1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 LTTng{nbsp}{revision} bears the name _Isseki Nicho_. The result of a
78 collaboration between http://www.dieuduciel.com/[Dieu du Ciel!] and
79 Nagano-based Shiga Kogen,
80 https://www.beeradvocate.com/beer/profile/1141/53111/[_**Isseki
81 Nicho**_] is a strong Imperial Dark Saison offering a rich roasted malt
82 flavor combined with a complex fruity finish typical of Saison yeasts.
84 New features and changes in LTTng{nbsp}{revision}:
86 * **Tracing control**:
87 ** You can attach <<java-application-context,Java application-specific
88 context fields>> to a <<channel,channel>> with the
89 man:lttng-add-context(1) command:
94 lttng add-context --jul --type='$app.retriever:cur_msg_id'
98 Here, `$app` is the prefix of all application-specific context fields,
99 `retriever` names a _context information retriever_ defined at the
100 application level, and `cur_msg_id` names a context field read from this
103 Both the `java.util.logging` and Apache log4j <<domain,tracing domains>>
106 ** You can use Java application-specific <<adding-context,context>>
107 fields in the <<enabling-disabling-events,filter expression>> of an
108 <<event,event rule>>:
113 lttng enable-event --log4j my_logger \
114 --filter='$app.retriever:cur_msg_id == 23'
118 ** New `lttng status` command which is the equivalent of +lttng list
119 __CUR__+, where +__CUR__+ is the name of the current
120 <<tracing-session,tracing session>>.
122 See man:lttng-status(1).
124 ** New `lttng metadata regenerate` command to
125 <<metadata-regenerate,regenerate the metadata file of an LTTng
126 trace>> at any moment. This command is meant to be used to resample
127 the wall time following a major
128 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
129 so that a system which boots with an incorrect wall time can be
130 traced before its wall time is NTP-corrected.
132 See man:lttng-metadata(1).
134 ** New command-line interface warnings when <<event,event records>> or
135 whole sub-buffers are
136 <<channel-overwrite-mode-vs-discard-mode,lost>>. The warning messages
137 are printed when a <<tracing-session,tracing session>> is
138 <<basic-tracing-session-control,stopped>> (man:lttng-stop(1)
141 * **User space tracing**:
142 ** Shared object base address dump in order to map <<event,event
143 records>> to original source location (file and line number).
145 If you attach the `ip` and `vpid` <<adding-context,context fields>> to a
146 user space <<channel,channel>> and if you use the
147 <<liblttng-ust-dl,path:{liblttng-ust-dl.so} helper>>, you can retrieve
148 the source location where a given event record was generated.
150 The http://diamon.org/babeltrace/[Babeltrace] trace viewer supports this
151 state dump and those context fields since version 1.4 to print the
152 source location of a given event record. http://tracecompass.org/[Trace
153 Compass] also supports this since version 2.0.
155 ** A <<java-application,Java application>> which uses
156 `java.util.logging` now adds an LTTng-UST log handler to the desired
159 The previous workflow was to initialize the LTTng-UST Java agent
160 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
161 an LTTng-UST log handler to the root loggers.
163 ** A <<java-application,Java application>> which uses Apache log4j now
164 adds an LTTng-UST log appender to the desired log4j loggers.
166 The previous workflow was to initialize the LTTng-UST Java agent
167 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
168 an LTTng-UST appender to the root loggers.
170 ** Any <<java-application,Java application>> can provide
171 <<java-application-context,dynamic context fields>> while running
172 thanks to a new API provided by the <<lttng-ust-agents,LTTng-UST Java
173 agent>>. You can require LTTng to record specific context fields in
174 event records, and you can use them in the filter expression of
175 <<event,event rules>>.
177 * **Linux kernel tracing**:
178 ** The LTTng kernel modules can now be built into a Linux kernel image,
179 that is, not as loadable modules.
182 https://github.com/lttng/lttng-modules/blob/stable-{revision}/README.md#kernel-built-in-support[`README.md`]
185 ** New instrumentation:
186 *** ARM64 architecture support.
188 *** x86 `irq_vectors`.
189 ** New <<adding-context,context fields>>:
192 *** `need_reschedule`
193 *** `migratable` (specific to RT-Preempt)
194 ** Clock source plugin support for advanced cases where a custom source
195 of time is needed to timestamp LTTng event records.
197 See https://github.com/lttng/lttng-modules/blob/stable-{revision}/lttng-clock.h[`lttng-clock.h`]
198 for an overview of the small API.
201 ** The link:/man[man pages] of the man:lttng(1) command-line tool are
202 split into one man page per command (à la Git), for example:
207 man lttng-enable-event
211 You can also use the `--help` option of any man:lttng(1) command to
214 The content and formatting of all the LTTng man pages has improved
221 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
222 generation_ is a modern toolkit for tracing Linux systems and
223 applications. So your first question might be:
230 As the history of software engineering progressed and led to what
231 we now take for granted--complex, numerous and
232 interdependent software applications running in parallel on
233 sophisticated operating systems like Linux--the authors of such
234 components, software developers, began feeling a natural
235 urge to have tools that would ensure the robustness and good performance
236 of their masterpieces.
238 One major achievement in this field is, inarguably, the
239 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
240 an essential tool for developers to find and fix bugs. But even the best
241 debugger won't help make your software run faster, and nowadays, faster
242 software means either more work done by the same hardware, or cheaper
243 hardware for the same work.
245 A _profiler_ is often the tool of choice to identify performance
246 bottlenecks. Profiling is suitable to identify _where_ performance is
247 lost in a given software. The profiler outputs a profile, a statistical
248 summary of observed events, which you may use to discover which
249 functions took the most time to execute. However, a profiler won't
250 report _why_ some identified functions are the bottleneck. Bottlenecks
251 might only occur when specific conditions are met, conditions that are
252 sometimes impossible to capture by a statistical profiler, or impossible
253 to reproduce with an application altered by the overhead of an
254 event-based profiler. For a thorough investigation of software
255 performance issues, a history of execution is essential, with the
256 recorded values of variables and context fields you choose, and
257 with as little influence as possible on the instrumented software. This
258 is where tracing comes in handy.
260 _Tracing_ is a technique used to understand what goes on in a running
261 software system. The software used for tracing is called a _tracer_,
262 which is conceptually similar to a tape recorder. When recording,
263 specific instrumentation points placed in the software source code
264 generate events that are saved on a giant tape: a _trace_ file. You
265 can trace user applications and the operating system at the same time,
266 opening the possibility of resolving a wide range of problems that would
267 otherwise be extremely challenging.
269 Tracing is often compared to _logging_. However, tracers and loggers are
270 two different tools, serving two different purposes. Tracers are
271 designed to record much lower-level events that occur much more
272 frequently than log messages, often in the range of thousands per
273 second, with very little execution overhead. Logging is more appropriate
274 for a very high-level analysis of less frequent events: user accesses,
275 exceptional conditions (errors and warnings, for example), database
276 transactions, instant messaging communications, and such. Simply put,
277 logging is one of the many use cases that can be satisfied with tracing.
279 The list of recorded events inside a trace file can be read manually
280 like a log file for the maximum level of detail, but it is generally
281 much more interesting to perform application-specific analyses to
282 produce reduced statistics and graphs that are useful to resolve a
283 given problem. Trace viewers and analyzers are specialized tools
286 In the end, this is what LTTng is: a powerful, open source set of
287 tools to trace the Linux kernel and user applications at the same time.
288 LTTng is composed of several components actively maintained and
289 developed by its link:/community/#where[community].
292 [[lttng-alternatives]]
293 === Alternatives to noch:{LTTng}
295 Excluding proprietary solutions, a few competing software tracers
298 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
299 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
300 user scripts and is responsible for loading code into the
301 Linux kernel for further execution and collecting the outputted data.
302 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
303 subsystem in the Linux kernel in which a virtual machine can execute
304 programs passed from the user space to the kernel. You can attach
305 such programs to tracepoints and KProbes thanks to a system call, and
306 they can output data to the user space when executed thanks to
307 different mechanisms (pipe, VM register values, and eBPF maps, to name
309 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
310 is the de facto function tracer of the Linux kernel. Its user
311 interface is a set of special files in sysfs.
312 * https://perf.wiki.kernel.org/[perf] is
313 a performance analyzing tool for Linux which supports hardware
314 performance counters, tracepoints, as well as other counters and
315 types of probes. perf's controlling utility is the cmd:perf command
317 * http://linux.die.net/man/1/strace[strace]
318 is a command-line utility which records system calls made by a
319 user process, as well as signal deliveries and changes of process
320 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
321 to fulfill its function.
322 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
323 analyze Linux kernel events. You write scripts, or _chisels_ in
324 sysdig's jargon, in Lua and sysdig executes them while the system is
325 being traced or afterwards. sysdig's interface is the cmd:sysdig
326 command-line tool as well as the curses-based cmd:csysdig tool.
327 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
328 user space tracer which uses custom user scripts to produce plain text
329 traces. SystemTap converts the scripts to the C language, and then
330 compiles them as Linux kernel modules which are loaded to produce
331 trace data. SystemTap's primary user interface is the cmd:stap
334 The main distinctive features of LTTng is that it produces correlated
335 kernel and user space traces, as well as doing so with the lowest
336 overhead amongst other solutions. It produces trace files in the
337 http://diamon.org/ctf[CTF] format, a file format optimized
338 for the production and analyses of multi-gigabyte data.
340 LTTng is the result of more than 10 years of active open source
341 development by a community of passionate developers.
342 LTTng{nbsp}{revision} is currently available on major desktop and server
345 The main interface for tracing control is a single command-line tool
346 named cmd:lttng. The latter can create several tracing sessions, enable
347 and disable events on the fly, filter events efficiently with custom
348 user expressions, start and stop tracing, and much more. LTTng can
349 record the traces on the file system or send them over the network, and
350 keep them totally or partially. You can view the traces once tracing
351 becomes inactive or in real-time.
353 <<installing-lttng,Install LTTng now>> and
354 <<getting-started,start tracing>>!
360 **LTTng** is a set of software <<plumbing,components>> which interact to
361 <<instrumenting,instrument>> the Linux kernel and user applications, and
362 to <<controlling-tracing,control tracing>> (start and stop
363 tracing, enable and disable event rules, and the rest). Those
364 components are bundled into the following packages:
366 * **LTTng-tools**: Libraries and command-line interface to
368 * **LTTng-modules**: Linux kernel modules to instrument and
370 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
371 trace user applications.
373 Most distributions mark the LTTng-modules and LTTng-UST packages as
374 optional when installing LTTng-tools (which is always required). In the
375 following sections, we always provide the steps to install all three,
378 * You only need to install LTTng-modules if you intend to trace the
380 * You only need to install LTTng-UST if you intend to trace user
384 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 2 December 2016.
386 |Distribution |Available in releases |Alternatives
388 |https://www.ubuntu.com/[Ubuntu]
389 |<<ubuntu,Ubuntu{nbsp}16.10 _Yakkety Yak_>>.
390 |LTTng{nbsp}{revision} for Ubuntu{nbsp}14.04 _Trusty Tahr_
391 and Ubuntu{nbsp}16.04 _Xenial Xerus_:
392 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
394 LTTng{nbsp}2.9 for Ubuntu{nbsp}14.04 _Trusty Tahr_
395 and Ubuntu{nbsp}16.04 _Xenial Xerus_:
396 link:/docs/v2.9#doc-ubuntu-ppa[use the LTTng Stable{nbsp}2.9 PPA].
398 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
399 other Ubuntu releases.
401 |https://getfedora.org/[Fedora]
402 |<<fedora,Fedora{nbsp}25>>.
403 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
404 other Fedora releases.
406 |https://www.debian.org/[Debian]
407 |<<debian,Debian "stretch" (testing)>>.
408 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
409 previous Debian releases.
411 |https://www.opensuse.org/[openSUSE]
413 |<<building-from-source,Build LTTng{nbsp}{revision} from source>>.
415 |https://www.archlinux.org/[Arch Linux]
417 |link:/docs/v2.9#doc-arch-linux[LTTng{nbsp}2.9 from the AUR].
419 |https://alpinelinux.org/[Alpine Linux]
420 |<<alpine-linux,Alpine Linux "edge">>.
421 |LTTng{nbsp}{revision} for Alpine Linux{nbsp}3.5 (not released yet).
423 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
424 other Alpine Linux releases.
426 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
427 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
430 |https://buildroot.org/[Buildroot]
431 |<<buildroot,Buildroot 2016.11>>.
432 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
433 other Buildroot releases.
435 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
436 https://www.yoctoproject.org/[Yocto]
437 |<<oe-yocto,Yocto Project{nbsp}2.2 _Morty_>> (`openembedded-core` layer).
438 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
439 other OpenEmbedded releases.
444 === [[ubuntu-official-repositories]]Ubuntu
446 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.10 _Yakkety Yak_.
447 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
448 Stable{nbsp}{revision} PPA>>.
450 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}16.10 _Yakkety Yak_:
452 . Install the main LTTng{nbsp}{revision} packages:
457 sudo apt-get install lttng-tools
458 sudo apt-get install lttng-modules-dkms
459 sudo apt-get install liblttng-ust-dev
463 . **If you need to instrument and trace
464 <<java-application,Java applications>>**, install the LTTng-UST
470 sudo apt-get install liblttng-ust-agent-java
474 . **If you need to instrument and trace
475 <<python-application,Python{nbsp}3 applications>>**, install the
476 LTTng-UST Python agent:
481 sudo apt-get install python3-lttngust
487 ==== noch:{LTTng} Stable {revision} PPA
489 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
490 Stable{nbsp}{revision} PPA] offers the latest stable
491 LTTng{nbsp}{revision} packages for:
493 * Ubuntu{nbsp}14.04 _Trusty Tahr_
494 * Ubuntu{nbsp}16.04 _Xenial Xerus_
496 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
498 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
504 sudo apt-add-repository ppa:lttng/stable-2.8
509 . Install the main LTTng{nbsp}{revision} packages:
514 sudo apt-get install lttng-tools
515 sudo apt-get install lttng-modules-dkms
516 sudo apt-get install liblttng-ust-dev
520 . **If you need to instrument and trace
521 <<java-application,Java applications>>**, install the LTTng-UST
527 sudo apt-get install liblttng-ust-agent-java
531 . **If you need to instrument and trace
532 <<python-application,Python{nbsp}3 applications>>**, install the
533 LTTng-UST Python agent:
538 sudo apt-get install python3-lttngust
546 To install LTTng{nbsp}{revision} on Fedora{nbsp}25:
548 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
554 sudo yum install lttng-tools
555 sudo yum install lttng-ust
559 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
565 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
566 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
567 cd lttng-modules-2.8.* &&
569 sudo make modules_install &&
575 .Java and Python application instrumentation and tracing
577 If you need to instrument and trace <<java-application,Java
578 applications>> on openSUSE, you need to build and install
579 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
580 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
581 `--enable-java-agent-all` options to the `configure` script, depending
582 on which Java logging framework you use.
584 If you need to instrument and trace <<python-application,Python
585 applications>> on openSUSE, you need to build and install
586 LTTng-UST{nbsp}{revision} from source and pass the
587 `--enable-python-agent` option to the `configure` script.
594 To install LTTng{nbsp}{revision} on Debian "stretch" (testing):
596 . Install the main LTTng{nbsp}{revision} packages:
601 sudo apt-get install lttng-modules-dkms
602 sudo apt-get install liblttng-ust-dev
603 sudo apt-get install lttng-tools
607 . **If you need to instrument and trace <<java-application,Java
608 applications>>**, install the LTTng-UST Java agent:
613 sudo apt-get install liblttng-ust-agent-java
617 . **If you need to instrument and trace <<python-application,Python
618 applications>>**, install the LTTng-UST Python agent:
623 sudo apt-get install python3-lttngust
631 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
634 . Make sure your system is
635 https://wiki.alpinelinux.org/wiki/Edge[configured for "edge"].
636 . Enable the _testing_ repository by uncommenting the corresponding
637 line in path:{/etc/apk/repositories}.
638 . Add the LTTng packages:
643 sudo apk add lttng-tools
644 sudo apk add lttng-ust-dev
648 To install LTTng-modules{nbsp}{revision} (Linux kernel tracing support)
649 on Alpine Linux "edge":
651 . Add the vanilla Linux kernel:
656 apk add linux-vanilla linux-vanilla-dev
660 . Reboot with the vanilla Linux kernel.
661 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
667 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
668 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
669 cd lttng-modules-2.8.* &&
671 sudo make modules_install &&
677 [[enterprise-distributions]]
678 === RHEL, SUSE, and other enterprise distributions
680 To install LTTng on enterprise Linux distributions, such as Red Hat
681 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
682 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
688 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2016.11:
690 . Launch the Buildroot configuration tool:
699 . In **Kernel**, check **Linux kernel**.
700 . In **Toolchain**, check **Enable WCHAR support**.
701 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
702 check **lttng-modules** and **lttng-tools**.
703 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
704 **Other**, check **lttng-libust**.
708 === OpenEmbedded and Yocto
710 LTTng{nbsp}{revision} recipes are available in the
711 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
712 layer for Yocto Project{nbsp}2.2 _Morty_ under the following names:
718 With BitBake, the simplest way to include LTTng recipes in your target
719 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
722 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
727 . Select a machine and an image recipe.
728 . Click **Edit image recipe**.
729 . Under the **All recipes** tab, search for **lttng**.
730 . Check the desired LTTng recipes.
733 .Java and Python application instrumentation and tracing
735 If you need to instrument and trace <<java-application,Java
736 applications>> on openSUSE, you need to build and install
737 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
738 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
739 `--enable-java-agent-all` options to the `configure` script, depending
740 on which Java logging framework you use.
742 If you need to instrument and trace <<python-application,Python
743 applications>> on openSUSE, you need to build and install
744 LTTng-UST{nbsp}{revision} from source and pass the
745 `--enable-python-agent` option to the `configure` script.
749 [[building-from-source]]
750 === Build from source
752 To build and install LTTng{nbsp}{revision} from source:
754 . Using your distribution's package manager, or from source, install
755 the following dependencies of LTTng-tools and LTTng-UST:
758 * https://sourceforge.net/projects/libuuid/[libuuid]
759 * http://directory.fsf.org/wiki/Popt[popt]
760 * http://liburcu.org/[Userspace RCU]
761 * http://www.xmlsoft.org/[libxml2]
764 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
770 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
771 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
772 cd lttng-modules-2.8.* &&
774 sudo make modules_install &&
779 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
785 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
786 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
787 cd lttng-ust-2.8.* &&
797 .Java and Python application tracing
799 If you need to instrument and trace <<java-application,Java
800 applications>>, pass the `--enable-java-agent-jul`,
801 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
802 `configure` script, depending on which Java logging framework you use.
804 If you need to instrument and trace <<python-application,Python
805 applications>>, pass the `--enable-python-agent` option to the
806 `configure` script. You can set the `PYTHON` environment variable to the
807 path to the Python interpreter for which to install the LTTng-UST Python
815 By default, LTTng-UST libraries are installed to
816 dir:{/usr/local/lib}, which is the de facto directory in which to
817 keep self-compiled and third-party libraries.
819 When <<building-tracepoint-providers-and-user-application,linking an
820 instrumented user application with `liblttng-ust`>>:
822 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
824 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
825 man:gcc(1), man:g++(1), or man:clang(1).
829 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
835 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
836 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
837 cd lttng-tools-2.8.* &&
845 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
846 previous steps automatically for a given version of LTTng and confine
847 the installed files in a specific directory. This can be useful to test
848 LTTng without installing it on your system.
854 This is a short guide to get started quickly with LTTng kernel and user
857 Before you follow this guide, make sure to <<installing-lttng,install>>
860 This tutorial walks you through the steps to:
862 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
863 . <<tracing-your-own-user-application,Trace a user application>> written
865 . <<viewing-and-analyzing-your-traces,View and analyze the
869 [[tracing-the-linux-kernel]]
870 === Trace the Linux kernel
872 The following command lines start with cmd:sudo because you need root
873 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
874 your Unix user is a member of the <<tracing-group,tracing group>>.
876 . Create a <<tracing-session,tracing session>> which writes its traces
877 to dir:{/tmp/my-kernel-trace}:
882 sudo lttng create my-kernel-session --output=/tmp/my-kernel-trace
886 . List the available kernel tracepoints and system calls:
892 lttng list --kernel --syscall
896 . Create <<event,event rules>> which match the desired instrumentation
897 point names, for example the `sched_switch` and `sched_process_fork`
898 tracepoints, and the man:open(2) and man:close(2) system calls:
903 sudo lttng enable-event --kernel sched_switch,sched_process_fork
904 sudo lttng enable-event --kernel --syscall open,close
908 You can also create an event rule which matches _all_ the Linux kernel
909 tracepoints (this will generate a lot of data when tracing):
914 sudo lttng enable-event --kernel --all
918 . <<basic-tracing-session-control,Start tracing>>:
927 . Do some operation on your system for a few seconds. For example,
928 load a website, or list the files of a directory.
929 . <<basic-tracing-session-control,Stop tracing>> and destroy the
940 The man:lttng-destroy(1) command does not destroy the trace data; it
941 only destroys the state of the tracing session.
943 . For the sake of this example, make the recorded trace accessible to
949 sudo chown -R $(whoami) /tmp/my-kernel-trace
953 See <<viewing-and-analyzing-your-traces,View and analyze the
954 recorded events>> to view the recorded events.
957 [[tracing-your-own-user-application]]
958 === Trace a user application
960 This section steps you through a simple example to trace a
961 _Hello world_ program written in C.
963 To create the traceable user application:
965 . Create the tracepoint provider header file, which defines the
966 tracepoints and the events they can generate:
972 #undef TRACEPOINT_PROVIDER
973 #define TRACEPOINT_PROVIDER hello_world
975 #undef TRACEPOINT_INCLUDE
976 #define TRACEPOINT_INCLUDE "./hello-tp.h"
978 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
981 #include <lttng/tracepoint.h>
991 ctf_string(my_string_field, my_string_arg)
992 ctf_integer(int, my_integer_field, my_integer_arg)
996 #endif /* _HELLO_TP_H */
998 #include <lttng/tracepoint-event.h>
1002 . Create the tracepoint provider package source file:
1008 #define TRACEPOINT_CREATE_PROBES
1009 #define TRACEPOINT_DEFINE
1011 #include "hello-tp.h"
1015 . Build the tracepoint provider package:
1020 gcc -c -I. hello-tp.c
1024 . Create the _Hello World_ application source file:
1031 #include "hello-tp.h"
1033 int main(int argc, char *argv[])
1037 puts("Hello, World!\nPress Enter to continue...");
1040 * The following getchar() call is only placed here for the purpose
1041 * of this demonstration, to pause the application in order for
1042 * you to have time to list its tracepoints. It is not
1048 * A tracepoint() call.
1050 * Arguments, as defined in hello-tp.h:
1052 * 1. Tracepoint provider name (required)
1053 * 2. Tracepoint name (required)
1054 * 3. my_integer_arg (first user-defined argument)
1055 * 4. my_string_arg (second user-defined argument)
1057 * Notice the tracepoint provider and tracepoint names are
1058 * NOT strings: they are in fact parts of variables that the
1059 * macros in hello-tp.h create.
1061 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
1063 for (x = 0; x < argc; ++x) {
1064 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
1067 puts("Quitting now!");
1068 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
1075 . Build the application:
1084 . Link the application with the tracepoint provider package,
1085 `liblttng-ust`, and `libdl`:
1090 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1094 Here's the whole build process:
1097 .User space tracing tutorial's build steps.
1098 image::ust-flow.png[]
1100 To trace the user application:
1102 . Run the application with a few arguments:
1107 ./hello world and beyond
1116 Press Enter to continue...
1120 . Start an LTTng <<lttng-sessiond,session daemon>>:
1125 lttng-sessiond --daemonize
1129 Note that a session daemon might already be running, for example as
1130 a service that the distribution's service manager started.
1132 . List the available user space tracepoints:
1137 lttng list --userspace
1141 You see the `hello_world:my_first_tracepoint` tracepoint listed
1142 under the `./hello` process.
1144 . Create a <<tracing-session,tracing session>>:
1149 lttng create my-user-space-session
1153 . Create an <<event,event rule>> which matches the
1154 `hello_world:my_first_tracepoint` event name:
1159 lttng enable-event --userspace hello_world:my_first_tracepoint
1163 . <<basic-tracing-session-control,Start tracing>>:
1172 . Go back to the running `hello` application and press Enter. The
1173 program executes all `tracepoint()` instrumentation points and exits.
1174 . <<basic-tracing-session-control,Stop tracing>> and destroy the
1185 The man:lttng-destroy(1) command does not destroy the trace data; it
1186 only destroys the state of the tracing session.
1188 By default, LTTng saves the traces in
1189 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1190 where +__name__+ is the tracing session name. The
1191 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1193 See <<viewing-and-analyzing-your-traces,View and analyze the
1194 recorded events>> to view the recorded events.
1197 [[viewing-and-analyzing-your-traces]]
1198 === View and analyze the recorded events
1200 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1201 kernel>> and <<tracing-your-own-user-application,Trace a user
1202 application>> tutorials, you can inspect the recorded events.
1204 Many tools are available to read LTTng traces:
1206 * **cmd:babeltrace** is a command-line utility which converts trace
1207 formats; it supports the format that LTTng produces, CTF, as well as a
1208 basic text output which can be ++grep++ed. The cmd:babeltrace command
1209 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1210 * Babeltrace also includes
1211 **https://www.python.org/[Python] bindings** so
1212 that you can easily open and read an LTTng trace with your own script,
1213 benefiting from the power of Python.
1214 * http://tracecompass.org/[**Trace Compass**]
1215 is a graphical user interface for viewing and analyzing any type of
1216 logs or traces, including LTTng's.
1217 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1218 project which includes many high-level analyses of LTTng kernel
1219 traces, like scheduling statistics, interrupt frequency distribution,
1220 top CPU usage, and more.
1222 NOTE: This section assumes that the traces recorded during the previous
1223 tutorials were saved to their default location, in the
1224 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1225 environment variable defaults to `$HOME` if not set.
1228 [[viewing-and-analyzing-your-traces-bt]]
1229 ==== Use the cmd:babeltrace command-line tool
1231 The simplest way to list all the recorded events of a trace is to pass
1232 its path to cmd:babeltrace with no options:
1236 babeltrace ~/lttng-traces/my-user-space-session*
1239 cmd:babeltrace finds all traces recursively within the given path and
1240 prints all their events, merging them in chronological order.
1242 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1247 babeltrace /tmp/my-kernel-trace | grep _switch
1250 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1251 count the recorded events:
1255 babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1259 [[viewing-and-analyzing-your-traces-bt-python]]
1260 ==== Use the Babeltrace Python bindings
1262 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1263 is useful to isolate events by simple matching using man:grep(1) and
1264 similar utilities. However, more elaborate filters, such as keeping only
1265 event records with a field value falling within a specific range, are
1266 not trivial to write using a shell. Moreover, reductions and even the
1267 most basic computations involving multiple event records are virtually
1268 impossible to implement.
1270 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1271 to read the event records of an LTTng trace sequentially and compute the
1272 desired information.
1274 The following script accepts an LTTng Linux kernel trace path as its
1275 first argument and prints the short names of the top 5 running processes
1276 on CPU 0 during the whole trace:
1281 from collections import Counter
1287 if len(sys.argv) != 2:
1288 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1289 print(msg, file=sys.stderr)
1292 # A trace collection contains one or more traces
1293 col = babeltrace.TraceCollection()
1295 # Add the trace provided by the user (LTTng traces always have
1297 if col.add_trace(sys.argv[1], 'ctf') is None:
1298 raise RuntimeError('Cannot add trace')
1300 # This counter dict contains execution times:
1302 # task command name -> total execution time (ns)
1303 exec_times = Counter()
1305 # This contains the last `sched_switch` timestamp
1309 for event in col.events:
1310 # Keep only `sched_switch` events
1311 if event.name != 'sched_switch':
1314 # Keep only events which happened on CPU 0
1315 if event['cpu_id'] != 0:
1319 cur_ts = event.timestamp
1325 # Previous task command (short) name
1326 prev_comm = event['prev_comm']
1328 # Initialize entry in our dict if not yet done
1329 if prev_comm not in exec_times:
1330 exec_times[prev_comm] = 0
1332 # Compute previous command execution time
1333 diff = cur_ts - last_ts
1335 # Update execution time of this command
1336 exec_times[prev_comm] += diff
1338 # Update last timestamp
1342 for name, ns in exec_times.most_common(5):
1344 print('{:20}{} s'.format(name, s))
1349 if __name__ == '__main__':
1350 sys.exit(0 if top5proc() else 1)
1357 python3 top5proc.py /tmp/my-kernel-trace/kernel
1363 swapper/0 48.607245889 s
1364 chromium 7.192738188 s
1365 pavucontrol 0.709894415 s
1366 Compositor 0.660867933 s
1367 Xorg.bin 0.616753786 s
1370 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1371 weren't using the CPU that much when tracing, its first position in the
1376 == [[understanding-lttng]]Core concepts
1378 From a user's perspective, the LTTng system is built on a few concepts,
1379 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1380 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1381 Understanding how those objects relate to eachother is key in mastering
1384 The core concepts are:
1386 * <<tracing-session,Tracing session>>
1387 * <<domain,Tracing domain>>
1388 * <<channel,Channel and ring buffer>>
1389 * <<"event","Instrumentation point, event rule, event, and event record">>
1395 A _tracing session_ is a stateful dialogue between you and
1396 a <<lttng-sessiond,session daemon>>. You can
1397 <<creating-destroying-tracing-sessions,create a new tracing
1398 session>> with the `lttng create` command.
1400 Anything that you do when you control LTTng tracers happens within a
1401 tracing session. In particular, a tracing session:
1404 * Has its own set of trace files.
1405 * Has its own state of activity (started or stopped).
1406 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1408 * Has its own <<channel,channels>> which have their own
1409 <<event,event rules>>.
1412 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1413 image::concepts.png[]
1415 Those attributes and objects are completely isolated between different
1418 A tracing session is analogous to a cash machine session:
1419 the operations you do on the banking system through the cash machine do
1420 not alter the data of other users of the same system. In the case of
1421 the cash machine, a session lasts as long as your bank card is inside.
1422 In the case of LTTng, a tracing session lasts from the `lttng create`
1423 command to the `lttng destroy` command.
1426 .Each Unix user has its own set of tracing sessions.
1427 image::many-sessions.png[]
1430 [[tracing-session-mode]]
1431 ==== Tracing session mode
1433 LTTng can send the generated trace data to different locations. The
1434 _tracing session mode_ dictates where to send it. The following modes
1435 are available in LTTng{nbsp}{revision}:
1438 LTTng writes the traces to the file system of the machine being traced
1441 Network streaming mode::
1442 LTTng sends the traces over the network to a
1443 <<lttng-relayd,relay daemon>> running on a remote system.
1446 LTTng does not write the traces by default. Instead, you can request
1447 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1448 current tracing buffers, and to write it to the target's file system
1449 or to send it over the network to a <<lttng-relayd,relay daemon>>
1450 running on a remote system.
1453 This mode is similar to the network streaming mode, but a live
1454 trace viewer can connect to the distant relay daemon to
1455 <<lttng-live,view event records as LTTng generates them>> by
1462 A _tracing domain_ is a namespace for event sources. A tracing domain
1463 has its own properties and features.
1465 There are currently five available tracing domains:
1469 * `java.util.logging` (JUL)
1473 You must specify a tracing domain when using some commands to avoid
1474 ambiguity. For example, since all the domains support named tracepoints
1475 as event sources (instrumentation points that you manually insert in the
1476 source code), you need to specify a tracing domain when
1477 <<enabling-disabling-events,creating an event rule>> because all the
1478 tracing domains could have tracepoints with the same names.
1480 Some features are reserved to specific tracing domains. Dynamic function
1481 entry and return instrumentation points, for example, are currently only
1482 supported in the Linux kernel tracing domain, but support for other
1483 tracing domains could be added in the future.
1485 You can create <<channel,channels>> in the Linux kernel and user space
1486 tracing domains. The other tracing domains have a single default
1491 === Channel and ring buffer
1493 A _channel_ is an object which is responsible for a set of ring buffers.
1494 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1495 tracer emits an event, it can record it to one or more
1496 sub-buffers. The attributes of a channel determine what to do when
1497 there's no space left for a new event record because all sub-buffers
1498 are full, where to send a full sub-buffer, and other behaviours.
1500 A channel is always associated to a <<domain,tracing domain>>. The
1501 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1502 a default channel which you cannot configure.
1504 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1505 an event, it records it to the sub-buffers of all
1506 the enabled channels with a satisfied event rule, as long as those
1507 channels are part of active <<tracing-session,tracing sessions>>.
1510 [[channel-buffering-schemes]]
1511 ==== Per-user vs. per-process buffering schemes
1513 A channel has at least one ring buffer _per CPU_. LTTng always
1514 records an event to the ring buffer associated to the CPU on which it
1517 Two _buffering schemes_ are available when you
1518 <<enabling-disabling-channels,create a channel>> in the
1519 user space <<domain,tracing domain>>:
1521 Per-user buffering::
1522 Allocate one set of ring buffers--one per CPU--shared by all the
1523 instrumented processes of each Unix user.
1527 .Per-user buffering scheme.
1528 image::per-user-buffering.png[]
1531 Per-process buffering::
1532 Allocate one set of ring buffers--one per CPU--for each
1533 instrumented process.
1537 .Per-process buffering scheme.
1538 image::per-process-buffering.png[]
1541 The per-process buffering scheme tends to consume more memory than the
1542 per-user option because systems generally have more instrumented
1543 processes than Unix users running instrumented processes. However, the
1544 per-process buffering scheme ensures that one process having a high
1545 event throughput won't fill all the shared sub-buffers of the same
1548 The Linux kernel tracing domain has only one available buffering scheme
1549 which is to allocate a single set of ring buffers for the whole system.
1550 This scheme is similar to the per-user option, but with a single, global
1551 user "running" the kernel.
1554 [[channel-overwrite-mode-vs-discard-mode]]
1555 ==== Overwrite vs. discard event loss modes
1557 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1558 arc in the following animation) of a specific channel's ring buffer.
1559 When there's no space left in a sub-buffer, the tracer marks it as
1560 consumable (red) and another, empty sub-buffer starts receiving the
1561 following event records. A <<lttng-consumerd,consumer daemon>>
1562 eventually consumes the marked sub-buffer (returns to white).
1565 [role="docsvg-channel-subbuf-anim"]
1570 In an ideal world, sub-buffers are consumed faster than they are filled,
1571 as is the case in the previous animation. In the real world,
1572 however, all sub-buffers can be full at some point, leaving no space to
1573 record the following events.
1575 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1576 available, it is acceptable to lose event records when the alternative
1577 would be to cause substantial delays in the instrumented application's
1578 execution. LTTng privileges performance over integrity; it aims at
1579 perturbing the traced system as little as possible in order to make
1580 tracing of subtle race conditions and rare interrupt cascades possible.
1582 When it comes to losing event records because no empty sub-buffer is
1583 available, the channel's _event loss mode_ determines what to do. The
1584 available event loss modes are:
1587 Drop the newest event records until a the tracer
1588 releases a sub-buffer.
1591 Clear the sub-buffer containing the oldest event records and start
1592 writing the newest event records there.
1594 This mode is sometimes called _flight recorder mode_ because it's
1596 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1597 always keep a fixed amount of the latest data.
1599 Which mechanism you should choose depends on your context: prioritize
1600 the newest or the oldest event records in the ring buffer?
1602 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1603 as soon as a there's no space left for a new event record, whereas in
1604 discard mode, the tracer only discards the event record that doesn't
1607 In discard mode, LTTng increments a count of lost event records when
1608 an event record is lost and saves this count to the trace. In
1609 overwrite mode, LTTng keeps no information when it overwrites a
1610 sub-buffer before consuming it.
1612 There are a few ways to decrease your probability of losing event
1614 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1615 how you can fine-une the sub-buffer count and size of a channel to
1616 virtually stop losing event records, though at the cost of greater
1620 [[channel-subbuf-size-vs-subbuf-count]]
1621 ==== Sub-buffer count and size
1623 When you <<enabling-disabling-channels,create a channel>>, you can
1624 set its number of sub-buffers and their size.
1626 Note that there is noticeable CPU overhead introduced when
1627 switching sub-buffers (marking a full one as consumable and switching
1628 to an empty one for the following events to be recorded). Knowing this,
1629 the following list presents a few practical situations along with how
1630 to configure the sub-buffer count and size for them:
1632 * **High event throughput**: In general, prefer bigger sub-buffers to
1633 lower the risk of losing event records.
1635 Having bigger sub-buffers also ensures a lower sub-buffer switching
1638 The number of sub-buffers is only meaningful if you create the channel
1639 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1640 other sub-buffers are left unaltered.
1642 * **Low event throughput**: In general, prefer smaller sub-buffers
1643 since the risk of losing event records is low.
1645 Because events occur less frequently, the sub-buffer switching frequency
1646 should remain low and thus the tracer's overhead should not be a
1649 * **Low memory system**: If your target system has a low memory
1650 limit, prefer fewer first, then smaller sub-buffers.
1652 Even if the system is limited in memory, you want to keep the
1653 sub-buffers as big as possible to avoid a high sub-buffer switching
1656 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1657 which means event data is very compact. For example, the average
1658 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1659 sub-buffer size of 1{nbsp}MiB is considered big.
1661 The previous situations highlight the major trade-off between a few big
1662 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1663 frequency vs. how much data is lost in overwrite mode. Assuming a
1664 constant event throughput and using the overwrite mode, the two
1665 following configurations have the same ring buffer total size:
1668 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1673 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1674 switching frequency, but if a sub-buffer overwrite happens, half of
1675 the event records so far (4{nbsp}MiB) are definitely lost.
1676 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1677 overhead as the previous configuration, but if a sub-buffer
1678 overwrite happens, only the eighth of event records so far are
1681 In discard mode, the sub-buffers count parameter is pointless: use two
1682 sub-buffers and set their size according to the requirements of your
1686 [[channel-switch-timer]]
1687 ==== Switch timer period
1689 The _switch timer period_ is an important configurable attribute of
1690 a channel to ensure periodic sub-buffer flushing.
1692 When the _switch timer_ expires, a sub-buffer switch happens. You can
1693 set the switch timer period attribute when you
1694 <<enabling-disabling-channels,create a channel>> to ensure that event
1695 data is consumed and committed to trace files or to a distant relay
1696 daemon periodically in case of a low event throughput.
1699 [role="docsvg-channel-switch-timer"]
1704 This attribute is also convenient when you use big sub-buffers to cope
1705 with a sporadic high event throughput, even if the throughput is
1709 [[channel-read-timer]]
1710 ==== Read timer period
1712 By default, the LTTng tracers use a notification mechanism to signal a
1713 full sub-buffer so that a consumer daemon can consume it. When such
1714 notifications must be avoided, for example in real-time applications,
1715 you can use the channel's _read timer_ instead. When the read timer
1716 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1717 consumable sub-buffers.
1720 [[tracefile-rotation]]
1721 ==== Trace file count and size
1723 By default, trace files can grow as large as needed. You can set the
1724 maximum size of each trace file that a channel writes when you
1725 <<enabling-disabling-channels,create a channel>>. When the size of
1726 a trace file reaches the channel's fixed maximum size, LTTng creates
1727 another file to contain the next event records. LTTng appends a file
1728 count to each trace file name in this case.
1730 If you set the trace file size attribute when you create a channel, the
1731 maximum number of trace files that LTTng creates is _unlimited_ by
1732 default. To limit them, you can also set a maximum number of trace
1733 files. When the number of trace files reaches the channel's fixed
1734 maximum count, the oldest trace file is overwritten. This mechanism is
1735 called _trace file rotation_.
1739 === Instrumentation point, event rule, event, and event record
1741 An _event rule_ is a set of conditions which must be **all** satisfied
1742 for LTTng to record an occuring event.
1744 You set the conditions when you <<enabling-disabling-events,create
1747 You always attach an event rule to <<channel,channel>> when you create
1750 When an event passes the conditions of an event rule, LTTng records it
1751 in one of the attached channel's sub-buffers.
1753 The available conditions, as of LTTng{nbsp}{revision}, are:
1755 * The event rule _is enabled_.
1756 * The instrumentation point's type _is{nbsp}T_.
1757 * The instrumentation point's name (sometimes called _event name_)
1758 _matches{nbsp}N_, but _is not{nbsp}E_.
1759 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1760 _is exactly{nbsp}L_.
1761 * The fields of the event's payload _satisfy_ a filter
1762 expression{nbsp}__F__.
1764 As you can see, all the conditions but the dynamic filter are related to
1765 the event rule's status or to the instrumentation point, not to the
1766 occurring events. This is why, without a filter, checking if an event
1767 passes an event rule is not a dynamic task: when you create or modify an
1768 event rule, all the tracers of its tracing domain enable or disable the
1769 instrumentation points themselves once. This is possible because the
1770 attributes of an instrumentation point (type, name, and log level) are
1771 defined statically. In other words, without a dynamic filter, the tracer
1772 _does not evaluate_ the arguments of an instrumentation point unless it
1773 matches an enabled event rule.
1775 Note that, for LTTng to record an event, the <<channel,channel>> to
1776 which a matching event rule is attached must also be enabled, and the
1777 tracing session owning this channel must be active.
1780 .Logical path from an instrumentation point to an event record.
1781 image::event-rule.png[]
1783 .Event, event record, or event rule?
1785 With so many similar terms, it's easy to get confused.
1787 An **event** is the consequence of the execution of an _instrumentation
1788 point_, like a tracepoint that you manually place in some source code,
1789 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1790 time. Different actions can be taken upon the occurance of an event,
1791 like record the event's payload to a buffer.
1793 An **event record** is the representation of an event in a sub-buffer. A
1794 tracer is responsible for capturing the payload of an event, current
1795 context variables, the event's ID, and the event's timestamp. LTTng
1796 can append this sub-buffer to a trace file.
1798 An **event rule** is a set of conditions which must all be satisfied for
1799 LTTng to record an occuring event. Events still occur without
1800 satisfying event rules, but LTTng does not record them.
1805 == Components of noch:{LTTng}
1807 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1808 to call LTTng a simple _tool_ since it is composed of multiple
1809 interacting components. This section describes those components,
1810 explains their respective roles, and shows how they connect together to
1811 form the LTTng ecosystem.
1813 The following diagram shows how the most important components of LTTng
1814 interact with user applications, the Linux kernel, and you:
1817 .Control and trace data paths between LTTng components.
1818 image::plumbing.png[]
1820 The LTTng project incorporates:
1822 * **LTTng-tools**: Libraries and command-line interface to
1823 control tracing sessions.
1824 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1825 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1826 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1827 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1828 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1829 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1831 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1832 headers to instrument and trace any native user application.
1833 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1834 *** `liblttng-ust-libc-wrapper`
1835 *** `liblttng-ust-pthread-wrapper`
1836 *** `liblttng-ust-cyg-profile`
1837 *** `liblttng-ust-cyg-profile-fast`
1838 *** `liblttng-ust-dl`
1839 ** User space tracepoint provider source files generator command-line
1840 tool (man:lttng-gen-tp(1)).
1841 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1842 Java applications using `java.util.logging` or
1843 Apache log4j 1.2 logging.
1844 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1845 Python applications using the standard `logging` package.
1846 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1848 ** LTTng kernel tracer module.
1849 ** Tracing ring buffer kernel modules.
1850 ** Probe kernel modules.
1851 ** LTTng logger kernel module.
1855 === Tracing control command-line interface
1858 .The tracing control command-line interface.
1859 image::plumbing-lttng-cli.png[]
1861 The _man:lttng(1) command-line tool_ is the standard user interface to
1862 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1863 is part of LTTng-tools.
1865 The cmd:lttng tool is linked with
1866 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1867 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1869 The cmd:lttng tool has a Git-like interface:
1873 lttng <general options> <command> <command options>
1876 The <<controlling-tracing,Tracing control>> section explores the
1877 available features of LTTng using the cmd:lttng tool.
1880 [[liblttng-ctl-lttng]]
1881 === Tracing control library
1884 .The tracing control library.
1885 image::plumbing-liblttng-ctl.png[]
1887 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1888 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1889 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1891 The <<lttng-cli,cmd:lttng command-line tool>>
1892 is linked with `liblttng-ctl`.
1894 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1899 #include <lttng/lttng.h>
1902 Some objects are referenced by name (C string), such as tracing
1903 sessions, but most of them require to create a handle first using
1904 `lttng_create_handle()`.
1906 The best available developer documentation for `liblttng-ctl` is, as of
1907 LTTng{nbsp}{revision}, its installed header files. Every function and
1908 structure is thoroughly documented.
1912 === User space tracing library
1915 .The user space tracing library.
1916 image::plumbing-liblttng-ust.png[]
1918 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1919 is the LTTng user space tracer. It receives commands from a
1920 <<lttng-sessiond,session daemon>>, for example to
1921 enable and disable specific instrumentation points, and writes event
1922 records to ring buffers shared with a
1923 <<lttng-consumerd,consumer daemon>>.
1924 `liblttng-ust` is part of LTTng-UST.
1926 Public C header files are installed beside `liblttng-ust` to
1927 instrument any <<c-application,C or $$C++$$ application>>.
1929 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1930 packages, use their own library providing tracepoints which is
1931 linked with `liblttng-ust`.
1933 An application or library does not have to initialize `liblttng-ust`
1934 manually: its constructor does the necessary tasks to properly register
1935 to a session daemon. The initialization phase also enables the
1936 instrumentation points matching the <<event,event rules>> that you
1940 [[lttng-ust-agents]]
1941 === User space tracing agents
1944 .The user space tracing agents.
1945 image::plumbing-lttng-ust-agents.png[]
1947 The _LTTng-UST Java and Python agents_ are regular Java and Python
1948 packages which add LTTng tracing capabilities to the
1949 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1951 In the case of Java, the
1952 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1953 core logging facilities] and
1954 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1955 Note that Apache Log4{nbsp}2 is not supported.
1957 In the case of Python, the standard
1958 https://docs.python.org/3/library/logging.html[`logging`] package
1959 is supported. Both Python 2 and Python 3 modules can import the
1960 LTTng-UST Python agent package.
1962 The applications using the LTTng-UST agents are in the
1963 `java.util.logging` (JUL),
1964 log4j, and Python <<domain,tracing domains>>.
1966 Both agents use the same mechanism to trace the log statements. When an
1967 agent is initialized, it creates a log handler that attaches to the root
1968 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1969 When the application executes a log statement, it is passed to the
1970 agent's log handler by the root logger. The agent's log handler calls a
1971 native function in a tracepoint provider package shared library linked
1972 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1973 other fields, like its logger name and its log level. This native
1974 function contains a user space instrumentation point, hence tracing the
1977 The log level condition of an
1978 <<event,event rule>> is considered when tracing
1979 a Java or a Python application, and it's compatible with the standard
1980 JUL, log4j, and Python log levels.
1984 === LTTng kernel modules
1987 .The LTTng kernel modules.
1988 image::plumbing-lttng-modules.png[]
1990 The _LTTng kernel modules_ are a set of Linux kernel modules
1991 which implement the kernel tracer of the LTTng project. The LTTng
1992 kernel modules are part of LTTng-modules.
1994 The LTTng kernel modules include:
1996 * A set of _probe_ modules.
1998 Each module attaches to a specific subsystem
1999 of the Linux kernel using its tracepoint instrument points. There are
2000 also modules to attach to the entry and return points of the Linux
2001 system call functions.
2003 * _Ring buffer_ modules.
2005 A ring buffer implementation is provided as kernel modules. The LTTng
2006 kernel tracer writes to the ring buffer; a
2007 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
2009 * The _LTTng kernel tracer_ module.
2010 * The _LTTng logger_ module.
2012 The LTTng logger module implements the special path:{/proc/lttng-logger}
2013 file so that any executable can generate LTTng events by opening and
2014 writing to this file.
2016 See <<proc-lttng-logger-abi,LTTng logger>>.
2018 Generally, you do not have to load the LTTng kernel modules manually
2019 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
2020 daemon>> loads the necessary modules when starting. If you have extra
2021 probe modules, you can specify to load them to the session daemon on
2024 The LTTng kernel modules are installed in
2025 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
2026 the kernel release (see `uname --kernel-release`).
2033 .The session daemon.
2034 image::plumbing-sessiond.png[]
2036 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
2037 managing tracing sessions and for controlling the various components of
2038 LTTng. The session daemon is part of LTTng-tools.
2040 The session daemon sends control requests to and receives control
2043 * The <<lttng-ust,user space tracing library>>.
2045 Any instance of the user space tracing library first registers to
2046 a session daemon. Then, the session daemon can send requests to
2047 this instance, such as:
2050 ** Get the list of tracepoints.
2051 ** Share an <<event,event rule>> so that the user space tracing library
2052 can enable or disable tracepoints. Amongst the possible conditions
2053 of an event rule is a filter expression which `liblttng-ust` evalutes
2054 when an event occurs.
2055 ** Share <<channel,channel>> attributes and ring buffer locations.
2058 The session daemon and the user space tracing library use a Unix
2059 domain socket for their communication.
2061 * The <<lttng-ust-agents,user space tracing agents>>.
2063 Any instance of a user space tracing agent first registers to
2064 a session daemon. Then, the session daemon can send requests to
2065 this instance, such as:
2068 ** Get the list of loggers.
2069 ** Enable or disable a specific logger.
2072 The session daemon and the user space tracing agent use a TCP connection
2073 for their communication.
2075 * The <<lttng-modules,LTTng kernel tracer>>.
2076 * The <<lttng-consumerd,consumer daemon>>.
2078 The session daemon sends requests to the consumer daemon to instruct
2079 it where to send the trace data streams, amongst other information.
2081 * The <<lttng-relayd,relay daemon>>.
2083 The session daemon receives commands from the
2084 <<liblttng-ctl-lttng,tracing control library>>.
2086 The root session daemon loads the appropriate
2087 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2088 a <<lttng-consumerd,consumer daemon>> as soon as you create
2089 an <<event,event rule>>.
2091 The session daemon does not send and receive trace data: this is the
2092 role of the <<lttng-consumerd,consumer daemon>> and
2093 <<lttng-relayd,relay daemon>>. It does, however, generate the
2094 http://diamon.org/ctf/[CTF] metadata stream.
2096 Each Unix user can have its own session daemon instance. The
2097 tracing sessions managed by different session daemons are completely
2100 The root user's session daemon is the only one which is
2101 allowed to control the LTTng kernel tracer, and its spawned consumer
2102 daemon is the only one which is allowed to consume trace data from the
2103 LTTng kernel tracer. Note, however, that any Unix user which is a member
2104 of the <<tracing-group,tracing group>> is allowed
2105 to create <<channel,channels>> in the
2106 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2109 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2110 session daemon when using its `create` command if none is currently
2111 running. You can also start the session daemon manually.
2118 .The consumer daemon.
2119 image::plumbing-consumerd.png[]
2121 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
2122 ring buffers with user applications or with the LTTng kernel modules to
2123 collect trace data and send it to some location (on disk or to a
2124 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2125 is part of LTTng-tools.
2127 You do not start a consumer daemon manually: a consumer daemon is always
2128 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2129 <<event,event rule>>, that is, before you start tracing. When you kill
2130 its owner session daemon, the consumer daemon also exits because it is
2131 the session daemon's child process. Command-line options of
2132 man:lttng-sessiond(8) target the consumer daemon process.
2134 There are up to two running consumer daemons per Unix user, whereas only
2135 one session daemon can run per user. This is because each process can be
2136 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2137 and 64-bit processes, it is more efficient to have separate
2138 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2139 exception: it can have up to _three_ running consumer daemons: 32-bit
2140 and 64-bit instances for its user applications, and one more
2141 reserved for collecting kernel trace data.
2149 image::plumbing-relayd.png[]
2151 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2152 between remote session and consumer daemons, local trace files, and a
2153 remote live trace viewer. The relay daemon is part of LTTng-tools.
2155 The main purpose of the relay daemon is to implement a receiver of
2156 <<sending-trace-data-over-the-network,trace data over the network>>.
2157 This is useful when the target system does not have much file system
2158 space to record trace files locally.
2160 The relay daemon is also a server to which a
2161 <<lttng-live,live trace viewer>> can
2162 connect. The live trace viewer sends requests to the relay daemon to
2163 receive trace data as the target system emits events. The
2164 communication protocol is named _LTTng live_; it is used over TCP
2167 Note that you can start the relay daemon on the target system directly.
2168 This is the setup of choice when the use case is to view events as
2169 the target system emits them without the need of a remote system.
2173 == [[using-lttng]]Instrumentation
2175 There are many examples of tracing and monitoring in our everyday life:
2177 * You have access to real-time and historical weather reports and
2178 forecasts thanks to weather stations installed around the country.
2179 * You know your heart is safe thanks to an electrocardiogram.
2180 * You make sure not to drive your car too fast and to have enough fuel
2181 to reach your destination thanks to gauges visible on your dashboard.
2183 All the previous examples have something in common: they rely on
2184 **instruments**. Without the electrodes attached to the surface of your
2185 body's skin, cardiac monitoring is futile.
2187 LTTng, as a tracer, is no different from those real life examples. If
2188 you're about to trace a software system or, in other words, record its
2189 history of execution, you better have **instrumentation points** in the
2190 subject you're tracing, that is, the actual software.
2192 Various ways were developed to instrument a piece of software for LTTng
2193 tracing. The most straightforward one is to manually place
2194 instrumentation points, called _tracepoints_, in the software's source
2195 code. It is also possible to add instrumentation points dynamically in
2196 the Linux kernel <<domain,tracing domain>>.
2198 If you're only interested in tracing the Linux kernel, your
2199 instrumentation needs are probably already covered by LTTng's built-in
2200 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2201 user application which is already instrumented for LTTng tracing.
2202 In such cases, you can skip this whole section and read the topics of
2203 the <<controlling-tracing,Tracing control>> section.
2205 Many methods are available to instrument a piece of software for LTTng
2208 * <<c-application,User space instrumentation for C and $$C++$$
2210 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2211 * <<java-application,User space Java agent>>.
2212 * <<python-application,User space Python agent>>.
2213 * <<proc-lttng-logger-abi,LTTng logger>>.
2214 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2218 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2220 The procedure to instrument a C or $$C++$$ user application with
2221 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2223 . <<tracepoint-provider,Create the source files of a tracepoint provider
2225 . <<probing-the-application-source-code,Add tracepoints to
2226 the application's source code>>.
2227 . <<building-tracepoint-providers-and-user-application,Build and link
2228 a tracepoint provider package and the user application>>.
2230 If you need quick, man:printf(3)-like instrumentation, you can skip
2231 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2234 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2235 instrument a user application with `liblttng-ust`.
2238 [[tracepoint-provider]]
2239 ==== Create the source files of a tracepoint provider package
2241 A _tracepoint provider_ is a set of compiled functions which provide
2242 **tracepoints** to an application, the type of instrumentation point
2243 supported by LTTng-UST. Those functions can emit events with
2244 user-defined fields and serialize those events as event records to one
2245 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2246 macro, which you <<probing-the-application-source-code,insert in a user
2247 application's source code>>, calls those functions.
2249 A _tracepoint provider package_ is an object file (`.o`) or a shared
2250 library (`.so`) which contains one or more tracepoint providers.
2251 Its source files are:
2253 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2254 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2256 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2257 the LTTng user space tracer, at run time.
2260 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2261 image::ust-app.png[]
2263 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2264 skip creating and using a tracepoint provider and use
2265 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2269 ===== Create a tracepoint provider header file template
2271 A _tracepoint provider header file_ contains the tracepoint
2272 definitions of a tracepoint provider.
2274 To create a tracepoint provider header file:
2276 . Start from this template:
2280 .Tracepoint provider header file template (`.h` file extension).
2282 #undef TRACEPOINT_PROVIDER
2283 #define TRACEPOINT_PROVIDER provider_name
2285 #undef TRACEPOINT_INCLUDE
2286 #define TRACEPOINT_INCLUDE "./tp.h"
2288 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2291 #include <lttng/tracepoint.h>
2294 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2295 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2300 #include <lttng/tracepoint-event.h>
2306 * `provider_name` with the name of your tracepoint provider.
2307 * `"tp.h"` with the name of your tracepoint provider header file.
2309 . Below the `#include <lttng/tracepoint.h>` line, put your
2310 <<defining-tracepoints,tracepoint definitions>>.
2312 Your tracepoint provider name must be unique amongst all the possible
2313 tracepoint provider names used on the same target system. We
2314 suggest to include the name of your project or company in the name,
2315 for example, `org_lttng_my_project_tpp`.
2317 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2318 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2319 write are the <<defining-tracepoints,tracepoint definitions>>.
2322 [[defining-tracepoints]]
2323 ===== Create a tracepoint definition
2325 A _tracepoint definition_ defines, for a given tracepoint:
2327 * Its **input arguments**. They are the macro parameters that the
2328 `tracepoint()` macro accepts for this particular tracepoint
2329 in the user application's source code.
2330 * Its **output event fields**. They are the sources of event fields
2331 that form the payload of any event that the execution of the
2332 `tracepoint()` macro emits for this particular tracepoint.
2334 You can create a tracepoint definition by using the
2335 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2337 <<tpp-header,tracepoint provider header file template>>.
2339 The syntax of the `TRACEPOINT_EVENT()` macro is:
2342 .`TRACEPOINT_EVENT()` macro syntax.
2345 /* Tracepoint provider name */
2348 /* Tracepoint name */
2351 /* Input arguments */
2356 /* Output event fields */
2365 * `provider_name` with your tracepoint provider name.
2366 * `tracepoint_name` with your tracepoint name.
2367 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2368 * `fields` with the <<tpp-def-output-fields,output event field>>
2371 This tracepoint emits events named `provider_name:tracepoint_name`.
2374 .Event name's length limitation
2376 The concatenation of the tracepoint provider name and the
2377 tracepoint name must not exceed **254 characters**. If it does, the
2378 instrumented application compiles and runs, but LTTng throws multiple
2379 warnings and you could experience serious issues.
2382 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2385 .`TP_ARGS()` macro syntax.
2394 * `type` with the C type of the argument.
2395 * `arg_name` with the argument name.
2397 You can repeat `type` and `arg_name` up to 10 times to have
2398 more than one argument.
2400 .`TP_ARGS()` usage with three arguments.
2412 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2413 tracepoint definition with no input arguments.
2415 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2416 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2417 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2418 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2421 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2422 C expression that the tracer evalutes at the `tracepoint()` macro site
2423 in the application's source code. This expression provides a field's
2424 source of data. The argument expression can include input argument names
2425 listed in the `TP_ARGS()` macro.
2427 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2428 must be unique within a given tracepoint definition.
2430 Here's a complete tracepoint definition example:
2432 .Tracepoint definition.
2434 The following tracepoint definition defines a tracepoint which takes
2435 three input arguments and has four output event fields.
2439 #include "my-custom-structure.h"
2445 const struct my_custom_structure*, my_custom_structure,
2450 ctf_string(query_field, query)
2451 ctf_float(double, ratio_field, ratio)
2452 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2453 ctf_integer(int, send_size, my_custom_structure->send_size)
2458 You can refer to this tracepoint definition with the `tracepoint()`
2459 macro in your application's source code like this:
2463 tracepoint(my_provider, my_tracepoint,
2464 my_structure, some_ratio, the_query);
2468 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2469 if they satisfy an enabled <<event,event rule>>.
2472 [[using-tracepoint-classes]]
2473 ===== Use a tracepoint class
2475 A _tracepoint class_ is a class of tracepoints which share the same
2476 output event field definitions. A _tracepoint instance_ is one
2477 instance of such a defined tracepoint class, with its own tracepoint
2480 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2481 shorthand which defines both a tracepoint class and a tracepoint
2482 instance at the same time.
2484 When you build a tracepoint provider package, the C or $$C++$$ compiler
2485 creates one serialization function for each **tracepoint class**. A
2486 serialization function is responsible for serializing the event fields
2487 of a tracepoint to a sub-buffer when tracing.
2489 For various performance reasons, when your situation requires multiple
2490 tracepoint definitions with different names, but with the same event
2491 fields, we recommend that you manually create a tracepoint class
2492 and instantiate as many tracepoint instances as needed. One positive
2493 effect of such a design, amongst other advantages, is that all
2494 tracepoint instances of the same tracepoint class reuse the same
2495 serialization function, thus reducing
2496 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2498 .Use a tracepoint class and tracepoint instances.
2500 Consider the following three tracepoint definitions:
2512 ctf_integer(int, userid, userid)
2513 ctf_integer(size_t, len, len)
2525 ctf_integer(int, userid, userid)
2526 ctf_integer(size_t, len, len)
2538 ctf_integer(int, userid, userid)
2539 ctf_integer(size_t, len, len)
2544 In this case, we create three tracepoint classes, with one implicit
2545 tracepoint instance for each of them: `get_account`, `get_settings`, and
2546 `get_transaction`. However, they all share the same event field names
2547 and types. Hence three identical, yet independent serialization
2548 functions are created when you build the tracepoint provider package.
2550 A better design choice is to define a single tracepoint class and three
2551 tracepoint instances:
2555 /* The tracepoint class */
2556 TRACEPOINT_EVENT_CLASS(
2557 /* Tracepoint provider name */
2560 /* Tracepoint class name */
2563 /* Input arguments */
2569 /* Output event fields */
2571 ctf_integer(int, userid, userid)
2572 ctf_integer(size_t, len, len)
2576 /* The tracepoint instances */
2577 TRACEPOINT_EVENT_INSTANCE(
2578 /* Tracepoint provider name */
2581 /* Tracepoint class name */
2584 /* Tracepoint name */
2587 /* Input arguments */
2593 TRACEPOINT_EVENT_INSTANCE(
2602 TRACEPOINT_EVENT_INSTANCE(
2615 [[assigning-log-levels]]
2616 ===== Assign a log level to a tracepoint definition
2618 You can assign an optional _log level_ to a
2619 <<defining-tracepoints,tracepoint definition>>.
2621 Assigning different levels of severity to tracepoint definitions can
2622 be useful: when you <<enabling-disabling-events,create an event rule>>,
2623 you can target tracepoints having a log level as severe as a specific
2626 The concept of LTTng-UST log levels is similar to the levels found
2627 in typical logging frameworks:
2629 * In a logging framework, the log level is given by the function
2630 or method name you use at the log statement site: `debug()`,
2631 `info()`, `warn()`, `error()`, and so on.
2632 * In LTTng-UST, you statically assign the log level to a tracepoint
2633 definition; any `tracepoint()` macro invocation which refers to
2634 this definition has this log level.
2636 You can assign a log level to a tracepoint definition with the
2637 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2638 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2639 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2642 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2645 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2647 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2652 * `provider_name` with the tracepoint provider name.
2653 * `tracepoint_name` with the tracepoint name.
2654 * `log_level` with the log level to assign to the tracepoint
2655 definition named `tracepoint_name` in the `provider_name`
2656 tracepoint provider.
2658 See man:lttng-ust(3) for a list of available log level names.
2660 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2664 /* Tracepoint definition */
2673 ctf_integer(int, userid, userid)
2674 ctf_integer(size_t, len, len)
2678 /* Log level assignment */
2679 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2685 ===== Create a tracepoint provider package source file
2687 A _tracepoint provider package source file_ is a C source file which
2688 includes a <<tpp-header,tracepoint provider header file>> to expand its
2689 macros into event serialization and other functions.
2691 You can always use the following tracepoint provider package source
2695 .Tracepoint provider package source file template.
2697 #define TRACEPOINT_CREATE_PROBES
2702 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2703 header file>> name. You may also include more than one tracepoint
2704 provider header file here to create a tracepoint provider package
2705 holding more than one tracepoint providers.
2708 [[probing-the-application-source-code]]
2709 ==== Add tracepoints to an application's source code
2711 Once you <<tpp-header,create a tracepoint provider header file>>, you
2712 can use the `tracepoint()` macro in your application's
2713 source code to insert the tracepoints that this header
2714 <<defining-tracepoints,defines>>.
2716 The `tracepoint()` macro takes at least two parameters: the tracepoint
2717 provider name and the tracepoint name. The corresponding tracepoint
2718 definition defines the other parameters.
2720 .`tracepoint()` usage.
2722 The following <<defining-tracepoints,tracepoint definition>> defines a
2723 tracepoint which takes two input arguments and has two output event
2727 .Tracepoint provider header file.
2729 #include "my-custom-structure.h"
2736 const char*, cmd_name
2739 ctf_string(cmd_name, cmd_name)
2740 ctf_integer(int, number_of_args, argc)
2745 You can refer to this tracepoint definition with the `tracepoint()`
2746 macro in your application's source code like this:
2749 .Application's source file.
2753 int main(int argc, char* argv[])
2755 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2761 Note how the application's source code includes
2762 the tracepoint provider header file containing the tracepoint
2763 definitions to use, path:{tp.h}.
2766 .`tracepoint()` usage with a complex tracepoint definition.
2768 Consider this complex tracepoint definition, where multiple event
2769 fields refer to the same input arguments in their argument expression
2773 .Tracepoint provider header file.
2775 /* For `struct stat` */
2776 #include <sys/types.h>
2777 #include <sys/stat.h>
2789 ctf_integer(int, my_constant_field, 23 + 17)
2790 ctf_integer(int, my_int_arg_field, my_int_arg)
2791 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2792 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2793 my_str_arg[2] + my_str_arg[3])
2794 ctf_string(my_str_arg_field, my_str_arg)
2795 ctf_integer_hex(off_t, size_field, st->st_size)
2796 ctf_float(double, size_dbl_field, (double) st->st_size)
2797 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2798 size_t, strlen(my_str_arg) / 2)
2803 You can refer to this tracepoint definition with the `tracepoint()`
2804 macro in your application's source code like this:
2807 .Application's source file.
2809 #define TRACEPOINT_DEFINE
2816 stat("/etc/fstab", &s);
2817 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2823 If you look at the event record that LTTng writes when tracing this
2824 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2825 it should look like this:
2827 .Event record fields
2829 |Field's name |Field's value
2830 |`my_constant_field` |40
2831 |`my_int_arg_field` |23
2832 |`my_int_arg_field2` |529
2834 |`my_str_arg_field` |`Hello, World!`
2835 |`size_field` |0x12d
2836 |`size_dbl_field` |301.0
2837 |`half_my_str_arg_field` |`Hello,`
2841 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2842 compute--they use the call stack, for example. To avoid this
2843 computation when the tracepoint is disabled, you can use the
2844 `tracepoint_enabled()` and `do_tracepoint()` macros.
2846 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2850 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2852 tracepoint_enabled(provider_name, tracepoint_name)
2853 do_tracepoint(provider_name, tracepoint_name, ...)
2858 * `provider_name` with the tracepoint provider name.
2859 * `tracepoint_name` with the tracepoint name.
2861 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2862 `tracepoint_name` from the provider named `provider_name` is enabled
2865 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2866 if the tracepoint is enabled. Using `tracepoint()` with
2867 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2868 the `tracepoint_enabled()` check, thus a race condition is
2869 possible in this situation:
2872 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2874 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2875 stuff = prepare_stuff();
2878 tracepoint(my_provider, my_tracepoint, stuff);
2881 If the tracepoint is enabled after the condition, then `stuff` is not
2882 prepared: the emitted event will either contain wrong data, or the whole
2883 application could crash (segmentation fault, for example).
2885 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2886 `STAP_PROBEV()` call. If you need it, you must emit
2890 [[building-tracepoint-providers-and-user-application]]
2891 ==== Build and link a tracepoint provider package and an application
2893 Once you have one or more <<tpp-header,tracepoint provider header
2894 files>> and a <<tpp-source,tracepoint provider package source file>>,
2895 you can create the tracepoint provider package by compiling its source
2896 file. From here, multiple build and run scenarios are possible. The
2897 following table shows common application and library configurations
2898 along with the required command lines to achieve them.
2900 In the following diagrams, we use the following file names:
2903 Executable application.
2906 Application's object file.
2909 Tracepoint provider package object file.
2912 Tracepoint provider package archive file.
2915 Tracepoint provider package shared object file.
2918 User library object file.
2921 User library shared object file.
2923 We use the following symbols in the diagrams of table below:
2926 .Symbols used in the build scenario diagrams.
2927 image::ust-sit-symbols.png[]
2929 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2930 variable in the following instructions.
2932 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2933 .Common tracepoint provider package scenarios.
2935 |Scenario |Instructions
2938 The instrumented application is statically linked with
2939 the tracepoint provider package object.
2941 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2944 include::../common/ust-sit-step-tp-o.txt[]
2946 To build the instrumented application:
2948 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2953 #define TRACEPOINT_DEFINE
2957 . Compile the application source file:
2966 . Build the application:
2971 gcc -o app app.o tpp.o -llttng-ust -ldl
2975 To run the instrumented application:
2977 * Start the application:
2987 The instrumented application is statically linked with the
2988 tracepoint provider package archive file.
2990 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2993 To create the tracepoint provider package archive file:
2995 . Compile the <<tpp-source,tracepoint provider package source file>>:
3004 . Create the tracepoint provider package archive file:
3013 To build the instrumented application:
3015 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3020 #define TRACEPOINT_DEFINE
3024 . Compile the application source file:
3033 . Build the application:
3038 gcc -o app app.o tpp.a -llttng-ust -ldl
3042 To run the instrumented application:
3044 * Start the application:
3054 The instrumented application is linked with the tracepoint provider
3055 package shared object.
3057 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
3060 include::../common/ust-sit-step-tp-so.txt[]
3062 To build the instrumented application:
3064 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3069 #define TRACEPOINT_DEFINE
3073 . Compile the application source file:
3082 . Build the application:
3087 gcc -o app app.o -ldl -L. -ltpp
3091 To run the instrumented application:
3093 * Start the application:
3103 The tracepoint provider package shared object is preloaded before the
3104 instrumented application starts.
3106 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3109 include::../common/ust-sit-step-tp-so.txt[]
3111 To build the instrumented application:
3113 . In path:{app.c}, before including path:{tpp.h}, add the
3119 #define TRACEPOINT_DEFINE
3120 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3124 . Compile the application source file:
3133 . Build the application:
3138 gcc -o app app.o -ldl
3142 To run the instrumented application with tracing support:
3144 * Preload the tracepoint provider package shared object and
3145 start the application:
3150 LD_PRELOAD=./libtpp.so ./app
3154 To run the instrumented application without tracing support:
3156 * Start the application:
3166 The instrumented application dynamically loads the tracepoint provider
3167 package shared object.
3169 See the <<dlclose-warning,warning about `dlclose()`>>.
3171 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3174 include::../common/ust-sit-step-tp-so.txt[]
3176 To build the instrumented application:
3178 . In path:{app.c}, before including path:{tpp.h}, add the
3184 #define TRACEPOINT_DEFINE
3185 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3189 . Compile the application source file:
3198 . Build the application:
3203 gcc -o app app.o -ldl
3207 To run the instrumented application:
3209 * Start the application:
3219 The application is linked with the instrumented user library.
3221 The instrumented user library is statically linked with the tracepoint
3222 provider package object file.
3224 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3227 include::../common/ust-sit-step-tp-o-fpic.txt[]
3229 To build the instrumented user library:
3231 . In path:{emon.c}, before including path:{tpp.h}, add the
3237 #define TRACEPOINT_DEFINE
3241 . Compile the user library source file:
3246 gcc -I. -fpic -c emon.c
3250 . Build the user library shared object:
3255 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3259 To build the application:
3261 . Compile the application source file:
3270 . Build the application:
3275 gcc -o app app.o -L. -lemon
3279 To run the application:
3281 * Start the application:
3291 The application is linked with the instrumented user library.
3293 The instrumented user library is linked with the tracepoint provider
3294 package shared object.
3296 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3299 include::../common/ust-sit-step-tp-so.txt[]
3301 To build the instrumented user library:
3303 . In path:{emon.c}, before including path:{tpp.h}, add the
3309 #define TRACEPOINT_DEFINE
3313 . Compile the user library source file:
3318 gcc -I. -fpic -c emon.c
3322 . Build the user library shared object:
3327 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3331 To build the application:
3333 . Compile the application source file:
3342 . Build the application:
3347 gcc -o app app.o -L. -lemon
3351 To run the application:
3353 * Start the application:
3363 The tracepoint provider package shared object is preloaded before the
3366 The application is linked with the instrumented user library.
3368 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3371 include::../common/ust-sit-step-tp-so.txt[]
3373 To build the instrumented user library:
3375 . In path:{emon.c}, before including path:{tpp.h}, add the
3381 #define TRACEPOINT_DEFINE
3382 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3386 . Compile the user library source file:
3391 gcc -I. -fpic -c emon.c
3395 . Build the user library shared object:
3400 gcc -shared -o libemon.so emon.o -ldl
3404 To build the application:
3406 . Compile the application source file:
3415 . Build the application:
3420 gcc -o app app.o -L. -lemon
3424 To run the application with tracing support:
3426 * Preload the tracepoint provider package shared object and
3427 start the application:
3432 LD_PRELOAD=./libtpp.so ./app
3436 To run the application without tracing support:
3438 * Start the application:
3448 The application is linked with the instrumented user library.
3450 The instrumented user library dynamically loads the tracepoint provider
3451 package shared object.
3453 See the <<dlclose-warning,warning about `dlclose()`>>.
3455 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3458 include::../common/ust-sit-step-tp-so.txt[]
3460 To build the instrumented user library:
3462 . In path:{emon.c}, before including path:{tpp.h}, add the
3468 #define TRACEPOINT_DEFINE
3469 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3473 . Compile the user library source file:
3478 gcc -I. -fpic -c emon.c
3482 . Build the user library shared object:
3487 gcc -shared -o libemon.so emon.o -ldl
3491 To build the application:
3493 . Compile the application source file:
3502 . Build the application:
3507 gcc -o app app.o -L. -lemon
3511 To run the application:
3513 * Start the application:
3523 The application dynamically loads the instrumented user library.
3525 The instrumented user library is linked with the tracepoint provider
3526 package shared object.
3528 See the <<dlclose-warning,warning about `dlclose()`>>.
3530 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3533 include::../common/ust-sit-step-tp-so.txt[]
3535 To build the instrumented user library:
3537 . In path:{emon.c}, before including path:{tpp.h}, add the
3543 #define TRACEPOINT_DEFINE
3547 . Compile the user library source file:
3552 gcc -I. -fpic -c emon.c
3556 . Build the user library shared object:
3561 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3565 To build the application:
3567 . Compile the application source file:
3576 . Build the application:
3581 gcc -o app app.o -ldl -L. -lemon
3585 To run the application:
3587 * Start the application:
3597 The application dynamically loads the instrumented user library.
3599 The instrumented user library dynamically loads the tracepoint provider
3600 package shared object.
3602 See the <<dlclose-warning,warning about `dlclose()`>>.
3604 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3607 include::../common/ust-sit-step-tp-so.txt[]
3609 To build the instrumented user library:
3611 . In path:{emon.c}, before including path:{tpp.h}, add the
3617 #define TRACEPOINT_DEFINE
3618 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3622 . Compile the user library source file:
3627 gcc -I. -fpic -c emon.c
3631 . Build the user library shared object:
3636 gcc -shared -o libemon.so emon.o -ldl
3640 To build the application:
3642 . Compile the application source file:
3651 . Build the application:
3656 gcc -o app app.o -ldl -L. -lemon
3660 To run the application:
3662 * Start the application:
3672 The tracepoint provider package shared object is preloaded before the
3675 The application dynamically loads the instrumented user library.
3677 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3680 include::../common/ust-sit-step-tp-so.txt[]
3682 To build the instrumented user library:
3684 . In path:{emon.c}, before including path:{tpp.h}, add the
3690 #define TRACEPOINT_DEFINE
3691 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3695 . Compile the user library source file:
3700 gcc -I. -fpic -c emon.c
3704 . Build the user library shared object:
3709 gcc -shared -o libemon.so emon.o -ldl
3713 To build the application:
3715 . Compile the application source file:
3724 . Build the application:
3729 gcc -o app app.o -L. -lemon
3733 To run the application with tracing support:
3735 * Preload the tracepoint provider package shared object and
3736 start the application:
3741 LD_PRELOAD=./libtpp.so ./app
3745 To run the application without tracing support:
3747 * Start the application:
3757 The application is statically linked with the tracepoint provider
3758 package object file.
3760 The application is linked with the instrumented user library.
3762 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3765 include::../common/ust-sit-step-tp-o.txt[]
3767 To build the instrumented user library:
3769 . In path:{emon.c}, before including path:{tpp.h}, add the
3775 #define TRACEPOINT_DEFINE
3779 . Compile the user library source file:
3784 gcc -I. -fpic -c emon.c
3788 . Build the user library shared object:
3793 gcc -shared -o libemon.so emon.o
3797 To build the application:
3799 . Compile the application source file:
3808 . Build the application:
3813 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3817 To run the instrumented application:
3819 * Start the application:
3829 The application is statically linked with the tracepoint provider
3830 package object file.
3832 The application dynamically loads the instrumented user library.
3834 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3837 include::../common/ust-sit-step-tp-o.txt[]
3839 To build the application:
3841 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3846 #define TRACEPOINT_DEFINE
3850 . Compile the application source file:
3859 . Build the application:
3864 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3869 The `--export-dynamic` option passed to the linker is necessary for the
3870 dynamically loaded library to ``see'' the tracepoint symbols defined in
3873 To build the instrumented user library:
3875 . Compile the user library source file:
3880 gcc -I. -fpic -c emon.c
3884 . Build the user library shared object:
3889 gcc -shared -o libemon.so emon.o
3893 To run the application:
3895 * Start the application:
3907 .Do not use man:dlclose(3) on a tracepoint provider package
3909 Never use man:dlclose(3) on any shared object which:
3911 * Is linked with, statically or dynamically, a tracepoint provider
3913 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3914 package shared object.
3916 This is currently considered **unsafe** due to a lack of reference
3917 counting from LTTng-UST to the shared object.
3919 A known workaround (available since glibc 2.2) is to use the
3920 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3921 effect of not unloading the loaded shared object, even if man:dlclose(3)
3924 You can also preload the tracepoint provider package shared object with
3925 the env:LD_PRELOAD environment variable to overcome this limitation.
3929 [[using-lttng-ust-with-daemons]]
3930 ===== Use noch:{LTTng-UST} with daemons
3932 If your instrumented application calls man:fork(2), man:clone(2),
3933 or BSD's man:rfork(2), without a following man:exec(3)-family
3934 system call, you must preload the path:{liblttng-ust-fork.so} shared
3935 object when starting the application.
3939 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3942 If your tracepoint provider package is
3943 a shared library which you also preload, you must put both
3944 shared objects in env:LD_PRELOAD:
3948 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3952 [[lttng-ust-pkg-config]]
3953 ===== Use noch:{pkg-config}
3955 On some distributions, LTTng-UST ships with a
3956 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3957 metadata file. If this is your case, then you can use cmd:pkg-config to
3958 build an application on the command line:
3962 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3966 [[instrumenting-32-bit-app-on-64-bit-system]]
3967 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3969 In order to trace a 32-bit application running on a 64-bit system,
3970 LTTng must use a dedicated 32-bit
3971 <<lttng-consumerd,consumer daemon>>.
3973 The following steps show how to build and install a 32-bit consumer
3974 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3975 build and install the 32-bit LTTng-UST libraries, and how to build and
3976 link an instrumented 32-bit application in that context.
3978 To build a 32-bit instrumented application for a 64-bit target system,
3979 assuming you have a fresh target system with no installed Userspace RCU
3982 . Download, build, and install a 32-bit version of Userspace RCU:
3988 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3989 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3990 cd userspace-rcu-0.9.* &&
3991 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3993 sudo make install &&
3998 . Using your distribution's package manager, or from source, install
3999 the following 32-bit versions of the following dependencies of
4000 LTTng-tools and LTTng-UST:
4003 * https://sourceforge.net/projects/libuuid/[libuuid]
4004 * http://directory.fsf.org/wiki/Popt[popt]
4005 * http://www.xmlsoft.org/[libxml2]
4008 . Download, build, and install a 32-bit version of the latest
4009 LTTng-UST{nbsp}{revision}:
4015 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
4016 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
4017 cd lttng-ust-2.8.* &&
4018 ./configure --libdir=/usr/local/lib32 \
4019 CFLAGS=-m32 CXXFLAGS=-m32 \
4020 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4022 sudo make install &&
4029 Depending on your distribution,
4030 32-bit libraries could be installed at a different location than
4031 `/usr/lib32`. For example, Debian is known to install
4032 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
4034 In this case, make sure to set `LDFLAGS` to all the
4035 relevant 32-bit library paths, for example:
4039 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
4043 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
4044 the 32-bit consumer daemon:
4050 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4051 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4052 cd lttng-tools-2.8.* &&
4053 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
4054 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4056 cd src/bin/lttng-consumerd &&
4057 sudo make install &&
4062 . From your distribution or from source,
4063 <<installing-lttng,install>> the 64-bit versions of
4064 LTTng-UST and Userspace RCU.
4065 . Download, build, and install the 64-bit version of the
4066 latest LTTng-tools{nbsp}{revision}:
4072 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
4073 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
4074 cd lttng-tools-2.8.* &&
4075 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
4076 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
4078 sudo make install &&
4083 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
4084 when linking your 32-bit application:
4087 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4088 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4091 For example, let's rebuild the quick start example in
4092 <<tracing-your-own-user-application,Trace a user application>> as an
4093 instrumented 32-bit application:
4098 gcc -m32 -c -I. hello-tp.c
4100 gcc -m32 -o hello hello.o hello-tp.o \
4101 -L/usr/lib32 -L/usr/local/lib32 \
4102 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4107 No special action is required to execute the 32-bit application and
4108 to trace it: use the command-line man:lttng(1) tool as usual.
4115 man:tracef(3) is a small LTTng-UST API designed for quick,
4116 man:printf(3)-like instrumentation without the burden of
4117 <<tracepoint-provider,creating>> and
4118 <<building-tracepoint-providers-and-user-application,building>>
4119 a tracepoint provider package.
4121 To use `tracef()` in your application:
4123 . In the C or C++ source files where you need to use `tracef()`,
4124 include `<lttng/tracef.h>`:
4129 #include <lttng/tracef.h>
4133 . In the application's source code, use `tracef()` like you would use
4141 tracef("my message: %d (%s)", my_integer, my_string);
4147 . Link your application with `liblttng-ust`:
4152 gcc -o app app.c -llttng-ust
4156 To trace the events that `tracef()` calls emit:
4158 * <<enabling-disabling-events,Create an event rule>> which matches the
4159 `lttng_ust_tracef:*` event name:
4164 lttng enable-event --userspace 'lttng_ust_tracef:*'
4169 .Limitations of `tracef()`
4171 The `tracef()` utility function was developed to make user space tracing
4172 super simple, albeit with notable disadvantages compared to
4173 <<defining-tracepoints,user-defined tracepoints>>:
4175 * All the emitted events have the same tracepoint provider and
4176 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4177 * There is no static type checking.
4178 * The only event record field you actually get, named `msg`, is a string
4179 potentially containing the values you passed to `tracef()`
4180 using your own format string. This also means that you cannot filter
4181 events with a custom expression at run time because there are no
4183 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4184 function behind the scenes to format the strings at run time, its
4185 expected performance is lower than with user-defined tracepoints,
4186 which do not require a conversion to a string.
4188 Taking this into consideration, `tracef()` is useful for some quick
4189 prototyping and debugging, but you should not consider it for any
4190 permanent and serious applicative instrumentation.
4196 ==== Use `tracelog()`
4198 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4199 the difference that it accepts an additional log level parameter.
4201 The goal of `tracelog()` is to ease the migration from logging to
4204 To use `tracelog()` in your application:
4206 . In the C or C++ source files where you need to use `tracelog()`,
4207 include `<lttng/tracelog.h>`:
4212 #include <lttng/tracelog.h>
4216 . In the application's source code, use `tracelog()` like you would use
4217 man:printf(3), except for the first parameter which is the log
4225 tracelog(TRACE_WARNING, "my message: %d (%s)",
4226 my_integer, my_string);
4232 See man:lttng-ust(3) for a list of available log level names.
4234 . Link your application with `liblttng-ust`:
4239 gcc -o app app.c -llttng-ust
4243 To trace the events that `tracelog()` calls emit with a log level
4244 _as severe as_ a specific log level:
4246 * <<enabling-disabling-events,Create an event rule>> which matches the
4247 `lttng_ust_tracelog:*` event name and a minimum level
4253 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4254 --loglevel=TRACE_WARNING
4258 To trace the events that `tracelog()` calls emit with a
4259 _specific log level_:
4261 * Create an event rule which matches the `lttng_ust_tracelog:*`
4262 event name and a specific log level:
4267 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4268 --loglevel-only=TRACE_INFO
4273 [[prebuilt-ust-helpers]]
4274 === Prebuilt user space tracing helpers
4276 The LTTng-UST package provides a few helpers in the form or preloadable
4277 shared objects which automatically instrument system functions and
4280 The helper shared objects are normally found in dir:{/usr/lib}. If you
4281 built LTTng-UST <<building-from-source,from source>>, they are probably
4282 located in dir:{/usr/local/lib}.
4284 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4287 path:{liblttng-ust-libc-wrapper.so}::
4288 path:{liblttng-ust-pthread-wrapper.so}::
4289 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4290 memory and POSIX threads function tracing>>.
4292 path:{liblttng-ust-cyg-profile.so}::
4293 path:{liblttng-ust-cyg-profile-fast.so}::
4294 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4296 path:{liblttng-ust-dl.so}::
4297 <<liblttng-ust-dl,Dynamic linker tracing>>.
4299 To use a user space tracing helper with any user application:
4301 * Preload the helper shared object when you start the application:
4306 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4310 You can preload more than one helper:
4315 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4321 [[liblttng-ust-libc-pthread-wrapper]]
4322 ==== Instrument C standard library memory and POSIX threads functions
4324 The path:{liblttng-ust-libc-wrapper.so} and
4325 path:{liblttng-ust-pthread-wrapper.so} helpers
4326 add instrumentation to some C standard library and POSIX
4330 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4332 |TP provider name |TP name |Instrumented function
4334 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4335 |`calloc` |man:calloc(3)
4336 |`realloc` |man:realloc(3)
4337 |`free` |man:free(3)
4338 |`memalign` |man:memalign(3)
4339 |`posix_memalign` |man:posix_memalign(3)
4343 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4345 |TP provider name |TP name |Instrumented function
4347 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4348 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4349 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4350 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4353 When you preload the shared object, it replaces the functions listed
4354 in the previous tables by wrappers which contain tracepoints and call
4355 the replaced functions.
4358 [[liblttng-ust-cyg-profile]]
4359 ==== Instrument function entry and exit
4361 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4362 to the entry and exit points of functions.
4364 man:gcc(1) and man:clang(1) have an option named
4365 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4366 which generates instrumentation calls for entry and exit to functions.
4367 The LTTng-UST function tracing helpers,
4368 path:{liblttng-ust-cyg-profile.so} and
4369 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4370 to add tracepoints to the two generated functions (which contain
4371 `cyg_profile` in their names, hence the helper's name).
4373 To use the LTTng-UST function tracing helper, the source files to
4374 instrument must be built using the `-finstrument-functions` compiler
4377 There are two versions of the LTTng-UST function tracing helper:
4379 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4380 that you should only use when it can be _guaranteed_ that the
4381 complete event stream is recorded without any lost event record.
4382 Any kind of duplicate information is left out.
4384 Assuming no event record is lost, having only the function addresses on
4385 entry is enough to create a call graph, since an event record always
4386 contains the ID of the CPU that generated it.
4388 You can use a tool like man:addr2line(1) to convert function addresses
4389 back to source file names and line numbers.
4391 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4392 which also works in use cases where event records might get discarded or
4393 not recorded from application startup.
4394 In these cases, the trace analyzer needs more information to be
4395 able to reconstruct the program flow.
4397 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4398 points of this helper.
4400 All the tracepoints that this helper provides have the
4401 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4403 TIP: It's sometimes a good idea to limit the number of source files that
4404 you compile with the `-finstrument-functions` option to prevent LTTng
4405 from writing an excessive amount of trace data at run time. When using
4406 man:gcc(1), you can use the
4407 `-finstrument-functions-exclude-function-list` option to avoid
4408 instrument entries and exits of specific function names.
4413 ==== Instrument the dynamic linker
4415 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4416 man:dlopen(3) and man:dlclose(3) function calls.
4418 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4423 [[java-application]]
4424 === User space Java agent
4426 You can instrument any Java application which uses one of the following
4429 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4430 (JUL) core logging facilities.
4431 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4432 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4435 .LTTng-UST Java agent imported by a Java application.
4436 image::java-app.png[]
4438 Note that the methods described below are new in LTTng{nbsp}{revision}.
4439 Previous LTTng versions use another technique.
4441 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4442 and https://ci.lttng.org/[continuous integration], thus this version is
4443 directly supported. However, the LTTng-UST Java agent is also tested
4444 with OpenJDK{nbsp}7.
4449 ==== Use the LTTng-UST Java agent for `java.util.logging`
4451 To use the LTTng-UST Java agent in a Java application which uses
4452 `java.util.logging` (JUL):
4454 . In the Java application's source code, import the LTTng-UST
4455 log handler package for `java.util.logging`:
4460 import org.lttng.ust.agent.jul.LttngLogHandler;
4464 . Create an LTTng-UST JUL log handler:
4469 Handler lttngUstLogHandler = new LttngLogHandler();
4473 . Add this handler to the JUL loggers which should emit LTTng events:
4478 Logger myLogger = Logger.getLogger("some-logger");
4480 myLogger.addHandler(lttngUstLogHandler);
4484 . Use `java.util.logging` log statements and configuration as usual.
4485 The loggers with an attached LTTng-UST log handler can emit
4488 . Before exiting the application, remove the LTTng-UST log handler from
4489 the loggers attached to it and call its `close()` method:
4494 myLogger.removeHandler(lttngUstLogHandler);
4495 lttngUstLogHandler.close();
4499 This is not strictly necessary, but it is recommended for a clean
4500 disposal of the handler's resources.
4502 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4503 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4505 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4506 path] when you build the Java application.
4508 The JAR files are typically located in dir:{/usr/share/java}.
4510 IMPORTANT: The LTTng-UST Java agent must be
4511 <<installing-lttng,installed>> for the logging framework your
4514 .Use the LTTng-UST Java agent for `java.util.logging`.
4519 import java.io.IOException;
4520 import java.util.logging.Handler;
4521 import java.util.logging.Logger;
4522 import org.lttng.ust.agent.jul.LttngLogHandler;
4526 private static final int answer = 42;
4528 public static void main(String[] argv) throws Exception
4531 Logger logger = Logger.getLogger("jello");
4533 // Create an LTTng-UST log handler
4534 Handler lttngUstLogHandler = new LttngLogHandler();
4536 // Add the LTTng-UST log handler to our logger
4537 logger.addHandler(lttngUstLogHandler);
4540 logger.info("some info");
4541 logger.warning("some warning");
4543 logger.finer("finer information; the answer is " + answer);
4545 logger.severe("error!");
4547 // Not mandatory, but cleaner
4548 logger.removeHandler(lttngUstLogHandler);
4549 lttngUstLogHandler.close();
4558 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4561 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4562 <<enabling-disabling-events,create an event rule>> matching the
4563 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4568 lttng enable-event --jul jello
4572 Run the compiled class:
4576 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4579 <<basic-tracing-session-control,Stop tracing>> and inspect the
4589 In the resulting trace, an <<event,event record>> generated by a Java
4590 application using `java.util.logging` is named `lttng_jul:event` and
4591 has the following fields:
4594 Log record's message.
4600 Name of the class in which the log statement was executed.
4603 Name of the method in which the log statement was executed.
4606 Logging time (timestamp in milliseconds).
4609 Log level integer value.
4612 ID of the thread in which the log statement was executed.
4614 You can use the opt:lttng-enable-event(1):--loglevel or
4615 opt:lttng-enable-event(1):--loglevel-only option of the
4616 man:lttng-enable-event(1) command to target a range of JUL log levels
4617 or a specific JUL log level.
4622 ==== Use the LTTng-UST Java agent for Apache log4j
4624 To use the LTTng-UST Java agent in a Java application which uses
4627 . In the Java application's source code, import the LTTng-UST
4628 log appender package for Apache log4j:
4633 import org.lttng.ust.agent.log4j.LttngLogAppender;
4637 . Create an LTTng-UST log4j log appender:
4642 Appender lttngUstLogAppender = new LttngLogAppender();
4646 . Add this appender to the log4j loggers which should emit LTTng events:
4651 Logger myLogger = Logger.getLogger("some-logger");
4653 myLogger.addAppender(lttngUstLogAppender);
4657 . Use Apache log4j log statements and configuration as usual. The
4658 loggers with an attached LTTng-UST log appender can emit LTTng events.
4660 . Before exiting the application, remove the LTTng-UST log appender from
4661 the loggers attached to it and call its `close()` method:
4666 myLogger.removeAppender(lttngUstLogAppender);
4667 lttngUstLogAppender.close();
4671 This is not strictly necessary, but it is recommended for a clean
4672 disposal of the appender's resources.
4674 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4675 files, path:{lttng-ust-agent-common.jar} and
4676 path:{lttng-ust-agent-log4j.jar}, in the
4677 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4678 path] when you build the Java application.
4680 The JAR files are typically located in dir:{/usr/share/java}.
4682 IMPORTANT: The LTTng-UST Java agent must be
4683 <<installing-lttng,installed>> for the logging framework your
4686 .Use the LTTng-UST Java agent for Apache log4j.
4691 import org.apache.log4j.Appender;
4692 import org.apache.log4j.Logger;
4693 import org.lttng.ust.agent.log4j.LttngLogAppender;
4697 private static final int answer = 42;
4699 public static void main(String[] argv) throws Exception
4702 Logger logger = Logger.getLogger("jello");
4704 // Create an LTTng-UST log appender
4705 Appender lttngUstLogAppender = new LttngLogAppender();
4707 // Add the LTTng-UST log appender to our logger
4708 logger.addAppender(lttngUstLogAppender);
4711 logger.info("some info");
4712 logger.warn("some warning");
4714 logger.debug("debug information; the answer is " + answer);
4716 logger.fatal("error!");
4718 // Not mandatory, but cleaner
4719 logger.removeAppender(lttngUstLogAppender);
4720 lttngUstLogAppender.close();
4726 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4731 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4734 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4735 <<enabling-disabling-events,create an event rule>> matching the
4736 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4741 lttng enable-event --log4j jello
4745 Run the compiled class:
4749 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4752 <<basic-tracing-session-control,Stop tracing>> and inspect the
4762 In the resulting trace, an <<event,event record>> generated by a Java
4763 application using log4j is named `lttng_log4j:event` and
4764 has the following fields:
4767 Log record's message.
4773 Name of the class in which the log statement was executed.
4776 Name of the method in which the log statement was executed.
4779 Name of the file in which the executed log statement is located.
4782 Line number at which the log statement was executed.
4788 Log level integer value.
4791 Name of the Java thread in which the log statement was executed.
4793 You can use the opt:lttng-enable-event(1):--loglevel or
4794 opt:lttng-enable-event(1):--loglevel-only option of the
4795 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4796 or a specific log4j log level.
4800 [[java-application-context]]
4801 ==== Provide application-specific context fields in a Java application
4803 A Java application-specific context field is a piece of state provided
4804 by the application which <<adding-context,you can add>>, using the
4805 man:lttng-add-context(1) command, to each <<event,event record>>
4806 produced by the log statements of this application.
4808 For example, a given object might have a current request ID variable.
4809 You can create a context information retriever for this object and
4810 assign a name to this current request ID. You can then, using the
4811 man:lttng-add-context(1) command, add this context field by name to
4812 the JUL or log4j <<channel,channel>>.
4814 To provide application-specific context fields in a Java application:
4816 . In the Java application's source code, import the LTTng-UST
4817 Java agent context classes and interfaces:
4822 import org.lttng.ust.agent.context.ContextInfoManager;
4823 import org.lttng.ust.agent.context.IContextInfoRetriever;
4827 . Create a context information retriever class, that is, a class which
4828 implements the `IContextInfoRetriever` interface:
4833 class MyContextInfoRetriever implements IContextInfoRetriever
4836 public Object retrieveContextInfo(String key)
4838 if (key.equals("intCtx")) {
4840 } else if (key.equals("strContext")) {
4841 return "context value!";
4850 This `retrieveContextInfo()` method is the only member of the
4851 `IContextInfoRetriever` interface. Its role is to return the current
4852 value of a state by name to create a context field. The names of the
4853 context fields and which state variables they return depends on your
4856 All primitive types and objects are supported as context fields.
4857 When `retrieveContextInfo()` returns an object, the context field
4858 serializer calls its `toString()` method to add a string field to
4859 event records. The method can also return `null`, which means that
4860 no context field is available for the required name.
4862 . Register an instance of your context information retriever class to
4863 the context information manager singleton:
4868 IContextInfoRetriever cir = new MyContextInfoRetriever();
4869 ContextInfoManager cim = ContextInfoManager.getInstance();
4870 cim.registerContextInfoRetriever("retrieverName", cir);
4874 . Before exiting the application, remove your context information
4875 retriever from the context information manager singleton:
4880 ContextInfoManager cim = ContextInfoManager.getInstance();
4881 cim.unregisterContextInfoRetriever("retrieverName");
4885 This is not strictly necessary, but it is recommended for a clean
4886 disposal of some manager's resources.
4888 . Build your Java application with LTTng-UST Java agent support as
4889 usual, following the procedure for either the <<jul,JUL>> or
4890 <<log4j,Apache log4j>> framework.
4893 .Provide application-specific context fields in a Java application.
4898 import java.util.logging.Handler;
4899 import java.util.logging.Logger;
4900 import org.lttng.ust.agent.jul.LttngLogHandler;
4901 import org.lttng.ust.agent.context.ContextInfoManager;
4902 import org.lttng.ust.agent.context.IContextInfoRetriever;
4906 // Our context information retriever class
4907 private static class MyContextInfoRetriever
4908 implements IContextInfoRetriever
4911 public Object retrieveContextInfo(String key) {
4912 if (key.equals("intCtx")) {
4914 } else if (key.equals("strContext")) {
4915 return "context value!";
4922 private static final int answer = 42;
4924 public static void main(String args[]) throws Exception
4926 // Get the context information manager instance
4927 ContextInfoManager cim = ContextInfoManager.getInstance();
4929 // Create and register our context information retriever
4930 IContextInfoRetriever cir = new MyContextInfoRetriever();
4931 cim.registerContextInfoRetriever("myRetriever", cir);
4934 Logger logger = Logger.getLogger("jello");
4936 // Create an LTTng-UST log handler
4937 Handler lttngUstLogHandler = new LttngLogHandler();
4939 // Add the LTTng-UST log handler to our logger
4940 logger.addHandler(lttngUstLogHandler);
4943 logger.info("some info");
4944 logger.warning("some warning");
4946 logger.finer("finer information; the answer is " + answer);
4948 logger.severe("error!");
4950 // Not mandatory, but cleaner
4951 logger.removeHandler(lttngUstLogHandler);
4952 lttngUstLogHandler.close();
4953 cim.unregisterContextInfoRetriever("myRetriever");
4962 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4965 <<creating-destroying-tracing-sessions,Create a tracing session>>
4966 and <<enabling-disabling-events,create an event rule>> matching the
4972 lttng enable-event --jul jello
4975 <<adding-context,Add the application-specific context fields>> to the
4980 lttng add-context --jul --type='$app.myRetriever:intCtx'
4981 lttng add-context --jul --type='$app.myRetriever:strContext'
4984 <<basic-tracing-session-control,Start tracing>>:
4991 Run the compiled class:
4995 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4998 <<basic-tracing-session-control,Stop tracing>> and inspect the
5010 [[python-application]]
5011 === User space Python agent
5013 You can instrument a Python 2 or Python 3 application which uses the
5014 standard https://docs.python.org/3/library/logging.html[`logging`]
5017 Each log statement emits an LTTng event once the
5018 application module imports the
5019 <<lttng-ust-agents,LTTng-UST Python agent>> package.
5022 .A Python application importing the LTTng-UST Python agent.
5023 image::python-app.png[]
5025 To use the LTTng-UST Python agent:
5027 . In the Python application's source code, import the LTTng-UST Python
5037 The LTTng-UST Python agent automatically adds its logging handler to the
5038 root logger at import time.
5040 Any log statement that the application executes before this import does
5041 not emit an LTTng event.
5043 IMPORTANT: The LTTng-UST Python agent must be
5044 <<installing-lttng,installed>>.
5046 . Use log statements and logging configuration as usual.
5047 Since the LTTng-UST Python agent adds a handler to the _root_
5048 logger, you can trace any log statement from any logger.
5050 .Use the LTTng-UST Python agent.
5061 logging.basicConfig()
5062 logger = logging.getLogger('my-logger')
5065 logger.debug('debug message')
5066 logger.info('info message')
5067 logger.warn('warn message')
5068 logger.error('error message')
5069 logger.critical('critical message')
5073 if __name__ == '__main__':
5077 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
5078 logging handler which prints to the standard error stream, is not
5079 strictly required for LTTng-UST tracing to work, but in versions of
5080 Python preceding 3.2, you could see a warning message which indicates
5081 that no handler exists for the logger `my-logger`.
5083 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5084 <<enabling-disabling-events,create an event rule>> matching the
5085 `my-logger` Python logger, and <<basic-tracing-session-control,start
5091 lttng enable-event --python my-logger
5095 Run the Python script:
5102 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5112 In the resulting trace, an <<event,event record>> generated by a Python
5113 application is named `lttng_python:event` and has the following fields:
5116 Logging time (string).
5119 Log record's message.
5125 Name of the function in which the log statement was executed.
5128 Line number at which the log statement was executed.
5131 Log level integer value.
5134 ID of the Python thread in which the log statement was executed.
5137 Name of the Python thread in which the log statement was executed.
5139 You can use the opt:lttng-enable-event(1):--loglevel or
5140 opt:lttng-enable-event(1):--loglevel-only option of the
5141 man:lttng-enable-event(1) command to target a range of Python log levels
5142 or a specific Python log level.
5144 When an application imports the LTTng-UST Python agent, the agent tries
5145 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5146 <<start-sessiond,start the session daemon>> _before_ you run the Python
5147 application. If a session daemon is found, the agent tries to register
5148 to it during 5{nbsp}seconds, after which the application continues
5149 without LTTng tracing support. You can override this timeout value with
5150 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5153 If the session daemon stops while a Python application with an imported
5154 LTTng-UST Python agent runs, the agent retries to connect and to
5155 register to a session daemon every 3{nbsp}seconds. You can override this
5156 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5161 [[proc-lttng-logger-abi]]
5164 The `lttng-tracer` Linux kernel module, part of
5165 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5166 path:{/proc/lttng-logger} when it's loaded. Any application can write
5167 text data to this file to emit an LTTng event.
5170 .An application writes to the LTTng logger file to emit an LTTng event.
5171 image::lttng-logger.png[]
5173 The LTTng logger is the quickest method--not the most efficient,
5174 however--to add instrumentation to an application. It is designed
5175 mostly to instrument shell scripts:
5179 echo "Some message, some $variable" > /proc/lttng-logger
5182 Any event that the LTTng logger emits is named `lttng_logger` and
5183 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5184 other instrumentation points in the kernel tracing domain, **any Unix
5185 user** can <<enabling-disabling-events,create an event rule>> which
5186 matches its event name, not only the root user or users in the
5187 <<tracing-group,tracing group>>.
5189 To use the LTTng logger:
5191 * From any application, write text data to the path:{/proc/lttng-logger}
5194 The `msg` field of `lttng_logger` event records contains the
5197 NOTE: The maximum message length of an LTTng logger event is
5198 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5199 than one event to contain the remaining data.
5201 You should not use the LTTng logger to trace a user application which
5202 can be instrumented in a more efficient way, namely:
5204 * <<c-application,C and $$C++$$ applications>>.
5205 * <<java-application,Java applications>>.
5206 * <<python-application,Python applications>>.
5208 .Use the LTTng logger.
5213 echo 'Hello, World!' > /proc/lttng-logger
5215 df --human-readable --print-type / > /proc/lttng-logger
5218 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5219 <<enabling-disabling-events,create an event rule>> matching the
5220 `lttng_logger` Linux kernel tracepoint, and
5221 <<basic-tracing-session-control,start tracing>>:
5226 lttng enable-event --kernel lttng_logger
5230 Run the Bash script:
5237 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5248 [[instrumenting-linux-kernel]]
5249 === LTTng kernel tracepoints
5251 NOTE: This section shows how to _add_ instrumentation points to the
5252 Linux kernel. The kernel's subsystems are already thoroughly
5253 instrumented at strategic places for LTTng when you
5254 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5258 There are two methods to instrument the Linux kernel:
5260 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5261 tracepoint which uses the `TRACE_EVENT()` API.
5263 Choose this if you want to instrumentation a Linux kernel tree with an
5264 instrumentation point compatible with ftrace, perf, and SystemTap.
5266 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5267 instrument an out-of-tree kernel module.
5269 Choose this if you don't need ftrace, perf, or SystemTap support.
5273 [[linux-add-lttng-layer]]
5274 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5276 This section shows how to add an LTTng layer to existing ftrace
5277 instrumentation using the `TRACE_EVENT()` API.
5279 This section does not document the `TRACE_EVENT()` macro. You can
5280 read the following articles to learn more about this API:
5282 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5283 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5284 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5286 The following procedure assumes that your ftrace tracepoints are
5287 correctly defined in their own header and that they are created in
5288 one source file using the `CREATE_TRACE_POINTS` definition.
5290 To add an LTTng layer over an existing ftrace tracepoint:
5292 . Make sure the following kernel configuration options are
5298 * `CONFIG_HIGH_RES_TIMERS`
5299 * `CONFIG_TRACEPOINTS`
5302 . Build the Linux source tree with your custom ftrace tracepoints.
5303 . Boot the resulting Linux image on your target system.
5305 Confirm that the tracepoints exist by looking for their names in the
5306 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5307 is your subsystem's name.
5309 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5315 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
5316 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
5317 cd lttng-modules-2.8.*
5321 . In dir:{instrumentation/events/lttng-module}, relative to the root
5322 of the LTTng-modules source tree, create a header file named
5323 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5324 LTTng-modules tracepoint definitions using the LTTng-modules
5327 Start with this template:
5331 .path:{instrumentation/events/lttng-module/my_subsys.h}
5334 #define TRACE_SYSTEM my_subsys
5336 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5337 #define _LTTNG_MY_SUBSYS_H
5339 #include "../../../probes/lttng-tracepoint-event.h"
5340 #include <linux/tracepoint.h>
5342 LTTNG_TRACEPOINT_EVENT(
5344 * Format is identical to TRACE_EVENT()'s version for the three
5345 * following macro parameters:
5348 TP_PROTO(int my_int, const char *my_string),
5349 TP_ARGS(my_int, my_string),
5351 /* LTTng-modules specific macros */
5353 ctf_integer(int, my_int_field, my_int)
5354 ctf_string(my_bar_field, my_bar)
5358 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5360 #include "../../../probes/define_trace.h"
5364 The entries in the `TP_FIELDS()` section are the list of fields for the
5365 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5366 ftrace's `TRACE_EVENT()` macro.
5368 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5369 complete description of the available `ctf_*()` macros.
5371 . Create the LTTng-modules probe's kernel module C source file,
5372 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5377 .path:{probes/lttng-probe-my-subsys.c}
5379 #include <linux/module.h>
5380 #include "../lttng-tracer.h"
5383 * Build-time verification of mismatch between mainline
5384 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5385 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5387 #include <trace/events/my_subsys.h>
5389 /* Create LTTng tracepoint probes */
5390 #define LTTNG_PACKAGE_BUILD
5391 #define CREATE_TRACE_POINTS
5392 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5394 #include "../instrumentation/events/lttng-module/my_subsys.h"
5396 MODULE_LICENSE("GPL and additional rights");
5397 MODULE_AUTHOR("Your name <your-email>");
5398 MODULE_DESCRIPTION("LTTng my_subsys probes");
5399 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5400 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5401 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5402 LTTNG_MODULES_EXTRAVERSION);
5406 . Edit path:{probes/Makefile} and add your new kernel module object
5407 next to the existing ones:
5411 .path:{probes/Makefile}
5415 obj-m += lttng-probe-module.o
5416 obj-m += lttng-probe-power.o
5418 obj-m += lttng-probe-my-subsys.o
5424 . Build and install the LTTng kernel modules:
5429 make KERNELDIR=/path/to/linux
5430 sudo make modules_install
5434 Replace `/path/to/linux` with the path to the Linux source tree where
5435 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5437 Note that you can also use the
5438 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5439 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5440 C code that need to be executed before the event fields are recorded.
5442 The best way to learn how to use the previous LTTng-modules macros is to
5443 inspect the existing LTTng-modules tracepoint definitions in the
5444 dir:{instrumentation/events/lttng-module} header files. Compare them
5445 with the Linux kernel mainline versions in the
5446 dir:{include/trace/events} directory of the Linux source tree.
5450 [[lttng-tracepoint-event-code]]
5451 ===== Use custom C code to access the data for tracepoint fields
5453 Although we recommended to always use the
5454 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5455 the arguments and fields of an LTTng-modules tracepoint when possible,
5456 sometimes you need a more complex process to access the data that the
5457 tracer records as event record fields. In other words, you need local
5458 variables and multiple C{nbsp}statements instead of simple
5459 argument-based expressions that you pass to the
5460 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5462 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5463 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5464 a block of C{nbsp}code to be executed before LTTng records the fields.
5465 The structure of this macro is:
5468 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5470 LTTNG_TRACEPOINT_EVENT_CODE(
5472 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5473 * version for the following three macro parameters:
5476 TP_PROTO(int my_int, const char *my_string),
5477 TP_ARGS(my_int, my_string),
5479 /* Declarations of custom local variables */
5482 unsigned long b = 0;
5483 const char *name = "(undefined)";
5484 struct my_struct *my_struct;
5488 * Custom code which uses both tracepoint arguments
5489 * (in TP_ARGS()) and local variables (in TP_locvar()).
5491 * Local variables are actually members of a structure pointed
5492 * to by the special variable tp_locvar.
5496 tp_locvar->a = my_int + 17;
5497 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5498 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5499 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5500 put_my_struct(tp_locvar->my_struct);
5509 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5510 * version for this, except that tp_locvar members can be
5511 * used in the argument expression parameters of
5512 * the ctf_*() macros.
5515 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5516 ctf_integer(int, my_struct_a, tp_locvar->a)
5517 ctf_string(my_string_field, my_string)
5518 ctf_string(my_struct_name, tp_locvar->name)
5523 IMPORTANT: The C code defined in `TP_code()` must not have any side
5524 effects when executed. In particular, the code must not allocate
5525 memory or get resources without deallocating this memory or putting
5526 those resources afterwards.
5529 [[instrumenting-linux-kernel-tracing]]
5530 ==== Load and unload a custom probe kernel module
5532 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5533 kernel module>> in the kernel before it can emit LTTng events.
5535 To load the default probe kernel modules and a custom probe kernel
5538 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5539 probe modules to load when starting a root <<lttng-sessiond,session
5543 .Load the `my_subsys`, `usb`, and the default probe modules.
5547 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5552 You only need to pass the subsystem name, not the whole kernel module
5555 To load _only_ a given custom probe kernel module:
5557 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5558 modules to load when starting a root session daemon:
5561 .Load only the `my_subsys` and `usb` probe modules.
5565 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5570 To confirm that a probe module is loaded:
5577 lsmod | grep lttng_probe_usb
5581 To unload the loaded probe modules:
5583 * Kill the session daemon with `SIGTERM`:
5588 sudo pkill lttng-sessiond
5592 You can also use man:modprobe(8)'s `--remove` option if the session
5593 daemon terminates abnormally.
5596 [[controlling-tracing]]
5599 Once an application or a Linux kernel is
5600 <<instrumenting,instrumented>> for LTTng tracing,
5603 This section is divided in topics on how to use the various
5604 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5605 command-line tool>>, to _control_ the LTTng daemons and tracers.
5607 NOTE: In the following subsections, we refer to an man:lttng(1) command
5608 using its man page name. For example, instead of _Run the `create`
5609 command to..._, we use _Run the man:lttng-create(1) command to..._.
5613 === Start a session daemon
5615 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5616 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5619 You will see the following error when you run a command while no session
5623 Error: No session daemon is available
5626 The only command that automatically runs a session daemon is
5627 man:lttng-create(1), which you use to
5628 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5629 this is most of the time the first operation that you do, sometimes it's
5630 not. Some examples are:
5632 * <<list-instrumentation-points,List the available instrumentation points>>.
5633 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5635 [[tracing-group]] Each Unix user must have its own running session
5636 daemon to trace user applications. The session daemon that the root user
5637 starts is the only one allowed to control the LTTng kernel tracer. Users
5638 that are part of the _tracing group_ can control the root session
5639 daemon. The default tracing group name is `tracing`; you can set it to
5640 something else with the opt:lttng-sessiond(8):--group option when you
5641 start the root session daemon.
5643 To start a user session daemon:
5645 * Run man:lttng-sessiond(8):
5650 lttng-sessiond --daemonize
5654 To start the root session daemon:
5656 * Run man:lttng-sessiond(8) as the root user:
5661 sudo lttng-sessiond --daemonize
5665 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5666 start the session daemon in foreground.
5668 To stop a session daemon, use man:kill(1) on its process ID (standard
5671 Note that some Linux distributions could manage the LTTng session daemon
5672 as a service. In this case, you should use the service manager to
5673 start, restart, and stop session daemons.
5676 [[creating-destroying-tracing-sessions]]
5677 === Create and destroy a tracing session
5679 Almost all the LTTng control operations happen in the scope of
5680 a <<tracing-session,tracing session>>, which is the dialogue between the
5681 <<lttng-sessiond,session daemon>> and you.
5683 To create a tracing session with a generated name:
5685 * Use the man:lttng-create(1) command:
5694 The created tracing session's name is `auto` followed by the
5697 To create a tracing session with a specific name:
5699 * Use the optional argument of the man:lttng-create(1) command:
5704 lttng create my-session
5708 Replace `my-session` with the specific tracing session name.
5710 LTTng appends the creation date to the created tracing session's name.
5712 LTTng writes the traces of a tracing session in
5713 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5714 name of the tracing session. Note that the env:LTTNG_HOME environment
5715 variable defaults to `$HOME` if not set.
5717 To output LTTng traces to a non-default location:
5719 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5724 lttng create --output=/tmp/some-directory my-session
5728 You may create as many tracing sessions as you wish.
5730 To list all the existing tracing sessions for your Unix user:
5732 * Use the man:lttng-list(1) command:
5741 When you create a tracing session, it is set as the _current tracing
5742 session_. The following man:lttng(1) commands operate on the current
5743 tracing session when you don't specify one:
5745 [role="list-3-cols"]
5761 To change the current tracing session:
5763 * Use the man:lttng-set-session(1) command:
5768 lttng set-session new-session
5772 Replace `new-session` by the name of the new current tracing session.
5774 When you are done tracing in a given tracing session, you can destroy
5775 it. This operation frees the resources taken by the tracing session
5776 to destroy; it does not destroy the trace data that LTTng wrote for
5777 this tracing session.
5779 To destroy the current tracing session:
5781 * Use the man:lttng-destroy(1) command:
5791 [[list-instrumentation-points]]
5792 === List the available instrumentation points
5794 The <<lttng-sessiond,session daemon>> can query the running instrumented
5795 user applications and the Linux kernel to get a list of available
5796 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5797 they are tracepoints and system calls. For the user space tracing
5798 domain, they are tracepoints. For the other tracing domains, they are
5801 To list the available instrumentation points:
5803 * Use the man:lttng-list(1) command with the requested tracing domain's
5807 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5808 must be a root user, or it must be a member of the
5809 <<tracing-group,tracing group>>).
5810 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5811 kernel system calls (your Unix user must be a root user, or it must be
5812 a member of the tracing group).
5813 * opt:lttng-list(1):--userspace: user space tracepoints.
5814 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5815 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5816 * opt:lttng-list(1):--python: Python loggers.
5819 .List the available user space tracepoints.
5823 lttng list --userspace
5827 .List the available Linux kernel system call tracepoints.
5831 lttng list --kernel --syscall
5836 [[enabling-disabling-events]]
5837 === Create and enable an event rule
5839 Once you <<creating-destroying-tracing-sessions,create a tracing
5840 session>>, you can create <<event,event rules>> with the
5841 man:lttng-enable-event(1) command.
5843 You specify each condition with a command-line option. The available
5844 condition options are shown in the following table.
5846 [role="growable",cols="asciidoc,asciidoc,default"]
5847 .Condition command-line options for the man:lttng-enable-event(1) command.
5849 |Option |Description |Applicable tracing domains
5855 . +--probe=__ADDR__+
5856 . +--function=__ADDR__+
5859 Instead of using the default _tracepoint_ instrumentation type, use:
5861 . A Linux system call.
5862 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5863 . The entry and return points of a Linux function (symbol or address).
5867 |First positional argument.
5870 Tracepoint or system call name. In the case of a Linux KProbe or
5871 function, this is a custom name given to the event rule. With the
5872 JUL, log4j, and Python domains, this is a logger name.
5874 With a tracepoint, logger, or system call name, the last character
5875 can be `*` to match anything that remains.
5882 . +--loglevel=__LEVEL__+
5883 . +--loglevel-only=__LEVEL__+
5886 . Match only tracepoints or log statements with a logging level at
5887 least as severe as +__LEVEL__+.
5888 . Match only tracepoints or log statements with a logging level
5889 equal to +__LEVEL__+.
5891 See man:lttng-enable-event(1) for the list of available logging level
5894 |User space, JUL, log4j, and Python.
5896 |+--exclude=__EXCLUSIONS__+
5899 When you use a `*` character at the end of the tracepoint or logger
5900 name (first positional argument), exclude the specific names in the
5901 comma-delimited list +__EXCLUSIONS__+.
5904 User space, JUL, log4j, and Python.
5906 |+--filter=__EXPR__+
5909 Match only events which satisfy the expression +__EXPR__+.
5911 See man:lttng-enable-event(1) to learn more about the syntax of a
5918 You attach an event rule to a <<channel,channel>> on creation. If you do
5919 not specify the channel with the opt:lttng-enable-event(1):--channel
5920 option, and if the event rule to create is the first in its
5921 <<domain,tracing domain>> for a given tracing session, then LTTng
5922 creates a _default channel_ for you. This default channel is reused in
5923 subsequent invocations of the man:lttng-enable-event(1) command for the
5924 same tracing domain.
5926 An event rule is always enabled at creation time.
5928 The following examples show how you can combine the previous
5929 command-line options to create simple to more complex event rules.
5931 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5935 lttng enable-event --kernel sched_switch
5939 .Create an event rule matching four Linux kernel system calls (default channel).
5943 lttng enable-event --kernel --syscall open,write,read,close
5947 .Create event rules matching tracepoints with filter expressions (default channel).
5951 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5956 lttng enable-event --kernel --all \
5957 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5962 lttng enable-event --jul my_logger \
5963 --filter='$app.retriever:cur_msg_id > 3'
5966 IMPORTANT: Make sure to always quote the filter string when you
5967 use man:lttng(1) from a shell.
5970 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5974 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5977 IMPORTANT: Make sure to always quote the wildcard character when you
5978 use man:lttng(1) from a shell.
5981 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5985 lttng enable-event --python my-app.'*' \
5986 --exclude='my-app.module,my-app.hello'
5990 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5994 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5998 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
6002 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
6006 The event rules of a given channel form a whitelist: as soon as an
6007 emitted event passes one of them, LTTng can record the event. For
6008 example, an event named `my_app:my_tracepoint` emitted from a user space
6009 tracepoint with a `TRACE_ERROR` log level passes both of the following
6014 lttng enable-event --userspace my_app:my_tracepoint
6015 lttng enable-event --userspace my_app:my_tracepoint \
6016 --loglevel=TRACE_INFO
6019 The second event rule is redundant: the first one includes
6023 [[disable-event-rule]]
6024 === Disable an event rule
6026 To disable an event rule that you <<enabling-disabling-events,created>>
6027 previously, use the man:lttng-disable-event(1) command. This command
6028 disables _all_ the event rules (of a given tracing domain and channel)
6029 which match an instrumentation point. The other conditions are not
6030 supported as of LTTng{nbsp}{revision}.
6032 The LTTng tracer does not record an emitted event which passes
6033 a _disabled_ event rule.
6035 .Disable an event rule matching a Python logger (default channel).
6039 lttng disable-event --python my-logger
6043 .Disable an event rule matching all `java.util.logging` loggers (default channel).
6047 lttng disable-event --jul '*'
6051 .Disable _all_ the event rules of the default channel.
6053 The opt:lttng-disable-event(1):--all-events option is not, like the
6054 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
6055 equivalent of the event name `*` (wildcard): it disables _all_ the event
6056 rules of a given channel.
6060 lttng disable-event --jul --all-events
6064 NOTE: You cannot delete an event rule once you create it.
6068 === Get the status of a tracing session
6070 To get the status of the current tracing session, that is, its
6071 parameters, its channels, event rules, and their attributes:
6073 * Use the man:lttng-status(1) command:
6083 To get the status of any tracing session:
6085 * Use the man:lttng-list(1) command with the tracing session's name:
6090 lttng list my-session
6094 Replace `my-session` with the desired tracing session's name.
6097 [[basic-tracing-session-control]]
6098 === Start and stop a tracing session
6100 Once you <<creating-destroying-tracing-sessions,create a tracing
6102 <<enabling-disabling-events,create one or more event rules>>,
6103 you can start and stop the tracers for this tracing session.
6105 To start tracing in the current tracing session:
6107 * Use the man:lttng-start(1) command:
6116 LTTng is very flexible: you can launch user applications before
6117 or after the you start the tracers. The tracers only record the events
6118 if they pass enabled event rules and if they occur while the tracers are
6121 To stop tracing in the current tracing session:
6123 * Use the man:lttng-stop(1) command:
6132 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6133 records>> or lost sub-buffers since the last time you ran
6134 man:lttng-start(1), warnings are printed when you run the
6135 man:lttng-stop(1) command.
6138 [[enabling-disabling-channels]]
6139 === Create a channel
6141 Once you create a tracing session, you can create a <<channel,channel>>
6142 with the man:lttng-enable-channel(1) command.
6144 Note that LTTng automatically creates a default channel when, for a
6145 given <<domain,tracing domain>>, no channels exist and you
6146 <<enabling-disabling-events,create>> the first event rule. This default
6147 channel is named `channel0` and its attributes are set to reasonable
6148 values. Therefore, you only need to create a channel when you need
6149 non-default attributes.
6151 You specify each non-default channel attribute with a command-line
6152 option when you use the man:lttng-enable-channel(1) command. The
6153 available command-line options are:
6155 [role="growable",cols="asciidoc,asciidoc"]
6156 .Command-line options for the man:lttng-enable-channel(1) command.
6158 |Option |Description
6164 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
6165 the default _discard_ mode.
6167 |`--buffers-pid` (user space tracing domain only)
6170 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6171 instead of the default per-user buffering scheme.
6173 |+--subbuf-size=__SIZE__+
6176 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6177 either for each Unix user (default), or for each instrumented process.
6179 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6181 |+--num-subbuf=__COUNT__+
6184 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6185 for each Unix user (default), or for each instrumented process.
6187 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6189 |+--tracefile-size=__SIZE__+
6192 Set the maximum size of each trace file that this channel writes within
6193 a stream to +__SIZE__+ bytes instead of no maximum.
6195 See <<tracefile-rotation,Trace file count and size>>.
6197 |+--tracefile-count=__COUNT__+
6200 Limit the number of trace files that this channel creates to
6201 +__COUNT__+ channels instead of no limit.
6203 See <<tracefile-rotation,Trace file count and size>>.
6205 |+--switch-timer=__PERIODUS__+
6208 Set the <<channel-switch-timer,switch timer period>>
6209 to +__PERIODUS__+{nbsp}µs.
6211 |+--read-timer=__PERIODUS__+
6214 Set the <<channel-read-timer,read timer period>>
6215 to +__PERIODUS__+{nbsp}µs.
6217 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6220 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6224 You can only create a channel in the Linux kernel and user space
6225 <<domain,tracing domains>>: other tracing domains have their own channel
6226 created on the fly when <<enabling-disabling-events,creating event
6231 Because of a current LTTng limitation, you must create all channels
6232 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6233 tracing session, that is, before the first time you run
6236 Since LTTng automatically creates a default channel when you use the
6237 man:lttng-enable-event(1) command with a specific tracing domain, you
6238 cannot, for example, create a Linux kernel event rule, start tracing,
6239 and then create a user space event rule, because no user space channel
6240 exists yet and it's too late to create one.
6242 For this reason, make sure to configure your channels properly
6243 before starting the tracers for the first time!
6246 The following examples show how you can combine the previous
6247 command-line options to create simple to more complex channels.
6249 .Create a Linux kernel channel with default attributes.
6253 lttng enable-channel --kernel my-channel
6257 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6261 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6262 --buffers-pid my-channel
6266 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6270 lttng enable-channel --kernel --tracefile-count=8 \
6271 --tracefile-size=4194304 my-channel
6275 .Create a user space channel in overwrite (or _flight recorder_) mode.
6279 lttng enable-channel --userspace --overwrite my-channel
6283 You can <<enabling-disabling-events,create>> the same event rule in
6284 two different channels:
6288 lttng enable-event --userspace --channel=my-channel app:tp
6289 lttng enable-event --userspace --channel=other-channel app:tp
6292 If both channels are enabled, when a tracepoint named `app:tp` is
6293 reached, LTTng records two events, one for each channel.
6297 === Disable a channel
6299 To disable a specific channel that you <<enabling-disabling-channels,created>>
6300 previously, use the man:lttng-disable-channel(1) command.
6302 .Disable a specific Linux kernel channel.
6306 lttng disable-channel --kernel my-channel
6310 The state of a channel precedes the individual states of event rules
6311 attached to it: event rules which belong to a disabled channel, even if
6312 they are enabled, are also considered disabled.
6316 === Add context fields to a channel
6318 Event record fields in trace files provide important information about
6319 events that occured previously, but sometimes some external context may
6320 help you solve a problem faster. Examples of context fields are:
6322 * The **process ID**, **thread ID**, **process name**, and
6323 **process priority** of the thread in which the event occurs.
6324 * The **hostname** of the system on which the event occurs.
6325 * The current values of many possible **performance counters** using
6327 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6329 ** Branch instructions, misses, and loads.
6331 * Any context defined at the application level (supported for the
6332 JUL and log4j <<domain,tracing domains>>).
6334 To get the full list of available context fields, see
6335 `lttng add-context --list`. Some context fields are reserved for a
6336 specific <<domain,tracing domain>> (Linux kernel or user space).
6338 You add context fields to <<channel,channels>>. All the events
6339 that a channel with added context fields records contain those fields.
6341 To add context fields to one or all the channels of a given tracing
6344 * Use the man:lttng-add-context(1) command.
6346 .Add context fields to all the channels of the current tracing session.
6348 The following command line adds the virtual process identifier and
6349 the per-thread CPU cycles count fields to all the user space channels
6350 of the current tracing session.
6354 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6358 .Add a context field to a specific channel.
6360 The following command line adds the thread identifier context field
6361 to the Linux kernel channel named `my-channel` in the current
6366 lttng add-context --kernel --channel=my-channel --type=tid
6370 .Add an application-specific context field to a specific channel.
6372 The following command line adds the `cur_msg_id` context field of the
6373 `retriever` context retriever for all the instrumented
6374 <<java-application,Java applications>> recording <<event,event records>>
6375 in the channel named `my-channel`:
6379 lttng add-context --kernel --channel=my-channel \
6380 --type='$app:retriever:cur_msg_id'
6383 IMPORTANT: Make sure to always quote the `$` character when you
6384 use man:lttng-add-context(1) from a shell.
6387 NOTE: You cannot remove context fields from a channel once you add it.
6392 === Track process IDs
6394 It's often useful to allow only specific process IDs (PIDs) to emit
6395 events. For example, you may wish to record all the system calls made by
6396 a given process (à la http://linux.die.net/man/1/strace[strace]).
6398 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6399 purpose. Both commands operate on a whitelist of process IDs. You _add_
6400 entries to this whitelist with the man:lttng-track(1) command and remove
6401 entries with the man:lttng-untrack(1) command. Any process which has one
6402 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6403 an enabled <<event,event rule>>.
6405 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6406 process with a given tracked ID exit and another process be given this
6407 ID, then the latter would also be allowed to emit events.
6409 .Track and untrack process IDs.
6411 For the sake of the following example, assume the target system has 16
6415 <<creating-destroying-tracing-sessions,create a tracing session>>,
6416 the whitelist contains all the possible PIDs:
6419 .All PIDs are tracked.
6420 image::track-all.png[]
6422 When the whitelist is full and you use the man:lttng-track(1) command to
6423 specify some PIDs to track, LTTng first clears the whitelist, then it
6424 tracks the specific PIDs. After:
6428 lttng track --pid=3,4,7,10,13
6434 .PIDs 3, 4, 7, 10, and 13 are tracked.
6435 image::track-3-4-7-10-13.png[]
6437 You can add more PIDs to the whitelist afterwards:
6441 lttng track --pid=1,15,16
6447 .PIDs 1, 15, and 16 are added to the whitelist.
6448 image::track-1-3-4-7-10-13-15-16.png[]
6450 The man:lttng-untrack(1) command removes entries from the PID tracker's
6451 whitelist. Given the previous example, the following command:
6455 lttng untrack --pid=3,7,10,13
6458 leads to this whitelist:
6461 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6462 image::track-1-4-15-16.png[]
6464 LTTng can track all possible PIDs again using the opt:track(1):--all
6469 lttng track --pid --all
6472 The result is, again:
6475 .All PIDs are tracked.
6476 image::track-all.png[]
6479 .Track only specific PIDs
6481 A very typical use case with PID tracking is to start with an empty
6482 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6483 then add PIDs manually while tracers are active. You can accomplish this
6484 by using the opt:lttng-untrack(1):--all option of the
6485 man:lttng-untrack(1) command to clear the whitelist after you
6486 <<creating-destroying-tracing-sessions,create a tracing session>>:
6490 lttng untrack --pid --all
6496 .No PIDs are tracked.
6497 image::untrack-all.png[]
6499 If you trace with this whitelist configuration, the tracer records no
6500 events for this <<domain,tracing domain>> because no processes are
6501 tracked. You can use the man:lttng-track(1) command as usual to track
6502 specific PIDs, for example:
6506 lttng track --pid=6,11
6512 .PIDs 6 and 11 are tracked.
6513 image::track-6-11.png[]
6518 [[saving-loading-tracing-session]]
6519 === Save and load tracing session configurations
6521 Configuring a <<tracing-session,tracing session>> can be long. Some of
6522 the tasks involved are:
6524 * <<enabling-disabling-channels,Create channels>> with
6525 specific attributes.
6526 * <<adding-context,Add context fields>> to specific channels.
6527 * <<enabling-disabling-events,Create event rules>> with specific log
6528 level and filter conditions.
6530 If you use LTTng to solve real world problems, chances are you have to
6531 record events using the same tracing session setup over and over,
6532 modifying a few variables each time in your instrumented program
6533 or environment. To avoid constant tracing session reconfiguration,
6534 the man:lttng(1) command-line tool can save and load tracing session
6535 configurations to/from XML files.
6537 To save a given tracing session configuration:
6539 * Use the man:lttng-save(1) command:
6544 lttng save my-session
6548 Replace `my-session` with the name of the tracing session to save.
6550 LTTng saves tracing session configurations to
6551 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6552 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6553 the opt:lttng-save(1):--output-path option to change this destination
6556 LTTng saves all configuration parameters, for example:
6558 * The tracing session name.
6559 * The trace data output path.
6560 * The channels with their state and all their attributes.
6561 * The context fields you added to channels.
6562 * The event rules with their state, log level and filter conditions.
6564 To load a tracing session:
6566 * Use the man:lttng-load(1) command:
6571 lttng load my-session
6575 Replace `my-session` with the name of the tracing session to load.
6577 When LTTng loads a configuration, it restores your saved tracing session
6578 as if you just configured it manually.
6580 See man:lttng(1) for the complete list of command-line options. You
6581 can also save and load all many sessions at a time, and decide in which
6582 directory to output the XML files.
6585 [[sending-trace-data-over-the-network]]
6586 === Send trace data over the network
6588 LTTng can send the recorded trace data to a remote system over the
6589 network instead of writing it to the local file system.
6591 To send the trace data over the network:
6593 . On the _remote_ system (which can also be the target system),
6594 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6603 . On the _target_ system, create a tracing session configured to
6604 send trace data over the network:
6609 lttng create my-session --set-url=net://remote-system
6613 Replace `remote-system` by the host name or IP address of the
6614 remote system. See man:lttng-create(1) for the exact URL format.
6616 . On the target system, use the man:lttng(1) command-line tool as usual.
6617 When tracing is active, the target's consumer daemon sends sub-buffers
6618 to the relay daemon running on the remote system intead of flushing
6619 them to the local file system. The relay daemon writes the received
6620 packets to the local file system.
6622 The relay daemon writes trace files to
6623 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6624 +__hostname__+ is the host name of the target system and +__session__+
6625 is the tracing session name. Note that the env:LTTNG_HOME environment
6626 variable defaults to `$HOME` if not set. Use the
6627 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6628 trace files to another base directory.
6633 === View events as LTTng emits them (noch:{LTTng} live)
6635 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6636 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6637 display events as LTTng emits them on the target system while tracing is
6640 The relay daemon creates a _tee_: it forwards the trace data to both
6641 the local file system and to connected live viewers:
6644 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6649 . On the _target system_, create a <<tracing-session,tracing session>>
6655 lttng create --live my-session
6659 This spawns a local relay daemon.
6661 . Start the live viewer and configure it to connect to the relay
6662 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6667 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6674 * `hostname` with the host name of the target system.
6675 * `my-session` with the name of the tracing session to view.
6678 . Configure the tracing session as usual with the man:lttng(1)
6679 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6681 You can list the available live tracing sessions with Babeltrace:
6685 babeltrace --input-format=lttng-live net://localhost
6688 You can start the relay daemon on another system. In this case, you need
6689 to specify the relay daemon's URL when you create the tracing session
6690 with the opt:lttng-create(1):--set-url option. You also need to replace
6691 `localhost` in the procedure above with the host name of the system on
6692 which the relay daemon is running.
6694 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6695 command-line options.
6699 [[taking-a-snapshot]]
6700 === Take a snapshot of the current sub-buffers of a tracing session
6702 The normal behavior of LTTng is to append full sub-buffers to growing
6703 trace data files. This is ideal to keep a full history of the events
6704 that occurred on the target system, but it can
6705 represent too much data in some situations. For example, you may wish
6706 to trace your application continuously until some critical situation
6707 happens, in which case you only need the latest few recorded
6708 events to perform the desired analysis, not multi-gigabyte trace files.
6710 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6711 current sub-buffers of a given <<tracing-session,tracing session>>.
6712 LTTng can write the snapshot to the local file system or send it over
6717 . Create a tracing session in _snapshot mode_:
6722 lttng create --snapshot my-session
6726 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6727 <<channel,channels>> created in this mode is automatically set to
6728 _overwrite_ (flight recorder mode).
6730 . Configure the tracing session as usual with the man:lttng(1)
6731 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6733 . **Optional**: When you need to take a snapshot,
6734 <<basic-tracing-session-control,stop tracing>>.
6736 You can take a snapshot when the tracers are active, but if you stop
6737 them first, you are sure that the data in the sub-buffers does not
6738 change before you actually take the snapshot.
6745 lttng snapshot record --name=my-first-snapshot
6749 LTTng writes the current sub-buffers of all the current tracing
6750 session's channels to trace files on the local file system. Those trace
6751 files have `my-first-snapshot` in their name.
6753 There is no difference between the format of a normal trace file and the
6754 format of a snapshot: viewers of LTTng traces also support LTTng
6757 By default, LTTng writes snapshot files to the path shown by
6758 `lttng snapshot list-output`. You can change this path or decide to send
6759 snapshots over the network using either:
6761 . An output path or URL that you specify when you create the
6763 . An snapshot output path or URL that you add using
6764 `lttng snapshot add-output`
6765 . An output path or URL that you provide directly to the
6766 `lttng snapshot record` command.
6768 Method 3 overrides method 2, which overrides method 1. When you
6769 specify a URL, a relay daemon must listen on a remote system (see
6770 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6775 === Use the machine interface
6777 With any command of the man:lttng(1) command-line tool, you can set the
6778 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6779 XML machine interface output, for example:
6783 lttng --mi=xml enable-event --kernel --syscall open
6786 A schema definition (XSD) is
6787 https://github.com/lttng/lttng-tools/blob/stable-2.8/src/common/mi-lttng-3.0.xsd[available]
6788 to ease the integration with external tools as much as possible.
6792 [[metadata-regenerate]]
6793 === Regenerate the metadata of an LTTng trace
6795 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6796 data stream files and a metadata file. This metadata file contains,
6797 amongst other things, information about the offset of the clock sources
6798 used to timestamp <<event,event records>> when tracing.
6800 If, once a <<tracing-session,tracing session>> is
6801 <<basic-tracing-session-control,started>>, a major
6802 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6803 happens, the trace's clock offset also needs to be updated. You
6804 can use the man:lttng-metadata(1) command to do so.
6806 The main use case of this command is to allow a system to boot with
6807 an incorrect wall time and trace it with LTTng before its wall time
6808 is corrected. Once the system is known to be in a state where its
6809 wall time is correct, it can run `lttng metadata regenerate`.
6811 To regenerate the metadata of an LTTng trace:
6813 * Use the `regenerate` action of the man:lttng-metadata(1) command:
6818 lttng metadata regenerate
6824 `lttng metadata regenerate` has the following limitations:
6826 * Tracing session <<creating-destroying-tracing-sessions,created>>
6828 * User space <<channel,channels>>, if any, using
6829 <<channel-buffering-schemes,per-user buffering>>.
6834 [[persistent-memory-file-systems]]
6835 === Record trace data on persistent memory file systems
6837 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6838 (NVRAM) is random-access memory that retains its information when power
6839 is turned off (non-volatile). Systems with such memory can store data
6840 structures in RAM and retrieve them after a reboot, without flushing
6841 to typical _storage_.
6843 Linux supports NVRAM file systems thanks to either
6844 http://pramfs.sourceforge.net/[PRAMFS] or
6845 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6846 (requires Linux 4.1+).
6848 This section does not describe how to operate such file systems;
6849 we assume that you have a working persistent memory file system.
6851 When you create a <<tracing-session,tracing session>>, you can specify
6852 the path of the shared memory holding the sub-buffers. If you specify a
6853 location on an NVRAM file system, then you can retrieve the latest
6854 recorded trace data when the system reboots after a crash.
6856 To record trace data on a persistent memory file system and retrieve the
6857 trace data after a system crash:
6859 . Create a tracing session with a sub-buffer shared memory path located
6860 on an NVRAM file system:
6865 lttng create --shm-path=/path/to/shm
6869 . Configure the tracing session as usual with the man:lttng(1)
6870 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6872 . After a system crash, use the man:lttng-crash(1) command-line tool to
6873 view the trace data recorded on the NVRAM file system:
6878 lttng-crash /path/to/shm
6882 The binary layout of the ring buffer files is not exactly the same as
6883 the trace files layout. This is why you need to use man:lttng-crash(1)
6884 instead of your preferred trace viewer directly.
6886 To convert the ring buffer files to LTTng trace files:
6888 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6893 lttng-crash --extract=/path/to/trace /path/to/shm
6901 [[lttng-modules-ref]]
6902 === noch:{LTTng-modules}
6905 [[lttng-modules-tp-fields]]
6906 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6908 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6909 tracepoint fields, which must be listed within `TP_FIELDS()` in
6910 `LTTNG_TRACEPOINT_EVENT()`, are:
6912 [role="func-desc growable",cols="asciidoc,asciidoc"]
6913 .Available macros to define LTTng-modules tracepoint fields
6915 |Macro |Description and parameters
6918 +ctf_integer(__t__, __n__, __e__)+
6920 +ctf_integer_nowrite(__t__, __n__, __e__)+
6922 +ctf_user_integer(__t__, __n__, __e__)+
6924 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6926 Standard integer, displayed in base 10.
6929 Integer C type (`int`, `long`, `size_t`, ...).
6935 Argument expression.
6938 +ctf_integer_hex(__t__, __n__, __e__)+
6940 +ctf_user_integer_hex(__t__, __n__, __e__)+
6942 Standard integer, displayed in base 16.
6951 Argument expression.
6953 |+ctf_integer_oct(__t__, __n__, __e__)+
6955 Standard integer, displayed in base 8.
6964 Argument expression.
6967 +ctf_integer_network(__t__, __n__, __e__)+
6969 +ctf_user_integer_network(__t__, __n__, __e__)+
6971 Integer in network byte order (big-endian), displayed in base 10.
6980 Argument expression.
6983 +ctf_integer_network_hex(__t__, __n__, __e__)+
6985 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6987 Integer in network byte order, displayed in base 16.
6996 Argument expression.
6999 +ctf_string(__n__, __e__)+
7001 +ctf_string_nowrite(__n__, __e__)+
7003 +ctf_user_string(__n__, __e__)+
7005 +ctf_user_string_nowrite(__n__, __e__)+
7007 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7013 Argument expression.
7016 +ctf_array(__t__, __n__, __e__, __s__)+
7018 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7020 +ctf_user_array(__t__, __n__, __e__, __s__)+
7022 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7024 Statically-sized array of integers.
7027 Array element C type.
7033 Argument expression.
7039 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7041 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7043 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7045 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7047 Statically-sized array of bits.
7049 The type of +__e__+ must be an integer type. +__s__+ is the number
7050 of elements of such type in +__e__+, not the number of bits.
7053 Array element C type.
7059 Argument expression.
7065 +ctf_array_text(__t__, __n__, __e__, __s__)+
7067 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7069 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7071 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7073 Statically-sized array, printed as text.
7075 The string does not need to be null-terminated.
7078 Array element C type (always `char`).
7084 Argument expression.
7090 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7092 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7094 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7096 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7098 Dynamically-sized array of integers.
7100 The type of +__E__+ must be unsigned.
7103 Array element C type.
7109 Argument expression.
7112 Length expression C type.
7118 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7120 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7122 Dynamically-sized array of integers, displayed in base 16.
7124 The type of +__E__+ must be unsigned.
7127 Array element C type.
7133 Argument expression.
7136 Length expression C type.
7141 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7143 Dynamically-sized array of integers in network byte order (big-endian),
7144 displayed in base 10.
7146 The type of +__E__+ must be unsigned.
7149 Array element C type.
7155 Argument expression.
7158 Length expression C type.
7164 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7166 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7168 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7170 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7172 Dynamically-sized array of bits.
7174 The type of +__e__+ must be an integer type. +__s__+ is the number
7175 of elements of such type in +__e__+, not the number of bits.
7177 The type of +__E__+ must be unsigned.
7180 Array element C type.
7186 Argument expression.
7189 Length expression C type.
7195 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7197 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7199 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7201 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7203 Dynamically-sized array, displayed as text.
7205 The string does not need to be null-terminated.
7207 The type of +__E__+ must be unsigned.
7209 The behaviour is undefined if +__e__+ is `NULL`.
7212 Sequence element C type (always `char`).
7218 Argument expression.
7221 Length expression C type.
7227 Use the `_user` versions when the argument expression, `e`, is
7228 a user space address. In the cases of `ctf_user_integer*()` and
7229 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7232 The `_nowrite` versions omit themselves from the session trace, but are
7233 otherwise identical. This means the `_nowrite` fields won't be written
7234 in the recorded trace. Their primary purpose is to make some
7235 of the event context available to the
7236 <<enabling-disabling-events,event filters>> without having to
7237 commit the data to sub-buffers.
7243 Terms related to LTTng and to tracing in general:
7246 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7247 the cmd:babeltrace command, some libraries, and Python bindings.
7249 <<channel-buffering-schemes,buffering scheme>>::
7250 A layout of sub-buffers applied to a given channel.
7252 <<channel,channel>>::
7253 An entity which is responsible for a set of ring buffers.
7255 <<event,Event rules>> are always attached to a specific channel.
7258 A reference of time for a tracer.
7260 <<lttng-consumerd,consumer daemon>>::
7261 A process which is responsible for consuming the full sub-buffers
7262 and write them to a file system or send them over the network.
7264 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7265 mode in which the tracer _discards_ new event records when there's no
7266 sub-buffer space left to store them.
7269 The consequence of the execution of an instrumentation
7270 point, like a tracepoint that you manually place in some source code,
7271 or a Linux kernel KProbe.
7273 An event is said to _occur_ at a specific time. Different actions can
7274 be taken upon the occurance of an event, like record the event's payload
7277 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7278 The mechanism by which event records of a given channel are lost
7279 (not recorded) when there is no sub-buffer space left to store them.
7281 [[def-event-name]]event name::
7282 The name of an event, which is also the name of the event record.
7283 This is also called the _instrumentation point name_.
7286 A record, in a trace, of the payload of an event which occured.
7288 <<event,event rule>>::
7289 Set of conditions which must be satisfied for one or more occuring
7290 events to be recorded.
7292 `java.util.logging`::
7294 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7296 <<instrumenting,instrumentation>>::
7297 The use of LTTng probes to make a piece of software traceable.
7299 instrumentation point::
7300 A point in the execution path of a piece of software that, when
7301 reached by this execution, can emit an event.
7303 instrumentation point name::
7304 See _<<def-event-name,event name>>_.
7307 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7308 developed by the Apache Software Foundation.
7311 Level of severity of a log statement or user space
7312 instrumentation point.
7315 The _Linux Trace Toolkit: next generation_ project.
7317 <<lttng-cli,cmd:lttng>>::
7318 A command-line tool provided by the LTTng-tools project which you
7319 can use to send and receive control messages to and from a
7323 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7324 which is a set of analyzing programs that are used to obtain a
7325 higher level view of an LTTng trace.
7327 cmd:lttng-consumerd::
7328 The name of the consumer daemon program.
7331 A utility provided by the LTTng-tools project which can convert
7332 ring buffer files (usually
7333 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7336 LTTng Documentation::
7339 <<lttng-live,LTTng live>>::
7340 A communication protocol between the relay daemon and live viewers
7341 which makes it possible to see events "live", as they are received by
7344 <<lttng-modules,LTTng-modules>>::
7345 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7346 which contains the Linux kernel modules to make the Linux kernel
7347 instrumentation points available for LTTng tracing.
7350 The name of the relay daemon program.
7352 cmd:lttng-sessiond::
7353 The name of the session daemon program.
7356 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7357 contains the various programs and libraries used to
7358 <<controlling-tracing,control tracing>>.
7360 <<lttng-ust,LTTng-UST>>::
7361 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7362 contains libraries to instrument user applications.
7364 <<lttng-ust-agents,LTTng-UST Java agent>>::
7365 A Java package provided by the LTTng-UST project to allow the
7366 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7369 <<lttng-ust-agents,LTTng-UST Python agent>>::
7370 A Python package provided by the LTTng-UST project to allow the
7371 LTTng instrumentation of Python logging statements.
7373 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7374 The event loss mode in which new event records overwrite older
7375 event records when there's no sub-buffer space left to store them.
7377 <<channel-buffering-schemes,per-process buffering>>::
7378 A buffering scheme in which each instrumented process has its own
7379 sub-buffers for a given user space channel.
7381 <<channel-buffering-schemes,per-user buffering>>::
7382 A buffering scheme in which all the processes of a Unix user share the
7383 same sub-buffer for a given user space channel.
7385 <<lttng-relayd,relay daemon>>::
7386 A process which is responsible for receiving the trace data sent by
7387 a distant consumer daemon.
7390 A set of sub-buffers.
7392 <<lttng-sessiond,session daemon>>::
7393 A process which receives control commands from you and orchestrates
7394 the tracers and various LTTng daemons.
7396 <<taking-a-snapshot,snapshot>>::
7397 A copy of the current data of all the sub-buffers of a given tracing
7398 session, saved as trace files.
7401 One part of an LTTng ring buffer which contains event records.
7404 The time information attached to an event when it is emitted.
7407 A set of files which are the concatenations of one or more
7408 flushed sub-buffers.
7411 The action of recording the events emitted by an application
7412 or by a system, or to initiate such recording by controlling
7416 The http://tracecompass.org[Trace Compass] project and application.
7419 An instrumentation point using the tracepoint mechanism of the Linux
7420 kernel or of LTTng-UST.
7422 tracepoint definition::
7423 The definition of a single tracepoint.
7426 The name of a tracepoint.
7428 tracepoint provider::
7429 A set of functions providing tracepoints to an instrumented user
7432 Not to be confused with a _tracepoint provider package_: many tracepoint
7433 providers can exist within a tracepoint provider package.
7435 tracepoint provider package::
7436 One or more tracepoint providers compiled as an object file or as
7440 A software which records emitted events.
7442 <<domain,tracing domain>>::
7443 A namespace for event sources.
7445 <<tracing-group,tracing group>>::
7446 The Unix group in which a Unix user can be to be allowed to trace the
7449 <<tracing-session,tracing session>>::
7450 A stateful dialogue between you and a <<lttng-sessiond,session
7454 An application running in user space, as opposed to a Linux kernel
7455 module, for example.