1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 * **Tracing control**:
78 ** You can attach <<java-application-context,Java application-specific
79 context fields>> to a <<channel,channel>> with the
80 man:lttng-add-context(1) command:
85 lttng add-context --jul --type='$app.retriever:cur_msg_id'
89 Here, `$app` is the prefix of all application-specific context fields,
90 `retriever` names a _context information retriever_ defined at the
91 application level, and `cur_msg_id` names a context field read from this
94 Both the `java.util.logging` and Apache log4j <<domain,tracing domains>>
97 ** You can use Java application-specific <<adding-context,context>>
98 fields in the <<enabling-disabling-events,filter expression>> of an
104 lttng enable-event --log4j my_logger \
105 --filter='$app.retriever:cur_msg_id == 23'
109 ** New `lttng status` command which is the equivalent of +lttng list
110 __CUR__+, where +__CUR__+ is the name of the current
111 <<tracing-session,tracing session>>.
113 See man:lttng-status(1).
115 ** New `lttng metadata regenerate` command to regenerate the metadata
116 file of an LTTng trace at any moment. This command is meant to be
117 used to resample the wall time following a major
118 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
119 so that a system which boots with an incorrect wall time can be
120 traced before its wall time is NTP-corrected.
122 See man:lttng-metadata(1).
124 ** New command-line interface warnings when <<event,event records>> or
125 whole sub-buffers are
126 <<channel-overwrite-mode-vs-discard-mode,lost>>. The warning messages
127 are printed when a <<tracing-session,tracing session>> is
128 <<basic-tracing-session-control,stopped>> (man:lttng-stop(1)
131 * **User space tracing**:
132 ** Shared object base address dump in order to map <<event,event
133 records>> to original source location (file and line number).
135 If you attach the `ip` and `vpid` <<adding-context,context fields>> to a
136 user space <<channel,channel>> and if you use the
137 <<liblttng-ust-dl,path:{liblttng-ust-dl.so} helper>>, you can retrieve
138 the source location where a given event record was generated.
140 The http://diamon.org/babeltrace/[Babeltrace] trace viewer supports this
141 state dump and those context fields since version 1.4 to print the
142 source location of a given event record. http://tracecompass.org/[Trace
143 Compass] also supports this since version 2.0.
145 ** A <<java-application,Java application>> which uses
146 `java.util.logging` now adds an LTTng-UST log handler to the desired
149 The previous workflow was to initialize the LTTng-UST Java agent
150 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
151 an LTTng-UST log handler to the root loggers.
153 ** A <<java-application,Java application>> which uses Apache log4j now
154 adds an LTTng-UST log appender to the desired log4j loggers.
156 The previous workflow was to initialize the LTTng-UST Java agent
157 by calling `LTTngAgent.getLTTngAgent()`. This had the effect of adding
158 an LTTng-UST appender to the root loggers.
160 ** Any <<java-application,Java application>> can provide
161 <<java-application-context,dynamic context fields>> while running
162 thanks to a new API provided by the <<lttng-ust-agents,LTTng-UST Java
163 agent>>. You can require LTTng to record specific context fields in
164 event records, and you can use them in the filter expression of
165 <<event,event rules>>.
167 * **Linux kernel tracing**:
168 ** The LTTng kernel modules can now be built into a Linux kernel image,
169 that is, not as loadable modules.
172 https://github.com/lttng/lttng-modules/blob/stable-{revision}/README.md#kernel-built-in-support[`README.md`]
175 ** New instrumentation:
176 *** ARM64 architecture support.
178 *** x86 `irq_vectors`.
179 ** New <<adding-context,context fields>>:
182 *** `need_reschedule`
183 *** `migratable` (specific to RT-Preempt)
184 ** Clock source plugin support for advanced cases where a custom source
185 of time is needed to timestamp LTTng event records.
187 See https://github.com/lttng/lttng-modules/blob/stable-{revision}/lttng-clock.h[`lttng-clock.h`]
188 for an overview of the small API.
191 ** The link:/man[man pages] of the man:lttng(1) command-line tool are
192 split into one man page per command (à la Git), for example:
197 man lttng-enable-event
201 You can also use the `--help` option of any man:lttng(1) command to
204 The content and formatting of all the LTTng man pages has improved
211 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
212 generation_ is a modern toolkit for tracing Linux systems and
213 applications. So your first question might be:
220 As the history of software engineering progressed and led to what
221 we now take for granted--complex, numerous and
222 interdependent software applications running in parallel on
223 sophisticated operating systems like Linux--the authors of such
224 components, software developers, began feeling a natural
225 urge to have tools that would ensure the robustness and good performance
226 of their masterpieces.
228 One major achievement in this field is, inarguably, the
229 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
230 an essential tool for developers to find and fix bugs. But even the best
231 debugger won't help make your software run faster, and nowadays, faster
232 software means either more work done by the same hardware, or cheaper
233 hardware for the same work.
235 A _profiler_ is often the tool of choice to identify performance
236 bottlenecks. Profiling is suitable to identify _where_ performance is
237 lost in a given software. The profiler outputs a profile, a statistical
238 summary of observed events, which you may use to discover which
239 functions took the most time to execute. However, a profiler won't
240 report _why_ some identified functions are the bottleneck. Bottlenecks
241 might only occur when specific conditions are met, conditions that are
242 sometimes impossible to capture by a statistical profiler, or impossible
243 to reproduce with an application altered by the overhead of an
244 event-based profiler. For a thorough investigation of software
245 performance issues, a history of execution is essential, with the
246 recorded values of variables and context fields you choose, and
247 with as little influence as possible on the instrumented software. This
248 is where tracing comes in handy.
250 _Tracing_ is a technique used to understand what goes on in a running
251 software system. The software used for tracing is called a _tracer_,
252 which is conceptually similar to a tape recorder. When recording,
253 specific instrumentation points placed in the software source code
254 generate events that are saved on a giant tape: a _trace_ file. You
255 can trace user applications and the operating system at the same time,
256 opening the possibility of resolving a wide range of problems that would
257 otherwise be extremely challenging.
259 Tracing is often compared to _logging_. However, tracers and loggers are
260 two different tools, serving two different purposes. Tracers are
261 designed to record much lower-level events that occur much more
262 frequently than log messages, often in the range of thousands per
263 second, with very little execution overhead. Logging is more appropriate
264 for a very high-level analysis of less frequent events: user accesses,
265 exceptional conditions (errors and warnings, for example), database
266 transactions, instant messaging communications, and such. Simply put,
267 logging is one of the many use cases that can be satisfied with tracing.
269 The list of recorded events inside a trace file can be read manually
270 like a log file for the maximum level of detail, but it is generally
271 much more interesting to perform application-specific analyses to
272 produce reduced statistics and graphs that are useful to resolve a
273 given problem. Trace viewers and analyzers are specialized tools
276 In the end, this is what LTTng is: a powerful, open source set of
277 tools to trace the Linux kernel and user applications at the same time.
278 LTTng is composed of several components actively maintained and
279 developed by its link:/community/#where[community].
282 [[lttng-alternatives]]
283 === Alternatives to noch:{LTTng}
285 Excluding proprietary solutions, a few competing software tracers
288 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
289 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
290 user scripts and is responsible for loading code into the
291 Linux kernel for further execution and collecting the outputted data.
292 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
293 subsystem in the Linux kernel in which a virtual machine can execute
294 programs passed from the user space to the kernel. You can attach
295 such programs to tracepoints and KProbes thanks to a system call, and
296 they can output data to the user space when executed thanks to
297 different mechanisms (pipe, VM register values, and eBPF maps, to name
299 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
300 is the de facto function tracer of the Linux kernel. Its user
301 interface is a set of special files in sysfs.
302 * https://perf.wiki.kernel.org/[perf] is
303 a performance analyzing tool for Linux which supports hardware
304 performance counters, tracepoints, as well as other counters and
305 types of probes. perf's controlling utility is the cmd:perf command
307 * http://linux.die.net/man/1/strace[strace]
308 is a command-line utility which records system calls made by a
309 user process, as well as signal deliveries and changes of process
310 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
311 to fulfill its function.
312 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
313 analyze Linux kernel events. You write scripts, or _chisels_ in
314 sysdig's jargon, in Lua and sysdig executes them while the system is
315 being traced or afterwards. sysdig's interface is the cmd:sysdig
316 command-line tool as well as the curses-based cmd:csysdig tool.
317 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
318 user space tracer which uses custom user scripts to produce plain text
319 traces. SystemTap converts the scripts to the C language, and then
320 compiles them as Linux kernel modules which are loaded to produce
321 trace data. SystemTap's primary user interface is the cmd:stap
324 The main distinctive features of LTTng is that it produces correlated
325 kernel and user space traces, as well as doing so with the lowest
326 overhead amongst other solutions. It produces trace files in the
327 http://diamon.org/ctf[CTF] format, a file format optimized
328 for the production and analyses of multi-gigabyte data.
330 LTTng is the result of more than 10 years of active open source
331 development by a community of passionate developers.
332 LTTng{nbsp}{revision} is currently available on major desktop and server
335 The main interface for tracing control is a single command-line tool
336 named cmd:lttng. The latter can create several tracing sessions, enable
337 and disable events on the fly, filter events efficiently with custom
338 user expressions, start and stop tracing, and much more. LTTng can
339 record the traces on the file system or send them over the network, and
340 keep them totally or partially. You can view the traces once tracing
341 becomes inactive or in real-time.
343 <<installing-lttng,Install LTTng now>> and
344 <<getting-started,start tracing>>!
350 **LTTng** is a set of software <<plumbing,components>> which interact to
351 <<instrumenting,instrument>> the Linux kernel and user applications, and
352 to <<controlling-tracing,control tracing>> (start and stop
353 tracing, enable and disable event rules, and the rest). Those
354 components are bundled into the following packages:
356 * **LTTng-tools**: Libraries and command-line interface to
358 * **LTTng-modules**: Linux kernel modules to instrument and
360 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
361 trace user applications.
363 Most distributions mark the LTTng-modules and LTTng-UST packages as
364 optional when installing LTTng-tools (which is always required). In the
365 following sections, we always provide the steps to install all three,
368 * You only need to install LTTng-modules if you intend to trace the
370 * You only need to install LTTng-UST if you intend to trace user
374 .Availability of LTTng{nbsp}{revision} for major Linux distributions.
376 |Distribution |Available in releases |Alternatives
379 |<<ubuntu,Ubuntu{nbsp}16.10 _Yakkety Yak_>>
380 |LTTng{nbsp}{revision} for Ubuntu{nbsp}14.04 _Trusty Tahr_
381 and Ubuntu{nbsp}16.04 _Xenial Xerus_:
382 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
384 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
385 other Ubuntu releases.
389 |LTTng{nbsp}{revision} for Fedora{nbsp}25 and Fedora{nbsp}26 (not
392 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
393 other Fedora releases.
396 |<<debian,Debian "stretch" (testing)>>
398 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
399 previous Debian releases.
403 |LTTng{nbsp}2.7 for openSUSE Leap{nbsp}42.1.
405 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
406 other openSUSE releases.
409 |Latest AUR packages.
413 |<<alpine-linux,Alpine Linux "edge">>
414 |LTTng{nbsp}{revision} for Alpine Linux{nbsp}3.5 (not released yet).
416 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
417 other Alpine Linux releases.
420 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
425 |LTTng{nbsp}{revision} for Buildroot{nbsp}2016.11 (not released yet).
427 LTTng{nbsp}2.7 for Buildroot{nbsp}2016.02, Buildroot{nbsp}2016.05,
428 and Buildroot{nbsp}2016.08.
430 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
431 other Buildroot releases.
433 |OpenEmbedded and Yocto
434 |<<oe-yocto,Yocto Project{nbsp}2.2 _Morty_>> (`openembedded-core` layer)
435 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
436 other OpenEmbedded releases.
441 === [[ubuntu-official-repositories]]Ubuntu
443 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.10 _Yakkety Yak_.
444 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
445 Stable{nbsp}{revision} PPA>>.
447 To install LTTng{nbsp}{revision} on Ubuntu 16.10{nbsp}_Yakkety Yak_:
449 . Install the main LTTng{nbsp}{revision} packages:
454 sudo apt-get install lttng-tools
455 sudo apt-get install lttng-modules-dkms
456 sudo apt-get install liblttng-ust-dev
460 . **If you need to instrument and trace
461 <<java-application,Java applications>>**, install the LTTng-UST
467 sudo apt-get install liblttng-ust-agent-java
471 . **If you need to instrument and trace
472 <<python-application,Python applications>>**, install the
473 LTTng-UST Python agent:
478 sudo apt-get install python3-lttngust
484 ==== noch:{LTTng} Stable {revision} PPA
486 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
487 Stable{nbsp}{revision} PPA] offers the latest stable
488 LTTng{nbsp}{revision} packages for:
490 * Ubuntu{nbsp}14.04 _Trusty Tahr_
491 * Ubuntu{nbsp}16.04 _Xenial Xerus_
493 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
495 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
501 sudo apt-add-repository ppa:lttng/stable-2.8
506 . Install the main LTTng{nbsp}{revision} packages:
511 sudo apt-get install lttng-tools
512 sudo apt-get install lttng-modules-dkms
513 sudo apt-get install liblttng-ust-dev
517 . **If you need to instrument and trace
518 <<java-application,Java applications>>**, install the LTTng-UST
524 sudo apt-get install liblttng-ust-agent-java
528 . **If you need to instrument and trace
529 <<python-application,Python applications>>**, install the
530 LTTng-UST Python agent:
535 sudo apt-get install python3-lttngust
543 To install LTTng{nbsp}{revision} on Debian "stretch" (testing):
545 . Install the main LTTng{nbsp}{revision} packages:
550 sudo apt-get install lttng-modules-dkms
551 sudo apt-get install liblttng-ust-dev
552 sudo apt-get install lttng-tools
556 . **If you need to instrument and trace <<java-application,Java
557 applications>>**, install the LTTng-UST Java agent:
562 sudo apt-get install liblttng-ust-agent-java
566 . **If you need to instrument and trace <<python-application,Python
567 applications>>**, install the LTTng-UST Python agent:
572 sudo apt-get install python3-lttngust
580 To install LTTng{nbsp}{revision} (tracing control and user space
581 tracing) on Alpine Linux "edge":
583 . Make sure your system is
584 https://wiki.alpinelinux.org/wiki/Edge[configured for "edge"].
585 . Enable the _testing_ repository by uncommenting the corresponding
586 line in path:{/etc/apk/repositories}.
587 . Add the LTTng packages:
592 sudo apk add lttng-tools
593 sudo apk add lttng-ust-dev
597 To install LTTng-modules{nbsp}{revision} (Linux kernel tracing support)
598 on Alpine Linux "edge":
600 . Add the vanilla Linux kernel:
605 apk add linux-vanilla linux-vanilla-dev
609 . Reboot with the vanilla Linux kernel.
610 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
616 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
617 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
618 cd lttng-modules-2.8.* &&
620 sudo make modules_install &&
627 === OpenEmbedded and Yocto
629 LTTng{nbsp}{revision} recipes are available in the
630 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
631 layer for Yocto Project{nbsp}2.2 _Morty_ under the following names:
637 With BitBake, the simplest way to include LTTng recipes in your target
638 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
641 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
646 . Select a machine and an image recipe.
647 . Click **Edit image recipe**.
648 . Under the **All recipes** tab, search for **lttng**.
649 . Check the desired LTTng recipes.
652 .Java and Python application instrumentation and tracing
654 If you need to instrument and trace <<java-application,Java
655 applications>> on openSUSE, you need to build and install
656 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
657 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
658 `--enable-java-agent-all` options to the `configure` script, depending
659 on which Java logging framework you use.
661 If you need to instrument and trace <<python-application,Python
662 applications>> on openSUSE, you need to build and install
663 LTTng-UST{nbsp}{revision} from source and pass the
664 `--enable-python-agent` option to the `configure` script.
668 [[enterprise-distributions]]
669 === RHEL, SUSE, and other enterprise distributions
671 To install LTTng on enterprise Linux distributions, such as Red Hat
672 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
673 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
676 [[building-from-source]]
677 === Build from source
679 To build and install LTTng{nbsp}{revision} from source:
681 . Using your distribution's package manager, or from source, install
682 the following dependencies of LTTng-tools and LTTng-UST:
685 * https://sourceforge.net/projects/libuuid/[libuuid]
686 * http://directory.fsf.org/wiki/Popt[popt]
687 * http://liburcu.org/[Userspace RCU]
688 * http://www.xmlsoft.org/[libxml2]
691 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
697 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
698 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
699 cd lttng-modules-2.8.* &&
701 sudo make modules_install &&
706 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
712 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
713 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
714 cd lttng-ust-2.8.* &&
724 .Java and Python application tracing
726 If you need to instrument and trace <<java-application,Java
727 applications>>, pass the `--enable-java-agent-jul`,
728 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
729 `configure` script, depending on which Java logging framework you use.
731 If you need to instrument and trace <<python-application,Python
732 applications>>, pass the `--enable-python-agent` option to the
733 `configure` script. You can set the `PYTHON` environment variable to the
734 path to the Python interpreter for which to install the LTTng-UST Python
742 By default, LTTng-UST libraries are installed to
743 dir:{/usr/local/lib}, which is the de facto directory in which to
744 keep self-compiled and third-party libraries.
746 When <<building-tracepoint-providers-and-user-application,linking an
747 instrumented user application with `liblttng-ust`>>:
749 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
751 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
752 man:gcc(1), man:g++(1), or man:clang(1).
756 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
762 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
763 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
764 cd lttng-tools-2.8.* &&
772 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
773 previous steps automatically for a given version of LTTng and confine
774 the installed files in a specific directory. This can be useful to test
775 LTTng without installing it on your system.
781 This is a short guide to get started quickly with LTTng kernel and user
784 Before you follow this guide, make sure to <<installing-lttng,install>>
787 This tutorial walks you through the steps to:
789 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
790 . <<tracing-your-own-user-application,Trace a user application>> written
792 . <<viewing-and-analyzing-your-traces,View and analyze the
796 [[tracing-the-linux-kernel]]
797 === Trace the Linux kernel
799 The following command lines start with cmd:sudo because you need root
800 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
801 your Unix user is a member of the <<lttng-sessiond,tracing group>>.
803 . Create a <<tracing-session,tracing session>>:
808 sudo lttng create my-kernel-session
812 . List the available kernel tracepoints and system calls:
821 . Create an <<event,event rule>> which matches the desired event names,
822 for example `sched_switch` and `sched_process_fork`:
827 sudo lttng enable-event --kernel sched_switch,sched_process_fork
831 You can also create an event rule which _matches_ all the Linux kernel
832 tracepoints (this will generate a lot of data when tracing):
837 sudo lttng enable-event --kernel --all
850 . Do some operation on your system for a few seconds. For example,
851 load a website, or list the files of a directory.
852 . Stop tracing and destroy the tracing session:
862 The man:lttng-destroy(1) command does not destroy the trace data; it
863 only destroys the state of the tracing session.
865 By default, LTTng saves the traces in
866 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
867 where +__name__+ is the tracing session name. Note that the
868 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
870 See <<viewing-and-analyzing-your-traces,View and analyze the
871 recorded events>> to view the recorded events.
874 [[tracing-your-own-user-application]]
875 === Trace a user application
877 This section steps you through a simple example to trace a
878 _Hello world_ program written in C.
880 To create the traceable user application:
882 . Create the tracepoint provider header file, which defines the
883 tracepoints and the events they can generate:
889 #undef TRACEPOINT_PROVIDER
890 #define TRACEPOINT_PROVIDER hello_world
892 #undef TRACEPOINT_INCLUDE
893 #define TRACEPOINT_INCLUDE "./hello-tp.h"
895 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
898 #include <lttng/tracepoint.h>
908 ctf_string(my_string_field, my_string_arg)
909 ctf_integer(int, my_integer_field, my_integer_arg)
913 #endif /* _HELLO_TP_H */
915 #include <lttng/tracepoint-event.h>
919 . Create the tracepoint provider package source file:
925 #define TRACEPOINT_CREATE_PROBES
926 #define TRACEPOINT_DEFINE
928 #include "hello-tp.h"
932 . Build the tracepoint provider package:
937 gcc -c -I. hello-tp.c
941 . Create the _Hello World_ application source file:
948 #include "hello-tp.h"
950 int main(int argc, char *argv[])
954 puts("Hello, World!\nPress Enter to continue...");
957 * The following getchar() call is only placed here for the purpose
958 * of this demonstration, to pause the application in order for
959 * you to have time to list its tracepoints. It is not
965 * A tracepoint() call.
967 * Arguments, as defined in hello-tp.h:
969 * 1. Tracepoint provider name (required)
970 * 2. Tracepoint name (required)
971 * 3. my_integer_arg (first user-defined argument)
972 * 4. my_string_arg (second user-defined argument)
974 * Notice the tracepoint provider and tracepoint names are
975 * NOT strings: they are in fact parts of variables that the
976 * macros in hello-tp.h create.
978 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
980 for (x = 0; x < argc; ++x) {
981 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
984 puts("Quitting now!");
985 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
992 . Build the application:
1001 . Link the application with the tracepoint provider package,
1002 `liblttng-ust`, and `libdl`:
1007 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1011 Here's the whole build process:
1014 .User space tracing tutorial's build steps.
1015 image::ust-flow.png[]
1017 To trace the user application:
1019 . Run the application with a few arguments:
1024 ./hello world and beyond
1033 Press Enter to continue...
1037 . Start an LTTng <<lttng-sessiond,session daemon>>:
1042 lttng-sessiond --daemonize
1046 Note that a session daemon might already be running, for example as
1047 a service that the distribution's service manager started.
1049 . List the available user space tracepoints:
1054 lttng list --userspace
1058 You see the `hello_world:my_first_tracepoint` tracepoint listed
1059 under the `./hello` process.
1061 . Create a <<tracing-session,tracing session>>:
1066 lttng create my-user-space-session
1070 . Create an <<event,event rule>> which matches the
1071 `hello_world:my_first_tracepoint` event name:
1076 lttng enable-event --userspace hello_world:my_first_tracepoint
1089 . Go back to the running `hello` application and press Enter. The
1090 program executes all `tracepoint()` instrumentation points and exits.
1091 . Stop tracing and destroy the tracing session:
1101 The man:lttng-destroy(1) command does not destroy the trace data; it
1102 only destroys the state of the tracing session.
1104 By default, LTTng saves the traces in
1105 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1106 where +__name__+ is the tracing session name. Note that the
1107 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1109 See <<viewing-and-analyzing-your-traces,View and analyze the
1110 recorded events>> to view the recorded events.
1113 [[viewing-and-analyzing-your-traces]]
1114 === View and analyze the recorded events
1116 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1117 kernel>> and <<tracing-your-own-user-application,Trace a user
1118 application>> tutorials, you can inspect the recorded events.
1120 Many tools are available to read LTTng traces:
1122 * **cmd:babeltrace** is a command-line utility which converts trace
1123 formats; it supports the format that LTTng produces, CTF, as well as a
1124 basic text output which can be ++grep++ed. The cmd:babeltrace command
1125 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1126 * Babeltrace also includes
1127 **https://www.python.org/[Python] bindings** so
1128 that you can easily open and read an LTTng trace with your own script,
1129 benefiting from the power of Python.
1130 * http://tracecompass.org/[**Trace Compass**]
1131 is a graphical user interface for viewing and analyzing any type of
1132 logs or traces, including LTTng's.
1133 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1134 project which includes many high-level analyses of LTTng kernel
1135 traces, like scheduling statistics, interrupt frequency distribution,
1136 top CPU usage, and more.
1138 NOTE: This section assumes that the traces recorded during the previous
1139 tutorials were saved to their default location, in the
1140 dir:{$LTTNG_HOME/lttng-traces} directory. Note that the env:LTTNG_HOME
1141 environment variable defaults to `$HOME` if not set.
1144 [[viewing-and-analyzing-your-traces-bt]]
1145 ==== Use the cmd:babeltrace command-line tool
1147 The simplest way to list all the recorded events of a trace is to pass
1148 its path to cmd:babeltrace with no options:
1152 babeltrace ~/lttng-traces/my-user-space-session*
1155 cmd:babeltrace finds all traces recursively within the given path and
1156 prints all their events, merging them in chronological order.
1158 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1163 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_
1166 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1167 count the recorded events:
1171 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_read | wc --lines
1175 [[viewing-and-analyzing-your-traces-bt-python]]
1176 ==== Use the Babeltrace Python bindings
1178 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1179 is useful to isolate events by simple matching using man:grep(1) and
1180 similar utilities. However, more elaborate filters, such as keeping only
1181 event records with a field value falling within a specific range, are
1182 not trivial to write using a shell. Moreover, reductions and even the
1183 most basic computations involving multiple event records are virtually
1184 impossible to implement.
1186 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1187 to read the event records of an LTTng trace sequentially and compute the
1188 desired information.
1190 The following script accepts an LTTng Linux kernel trace path as its
1191 first argument and prints the short names of the top 5 running processes
1192 on CPU 0 during the whole trace:
1197 from collections import Counter
1203 if len(sys.argv) != 2:
1204 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1205 print(msg, file=sys.stderr)
1208 # A trace collection contains one or more traces
1209 col = babeltrace.TraceCollection()
1211 # Add the trace provided by the user (LTTng traces always have
1213 if col.add_trace(sys.argv[1], 'ctf') is None:
1214 raise RuntimeError('Cannot add trace')
1216 # This counter dict contains execution times:
1218 # task command name -> total execution time (ns)
1219 exec_times = Counter()
1221 # This contains the last `sched_switch` timestamp
1225 for event in col.events:
1226 # Keep only `sched_switch` events
1227 if event.name != 'sched_switch':
1230 # Keep only events which happened on CPU 0
1231 if event['cpu_id'] != 0:
1235 cur_ts = event.timestamp
1241 # Previous task command (short) name
1242 prev_comm = event['prev_comm']
1244 # Initialize entry in our dict if not yet done
1245 if prev_comm not in exec_times:
1246 exec_times[prev_comm] = 0
1248 # Compute previous command execution time
1249 diff = cur_ts - last_ts
1251 # Update execution time of this command
1252 exec_times[prev_comm] += diff
1254 # Update last timestamp
1258 for name, ns in exec_times.most_common(5):
1260 print('{:20}{} s'.format(name, s))
1265 if __name__ == '__main__':
1266 sys.exit(0 if top5proc() else 1)
1273 python3 top5proc.py ~/lttng-traces/my-kernel-session*/kernel
1279 swapper/0 48.607245889 s
1280 chromium 7.192738188 s
1281 pavucontrol 0.709894415 s
1282 Compositor 0.660867933 s
1283 Xorg.bin 0.616753786 s
1286 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1287 weren't using the CPU that much when tracing, its first position in the
1292 == [[understanding-lttng]]Core concepts
1294 From a user's perspective, the LTTng system is built on a few concepts,
1295 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1296 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1297 Understanding how those objects relate to eachother is key in mastering
1300 The core concepts are:
1302 * <<tracing-session,Tracing session>>
1303 * <<domain,Tracing domain>>
1304 * <<channel,Channel and ring buffer>>
1305 * <<"event","Instrumentation point, event rule, event, and event record">>
1311 A _tracing session_ is a stateful dialogue between you and
1312 a <<lttng-sessiond,session daemon>>. You can
1313 <<creating-destroying-tracing-sessions,create a new tracing
1314 session>> with the `lttng create` command.
1316 Anything that you do when you control LTTng tracers happens within a
1317 tracing session. In particular, a tracing session:
1320 * Has its own set of trace files.
1321 * Has its own state of activity (started or stopped).
1322 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1324 * Has its own <<channel,channels>> which have their own
1325 <<event,event rules>>.
1328 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1329 image::concepts.png[]
1331 Those attributes and objects are completely isolated between different
1334 A tracing session is analogous to a cash machine session:
1335 the operations you do on the banking system through the cash machine do
1336 not alter the data of other users of the same system. In the case of
1337 the cash machine, a session lasts as long as your bank card is inside.
1338 In the case of LTTng, a tracing session lasts from the `lttng create`
1339 command to the `lttng destroy` command.
1342 .Each Unix user has its own set of tracing sessions.
1343 image::many-sessions.png[]
1346 [[tracing-session-mode]]
1347 ==== Tracing session mode
1349 LTTng can send the generated trace data to different locations. The
1350 _tracing session mode_ dictates where to send it. The following modes
1351 are available in LTTng{nbsp}{revision}:
1354 LTTng writes the traces to the file system of the machine being traced
1357 Network streaming mode::
1358 LTTng sends the traces over the network to a
1359 <<lttng-relayd,relay daemon>> running on a remote system.
1362 LTTng does not write the traces by default. Instead, you can request
1363 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1364 current tracing buffers, and to write it to the target's file system
1365 or to send it over the network to a <<lttng-relayd,relay daemon>>
1366 running on a remote system.
1369 This mode is similar to the network streaming mode, but a live
1370 trace viewer can connect to the distant relay daemon to
1371 <<lttng-live,view event records as LTTng generates them>> by
1378 A _tracing domain_ is a namespace for event sources. A tracing domain
1379 has its own properties and features.
1381 There are currently five available tracing domains:
1385 * `java.util.logging` (JUL)
1389 You must specify a tracing domain when using some commands to avoid
1390 ambiguity. For example, since all the domains support named tracepoints
1391 as event sources (instrumentation points that you manually insert in the
1392 source code), you need to specify a tracing domain when
1393 <<enabling-disabling-events,creating an event rule>> because all the
1394 tracing domains could have tracepoints with the same names.
1396 Some features are reserved to specific tracing domains. Dynamic function
1397 entry and return instrumentation points, for example, are currently only
1398 supported in the Linux kernel tracing domain, but support for other
1399 tracing domains could be added in the future.
1401 You can create <<channel,channels>> in the Linux kernel and user space
1402 tracing domains. The other tracing domains have a single default
1407 === Channel and ring buffer
1409 A _channel_ is an object which is responsible for a set of ring buffers.
1410 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1411 tracer emits an event, it can record it to one or more
1412 sub-buffers. The attributes of a channel determine what to do when
1413 there's no space left for a new event record because all sub-buffers
1414 are full, where to send a full sub-buffer, and other behaviours.
1416 A channel is always associated to a <<domain,tracing domain>>. The
1417 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1418 a default channel which you cannot configure.
1420 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1421 an event, it records it to the sub-buffers of all
1422 the enabled channels with a satisfied event rule, as long as those
1423 channels are part of active <<tracing-session,tracing sessions>>.
1426 [[channel-buffering-schemes]]
1427 ==== Per-user vs. per-process buffering schemes
1429 A channel has at least one ring buffer _per CPU_. LTTng always
1430 records an event to the ring buffer associated to the CPU on which it
1433 Two _buffering schemes_ are available when you
1434 <<enabling-disabling-channels,create a channel>> in the
1435 user space <<domain,tracing domain>>:
1437 Per-user buffering::
1438 Allocate one set of ring buffers--one per CPU--shared by all the
1439 instrumented processes of each Unix user.
1443 .Per-user buffering scheme.
1444 image::per-user-buffering.png[]
1447 Per-process buffering::
1448 Allocate one set of ring buffers--one per CPU--for each
1449 instrumented process.
1453 .Per-process buffering scheme.
1454 image::per-process-buffering.png[]
1457 The per-process buffering scheme tends to consume more memory than the
1458 per-user option because systems generally have more instrumented
1459 processes than Unix users running instrumented processes. However, the
1460 per-process buffering scheme ensures that one process having a high
1461 event throughput won't fill all the shared sub-buffers of the same
1464 The Linux kernel tracing domain has only one available buffering scheme
1465 which is to allocate a single set of ring buffers for the whole system.
1466 This scheme is similar to the per-user option, but with a single, global
1467 user "running" the kernel.
1470 [[channel-overwrite-mode-vs-discard-mode]]
1471 ==== Overwrite vs. discard event loss modes
1473 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1474 arc in the following animation) of a specific channel's ring buffer.
1475 When there's no space left in a sub-buffer, the tracer marks it as
1476 consumable (red) and another, empty sub-buffer starts receiving the
1477 following event records. A <<lttng-consumerd,consumer daemon>>
1478 eventually consumes the marked sub-buffer (returns to white).
1481 [role="docsvg-channel-subbuf-anim"]
1486 In an ideal world, sub-buffers are consumed faster than they are filled,
1487 as is the case in the previous animation. In the real world,
1488 however, all sub-buffers can be full at some point, leaving no space to
1489 record the following events.
1491 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1492 available, it is acceptable to lose event records when the alternative
1493 would be to cause substantial delays in the instrumented application's
1494 execution. LTTng privileges performance over integrity; it aims at
1495 perturbing the traced system as little as possible in order to make
1496 tracing of subtle race conditions and rare interrupt cascades possible.
1498 When it comes to losing event records because no empty sub-buffer is
1499 available, the channel's _event loss mode_ determines what to do. The
1500 available event loss modes are:
1503 Drop the newest event records until a the tracer
1504 releases a sub-buffer.
1507 Clear the sub-buffer containing the oldest event records and start
1508 writing the newest event records there.
1510 This mode is sometimes called _flight recorder mode_ because it's
1512 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1513 always keep a fixed amount of the latest data.
1515 Which mechanism you should choose depends on your context: prioritize
1516 the newest or the oldest event records in the ring buffer?
1518 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1519 as soon as a there's no space left for a new event record, whereas in
1520 discard mode, the tracer only discards the event record that doesn't
1523 In discard mode, LTTng increments a count of lost event records when
1524 an event record is lost and saves this count to the trace. In
1525 overwrite mode, LTTng keeps no information when it overwrites a
1526 sub-buffer before consuming it.
1528 There are a few ways to decrease your probability of losing event
1530 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1531 how you can fine-une the sub-buffer count and size of a channel to
1532 virtually stop losing event records, though at the cost of greater
1536 [[channel-subbuf-size-vs-subbuf-count]]
1537 ==== Sub-buffer count and size
1539 When you <<enabling-disabling-channels,create a channel>>, you can
1540 set its number of sub-buffers and their size.
1542 Note that there is noticeable CPU overhead introduced when
1543 switching sub-buffers (marking a full one as consumable and switching
1544 to an empty one for the following events to be recorded). Knowing this,
1545 the following list presents a few practical situations along with how
1546 to configure the sub-buffer count and size for them:
1548 * **High event throughput**: In general, prefer bigger sub-buffers to
1549 lower the risk of losing event records.
1551 Having bigger sub-buffers also ensures a lower sub-buffer switching
1554 The number of sub-buffers is only meaningful if you create the channel
1555 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1556 other sub-buffers are left unaltered.
1558 * **Low event throughput**: In general, prefer smaller sub-buffers
1559 since the risk of losing event records is low.
1561 Because events occur less frequently, the sub-buffer switching frequency
1562 should remain low and thus the tracer's overhead should not be a
1565 * **Low memory system**: If your target system has a low memory
1566 limit, prefer fewer first, then smaller sub-buffers.
1568 Even if the system is limited in memory, you want to keep the
1569 sub-buffers as big as possible to avoid a high sub-buffer switching
1572 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1573 which means event data is very compact. For example, the average
1574 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1575 sub-buffer size of 1{nbsp}MiB is considered big.
1577 The previous situations highlight the major trade-off between a few big
1578 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1579 frequency vs. how much data is lost in overwrite mode. Assuming a
1580 constant event throughput and using the overwrite mode, the two
1581 following configurations have the same ring buffer total size:
1584 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1589 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1590 switching frequency, but if a sub-buffer overwrite happens, half of
1591 the event records so far (4{nbsp}MiB) are definitely lost.
1592 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1593 overhead as the previous configuration, but if a sub-buffer
1594 overwrite happens, only the eighth of event records so far are
1597 In discard mode, the sub-buffers count parameter is pointless: use two
1598 sub-buffers and set their size according to the requirements of your
1602 [[channel-switch-timer]]
1603 ==== Switch timer period
1605 The _switch timer period_ is an important configurable attribute of
1606 a channel to ensure periodic sub-buffer flushing.
1608 When the _switch timer_ expires, a sub-buffer switch happens. You can
1609 set the switch timer period attribute when you
1610 <<enabling-disabling-channels,create a channel>> to ensure that event
1611 data is consumed and committed to trace files or to a distant relay
1612 daemon periodically in case of a low event throughput.
1615 [role="docsvg-channel-switch-timer"]
1620 This attribute is also convenient when you use big sub-buffers to cope
1621 with a sporadic high event throughput, even if the throughput is
1625 [[channel-read-timer]]
1626 ==== Read timer period
1628 By default, the LTTng tracers use a notification mechanism to signal a
1629 full sub-buffer so that a consumer daemon can consume it. When such
1630 notifications must be avoided, for example in real-time applications,
1631 you can use the channel's _read timer_ instead. When the read timer
1632 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1633 consumable sub-buffers.
1636 [[tracefile-rotation]]
1637 ==== Trace file count and size
1639 By default, trace files can grow as large as needed. You can set the
1640 maximum size of each trace file that a channel writes when you
1641 <<enabling-disabling-channels,create a channel>>. When the size of
1642 a trace file reaches the channel's fixed maximum size, LTTng creates
1643 another file to contain the next event records. LTTng appends a file
1644 count to each trace file name in this case.
1646 If you set the trace file size attribute when you create a channel, the
1647 maximum number of trace files that LTTng creates is _unlimited_ by
1648 default. To limit them, you can also set a maximum number of trace
1649 files. When the number of trace files reaches the channel's fixed
1650 maximum count, the oldest trace file is overwritten. This mechanism is
1651 called _trace file rotation_.
1655 === Instrumentation point, event rule, event, and event record
1657 An _event rule_ is a set of conditions which must be **all** satisfied
1658 for LTTng to record an occuring event.
1660 You set the conditions when you <<enabling-disabling-events,create
1663 You always attach an event rule to <<channel,channel>> when you create
1666 When an event passes the conditions of an event rule, LTTng records it
1667 in one of the attached channel's sub-buffers.
1669 The available conditions, as of LTTng{nbsp}{revision}, are:
1671 * The event rule _is enabled_.
1672 * The instrumentation point's type _is{nbsp}T_.
1673 * The instrumentation point's name (sometimes called _event name_)
1674 _matches{nbsp}N_, but _is not{nbsp}E_.
1675 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1676 _is exactly{nbsp}L_.
1677 * The fields of the event's payload _satisfy_ a filter
1678 expression{nbsp}__F__.
1680 As you can see, all the conditions but the dynamic filter are related to
1681 the event rule's status or to the instrumentation point, not to the
1682 occurring events. This is why, without a filter, checking if an event
1683 passes an event rule is not a dynamic task: when you create or modify an
1684 event rule, all the tracers of its tracing domain enable or disable the
1685 instrumentation points themselves once. This is possible because the
1686 attributes of an instrumentation point (type, name, and log level) are
1687 defined statically. In other words, without a dynamic filter, the tracer
1688 _does not evaluate_ the arguments of an instrumentation point unless it
1689 matches an enabled event rule.
1691 Note that, for LTTng to record an event, the <<channel,channel>> to
1692 which a matching event rule is attached must also be enabled, and the
1693 tracing session owning this channel must be active.
1696 .Logical path from an instrumentation point to an event record.
1697 image::event-rule.png[]
1699 .Event, event record, or event rule?
1701 With so many similar terms, it's easy to get confused.
1703 An **event** is the consequence of the execution of an _instrumentation
1704 point_, like a tracepoint that you manually place in some source code,
1705 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1706 time. Different actions can be taken upon the occurance of an event,
1707 like record the event's payload to a buffer.
1709 An **event record** is the representation of an event in a sub-buffer. A
1710 tracer is responsible for capturing the payload of an event, current
1711 context variables, the event's ID, and the event's timestamp. LTTng
1712 can append this sub-buffer to a trace file.
1714 An **event rule** is a set of conditions which must all be satisfied for
1715 LTTng to record an occuring event. Events still occur without
1716 satisfying event rules, but LTTng does not record them.
1721 == Components of noch:{LTTng}
1723 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1724 to call LTTng a simple _tool_ since it is composed of multiple
1725 interacting components. This section describes those components,
1726 explains their respective roles, and shows how they connect together to
1727 form the LTTng ecosystem.
1729 The following diagram shows how the most important components of LTTng
1730 interact with user applications, the Linux kernel, and you:
1733 .Control and trace data paths between LTTng components.
1734 image::plumbing.png[]
1736 The LTTng project incorporates:
1738 * **LTTng-tools**: Libraries and command-line interface to
1739 control tracing sessions.
1740 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1741 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1742 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1743 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1744 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1745 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1747 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1748 headers to instrument and trace any native user application.
1749 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1750 *** `liblttng-ust-libc-wrapper`
1751 *** `liblttng-ust-pthread-wrapper`
1752 *** `liblttng-ust-cyg-profile`
1753 *** `liblttng-ust-cyg-profile-fast`
1754 *** `liblttng-ust-dl`
1755 ** User space tracepoint provider source files generator command-line
1756 tool (man:lttng-gen-tp(1)).
1757 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1758 Java applications using `java.util.logging` or
1759 Apache log4j 1.2 logging.
1760 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1761 Python applications using the standard `logging` package.
1762 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1764 ** LTTng kernel tracer module.
1765 ** Tracing ring buffer kernel modules.
1766 ** Probe kernel modules.
1767 ** LTTng logger kernel module.
1771 === Tracing control command-line interface
1774 .The tracing control command-line interface.
1775 image::plumbing-lttng-cli.png[]
1777 The _man:lttng(1) command-line tool_ is the standard user interface to
1778 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1779 is part of LTTng-tools.
1781 The cmd:lttng tool is linked with
1782 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1783 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1785 The cmd:lttng tool has a Git-like interface:
1789 lttng <general options> <command> <command options>
1792 The <<controlling-tracing,Tracing control>> section explores the
1793 available features of LTTng using the cmd:lttng tool.
1796 [[liblttng-ctl-lttng]]
1797 === Tracing control library
1800 .The tracing control library.
1801 image::plumbing-liblttng-ctl.png[]
1803 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1804 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1805 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1807 The <<lttng-cli,cmd:lttng command-line tool>>
1808 is linked with `liblttng-ctl`.
1810 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1815 #include <lttng/lttng.h>
1818 Some objects are referenced by name (C string), such as tracing
1819 sessions, but most of them require to create a handle first using
1820 `lttng_create_handle()`.
1822 The best available developer documentation for `liblttng-ctl` is, as of
1823 LTTng{nbsp}{revision}, its installed header files. Every function and
1824 structure is thoroughly documented.
1828 === User space tracing library
1831 .The user space tracing library.
1832 image::plumbing-liblttng-ust.png[]
1834 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1835 is the LTTng user space tracer. It receives commands from a
1836 <<lttng-sessiond,session daemon>>, for example to
1837 enable and disable specific instrumentation points, and writes event
1838 records to ring buffers shared with a
1839 <<lttng-consumerd,consumer daemon>>.
1840 `liblttng-ust` is part of LTTng-UST.
1842 Public C header files are installed beside `liblttng-ust` to
1843 instrument any <<c-application,C or $$C++$$ application>>.
1845 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1846 packages, use their own library providing tracepoints which is
1847 linked with `liblttng-ust`.
1849 An application or library does not have to initialize `liblttng-ust`
1850 manually: its constructor does the necessary tasks to properly register
1851 to a session daemon. The initialization phase also enables the
1852 instrumentation points matching the <<event,event rules>> that you
1856 [[lttng-ust-agents]]
1857 === User space tracing agents
1860 .The user space tracing agents.
1861 image::plumbing-lttng-ust-agents.png[]
1863 The _LTTng-UST Java and Python agents_ are regular Java and Python
1864 packages which add LTTng tracing capabilities to the
1865 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1867 In the case of Java, the
1868 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1869 core logging facilities] and
1870 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1871 Note that Apache Log4{nbsp}2 is not supported.
1873 In the case of Python, the standard
1874 https://docs.python.org/3/library/logging.html[`logging`] package
1875 is supported. Both Python 2 and Python 3 modules can import the
1876 LTTng-UST Python agent package.
1878 The applications using the LTTng-UST agents are in the
1879 `java.util.logging` (JUL),
1880 log4j, and Python <<domain,tracing domains>>.
1882 Both agents use the same mechanism to trace the log statements. When an
1883 agent is initialized, it creates a log handler that attaches to the root
1884 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1885 When the application executes a log statement, it is passed to the
1886 agent's log handler by the root logger. The agent's log handler calls a
1887 native function in a tracepoint provider package shared library linked
1888 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1889 other fields, like its logger name and its log level. This native
1890 function contains a user space instrumentation point, hence tracing the
1893 The log level condition of an
1894 <<event,event rule>> is considered when tracing
1895 a Java or a Python application, and it's compatible with the standard
1896 JUL, log4j, and Python log levels.
1900 === LTTng kernel modules
1903 .The LTTng kernel modules.
1904 image::plumbing-lttng-modules.png[]
1906 The _LTTng kernel modules_ are a set of Linux kernel modules
1907 which implement the kernel tracer of the LTTng project. The LTTng
1908 kernel modules are part of LTTng-modules.
1910 The LTTng kernel modules include:
1912 * A set of _probe_ modules.
1914 Each module attaches to a specific subsystem
1915 of the Linux kernel using its tracepoint instrument points. There are
1916 also modules to attach to the entry and return points of the Linux
1917 system call functions.
1919 * _Ring buffer_ modules.
1921 A ring buffer implementation is provided as kernel modules. The LTTng
1922 kernel tracer writes to the ring buffer; a
1923 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1925 * The _LTTng kernel tracer_ module.
1926 * The _LTTng logger_ module.
1928 The LTTng logger module implements the special path:{/proc/lttng-logger}
1929 file so that any executable can generate LTTng events by opening and
1930 writing to this file.
1932 See <<proc-lttng-logger-abi,LTTng logger>>.
1934 Generally, you do not have to load the LTTng kernel modules manually
1935 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1936 daemon>> loads the necessary modules when starting. If you have extra
1937 probe modules, you can specify to load them to the session daemon on
1940 The LTTng kernel modules are installed in
1941 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1942 the kernel release (see `uname --kernel-release`).
1949 .The session daemon.
1950 image::plumbing-sessiond.png[]
1952 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1953 managing tracing sessions and for controlling the various components of
1954 LTTng. The session daemon is part of LTTng-tools.
1956 The session daemon sends control requests to and receives control
1959 * The <<lttng-ust,user space tracing library>>.
1961 Any instance of the user space tracing library first registers to
1962 a session daemon. Then, the session daemon can send requests to
1963 this instance, such as:
1966 ** Get the list of tracepoints.
1967 ** Share an <<event,event rule>> so that the user space tracing library
1968 can enable or disable tracepoints. Amongst the possible conditions
1969 of an event rule is a filter expression which `liblttng-ust` evalutes
1970 when an event occurs.
1971 ** Share <<channel,channel>> attributes and ring buffer locations.
1974 The session daemon and the user space tracing library use a Unix
1975 domain socket for their communication.
1977 * The <<lttng-ust-agents,user space tracing agents>>.
1979 Any instance of a user space tracing agent first registers to
1980 a session daemon. Then, the session daemon can send requests to
1981 this instance, such as:
1984 ** Get the list of loggers.
1985 ** Enable or disable a specific logger.
1988 The session daemon and the user space tracing agent use a TCP connection
1989 for their communication.
1991 * The <<lttng-modules,LTTng kernel tracer>>.
1992 * The <<lttng-consumerd,consumer daemon>>.
1994 The session daemon sends requests to the consumer daemon to instruct
1995 it where to send the trace data streams, amongst other information.
1997 * The <<lttng-relayd,relay daemon>>.
1999 The session daemon receives commands from the
2000 <<liblttng-ctl-lttng,tracing control library>>.
2002 The root session daemon loads the appropriate
2003 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2004 a <<lttng-consumerd,consumer daemon>> as soon as you create
2005 an <<event,event rule>>.
2007 The session daemon does not send and receive trace data: this is the
2008 role of the <<lttng-consumerd,consumer daemon>> and
2009 <<lttng-relayd,relay daemon>>. It does, however, generate the
2010 http://diamon.org/ctf/[CTF] metadata stream.
2012 Each Unix user can have its own session daemon instance. The
2013 tracing sessions managed by different session daemons are completely
2016 The root user's session daemon is the only one which is
2017 allowed to control the LTTng kernel tracer, and its spawned consumer
2018 daemon is the only one which is allowed to consume trace data from the
2019 LTTng kernel tracer. Note, however, that any Unix user which is a member
2020 of the <<tracing-group,tracing group>> is allowed
2021 to create <<channel,channels>> in the
2022 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2025 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2026 session daemon when using its `create` command if none is currently
2027 running. You can also start the session daemon manually.
2034 .The consumer daemon.
2035 image::plumbing-consumerd.png[]
2037 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
2038 ring buffers with user applications or with the LTTng kernel modules to
2039 collect trace data and send it to some location (on disk or to a
2040 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2041 is part of LTTng-tools.
2043 You do not start a consumer daemon manually: a consumer daemon is always
2044 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2045 <<event,event rule>>, that is, before you start tracing. When you kill
2046 its owner session daemon, the consumer daemon also exits because it is
2047 the session daemon's child process. Command-line options of
2048 man:lttng-sessiond(8) target the consumer daemon process.
2050 There are up to two running consumer daemons per Unix user, whereas only
2051 one session daemon can run per user. This is because each process can be
2052 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2053 and 64-bit processes, it is more efficient to have separate
2054 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2055 exception: it can have up to _three_ running consumer daemons: 32-bit
2056 and 64-bit instances for its user applications, and one more
2057 reserved for collecting kernel trace data.
2065 image::plumbing-relayd.png[]
2067 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2068 between remote session and consumer daemons, local trace files, and a
2069 remote live trace viewer. The relay daemon is part of LTTng-tools.
2071 The main purpose of the relay daemon is to implement a receiver of
2072 <<sending-trace-data-over-the-network,trace data over the network>>.
2073 This is useful when the target system does not have much file system
2074 space to record trace files locally.
2076 The relay daemon is also a server to which a
2077 <<lttng-live,live trace viewer>> can
2078 connect. The live trace viewer sends requests to the relay daemon to
2079 receive trace data as the target system emits events. The
2080 communication protocol is named _LTTng live_; it is used over TCP
2083 Note that you can start the relay daemon on the target system directly.
2084 This is the setup of choice when the use case is to view events as
2085 the target system emits them without the need of a remote system.
2089 == [[using-lttng]]Instrumentation
2091 There are many examples of tracing and monitoring in our everyday life:
2093 * You have access to real-time and historical weather reports and
2094 forecasts thanks to weather stations installed around the country.
2095 * You know your heart is safe thanks to an electrocardiogram.
2096 * You make sure not to drive your car too fast and to have enough fuel
2097 to reach your destination thanks to gauges visible on your dashboard.
2099 All the previous examples have something in common: they rely on
2100 **instruments**. Without the electrodes attached to the surface of your
2101 body's skin, cardiac monitoring is futile.
2103 LTTng, as a tracer, is no different from those real life examples. If
2104 you're about to trace a software system or, in other words, record its
2105 history of execution, you better have **instrumentation points** in the
2106 subject you're tracing, that is, the actual software.
2108 Various ways were developed to instrument a piece of software for LTTng
2109 tracing. The most straightforward one is to manually place
2110 instrumentation points, called _tracepoints_, in the software's source
2111 code. It is also possible to add instrumentation points dynamically in
2112 the Linux kernel <<domain,tracing domain>>.
2114 If you're only interested in tracing the Linux kernel, your
2115 instrumentation needs are probably already covered by LTTng's built-in
2116 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2117 user application which is already instrumented for LTTng tracing.
2118 In such cases, you can skip this whole section and read the topics of
2119 the <<controlling-tracing,Tracing control>> section.
2121 Many methods are available to instrument a piece of software for LTTng
2124 * <<c-application,User space instrumentation for C and $$C++$$
2126 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2127 * <<java-application,User space Java agent>>.
2128 * <<python-application,User space Python agent>>.
2129 * <<proc-lttng-logger-abi,LTTng logger>>.
2130 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2134 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2136 The procedure to instrument a C or $$C++$$ user application with
2137 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2139 . <<tracepoint-provider,Create the source files of a tracepoint provider
2141 . <<probing-the-application-source-code,Add tracepoints to
2142 the application's source code>>.
2143 . <<building-tracepoint-providers-and-user-application,Build and link
2144 a tracepoint provider package and the user application>>.
2146 If you need quick, man:printf(3)-like instrumentation, you can skip
2147 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2150 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2151 instrument a user application with `liblttng-ust`.
2154 [[tracepoint-provider]]
2155 ==== Create the source files of a tracepoint provider package
2157 A _tracepoint provider_ is a set of compiled functions which provide
2158 **tracepoints** to an application, the type of instrumentation point
2159 supported by LTTng-UST. Those functions can emit events with
2160 user-defined fields and serialize those events as event records to one
2161 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2162 macro, which you <<probing-the-application-source-code,insert in a user
2163 application's source code>>, calls those functions.
2165 A _tracepoint provider package_ is an object file (`.o`) or a shared
2166 library (`.so`) which contains one or more tracepoint providers.
2167 Its source files are:
2169 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2170 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2172 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2173 the LTTng user space tracer, at run time.
2176 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2177 image::ust-app.png[]
2179 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2180 skip creating and using a tracepoint provider and use
2181 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2185 ===== Create a tracepoint provider header file template
2187 A _tracepoint provider header file_ contains the tracepoint
2188 definitions of a tracepoint provider.
2190 To create a tracepoint provider header file:
2192 . Start from this template:
2196 .Tracepoint provider header file template (`.h` file extension).
2198 #undef TRACEPOINT_PROVIDER
2199 #define TRACEPOINT_PROVIDER provider_name
2201 #undef TRACEPOINT_INCLUDE
2202 #define TRACEPOINT_INCLUDE "./tp.h"
2204 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2207 #include <lttng/tracepoint.h>
2210 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2211 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2216 #include <lttng/tracepoint-event.h>
2222 * `provider_name` with the name of your tracepoint provider.
2223 * `"tp.h"` with the name of your tracepoint provider header file.
2225 . Below the `#include <lttng/tracepoint.h>` line, put your
2226 <<defining-tracepoints,tracepoint definitions>>.
2228 Your tracepoint provider name must be unique amongst all the possible
2229 tracepoint provider names used on the same target system. We
2230 suggest to include the name of your project or company in the name,
2231 for example, `org_lttng_my_project_tpp`.
2233 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2234 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2235 write are the <<defining-tracepoints,tracepoint definitions>>.
2238 [[defining-tracepoints]]
2239 ===== Create a tracepoint definition
2241 A _tracepoint definition_ defines, for a given tracepoint:
2243 * Its **input arguments**. They are the macro parameters that the
2244 `tracepoint()` macro accepts for this particular tracepoint
2245 in the user application's source code.
2246 * Its **output event fields**. They are the sources of event fields
2247 that form the payload of any event that the execution of the
2248 `tracepoint()` macro emits for this particular tracepoint.
2250 You can create a tracepoint definition by using the
2251 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2253 <<tpp-header,tracepoint provider header file template>>.
2255 The syntax of the `TRACEPOINT_EVENT()` macro is:
2258 .`TRACEPOINT_EVENT()` macro syntax.
2261 /* Tracepoint provider name */
2264 /* Tracepoint name */
2267 /* Input arguments */
2272 /* Output event fields */
2281 * `provider_name` with your tracepoint provider name.
2282 * `tracepoint_name` with your tracepoint name.
2283 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2284 * `fields` with the <<tpp-def-output-fields,output event field>>
2287 This tracepoint emits events named `provider_name:tracepoint_name`.
2290 .Event name's length limitation
2292 The concatenation of the tracepoint provider name and the
2293 tracepoint name must not exceed **254 characters**. If it does, the
2294 instrumented application compiles and runs, but LTTng throws multiple
2295 warnings and you could experience serious issues.
2298 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2301 .`TP_ARGS()` macro syntax.
2310 * `type` with the C type of the argument.
2311 * `arg_name` with the argument name.
2313 You can repeat `type` and `arg_name` up to 10 times to have
2314 more than one argument.
2316 .`TP_ARGS()` usage with three arguments.
2328 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2329 tracepoint definition with no input arguments.
2331 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2332 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2333 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2334 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2337 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2338 C expression that the tracer evalutes at the `tracepoint()` macro site
2339 in the application's source code. This expression provides a field's
2340 source of data. The argument expression can include input argument names
2341 listed in the `TP_ARGS()` macro.
2343 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2344 must be unique within a given tracepoint definition.
2346 Here's a complete tracepoint definition example:
2348 .Tracepoint definition.
2350 The following tracepoint definition defines a tracepoint which takes
2351 three input arguments and has four output event fields.
2355 #include "my-custom-structure.h"
2361 const struct my_custom_structure*, my_custom_structure,
2366 ctf_string(query_field, query)
2367 ctf_float(double, ratio_field, ratio)
2368 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2369 ctf_integer(int, send_size, my_custom_structure->send_size)
2374 You can refer to this tracepoint definition with the `tracepoint()`
2375 macro in your application's source code like this:
2379 tracepoint(my_provider, my_tracepoint,
2380 my_structure, some_ratio, the_query);
2384 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2385 if they satisfy an enabled <<event,event rule>>.
2388 [[using-tracepoint-classes]]
2389 ===== Use a tracepoint class
2391 A _tracepoint class_ is a class of tracepoints which share the same
2392 output event field definitions. A _tracepoint instance_ is one
2393 instance of such a defined tracepoint class, with its own tracepoint
2396 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2397 shorthand which defines both a tracepoint class and a tracepoint
2398 instance at the same time.
2400 When you build a tracepoint provider package, the C or $$C++$$ compiler
2401 creates one serialization function for each **tracepoint class**. A
2402 serialization function is responsible for serializing the event fields
2403 of a tracepoint to a sub-buffer when tracing.
2405 For various performance reasons, when your situation requires multiple
2406 tracepoint definitions with different names, but with the same event
2407 fields, we recommend that you manually create a tracepoint class
2408 and instantiate as many tracepoint instances as needed. One positive
2409 effect of such a design, amongst other advantages, is that all
2410 tracepoint instances of the same tracepoint class reuse the same
2411 serialization function, thus reducing
2412 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2414 .Use a tracepoint class and tracepoint instances.
2416 Consider the following three tracepoint definitions:
2428 ctf_integer(int, userid, userid)
2429 ctf_integer(size_t, len, len)
2441 ctf_integer(int, userid, userid)
2442 ctf_integer(size_t, len, len)
2454 ctf_integer(int, userid, userid)
2455 ctf_integer(size_t, len, len)
2460 In this case, we create three tracepoint classes, with one implicit
2461 tracepoint instance for each of them: `get_account`, `get_settings`, and
2462 `get_transaction`. However, they all share the same event field names
2463 and types. Hence three identical, yet independent serialization
2464 functions are created when you build the tracepoint provider package.
2466 A better design choice is to define a single tracepoint class and three
2467 tracepoint instances:
2471 /* The tracepoint class */
2472 TRACEPOINT_EVENT_CLASS(
2473 /* Tracepoint provider name */
2476 /* Tracepoint class name */
2479 /* Input arguments */
2485 /* Output event fields */
2487 ctf_integer(int, userid, userid)
2488 ctf_integer(size_t, len, len)
2492 /* The tracepoint instances */
2493 TRACEPOINT_EVENT_INSTANCE(
2494 /* Tracepoint provider name */
2497 /* Tracepoint class name */
2500 /* Tracepoint name */
2503 /* Input arguments */
2509 TRACEPOINT_EVENT_INSTANCE(
2518 TRACEPOINT_EVENT_INSTANCE(
2531 [[assigning-log-levels]]
2532 ===== Assign a log level to a tracepoint definition
2534 You can assign an optional _log level_ to a
2535 <<defining-tracepoints,tracepoint definition>>.
2537 Assigning different levels of severity to tracepoint definitions can
2538 be useful: when you <<enabling-disabling-events,create an event rule>>,
2539 you can target tracepoints having a log level as severe as a specific
2542 The concept of LTTng-UST log levels is similar to the levels found
2543 in typical logging frameworks:
2545 * In a logging framework, the log level is given by the function
2546 or method name you use at the log statement site: `debug()`,
2547 `info()`, `warn()`, `error()`, and so on.
2548 * In LTTng-UST, you statically assign the log level to a tracepoint
2549 definition; any `tracepoint()` macro invocation which refers to
2550 this definition has this log level.
2552 You can assign a log level to a tracepoint definition with the
2553 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2554 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2555 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2558 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2561 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2563 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2568 * `provider_name` with the tracepoint provider name.
2569 * `tracepoint_name` with the tracepoint name.
2570 * `log_level` with the log level to assign to the tracepoint
2571 definition named `tracepoint_name` in the `provider_name`
2572 tracepoint provider.
2574 See man:lttng-ust(3) for a list of available log level names.
2576 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2580 /* Tracepoint definition */
2589 ctf_integer(int, userid, userid)
2590 ctf_integer(size_t, len, len)
2594 /* Log level assignment */
2595 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2601 ===== Create a tracepoint provider package source file
2603 A _tracepoint provider package source file_ is a C source file which
2604 includes a <<tpp-header,tracepoint provider header file>> to expand its
2605 macros into event serialization and other functions.
2607 You can always use the following tracepoint provider package source
2611 .Tracepoint provider package source file template.
2613 #define TRACEPOINT_CREATE_PROBES
2618 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2619 header file>> name. You may also include more than one tracepoint
2620 provider header file here to create a tracepoint provider package
2621 holding more than one tracepoint providers.
2624 [[probing-the-application-source-code]]
2625 ==== Add tracepoints to an application's source code
2627 Once you <<tpp-header,create a tracepoint provider header file>>, you
2628 can use the `tracepoint()` macro in your application's
2629 source code to insert the tracepoints that this header
2630 <<defining-tracepoints,defined>> defines.
2632 The `tracepoint()` macro takes at least two parameters: the tracepoint
2633 provider name and the tracepoint name. The corresponding tracepoint
2634 definition defines the other parameters.
2636 .`tracepoint()` usage.
2638 The following <<defining-tracepoints,tracepoint definition>> defines a
2639 tracepoint which takes two input arguments and has two output event
2643 .Tracepoint provider header file.
2645 #include "my-custom-structure.h"
2652 const char*, cmd_name
2655 ctf_string(cmd_name, cmd_name)
2656 ctf_integer(int, number_of_args, argc)
2661 You can refer to this tracepoint definition with the `tracepoint()`
2662 macro in your application's source code like this:
2665 .Application's source file.
2669 int main(int argc, char* argv[])
2671 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2677 Note how the application's source code includes
2678 the tracepoint provider header file containing the tracepoint
2679 definitions to use, path:{tp.h}.
2682 .`tracepoint()` usage with a complex tracepoint definition.
2684 Consider this complex tracepoint definition, where multiple event
2685 fields refer to the same input arguments in their argument expression
2689 .Tracepoint provider header file.
2691 /* For `struct stat` */
2692 #include <sys/types.h>
2693 #include <sys/stat.h>
2705 ctf_integer(int, my_constant_field, 23 + 17)
2706 ctf_integer(int, my_int_arg_field, my_int_arg)
2707 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2708 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2709 my_str_arg[2] + my_str_arg[3])
2710 ctf_string(my_str_arg_field, my_str_arg)
2711 ctf_integer_hex(off_t, size_field, st->st_size)
2712 ctf_float(double, size_dbl_field, (double) st->st_size)
2713 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2714 size_t, strlen(my_str_arg) / 2)
2719 You can refer to this tracepoint definition with the `tracepoint()`
2720 macro in your application's source code like this:
2723 .Application's source file.
2725 #define TRACEPOINT_DEFINE
2732 stat("/etc/fstab", &s);
2733 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2739 If you look at the event record that LTTng writes when tracing this
2740 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2741 it should look like this:
2743 .Event record fields
2745 |Field's name |Field's value
2746 |`my_constant_field` |40
2747 |`my_int_arg_field` |23
2748 |`my_int_arg_field2` |529
2750 |`my_str_arg_field` |`Hello, World!`
2751 |`size_field` |0x12d
2752 |`size_dbl_field` |301.0
2753 |`half_my_str_arg_field` |`Hello,`
2757 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2758 compute--they use the call stack, for example. To avoid this
2759 computation when the tracepoint is disabled, you can use the
2760 `tracepoint_enabled()` and `do_tracepoint()` macros.
2762 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2766 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2768 tracepoint_enabled(provider_name, tracepoint_name)
2769 do_tracepoint(provider_name, tracepoint_name, ...)
2774 * `provider_name` with the tracepoint provider name.
2775 * `tracepoint_name` with the tracepoint name.
2777 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2778 `tracepoint_name` from the provider named `provider_name` is enabled
2781 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2782 if the tracepoint is enabled. Using `tracepoint()` with
2783 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2784 the `tracepoint_enabled()` check, thus a race condition is
2785 possible in this situation:
2788 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2790 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2791 stuff = prepare_stuff();
2794 tracepoint(my_provider, my_tracepoint, stuff);
2797 If the tracepoint is enabled after the condition, then `stuff` is not
2798 prepared: the emitted event will either contain wrong data, or the whole
2799 application could crash (segmentation fault, for example).
2801 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2802 `STAP_PROBEV()` call. If you need it, you must emit
2806 [[building-tracepoint-providers-and-user-application]]
2807 ==== Build and link a tracepoint provider package and an application
2809 Once you have one or more <<tpp-header,tracepoint provider header
2810 files>> and a <<tpp-source,tracepoint provider package source file>>,
2811 you can create the tracepoint provider package by compiling its source
2812 file. From here, multiple build and run scenarios are possible. The
2813 following table shows common application and library configurations
2814 along with the required command lines to achieve them.
2816 In the following diagrams, we use the following file names:
2819 Executable application.
2822 Application's object file.
2825 Tracepoint provider package object file.
2828 Tracepoint provider package archive file.
2831 Tracepoint provider package shared object file.
2834 User library object file.
2837 User library shared object file.
2839 The red star indicates that this object file is instrumented
2840 (contains code which uses the `tracepoint()` macro). The spring
2841 symbol between the application and a library means the application is
2842 linked with the library at build time.
2844 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2845 variable in the following instructions.
2847 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2848 .Common tracepoint provider package scenarios.
2850 |Scenario |Instructions
2853 The instrumented application is statically linked with
2854 the tracepoint provider package object.
2856 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2859 include::../common/ust-sit-step-tp-o.txt[]
2861 To build the instrumented application:
2863 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2868 #define TRACEPOINT_DEFINE
2872 . Compile the application source file:
2881 . Build the application:
2886 gcc -o app app.o tpp.o -llttng-ust -ldl
2890 To run the instrumented application:
2892 * Start the application:
2902 The instrumented application is statically linked with the
2903 tracepoint provider package archive file.
2905 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2908 To create the tracepoint provider package archive file:
2910 . Compile the <<tpp-source,tracepoint provider package source file>>:
2919 . Create the tracepoint provider package archive file:
2928 To build the instrumented application:
2930 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2935 #define TRACEPOINT_DEFINE
2939 . Compile the application source file:
2948 . Build the application:
2953 gcc -o app app.o tpp.a -llttng-ust -ldl
2957 To run the instrumented application:
2959 * Start the application:
2969 The instrumented application is linked with the tracepoint provider
2970 package shared object.
2972 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2975 include::../common/ust-sit-step-tp-so.txt[]
2977 To build the instrumented application:
2979 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2984 #define TRACEPOINT_DEFINE
2988 . Compile the application source file:
2997 . Build the application:
3002 gcc -o app app.o -ldl -L. -ltpp
3006 To run the instrumented application:
3008 * Start the application:
3018 The tracepoint provider package shared object is preloaded before the
3019 instrumented application starts.
3021 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3024 include::../common/ust-sit-step-tp-so.txt[]
3026 To build the instrumented application:
3028 . In path:{app.c}, before including path:{tpp.h}, add the
3034 #define TRACEPOINT_DEFINE
3035 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3039 . Compile the application source file:
3048 . Build the application:
3053 gcc -o app app.o -ldl
3057 To run the instrumented application with tracing support:
3059 * Preload the tracepoint provider package shared object and
3060 start the application:
3065 LD_PRELOAD=./libtpp.so ./app
3069 To run the instrumented application without tracing support:
3071 * Start the application:
3081 The instrumented application dynamically loads the tracepoint provider
3082 package shared object.
3084 See the <<dlclose-warning,warning about `dlclose()`>>.
3086 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3089 include::../common/ust-sit-step-tp-so.txt[]
3091 To build the instrumented application:
3093 . In path:{app.c}, before including path:{tpp.h}, add the
3099 #define TRACEPOINT_DEFINE
3100 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3104 . Compile the application source file:
3113 . Build the application:
3118 gcc -o app app.o -ldl
3122 To run the instrumented application:
3124 * Start the application:
3134 The application is linked with the instrumented user library.
3136 The instrumented user library is statically linked with the tracepoint
3137 provider package object file.
3139 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3142 include::../common/ust-sit-step-tp-o-fpic.txt[]
3144 To build the instrumented user library:
3146 . In path:{emon.c}, before including path:{tpp.h}, add the
3152 #define TRACEPOINT_DEFINE
3156 . Compile the user library source file:
3161 gcc -I. -fpic -c emon.c
3165 . Build the user library shared object:
3170 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3174 To build the application:
3176 . Compile the application source file:
3185 . Build the application:
3190 gcc -o app app.o -L. -lemon
3194 To run the application:
3196 * Start the application:
3206 The application is linked with the instrumented user library.
3208 The instrumented user library is linked with the tracepoint provider
3209 package shared object.
3211 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3214 include::../common/ust-sit-step-tp-so.txt[]
3216 To build the instrumented user library:
3218 . In path:{emon.c}, before including path:{tpp.h}, add the
3224 #define TRACEPOINT_DEFINE
3228 . Compile the user library source file:
3233 gcc -I. -fpic -c emon.c
3237 . Build the user library shared object:
3242 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3246 To build the application:
3248 . Compile the application source file:
3257 . Build the application:
3262 gcc -o app app.o -L. -lemon
3266 To run the application:
3268 * Start the application:
3278 The tracepoint provider package shared object is preloaded before the
3281 The application is linked with the instrumented user library.
3283 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3286 include::../common/ust-sit-step-tp-so.txt[]
3288 To build the instrumented user library:
3290 . In path:{emon.c}, before including path:{tpp.h}, add the
3296 #define TRACEPOINT_DEFINE
3297 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3301 . Compile the user library source file:
3306 gcc -I. -fpic -c emon.c
3310 . Build the user library shared object:
3315 gcc -shared -o libemon.so emon.o -ldl
3319 To build the application:
3321 . Compile the application source file:
3330 . Build the application:
3335 gcc -o app app.o -L. -lemon
3339 To run the application with tracing support:
3341 * Preload the tracepoint provider package shared object and
3342 start the application:
3347 LD_PRELOAD=./libtpp.so ./app
3351 To run the application without tracing support:
3353 * Start the application:
3363 The application is linked with the instrumented user library.
3365 The instrumented user library dynamically loads the tracepoint provider
3366 package shared object.
3368 See the <<dlclose-warning,warning about `dlclose()`>>.
3370 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3373 include::../common/ust-sit-step-tp-so.txt[]
3375 To build the instrumented user library:
3377 . In path:{emon.c}, before including path:{tpp.h}, add the
3383 #define TRACEPOINT_DEFINE
3384 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3388 . Compile the user library source file:
3393 gcc -I. -fpic -c emon.c
3397 . Build the user library shared object:
3402 gcc -shared -o libemon.so emon.o -ldl
3406 To build the application:
3408 . Compile the application source file:
3417 . Build the application:
3422 gcc -o app app.o -L. -lemon
3426 To run the application:
3428 * Start the application:
3438 The application dynamically loads the instrumented user library.
3440 The instrumented user library is linked with the tracepoint provider
3441 package shared object.
3443 See the <<dlclose-warning,warning about `dlclose()`>>.
3445 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3448 include::../common/ust-sit-step-tp-so.txt[]
3450 To build the instrumented user library:
3452 . In path:{emon.c}, before including path:{tpp.h}, add the
3458 #define TRACEPOINT_DEFINE
3462 . Compile the user library source file:
3467 gcc -I. -fpic -c emon.c
3471 . Build the user library shared object:
3476 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3480 To build the application:
3482 . Compile the application source file:
3491 . Build the application:
3496 gcc -o app app.o -ldl -L. -lemon
3500 To run the application:
3502 * Start the application:
3512 The application dynamically loads the instrumented user library.
3514 The instrumented user library dynamically loads the tracepoint provider
3515 package shared object.
3517 See the <<dlclose-warning,warning about `dlclose()`>>.
3519 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3522 include::../common/ust-sit-step-tp-so.txt[]
3524 To build the instrumented user library:
3526 . In path:{emon.c}, before including path:{tpp.h}, add the
3532 #define TRACEPOINT_DEFINE
3533 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3537 . Compile the user library source file:
3542 gcc -I. -fpic -c emon.c
3546 . Build the user library shared object:
3551 gcc -shared -o libemon.so emon.o -ldl
3555 To build the application:
3557 . Compile the application source file:
3566 . Build the application:
3571 gcc -o app app.o -ldl -L. -lemon
3575 To run the application:
3577 * Start the application:
3587 The tracepoint provider package shared object is preloaded before the
3590 The application dynamically loads the instrumented user library.
3592 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3595 include::../common/ust-sit-step-tp-so.txt[]
3597 To build the instrumented user library:
3599 . In path:{emon.c}, before including path:{tpp.h}, add the
3605 #define TRACEPOINT_DEFINE
3606 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3610 . Compile the user library source file:
3615 gcc -I. -fpic -c emon.c
3619 . Build the user library shared object:
3624 gcc -shared -o libemon.so emon.o -ldl
3628 To build the application:
3630 . Compile the application source file:
3639 . Build the application:
3644 gcc -o app app.o -L. -lemon
3648 To run the application with tracing support:
3650 * Preload the tracepoint provider package shared object and
3651 start the application:
3656 LD_PRELOAD=./libtpp.so ./app
3660 To run the application without tracing support:
3662 * Start the application:
3672 The application is statically linked with the tracepoint provider
3673 package object file.
3675 The application is linked with the instrumented user library.
3677 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3680 include::../common/ust-sit-step-tp-o.txt[]
3682 To build the instrumented user library:
3684 . In path:{emon.c}, before including path:{tpp.h}, add the
3690 #define TRACEPOINT_DEFINE
3694 . Compile the user library source file:
3699 gcc -I. -fpic -c emon.c
3703 . Build the user library shared object:
3708 gcc -shared -o libemon.so emon.o
3712 To build the application:
3714 . Compile the application source file:
3723 . Build the application:
3728 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3732 To run the instrumented application:
3734 * Start the application:
3744 The application is statically linked with the tracepoint provider
3745 package object file.
3747 The application dynamically loads the instrumented user library.
3749 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3752 include::../common/ust-sit-step-tp-o.txt[]
3754 To build the application:
3756 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3761 #define TRACEPOINT_DEFINE
3765 . Compile the application source file:
3774 . Build the application:
3779 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3784 The `--export-dynamic` option passed to the linker is necessary for the
3785 dynamically loaded library to ``see'' the tracepoint symbols defined in
3788 To build the instrumented user library:
3790 . Compile the user library source file:
3795 gcc -I. -fpic -c emon.c
3799 . Build the user library shared object:
3804 gcc -shared -o libemon.so emon.o
3808 To run the application:
3810 * Start the application:
3822 .Do not use man:dlclose(3) on a tracepoint provider package
3824 Never use man:dlclose(3) on any shared object which:
3826 * Is linked with, statically or dynamically, a tracepoint provider
3828 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3829 package shared object.
3831 This is currently considered **unsafe** due to a lack of reference
3832 counting from LTTng-UST to the shared object.
3834 A known workaround (available since glibc 2.2) is to use the
3835 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3836 effect of not unloading the loaded shared object, even if man:dlclose(3)
3839 You can also preload the tracepoint provider package shared object with
3840 the env:LD_PRELOAD environment variable to overcome this limitation.
3844 [[using-lttng-ust-with-daemons]]
3845 ===== Use noch:{LTTng-UST} with daemons
3847 If your instrumented application calls man:fork(2), man:clone(2),
3848 or BSD's man:rfork(2), without a following man:exec(3)-family
3849 system call, you must preload the path:{liblttng-ust-fork.so} shared
3850 object when starting the application.
3854 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3857 If your tracepoint provider package is
3858 a shared library which you also preload, you must put both
3859 shared objects in env:LD_PRELOAD:
3863 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3867 [[lttng-ust-pkg-config]]
3868 ===== Use noch:{pkg-config}
3870 On some distributions, LTTng-UST ships with a
3871 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3872 metadata file. If this is your case, then you can use cmd:pkg-config to
3873 build an application on the command line:
3877 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3881 [[instrumenting-32-bit-app-on-64-bit-system]]
3882 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3884 In order to trace a 32-bit application running on a 64-bit system,
3885 LTTng must use a dedicated 32-bit
3886 <<lttng-consumerd,consumer daemon>>.
3888 The following steps show how to build and install a 32-bit consumer
3889 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3890 build and install the 32-bit LTTng-UST libraries, and how to build and
3891 link an instrumented 32-bit application in that context.
3893 To build a 32-bit instrumented application for a 64-bit target system,
3894 assuming you have a fresh target system with no installed Userspace RCU
3897 . Download, build, and install a 32-bit version of Userspace RCU:
3903 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3904 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3905 cd userspace-rcu-0.9.* &&
3906 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3908 sudo make install &&
3913 . Using your distribution's package manager, or from source, install
3914 the following 32-bit versions of the following dependencies of
3915 LTTng-tools and LTTng-UST:
3918 * https://sourceforge.net/projects/libuuid/[libuuid]
3919 * http://directory.fsf.org/wiki/Popt[popt]
3920 * http://www.xmlsoft.org/[libxml2]
3923 . Download, build, and install a 32-bit version of the latest
3924 LTTng-UST{nbsp}{revision}:
3930 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.8.tar.bz2 &&
3931 tar -xf lttng-ust-latest-2.8.tar.bz2 &&
3932 cd lttng-ust-2.8.* &&
3933 ./configure --libdir=/usr/local/lib32 \
3934 CFLAGS=-m32 CXXFLAGS=-m32 \
3935 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3937 sudo make install &&
3944 Depending on your distribution,
3945 32-bit libraries could be installed at a different location than
3946 `/usr/lib32`. For example, Debian is known to install
3947 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3949 In this case, make sure to set `LDFLAGS` to all the
3950 relevant 32-bit library paths, for example:
3954 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3958 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3959 the 32-bit consumer daemon:
3965 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
3966 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
3967 cd lttng-tools-2.8.* &&
3968 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3969 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3971 cd src/bin/lttng-consumerd &&
3972 sudo make install &&
3977 . From your distribution or from source,
3978 <<installing-lttng,install>> the 64-bit versions of
3979 LTTng-UST and Userspace RCU.
3980 . Download, build, and install the 64-bit version of the
3981 latest LTTng-tools{nbsp}{revision}:
3987 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.8.tar.bz2 &&
3988 tar -xf lttng-tools-latest-2.8.tar.bz2 &&
3989 cd lttng-tools-2.8.* &&
3990 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3991 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3993 sudo make install &&
3998 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3999 when linking your 32-bit application:
4002 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4003 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4006 For example, let's rebuild the quick start example in
4007 <<tracing-your-own-user-application,Trace a user application>> as an
4008 instrumented 32-bit application:
4013 gcc -m32 -c -I. hello-tp.c
4015 gcc -m32 -o hello hello.o hello-tp.o \
4016 -L/usr/lib32 -L/usr/local/lib32 \
4017 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4022 No special action is required to execute the 32-bit application and
4023 to trace it: use the command-line man:lttng(1) tool as usual.
4030 man:tracef(3) is a small LTTng-UST API designed for quick,
4031 man:printf(3)-like instrumentation without the burden of
4032 <<tracepoint-provider,creating>> and
4033 <<building-tracepoint-providers-and-user-application,building>>
4034 a tracepoint provider package.
4036 To use `tracef()` in your application:
4038 . In the C or C++ source files where you need to use `tracef()`,
4039 include `<lttng/tracef.h>`:
4044 #include <lttng/tracef.h>
4048 . In the application's source code, use `tracef()` like you would use
4056 tracef("my message: %d (%s)", my_integer, my_string);
4062 . Link your application with `liblttng-ust`:
4067 gcc -o app app.c -llttng-ust
4071 To trace the events that `tracef()` calls emit:
4073 * <<enabling-disabling-events,Create an event rule>> which matches the
4074 `lttng_ust_tracef:*` event name:
4079 lttng enable-event --userspace 'lttng_ust_tracef:*'
4084 .Limitations of `tracef()`
4086 The `tracef()` utility function was developed to make user space tracing
4087 super simple, albeit with notable disadvantages compared to
4088 <<defining-tracepoints,user-defined tracepoints>>:
4090 * All the emitted events have the same tracepoint provider and
4091 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4092 * There is no static type checking.
4093 * The only event record field you actually get, named `msg`, is a string
4094 potentially containing the values you passed to `tracef()`
4095 using your own format string. This also means that you cannot filter
4096 events with a custom expression at run time because there are no
4098 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4099 function behind the scenes to format the strings at run time, its
4100 expected performance is lower than with user-defined tracepoints,
4101 which do not require a conversion to a string.
4103 Taking this into consideration, `tracef()` is useful for some quick
4104 prototyping and debugging, but you should not consider it for any
4105 permanent and serious applicative instrumentation.
4111 ==== Use `tracelog()`
4113 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4114 the difference that it accepts an additional log level parameter.
4116 The goal of `tracelog()` is to ease the migration from logging to
4119 To use `tracelog()` in your application:
4121 . In the C or C++ source files where you need to use `tracelog()`,
4122 include `<lttng/tracelog.h>`:
4127 #include <lttng/tracelog.h>
4131 . In the application's source code, use `tracelog()` like you would use
4132 man:printf(3), except for the first parameter which is the log
4140 tracelog(TRACE_WARNING, "my message: %d (%s)",
4141 my_integer, my_string);
4147 See man:lttng-ust(3) for a list of available log level names.
4149 . Link your application with `liblttng-ust`:
4154 gcc -o app app.c -llttng-ust
4158 To trace the events that `tracelog()` calls emit with a log level
4159 _as severe as_ a specific log level:
4161 * <<enabling-disabling-events,Create an event rule>> which matches the
4162 `lttng_ust_tracelog:*` event name and a minimum level
4168 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4169 --loglevel=TRACE_WARNING
4173 To trace the events that `tracelog()` calls emit with a
4174 _specific log level_:
4176 * Create an event rule which matches the `lttng_ust_tracelog:*`
4177 event name and a specific log level:
4182 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4183 --loglevel-only=TRACE_INFO
4188 [[prebuilt-ust-helpers]]
4189 === Prebuilt user space tracing helpers
4191 The LTTng-UST package provides a few helpers in the form or preloadable
4192 shared objects which automatically instrument system functions and
4195 The helper shared objects are normally found in dir:{/usr/lib}. If you
4196 built LTTng-UST <<building-from-source,from source>>, they are probably
4197 located in dir:{/usr/local/lib}.
4199 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4202 path:{liblttng-ust-libc-wrapper.so}::
4203 path:{liblttng-ust-pthread-wrapper.so}::
4204 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4205 memory and POSIX threads function tracing>>.
4207 path:{liblttng-ust-cyg-profile.so}::
4208 path:{liblttng-ust-cyg-profile-fast.so}::
4209 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4211 path:{liblttng-ust-dl.so}::
4212 <<liblttng-ust-dl,Dynamic linker tracing>>.
4214 To use a user space tracing helper with any user application:
4216 * Preload the helper shared object when you start the application:
4221 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4225 You can preload more than one helper:
4230 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4236 [[liblttng-ust-libc-pthread-wrapper]]
4237 ==== Instrument C standard library memory and POSIX threads functions
4239 The path:{liblttng-ust-libc-wrapper.so} and
4240 path:{liblttng-ust-pthread-wrapper.so} helpers
4241 add instrumentation to some C standard library and POSIX
4245 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4247 |TP provider name |TP name |Instrumented function
4249 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4250 |`calloc` |man:calloc(3)
4251 |`realloc` |man:realloc(3)
4252 |`free` |man:free(3)
4253 |`memalign` |man:memalign(3)
4254 |`posix_memalign` |man:posix_memalign(3)
4258 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4260 |TP provider name |TP name |Instrumented function
4262 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4263 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4264 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4265 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4268 When you preload the shared object, it replaces the functions listed
4269 in the previous tables by wrappers which contain tracepoints and call
4270 the replaced functions.
4273 [[liblttng-ust-cyg-profile]]
4274 ==== Instrument function entry and exit
4276 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4277 to the entry and exit points of functions.
4279 man:gcc(1) and man:clang(1) have an option named
4280 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4281 which generates instrumentation calls for entry and exit to functions.
4282 The LTTng-UST function tracing helpers,
4283 path:{liblttng-ust-cyg-profile.so} and
4284 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4285 to add tracepoints to the two generated functions (which contain
4286 `cyg_profile` in their names, hence the helper's name).
4288 To use the LTTng-UST function tracing helper, the source files to
4289 instrument must be built using the `-finstrument-functions` compiler
4292 There are two versions of the LTTng-UST function tracing helper:
4294 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4295 that you should only use when it can be _guaranteed_ that the
4296 complete event stream is recorded without any lost event record.
4297 Any kind of duplicate information is left out.
4299 Assuming no event record is lost, having only the function addresses on
4300 entry is enough to create a call graph, since an event record always
4301 contains the ID of the CPU that generated it.
4303 You can use a tool like
4304 https://sourceware.org/binutils/docs/binutils/addr2line.html[cmd:addr2line]
4305 to convert function addresses back to source file names and
4308 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4309 which also works in use cases where event records might get discarded or
4310 not recorded from application startup.
4311 In these cases, the trace analyzer needs more information to be
4312 able to reconstruct the program flow.
4314 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4315 points of this helper.
4317 All the tracepoints that this helper provides have the
4318 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4320 TIP: It's sometimes a good idea to limit the number of source files that
4321 you compile with the `-finstrument-functions` option to prevent LTTng
4322 from writing an excessive amount of trace data at run time. When using
4323 man:gcc(1), you can use the
4324 `-finstrument-functions-exclude-function-list` option to avoid
4325 instrument entries and exits of specific function names.
4330 ==== Instrument the dynamic linker
4332 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4333 man:dlopen(3) and man:dlclose(3) function calls.
4335 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4340 [[java-application]]
4341 === User space Java agent
4343 You can instrument any Java application which uses one of the following
4346 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4347 (JUL) core logging facilities.
4348 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4349 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4352 .LTTng-UST Java agent imported by a Java application.
4353 image::java-app.png[]
4355 Note that the methods described below are new in LTTng{nbsp}{revision}.
4356 Previous LTTng versions use another technique.
4358 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4359 and https://ci.lttng.org/[continuous integration], thus this version is
4360 directly supported. However, the LTTng-UST Java agent is also tested
4361 with OpenJDK{nbsp}7.
4366 ==== Use the LTTng-UST Java agent for `java.util.logging`
4368 To use the LTTng-UST Java agent in a Java application which uses
4369 `java.util.logging` (JUL):
4371 . In the Java application's source code, import the LTTng-UST
4372 log handler package for `java.util.logging`:
4377 import org.lttng.ust.agent.jul.LttngLogHandler;
4381 . Create an LTTng-UST JUL log handler:
4386 Handler lttngUstLogHandler = new LttngLogHandler();
4390 . Add this handler to the JUL loggers which should emit LTTng events:
4395 Logger myLogger = Logger.getLogger("some-logger");
4397 myLogger.addHandler(lttngUstLogHandler);
4401 . Use `java.util.logging` log statements and configuration as usual.
4402 The loggers with an attached LTTng-UST log handler can emit
4405 . Before exiting the application, remove the LTTng-UST log handler from
4406 the loggers attached to it and call its `close()` method:
4411 myLogger.removeHandler(lttngUstLogHandler);
4412 lttngUstLogHandler.close();
4416 This is not strictly necessary, but it is recommended for a clean
4417 disposal of the handler's resources.
4419 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4420 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4422 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4423 path] when you build the Java application.
4425 The JAR files are typically located in dir:{/usr/share/java}.
4427 IMPORTANT: The LTTng-UST Java agent must be
4428 <<installing-lttng,installed>> for the logging framework your
4431 .Use the LTTng-UST Java agent for `java.util.logging`.
4436 import java.io.IOException;
4437 import java.util.logging.Handler;
4438 import java.util.logging.Logger;
4439 import org.lttng.ust.agent.jul.LttngLogHandler;
4443 private static final int answer = 42;
4445 public static void main(String[] argv) throws Exception
4448 Logger logger = Logger.getLogger("jello");
4450 // Create an LTTng-UST log handler
4451 Handler lttngUstLogHandler = new LttngLogHandler();
4453 // Add the LTTng-UST log handler to our logger
4454 logger.addHandler(lttngUstLogHandler);
4457 logger.info("some info");
4458 logger.warning("some warning");
4460 logger.finer("finer information; the answer is " + answer);
4462 logger.severe("error!");
4464 // Not mandatory, but cleaner
4465 logger.removeHandler(lttngUstLogHandler);
4466 lttngUstLogHandler.close();
4475 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4478 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4479 <<enabling-disabling-events,create an event rule>> matching the
4480 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4485 lttng enable-event --jul jello
4489 Run the compiled class:
4493 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4496 <<basic-tracing-session-control,Stop tracing>> and inspect the
4506 You can use the opt:lttng-enable-event(1):--loglevel or
4507 opt:lttng-enable-event(1):--loglevel-only option of the
4508 man:lttng-enable-event(1) command to target a range of JUL log levels
4509 or a specific JUL log level.
4514 ==== Use the LTTng-UST Java agent for Apache log4j
4516 To use the LTTng-UST Java agent in a Java application which uses
4519 . In the Java application's source code, import the LTTng-UST
4520 log appender package for Apache log4j:
4525 import org.lttng.ust.agent.log4j.LttngLogAppender;
4529 . Create an LTTng-UST log4j log appender:
4534 Appender lttngUstLogAppender = new LttngLogAppender();
4538 . Add this appender to the log4j loggers which should emit LTTng events:
4543 Logger myLogger = Logger.getLogger("some-logger");
4545 myLogger.addAppender(lttngUstLogAppender);
4549 . Use Apache log4j log statements and configuration as usual. The
4550 loggers with an attached LTTng-UST log appender can emit LTTng events.
4552 . Before exiting the application, remove the LTTng-UST log appender from
4553 the loggers attached to it and call its `close()` method:
4558 myLogger.removeAppender(lttngUstLogAppender);
4559 lttngUstLogAppender.close();
4563 This is not strictly necessary, but it is recommended for a clean
4564 disposal of the appender's resources.
4566 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4567 files, path:{lttng-ust-agent-common.jar} and
4568 path:{lttng-ust-agent-log4j.jar}, in the
4569 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4570 path] when you build the Java application.
4572 The JAR files are typically located in dir:{/usr/share/java}.
4574 IMPORTANT: The LTTng-UST Java agent must be
4575 <<installing-lttng,installed>> for the logging framework your
4578 .Use the LTTng-UST Java agent for Apache log4j.
4583 import org.apache.log4j.Appender;
4584 import org.apache.log4j.Logger;
4585 import org.lttng.ust.agent.log4j.LttngLogAppender;
4589 private static final int answer = 42;
4591 public static void main(String[] argv) throws Exception
4594 Logger logger = Logger.getLogger("jello");
4596 // Create an LTTng-UST log appender
4597 Appender lttngUstLogAppender = new LttngLogAppender();
4599 // Add the LTTng-UST log appender to our logger
4600 logger.addAppender(lttngUstLogAppender);
4603 logger.info("some info");
4604 logger.warn("some warning");
4606 logger.debug("debug information; the answer is " + answer);
4608 logger.fatal("error!");
4610 // Not mandatory, but cleaner
4611 logger.removeAppender(lttngUstLogAppender);
4612 lttngUstLogAppender.close();
4618 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4623 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4626 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4627 <<enabling-disabling-events,create an event rule>> matching the
4628 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4633 lttng enable-event --log4j jello
4637 Run the compiled class:
4641 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4644 <<basic-tracing-session-control,Stop tracing>> and inspect the
4654 You can use the opt:lttng-enable-event(1):--loglevel or
4655 opt:lttng-enable-event(1):--loglevel-only option of the
4656 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4657 or a specific log4j log level.
4661 [[java-application-context]]
4662 ==== Provide application-specific context fields in a Java application
4664 A Java application-specific context field is a piece of state provided
4665 by the application which <<adding-context,you can add>>, using the
4666 man:lttng-add-context(1) command, to each <<event,event record>>
4667 produced by the log statements of this application.
4669 For example, a given object might have a current request ID variable.
4670 You can create a context information retriever for this object and
4671 assign a name to this current request ID. You can then, using the
4672 man:lttng-add-context(1) command, add this context field by name to
4673 the JUL or log4j <<channel,channel>>.
4675 To provide application-specific context fields in a Java application:
4677 . In the Java application's source code, import the LTTng-UST
4678 Java agent context classes and interfaces:
4683 import org.lttng.ust.agent.context.ContextInfoManager;
4684 import org.lttng.ust.agent.context.IContextInfoRetriever;
4688 . Create a context information retriever class, that is, a class which
4689 implements the `IContextInfoRetriever` interface:
4694 class MyContextInfoRetriever implements IContextInfoRetriever
4697 public Object retrieveContextInfo(String key)
4699 if (key.equals("intCtx")) {
4701 } else if (key.equals("strContext")) {
4702 return "context value!";
4711 This `retrieveContextInfo()` method is the only member of the
4712 `IContextInfoRetriever` interface. Its role is to return the current
4713 value of a state by name to create a context field. The names of the
4714 context fields and which state variables they return depends on your
4717 All primitive types and objects are supported as context fields.
4718 When `retrieveContextInfo()` returns an object, the context field
4719 serializer calls its `toString()` method to add a string field to
4720 event records. The method can also return `null`, which means that
4721 no context field is available for the required name.
4723 . Register an instance of your context information retriever class to
4724 the context information manager singleton:
4729 IContextInfoRetriever cir = new MyContextInfoRetriever();
4730 ContextInfoManager cim = ContextInfoManager.getInstance();
4731 cim.registerContextInfoRetriever("retrieverName", cir);
4735 . Before exiting the application, remove your context information
4736 retriever from the context information manager singleton:
4741 ContextInfoManager cim = ContextInfoManager.getInstance();
4742 cim.unregisterContextInfoRetriever("retrieverName");
4746 This is not strictly necessary, but it is recommended for a clean
4747 disposal of some manager's resources.
4749 . Build your Java application with LTTng-UST Java agent support as
4750 usual, following the procedure for either the <<jul,JUL>> or
4751 <<log4j,Apache log4j>> framework.
4754 .Provide application-specific context fields in a Java application.
4759 import java.util.logging.Handler;
4760 import java.util.logging.Logger;
4761 import org.lttng.ust.agent.jul.LttngLogHandler;
4762 import org.lttng.ust.agent.context.ContextInfoManager;
4763 import org.lttng.ust.agent.context.IContextInfoRetriever;
4767 // Our context information retriever class
4768 private static class MyContextInfoRetriever
4769 implements IContextInfoRetriever
4772 public Object retrieveContextInfo(String key) {
4773 if (key.equals("intCtx")) {
4775 } else if (key.equals("strContext")) {
4776 return "context value!";
4783 private static final int answer = 42;
4785 public static void main(String args[]) throws Exception
4787 // Get the context information manager instance
4788 ContextInfoManager cim = ContextInfoManager.getInstance();
4790 // Create and register our context information retriever
4791 IContextInfoRetriever cir = new MyContextInfoRetriever();
4792 cim.registerContextInfoRetriever("myRetriever", cir);
4795 Logger logger = Logger.getLogger("jello");
4797 // Create an LTTng-UST log handler
4798 Handler lttngUstLogHandler = new LttngLogHandler();
4800 // Add the LTTng-UST log handler to our logger
4801 logger.addHandler(lttngUstLogHandler);
4804 logger.info("some info");
4805 logger.warning("some warning");
4807 logger.finer("finer information; the answer is " + answer);
4809 logger.severe("error!");
4811 // Not mandatory, but cleaner
4812 logger.removeHandler(lttngUstLogHandler);
4813 lttngUstLogHandler.close();
4814 cim.unregisterContextInfoRetriever("myRetriever");
4823 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4826 <<creating-destroying-tracing-sessions,Create a tracing session>>
4827 and <<enabling-disabling-events,create an event rule>> matching the
4833 lttng enable-event --jul jello
4836 <<adding-context,Add the application-specific context fields>> to the
4841 lttng add-context --jul --type='$app.myRetriever:intCtx'
4842 lttng add-context --jul --type='$app.myRetriever:strContext'
4845 <<basic-tracing-session-control,Start tracing>>:
4852 Run the compiled class:
4856 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4859 <<basic-tracing-session-control,Stop tracing>> and inspect the
4871 [[python-application]]
4872 === User space Python agent
4874 You can instrument a Python 2 or Python 3 application which uses the
4875 standard https://docs.python.org/3/library/logging.html[`logging`]
4878 Each log statement emits an LTTng event once the
4879 application module imports the
4880 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4883 .A Python application importing the LTTng-UST Python agent.
4884 image::python-app.png[]
4886 To use the LTTng-UST Python agent:
4888 . In the Python application's source code, import the LTTng-UST Python
4898 The LTTng-UST Python agent automatically adds its logging handler to the
4899 root logger at import time.
4901 Any log statement that the application executes before this import does
4902 not emit an LTTng event.
4904 IMPORTANT: The LTTng-UST Python agent must be
4905 <<installing-lttng,installed>>.
4907 . Use log statements and logging configuration as usual.
4908 Since the LTTng-UST Python agent adds a handler to the _root_
4909 logger, you can trace any log statement from any logger.
4911 .Use the LTTng-UST Python agent.
4922 logging.basicConfig()
4923 logger = logging.getLogger('my-logger')
4926 logger.debug('debug message')
4927 logger.info('info message')
4928 logger.warn('warn message')
4929 logger.error('error message')
4930 logger.critical('critical message')
4934 if __name__ == '__main__':
4938 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4939 logging handler which prints to the standard error stream, is not
4940 strictly required for LTTng-UST tracing to work, but in versions of
4941 Python preceding 3.2, you could see a warning message which indicates
4942 that no handler exists for the logger `my-logger`.
4944 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4945 <<enabling-disabling-events,create an event rule>> matching the
4946 `my-logger` Python logger, and <<basic-tracing-session-control,start
4952 lttng enable-event --python my-logger
4956 Run the Python script:
4963 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4973 You can use the opt:lttng-enable-event(1):--loglevel or
4974 opt:lttng-enable-event(1):--loglevel-only option of the
4975 man:lttng-enable-event(1) command to target a range of Python log levels
4976 or a specific Python log level.
4978 When an application imports the LTTng-UST Python agent, the agent tries
4979 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4980 <<start-sessiond,start the session daemon>> _before_ you run the Python
4981 application. If a session daemon is found, the agent tries to register
4982 to it during 5{nbsp}seconds, after which the application continues
4983 without LTTng tracing support. You can override this timeout value with
4984 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4987 If the session daemon stops while a Python application with an imported
4988 LTTng-UST Python agent runs, the agent retries to connect and to
4989 register to a session daemon every 3{nbsp}seconds. You can override this
4990 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4995 [[proc-lttng-logger-abi]]
4998 The `lttng-tracer` Linux kernel module, part of
4999 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
5000 path:{/proc/lttng-logger} when it's loaded. Any application can write
5001 text data to this file to emit an LTTng event.
5004 .An application writes to the LTTng logger file to emit an LTTng event.
5005 image::lttng-logger.png[]
5007 The LTTng logger is the quickest method--not the most efficient,
5008 however--to add instrumentation to an application. It is designed
5009 mostly to instrument shell scripts:
5013 echo "Some message, some $variable" > /proc/lttng-logger
5016 Any event that the LTTng logger emits is named `lttng_logger` and
5017 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5018 other instrumentation points in the kernel tracing domain, **any Unix
5019 user** can <<enabling-disabling-events,create an event rule>> which
5020 matches its event name, not only the root user or users in the tracing
5023 To use the LTTng logger:
5025 * From any application, write text data to the path:{/proc/lttng-logger}
5028 The `msg` field of `lttng_logger` event records contains the
5031 NOTE: The maximum message length of an LTTng logger event is
5032 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5033 than one event to contain the remaining data.
5035 You should not use the LTTng logger to trace a user application which
5036 can be instrumented in a more efficient way, namely:
5038 * <<c-application,C and $$C++$$ applications>>.
5039 * <<java-application,Java applications>>.
5040 * <<python-application,Python applications>>.
5042 .Use the LTTng logger.
5047 echo 'Hello, World!' > /proc/lttng-logger
5049 df --human-readable --print-type / > /proc/lttng-logger
5052 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5053 <<enabling-disabling-events,create an event rule>> matching the
5054 `lttng_logger` Linux kernel tracepoint, and
5055 <<basic-tracing-session-control,start tracing>>:
5060 lttng enable-event --kernel lttng_logger
5064 Run the Bash script:
5071 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5082 [[instrumenting-linux-kernel]]
5083 === LTTng kernel tracepoints
5085 NOTE: This section shows how to _add_ instrumentation points to the
5086 Linux kernel. The kernel's subsystems are already thoroughly
5087 instrumented at strategic places for LTTng when you
5088 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5092 There are two methods to instrument the Linux kernel:
5094 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5095 tracepoint which uses the `TRACE_EVENT()` API.
5097 Choose this if you want to instrumentation a Linux kernel tree with an
5098 instrumentation point compatible with ftrace, perf, and SystemTap.
5100 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5101 instrument an out-of-tree kernel module.
5103 Choose this if you don't need ftrace, perf, or SystemTap support.
5107 [[linux-add-lttng-layer]]
5108 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5110 This section shows how to add an LTTng layer to existing ftrace
5111 instrumentation using the `TRACE_EVENT()` API.
5113 This section does not document the `TRACE_EVENT()` macro. You can
5114 read the following articles to learn more about this API:
5116 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5117 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5118 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5120 The following procedure assumes that your ftrace tracepoints are
5121 correctly defined in their own header and that they are created in
5122 one source file using the `CREATE_TRACE_POINTS` definition.
5124 To add an LTTng layer over an existing ftrace tracepoint:
5126 . Make sure the following kernel configuration options are
5132 * `CONFIG_HIGH_RES_TIMERS`
5133 * `CONFIG_TRACEPOINTS`
5136 . Build the Linux source tree with your custom ftrace tracepoints.
5137 . Boot the resulting Linux image on your target system.
5139 Confirm that the tracepoints exist by looking for their names in the
5140 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5141 is your subsystem's name.
5143 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5149 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
5150 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
5151 cd lttng-modules-2.8.*
5155 . In dir:{instrumentation/events/lttng-module}, relative to the root
5156 of the LTTng-modules source tree, create a header file named
5157 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5158 LTTng-modules tracepoint definitions using the LTTng-modules
5161 Start with this template:
5165 .path:{instrumentation/events/lttng-module/my_subsys.h}
5168 #define TRACE_SYSTEM my_subsys
5170 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5171 #define _LTTNG_MY_SUBSYS_H
5173 #include "../../../probes/lttng-tracepoint-event.h"
5174 #include <linux/tracepoint.h>
5176 LTTNG_TRACEPOINT_EVENT(
5178 * Format is identical to TRACE_EVENT()'s version for the three
5179 * following macro parameters:
5182 TP_PROTO(int my_int, const char *my_string),
5183 TP_ARGS(my_int, my_string),
5185 /* LTTng-modules specific macros */
5187 ctf_integer(int, my_int_field, my_int)
5188 ctf_string(my_bar_field, my_bar)
5192 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5194 #include "../../../probes/define_trace.h"
5198 The entries in the `TP_FIELDS()` section are the list of fields for the
5199 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5200 ftrace's `TRACE_EVENT()` macro.
5202 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5203 complete description of the available `ctf_*()` macros.
5205 . Create the LTTng-modules probe's kernel module C source file,
5206 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5211 .path:{probes/lttng-probe-my-subsys.c}
5213 #include <linux/module.h>
5214 #include "../lttng-tracer.h"
5217 * Build-time verification of mismatch between mainline
5218 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5219 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5221 #include <trace/events/my_subsys.h>
5223 /* Create LTTng tracepoint probes */
5224 #define LTTNG_PACKAGE_BUILD
5225 #define CREATE_TRACE_POINTS
5226 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5228 #include "../instrumentation/events/lttng-module/my_subsys.h"
5230 MODULE_LICENSE("GPL and additional rights");
5231 MODULE_AUTHOR("Your name <your-email>");
5232 MODULE_DESCRIPTION("LTTng my_subsys probes");
5233 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5234 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5235 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5236 LTTNG_MODULES_EXTRAVERSION);
5240 . Edit path:{probes/Makefile} and add your new kernel module object
5241 next to the existing ones:
5245 .path:{probes/Makefile}
5249 obj-m += lttng-probe-module.o
5250 obj-m += lttng-probe-power.o
5252 obj-m += lttng-probe-my-subsys.o
5258 . Build and install the LTTng kernel modules:
5263 make KERNELDIR=/path/to/linux
5264 sudo make modules_install
5268 Replace `/path/to/linux` with the path to the Linux source tree where
5269 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5271 Note that you can also use the
5272 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5273 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5274 C code that need to be executed before the event fields are recorded.
5276 The best way to learn how to use the previous LTTng-modules macros is to
5277 inspect the existing LTTng-modules tracepoint definitions in the
5278 dir:{instrumentation/events/lttng-module} header files. Compare them
5279 with the Linux kernel mainline versions in the
5280 dir:{include/trace/events} directory of the Linux source tree.
5284 [[lttng-tracepoint-event-code]]
5285 ===== Use custom C code to access the data for tracepoint fields
5287 Although we recommended to always use the
5288 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5289 the arguments and fields of an LTTng-modules tracepoint when possible,
5290 sometimes you need a more complex process to access the data that the
5291 tracer records as event record fields. In other words, you need local
5292 variables and multiple C{nbsp}statements instead of simple
5293 argument-based expressions that you pass to the
5294 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5296 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5297 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5298 a block of C{nbsp}code to be executed before LTTng records the fields.
5299 The structure of this macro is:
5302 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5304 LTTNG_TRACEPOINT_EVENT_CODE(
5306 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5307 * version for the following three macro parameters:
5310 TP_PROTO(int my_int, const char *my_string),
5311 TP_ARGS(my_int, my_string),
5313 /* Declarations of custom local variables */
5316 unsigned long b = 0;
5317 const char *name = "(undefined)";
5318 struct my_struct *my_struct;
5322 * Custom code which uses both tracepoint arguments
5323 * (in TP_ARGS()) and local variables (in TP_locvar()).
5325 * Local variables are actually members of a structure pointed
5326 * to by the special variable tp_locvar.
5330 tp_locvar->a = my_int + 17;
5331 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5332 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5333 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5334 put_my_struct(tp_locvar->my_struct);
5343 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5344 * version for this, except that tp_locvar members can be
5345 * used in the argument expression parameters of
5346 * the ctf_*() macros.
5349 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5350 ctf_integer(int, my_struct_a, tp_locvar->a)
5351 ctf_string(my_string_field, my_string)
5352 ctf_string(my_struct_name, tp_locvar->name)
5357 IMPORTANT: The C code defined in `TP_code()` must not have any side
5358 effects when executed. In particular, the code must not allocate
5359 memory or get resources without deallocating this memory or putting
5360 those resources afterwards.
5363 [[instrumenting-linux-kernel-tracing]]
5364 ==== Load and unload a custom probe kernel module
5366 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5367 kernel module>> in the kernel before it can emit LTTng events.
5369 To load the default probe kernel modules and a custom probe kernel
5372 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5373 probe modules to load when starting a root <<lttng-sessiond,session
5377 .Load the `my_subsys`, `usb`, and the default probe modules.
5381 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5386 You only need to pass the subsystem name, not the whole kernel module
5389 To load _only_ a given custom probe kernel module:
5391 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5392 modules to load when starting a root session daemon:
5395 .Load only the `my_subsys` and `usb` probe modules.
5399 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5404 To confirm that a probe module is loaded:
5411 lsmod | grep lttng_probe_usb
5415 To unload the loaded probe modules:
5417 * Kill the session daemon with `SIGTERM`:
5422 sudo pkill lttng-sessiond
5426 You can also use man:modprobe(8)'s `--remove` option if the session
5427 daemon terminates abnormally.
5430 [[controlling-tracing]]
5433 Once an application or a Linux kernel is
5434 <<instrumenting,instrumented>> for LTTng tracing,
5437 This section is divided in topics on how to use the various
5438 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5439 command-line tool>>, to _control_ the LTTng daemons and tracers.
5441 NOTE: In the following subsections, we refer to an man:lttng(1) command
5442 using its man page name. For example, instead of _Run the `create`
5443 command to..._, we use _Run the man:lttng-create(1) command to..._.
5447 === Start a session daemon
5449 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5450 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5453 You will see the following error when you run a command while no session
5457 Error: No session daemon is available
5460 The only command that automatically runs a session daemon is
5461 man:lttng-create(1), which you use to
5462 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5463 this is most of the time the first operation that you do, sometimes it's
5464 not. Some examples are:
5466 * <<list-instrumentation-points,List the available instrumentation points>>.
5467 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5469 [[tracing-group]] Each Unix user must have its own running session
5470 daemon to trace user applications. The session daemon that the root user
5471 starts is the only one allowed to control the LTTng kernel tracer. Users
5472 that are part of the _tracing group_ can control the root session
5473 daemon. The default tracing group name is `tracing`; you can set it to
5474 something else with the opt:lttng-sessiond(8):--group option when you
5475 start the root session daemon.
5477 To start a user session daemon:
5479 * Run man:lttng-sessiond(8):
5484 lttng-sessiond --daemonize
5488 To start the root session daemon:
5490 * Run man:lttng-sessiond(8) as the root user:
5495 sudo lttng-sessiond --daemonize
5499 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5500 start the session daemon in foreground.
5502 To stop a session daemon, use man:kill(1) on its process ID (standard
5505 Note that some Linux distributions could manage the LTTng session daemon
5506 as a service. In this case, you should use the service manager to
5507 start, restart, and stop session daemons.
5510 [[creating-destroying-tracing-sessions]]
5511 === Create and destroy a tracing session
5513 Almost all the LTTng control operations happen in the scope of
5514 a <<tracing-session,tracing session>>, which is the dialogue between the
5515 <<lttng-sessiond,session daemon>> and you.
5517 To create a tracing session with a generated name:
5519 * Use the man:lttng-create(1) command:
5528 The created tracing session's name is `auto` followed by the
5531 To create a tracing session with a specific name:
5533 * Use the optional argument of the man:lttng-create(1) command:
5538 lttng create my-session
5542 Replace `my-session` with the specific tracing session name.
5544 LTTng appends the creation date to the created tracing session's name.
5546 LTTng writes the traces of a tracing session in
5547 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5548 name of the tracing session. Note that the env:LTTNG_HOME environment
5549 variable defaults to `$HOME` if not set.
5551 To output LTTng traces to a non-default location:
5553 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5558 lttng create --output=/tmp/some-directory my-session
5562 You may create as many tracing sessions as you wish.
5564 To list all the existing tracing sessions for your Unix user:
5566 * Use the man:lttng-list(1) command:
5575 When you create a tracing session, it is set as the _current tracing
5576 session_. The following man:lttng(1) commands operate on the current
5577 tracing session when you don't specify one:
5579 [role="list-3-cols"]
5595 To change the current tracing session:
5597 * Use the man:lttng-set-session(1) command:
5602 lttng set-session new-session
5606 Replace `new-session` by the name of the new current tracing session.
5608 When you are done tracing in a given tracing session, you can destroy
5609 it. This operation frees the resources taken by the tracing session
5610 to destroy; it does not destroy the trace data that LTTng wrote for
5611 this tracing session.
5613 To destroy the current tracing session:
5615 * Use the man:lttng-destroy(1) command:
5625 [[list-instrumentation-points]]
5626 === List the available instrumentation points
5628 The <<lttng-sessiond,session daemon>> can query the running instrumented
5629 user applications and the Linux kernel to get a list of available
5630 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5631 they are tracepoints and system calls. For the user space tracing
5632 domain, they are tracepoints. For the other tracing domains, they are
5635 To list the available instrumentation points:
5637 * Use the man:lttng-list(1) command with the requested tracing domain's
5641 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5642 must be a root user, or it must be a member of the
5643 <<tracing-group,tracing group>>).
5644 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5645 kernel system calls (your Unix user must be a root user, or it must be
5646 a member of the tracing group).
5647 * opt:lttng-list(1):--userspace: user space tracepoints.
5648 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5649 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5650 * opt:lttng-list(1):--python: Python loggers.
5653 .List the available user space tracepoints.
5657 lttng list --userspace
5661 .List the available Linux kernel system call tracepoints.
5665 lttng list --kernel --syscall
5670 [[enabling-disabling-events]]
5671 === Create and enable an event rule
5673 Once you <<creating-destroying-tracing-sessions,create a tracing
5674 session>>, you can create <<event,event rules>> with the
5675 man:lttng-enable-event(1) command.
5677 You specify each condition with a command-line option. The available
5678 condition options are shown in the following table.
5680 [role="growable",cols="asciidoc,asciidoc,default"]
5681 .Condition command-line options for the man:lttng-enable-event(1) command.
5683 |Option |Description |Applicable tracing domains
5689 . +--probe=__ADDR__+
5690 . +--function=__ADDR__+
5693 Instead of using the default _tracepoint_ instrumentation type, use:
5695 . A Linux system call.
5696 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5697 . The entry and return points of a Linux function (symbol or address).
5701 |First positional argument.
5704 Tracepoint or system call name. In the case of a Linux KProbe or
5705 function, this is a custom name given to the event rule. With the
5706 JUL, log4j, and Python domains, this is a logger name.
5708 With a tracepoint, logger, or system call name, the last character
5709 can be `*` to match anything that remains.
5716 . +--loglevel=__LEVEL__+
5717 . +--loglevel-only=__LEVEL__+
5720 . Match only tracepoints or log statements with a logging level at
5721 least as severe as +__LEVEL__+.
5722 . Match only tracepoints or log statements with a logging level
5723 equal to +__LEVEL__+.
5725 See man:lttng-enable-event(1) for the list of available logging level
5728 |User space, JUL, log4j, and Python.
5730 |+--exclude=__EXCLUSIONS__+
5733 When you use a `*` character at the end of the tracepoint or logger
5734 name (first positional argument), exclude the specific names in the
5735 comma-delimited list +__EXCLUSIONS__+.
5738 User space, JUL, log4j, and Python.
5740 |+--filter=__EXPR__+
5743 Match only events which satisfy the expression +__EXPR__+.
5745 See man:lttng-enable-event(1) to learn more about the syntax of a
5752 You attach an event rule to a <<channel,channel>> on creation. If you do
5753 not specify the channel with the opt:lttng-enable-event(1):--channel
5754 option, and if the event rule to create is the first in its
5755 <<domain,tracing domain>> for a given tracing session, then LTTng
5756 creates a _default channel_ for you. This default channel is reused in
5757 subsequent invocations of the man:lttng-enable-event(1) command for the
5758 same tracing domain.
5760 An event rule is always enabled at creation time.
5762 The following examples show how you can combine the previous
5763 command-line options to create simple to more complex event rules.
5765 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5769 lttng enable-event --kernel sched_switch
5773 .Create an event rule matching four Linux kernel system calls (default channel).
5777 lttng enable-event --kernel --syscall open,write,read,close
5781 .Create event rules matching tracepoints with filter expressions (default channel).
5785 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5790 lttng enable-event --kernel --all \
5791 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5796 lttng enable-event --jul my_logger \
5797 --filter='$app.retriever:cur_msg_id > 3'
5800 IMPORTANT: Make sure to always quote the filter string when you
5801 use man:lttng(1) from a shell.
5804 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5808 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5811 IMPORTANT: Make sure to always quote the wildcard character when you
5812 use man:lttng(1) from a shell.
5815 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5819 lttng enable-event --python my-app.'*' \
5820 --exclude='my-app.module,my-app.hello'
5824 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5828 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5832 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5836 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5840 The event rules of a given channel form a whitelist: as soon as an
5841 emitted event passes one of them, LTTng can record the event. For
5842 example, an event named `my_app:my_tracepoint` emitted from a user space
5843 tracepoint with a `TRACE_ERROR` log level passes both of the following
5848 lttng enable-event --userspace my_app:my_tracepoint
5849 lttng enable-event --userspace my_app:my_tracepoint \
5850 --loglevel=TRACE_INFO
5853 The second event rule is redundant: the first one includes
5857 [[disable-event-rule]]
5858 === Disable an event rule
5860 To disable an event rule that you <<enabling-disabling-events,created>>
5861 previously, use the man:lttng-disable-event(1) command. This command
5862 disables _all_ the event rules (of a given tracing domain and channel)
5863 which match an instrumentation point. The other conditions are not
5864 supported as of LTTng{nbsp}{revision}.
5866 The LTTng tracer does not record an emitted event which passes
5867 a _disabled_ event rule.
5869 .Disable an event rule matching a Python logger (default channel).
5873 lttng disable-event --python my-logger
5877 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5881 lttng disable-event --jul '*'
5885 .Disable _all_ the event rules of the default channel.
5887 The opt:lttng-disable-event(1):--all-events option is not, like the
5888 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5889 equivalent of the event name `*` (wildcard): it disables _all_ the event
5890 rules of a given channel.
5894 lttng disable-event --jul --all-events
5898 NOTE: You cannot delete an event rule once you create it.
5902 === Get the status of a tracing session
5904 To get the status of the current tracing session, that is, its
5905 parameters, its channels, event rules, and their attributes:
5907 * Use the man:lttng-status(1) command:
5917 To get the status of any tracing session:
5919 * Use the man:lttng-list(1) command with the tracing session's name:
5924 lttng list my-session
5928 Replace `my-session` with the desired tracing session's name.
5931 [[basic-tracing-session-control]]
5932 === Start and stop a tracing session
5934 Once you <<creating-destroying-tracing-sessions,create a tracing
5936 <<enabling-disabling-events,create one or more event rules>>,
5937 you can start and stop the tracers for this tracing session.
5939 To start tracing in the current tracing session:
5941 * Use the man:lttng-start(1) command:
5950 LTTng is very flexible: you can launch user applications before
5951 or after the you start the tracers. The tracers only record the events
5952 if they pass enabled event rules and if they occur while the tracers are
5955 To stop tracing in the current tracing session:
5957 * Use the man:lttng-stop(1) command:
5966 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5967 records>> or lost sub-buffers since the last time you ran
5968 man:lttng-start(1), warnings are printed when you run the
5969 man:lttng-stop(1) command.
5972 [[enabling-disabling-channels]]
5973 === Create a channel
5975 Once you create a tracing session, you can create a <<channel,channel>>
5976 with the man:lttng-enable-channel(1) command.
5978 Note that LTTng automatically creates a default channel when, for a
5979 given <<domain,tracing domain>>, no channels exist and you
5980 <<enabling-disabling-events,create>> the first event rule. This default
5981 channel is named `channel0` and its attributes are set to reasonable
5982 values. Therefore, you only need to create a channel when you need
5983 non-default attributes.
5985 You specify each non-default channel attribute with a command-line
5986 option when you use the man:lttng-enable-channel(1) command. The
5987 available command-line options are:
5989 [role="growable",cols="asciidoc,asciidoc"]
5990 .Command-line options for the man:lttng-enable-channel(1) command.
5992 |Option |Description
5998 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5999 the default _discard_ mode.
6001 |`--buffers-pid` (user space tracing domain only)
6004 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6005 instead of the default per-user buffering scheme.
6007 |+--subbuf-size=__SIZE__+
6010 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6011 either for each Unix user (default), or for each instrumented process.
6013 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6015 |+--num-subbuf=__COUNT__+
6018 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6019 for each Unix user (default), or for each instrumented process.
6021 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6023 |+--tracefile-size=__SIZE__+
6026 Set the maximum size of each trace file that this channel writes within
6027 a stream to +__SIZE__+ bytes instead of no maximum.
6029 See <<tracefile-rotation,Trace file count and size>>.
6031 |+--tracefile-count=__COUNT__+
6034 Limit the number of trace files that this channel creates to
6035 +__COUNT__+ channels instead of no limit.
6037 See <<tracefile-rotation,Trace file count and size>>.
6039 |+--switch-timer=__PERIODUS__+
6042 Set the <<channel-switch-timer,switch timer period>>
6043 to +__PERIODUS__+{nbsp}µs.
6045 |+--read-timer=__PERIODUS__+
6048 Set the <<channel-read-timer,read timer period>>
6049 to +__PERIODUS__+{nbsp}µs.
6051 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6054 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6058 You can only create a channel in the Linux kernel and user space
6059 <<domain,tracing domains>>: other tracing domains have their own channel
6060 created on the fly when <<enabling-disabling-events,creating event
6065 Because of a current LTTng limitation, you must create all channels
6066 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6067 tracing session, that is, before the first time you run
6070 Since LTTng automatically creates a default channel when you use the
6071 man:lttng-enable-event(1) command with a specific tracing domain, you
6072 cannot, for example, create a Linux kernel event rule, start tracing,
6073 and then create a user space event rule, because no user space channel
6074 exists yet and it's too late to create one.
6076 For this reason, make sure to configure your channels properly
6077 before starting the tracers for the first time!
6080 The following examples show how you can combine the previous
6081 command-line options to create simple to more complex channels.
6083 .Create a Linux kernel channel with default attributes.
6087 lttng enable-channel --kernel my-channel
6091 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6095 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6096 --buffers-pid my-channel
6100 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6104 lttng enable-channel --kernel --tracefile-count=8 \
6105 --tracefile-size=4194304 my-channel
6109 .Create a user space channel in overwrite (or _flight recorder_) mode.
6113 lttng enable-channel --userspace --overwrite my-channel
6117 You can <<enabling-disabling-events,create>> the same event rule in
6118 two different channels:
6122 lttng enable-event --userspace --channel=my-channel app:tp
6123 lttng enable-event --userspace --channel=other-channel app:tp
6126 If both channels are enabled, when a tracepoint named `app:tp` is
6127 reached, LTTng records two events, one for each channel.
6131 === Disable a channel
6133 To disable a specific channel that you <<enabling-disabling-channels,created>>
6134 previously, use the man:lttng-disable-channel(1) command.
6136 .Disable a specific Linux kernel channel.
6140 lttng disable-channel --kernel my-channel
6144 The state of a channel precedes the individual states of event rules
6145 attached to it: event rules which belong to a disabled channel, even if
6146 they are enabled, are also considered disabled.
6150 === Add context fields to a channel
6152 Event record fields in trace files provide important information about
6153 events that occured previously, but sometimes some external context may
6154 help you solve a problem faster. Examples of context fields are:
6156 * The **process ID**, **thread ID**, **process name**, and
6157 **process priority** of the thread in which the event occurs.
6158 * The **hostname** of the system on which the event occurs.
6159 * The current values of many possible **performance counters** using
6161 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6163 ** Branch instructions, misses, and loads.
6165 * Any context defined at the application level (supported for the
6166 JUL and log4j <<domain,tracing domains>>).
6168 To get the full list of available context fields, see
6169 `lttng add-context --list`. Some context fields are reserved for a
6170 specific <<domain,tracing domain>> (Linux kernel or user space).
6172 You add context fields to <<channel,channels>>. All the events
6173 that a channel with added context fields records contain those fields.
6175 To add context fields to one or all the channels of a given tracing
6178 * Use the man:lttng-add-context(1) command.
6180 .Add context fields to all the channels of the current tracing session.
6182 The following command line adds the virtual process identifier and
6183 the per-thread CPU cycles count fields to all the user space channels
6184 of the current tracing session.
6188 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6192 .Add a context field to a specific channel.
6194 The following command line adds the thread identifier context field
6195 to the Linux kernel channel named `my-channel` in the current
6200 lttng add-context --kernel --channel=my-channel --type=tid
6204 .Add an application-specific context field to a specific channel.
6206 The following command line adds the `cur_msg_id` context field of the
6207 `retriever` context retriever for all the instrumented
6208 <<java-application,Java applications>> recording <<event,event records>>
6209 in the channel named `my-channel`:
6213 lttng add-context --kernel --channel=my-channel \
6214 --type='$app:retriever:cur_msg_id'
6217 IMPORTANT: Make sure to always quote the `$` character when you
6218 use man:lttng-add-context(1) from a shell.
6221 NOTE: You cannot remove context fields from a channel once you add it.
6226 === Track process IDs
6228 It's often useful to allow only specific process IDs (PIDs) to emit
6229 events. For example, you may wish to record all the system calls made by
6230 a given process (à la http://linux.die.net/man/1/strace[strace]).
6232 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6233 purpose. Both commands operate on a whitelist of process IDs. You _add_
6234 entries to this whitelist with the man:lttng-track(1) command and remove
6235 entries with the man:lttng-untrack(1) command. Any process which has one
6236 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6237 an enabled <<event,event rule>>.
6239 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6240 process with a given tracked ID exit and another process be given this
6241 ID, then the latter would also be allowed to emit events.
6243 .Track and untrack process IDs.
6245 For the sake of the following example, assume the target system has 16
6249 <<creating-destroying-tracing-sessions,create a tracing session>>,
6250 the whitelist contains all the possible PIDs:
6253 .All PIDs are tracked.
6254 image::track-all.png[]
6256 When the whitelist is full and you use the man:lttng-track(1) command to
6257 specify some PIDs to track, LTTng first clears the whitelist, then it
6258 tracks the specific PIDs. After:
6262 lttng track --pid=3,4,7,10,13
6268 .PIDs 3, 4, 7, 10, and 13 are tracked.
6269 image::track-3-4-7-10-13.png[]
6271 You can add more PIDs to the whitelist afterwards:
6275 lttng track --pid=1,15,16
6281 .PIDs 1, 15, and 16 are added to the whitelist.
6282 image::track-1-3-4-7-10-13-15-16.png[]
6284 The man:lttng-untrack(1) command removes entries from the PID tracker's
6285 whitelist. Given the previous example, the following command:
6289 lttng untrack --pid=3,7,10,13
6292 leads to this whitelist:
6295 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6296 image::track-1-4-15-16.png[]
6298 LTTng can track all possible PIDs again using the opt:track(1):--all
6303 lttng track --pid --all
6306 The result is, again:
6309 .All PIDs are tracked.
6310 image::track-all.png[]
6313 .Track only specific PIDs
6315 A very typical use case with PID tracking is to start with an empty
6316 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6317 then add PIDs manually while tracers are active. You can accomplish this
6318 by using the opt:lttng-untrack(1):--all option of the
6319 man:lttng-untrack(1) command to clear the whitelist after you
6320 <<creating-destroying-tracing-sessions,create a tracing session>>:
6324 lttng untrack --pid --all
6330 .No PIDs are tracked.
6331 image::untrack-all.png[]
6333 If you trace with this whitelist configuration, the tracer records no
6334 events for this <<domain,tracing domain>> because no processes are
6335 tracked. You can use the man:lttng-track(1) command as usual to track
6336 specific PIDs, for example:
6340 lttng track --pid=6,11
6346 .PIDs 6 and 11 are tracked.
6347 image::track-6-11.png[]
6352 [[saving-loading-tracing-session]]
6353 === Save and load tracing session configurations
6355 Configuring a <<tracing-session,tracing session>> can be long. Some of
6356 the tasks involved are:
6358 * <<enabling-disabling-channels,Create channels>> with
6359 specific attributes.
6360 * <<adding-context,Add context fields>> to specific channels.
6361 * <<enabling-disabling-events,Create event rules>> with specific log
6362 level and filter conditions.
6364 If you use LTTng to solve real world problems, chances are you have to
6365 record events using the same tracing session setup over and over,
6366 modifying a few variables each time in your instrumented program
6367 or environment. To avoid constant tracing session reconfiguration,
6368 the man:lttng(1) command-line tool can save and load tracing session
6369 configurations to/from XML files.
6371 To save a given tracing session configuration:
6373 * Use the man:lttng-save(1) command:
6378 lttng save my-session
6382 Replace `my-session` with the name of the tracing session to save.
6384 LTTng saves tracing session configurations to
6385 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6386 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6387 the opt:lttng-save(1):--output-path option to change this destination
6390 LTTng saves all configuration parameters, for example:
6392 * The tracing session name.
6393 * The trace data output path.
6394 * The channels with their state and all their attributes.
6395 * The context fields you added to channels.
6396 * The event rules with their state, log level and filter conditions.
6398 To load a tracing session:
6400 * Use the man:lttng-load(1) command:
6405 lttng load my-session
6409 Replace `my-session` with the name of the tracing session to load.
6411 When LTTng loads a configuration, it restores your saved tracing session
6412 as if you just configured it manually.
6414 See man:lttng(1) for the complete list of command-line options. You
6415 can also save and load all many sessions at a time, and decide in which
6416 directory to output the XML files.
6419 [[sending-trace-data-over-the-network]]
6420 === Send trace data over the network
6422 LTTng can send the recorded trace data to a remote system over the
6423 network instead of writing it to the local file system.
6425 To send the trace data over the network:
6427 . On the _remote_ system (which can also be the target system),
6428 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6437 . On the _target_ system, create a tracing session configured to
6438 send trace data over the network:
6443 lttng create my-session --set-url=net://remote-system
6447 Replace `remote-system` by the host name or IP address of the
6448 remote system. See man:lttng-create(1) for the exact URL format.
6450 . On the target system, use the man:lttng(1) command-line tool as usual.
6451 When tracing is active, the target's consumer daemon sends sub-buffers
6452 to the relay daemon running on the remote system intead of flushing
6453 them to the local file system. The relay daemon writes the received
6454 packets to the local file system.
6456 The relay daemon writes trace files to
6457 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6458 +__hostname__+ is the host name of the target system and +__session__+
6459 is the tracing session name. Note that the env:LTTNG_HOME environment
6460 variable defaults to `$HOME` if not set. Use the
6461 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6462 trace files to another base directory.
6467 === View events as LTTng emits them (noch:{LTTng} live)
6469 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6470 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6471 display events as LTTng emits them on the target system while tracing is
6474 The relay daemon creates a _tee_: it forwards the trace data to both
6475 the local file system and to connected live viewers:
6478 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6483 . On the _target system_, create a <<tracing-session,tracing session>>
6489 lttng create --live my-session
6493 This spawns a local relay daemon.
6495 . Start the live viewer and configure it to connect to the relay
6496 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6501 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6508 * `hostname` with the host name of the target system.
6509 * `my-session` with the name of the tracing session to view.
6512 . Configure the tracing session as usual with the man:lttng(1)
6513 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6515 You can list the available live tracing sessions with Babeltrace:
6519 babeltrace --input-format=lttng-live net://localhost
6522 You can start the relay daemon on another system. In this case, you need
6523 to specify the relay daemon's URL when you create the tracing session
6524 with the opt:lttng-create(1):--set-url option. You also need to replace
6525 `localhost` in the procedure above with the host name of the system on
6526 which the relay daemon is running.
6528 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6529 command-line options.
6533 [[taking-a-snapshot]]
6534 === Take a snapshot of the current sub-buffers of a tracing session
6536 The normal behavior of LTTng is to append full sub-buffers to growing
6537 trace data files. This is ideal to keep a full history of the events
6538 that occurred on the target system, but it can
6539 represent too much data in some situations. For example, you may wish
6540 to trace your application continuously until some critical situation
6541 happens, in which case you only need the latest few recorded
6542 events to perform the desired analysis, not multi-gigabyte trace files.
6544 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6545 current sub-buffers of a given <<tracing-session,tracing session>>.
6546 LTTng can write the snapshot to the local file system or send it over
6551 . Create a tracing session in _snapshot mode_:
6556 lttng create --snapshot my-session
6560 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6561 <<channel,channels>> created in this mode is automatically set to
6562 _overwrite_ (flight recorder mode).
6564 . Configure the tracing session as usual with the man:lttng(1)
6565 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6567 . **Optional**: When you need to take a snapshot,
6568 <<basic-tracing-session-control,stop tracing>>.
6570 You can take a snapshot when the tracers are active, but if you stop
6571 them first, you are sure that the data in the sub-buffers does not
6572 change before you actually take the snapshot.
6579 lttng snapshot record --name=my-first-snapshot
6583 LTTng writes the current sub-buffers of all the current tracing
6584 session's channels to trace files on the local file system. Those trace
6585 files have `my-first-snapshot` in their name.
6587 There is no difference between the format of a normal trace file and the
6588 format of a snapshot: viewers of LTTng traces also support LTTng
6591 By default, LTTng writes snapshot files to the path shown by
6592 `lttng snapshot list-output`. You can change this path or decide to send
6593 snapshots over the network using either:
6595 . An output path or URL that you specify when you create the
6597 . An snapshot output path or URL that you add using
6598 `lttng snapshot add-output`
6599 . An output path or URL that you provide directly to the
6600 `lttng snapshot record` command.
6602 Method 3 overrides method 2, which overrides method 1. When you
6603 specify a URL, a relay daemon must listen on a remote system (see
6604 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6609 === Use the machine interface
6611 With any command of the man:lttng(1) command-line tool, you can set the
6612 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6613 XML machine interface output, for example:
6617 lttng --mi=xml enable-event --kernel --syscall open
6620 A schema definition (XSD) is
6621 https://github.com/lttng/lttng-tools/blob/stable-2.8/src/common/mi-lttng-3.0.xsd[available]
6622 to ease the integration with external tools as much as possible.
6626 [[metadata-regenerate]]
6627 === Regenerate the metadata of an LTTng trace
6629 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6630 data stream files and a metadata file. This metadata file contains,
6631 amongst other things, information about the offset of the clock sources
6632 used to timestamp <<event,event records>> when tracing.
6634 If, once a <<tracing-session,tracing session>> is
6635 <<basic-tracing-session-control,started>>, a major
6636 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6637 happens, the trace's clock offset also needs to be updated. You
6638 can use the man:lttng-metadata(1) command to do so.
6640 The main use case of this command is to allow a system to boot with
6641 an incorrect wall time and trace it with LTTng before its wall time
6642 is corrected. Once the system is known to be in a state where its
6643 wall time is correct, it can run `lttng metadata regenerate`.
6645 To regenerate the metadata of an LTTng trace:
6647 * Use the `regenerate` action of the man:lttng-metadata(1) command:
6652 lttng metadata regenerate
6658 `lttng metadata regenerate` has the following limitations:
6660 * Tracing session <<creating-destroying-tracing-sessions,created>>
6662 * User space <<channel,channels>>, if any, using
6663 <<channel-buffering-schemes,per-user buffering>>.
6668 [[persistent-memory-file-systems]]
6669 === Record trace data on persistent memory file systems
6671 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6672 (NVRAM) is random-access memory that retains its information when power
6673 is turned off (non-volatile). Systems with such memory can store data
6674 structures in RAM and retrieve them after a reboot, without flushing
6675 to typical _storage_.
6677 Linux supports NVRAM file systems thanks to either
6678 http://pramfs.sourceforge.net/[PRAMFS] or
6679 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6680 (requires Linux 4.1+).
6682 This section does not describe how to operate such file systems;
6683 we assume that you have a working persistent memory file system.
6685 When you create a <<tracing-session,tracing session>>, you can specify
6686 the path of the shared memory holding the sub-buffers. If you specify a
6687 location on an NVRAM file system, then you can retrieve the latest
6688 recorded trace data when the system reboots after a crash.
6690 To record trace data on a persistent memory file system and retrieve the
6691 trace data after a system crash:
6693 . Create a tracing session with a sub-buffer shared memory path located
6694 on an NVRAM file system:
6699 lttng create --shm-path=/path/to/shm
6703 . Configure the tracing session as usual with the man:lttng(1)
6704 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6706 . After a system crash, use the man:lttng-crash(1) command-line tool to
6707 view the trace data recorded on the NVRAM file system:
6712 lttng-crash /path/to/shm
6716 The binary layout of the ring buffer files is not exactly the same as
6717 the trace files layout. This is why you need to use man:lttng-crash(1)
6718 instead of your preferred trace viewer directly.
6720 To convert the ring buffer files to LTTng trace files:
6722 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6727 lttng-crash --extract=/path/to/trace /path/to/shm
6735 [[lttng-modules-ref]]
6736 === noch:{LTTng-modules}
6739 [[lttng-modules-tp-fields]]
6740 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6742 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6743 tracepoint fields, which must be listed within `TP_FIELDS()` in
6744 `LTTNG_TRACEPOINT_EVENT()`, are:
6746 [role="func-desc growable",cols="asciidoc,asciidoc"]
6747 .Available macros to define LTTng-modules tracepoint fields
6749 |Macro |Description and parameters
6752 +ctf_integer(__t__, __n__, __e__)+
6754 +ctf_integer_nowrite(__t__, __n__, __e__)+
6756 +ctf_user_integer(__t__, __n__, __e__)+
6758 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6760 Standard integer, displayed in base 10.
6763 Integer C type (`int`, `long`, `size_t`, ...).
6769 Argument expression.
6772 +ctf_integer_hex(__t__, __n__, __e__)+
6774 +ctf_user_integer_hex(__t__, __n__, __e__)+
6776 Standard integer, displayed in base 16.
6785 Argument expression.
6787 |+ctf_integer_oct(__t__, __n__, __e__)+
6789 Standard integer, displayed in base 8.
6798 Argument expression.
6801 +ctf_integer_network(__t__, __n__, __e__)+
6803 +ctf_user_integer_network(__t__, __n__, __e__)+
6805 Integer in network byte order (big-endian), displayed in base 10.
6814 Argument expression.
6817 +ctf_integer_network_hex(__t__, __n__, __e__)+
6819 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6821 Integer in network byte order, displayed in base 16.
6830 Argument expression.
6833 +ctf_string(__n__, __e__)+
6835 +ctf_string_nowrite(__n__, __e__)+
6837 +ctf_user_string(__n__, __e__)+
6839 +ctf_user_string_nowrite(__n__, __e__)+
6841 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6847 Argument expression.
6850 +ctf_array(__t__, __n__, __e__, __s__)+
6852 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6854 +ctf_user_array(__t__, __n__, __e__, __s__)+
6856 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6858 Statically-sized array of integers.
6861 Array element C type.
6867 Argument expression.
6873 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
6875 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6877 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
6879 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6881 Statically-sized array of bits.
6883 The type of +__e__+ must be an integer type. +__s__+ is the number
6884 of elements of such type in +__e__+, not the number of bits.
6887 Array element C type.
6893 Argument expression.
6899 +ctf_array_text(__t__, __n__, __e__, __s__)+
6901 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6903 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6905 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6907 Statically-sized array, printed as text.
6909 The string does not need to be null-terminated.
6912 Array element C type (always `char`).
6918 Argument expression.
6924 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6926 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6928 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6930 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6932 Dynamically-sized array of integers.
6934 The type of +__E__+ must be unsigned.
6937 Array element C type.
6943 Argument expression.
6946 Length expression C type.
6952 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6954 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6956 Dynamically-sized array of integers, displayed in base 16.
6958 The type of +__E__+ must be unsigned.
6961 Array element C type.
6967 Argument expression.
6970 Length expression C type.
6975 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
6977 Dynamically-sized array of integers in network byte order (big-endian),
6978 displayed in base 10.
6980 The type of +__E__+ must be unsigned.
6983 Array element C type.
6989 Argument expression.
6992 Length expression C type.
6998 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7000 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7002 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7004 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7006 Dynamically-sized array of bits.
7008 The type of +__e__+ must be an integer type. +__s__+ is the number
7009 of elements of such type in +__e__+, not the number of bits.
7011 The type of +__E__+ must be unsigned.
7014 Array element C type.
7020 Argument expression.
7023 Length expression C type.
7029 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7031 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7033 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7035 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7037 Dynamically-sized array, displayed as text.
7039 The string does not need to be null-terminated.
7041 The type of +__E__+ must be unsigned.
7043 The behaviour is undefined if +__e__+ is `NULL`.
7046 Sequence element C type (always `char`).
7052 Argument expression.
7055 Length expression C type.
7061 Use the `_user` versions when the argument expression, `e`, is
7062 a user space address. In the cases of `ctf_user_integer*()` and
7063 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7066 The `_nowrite` versions omit themselves from the session trace, but are
7067 otherwise identical. This means the `_nowrite` fields won't be written
7068 in the recorded trace. Their primary purpose is to make some
7069 of the event context available to the
7070 <<enabling-disabling-events,event filters>> without having to
7071 commit the data to sub-buffers.
7077 Terms related to LTTng and to tracing in general:
7080 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7081 the cmd:babeltrace command, some libraries, and Python bindings.
7083 <<channel-buffering-schemes,buffering scheme>>::
7084 A layout of sub-buffers applied to a given channel.
7086 <<channel,channel>>::
7087 An entity which is responsible for a set of ring buffers.
7089 <<event,Event rules>> are always attached to a specific channel.
7092 A reference of time for a tracer.
7094 <<lttng-consumerd,consumer daemon>>::
7095 A process which is responsible for consuming the full sub-buffers
7096 and write them to a file system or send them over the network.
7098 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7099 mode in which the tracer _discards_ new event records when there's no
7100 sub-buffer space left to store them.
7103 The consequence of the execution of an instrumentation
7104 point, like a tracepoint that you manually place in some source code,
7105 or a Linux kernel KProbe.
7107 An event is said to _occur_ at a specific time. Different actions can
7108 be taken upon the occurance of an event, like record the event's payload
7111 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7112 The mechanism by which event records of a given channel are lost
7113 (not recorded) when there is no sub-buffer space left to store them.
7115 [[def-event-name]]event name::
7116 The name of an event, which is also the name of the event record.
7117 This is also called the _instrumentation point name_.
7120 A record, in a trace, of the payload of an event which occured.
7122 <<event,event rule>>::
7123 Set of conditions which must be satisfied for one or more occuring
7124 events to be recorded.
7126 `java.util.logging`::
7128 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7130 <<instrumenting,instrumentation>>::
7131 The use of LTTng probes to make a piece of software traceable.
7133 instrumentation point::
7134 A point in the execution path of a piece of software that, when
7135 reached by this execution, can emit an event.
7137 instrumentation point name::
7138 See _<<def-event-name,event name>>_.
7141 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7142 developed by the Apache Software Foundation.
7145 Level of severity of a log statement or user space
7146 instrumentation point.
7149 The _Linux Trace Toolkit: next generation_ project.
7151 <<lttng-cli,cmd:lttng>>::
7152 A command-line tool provided by the LTTng-tools project which you
7153 can use to send and receive control messages to and from a
7157 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7158 which is a set of analyzing programs that are used to obtain a
7159 higher level view of an LTTng trace.
7161 cmd:lttng-consumerd::
7162 The name of the consumer daemon program.
7165 A utility provided by the LTTng-tools project which can convert
7166 ring buffer files (usually
7167 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7170 LTTng Documentation::
7173 <<lttng-live,LTTng live>>::
7174 A communication protocol between the relay daemon and live viewers
7175 which makes it possible to see events "live", as they are received by
7178 <<lttng-modules,LTTng-modules>>::
7179 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7180 which contains the Linux kernel modules to make the Linux kernel
7181 instrumentation points available for LTTng tracing.
7184 The name of the relay daemon program.
7186 cmd:lttng-sessiond::
7187 The name of the session daemon program.
7190 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7191 contains the various programs and libraries used to
7192 <<controlling-tracing,control tracing>>.
7194 <<lttng-ust,LTTng-UST>>::
7195 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7196 contains libraries to instrument user applications.
7198 <<lttng-ust-agents,LTTng-UST Java agent>>::
7199 A Java package provided by the LTTng-UST project to allow the
7200 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7203 <<lttng-ust-agents,LTTng-UST Python agent>>::
7204 A Python package provided by the LTTng-UST project to allow the
7205 LTTng instrumentation of Python logging statements.
7207 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7208 The event loss mode in which new event records overwrite older
7209 event records when there's no sub-buffer space left to store them.
7211 <<channel-buffering-schemes,per-process buffering>>::
7212 A buffering scheme in which each instrumented process has its own
7213 sub-buffers for a given user space channel.
7215 <<channel-buffering-schemes,per-user buffering>>::
7216 A buffering scheme in which all the processes of a Unix user share the
7217 same sub-buffer for a given user space channel.
7219 <<lttng-relayd,relay daemon>>::
7220 A process which is responsible for receiving the trace data sent by
7221 a distant consumer daemon.
7224 A set of sub-buffers.
7226 <<lttng-sessiond,session daemon>>::
7227 A process which receives control commands from you and orchestrates
7228 the tracers and various LTTng daemons.
7230 <<taking-a-snapshot,snapshot>>::
7231 A copy of the current data of all the sub-buffers of a given tracing
7232 session, saved as trace files.
7235 One part of an LTTng ring buffer which contains event records.
7238 The time information attached to an event when it is emitted.
7241 A set of files which are the concatenations of one or more
7242 flushed sub-buffers.
7245 The action of recording the events emitted by an application
7246 or by a system, or to initiate such recording by controlling
7250 The http://tracecompass.org[Trace Compass] project and application.
7253 An instrumentation point using the tracepoint mechanism of the Linux
7254 kernel or of LTTng-UST.
7256 tracepoint definition::
7257 The definition of a single tracepoint.
7260 The name of a tracepoint.
7262 tracepoint provider::
7263 A set of functions providing tracepoints to an instrumented user
7266 Not to be confused with a _tracepoint provider package_: many tracepoint
7267 providers can exist within a tracepoint provider package.
7269 tracepoint provider package::
7270 One or more tracepoint providers compiled as an object file or as
7274 A software which records emitted events.
7276 <<domain,tracing domain>>::
7277 A namespace for event sources.
7280 The Unix group in which a Unix user can be to be allowed to trace the
7283 <<tracing-session,tracing session>>::
7284 A stateful dialogue between you and a <<lttng-sessiond,session
7288 An application running in user space, as opposed to a Linux kernel
7289 module, for example.