start the application:
3472 The application dynamically loads the instrumented user library.
3474 The instrumented user library is linked with the tracepoint provider
3475 package shared object.
3477 See the <<dlclose-warning,warning about `dlclose()`>>.
3479 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3482 include::../common/ust-sit-step-tp-so.txt[]
3484 To build the instrumented user library:
3486 . In path:{emon.c}, before including path:{tpp.h}, add the
3492 #define TRACEPOINT_DEFINE
3496 . Compile the user library source file:
3501 gcc -I. -fpic -c emon.c
3505 . Build the user library shared object:
3510 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3514 To build the application:
3516 . Compile the application source file:
3525 . Build the application:
3530 gcc -o app app.o -ldl -L. -lemon
3534 To run the application:
3536 * Start the application:
3546 The application dynamically loads the instrumented user library.
3548 The instrumented user library dynamically loads the tracepoint provider
3549 package shared object.
3551 See the <<dlclose-warning,warning about `dlclose()`>>.
3553 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3556 include::../common/ust-sit-step-tp-so.txt[]
3558 To build the instrumented user library:
3560 . In path:{emon.c}, before including path:{tpp.h}, add the
3566 #define TRACEPOINT_DEFINE
3567 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3571 . Compile the user library source file:
3576 gcc -I. -fpic -c emon.c
3580 . Build the user library shared object:
3585 gcc -shared -o libemon.so emon.o -ldl
3589 To build the application:
3591 . Compile the application source file:
3600 . Build the application:
3605 gcc -o app app.o -ldl -L. -lemon
3609 To run the application:
3611 * Start the application:
3621 The tracepoint provider package shared object is preloaded before the
3624 The application dynamically loads the instrumented user library.
3626 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3629 include::../common/ust-sit-step-tp-so.txt[]
3631 To build the instrumented user library:
3633 . In path:{emon.c}, before including path:{tpp.h}, add the
3639 #define TRACEPOINT_DEFINE
3640 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3644 . Compile the user library source file:
3649 gcc -I. -fpic -c emon.c
3653 . Build the user library shared object:
3658 gcc -shared -o libemon.so emon.o -ldl
3662 To build the application:
3664 . Compile the application source file:
3673 . Build the application:
3678 gcc -o app app.o -L. -lemon
3682 To run the application with tracing support:
3684 * Preload the tracepoint provider package shared object and
3685 start the application:
3690 LD_PRELOAD=./libtpp.so ./app
3694 To run the application without tracing support:
3696 * Start the application:
3706 The application is statically linked with the tracepoint provider
3707 package object file.
3709 The application is linked with the instrumented user library.
3711 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3714 include::../common/ust-sit-step-tp-o.txt[]
3716 To build the instrumented user library:
3718 . In path:{emon.c}, before including path:{tpp.h}, add the
3724 #define TRACEPOINT_DEFINE
3728 . Compile the user library source file:
3733 gcc -I. -fpic -c emon.c
3737 . Build the user library shared object:
3742 gcc -shared -o libemon.so emon.o
3746 To build the application:
3748 . Compile the application source file:
3757 . Build the application:
3762 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3766 To run the instrumented application:
3768 * Start the application:
3778 The application is statically linked with the tracepoint provider
3779 package object file.
3781 The application dynamically loads the instrumented user library.
3783 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3786 include::../common/ust-sit-step-tp-o.txt[]
3788 To build the application:
3790 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3795 #define TRACEPOINT_DEFINE
3799 . Compile the application source file:
3808 . Build the application:
3813 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3818 The `--export-dynamic` option passed to the linker is necessary for the
3819 dynamically loaded library to ``see'' the tracepoint symbols defined in
3822 To build the instrumented user library:
3824 . Compile the user library source file:
3829 gcc -I. -fpic -c emon.c
3833 . Build the user library shared object:
3838 gcc -shared -o libemon.so emon.o
3842 To run the application:
3844 * Start the application:
3856 .Do not use man:dlclose(3) on a tracepoint provider package
3858 Never use man:dlclose(3) on any shared object which:
3860 * Is linked with, statically or dynamically, a tracepoint provider
3862 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3863 package shared object.
3865 This is currently considered **unsafe** due to a lack of reference
3866 counting from LTTng-UST to the shared object.
3868 A known workaround (available since glibc 2.2) is to use the
3869 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3870 effect of not unloading the loaded shared object, even if man:dlclose(3)
3873 You can also preload the tracepoint provider package shared object with
3874 the env:LD_PRELOAD environment variable to overcome this limitation.
3878 [[using-lttng-ust-with-daemons]]
3879 ===== Use noch:{LTTng-UST} with daemons
3881 If your instrumented application calls man:fork(2), man:clone(2),
3882 or BSD's man:rfork(2), without a following man:exec(3)-family
3883 system call, you must preload the path:{liblttng-ust-fork.so} shared
3884 object when starting the application.
3888 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3891 If your tracepoint provider package is
3892 a shared library which you also preload, you must put both
3893 shared objects in env:LD_PRELOAD:
3897 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3901 [[lttng-ust-pkg-config]]
3902 ===== Use noch:{pkg-config}
3904 On some distributions, LTTng-UST ships with a
3905 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3906 metadata file. If this is your case, then you can use cmd:pkg-config to
3907 build an application on the command line:
3911 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3915 [[instrumenting-32-bit-app-on-64-bit-system]]
3916 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3918 In order to trace a 32-bit application running on a 64-bit system,
3919 LTTng must use a dedicated 32-bit
3920 <<lttng-consumerd,consumer daemon>>.
3922 The following steps show how to build and install a 32-bit consumer
3923 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3924 build and install the 32-bit LTTng-UST libraries, and how to build and
3925 link an instrumented 32-bit application in that context.
3927 To build a 32-bit instrumented application for a 64-bit target system,
3928 assuming you have a fresh target system with no installed Userspace RCU
3931 . Download, build, and install a 32-bit version of Userspace RCU:
3937 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3938 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3939 cd userspace-rcu-0.9.* &&
3940 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3942 sudo make install &&
3947 . Using your distribution's package manager, or from source, install
3948 the following 32-bit versions of the following dependencies of
3949 LTTng-tools and LTTng-UST:
3952 * https://sourceforge.net/projects/libuuid/[libuuid]
3953 * http://directory.fsf.org/wiki/Popt[popt]
3954 * http://www.xmlsoft.org/[libxml2]
3957 . Download, build, and install a 32-bit version of the latest
3958 LTTng-UST{nbsp}{revision}:
3964 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
3965 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
3966 cd lttng-ust-2.8.* &&
3967 ./configure --libdir=/usr/local/lib32 \
3968 CFLAGS=-m32 CXXFLAGS=-m32 \
3969 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3971 sudo make install &&
3978 Depending on your distribution,
3979 32-bit libraries could be installed at a different location than
3980 `/usr/lib32`. For example, Debian is known to install
3981 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3983 In this case, make sure to set `LDFLAGS` to all the
3984 relevant 32-bit library paths, for example:
3988 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3992 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3993 the 32-bit consumer daemon:
3999 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4000 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4001 cd lttng-tools-2.8.* &&
4002 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
4003 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4005 cd src/bin/lttng-consumerd &&
4006 sudo make install &&
4011 . From your distribution or from source,
4012 <<installing-lttng,install>> the 64-bit versions of
4013 LTTng-UST and Userspace RCU.
4014 . Download, build, and install the 64-bit version of the
4015 latest LTTng-tools{nbsp}{revision}:
4021 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4022 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4023 cd lttng-tools-2.8.* &&
4024 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
4025 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
4027 sudo make install &&
4032 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
4033 when linking your 32-bit application:
4036 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4037 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4040 For example, let's rebuild the quick start example in
4041 <<tracing-your-own-user-application,Trace a user application>> as an
4042 instrumented 32-bit application:
4047 gcc -m32 -c -I. hello-tp.c
4049 gcc -m32 -o hello hello.o hello-tp.o \
4050 -L/usr/lib32 -L/usr/local/lib32 \
4051 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4056 No special action is required to execute the 32-bit application and
4057 to trace it: use the command-line man:lttng(1) tool as usual.
4064 man:tracef(3) is a small LTTng-UST API designed for quick,
4065 man:printf(3)-like instrumentation without the burden of
4066 <<tracepoint-provider,creating>> and
4067 <<building-tracepoint-providers-and-user-application,building>>
4068 a tracepoint provider package.
4070 To use `tracef()` in your application:
4072 . In the C or C++ source files where you need to use `tracef()`,
4073 include `<lttng/tracef.h>`:
4078 #include <lttng/tracef.h>
4082 . In the application's source code, use `tracef()` like you would use
4090 tracef("my message: %d (%s)", my_integer, my_string);
4096 . Link your application with `liblttng-ust`:
4101 gcc -o app app.c -llttng-ust
4105 To trace the events that `tracef()` calls emit:
4107 * <<enabling-disabling-events,Create an event rule>> which matches the
4108 `lttng_ust_tracef:*` event name:
4113 lttng enable-event --userspace 'lttng_ust_tracef:*'
4118 .Limitations of `tracef()`
4120 The `tracef()` utility function was developed to make user space tracing
4121 super simple, albeit with notable disadvantages compared to
4122 <<defining-tracepoints,user-defined tracepoints>>:
4124 * All the emitted events have the same tracepoint provider and
4125 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4126 * There is no static type checking.
4127 * The only event record field you actually get, named `msg`, is a string
4128 potentially containing the values you passed to `tracef()`
4129 using your own format string. This also means that you cannot filter
4130 events with a custom expression at run time because there are no
4132 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4133 function behind the scenes to format the strings at run time, its
4134 expected performance is lower than with user-defined tracepoints,
4135 which do not require a conversion to a string.
4137 Taking this into consideration, `tracef()` is useful for some quick
4138 prototyping and debugging, but you should not consider it for any
4139 permanent and serious applicative instrumentation.
4145 ==== Use `tracelog()`
4147 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4148 the difference that it accepts an additional log level parameter.
4150 The goal of `tracelog()` is to ease the migration from logging to
4153 To use `tracelog()` in your application:
4155 . In the C or C++ source files where you need to use `tracelog()`,
4156 include `<lttng/tracelog.h>`:
4161 #include <lttng/tracelog.h>
4165 . In the application's source code, use `tracelog()` like you would use
4166 man:printf(3), except for the first parameter which is the log
4174 tracelog(TRACE_WARNING, "my message: %d (%s)",
4175 my_integer, my_string);
4181 See man:lttng-ust(3) for a list of available log level names.
4183 . Link your application with `liblttng-ust`:
4188 gcc -o app app.c -llttng-ust
4192 To trace the events that `tracelog()` calls emit with a log level
4193 _as severe as_ a specific log level:
4195 * <<enabling-disabling-events,Create an event rule>> which matches the
4196 `lttng_ust_tracelog:*` event name and a minimum level
4202 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4203 --loglevel=TRACE_WARNING
4207 To trace the events that `tracelog()` calls emit with a
4208 _specific log level_:
4210 * Create an event rule which matches the `lttng_ust_tracelog:*`
4211 event name and a specific log level:
4216 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4217 --loglevel-only=TRACE_INFO
4222 [[prebuilt-ust-helpers]]
4223 === Prebuilt user space tracing helpers
4225 The LTTng-UST package provides a few helpers in the form or preloadable
4226 shared objects which automatically instrument system functions and
4229 The helper shared objects are normally found in dir:{/usr/lib}. If you
4230 built LTTng-UST <<building-from-source,from source>>, they are probably
4231 located in dir:{/usr/local/lib}.
4233 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4236 path:{liblttng-ust-libc-wrapper.so}::
4237 path:{liblttng-ust-pthread-wrapper.so}::
4238 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4239 memory and POSIX threads function tracing>>.
4241 path:{liblttng-ust-cyg-profile.so}::
4242 path:{liblttng-ust-cyg-profile-fast.so}::
4243 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4245 path:{liblttng-ust-dl.so}::
4246 <<liblttng-ust-dl,Dynamic linker tracing>>.
4248 To use a user space tracing helper with any user application:
4250 * Preload the helper shared object when you start the application:
4255 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4259 You can preload more than one helper:
4264 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4270 [[liblttng-ust-libc-pthread-wrapper]]
4271 ==== Instrument C standard library memory and POSIX threads functions
4273 The path:{liblttng-ust-libc-wrapper.so} and
4274 path:{liblttng-ust-pthread-wrapper.so} helpers
4275 add instrumentation to some C standard library and POSIX
4279 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4281 |TP provider name |TP name |Instrumented function
4283 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4284 |`calloc` |man:calloc(3)
4285 |`realloc` |man:realloc(3)
4286 |`free` |man:free(3)
4287 |`memalign` |man:memalign(3)
4288 |`posix_memalign` |man:posix_memalign(3)
4292 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4294 |TP provider name |TP name |Instrumented function
4296 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4297 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4298 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4299 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4302 When you preload the shared object, it replaces the functions listed
4303 in the previous tables by wrappers which contain tracepoints and call
4304 the replaced functions.
4307 [[liblttng-ust-cyg-profile]]
4308 ==== Instrument function entry and exit
4310 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4311 to the entry and exit points of functions.
4313 man:gcc(1) and man:clang(1) have an option named
4314 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4315 which generates instrumentation calls for entry and exit to functions.
4316 The LTTng-UST function tracing helpers,
4317 path:{liblttng-ust-cyg-profile.so} and
4318 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4319 to add tracepoints to the two generated functions (which contain
4320 `cyg_profile` in their names, hence the helper's name).
4322 To use the LTTng-UST function tracing helper, the source files to
4323 instrument must be built using the `-finstrument-functions` compiler
4326 There are two versions of the LTTng-UST function tracing helper:
4328 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4329 that you should only use when it can be _guaranteed_ that the
4330 complete event stream is recorded without any lost event record.
4331 Any kind of duplicate information is left out.
4333 Assuming no event record is lost, having only the function addresses on
4334 entry is enough to create a call graph, since an event record always
4335 contains the ID of the CPU that generated it.
4337 You can use a tool like man:addr2line(1) to convert function addresses
4338 back to source file names and line numbers.
4340 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4341 which also works in use cases where event records might get discarded or
4342 not recorded from application startup.
4343 In these cases, the trace analyzer needs more information to be
4344 able to reconstruct the program flow.
4346 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4347 points of this helper.
4349 All the tracepoints that this helper provides have the
4350 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4352 TIP: It's sometimes a good idea to limit the number of source files that
4353 you compile with the `-finstrument-functions` option to prevent LTTng
4354 from writing an excessive amount of trace data at run time. When using
4355 man:gcc(1), you can use the
4356 `-finstrument-functions-exclude-function-list` option to avoid
4357 instrument entries and exits of specific function names.
4362 ==== Instrument the dynamic linker
4364 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4365 man:dlopen(3) and man:dlclose(3) function calls.
4367 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4372 [[java-application]]
4373 === User space Java agent
4375 You can instrument any Java application which uses one of the following
4378 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4379 (JUL) core logging facilities.
4380 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4381 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4384 .LTTng-UST Java agent imported by a Java application.
4385 image::java-app.png[]
4387 Note that the methods described below are new in LTTng{nbsp}{revision}.
4388 Previous LTTng versions use another technique.
4390 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4391 and https://ci.lttng.org/[continuous integration], thus this version is
4392 directly supported. However, the LTTng-UST Java agent is also tested
4393 with OpenJDK{nbsp}7.
4398 ==== Use the LTTng-UST Java agent for `java.util.logging`
4400 To use the LTTng-UST Java agent in a Java application which uses
4401 `java.util.logging` (JUL):
4403 . In the Java application's source code, import the LTTng-UST
4404 log handler package for `java.util.logging`:
4409 import org.lttng.ust.agent.jul.LttngLogHandler;
4413 . Create an LTTng-UST JUL log handler:
4418 Handler lttngUstLogHandler = new LttngLogHandler();
4422 . Add this handler to the JUL loggers which should emit LTTng events:
4427 Logger myLogger = Logger.getLogger("some-logger");
4429 myLogger.addHandler(lttngUstLogHandler);
4433 . Use `java.util.logging` log statements and configuration as usual.
4434 The loggers with an attached LTTng-UST log handler can emit
4437 . Before exiting the application, remove the LTTng-UST log handler from
4438 the loggers attached to it and call its `close()` method:
4443 myLogger.removeHandler(lttngUstLogHandler);
4444 lttngUstLogHandler.close();
4448 This is not strictly necessary, but it is recommended for a clean
4449 disposal of the handler's resources.
4451 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4452 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4454 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4455 path] when you build the Java application.
4457 The JAR files are typically located in dir:{/usr/share/java}.
4459 IMPORTANT: The LTTng-UST Java agent must be
4460 <<installing-lttng,installed>> for the logging framework your
4463 .Use the LTTng-UST Java agent for `java.util.logging`.
4468 import java.io.IOException;
4469 import java.util.logging.Handler;
4470 import java.util.logging.Logger;
4471 import org.lttng.ust.agent.jul.LttngLogHandler;
4475 private static final int answer = 42;
4477 public static void main(String[] argv) throws Exception
4480 Logger logger = Logger.getLogger("jello");
4482 // Create an LTTng-UST log handler
4483 Handler lttngUstLogHandler = new LttngLogHandler();
4485 // Add the LTTng-UST log handler to our logger
4486 logger.addHandler(lttngUstLogHandler);
4489 logger.info("some info");
4490 logger.warning("some warning");
4492 logger.finer("finer information; the answer is " + answer);
4494 logger.severe("error!");
4496 // Not mandatory, but cleaner
4497 logger.removeHandler(lttngUstLogHandler);
4498 lttngUstLogHandler.close();
4507 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4510 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4511 <<enabling-disabling-events,create an event rule>> matching the
4512 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4517 lttng enable-event --jul jello
4521 Run the compiled class:
4525 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4528 <<basic-tracing-session-control,Stop tracing>> and inspect the
4538 You can use the opt:lttng-enable-event(1):--loglevel or
4539 opt:lttng-enable-event(1):--loglevel-only option of the
4540 man:lttng-enable-event(1) command to target a range of JUL log levels
4541 or a specific JUL log level.
4546 ==== Use the LTTng-UST Java agent for Apache log4j
4548 To use the LTTng-UST Java agent in a Java application which uses
4551 . In the Java application's source code, import the LTTng-UST
4552 log appender package for Apache log4j:
4557 import org.lttng.ust.agent.log4j.LttngLogAppender;
4561 . Create an LTTng-UST log4j log appender:
4566 Appender lttngUstLogAppender = new LttngLogAppender();
4570 . Add this appender to the log4j loggers which should emit LTTng events:
4575 Logger myLogger = Logger.getLogger("some-logger");
4577 myLogger.addAppender(lttngUstLogAppender);
4581 . Use Apache log4j log statements and configuration as usual. The
4582 loggers with an attached LTTng-UST log appender can emit LTTng events.
4584 . Before exiting the application, remove the LTTng-UST log appender from
4585 the loggers attached to it and call its `close()` method:
4590 myLogger.removeAppender(lttngUstLogAppender);
4591 lttngUstLogAppender.close();
4595 This is not strictly necessary, but it is recommended for a clean
4596 disposal of the appender's resources.
4598 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4599 files, path:{lttng-ust-agent-common.jar} and
4600 path:{lttng-ust-agent-log4j.jar}, in the
4601 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4602 path] when you build the Java application.
4604 The JAR files are typically located in dir:{/usr/share/java}.
4606 IMPORTANT: The LTTng-UST Java agent must be
4607 <<installing-lttng,installed>> for the logging framework your
4610 .Use the LTTng-UST Java agent for Apache log4j.
4615 import org.apache.log4j.Appender;
4616 import org.apache.log4j.Logger;
4617 import org.lttng.ust.agent.log4j.LttngLogAppender;
4621 private static final int answer = 42;
4623 public static void main(String[] argv) throws Exception
4626 Logger logger = Logger.getLogger("jello");
4628 // Create an LTTng-UST log appender
4629 Appender lttngUstLogAppender = new LttngLogAppender();
4631 // Add the LTTng-UST log appender to our logger
4632 logger.addAppender(lttngUstLogAppender);
4635 logger.info("some info");
4636 logger.warn("some warning");
4638 logger.debug("debug information; the answer is " + answer);
4640 logger.fatal("error!");
4642 // Not mandatory, but cleaner
4643 logger.removeAppender(lttngUstLogAppender);
4644 lttngUstLogAppender.close();
4650 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4655 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4658 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4659 <<enabling-disabling-events,create an event rule>> matching the
4660 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4665 lttng enable-event --log4j jello
4669 Run the compiled class:
4673 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4676 <<basic-tracing-session-control,Stop tracing>> and inspect the
4686 You can use the opt:lttng-enable-event(1):--loglevel or
4687 opt:lttng-enable-event(1):--loglevel-only option of the
4688 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4689 or a specific log4j log level.
4693 [[java-application-context]]
4694 ==== Provide application-specific context fields in a Java application
4696 A Java application-specific context field is a piece of state provided
4697 by the application which <<adding-context,you can add>>, using the
4698 man:lttng-add-context(1) command, to each <<event,event record>>
4699 produced by the log statements of this application.
4701 For example, a given object might have a current request ID variable.
4702 You can create a context information retriever for this object and
4703 assign a name to this current request ID. You can then, using the
4704 man:lttng-add-context(1) command, add this context field by name to
4705 the JUL or log4j <<channel,channel>>.
4707 To provide application-specific context fields in a Java application:
4709 . In the Java application's source code, import the LTTng-UST
4710 Java agent context classes and interfaces:
4715 import org.lttng.ust.agent.context.ContextInfoManager;
4716 import org.lttng.ust.agent.context.IContextInfoRetriever;
4720 . Create a context information retriever class, that is, a class which
4721 implements the `IContextInfoRetriever` interface:
4726 class MyContextInfoRetriever implements IContextInfoRetriever
4729 public Object retrieveContextInfo(String key)
4731 if (key.equals("intCtx")) {
4733 } else if (key.equals("strContext")) {
4734 return "context value!";
4743 This `retrieveContextInfo()` method is the only member of the
4744 `IContextInfoRetriever` interface. Its role is to return the current
4745 value of a state by name to create a context field. The names of the
4746 context fields and which state variables they return depends on your
4749 All primitive types and objects are supported as context fields.
4750 When `retrieveContextInfo()` returns an object, the context field
4751 serializer calls its `toString()` method to add a string field to
4752 event records. The method can also return `null`, which means that
4753 no context field is available for the required name.
4755 . Register an instance of your context information retriever class to
4756 the context information manager singleton:
4761 IContextInfoRetriever cir = new MyContextInfoRetriever();
4762 ContextInfoManager cim = ContextInfoManager.getInstance();
4763 cim.registerContextInfoRetriever("retrieverName", cir);
4767 . Before exiting the application, remove your context information
4768 retriever from the context information manager singleton:
4773 ContextInfoManager cim = ContextInfoManager.getInstance();
4774 cim.unregisterContextInfoRetriever("retrieverName");
4778 This is not strictly necessary, but it is recommended for a clean
4779 disposal of some manager's resources.
4781 . Build your Java application with LTTng-UST Java agent support as
4782 usual, following the procedure for either the <<jul,JUL>> or
4783 <<log4j,Apache log4j>> framework.
4786 .Provide application-specific context fields in a Java application.
4791 import java.util.logging.Handler;
4792 import java.util.logging.Logger;
4793 import org.lttng.ust.agent.jul.LttngLogHandler;
4794 import org.lttng.ust.agent.context.ContextInfoManager;
4795 import org.lttng.ust.agent.context.IContextInfoRetriever;
4799 // Our context information retriever class
4800 private static class MyContextInfoRetriever
4801 implements IContextInfoRetriever
4804 public Object retrieveContextInfo(String key) {
4805 if (key.equals("intCtx")) {
4807 } else if (key.equals("strContext")) {
4808 return "context value!";
4815 private static final int answer = 42;
4817 public static void main(String args[]) throws Exception
4819 // Get the context information manager instance
4820 ContextInfoManager cim = ContextInfoManager.getInstance();
4822 // Create and register our context information retriever
4823 IContextInfoRetriever cir = new MyContextInfoRetriever();
4824 cim.registerContextInfoRetriever("myRetriever", cir);
4827 Logger logger = Logger.getLogger("jello");
4829 // Create an LTTng-UST log handler
4830 Handler lttngUstLogHandler = new LttngLogHandler();
4832 // Add the LTTng-UST log handler to our logger
4833 logger.addHandler(lttngUstLogHandler);
4836 logger.info("some info");
4837 logger.warning("some warning");
4839 logger.finer("finer information; the answer is " + answer);
4841 logger.severe("error!");
4843 // Not mandatory, but cleaner
4844 logger.removeHandler(lttngUstLogHandler);
4845 lttngUstLogHandler.close();
4846 cim.unregisterContextInfoRetriever("myRetriever");
4855 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4858 <<creating-destroying-tracing-sessions,Create a tracing session>>
4859 and <<enabling-disabling-events,create an event rule>> matching the
4865 lttng enable-event --jul jello
4868 <<adding-context,Add the application-specific context fields>> to the
4873 lttng add-context --jul --type='$app.myRetriever:intCtx'
4874 lttng add-context --jul --type='$app.myRetriever:strContext'
4877 <<basic-tracing-session-control,Start tracing>>:
4884 Run the compiled class:
4888 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4891 <<basic-tracing-session-control,Stop tracing>> and inspect the
4903 [[python-application]]
4904 === User space Python agent
4906 You can instrument a Python 2 or Python 3 application which uses the
4907 standard https://docs.python.org/3/library/logging.html[`logging`]
4910 Each log statement emits an LTTng event once the
4911 application module imports the
4912 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4915 .A Python application importing the LTTng-UST Python agent.
4916 image::python-app.png[]
4918 To use the LTTng-UST Python agent:
4920 . In the Python application's source code, import the LTTng-UST Python
4930 The LTTng-UST Python agent automatically adds its logging handler to the
4931 root logger at import time.
4933 Any log statement that the application executes before this import does
4934 not emit an LTTng event.
4936 IMPORTANT: The LTTng-UST Python agent must be
4937 <<installing-lttng,installed>>.
4939 . Use log statements and logging configuration as usual.
4940 Since the LTTng-UST Python agent adds a handler to the _root_
4941 logger, you can trace any log statement from any logger.
4943 .Use the LTTng-UST Python agent.
4954 logging.basicConfig()
4955 logger = logging.getLogger('my-logger')
4958 logger.debug('debug message')
4959 logger.info('info message')
4960 logger.warn('warn message')
4961 logger.error('error message')
4962 logger.critical('critical message')
4966 if __name__ == '__main__':
4970 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4971 logging handler which prints to the standard error stream, is not
4972 strictly required for LTTng-UST tracing to work, but in versions of
4973 Python preceding 3.2, you could see a warning message which indicates
4974 that no handler exists for the logger `my-logger`.
4976 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4977 <<enabling-disabling-events,create an event rule>> matching the
4978 `my-logger` Python logger, and <<basic-tracing-session-control,start
4984 lttng enable-event --python my-logger
4988 Run the Python script:
4995 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5005 You can use the opt:lttng-enable-event(1):--loglevel or
5006 opt:lttng-enable-event(1):--loglevel-only option of the
5007 man:lttng-enable-event(1) command to target a range of Python log levels
5008 or a specific Python log level.
5010 When an application imports the LTTng-UST Python agent, the agent tries
5011 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5012 <<start-sessiond,start the session daemon>> _before_ you run the Python
5013 application. If a session daemon is found, the agent tries to register
5014 to it during 5{nbsp}seconds, after which the application continues
5015 without LTTng tracing support. You can override this timeout value with
5016 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5019 If the session daemon stops while a Python application with an imported
5020 LTTng-UST Python agent runs, the agent retries to connect and to
5021 register to a session daemon every 3{nbsp}seconds. You can override this
5022 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5027 [[proc-lttng-logger-abi]]
5030 The `lttng-tracer` Linux kernel module, part of
5031 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5032 path:{/proc/lttng-logger} when it's loaded. Any application can write
5033 text data to this file to emit an LTTng event.
5036 .An application writes to the LTTng logger file to emit an LTTng event.
5037 image::lttng-logger.png[]
5039 The LTTng logger is the quickest method--not the most efficient,
5040 however--to add instrumentation to an application. It is designed
5041 mostly to instrument shell scripts:
5045 echo "Some message, some $variable" > /proc/lttng-logger
5048 Any event that the LTTng logger emits is named `lttng_logger` and
5049 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5050 other instrumentation points in the kernel tracing domain, **any Unix
5051 user** can <<enabling-disabling-events,create an event rule>> which
5052 matches its event name, not only the root user or users in the tracing
5055 To use the LTTng logger:
5057 * From any application, write text data to the path:{/proc/lttng-logger}
5060 The `msg` field of `lttng_logger` event records contains the
5063 NOTE: The maximum message length of an LTTng logger event is
5064 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5065 than one event to contain the remaining data.
5067 You should not use the LTTng logger to trace a user application which
5068 can be instrumented in a more efficient way, namely:
5070 * <<c-application,C and $$C++$$ applications>>.
5071 * <<java-application,Java applications>>.
5072 * <<python-application,Python applications>>.
5074 .Use the LTTng logger.
5079 echo 'Hello, World!' > /proc/lttng-logger
5081 df --human-readable --print-type / > /proc/lttng-logger
5084 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5085 <<enabling-disabling-events,create an event rule>> matching the
5086 `lttng_logger` Linux kernel tracepoint, and
5087 <<basic-tracing-session-control,start tracing>>:
5092 lttng enable-event --kernel lttng_logger
5096 Run the Bash script:
5103 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5114 [[instrumenting-linux-kernel]]
5115 === LTTng kernel tracepoints
5117 NOTE: This section shows how to _add_ instrumentation points to the
5118 Linux kernel. The kernel's subsystems are already thoroughly
5119 instrumented at strategic places for LTTng when you
5120 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5124 There are two methods to instrument the Linux kernel:
5126 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5127 tracepoint which uses the `TRACE_EVENT()` API.
5129 Choose this if you want to instrumentation a Linux kernel tree with an
5130 instrumentation point compatible with ftrace, perf, and SystemTap.
5132 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5133 instrument an out-of-tree kernel module.
5135 Choose this if you don't need ftrace, perf, or SystemTap support.
5139 [[linux-add-lttng-layer]]
5140 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5142 This section shows how to add an LTTng layer to existing ftrace
5143 instrumentation using the `TRACE_EVENT()` API.
5145 This section does not document the `TRACE_EVENT()` macro. You can
5146 read the following articles to learn more about this API:
5148 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5149 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5150 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5152 The following procedure assumes that your ftrace tracepoints are
5153 correctly defined in their own header and that they are created in
5154 one source file using the `CREATE_TRACE_POINTS` definition.
5156 To add an LTTng layer over an existing ftrace tracepoint:
5158 . Make sure the following kernel configuration options are
5164 * `CONFIG_HIGH_RES_TIMERS`
5165 * `CONFIG_TRACEPOINTS`
5168 . Build the Linux source tree with your custom ftrace tracepoints.
5169 . Boot the resulting Linux image on your target system.
5171 Confirm that the tracepoints exist by looking for their names in the
5172 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5173 is your subsystem's name.
5175 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5181 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
5182 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
5183 cd lttng-modules-2.8.*
5187 . In dir:{instrumentation/events/lttng-module}, relative to the root
5188 of the LTTng-modules source tree, create a header file named
5189 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5190 LTTng-modules tracepoint definitions using the LTTng-modules
5193 Start with this template:
5197 .path:{instrumentation/events/lttng-module/my_subsys.h}
5200 #define TRACE_SYSTEM my_subsys
5202 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5203 #define _LTTNG_MY_SUBSYS_H
5205 #include "../../../probes/lttng-tracepoint-event.h"
5206 #include <linux/tracepoint.h>
5208 LTTNG_TRACEPOINT_EVENT(
5210 * Format is identical to TRACE_EVENT()'s version for the three
5211 * following macro parameters:
5214 TP_PROTO(int my_int, const char *my_string),
5215 TP_ARGS(my_int, my_string),
5217 /* LTTng-modules specific macros */
5219 ctf_integer(int, my_int_field, my_int)
5220 ctf_string(my_bar_field, my_bar)
5224 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5226 #include "../../../probes/define_trace.h"
5230 The entries in the `TP_FIELDS()` section are the list of fields for the
5231 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5232 ftrace's `TRACE_EVENT()` macro.
5234 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5235 complete description of the available `ctf_*()` macros.
5237 . Create the LTTng-modules probe's kernel module C source file,
5238 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5243 .path:{probes/lttng-probe-my-subsys.c}
5245 #include <linux/module.h>
5246 #include "../lttng-tracer.h"
5249 * Build-time verification of mismatch between mainline
5250 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5251 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5253 #include <trace/events/my_subsys.h>
5255 /* Create LTTng tracepoint probes */
5256 #define LTTNG_PACKAGE_BUILD
5257 #define CREATE_TRACE_POINTS
5258 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5260 #include "../instrumentation/events/lttng-module/my_subsys.h"
5262 MODULE_LICENSE("GPL and additional rights");
5263 MODULE_AUTHOR("Your name <your-email>");
5264 MODULE_DESCRIPTION("LTTng my_subsys probes");
5265 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5266 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5267 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5268 LTTNG_MODULES_EXTRAVERSION);
5272 . Edit path:{probes/Makefile} and add your new kernel module object
5273 next to the existing ones:
5277 .path:{probes/Makefile}
5281 obj-m += lttng-probe-module.o
5282 obj-m += lttng-probe-power.o
5284 obj-m += lttng-probe-my-subsys.o
5290 . Build and install the LTTng kernel modules:
5295 make KERNELDIR=/path/to/linux
5296 sudo make modules_install
5300 Replace `/path/to/linux` with the path to the Linux source tree where
5301 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5303 Note that you can also use the
5304 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5305 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5306 C code that need to be executed before the event fields are recorded.
5308 The best way to learn how to use the previous LTTng-modules macros is to
5309 inspect the existing LTTng-modules tracepoint definitions in the
5310 dir:{instrumentation/events/lttng-module} header files. Compare them
5311 with the Linux kernel mainline versions in the
5312 dir:{include/trace/events} directory of the Linux source tree.
5316 [[lttng-tracepoint-event-code]]
5317 ===== Use custom C code to access the data for tracepoint fields
5319 Although we recommended to always use the
5320 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5321 the arguments and fields of an LTTng-modules tracepoint when possible,
5322 sometimes you need a more complex process to access the data that the
5323 tracer records as event record fields. In other words, you need local
5324 variables and multiple C{nbsp}statements instead of simple
5325 argument-based expressions that you pass to the
5326 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5328 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5329 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5330 a block of C{nbsp}code to be executed before LTTng records the fields.
5331 The structure of this macro is:
5334 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5336 LTTNG_TRACEPOINT_EVENT_CODE(
5338 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5339 * version for the following three macro parameters:
5342 TP_PROTO(int my_int, const char *my_string),
5343 TP_ARGS(my_int, my_string),
5345 /* Declarations of custom local variables */
5348 unsigned long b = 0;
5349 const char *name = "(undefined)";
5350 struct my_struct *my_struct;
5354 * Custom code which uses both tracepoint arguments
5355 * (in TP_ARGS()) and local variables (in TP_locvar()).
5357 * Local variables are actually members of a structure pointed
5358 * to by the special variable tp_locvar.
5362 tp_locvar->a = my_int + 17;
5363 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5364 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5365 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5366 put_my_struct(tp_locvar->my_struct);
5375 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5376 * version for this, except that tp_locvar members can be
5377 * used in the argument expression parameters of
5378 * the ctf_*() macros.
5381 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5382 ctf_integer(int, my_struct_a, tp_locvar->a)
5383 ctf_string(my_string_field, my_string)
5384 ctf_string(my_struct_name, tp_locvar->name)
5389 IMPORTANT: The C code defined in `TP_code()` must not have any side
5390 effects when executed. In particular, the code must not allocate
5391 memory or get resources without deallocating this memory or putting
5392 those resources afterwards.
5395 [[instrumenting-linux-kernel-tracing]]
5396 ==== Load and unload a custom probe kernel module
5398 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5399 kernel module>> in the kernel before it can emit LTTng events.
5401 To load the default probe kernel modules and a custom probe kernel
5404 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5405 probe modules to load when starting a root <<lttng-sessiond,session
5409 .Load the `my_subsys`, `usb`, and the default probe modules.
5413 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5418 You only need to pass the subsystem name, not the whole kernel module
5421 To load _only_ a given custom probe kernel module:
5423 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5424 modules to load when starting a root session daemon:
5427 .Load only the `my_subsys` and `usb` probe modules.
5431 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5436 To confirm that a probe module is loaded:
5443 lsmod | grep lttng_probe_usb
5447 To unload the loaded probe modules:
5449 * Kill the session daemon with `SIGTERM`:
5454 sudo pkill lttng-sessiond
5458 You can also use man:modprobe(8)'s `--remove` option if the session
5459 daemon terminates abnormally.
5462 [[controlling-tracing]]
5465 Once an application or a Linux kernel is
5466 <<instrumenting,instrumented>> for LTTng tracing,
5469 This section is divided in topics on how to use the various
5470 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5471 command-line tool>>, to _control_ the LTTng daemons and tracers.
5473 NOTE: In the following subsections, we refer to an man:lttng(1) command
5474 using its man page name. For example, instead of _Run the `create`
5475 command to..._, we use _Run the man:lttng-create(1) command to..._.
5479 === Start a session daemon
5481 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5482 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5485 You will see the following error when you run a command while no session
5489 Error: No session daemon is available
5492 The only command that automatically runs a session daemon is
5493 man:lttng-create(1), which you use to
5494 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5495 this is most of the time the first operation that you do, sometimes it's
5496 not. Some examples are:
5498 * <<list-instrumentation-points,List the available instrumentation points>>.
5499 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5501 [[tracing-group]] Each Unix user must have its own running session
5502 daemon to trace user applications. The session daemon that the root user
5503 starts is the only one allowed to control the LTTng kernel tracer. Users
5504 that are part of the _tracing group_ can control the root session
5505 daemon. The default tracing group name is `tracing`; you can set it to
5506 something else with the opt:lttng-sessiond(8):--group option when you
5507 start the root session daemon.
5509 To start a user session daemon:
5511 * Run man:lttng-sessiond(8):
5516 lttng-sessiond --daemonize
5520 To start the root session daemon:
5522 * Run man:lttng-sessiond(8) as the root user:
5527 sudo lttng-sessiond --daemonize
5531 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5532 start the session daemon in foreground.
5534 To stop a session daemon, use man:kill(1) on its process ID (standard
5537 Note that some Linux distributions could manage the LTTng session daemon
5538 as a service. In this case, you should use the service manager to
5539 start, restart, and stop session daemons.
5542 [[creating-destroying-tracing-sessions]]
5543 === Create and destroy a tracing session
5545 Almost all the LTTng control operations happen in the scope of
5546 a <<tracing-session,tracing session>>, which is the dialogue between the
5547 <<lttng-sessiond,session daemon>> and you.
5549 To create a tracing session with a generated name:
5551 * Use the man:lttng-create(1) command:
5560 The created tracing session's name is `auto` followed by the
5563 To create a tracing session with a specific name:
5565 * Use the optional argument of the man:lttng-create(1) command:
5570 lttng create my-session
5574 Replace `my-session` with the specific tracing session name.
5576 LTTng appends the creation date to the created tracing session's name.
5578 LTTng writes the traces of a tracing session in
5579 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5580 name of the tracing session. Note that the env:LTTNG_HOME environment
5581 variable defaults to `$HOME` if not set.
5583 To output LTTng traces to a non-default location:
5585 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5590 lttng create --output=/tmp/some-directory my-session
5594 You may create as many tracing sessions as you wish.
5596 To list all the existing tracing sessions for your Unix user:
5598 * Use the man:lttng-list(1) command:
5607 When you create a tracing session, it is set as the _current tracing
5608 session_. The following man:lttng(1) commands operate on the current
5609 tracing session when you don't specify one:
5611 [role="list-3-cols"]
5627 To change the current tracing session:
5629 * Use the man:lttng-set-session(1) command:
5634 lttng set-session new-session
5638 Replace `new-session` by the name of the new current tracing session.
5640 When you are done tracing in a given tracing session, you can destroy
5641 it. This operation frees the resources taken by the tracing session
5642 to destroy; it does not destroy the trace data that LTTng wrote for
5643 this tracing session.
5645 To destroy the current tracing session:
5647 * Use the man:lttng-destroy(1) command:
5657 [[list-instrumentation-points]]
5658 === List the available instrumentation points
5660 The <<lttng-sessiond,session daemon>> can query the running instrumented
5661 user applications and the Linux kernel to get a list of available
5662 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5663 they are tracepoints and system calls. For the user space tracing
5664 domain, they are tracepoints. For the other tracing domains, they are
5667 To list the available instrumentation points:
5669 * Use the man:lttng-list(1) command with the requested tracing domain's
5673 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5674 must be a root user, or it must be a member of the
5675 <<tracing-group,tracing group>>).
5676 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5677 kernel system calls (your Unix user must be a root user, or it must be
5678 a member of the tracing group).
5679 * opt:lttng-list(1):--userspace: user space tracepoints.
5680 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5681 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5682 * opt:lttng-list(1):--python: Python loggers.
5685 .List the available user space tracepoints.
5689 lttng list --userspace
5693 .List the available Linux kernel system call tracepoints.
5697 lttng list --kernel --syscall
5702 [[enabling-disabling-events]]
5703 === Create and enable an event rule
5705 Once you <<creating-destroying-tracing-sessions,create a tracing
5706 session>>, you can create <<event,event rules>> with the
5707 man:lttng-enable-event(1) command.
5709 You specify each condition with a command-line option. The available
5710 condition options are shown in the following table.
5712 [role="growable",cols="asciidoc,asciidoc,default"]
5713 .Condition command-line options for the man:lttng-enable-event(1) command.
5715 |Option |Description |Applicable tracing domains
5721 . +--probe=__ADDR__+
5722 . +--function=__ADDR__+
5725 Instead of using the default _tracepoint_ instrumentation type, use:
5727 . A Linux system call.
5728 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5729 . The entry and return points of a Linux function (symbol or address).
5733 |First positional argument.
5736 Tracepoint or system call name. In the case of a Linux KProbe or
5737 function, this is a custom name given to the event rule. With the
5738 JUL, log4j, and Python domains, this is a logger name.
5740 With a tracepoint, logger, or system call name, the last character
5741 can be `*` to match anything that remains.
5748 . +--loglevel=__LEVEL__+
5749 . +--loglevel-only=__LEVEL__+
5752 . Match only tracepoints or log statements with a logging level at
5753 least as severe as +__LEVEL__+.
5754 . Match only tracepoints or log statements with a logging level
5755 equal to +__LEVEL__+.
5757 See man:lttng-enable-event(1) for the list of available logging level
5760 |User space, JUL, log4j, and Python.
5762 |+--exclude=__EXCLUSIONS__+
5765 When you use a `*` character at the end of the tracepoint or logger
5766 name (first positional argument), exclude the specific names in the
5767 comma-delimited list +__EXCLUSIONS__+.
5770 User space, JUL, log4j, and Python.
5772 |+--filter=__EXPR__+
5775 Match only events which satisfy the expression +__EXPR__+.
5777 See man:lttng-enable-event(1) to learn more about the syntax of a
5784 You attach an event rule to a <<channel,channel>> on creation. If you do
5785 not specify the channel with the opt:lttng-enable-event(1):--channel
5786 option, and if the event rule to create is the first in its
5787 <<domain,tracing domain>> for a given tracing session, then LTTng
5788 creates a _default channel_ for you. This default channel is reused in
5789 subsequent invocations of the man:lttng-enable-event(1) command for the
5790 same tracing domain.
5792 An event rule is always enabled at creation time.
5794 The following examples show how you can combine the previous
5795 command-line options to create simple to more complex event rules.
5797 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5801 lttng enable-event --kernel sched_switch
5805 .Create an event rule matching four Linux kernel system calls (default channel).
5809 lttng enable-event --kernel --syscall open,write,read,close
5813 .Create event rules matching tracepoints with filter expressions (default channel).
5817 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5822 lttng enable-event --kernel --all \
5823 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5828 lttng enable-event --jul my_logger \
5829 --filter='$app.retriever:cur_msg_id > 3'
5832 IMPORTANT: Make sure to always quote the filter string when you
5833 use man:lttng(1) from a shell.
5836 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5840 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5843 IMPORTANT: Make sure to always quote the wildcard character when you
5844 use man:lttng(1) from a shell.
5847 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5851 lttng enable-event --python my-app.'*' \
5852 --exclude='my-app.module,my-app.hello'
5856 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5860 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5864 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5868 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5872 The event rules of a given channel form a whitelist: as soon as an
5873 emitted event passes one of them, LTTng can record the event. For
5874 example, an event named `my_app:my_tracepoint` emitted from a user space
5875 tracepoint with a `TRACE_ERROR` log level passes both of the following
5880 lttng enable-event --userspace my_app:my_tracepoint
5881 lttng enable-event --userspace my_app:my_tracepoint \
5882 --loglevel=TRACE_INFO
5885 The second event rule is redundant: the first one includes
5889 [[disable-event-rule]]
5890 === Disable an event rule
5892 To disable an event rule that you <<enabling-disabling-events,created>>
5893 previously, use the man:lttng-disable-event(1) command. This command
5894 disables _all_ the event rules (of a given tracing domain and channel)
5895 which match an instrumentation point. The other conditions are not
5896 supported as of LTTng{nbsp}{revision}.
5898 The LTTng tracer does not record an emitted event which passes
5899 a _disabled_ event rule.
5901 .Disable an event rule matching a Python logger (default channel).
5905 lttng disable-event --python my-logger
5909 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5913 lttng disable-event --jul '*'
5917 .Disable _all_ the event rules of the default channel.
5919 The opt:lttng-disable-event(1):--all-events option is not, like the
5920 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5921 equivalent of the event name `*` (wildcard): it disables _all_ the event
5922 rules of a given channel.
5926 lttng disable-event --jul --all-events
5930 NOTE: You cannot delete an event rule once you create it.
5934 === Get the status of a tracing session
5936 To get the status of the current tracing session, that is, its
5937 parameters, its channels, event rules, and their attributes:
5939 * Use the man:lttng-status(1) command:
5949 To get the status of any tracing session:
5951 * Use the man:lttng-list(1) command with the tracing session's name:
5956 lttng list my-session
5960 Replace `my-session` with the desired tracing session's name.
5963 [[basic-tracing-session-control]]
5964 === Start and stop a tracing session
5966 Once you <<creating-destroying-tracing-sessions,create a tracing
5968 <<enabling-disabling-events,create one or more event rules>>,
5969 you can start and stop the tracers for this tracing session.
5971 To start tracing in the current tracing session:
5973 * Use the man:lttng-start(1) command:
5982 LTTng is very flexible: you can launch user applications before
5983 or after the you start the tracers. The tracers only record the events
5984 if they pass enabled event rules and if they occur while the tracers are
5987 To stop tracing in the current tracing session:
5989 * Use the man:lttng-stop(1) command:
5998 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5999 records>> or lost sub-buffers since the last time you ran
6000 man:lttng-start(1), warnings are printed when you run the
6001 man:lttng-stop(1) command.
6004 [[enabling-disabling-channels]]
6005 === Create a channel
6007 Once you create a tracing session, you can create a <<channel,channel>>
6008 with the man:lttng-enable-channel(1) command.
6010 Note that LTTng automatically creates a default channel when, for a
6011 given <<domain,tracing domain>>, no channels exist and you
6012 <<enabling-disabling-events,create>> the first event rule. This default
6013 channel is named `channel0` and its attributes are set to reasonable
6014 values. Therefore, you only need to create a channel when you need
6015 non-default attributes.
6017 You specify each non-default channel attribute with a command-line
6018 option when you use the man:lttng-enable-channel(1) command. The
6019 available command-line options are:
6021 [role="growable",cols="asciidoc,asciidoc"]
6022 .Command-line options for the man:lttng-enable-channel(1) command.
6024 |Option |Description
6030 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
6031 the default _discard_ mode.
6033 |`--buffers-pid` (user space tracing domain only)
6036 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6037 instead of the default per-user buffering scheme.
6039 |+--subbuf-size=__SIZE__+
6042 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6043 either for each Unix user (default), or for each instrumented process.
6045 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6047 |+--num-subbuf=__COUNT__+
6050 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6051 for each Unix user (default), or for each instrumented process.
6053 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6055 |+--tracefile-size=__SIZE__+
6058 Set the maximum size of each trace file that this channel writes within
6059 a stream to +__SIZE__+ bytes instead of no maximum.
6061 See <<tracefile-rotation,Trace file count and size>>.
6063 |+--tracefile-count=__COUNT__+
6066 Limit the number of trace files that this channel creates to
6067 +__COUNT__+ channels instead of no limit.
6069 See <<tracefile-rotation,Trace file count and size>>.
6071 |+--switch-timer=__PERIODUS__+
6074 Set the <<channel-switch-timer,switch timer period>>
6075 to +__PERIODUS__+{nbsp}µs.
6077 |+--read-timer=__PERIODUS__+
6080 Set the <<channel-read-timer,read timer period>>
6081 to +__PERIODUS__+{nbsp}µs.
6083 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6086 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6090 You can only create a channel in the Linux kernel and user space
6091 <<domain,tracing domains>>: other tracing domains have their own channel
6092 created on the fly when <<enabling-disabling-events,creating event
6097 Because of a current LTTng limitation, you must create all channels
6098 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6099 tracing session, that is, before the first time you run
6102 Since LTTng automatically creates a default channel when you use the
6103 man:lttng-enable-event(1) command with a specific tracing domain, you
6104 cannot, for example, create a Linux kernel event rule, start tracing,
6105 and then create a user space event rule, because no user space channel
6106 exists yet and it's too late to create one.
6108 For this reason, make sure to configure your channels properly
6109 before starting the tracers for the first time!
6112 The following examples show how you can combine the previous
6113 command-line options to create simple to more complex channels.
6115 .Create a Linux kernel channel with default attributes.
6119 lttng enable-channel --kernel my-channel
6123 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6127 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6128 --buffers-pid my-channel
6132 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6136 lttng enable-channel --kernel --tracefile-count=8 \
6137 --tracefile-size=4194304 my-channel
6141 .Create a user space channel in overwrite (or _flight recorder_) mode.
6145 lttng enable-channel --userspace --overwrite my-channel
6149 You can <<enabling-disabling-events,create>> the same event rule in
6150 two different channels:
6154 lttng enable-event --userspace --channel=my-channel app:tp
6155 lttng enable-event --userspace --channel=other-channel app:tp
6158 If both channels are enabled, when a tracepoint named `app:tp` is
6159 reached, LTTng records two events, one for each channel.
6163 === Disable a channel
6165 To disable a specific channel that you <<enabling-disabling-channels,created>>
6166 previously, use the man:lttng-disable-channel(1) command.
6168 .Disable a specific Linux kernel channel.
6172 lttng disable-channel --kernel my-channel
6176 The state of a channel precedes the individual states of event rules
6177 attached to it: event rules which belong to a disabled channel, even if
6178 they are enabled, are also considered disabled.
6182 === Add context fields to a channel
6184 Event record fields in trace files provide important information about
6185 events that occured previously, but sometimes some external context may
6186 help you solve a problem faster. Examples of context fields are:
6188 * The **process ID**, **thread ID**, **process name**, and
6189 **process priority** of the thread in which the event occurs.
6190 * The **hostname** of the system on which the event occurs.
6191 * The current values of many possible **performance counters** using
6193 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6195 ** Branch instructions, misses, and loads.
6197 * Any context defined at the application level (supported for the
6198 JUL and log4j <<domain,tracing domains>>).
6200 To get the full list of available context fields, see
6201 `lttng add-context --list`. Some context fields are reserved for a
6202 specific <<domain,tracing domain>> (Linux kernel or user space).
6204 You add context fields to <<channel,channels>>. All the events
6205 that a channel with added context fields records contain those fields.
6207 To add context fields to one or all the channels of a given tracing
6210 * Use the man:lttng-add-context(1) command.
6212 .Add context fields to all the channels of the current tracing session.
6214 The following command line adds the virtual process identifier and
6215 the per-thread CPU cycles count fields to all the user space channels
6216 of the current tracing session.
6220 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6224 .Add a context field to a specific channel.
6226 The following command line adds the thread identifier context field
6227 to the Linux kernel channel named `my-channel` in the current
6232 lttng add-context --kernel --channel=my-channel --type=tid
6236 .Add an application-specific context field to a specific channel.
6238 The following command line adds the `cur_msg_id` context field of the
6239 `retriever` context retriever for all the instrumented
6240 <<java-application,Java applications>> recording <<event,event records>>
6241 in the channel named `my-channel`:
6245 lttng add-context --kernel --channel=my-channel \
6246 --type='$app:retriever:cur_msg_id'
6249 IMPORTANT: Make sure to always quote the `$` character when you
6250 use man:lttng-add-context(1) from a shell.
6253 NOTE: You cannot remove context fields from a channel once you add it.
6258 === Track process IDs
6260 It's often useful to allow only specific process IDs (PIDs) to emit
6261 events. For example, you may wish to record all the system calls made by
6262 a given process (à la http://linux.die.net/man/1/strace[strace]).
6264 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6265 purpose. Both commands operate on a whitelist of process IDs. You _add_
6266 entries to this whitelist with the man:lttng-track(1) command and remove
6267 entries with the man:lttng-untrack(1) command. Any process which has one
6268 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6269 an enabled <<event,event rule>>.
6271 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6272 process with a given tracked ID exit and another process be given this
6273 ID, then the latter would also be allowed to emit events.
6275 .Track and untrack process IDs.
6277 For the sake of the following example, assume the target system has 16
6281 <<creating-destroying-tracing-sessions,create a tracing session>>,
6282 the whitelist contains all the possible PIDs:
6285 .All PIDs are tracked.
6286 image::track-all.png[]
6288 When the whitelist is full and you use the man:lttng-track(1) command to
6289 specify some PIDs to track, LTTng first clears the whitelist, then it
6290 tracks the specific PIDs. After:
6294 lttng track --pid=3,4,7,10,13
6300 .PIDs 3, 4, 7, 10, and 13 are tracked.
6301 image::track-3-4-7-10-13.png[]
6303 You can add more PIDs to the whitelist afterwards:
6307 lttng track --pid=1,15,16
6313 .PIDs 1, 15, and 16 are added to the whitelist.
6314 image::track-1-3-4-7-10-13-15-16.png[]
6316 The man:lttng-untrack(1) command removes entries from the PID tracker's
6317 whitelist. Given the previous example, the following command:
6321 lttng untrack --pid=3,7,10,13
6324 leads to this whitelist:
6327 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6328 image::track-1-4-15-16.png[]
6330 LTTng can track all possible PIDs again using the opt:track(1):--all
6335 lttng track --pid --all
6338 The result is, again:
6341 .All PIDs are tracked.
6342 image::track-all.png[]
6345 .Track only specific PIDs
6347 A very typical use case with PID tracking is to start with an empty
6348 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6349 then add PIDs manually while tracers are active. You can accomplish this
6350 by using the opt:lttng-untrack(1):--all option of the
6351 man:lttng-untrack(1) command to clear the whitelist after you
6352 <<creating-destroying-tracing-sessions,create a tracing session>>:
6356 lttng untrack --pid --all
6362 .No PIDs are tracked.
6363 image::untrack-all.png[]
6365 If you trace with this whitelist configuration, the tracer records no
6366 events for this <<domain,tracing domain>> because no processes are
6367 tracked. You can use the man:lttng-track(1) command as usual to track
6368 specific PIDs, for example:
6372 lttng track --pid=6,11
6378 .PIDs 6 and 11 are tracked.
6379 image::track-6-11.png[]
6384 [[saving-loading-tracing-session]]
6385 === Save and load tracing session configurations
6387 Configuring a <<tracing-session,tracing session>> can be long. Some of
6388 the tasks involved are:
6390 * <<enabling-disabling-channels,Create channels>> with
6391 specific attributes.
6392 * <<adding-context,Add context fields>> to specific channels.
6393 * <<enabling-disabling-events,Create event rules>> with specific log
6394 level and filter conditions.
6396 If you use LTTng to solve real world problems, chances are you have to
6397 record events using the same tracing session setup over and over,
6398 modifying a few variables each time in your instrumented program
6399 or environment. To avoid constant tracing session reconfiguration,
6400 the man:lttng(1) command-line tool can save and load tracing session
6401 configurations to/from XML files.
6403 To save a given tracing session configuration:
6405 * Use the man:lttng-save(1) command:
6410 lttng save my-session
6414 Replace `my-session` with the name of the tracing session to save.
6416 LTTng saves tracing session configurations to
6417 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6418 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6419 the opt:lttng-save(1):--output-path option to change this destination
6422 LTTng saves all configuration parameters, for example:
6424 * The tracing session name.
6425 * The trace data output path.
6426 * The channels with their state and all their attributes.
6427 * The context fields you added to channels.
6428 * The event rules with their state, log level and filter conditions.
6430 To load a tracing session:
6432 * Use the man:lttng-load(1) command:
6437 lttng load my-session
6441 Replace `my-session` with the name of the tracing session to load.
6443 When LTTng loads a configuration, it restores your saved tracing session
6444 as if you just configured it manually.
6446 See man:lttng(1) for the complete list of command-line options. You
6447 can also save and load all many sessions at a time, and decide in which
6448 directory to output the XML files.
6451 [[sending-trace-data-over-the-network]]
6452 === Send trace data over the network
6454 LTTng can send the recorded trace data to a remote system over the
6455 network instead of writing it to the local file system.
6457 To send the trace data over the network:
6459 . On the _remote_ system (which can also be the target system),
6460 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6469 . On the _target_ system, create a tracing session configured to
6470 send trace data over the network:
6475 lttng create my-session --set-url=net://remote-system
6479 Replace `remote-system` by the host name or IP address of the
6480 remote system. See man:lttng-create(1) for the exact URL format.
6482 . On the target system, use the man:lttng(1) command-line tool as usual.
6483 When tracing is active, the target's consumer daemon sends sub-buffers
6484 to the relay daemon running on the remote system intead of flushing
6485 them to the local file system. The relay daemon writes the received
6486 packets to the local file system.
6488 The relay daemon writes trace files to
6489 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6490 +__hostname__+ is the host name of the target system and +__session__+
6491 is the tracing session name. Note that the env:LTTNG_HOME environment
6492 variable defaults to `$HOME` if not set. Use the
6493 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6494 trace files to another base directory.
6499 === View events as LTTng emits them (noch:{LTTng} live)
6501 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6502 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6503 display events as LTTng emits them on the target system while tracing is
6506 The relay daemon creates a _tee_: it forwards the trace data to both
6507 the local file system and to connected live viewers:
6510 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6515 . On the _target system_, create a <<tracing-session,tracing session>>
6521 lttng create --live my-session
6525 This spawns a local relay daemon.
6527 . Start the live viewer and configure it to connect to the relay
6528 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6533 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6540 * `hostname` with the host name of the target system.
6541 * `my-session` with the name of the tracing session to view.
6544 . Configure the tracing session as usual with the man:lttng(1)
6545 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6547 You can list the available live tracing sessions with Babeltrace:
6551 babeltrace --input-format=lttng-live net://localhost
6554 You can start the relay daemon on another system. In this case, you need
6555 to specify the relay daemon's URL when you create the tracing session
6556 with the opt:lttng-create(1):--set-url option. You also need to replace
6557 `localhost` in the procedure above with the host name of the system on
6558 which the relay daemon is running.
6560 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6561 command-line options.
6565 [[taking-a-snapshot]]
6566 === Take a snapshot of the current sub-buffers of a tracing session
6568 The normal behavior of LTTng is to append full sub-buffers to growing
6569 trace data files. This is ideal to keep a full history of the events
6570 that occurred on the target system, but it can
6571 represent too much data in some situations. For example, you may wish
6572 to trace your application continuously until some critical situation
6573 happens, in which case you only need the latest few recorded
6574 events to perform the desired analysis, not multi-gigabyte trace files.
6576 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6577 current sub-buffers of a given <<tracing-session,tracing session>>.
6578 LTTng can write the snapshot to the local file system or send it over
6583 . Create a tracing session in _snapshot mode_:
6588 lttng create --snapshot my-session
6592 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6593 <<channel,channels>> created in this mode is automatically set to
6594 _overwrite_ (flight recorder mode).
6596 . Configure the tracing session as usual with the man:lttng(1)
6597 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6599 . **Optional**: When you need to take a snapshot,
6600 <<basic-tracing-session-control,stop tracing>>.
6602 You can take a snapshot when the tracers are active, but if you stop
6603 them first, you are sure that the data in the sub-buffers does not
6604 change before you actually take the snapshot.
6611 lttng snapshot record --name=my-first-snapshot
6615 LTTng writes the current sub-buffers of all the current tracing
6616 session's channels to trace files on the local file system. Those trace
6617 files have `my-first-snapshot` in their name.
6619 There is no difference between the format of a normal trace file and the
6620 format of a snapshot: viewers of LTTng traces also support LTTng
6623 By default, LTTng writes snapshot files to the path shown by
6624 `lttng snapshot list-output`. You can change this path or decide to send
6625 snapshots over the network using either:
6627 . An output path or URL that you specify when you create the
6629 . An snapshot output path or URL that you add using
6630 `lttng snapshot add-output`
6631 . An output path or URL that you provide directly to the
6632 `lttng snapshot record` command.
6634 Method 3 overrides method 2, which overrides method 1. When you
6635 specify a URL, a relay daemon must listen on a remote system (see
6636 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6641 === Use the machine interface
6643 With any command of the man:lttng(1) command-line tool, you can set the
6644 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6645 XML machine interface output, for example:
6649 lttng --mi=xml enable-event --kernel --syscall open
6652 A schema definition (XSD) is
6653 https://github.com/lttng/lttng-tools/blob/stable-2.8/src/common/mi-lttng-3.0.xsd[available]
6654 to ease the integration with external tools as much as possible.
6658 [[metadata-regenerate]]
6659 === Regenerate the metadata of an LTTng trace
6661 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6662 data stream files and a metadata file. This metadata file contains,
6663 amongst other things, information about the offset of the clock sources
6664 used to timestamp <<event,event records>> when tracing.
6666 If, once a <<tracing-session,tracing session>> is
6667 <<basic-tracing-session-control,started>>, a major
6668 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6669 happens, the trace's clock offset also needs to be updated. You
6670 can use the man:lttng-metadata(1) command to do so.
6672 The main use case of this command is to allow a system to boot with
6673 an incorrect wall time and trace it with LTTng before its wall time
6674 is corrected. Once the system is known to be in a state where its
6675 wall time is correct, it can run `lttng metadata regenerate`.
6677 To regenerate the metadata of an LTTng trace:
6679 * Use the `regenerate` action of the man:lttng-metadata(1) command:
6684 lttng metadata regenerate
6690 `lttng metadata regenerate` has the following limitations:
6692 * Tracing session <<creating-destroying-tracing-sessions,created>>
6694 * User space <<channel,channels>>, if any, using
6695 <<channel-buffering-schemes,per-user buffering>>.
6700 [[persistent-memory-file-systems]]
6701 === Record trace data on persistent memory file systems
6703 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6704 (NVRAM) is random-access memory that retains its information when power
6705 is turned off (non-volatile). Systems with such memory can store data
6706 structures in RAM and retrieve them after a reboot, without flushing
6707 to typical _storage_.
6709 Linux supports NVRAM file systems thanks to either
6710 http://pramfs.sourceforge.net/[PRAMFS] or
6711 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6712 (requires Linux 4.1+).
6714 This section does not describe how to operate such file systems;
6715 we assume that you have a working persistent memory file system.
6717 When you create a <<tracing-session,tracing session>>, you can specify
6718 the path of the shared memory holding the sub-buffers. If you specify a
6719 location on an NVRAM file system, then you can retrieve the latest
6720 recorded trace data when the system reboots after a crash.
6722 To record trace data on a persistent memory file system and retrieve the
6723 trace data after a system crash:
6725 . Create a tracing session with a sub-buffer shared memory path located
6726 on an NVRAM file system:
6731 lttng create --shm-path=/path/to/shm
6735 . Configure the tracing session as usual with the man:lttng(1)
6736 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6738 . After a system crash, use the man:lttng-crash(1) command-line tool to
6739 view the trace data recorded on the NVRAM file system:
6744 lttng-crash /path/to/shm
6748 The binary layout of the ring buffer files is not exactly the same as
6749 the trace files layout. This is why you need to use man:lttng-crash(1)
6750 instead of your preferred trace viewer directly.
6752 To convert the ring buffer files to LTTng trace files:
6754 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6759 lttng-crash --extract=/path/to/trace /path/to/shm
6767 [[lttng-modules-ref]]
6768 === noch:{LTTng-modules}
6771 [[lttng-modules-tp-fields]]
6772 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6774 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6775 tracepoint fields, which must be listed within `TP_FIELDS()` in
6776 `LTTNG_TRACEPOINT_EVENT()`, are:
6778 [role="func-desc growable",cols="asciidoc,asciidoc"]
6779 .Available macros to define LTTng-modules tracepoint fields
6781 |Macro |Description and parameters
6784 +ctf_integer(__t__, __n__, __e__)+
6786 +ctf_integer_nowrite(__t__, __n__, __e__)+
6788 +ctf_user_integer(__t__, __n__, __e__)+
6790 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6792 Standard integer, displayed in base 10.
6795 Integer C type (`int`, `long`, `size_t`, ...).
6801 Argument expression.
6804 +ctf_integer_hex(__t__, __n__, __e__)+
6806 +ctf_user_integer_hex(__t__, __n__, __e__)+
6808 Standard integer, displayed in base 16.
6817 Argument expression.
6819 |+ctf_integer_oct(__t__, __n__, __e__)+
6821 Standard integer, displayed in base 8.
6830 Argument expression.
6833 +ctf_integer_network(__t__, __n__, __e__)+
6835 +ctf_user_integer_network(__t__, __n__, __e__)+
6837 Integer in network byte order (big-endian), displayed in base 10.
6846 Argument expression.
6849 +ctf_integer_network_hex(__t__, __n__, __e__)+
6851 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6853 Integer in network byte order, displayed in base 16.
6862 Argument expression.
6865 +ctf_string(__n__, __e__)+
6867 +ctf_string_nowrite(__n__, __e__)+
6869 +ctf_user_string(__n__, __e__)+
6871 +ctf_user_string_nowrite(__n__, __e__)+
6873 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6879 Argument expression.
6882 +ctf_array(__t__, __n__, __e__, __s__)+
6884 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6886 +ctf_user_array(__t__, __n__, __e__, __s__)+
6888 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6890 Statically-sized array of integers.
6893 Array element C type.
6899 Argument expression.
6905 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
6907 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6909 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
6911 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6913 Statically-sized array of bits.
6915 The type of +__e__+ must be an integer type. +__s__+ is the number
6916 of elements of such type in +__e__+, not the number of bits.
6919 Array element C type.
6925 Argument expression.
6931 +ctf_array_text(__t__, __n__, __e__, __s__)+
6933 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6935 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6937 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6939 Statically-sized array, printed as text.
6941 The string does not need to be null-terminated.
6944 Array element C type (always `char`).
6950 Argument expression.
6956 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6958 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6960 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6962 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6964 Dynamically-sized array of integers.
6966 The type of +__E__+ must be unsigned.
6969 Array element C type.
6975 Argument expression.
6978 Length expression C type.
6984 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6986 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6988 Dynamically-sized array of integers, displayed in base 16.
6990 The type of +__E__+ must be unsigned.
6993 Array element C type.
6999 Argument expression.
7002 Length expression C type.
7007 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7009 Dynamically-sized array of integers in network byte order (big-endian),
7010 displayed in base 10.
7012 The type of +__E__+ must be unsigned.
7015 Array element C type.
7021 Argument expression.
7024 Length expression C type.
7030 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7032 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7034 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7036 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7038 Dynamically-sized array of bits.
7040 The type of +__e__+ must be an integer type. +__s__+ is the number
7041 of elements of such type in +__e__+, not the number of bits.
7043 The type of +__E__+ must be unsigned.
7046 Array element C type.
7052 Argument expression.
7055 Length expression C type.
7061 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7063 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7065 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7067 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7069 Dynamically-sized array, displayed as text.
7071 The string does not need to be null-terminated.
7073 The type of +__E__+ must be unsigned.
7075 The behaviour is undefined if +__e__+ is `NULL`.
7078 Sequence element C type (always `char`).
7084 Argument expression.
7087 Length expression C type.
7093 Use the `_user` versions when the argument expression, `e`, is
7094 a user space address. In the cases of `ctf_user_integer*()` and
7095 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7098 The `_nowrite` versions omit themselves from the session trace, but are
7099 otherwise identical. This means the `_nowrite` fields won't be written
7100 in the recorded trace. Their primary purpose is to make some
7101 of the event context available to the
7102 <<enabling-disabling-events,event filters>> without having to
7103 commit the data to sub-buffers.
7109 Terms related to LTTng and to tracing in general:
7112 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7113 the cmd:babeltrace command, some libraries, and Python bindings.
7115 <<channel-buffering-schemes,buffering scheme>>::
7116 A layout of sub-buffers applied to a given channel.
7118 <<channel,channel>>::
7119 An entity which is responsible for a set of ring buffers.
7121 <<event,Event rules>> are always attached to a specific channel.
7124 A reference of time for a tracer.
7126 <<lttng-consumerd,consumer daemon>>::
7127 A process which is responsible for consuming the full sub-buffers
7128 and write them to a file system or send them over the network.
7130 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7131 mode in which the tracer _discards_ new event records when there's no
7132 sub-buffer space left to store them.
7135 The consequence of the execution of an instrumentation
7136 point, like a tracepoint that you manually place in some source code,
7137 or a Linux kernel KProbe.
7139 An event is said to _occur_ at a specific time. Different actions can
7140 be taken upon the occurance of an event, like record the event's payload
7143 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7144 The mechanism by which event records of a given channel are lost
7145 (not recorded) when there is no sub-buffer space left to store them.
7147 [[def-event-name]]event name::
7148 The name of an event, which is also the name of the event record.
7149 This is also called the _instrumentation point name_.
7152 A record, in a trace, of the payload of an event which occured.
7154 <<event,event rule>>::
7155 Set of conditions which must be satisfied for one or more occuring
7156 events to be recorded.
7158 `java.util.logging`::
7160 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7162 <<instrumenting,instrumentation>>::
7163 The use of LTTng probes to make a piece of software traceable.
7165 instrumentation point::
7166 A point in the execution path of a piece of software that, when
7167 reached by this execution, can emit an event.
7169 instrumentation point name::
7170 See _<<def-event-name,event name>>_.
7173 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7174 developed by the Apache Software Foundation.
7177 Level of severity of a log statement or user space
7178 instrumentation point.
7181 The _Linux Trace Toolkit: next generation_ project.
7183 <<lttng-cli,cmd:lttng>>::
7184 A command-line tool provided by the LTTng-tools project which you
7185 can use to send and receive control messages to and from a
7189 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7190 which is a set of analyzing programs that are used to obtain a
7191 higher level view of an LTTng trace.
7193 cmd:lttng-consumerd::
7194 The name of the consumer daemon program.
7197 A utility provided by the LTTng-tools project which can convert
7198 ring buffer files (usually
7199 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7202 LTTng Documentation::
7205 <<lttng-live,LTTng live>>::
7206 A communication protocol between the relay daemon and live viewers
7207 which makes it possible to see events "live", as they are received by
7210 <<lttng-modules,LTTng-modules>>::
7211 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7212 which contains the Linux kernel modules to make the Linux kernel
7213 instrumentation points available for LTTng tracing.
7216 The name of the relay daemon program.
7218 cmd:lttng-sessiond::
7219 The name of the session daemon program.
7222 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7223 contains the various programs and libraries used to
7224 <<controlling-tracing,control tracing>>.
7226 <<lttng-ust,LTTng-UST>>::
7227 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7228 contains libraries to instrument user applications.
7230 <<lttng-ust-agents,LTTng-UST Java agent>>::
7231 A Java package provided by the LTTng-UST project to allow the
7232 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7235 <<lttng-ust-agents,LTTng-UST Python agent>>::
7236 A Python package provided by the LTTng-UST project to allow the
7237 LTTng instrumentation of Python logging statements.
7239 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7240 The event loss mode in which new event records overwrite older
7241 event records when there's no sub-buffer space left to store them.
7243 <<channel-buffering-schemes,per-process buffering>>::
7244 A buffering scheme in which each instrumented process has its own
7245 sub-buffers for a given user space channel.
7247 <<channel-buffering-schemes,per-user buffering>>::
7248 A buffering scheme in which all the processes of a Unix user share the
7249 same sub-buffer for a given user space channel.
7251 <<lttng-relayd,relay daemon>>::
7252 A process which is responsible for receiving the trace data sent by
7253 a distant consumer daemon.
7256 A set of sub-buffers.
7258 <<lttng-sessiond,session daemon>>::
7259 A process which receives control commands from you and orchestrates
7260 the tracers and various LTTng daemons.
7262 <<taking-a-snapshot,snapshot>>::
7263 A copy of the current data of all the sub-buffers of a given tracing
7264 session, saved as trace files.
7267 One part of an LTTng ring buffer which contains event records.
7270 The time information attached to an event when it is emitted.
7273 A set of files which are the concatenations of one or more
7274 flushed sub-buffers.
7277 The action of recording the events emitted by an application
7278 or by a system, or to initiate such recording by controlling
7282 The http://tracecompass.org[Trace Compass] project and application.
7285 An instrumentation point using the tracepoint mechanism of the Linux
7286 kernel or of LTTng-UST.
7288 tracepoint definition::
7289 The definition of a single tracepoint.
7292 The name of a tracepoint.
7294 tracepoint provider::
7295 A set of functions providing tracepoints to an instrumented user
7298 Not to be confused with a _tracepoint provider package_: many tracepoint
7299 providers can exist within a tracepoint provider package.
7301 tracepoint provider package::
7302 One or more tracepoint providers compiled as an object file or as
7306 A software which records emitted events.
7308 <<domain,tracing domain>>::
7309 A namespace for event sources.
7312 The Unix group in which a Unix user can be to be allowed to trace the
7315 <<tracing-session,tracing session>>::
7316 A stateful dialogue between you and a <<lttng-sessiond,session
7320 An application running in user space, as opposed to a Linux kernel
7321 module, for example.