1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/warning-not-maintained.txt[]
13 include::../common/welcome.txt[]
16 include::../common/audience.txt[]
20 === What's in this documentation?
22 The LTTng Documentation is divided into the following sections:
24 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
25 rudiments of software tracing and the rationale behind the
28 You can skip this section if you’re familiar with software tracing and
29 with the LTTng project.
31 * **<<installing-lttng,Installation>>** describes the steps to
32 install the LTTng packages on common Linux distributions and from
35 You can skip this section if you already properly installed LTTng on
38 * **<<getting-started,Quick start>>** is a concise guide to
39 getting started quickly with LTTng kernel and user space tracing.
41 We recommend this section if you're new to LTTng or to software tracing
44 You can skip this section if you're not new to LTTng.
46 * **<<core-concepts,Core concepts>>** explains the concepts at
49 It's a good idea to become familiar with the core concepts
50 before attempting to use the toolkit.
52 * **<<plumbing,Components of LTTng>>** describes the various components
53 of the LTTng machinery, like the daemons, the libraries, and the
54 command-line interface.
55 * **<<instrumenting,Instrumentation>>** shows different ways to
56 instrument user applications and the Linux kernel.
58 Instrumenting source code is essential to provide a meaningful
61 You can skip this section if you do not have a programming background.
63 * **<<controlling-tracing,Tracing control>>** is divided into topics
64 which demonstrate how to use the vast array of features that
65 LTTng{nbsp}{revision} offers.
66 * **<<reference,Reference>>** contains reference links and tables.
67 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
68 to LTTng or to the field of software tracing.
71 include::../common/convention.txt[]
74 include::../common/acknowledgements.txt[]
78 == What's new in LTTng {revision}?
80 * **Tracing control**:
81 ** Dynamic filter support for <<event,event rules>> in the Linux kernel
82 <<domain,tracing domain>>. For example:
87 lttng enable-event --kernel irq_handler_entry --filter='irq == 28'
91 ** Wildcard support in the instrumentation point name of an event rule
92 in the Linux kernel tracing domain. For example:
97 lttng enable-event --kernel 'sched_*'
101 ** New `lttng track` and `lttng untrack` commands to make
102 <<pid-tracking,PID tracking>> super-fast for both the Linux kernel
103 and the user space tracing domains.
105 When LTTng _tracks_ one or more PIDs, only the processes having those PIDs
106 can emit events for a given tracing session.
108 ** New `--shm-path` option of the `lttng create` command to specify the
109 path where LTTng creates the shared memory holding the ring buffers.
111 This feature is useful when used with persistent memory file systems to
112 extract the latest recorded trace data in the event of a crash requiring
115 The new man:lttng-crash(1) command-line utility can extract trace data
116 from such a file (see <<persistent-memory-file-systems,Record trace data
117 on persistent memory file systems>>).
119 * **User space tracing**:
120 ** New <<python-application,LTTng-UST Python agent>> which makes it easy
121 to trace existing Python applications that are using the standard
122 https://docs.python.org/3/howto/logging.html[`logging` package].
124 This agent is compatible with both the Python 2 and Python 3 languages.
126 ** New <<tracelog,`tracelog()`>> facility to ease the migration from
129 `tracelog()` is similar to <<tracef,`tracef()`>>,
130 but it accepts an additional log level parameter.
132 ** Plugin support in LTTng-UST to provide a custom clock source and to
133 retrieve the current CPU number.
135 This feature exists for very advanced use cases.
138 https://github.com/lttng/lttng-ust/tree/stable-{revision}/doc/examples/clock-override[clock-override]
140 https://github.com/lttng/lttng-ust/tree/stable-{revision}/doc/examples/getcpu-override[getcpu-override]
141 examples for more details.
143 Moreover, LTTng{nbsp}{revision} boasts great stability, benifiting from
144 piles of bug fixes and more-than-welcome internal refactorings.
146 To learn more about the new features of LTTng{nbsp}{revision}, see
147 https://lttng.org/blog/2015/10/14/lttng-2.7-released/[the release announcement].
153 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
154 generation_ is a modern toolkit for tracing Linux systems and
155 applications. So your first question might be:
162 As the history of software engineering progressed and led to what
163 we now take for granted--complex, numerous and
164 interdependent software applications running in parallel on
165 sophisticated operating systems like Linux--the authors of such
166 components, software developers, began feeling a natural
167 urge to have tools that would ensure the robustness and good performance
168 of their masterpieces.
170 One major achievement in this field is, inarguably, the
171 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
172 an essential tool for developers to find and fix bugs. But even the best
173 debugger won't help make your software run faster, and nowadays, faster
174 software means either more work done by the same hardware, or cheaper
175 hardware for the same work.
177 A _profiler_ is often the tool of choice to identify performance
178 bottlenecks. Profiling is suitable to identify _where_ performance is
179 lost in a given software. The profiler outputs a profile, a statistical
180 summary of observed events, which you may use to discover which
181 functions took the most time to execute. However, a profiler won't
182 report _why_ some identified functions are the bottleneck. Bottlenecks
183 might only occur when specific conditions are met, conditions that are
184 sometimes impossible to capture by a statistical profiler, or impossible
185 to reproduce with an application altered by the overhead of an
186 event-based profiler. For a thorough investigation of software
187 performance issues, a history of execution is essential, with the
188 recorded values of variables and context fields you choose, and
189 with as little influence as possible on the instrumented software. This
190 is where tracing comes in handy.
192 _Tracing_ is a technique used to understand what goes on in a running
193 software system. The software used for tracing is called a _tracer_,
194 which is conceptually similar to a tape recorder. When recording,
195 specific instrumentation points placed in the software source code
196 generate events that are saved on a giant tape: a _trace_ file. You
197 can trace user applications and the operating system at the same time,
198 opening the possibility of resolving a wide range of problems that would
199 otherwise be extremely challenging.
201 Tracing is often compared to _logging_. However, tracers and loggers are
202 two different tools, serving two different purposes. Tracers are
203 designed to record much lower-level events that occur much more
204 frequently than log messages, often in the range of thousands per
205 second, with very little execution overhead. Logging is more appropriate
206 for a very high-level analysis of less frequent events: user accesses,
207 exceptional conditions (errors and warnings, for example), database
208 transactions, instant messaging communications, and such. Simply put,
209 logging is one of the many use cases that can be satisfied with tracing.
211 The list of recorded events inside a trace file can be read manually
212 like a log file for the maximum level of detail, but it is generally
213 much more interesting to perform application-specific analyses to
214 produce reduced statistics and graphs that are useful to resolve a
215 given problem. Trace viewers and analyzers are specialized tools
218 In the end, this is what LTTng is: a powerful, open source set of
219 tools to trace the Linux kernel and user applications at the same time.
220 LTTng is composed of several components actively maintained and
221 developed by its link:/community/#where[community].
224 [[lttng-alternatives]]
225 === Alternatives to noch:{LTTng}
227 Excluding proprietary solutions, a few competing software tracers
230 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
231 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
232 user scripts and is responsible for loading code into the
233 Linux kernel for further execution and collecting the outputted data.
234 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
235 subsystem in the Linux kernel in which a virtual machine can execute
236 programs passed from the user space to the kernel. You can attach
237 such programs to tracepoints and KProbes thanks to a system call, and
238 they can output data to the user space when executed thanks to
239 different mechanisms (pipe, VM register values, and eBPF maps, to name
241 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
242 is the de facto function tracer of the Linux kernel. Its user
243 interface is a set of special files in sysfs.
244 * https://perf.wiki.kernel.org/[perf] is
245 a performance analyzing tool for Linux which supports hardware
246 performance counters, tracepoints, as well as other counters and
247 types of probes. perf's controlling utility is the cmd:perf command
249 * http://linux.die.net/man/1/strace[strace]
250 is a command-line utility which records system calls made by a
251 user process, as well as signal deliveries and changes of process
252 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
253 to fulfill its function.
254 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
255 analyze Linux kernel events. You write scripts, or _chisels_ in
256 sysdig's jargon, in Lua and sysdig executes them while the system is
257 being traced or afterwards. sysdig's interface is the cmd:sysdig
258 command-line tool as well as the curses-based cmd:csysdig tool.
259 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
260 user space tracer which uses custom user scripts to produce plain text
261 traces. SystemTap converts the scripts to the C language, and then
262 compiles them as Linux kernel modules which are loaded to produce
263 trace data. SystemTap's primary user interface is the cmd:stap
266 The main distinctive features of LTTng is that it produces correlated
267 kernel and user space traces, as well as doing so with the lowest
268 overhead amongst other solutions. It produces trace files in the
269 http://diamon.org/ctf[CTF] format, a file format optimized
270 for the production and analyses of multi-gigabyte data.
272 LTTng is the result of more than 10 years of active open source
273 development by a community of passionate developers.
274 LTTng{nbsp}{revision} is currently available on major desktop and server
277 The main interface for tracing control is a single command-line tool
278 named cmd:lttng. The latter can create several tracing sessions, enable
279 and disable events on the fly, filter events efficiently with custom
280 user expressions, start and stop tracing, and much more. LTTng can
281 record the traces on the file system or send them over the network, and
282 keep them totally or partially. You can view the traces once tracing
283 becomes inactive or in real-time.
285 <<installing-lttng,Install LTTng now>> and
286 <<getting-started,start tracing>>!
292 include::../common/warning-installation-outdated.txt[]
294 **LTTng** is a set of software <<plumbing,components>> which interact to
295 <<instrumenting,instrument>> the Linux kernel and user applications, and
296 to <<controlling-tracing,control tracing>> (start and stop
297 tracing, enable and disable event rules, and the rest). Those
298 components are bundled into the following packages:
300 * **LTTng-tools**: Libraries and command-line interface to
302 * **LTTng-modules**: Linux kernel modules to instrument and
304 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
305 trace user applications.
307 Most distributions mark the LTTng-modules and LTTng-UST packages as
308 optional when installing LTTng-tools (which is always required). In the
309 following sections, we always provide the steps to install all three,
312 * You only need to install LTTng-modules if you intend to trace the
314 * You only need to install LTTng-UST if you intend to trace user
318 .Availability of LTTng{nbsp}{revision} for major Linux distributions.
320 |Distribution |Available in releases |Alternatives
323 |<<ubuntu,Ubuntu{nbsp}16.04 _Xenial Xerus_>>
324 |LTTng{nbsp}2.8 for Ubuntu{nbsp}16.10 _Yakkety Yak_.
326 LTTng{nbsp}{revision} for Ubuntu{nbsp}12.04 _Precise Pangolin_,
327 Ubuntu{nbsp}14.04 _Trusty Tahr_, and Ubuntu{nbsp}16.04 _Xenial Xerus_:
328 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
330 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
331 other Ubuntu releases.
335 |LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} for
336 Fedora{nbsp}25 and Fedora{nbsp}26 (both are not released yet).
338 <<building-from-source,Build LTTng-modules{nbsp}{revision} from
341 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
342 other Fedora releases.
346 |LTTng{nbsp}2.8 for Debian "stretch" (testing).
348 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
349 other Debian releases.
352 |<<opensuse,openSUSE Leap{nbsp}42.1>>
353 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
354 other openSUSE releases.
359 LTTng{nbsp}2.8 on the AUR.
361 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
365 |LTTng{nbsp}2.8 for Alpine Linux "edge".
367 LTTng{nbsp}2.8 for Alpine Linux{nbsp}3.5 (not released yet).
369 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
370 other Alpine Linux releases.
373 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
377 |<<"buildroot","Buildroot{nbsp}2016.02, Buildroot{nbsp}2016.05,
378 and Buildroot{nbsp}2016.08">>
379 |LTTng{nbsp}2.8 for Buildroot{nbsp}2016.11 (not released yet).
381 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
382 other Buildroot releases.
384 |OpenEmbedded and Yocto
385 |<<oe-yocto,Yocto Project{nbsp}2.1 _Krogoth_>> (`openembedded-core` layer)
386 |LTTng{nbsp}2.8 for Yocto Project{nbsp}2.2 _Morty_.
388 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
389 other Yocto releases.
394 === [[ubuntu-official-repositories]]Ubuntu
396 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.04 _Xenial Xerus_.
397 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
398 Stable{nbsp}{revision} PPA>>.
400 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}16.04 _Xenial Xerus_:
402 . Install the main LTTng{nbsp}{revision} packages:
407 sudo apt-get install lttng-tools
408 sudo apt-get install lttng-modules-dkms
409 sudo apt-get install liblttng-ust-dev
413 . **If you need to instrument and trace
414 <<java-application,Java applications>>**, install the LTTng-UST
420 sudo apt-get install liblttng-ust-agent-java
424 . **If you need to instrument and trace
425 <<python-application,Python{nbsp}3 applications>>**, install the
426 LTTng-UST Python agent:
431 sudo apt-get install python3-lttngust
437 ==== noch:{LTTng} Stable {revision} PPA
440 https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng Stable{nbsp}{revision} PPA]
441 offers the latest stable LTTng{nbsp}{revision} packages for:
443 * Ubuntu{nbsp}12.04 _Precise Pangolin_
444 * Ubuntu{nbsp}14.04 _Trusty Tahr_
445 * Ubuntu{nbsp}16.04 _Xenial Xerus_
447 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
450 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
456 sudo apt-add-repository ppa:lttng/stable-2.7
461 . Install the main LTTng{nbsp}{revision} packages:
466 sudo apt-get install lttng-tools
467 sudo apt-get install lttng-modules-dkms
468 sudo apt-get install liblttng-ust-dev
472 . **If you need to instrument and trace
473 <<java-application,Java applications>>**, install the LTTng-UST
479 sudo apt-get install liblttng-ust-agent-java
483 . **If you need to instrument and trace
484 <<python-application,Python{nbsp}3 applications>>**, install the
485 LTTng-UST Python agent:
490 sudo apt-get install python3-lttngust
496 === noch:{openSUSE}/RPM
498 To install LTTng{nbsp}{revision} on openSUSE Leap{nbsp}42.1:
500 * Install the main LTTng{nbsp}{revision} packages:
505 sudo zypper install lttng-tools
506 sudo zypper install lttng-modules
507 sudo zypper install lttng-ust-devel
512 .Java and Python application instrumentation and tracing
514 If you need to instrument and trace <<java-application,Java
515 applications>> on openSUSE, you need to build and install
516 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
517 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
518 `--enable-java-agent-all` options to the `configure` script, depending
519 on which Java logging framework you use.
521 If you need to instrument and trace <<python-application,Python
522 applications>> on openSUSE, you need to build and install
523 LTTng-UST{nbsp}{revision} from source and pass the
524 `--enable-python-agent` option to the `configure` script.
531 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2016.02,
532 Buildroot{nbsp}2016.05, or Buildroot{nbsp}2016.08:
534 . Launch the Buildroot configuration tool:
543 . In **Kernel**, check **Linux kernel**.
544 . In **Toolchain**, check **Enable WCHAR support**.
545 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
546 check **lttng-modules** and **lttng-tools**.
547 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
548 **Other**, check **lttng-libust**.
552 === OpenEmbedded and Yocto
554 LTTng{nbsp}{revision} recipes are available in the
555 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
556 layer for Yocto Project{nbsp}2.1 _Krogoth_ under the following names:
562 With BitBake, the simplest way to include LTTng recipes in your target
563 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
566 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
571 . Select a machine and an image recipe.
572 . Click **Edit image recipe**.
573 . Under the **All recipes** tab, search for **lttng**.
574 . Check the desired LTTng recipes.
577 .Java and Python application instrumentation and tracing
579 If you need to instrument and trace <<java-application,Java
580 applications>> on openSUSE, you need to build and install
581 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
582 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
583 `--enable-java-agent-all` options to the `configure` script, depending
584 on which Java logging framework you use.
586 If you need to instrument and trace <<python-application,Python
587 applications>> on openSUSE, you need to build and install
588 LTTng-UST{nbsp}{revision} from source and pass the
589 `--enable-python-agent` option to the `configure` script.
593 [[enterprise-distributions]]
594 === RHEL, SUSE, and other enterprise distributions
596 To install LTTng on enterprise Linux distributions, such as Red Hat
597 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
598 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
601 [[building-from-source]]
602 === Build from source
604 To build and install LTTng{nbsp}{revision} from source:
606 . Using your distribution's package manager, or from source, install
607 the following dependencies of LTTng-tools and LTTng-UST:
610 * https://sourceforge.net/projects/libuuid/[libuuid]
611 * http://directory.fsf.org/wiki/Popt[popt]
612 * http://liburcu.org/[Userspace RCU]
613 * http://www.xmlsoft.org/[libxml2]
616 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
622 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.7.tar.bz2 &&
623 tar -xf lttng-modules-latest-2.7.tar.bz2 &&
624 cd lttng-modules-2.7.* &&
626 sudo make modules_install &&
631 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
637 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.7.tar.bz2 &&
638 tar -xf lttng-ust-latest-2.7.tar.bz2 &&
639 cd lttng-ust-2.7.* &&
649 .Java and Python application tracing
651 If you need to instrument and trace <<java-application,Java
652 applications>>, pass the `--enable-java-agent-jul`,
653 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
654 `configure` script, depending on which Java logging framework you use.
656 If you need to instrument and trace <<python-application,Python
657 applications>>, pass the `--enable-python-agent` option to the
658 `configure` script. You can set the `PYTHON` environment variable to the
659 path to the Python interpreter for which to install the LTTng-UST Python
667 By default, LTTng-UST libraries are installed to
668 dir:{/usr/local/lib}, which is the de facto directory in which to
669 keep self-compiled and third-party libraries.
671 When <<building-tracepoint-providers-and-user-application,linking an
672 instrumented user application with `liblttng-ust`>>:
674 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
676 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
677 man:gcc(1), man:g++(1), or man:clang(1).
681 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
687 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
688 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
689 cd lttng-tools-2.7.* &&
697 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
698 previous steps automatically for a given version of LTTng and confine
699 the installed files in a specific directory. This can be useful to test
700 LTTng without installing it on your system.
706 This is a short guide to get started quickly with LTTng kernel and user
709 Before you follow this guide, make sure to <<installing-lttng,install>>
712 This tutorial walks you through the steps to:
714 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
715 . <<tracing-your-own-user-application,Trace a user application>> written
717 . <<viewing-and-analyzing-your-traces,View and analyze the
721 [[tracing-the-linux-kernel]]
722 === Trace the Linux kernel
724 The following command lines start with cmd:sudo because you need root
725 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
726 your Unix user is a member of the <<lttng-sessiond,tracing group>>.
728 . Create a <<tracing-session,tracing session>>:
733 sudo lttng create my-kernel-session
737 . List the available kernel tracepoints and system calls:
746 . Create an <<event,event rule>> which matches the desired event names,
747 for example `sched_switch` and `sched_process_fork`:
752 sudo lttng enable-event --kernel sched_switch,sched_process_fork
756 You can also create an event rule which _matches_ all the Linux kernel
757 tracepoints (this will generate a lot of data when tracing):
762 sudo lttng enable-event --kernel --all
775 . Do some operation on your system for a few seconds. For example,
776 load a website, or list the files of a directory.
777 . Stop tracing and destroy the tracing session:
787 The `destroy` command does not destroy the trace data; it only destroys
788 the state of the tracing session.
790 By default, LTTng saves the traces in
791 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
792 where +__name__+ is the tracing session name. Note that the
793 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
795 See <<viewing-and-analyzing-your-traces,View and analyze the
796 recorded events>> to view the recorded events.
799 [[tracing-your-own-user-application]]
800 === Trace a user application
802 This section steps you through a simple example to trace a
803 _Hello world_ program written in C.
805 To create the traceable user application:
807 . Create the tracepoint provider header file, which defines the
808 tracepoints and the events they can generate:
814 #undef TRACEPOINT_PROVIDER
815 #define TRACEPOINT_PROVIDER hello_world
817 #undef TRACEPOINT_INCLUDE
818 #define TRACEPOINT_INCLUDE "./hello-tp.h"
820 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
823 #include <lttng/tracepoint.h>
833 ctf_string(my_string_field, my_string_arg)
834 ctf_integer(int, my_integer_field, my_integer_arg)
838 #endif /* _HELLO_TP_H */
840 #include <lttng/tracepoint-event.h>
844 . Create the tracepoint provider package source file:
850 #define TRACEPOINT_CREATE_PROBES
851 #define TRACEPOINT_DEFINE
853 #include "hello-tp.h"
857 . Build the tracepoint provider package:
862 gcc -c -I. hello-tp.c
866 . Create the _Hello World_ application source file:
873 #include "hello-tp.h"
875 int main(int argc, char *argv[])
879 puts("Hello, World!\nPress Enter to continue...");
882 * The following getchar() call is only placed here for the purpose
883 * of this demonstration, to pause the application in order for
884 * you to have time to list its tracepoints. It is not
890 * A tracepoint() call.
892 * Arguments, as defined in hello-tp.h:
894 * 1. Tracepoint provider name (required)
895 * 2. Tracepoint name (required)
896 * 3. my_integer_arg (first user-defined argument)
897 * 4. my_string_arg (second user-defined argument)
899 * Notice the tracepoint provider and tracepoint names are
900 * NOT strings: they are in fact parts of variables that the
901 * macros in hello-tp.h create.
903 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
905 for (x = 0; x < argc; ++x) {
906 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
909 puts("Quitting now!");
910 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
917 . Build the application:
926 . Link the application with the tracepoint provider package,
927 `liblttng-ust`, and `libdl`:
932 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
936 Here's the whole build process:
939 .User space tracing tutorial's build steps.
940 image::ust-flow.png[]
942 To trace the user application:
944 . Run the application with a few arguments:
949 ./hello world and beyond
958 Press Enter to continue...
962 . Start an LTTng <<lttng-sessiond,session daemon>>:
967 lttng-sessiond --daemonize
971 Note that a session daemon might already be running, for example as
972 a service that the distribution's service manager started.
974 . List the available user space tracepoints:
979 lttng list --userspace
983 You see the `hello_world:my_first_tracepoint` tracepoint listed
984 under the `./hello` process.
986 . Create a <<tracing-session,tracing session>>:
991 lttng create my-user-space-session
995 . Create an <<event,event rule>> which matches the
996 `hello_world:my_first_tracepoint` event name:
1001 lttng enable-event --userspace hello_world:my_first_tracepoint
1014 . Go back to the running `hello` application and press Enter. The
1015 program executes all `tracepoint()` instrumentation points and exits.
1016 . Stop tracing and destroy the tracing session:
1026 The `destroy` command does not destroy the trace data; it only destroys
1027 the state of the tracing session.
1029 By default, LTTng saves the traces in
1030 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1031 where +__name__+ is the tracing session name. Note that the
1032 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1034 See <<viewing-and-analyzing-your-traces,View and analyze the
1035 recorded events>> to view the recorded events.
1038 [[viewing-and-analyzing-your-traces]]
1039 === View and analyze the recorded events
1041 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1042 kernel>> and <<tracing-your-own-user-application,Trace a user
1043 application>> tutorials, you can inspect the recorded events.
1045 Many tools are available to read LTTng traces:
1047 * **cmd:babeltrace** is a command-line utility which converts trace
1048 formats; it supports the format that LTTng produces, CTF, as well as a
1049 basic text output which can be ++grep++ed. The cmd:babeltrace command
1050 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1051 * Babeltrace also includes
1052 **https://www.python.org/[Python] bindings** so
1053 that you can easily open and read an LTTng trace with your own script,
1054 benefiting from the power of Python.
1055 * http://tracecompass.org/[**Trace Compass**]
1056 is a graphical user interface for viewing and analyzing any type of
1057 logs or traces, including LTTng's.
1058 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1059 project which includes many high-level analyses of LTTng kernel
1060 traces, like scheduling statistics, interrupt frequency distribution,
1061 top CPU usage, and more.
1063 NOTE: This section assumes that the traces recorded during the previous
1064 tutorials were saved to their default location, in the
1065 dir:{$LTTNG_HOME/lttng-traces} directory. Note that the env:LTTNG_HOME
1066 environment variable defaults to `$HOME` if not set.
1069 [[viewing-and-analyzing-your-traces-bt]]
1070 ==== Use the cmd:babeltrace command-line tool
1072 The simplest way to list all the recorded events of a trace is to pass
1073 its path to cmd:babeltrace with no options:
1077 babeltrace ~/lttng-traces/my-user-space-session*
1080 cmd:babeltrace finds all traces recursively within the given path and
1081 prints all their events, merging them in chronological order.
1083 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1088 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_
1091 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1092 count the recorded events:
1096 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_read | wc --lines
1100 [[viewing-and-analyzing-your-traces-bt-python]]
1101 ==== Use the Babeltrace Python bindings
1103 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1104 is useful to isolate events by simple matching using man:grep(1) and
1105 similar utilities. However, more elaborate filters, such as keeping only
1106 event records with a field value falling within a specific range, are
1107 not trivial to write using a shell. Moreover, reductions and even the
1108 most basic computations involving multiple event records are virtually
1109 impossible to implement.
1111 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1112 to read the event records of an LTTng trace sequentially and compute the
1113 desired information.
1115 The following script accepts an LTTng Linux kernel trace path as its
1116 first argument and prints the short names of the top 5 running processes
1117 on CPU 0 during the whole trace:
1122 from collections import Counter
1128 if len(sys.argv) != 2:
1129 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1130 print(msg, file=sys.stderr)
1133 # A trace collection contains one or more traces
1134 col = babeltrace.TraceCollection()
1136 # Add the trace provided by the user (LTTng traces always have
1138 if col.add_trace(sys.argv[1], 'ctf') is None:
1139 raise RuntimeError('Cannot add trace')
1141 # This counter dict contains execution times:
1143 # task command name -> total execution time (ns)
1144 exec_times = Counter()
1146 # This contains the last `sched_switch` timestamp
1150 for event in col.events:
1151 # Keep only `sched_switch` events
1152 if event.name != 'sched_switch':
1155 # Keep only events which happened on CPU 0
1156 if event['cpu_id'] != 0:
1160 cur_ts = event.timestamp
1166 # Previous task command (short) name
1167 prev_comm = event['prev_comm']
1169 # Initialize entry in our dict if not yet done
1170 if prev_comm not in exec_times:
1171 exec_times[prev_comm] = 0
1173 # Compute previous command execution time
1174 diff = cur_ts - last_ts
1176 # Update execution time of this command
1177 exec_times[prev_comm] += diff
1179 # Update last timestamp
1183 for name, ns in exec_times.most_common(5):
1185 print('{:20}{} s'.format(name, s))
1190 if __name__ == '__main__':
1191 sys.exit(0 if top5proc() else 1)
1198 python3 top5proc.py ~/lttng-traces/my-kernel-session*/kernel
1204 swapper/0 48.607245889 s
1205 chromium 7.192738188 s
1206 pavucontrol 0.709894415 s
1207 Compositor 0.660867933 s
1208 Xorg.bin 0.616753786 s
1211 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1212 weren't using the CPU that much when tracing, its first position in the
1217 == [[understanding-lttng]]Core concepts
1219 From a user's perspective, the LTTng system is built on a few concepts,
1220 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1221 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1222 Understanding how those objects relate to eachother is key in mastering
1225 The core concepts are:
1227 * <<tracing-session,Tracing session>>
1228 * <<domain,Tracing domain>>
1229 * <<channel,Channel and ring buffer>>
1230 * <<"event","Instrumentation point, event rule, event, and event record">>
1236 A _tracing session_ is a stateful dialogue between you and
1237 a <<lttng-sessiond,session daemon>>. You can
1238 <<creating-destroying-tracing-sessions,create a new tracing
1239 session>> with the `lttng create` command.
1241 Anything that you do when you control LTTng tracers happens within a
1242 tracing session. In particular, a tracing session:
1245 * Has its own set of trace files.
1246 * Has its own state of activity (started or stopped).
1247 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1249 * Has its own <<channel,channels>> which have their own
1250 <<event,event rules>>.
1253 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1254 image::concepts.png[]
1256 Those attributes and objects are completely isolated between different
1259 A tracing session is analogous to a cash machine session:
1260 the operations you do on the banking system through the cash machine do
1261 not alter the data of other users of the same system. In the case of
1262 the cash machine, a session lasts as long as your bank card is inside.
1263 In the case of LTTng, a tracing session lasts from the `lttng create`
1264 command to the `lttng destroy` command.
1267 .Each Unix user has its own set of tracing sessions.
1268 image::many-sessions.png[]
1271 [[tracing-session-mode]]
1272 ==== Tracing session mode
1274 LTTng can send the generated trace data to different locations. The
1275 _tracing session mode_ dictates where to send it. The following modes
1276 are available in LTTng{nbsp}{revision}:
1279 LTTng writes the traces to the file system of the machine being traced
1282 Network streaming mode::
1283 LTTng sends the traces over the network to a
1284 <<lttng-relayd,relay daemon>> running on a remote system.
1287 LTTng does not write the traces by default. Instead, you can request
1288 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1289 current tracing buffers, and to write it to the target's file system
1290 or to send it over the network to a <<lttng-relayd,relay daemon>>
1291 running on a remote system.
1294 This mode is similar to the network streaming mode, but a live
1295 trace viewer can connect to the distant relay daemon to
1296 <<lttng-live,view event records as LTTng generates them>> by
1303 A _tracing domain_ is a namespace for event sources. A tracing domain
1304 has its own properties and features.
1306 There are currently five available tracing domains:
1310 * `java.util.logging` (JUL)
1314 You must specify a tracing domain when using some commands to avoid
1315 ambiguity. For example, since all the domains support named tracepoints
1316 as event sources (instrumentation points that you manually insert in the
1317 source code), you need to specify a tracing domain when
1318 <<enabling-disabling-events,creating an event rule>> because all the
1319 tracing domains could have tracepoints with the same names.
1321 Some features are reserved to specific tracing domains. Dynamic function
1322 entry and return instrumentation points, for example, are currently only
1323 supported in the Linux kernel tracing domain, but support for other
1324 tracing domains could be added in the future.
1326 You can create <<channel,channels>> in the Linux kernel and user space
1327 tracing domains. The other tracing domains have a single default
1332 === Channel and ring buffer
1334 A _channel_ is an object which is responsible for a set of ring buffers.
1335 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1336 tracer emits an event, it can record it to one or more
1337 sub-buffers. The attributes of a channel determine what to do when
1338 there's no space left for a new event record because all sub-buffers
1339 are full, where to send a full sub-buffer, and other behaviours.
1341 A channel is always associated to a <<domain,tracing domain>>. The
1342 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1343 a default channel which you cannot configure.
1345 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1346 an event, it records it to the sub-buffers of all
1347 the enabled channels with a satisfied event rule, as long as those
1348 channels are part of active <<tracing-session,tracing sessions>>.
1351 [[channel-buffering-schemes]]
1352 ==== Per-user vs. per-process buffering schemes
1354 A channel has at least one ring buffer _per CPU_. LTTng always
1355 records an event to the ring buffer associated to the CPU on which it
1358 Two _buffering schemes_ are available when you
1359 <<enabling-disabling-channels,create a channel>> in the
1360 user space <<domain,tracing domain>>:
1362 Per-user buffering::
1363 Allocate one set of ring buffers--one per CPU--shared by all the
1364 instrumented processes of each Unix user.
1368 .Per-user buffering scheme.
1369 image::per-user-buffering.png[]
1372 Per-process buffering::
1373 Allocate one set of ring buffers--one per CPU--for each
1374 instrumented process.
1378 .Per-process buffering scheme.
1379 image::per-process-buffering.png[]
1382 The per-process buffering scheme tends to consume more memory than the
1383 per-user option because systems generally have more instrumented
1384 processes than Unix users running instrumented processes. However, the
1385 per-process buffering scheme ensures that one process having a high
1386 event throughput won't fill all the shared sub-buffers of the same
1389 The Linux kernel tracing domain has only one available buffering scheme
1390 which is to allocate a single set of ring buffers for the whole system.
1391 This scheme is similar to the per-user option, but with a single, global
1392 user "running" the kernel.
1395 [[channel-overwrite-mode-vs-discard-mode]]
1396 ==== Overwrite vs. discard event loss modes
1398 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1399 arc in the following animation) of a specific channel's ring buffer.
1400 When there's no space left in a sub-buffer, the tracer marks it as
1401 consumable (red) and another, empty sub-buffer starts receiving the
1402 following event records. A <<lttng-consumerd,consumer daemon>>
1403 eventually consumes the marked sub-buffer (returns to white).
1406 [role="docsvg-channel-subbuf-anim"]
1411 In an ideal world, sub-buffers are consumed faster than they are filled,
1412 as is the case in the previous animation. In the real world,
1413 however, all sub-buffers can be full at some point, leaving no space to
1414 record the following events.
1416 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1417 available, it is acceptable to lose event records when the alternative
1418 would be to cause substantial delays in the instrumented application's
1419 execution. LTTng privileges performance over integrity; it aims at
1420 perturbing the traced system as little as possible in order to make
1421 tracing of subtle race conditions and rare interrupt cascades possible.
1423 When it comes to losing event records because no empty sub-buffer is
1424 available, the channel's _event loss mode_ determines what to do. The
1425 available event loss modes are:
1428 Drop the newest event records until a the tracer
1429 releases a sub-buffer.
1432 Clear the sub-buffer containing the oldest event records and start
1433 writing the newest event records there.
1435 This mode is sometimes called _flight recorder mode_ because it's
1437 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1438 always keep a fixed amount of the latest data.
1440 Which mechanism you should choose depends on your context: prioritize
1441 the newest or the oldest event records in the ring buffer?
1443 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1444 as soon as a there's no space left for a new event record, whereas in
1445 discard mode, the tracer only discards the event record that doesn't
1448 In discard mode, LTTng increments a count of lost event records when
1449 an event record is lost and saves this count to the trace. In
1450 overwrite mode, LTTng keeps no information when it overwrites a
1451 sub-buffer before consuming it.
1453 There are a few ways to decrease your probability of losing event
1455 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1456 how you can fine-une the sub-buffer count and size of a channel to
1457 virtually stop losing event records, though at the cost of greater
1461 [[channel-subbuf-size-vs-subbuf-count]]
1462 ==== Sub-buffer count and size
1464 When you <<enabling-disabling-channels,create a channel>>, you can
1465 set its number of sub-buffers and their size.
1467 Note that there is noticeable CPU overhead introduced when
1468 switching sub-buffers (marking a full one as consumable and switching
1469 to an empty one for the following events to be recorded). Knowing this,
1470 the following list presents a few practical situations along with how
1471 to configure the sub-buffer count and size for them:
1473 * **High event throughput**: In general, prefer bigger sub-buffers to
1474 lower the risk of losing event records.
1476 Having bigger sub-buffers also ensures a lower sub-buffer switching
1479 The number of sub-buffers is only meaningful if you create the channel
1480 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1481 other sub-buffers are left unaltered.
1483 * **Low event throughput**: In general, prefer smaller sub-buffers
1484 since the risk of losing event records is low.
1486 Because events occur less frequently, the sub-buffer switching frequency
1487 should remain low and thus the tracer's overhead should not be a
1490 * **Low memory system**: If your target system has a low memory
1491 limit, prefer fewer first, then smaller sub-buffers.
1493 Even if the system is limited in memory, you want to keep the
1494 sub-buffers as big as possible to avoid a high sub-buffer switching
1497 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1498 which means event data is very compact. For example, the average
1499 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1500 sub-buffer size of 1{nbsp}MiB is considered big.
1502 The previous situations highlight the major trade-off between a few big
1503 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1504 frequency vs. how much data is lost in overwrite mode. Assuming a
1505 constant event throughput and using the overwrite mode, the two
1506 following configurations have the same ring buffer total size:
1509 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1514 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1515 switching frequency, but if a sub-buffer overwrite happens, half of
1516 the event records so far (4{nbsp}MiB) are definitely lost.
1517 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1518 overhead as the previous configuration, but if a sub-buffer
1519 overwrite happens, only the eighth of event records so far are
1522 In discard mode, the sub-buffers count parameter is pointless: use two
1523 sub-buffers and set their size according to the requirements of your
1527 [[channel-switch-timer]]
1528 ==== Switch timer period
1530 The _switch timer period_ is an important configurable attribute of
1531 a channel to ensure periodic sub-buffer flushing.
1533 When the _switch timer_ expires, a sub-buffer switch happens. You can
1534 set the switch timer period attribute when you
1535 <<enabling-disabling-channels,create a channel>> to ensure that event
1536 data is consumed and committed to trace files or to a distant relay
1537 daemon periodically in case of a low event throughput.
1540 [role="docsvg-channel-switch-timer"]
1545 This attribute is also convenient when you use big sub-buffers to cope
1546 with a sporadic high event throughput, even if the throughput is
1550 [[channel-read-timer]]
1551 ==== Read timer period
1553 By default, the LTTng tracers use a notification mechanism to signal a
1554 full sub-buffer so that a consumer daemon can consume it. When such
1555 notifications must be avoided, for example in real-time applications,
1556 you can use the channel's _read timer_ instead. When the read timer
1557 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1558 consumable sub-buffers.
1561 [[tracefile-rotation]]
1562 ==== Trace file count and size
1564 By default, trace files can grow as large as needed. You can set the
1565 maximum size of each trace file that a channel writes when you
1566 <<enabling-disabling-channels,create a channel>>. When the size of
1567 a trace file reaches the channel's fixed maximum size, LTTng creates
1568 another file to contain the next event records. LTTng appends a file
1569 count to each trace file name in this case.
1571 If you set the trace file size attribute when you create a channel, the
1572 maximum number of trace files that LTTng creates is _unlimited_ by
1573 default. To limit them, you can also set a maximum number of trace
1574 files. When the number of trace files reaches the channel's fixed
1575 maximum count, the oldest trace file is overwritten. This mechanism is
1576 called _trace file rotation_.
1580 === Instrumentation point, event rule, event, and event record
1582 An _event rule_ is a set of conditions which must be **all** satisfied
1583 for LTTng to record an occuring event.
1585 You set the conditions when you <<enabling-disabling-events,create
1588 You always attach an event rule to <<channel,channel>> when you create
1591 When an event passes the conditions of an event rule, LTTng records it
1592 in one of the attached channel's sub-buffers.
1594 The available conditions, as of LTTng{nbsp}{revision}, are:
1596 * The event rule _is enabled_.
1597 * The instrumentation point's type _is{nbsp}T_.
1598 * The instrumentation point's name (sometimes called _event name_)
1599 _matches{nbsp}N_, but _is not{nbsp}E_.
1600 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1601 _is exactly{nbsp}L_.
1602 * The fields of the event's payload _satisfy_ a filter
1603 expression{nbsp}__F__.
1605 As you can see, all the conditions but the dynamic filter are related to
1606 the event rule's status or to the instrumentation point, not to the
1607 occurring events. This is why, without a filter, checking if an event
1608 passes an event rule is not a dynamic task: when you create or modify an
1609 event rule, all the tracers of its tracing domain enable or disable the
1610 instrumentation points themselves once. This is possible because the
1611 attributes of an instrumentation point (type, name, and log level) are
1612 defined statically. In other words, without a dynamic filter, the tracer
1613 _does not evaluate_ the arguments of an instrumentation point unless it
1614 matches an enabled event rule.
1616 Note that, for LTTng to record an event, the <<channel,channel>> to
1617 which a matching event rule is attached must also be enabled, and the
1618 tracing session owning this channel must be active.
1621 .Logical path from an instrumentation point to an event record.
1622 image::event-rule.png[]
1624 .Event, event record, or event rule?
1626 With so many similar terms, it's easy to get confused.
1628 An **event** is the consequence of the execution of an _instrumentation
1629 point_, like a tracepoint that you manually place in some source code,
1630 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1631 time. Different actions can be taken upon the occurrence of an event,
1632 like record the event's payload to a buffer.
1634 An **event record** is the representation of an event in a sub-buffer. A
1635 tracer is responsible for capturing the payload of an event, current
1636 context variables, the event's ID, and the event's timestamp. LTTng
1637 can append this sub-buffer to a trace file.
1639 An **event rule** is a set of conditions which must all be satisfied for
1640 LTTng to record an occuring event. Events still occur without
1641 satisfying event rules, but LTTng does not record them.
1646 == Components of noch:{LTTng}
1648 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1649 to call LTTng a simple _tool_ since it is composed of multiple
1650 interacting components. This section describes those components,
1651 explains their respective roles, and shows how they connect together to
1652 form the LTTng ecosystem.
1654 The following diagram shows how the most important components of LTTng
1655 interact with user applications, the Linux kernel, and you:
1658 .Control and trace data paths between LTTng components.
1659 image::plumbing.png[]
1661 The LTTng project incorporates:
1663 * **LTTng-tools**: Libraries and command-line interface to
1664 control tracing sessions.
1665 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1666 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1667 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1668 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1669 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1670 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1672 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1673 headers to instrument and trace any native user application.
1674 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1675 *** `liblttng-ust-libc-wrapper`
1676 *** `liblttng-ust-pthread-wrapper`
1677 *** `liblttng-ust-cyg-profile`
1678 *** `liblttng-ust-cyg-profile-fast`
1679 *** `liblttng-ust-dl`
1680 ** User space tracepoint provider source files generator command-line
1681 tool (man:lttng-gen-tp(1)).
1682 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1683 Java applications using `java.util.logging` or
1684 Apache log4j 1.2 logging.
1685 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1686 Python applications using the standard `logging` package.
1687 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1689 ** LTTng kernel tracer module.
1690 ** Tracing ring buffer kernel modules.
1691 ** Probe kernel modules.
1692 ** LTTng logger kernel module.
1696 === Tracing control command-line interface
1699 .The tracing control command-line interface.
1700 image::plumbing-lttng-cli.png[]
1702 The _man:lttng(1) command-line tool_ is the standard user interface to
1703 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1704 is part of LTTng-tools.
1706 The cmd:lttng tool is linked with
1707 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1708 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1710 The cmd:lttng tool has a Git-like interface:
1714 lttng <general options> <command> <command options>
1717 The <<controlling-tracing,Tracing control>> section explores the
1718 available features of LTTng using the cmd:lttng tool.
1721 [[liblttng-ctl-lttng]]
1722 === Tracing control library
1725 .The tracing control library.
1726 image::plumbing-liblttng-ctl.png[]
1728 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1729 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1730 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1732 The <<lttng-cli,cmd:lttng command-line tool>>
1733 is linked with `liblttng-ctl`.
1735 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1740 #include <lttng/lttng.h>
1743 Some objects are referenced by name (C string), such as tracing
1744 sessions, but most of them require to create a handle first using
1745 `lttng_create_handle()`.
1747 The best available developer documentation for `liblttng-ctl` is, as of
1748 LTTng{nbsp}{revision}, its installed header files. Every function and
1749 structure is thoroughly documented.
1753 === User space tracing library
1756 .The user space tracing library.
1757 image::plumbing-liblttng-ust.png[]
1759 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1760 is the LTTng user space tracer. It receives commands from a
1761 <<lttng-sessiond,session daemon>>, for example to
1762 enable and disable specific instrumentation points, and writes event
1763 records to ring buffers shared with a
1764 <<lttng-consumerd,consumer daemon>>.
1765 `liblttng-ust` is part of LTTng-UST.
1767 Public C header files are installed beside `liblttng-ust` to
1768 instrument any <<c-application,C or $$C++$$ application>>.
1770 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1771 packages, use their own library providing tracepoints which is
1772 linked with `liblttng-ust`.
1774 An application or library does not have to initialize `liblttng-ust`
1775 manually: its constructor does the necessary tasks to properly register
1776 to a session daemon. The initialization phase also enables the
1777 instrumentation points matching the <<event,event rules>> that you
1781 [[lttng-ust-agents]]
1782 === User space tracing agents
1785 .The user space tracing agents.
1786 image::plumbing-lttng-ust-agents.png[]
1788 The _LTTng-UST Java and Python agents_ are regular Java and Python
1789 packages which add LTTng tracing capabilities to the
1790 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1792 In the case of Java, the
1793 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1794 core logging facilities] and
1795 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1796 Note that Apache Log4{nbsp}2 is not supported.
1798 In the case of Python, the standard
1799 https://docs.python.org/3/library/logging.html[`logging`] package
1800 is supported. Both Python 2 and Python 3 modules can import the
1801 LTTng-UST Python agent package.
1803 The applications using the LTTng-UST agents are in the
1804 `java.util.logging` (JUL),
1805 log4j, and Python <<domain,tracing domains>>.
1807 Both agents use the same mechanism to trace the log statements. When an
1808 agent is initialized, it creates a log handler that attaches to the root
1809 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1810 When the application executes a log statement, it is passed to the
1811 agent's log handler by the root logger. The agent's log handler calls a
1812 native function in a tracepoint provider package shared library linked
1813 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1814 other fields, like its logger name and its log level. This native
1815 function contains a user space instrumentation point, hence tracing the
1818 The log level condition of an
1819 <<event,event rule>> is considered when tracing
1820 a Java or a Python application, and it's compatible with the standard
1821 JUL, log4j, and Python log levels.
1825 === LTTng kernel modules
1828 .The LTTng kernel modules.
1829 image::plumbing-lttng-modules.png[]
1831 The _LTTng kernel modules_ are a set of Linux kernel modules
1832 which implement the kernel tracer of the LTTng project. The LTTng
1833 kernel modules are part of LTTng-modules.
1835 The LTTng kernel modules include:
1837 * A set of _probe_ modules.
1839 Each module attaches to a specific subsystem
1840 of the Linux kernel using its tracepoint instrument points. There are
1841 also modules to attach to the entry and return points of the Linux
1842 system call functions.
1844 * _Ring buffer_ modules.
1846 A ring buffer implementation is provided as kernel modules. The LTTng
1847 kernel tracer writes to the ring buffer; a
1848 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1850 * The _LTTng kernel tracer_ module.
1851 * The _LTTng logger_ module.
1853 The LTTng logger module implements the special path:{/proc/lttng-logger}
1854 file so that any executable can generate LTTng events by opening and
1855 writing to this file.
1857 See <<proc-lttng-logger-abi,LTTng logger>>.
1859 Generally, you do not have to load the LTTng kernel modules manually
1860 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1861 daemon>> loads the necessary modules when starting. If you have extra
1862 probe modules, you can specify to load them to the session daemon on
1865 The LTTng kernel modules are installed in
1866 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1867 the kernel release (see `uname --kernel-release`).
1874 .The session daemon.
1875 image::plumbing-sessiond.png[]
1877 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1878 managing tracing sessions and for controlling the various components of
1879 LTTng. The session daemon is part of LTTng-tools.
1881 The session daemon sends control requests to and receives control
1884 * The <<lttng-ust,user space tracing library>>.
1886 Any instance of the user space tracing library first registers to
1887 a session daemon. Then, the session daemon can send requests to
1888 this instance, such as:
1891 ** Get the list of tracepoints.
1892 ** Share an <<event,event rule>> so that the user space tracing library
1893 can enable or disable tracepoints. Amongst the possible conditions
1894 of an event rule is a filter expression which `liblttng-ust` evalutes
1895 when an event occurs.
1896 ** Share <<channel,channel>> attributes and ring buffer locations.
1899 The session daemon and the user space tracing library use a Unix
1900 domain socket for their communication.
1902 * The <<lttng-ust-agents,user space tracing agents>>.
1904 Any instance of a user space tracing agent first registers to
1905 a session daemon. Then, the session daemon can send requests to
1906 this instance, such as:
1909 ** Get the list of loggers.
1910 ** Enable or disable a specific logger.
1913 The session daemon and the user space tracing agent use a TCP connection
1914 for their communication.
1916 * The <<lttng-modules,LTTng kernel tracer>>.
1917 * The <<lttng-consumerd,consumer daemon>>.
1919 The session daemon sends requests to the consumer daemon to instruct
1920 it where to send the trace data streams, amongst other information.
1922 * The <<lttng-relayd,relay daemon>>.
1924 The session daemon receives commands from the
1925 <<liblttng-ctl-lttng,tracing control library>>.
1927 The root session daemon loads the appropriate
1928 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1929 a <<lttng-consumerd,consumer daemon>> as soon as you create
1930 an <<event,event rule>>.
1932 The session daemon does not send and receive trace data: this is the
1933 role of the <<lttng-consumerd,consumer daemon>> and
1934 <<lttng-relayd,relay daemon>>. It does, however, generate the
1935 http://diamon.org/ctf/[CTF] metadata stream.
1937 Each Unix user can have its own session daemon instance. The
1938 tracing sessions managed by different session daemons are completely
1941 The root user's session daemon is the only one which is
1942 allowed to control the LTTng kernel tracer, and its spawned consumer
1943 daemon is the only one which is allowed to consume trace data from the
1944 LTTng kernel tracer. Note, however, that any Unix user which is a member
1945 of the <<tracing-group,tracing group>> is allowed
1946 to create <<channel,channels>> in the
1947 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1950 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1951 session daemon when using its `create` command if none is currently
1952 running. You can also start the session daemon manually.
1959 .The consumer daemon.
1960 image::plumbing-consumerd.png[]
1962 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
1963 ring buffers with user applications or with the LTTng kernel modules to
1964 collect trace data and send it to some location (on disk or to a
1965 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1966 is part of LTTng-tools.
1968 You do not start a consumer daemon manually: a consumer daemon is always
1969 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1970 <<event,event rule>>, that is, before you start tracing. When you kill
1971 its owner session daemon, the consumer daemon also exits because it is
1972 the session daemon's child process. Command-line options of
1973 man:lttng-sessiond(8) target the consumer daemon process.
1975 There are up to two running consumer daemons per Unix user, whereas only
1976 one session daemon can run per user. This is because each process can be
1977 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1978 and 64-bit processes, it is more efficient to have separate
1979 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1980 exception: it can have up to _three_ running consumer daemons: 32-bit
1981 and 64-bit instances for its user applications, and one more
1982 reserved for collecting kernel trace data.
1990 image::plumbing-relayd.png[]
1992 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1993 between remote session and consumer daemons, local trace files, and a
1994 remote live trace viewer. The relay daemon is part of LTTng-tools.
1996 The main purpose of the relay daemon is to implement a receiver of
1997 <<sending-trace-data-over-the-network,trace data over the network>>.
1998 This is useful when the target system does not have much file system
1999 space to record trace files locally.
2001 The relay daemon is also a server to which a
2002 <<lttng-live,live trace viewer>> can
2003 connect. The live trace viewer sends requests to the relay daemon to
2004 receive trace data as the target system emits events. The
2005 communication protocol is named _LTTng live_; it is used over TCP
2008 Note that you can start the relay daemon on the target system directly.
2009 This is the setup of choice when the use case is to view events as
2010 the target system emits them without the need of a remote system.
2014 == [[using-lttng]]Instrumentation
2016 There are many examples of tracing and monitoring in our everyday life:
2018 * You have access to real-time and historical weather reports and
2019 forecasts thanks to weather stations installed around the country.
2020 * You know your heart is safe thanks to an electrocardiogram.
2021 * You make sure not to drive your car too fast and to have enough fuel
2022 to reach your destination thanks to gauges visible on your dashboard.
2024 All the previous examples have something in common: they rely on
2025 **instruments**. Without the electrodes attached to the surface of your
2026 body's skin, cardiac monitoring is futile.
2028 LTTng, as a tracer, is no different from those real life examples. If
2029 you're about to trace a software system or, in other words, record its
2030 history of execution, you better have **instrumentation points** in the
2031 subject you're tracing, that is, the actual software.
2033 Various ways were developed to instrument a piece of software for LTTng
2034 tracing. The most straightforward one is to manually place
2035 instrumentation points, called _tracepoints_, in the software's source
2036 code. It is also possible to add instrumentation points dynamically in
2037 the Linux kernel <<domain,tracing domain>>.
2039 If you're only interested in tracing the Linux kernel, your
2040 instrumentation needs are probably already covered by LTTng's built-in
2041 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2042 user application which is already instrumented for LTTng tracing.
2043 In such cases, you can skip this whole section and read the topics of
2044 the <<controlling-tracing,Tracing control>> section.
2046 Many methods are available to instrument a piece of software for LTTng
2049 * <<c-application,User space instrumentation for C and $$C++$$
2051 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2052 * <<java-application,User space Java agent>>.
2053 * <<python-application,User space Python agent>>.
2054 * <<proc-lttng-logger-abi,LTTng logger>>.
2055 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2059 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2061 The procedure to instrument a C or $$C++$$ user application with
2062 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2064 . <<tracepoint-provider,Create the source files of a tracepoint provider
2066 . <<probing-the-application-source-code,Add tracepoints to
2067 the application's source code>>.
2068 . <<building-tracepoint-providers-and-user-application,Build and link
2069 a tracepoint provider package and the user application>>.
2071 If you need quick, man:printf(3)-like instrumentation, you can skip
2072 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2075 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2076 instrument a user application with `liblttng-ust`.
2079 [[tracepoint-provider]]
2080 ==== Create the source files of a tracepoint provider package
2082 A _tracepoint provider_ is a set of compiled functions which provide
2083 **tracepoints** to an application, the type of instrumentation point
2084 supported by LTTng-UST. Those functions can emit events with
2085 user-defined fields and serialize those events as event records to one
2086 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2087 macro, which you <<probing-the-application-source-code,insert in a user
2088 application's source code>>, calls those functions.
2090 A _tracepoint provider package_ is an object file (`.o`) or a shared
2091 library (`.so`) which contains one or more tracepoint providers.
2092 Its source files are:
2094 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2095 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2097 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2098 the LTTng user space tracer, at run time.
2101 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2102 image::ust-app.png[]
2104 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2105 skip creating and using a tracepoint provider and use
2106 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2110 ===== Create a tracepoint provider header file template
2112 A _tracepoint provider header file_ contains the tracepoint
2113 definitions of a tracepoint provider.
2115 To create a tracepoint provider header file:
2117 . Start from this template:
2121 .Tracepoint provider header file template (`.h` file extension).
2123 #undef TRACEPOINT_PROVIDER
2124 #define TRACEPOINT_PROVIDER provider_name
2126 #undef TRACEPOINT_INCLUDE
2127 #define TRACEPOINT_INCLUDE "./tp.h"
2129 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2132 #include <lttng/tracepoint.h>
2135 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2136 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2141 #include <lttng/tracepoint-event.h>
2147 * `provider_name` with the name of your tracepoint provider.
2148 * `"tp.h"` with the name of your tracepoint provider header file.
2150 . Below the `#include <lttng/tracepoint.h>` line, put your
2151 <<defining-tracepoints,tracepoint definitions>>.
2153 Your tracepoint provider name must be unique amongst all the possible
2154 tracepoint provider names used on the same target system. We
2155 suggest to include the name of your project or company in the name,
2156 for example, `org_lttng_my_project_tpp`.
2158 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2159 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2160 write are the <<defining-tracepoints,tracepoint definitions>>.
2163 [[defining-tracepoints]]
2164 ===== Create a tracepoint definition
2166 A _tracepoint definition_ defines, for a given tracepoint:
2168 * Its **input arguments**. They are the macro parameters that the
2169 `tracepoint()` macro accepts for this particular tracepoint
2170 in the user application's source code.
2171 * Its **output event fields**. They are the sources of event fields
2172 that form the payload of any event that the execution of the
2173 `tracepoint()` macro emits for this particular tracepoint.
2175 You can create a tracepoint definition by using the
2176 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2178 <<tpp-header,tracepoint provider header file template>>.
2180 The syntax of the `TRACEPOINT_EVENT()` macro is:
2183 .`TRACEPOINT_EVENT()` macro syntax.
2186 /* Tracepoint provider name */
2189 /* Tracepoint name */
2192 /* Input arguments */
2197 /* Output event fields */
2206 * `provider_name` with your tracepoint provider name.
2207 * `tracepoint_name` with your tracepoint name.
2208 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2209 * `fields` with the <<tpp-def-output-fields,output event field>>
2212 This tracepoint emits events named `provider_name:tracepoint_name`.
2215 .Event name's length limitation
2217 The concatenation of the tracepoint provider name and the
2218 tracepoint name must not exceed **254 characters**. If it does, the
2219 instrumented application compiles and runs, but LTTng throws multiple
2220 warnings and you could experience serious issues.
2223 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2226 .`TP_ARGS()` macro syntax.
2235 * `type` with the C type of the argument.
2236 * `arg_name` with the argument name.
2238 You can repeat `type` and `arg_name` up to 10 times to have
2239 more than one argument.
2241 .`TP_ARGS()` usage with three arguments.
2253 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2254 tracepoint definition with no input arguments.
2256 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2257 `ctf_*()` macros. Each `ctf_*()` macro defines one event field.
2258 See <<liblttng-ust-tp-fields,Tracepoint fields macros>> for a
2259 complete description of the available `ctf_*()` macros.
2260 A `ctf_*()` macro specifies the type, size, and byte order of
2263 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2264 C expression that the tracer evalutes at the `tracepoint()` macro site
2265 in the application's source code. This expression provides a field's
2266 source of data. The argument expression can include input argument names
2267 listed in the `TP_ARGS()` macro.
2269 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2270 must be unique within a given tracepoint definition.
2272 Here's a complete tracepoint definition example:
2274 .Tracepoint definition.
2276 The following tracepoint definition defines a tracepoint which takes
2277 three input arguments and has four output event fields.
2281 #include "my-custom-structure.h"
2287 const struct my_custom_structure*, my_custom_structure,
2292 ctf_string(query_field, query)
2293 ctf_float(double, ratio_field, ratio)
2294 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2295 ctf_integer(int, send_size, my_custom_structure->send_size)
2300 You can refer to this tracepoint definition with the `tracepoint()`
2301 macro in your application's source code like this:
2305 tracepoint(my_provider, my_tracepoint,
2306 my_structure, some_ratio, the_query);
2310 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2311 if they satisfy an enabled <<event,event rule>>.
2314 [[using-tracepoint-classes]]
2315 ===== Use a tracepoint class
2317 A _tracepoint class_ is a class of tracepoints which share the same
2318 output event field definitions. A _tracepoint instance_ is one
2319 instance of such a defined tracepoint class, with its own tracepoint
2322 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2323 shorthand which defines both a tracepoint class and a tracepoint
2324 instance at the same time.
2326 When you build a tracepoint provider package, the C or $$C++$$ compiler
2327 creates one serialization function for each **tracepoint class**. A
2328 serialization function is responsible for serializing the event fields
2329 of a tracepoint to a sub-buffer when tracing.
2331 For various performance reasons, when your situation requires multiple
2332 tracepoint definitions with different names, but with the same event
2333 fields, we recommend that you manually create a tracepoint class
2334 and instantiate as many tracepoint instances as needed. One positive
2335 effect of such a design, amongst other advantages, is that all
2336 tracepoint instances of the same tracepoint class reuse the same
2337 serialization function, thus reducing
2338 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2340 .Use a tracepoint class and tracepoint instances.
2342 Consider the following three tracepoint definitions:
2354 ctf_integer(int, userid, userid)
2355 ctf_integer(size_t, len, len)
2367 ctf_integer(int, userid, userid)
2368 ctf_integer(size_t, len, len)
2380 ctf_integer(int, userid, userid)
2381 ctf_integer(size_t, len, len)
2386 In this case, we create three tracepoint classes, with one implicit
2387 tracepoint instance for each of them: `get_account`, `get_settings`, and
2388 `get_transaction`. However, they all share the same event field names
2389 and types. Hence three identical, yet independent serialization
2390 functions are created when you build the tracepoint provider package.
2392 A better design choice is to define a single tracepoint class and three
2393 tracepoint instances:
2397 /* The tracepoint class */
2398 TRACEPOINT_EVENT_CLASS(
2399 /* Tracepoint provider name */
2402 /* Tracepoint class name */
2405 /* Input arguments */
2411 /* Output event fields */
2413 ctf_integer(int, userid, userid)
2414 ctf_integer(size_t, len, len)
2418 /* The tracepoint instances */
2419 TRACEPOINT_EVENT_INSTANCE(
2420 /* Tracepoint provider name */
2423 /* Tracepoint class name */
2426 /* Tracepoint name */
2429 /* Input arguments */
2435 TRACEPOINT_EVENT_INSTANCE(
2444 TRACEPOINT_EVENT_INSTANCE(
2457 [[assigning-log-levels]]
2458 ===== Assign a log level to a tracepoint definition
2460 You can assign an optional _log level_ to a
2461 <<defining-tracepoints,tracepoint definition>>.
2463 Assigning different levels of severity to tracepoint definitions can
2464 be useful: when you <<enabling-disabling-events,create an event rule>>,
2465 you can target tracepoints having a log level as severe as a specific
2468 The concept of LTTng-UST log levels is similar to the levels found
2469 in typical logging frameworks:
2471 * In a logging framework, the log level is given by the function
2472 or method name you use at the log statement site: `debug()`,
2473 `info()`, `warn()`, `error()`, and so on.
2474 * In LTTng-UST, you statically assign the log level to a tracepoint
2475 definition; any `tracepoint()` macro invocation which refers to
2476 this definition has this log level.
2478 You can assign a log level to a tracepoint definition with the
2479 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2480 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2481 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2484 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2487 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2489 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2494 * `provider_name` with the tracepoint provider name.
2495 * `tracepoint_name` with the tracepoint name.
2496 * `log_level` with the log level to assign to the tracepoint
2497 definition named `tracepoint_name` in the `provider_name`
2498 tracepoint provider.
2500 See <<liblttng-ust-tracepoint-loglevel,Tracepoint log levels>> for
2501 a list of available log level names.
2503 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2507 /* Tracepoint definition */
2516 ctf_integer(int, userid, userid)
2517 ctf_integer(size_t, len, len)
2521 /* Log level assignment */
2522 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2528 ===== Create a tracepoint provider package source file
2530 A _tracepoint provider package source file_ is a C source file which
2531 includes a <<tpp-header,tracepoint provider header file>> to expand its
2532 macros into event serialization and other functions.
2534 You can always use the following tracepoint provider package source
2538 .Tracepoint provider package source file template.
2540 #define TRACEPOINT_CREATE_PROBES
2545 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2546 header file>> name. You may also include more than one tracepoint
2547 provider header file here to create a tracepoint provider package
2548 holding more than one tracepoint providers.
2551 [[probing-the-application-source-code]]
2552 ==== Add tracepoints to an application's source code
2554 Once you <<tpp-header,create a tracepoint provider header file>>, you
2555 can use the `tracepoint()` macro in your application's
2556 source code to insert the tracepoints that this header
2557 <<defining-tracepoints,defines>>.
2559 The `tracepoint()` macro takes at least two parameters: the tracepoint
2560 provider name and the tracepoint name. The corresponding tracepoint
2561 definition defines the other parameters.
2563 .`tracepoint()` usage.
2565 The following <<defining-tracepoints,tracepoint definition>> defines a
2566 tracepoint which takes two input arguments and has two output event
2570 .Tracepoint provider header file.
2572 #include "my-custom-structure.h"
2579 const char*, cmd_name
2582 ctf_string(cmd_name, cmd_name)
2583 ctf_integer(int, number_of_args, argc)
2588 You can refer to this tracepoint definition with the `tracepoint()`
2589 macro in your application's source code like this:
2592 .Application's source file.
2596 int main(int argc, char* argv[])
2598 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2604 Note how the application's source code includes
2605 the tracepoint provider header file containing the tracepoint
2606 definitions to use, path:{tp.h}.
2609 .`tracepoint()` usage with a complex tracepoint definition.
2611 Consider this complex tracepoint definition, where multiple event
2612 fields refer to the same input arguments in their argument expression
2616 .Tracepoint provider header file.
2618 /* For `struct stat` */
2619 #include <sys/types.h>
2620 #include <sys/stat.h>
2632 ctf_integer(int, my_constant_field, 23 + 17)
2633 ctf_integer(int, my_int_arg_field, my_int_arg)
2634 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2635 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2636 my_str_arg[2] + my_str_arg[3])
2637 ctf_string(my_str_arg_field, my_str_arg)
2638 ctf_integer_hex(off_t, size_field, st->st_size)
2639 ctf_float(double, size_dbl_field, (double) st->st_size)
2640 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2641 size_t, strlen(my_str_arg) / 2)
2646 You can refer to this tracepoint definition with the `tracepoint()`
2647 macro in your application's source code like this:
2650 .Application's source file.
2652 #define TRACEPOINT_DEFINE
2659 stat("/etc/fstab", &s);
2660 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2666 If you look at the event record that LTTng writes when tracing this
2667 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2668 it should look like this:
2670 .Event record fields
2672 |Field's name |Field's value
2673 |`my_constant_field` |40
2674 |`my_int_arg_field` |23
2675 |`my_int_arg_field2` |529
2677 |`my_str_arg_field` |`Hello, World!`
2678 |`size_field` |0x12d
2679 |`size_dbl_field` |301.0
2680 |`half_my_str_arg_field` |`Hello,`
2684 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2685 compute--they use the call stack, for example. To avoid this
2686 computation when the tracepoint is disabled, you can use the
2687 `tracepoint_enabled()` and `do_tracepoint()` macros.
2689 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2693 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2695 tracepoint_enabled(provider_name, tracepoint_name)
2696 do_tracepoint(provider_name, tracepoint_name, ...)
2701 * `provider_name` with the tracepoint provider name.
2702 * `tracepoint_name` with the tracepoint name.
2704 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2705 `tracepoint_name` from the provider named `provider_name` is enabled
2708 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2709 if the tracepoint is enabled. Using `tracepoint()` with
2710 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2711 the `tracepoint_enabled()` check, thus a race condition is
2712 possible in this situation:
2715 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2717 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2718 stuff = prepare_stuff();
2721 tracepoint(my_provider, my_tracepoint, stuff);
2724 If the tracepoint is enabled after the condition, then `stuff` is not
2725 prepared: the emitted event will either contain wrong data, or the whole
2726 application could crash (segmentation fault, for example).
2728 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2729 `STAP_PROBEV()` call. If you need it, you must emit
2733 [[building-tracepoint-providers-and-user-application]]
2734 ==== Build and link a tracepoint provider package and an application
2736 Once you have one or more <<tpp-header,tracepoint provider header
2737 files>> and a <<tpp-source,tracepoint provider package source file>>,
2738 you can create the tracepoint provider package by compiling its source
2739 file. From here, multiple build and run scenarios are possible. The
2740 following table shows common application and library configurations
2741 along with the required command lines to achieve them.
2743 In the following diagrams, we use the following file names:
2746 Executable application.
2749 Application's object file.
2752 Tracepoint provider package object file.
2755 Tracepoint provider package archive file.
2758 Tracepoint provider package shared object file.
2761 User library object file.
2764 User library shared object file.
2766 We use the following symbols in the diagrams of table below:
2769 .Symbols used in the build scenario diagrams.
2770 image::ust-sit-symbols.png[]
2772 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2773 variable in the following instructions.
2775 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2776 .Common tracepoint provider package scenarios.
2778 |Scenario |Instructions
2781 The instrumented application is statically linked with
2782 the tracepoint provider package object.
2784 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2787 include::../common/ust-sit-step-tp-o.txt[]
2789 To build the instrumented application:
2791 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2796 #define TRACEPOINT_DEFINE
2800 . Compile the application source file:
2809 . Build the application:
2814 gcc -o app app.o tpp.o -llttng-ust -ldl
2818 To run the instrumented application:
2820 * Start the application:
2830 The instrumented application is statically linked with the
2831 tracepoint provider package archive file.
2833 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2836 To create the tracepoint provider package archive file:
2838 . Compile the <<tpp-source,tracepoint provider package source file>>:
2847 . Create the tracepoint provider package archive file:
2856 To build the instrumented application:
2858 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2863 #define TRACEPOINT_DEFINE
2867 . Compile the application source file:
2876 . Build the application:
2881 gcc -o app app.o tpp.a -llttng-ust -ldl
2885 To run the instrumented application:
2887 * Start the application:
2897 The instrumented application is linked with the tracepoint provider
2898 package shared object.
2900 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2903 include::../common/ust-sit-step-tp-so.txt[]
2905 To build the instrumented application:
2907 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2912 #define TRACEPOINT_DEFINE
2916 . Compile the application source file:
2925 . Build the application:
2930 gcc -o app app.o -ldl -L. -ltpp
2934 To run the instrumented application:
2936 * Start the application:
2946 The tracepoint provider package shared object is preloaded before the
2947 instrumented application starts.
2949 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2952 include::../common/ust-sit-step-tp-so.txt[]
2954 To build the instrumented application:
2956 . In path:{app.c}, before including path:{tpp.h}, add the
2962 #define TRACEPOINT_DEFINE
2963 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2967 . Compile the application source file:
2976 . Build the application:
2981 gcc -o app app.o -ldl
2985 To run the instrumented application with tracing support:
2987 * Preload the tracepoint provider package shared object and
2988 start the application:
2993 LD_PRELOAD=./libtpp.so ./app
2997 To run the instrumented application without tracing support:
2999 * Start the application:
3009 The instrumented application dynamically loads the tracepoint provider
3010 package shared object.
3012 See the <<dlclose-warning,warning about `dlclose()`>>.
3014 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3017 include::../common/ust-sit-step-tp-so.txt[]
3019 To build the instrumented application:
3021 . In path:{app.c}, before including path:{tpp.h}, add the
3027 #define TRACEPOINT_DEFINE
3028 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3032 . Compile the application source file:
3041 . Build the application:
3046 gcc -o app app.o -ldl
3050 To run the instrumented application:
3052 * Start the application:
3062 The application is linked with the instrumented user library.
3064 The instrumented user library is statically linked with the tracepoint
3065 provider package object file.
3067 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3070 include::../common/ust-sit-step-tp-o-fpic.txt[]
3072 To build the instrumented user library:
3074 . In path:{emon.c}, before including path:{tpp.h}, add the
3080 #define TRACEPOINT_DEFINE
3084 . Compile the user library source file:
3089 gcc -I. -fpic -c emon.c
3093 . Build the user library shared object:
3098 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3102 To build the application:
3104 . Compile the application source file:
3113 . Build the application:
3118 gcc -o app app.o -L. -lemon
3122 To run the application:
3124 * Start the application:
3134 The application is linked with the instrumented user library.
3136 The instrumented user library is linked with the tracepoint provider
3137 package shared object.
3139 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3142 include::../common/ust-sit-step-tp-so.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 -ldl -L. -ltpp
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 tracepoint provider package shared object is preloaded before the
3209 The application is linked with the instrumented user library.
3211 image::ust-sit+tp-so-preloaded+app-linked-with-lib+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
3225 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3229 . Compile the user library source file:
3234 gcc -I. -fpic -c emon.c
3238 . Build the user library shared object:
3243 gcc -shared -o libemon.so emon.o -ldl
3247 To build the application:
3249 . Compile the application source file:
3258 . Build the application:
3263 gcc -o app app.o -L. -lemon
3267 To run the application with tracing support:
3269 * Preload the tracepoint provider package shared object and
3270 start the application:
3275 LD_PRELOAD=./libtpp.so ./app
3279 To run the application without tracing support:
3281 * Start the application:
3291 The application is linked with the instrumented user library.
3293 The instrumented user library dynamically loads the tracepoint provider
3294 package shared object.
3296 See the <<dlclose-warning,warning about `dlclose()`>>.
3298 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3301 include::../common/ust-sit-step-tp-so.txt[]
3303 To build the instrumented user library:
3305 . In path:{emon.c}, before including path:{tpp.h}, add the
3311 #define TRACEPOINT_DEFINE
3312 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3316 . Compile the user library source file:
3321 gcc -I. -fpic -c emon.c
3325 . Build the user library shared object:
3330 gcc -shared -o libemon.so emon.o -ldl
3334 To build the application:
3336 . Compile the application source file:
3345 . Build the application:
3350 gcc -o app app.o -L. -lemon
3354 To run the application:
3356 * Start the application:
3366 The application dynamically loads the instrumented user library.
3368 The instrumented user library is linked with the tracepoint provider
3369 package shared object.
3371 See the <<dlclose-warning,warning about `dlclose()`>>.
3373 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3376 include::../common/ust-sit-step-tp-so.txt[]
3378 To build the instrumented user library:
3380 . In path:{emon.c}, before including path:{tpp.h}, add the
3386 #define TRACEPOINT_DEFINE
3390 . Compile the user library source file:
3395 gcc -I. -fpic -c emon.c
3399 . Build the user library shared object:
3404 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3408 To build the application:
3410 . Compile the application source file:
3419 . Build the application:
3424 gcc -o app app.o -ldl -L. -lemon
3428 To run the application:
3430 * Start the application:
3440 The application dynamically loads the instrumented user library.
3442 The instrumented user library dynamically loads the tracepoint provider
3443 package shared object.
3445 See the <<dlclose-warning,warning about `dlclose()`>>.
3447 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3450 include::../common/ust-sit-step-tp-so.txt[]
3452 To build the instrumented user library:
3454 . In path:{emon.c}, before including path:{tpp.h}, add the
3460 #define TRACEPOINT_DEFINE
3461 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3465 . Compile the user library source file:
3470 gcc -I. -fpic -c emon.c
3474 . Build the user library shared object:
3479 gcc -shared -o libemon.so emon.o -ldl
3483 To build the application:
3485 . Compile the application source file:
3494 . Build the application:
3499 gcc -o app app.o -ldl -L. -lemon
3503 To run the application:
3505 * Start the application:
3515 The tracepoint provider package shared object is preloaded before the
3518 The application dynamically loads the instrumented user library.
3520 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3523 include::../common/ust-sit-step-tp-so.txt[]
3525 To build the instrumented user library:
3527 . In path:{emon.c}, before including path:{tpp.h}, add the
3533 #define TRACEPOINT_DEFINE
3534 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3538 . Compile the user library source file:
3543 gcc -I. -fpic -c emon.c
3547 . Build the user library shared object:
3552 gcc -shared -o libemon.so emon.o -ldl
3556 To build the application:
3558 . Compile the application source file:
3567 . Build the application:
3572 gcc -o app app.o -L. -lemon
3576 To run the application with tracing support:
3578 * Preload the tracepoint provider package shared object and
3579 start the application:
3584 LD_PRELOAD=./libtpp.so ./app
3588 To run the application without tracing support:
3590 * Start the application:
3600 The application is statically linked with the tracepoint provider
3601 package object file.
3603 The application is linked with the instrumented user library.
3605 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3608 include::../common/ust-sit-step-tp-o.txt[]
3610 To build the instrumented user library:
3612 . In path:{emon.c}, before including path:{tpp.h}, add the
3618 #define TRACEPOINT_DEFINE
3622 . Compile the user library source file:
3627 gcc -I. -fpic -c emon.c
3631 . Build the user library shared object:
3636 gcc -shared -o libemon.so emon.o
3640 To build the application:
3642 . Compile the application source file:
3651 . Build the application:
3656 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3660 To run the instrumented application:
3662 * Start the application:
3672 The application is statically linked with the tracepoint provider
3673 package object file.
3675 The application dynamically loads the instrumented user library.
3677 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3680 include::../common/ust-sit-step-tp-o.txt[]
3682 To build the application:
3684 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3689 #define TRACEPOINT_DEFINE
3693 . Compile the application source file:
3702 . Build the application:
3707 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3712 The `--export-dynamic` option passed to the linker is necessary for the
3713 dynamically loaded library to ``see'' the tracepoint symbols defined in
3716 To build the instrumented user library:
3718 . Compile the user library source file:
3723 gcc -I. -fpic -c emon.c
3727 . Build the user library shared object:
3732 gcc -shared -o libemon.so emon.o
3736 To run the application:
3738 * Start the application:
3750 .Do not use man:dlclose(3) on a tracepoint provider package
3752 Never use man:dlclose(3) on any shared object which:
3754 * Is linked with, statically or dynamically, a tracepoint provider
3756 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3757 package shared object.
3759 This is currently considered **unsafe** due to a lack of reference
3760 counting from LTTng-UST to the shared object.
3762 A known workaround (available since glibc 2.2) is to use the
3763 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3764 effect of not unloading the loaded shared object, even if man:dlclose(3)
3767 You can also preload the tracepoint provider package shared object with
3768 the env:LD_PRELOAD environment variable to overcome this limitation.
3772 [[using-lttng-ust-with-daemons]]
3773 ===== Use noch:{LTTng-UST} with daemons
3775 If your instrumented application calls man:fork(2), man:clone(2),
3776 or BSD's man:rfork(2), without a following man:exec(3)-family
3777 system call, you must preload the path:{liblttng-ust-fork.so} shared
3778 object when starting the application.
3782 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3785 If your tracepoint provider package is
3786 a shared library which you also preload, you must put both
3787 shared objects in env:LD_PRELOAD:
3791 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3795 [[lttng-ust-pkg-config]]
3796 ===== Use noch:{pkg-config}
3798 On some distributions, LTTng-UST ships with a
3799 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3800 metadata file. If this is your case, then you can use cmd:pkg-config to
3801 build an application on the command line:
3805 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3809 [[instrumenting-32-bit-app-on-64-bit-system]]
3810 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3812 In order to trace a 32-bit application running on a 64-bit system,
3813 LTTng must use a dedicated 32-bit
3814 <<lttng-consumerd,consumer daemon>>.
3816 The following steps show how to build and install a 32-bit consumer
3817 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3818 build and install the 32-bit LTTng-UST libraries, and how to build and
3819 link an instrumented 32-bit application in that context.
3821 To build a 32-bit instrumented application for a 64-bit target system,
3822 assuming you have a fresh target system with no installed Userspace RCU
3825 . Download, build, and install a 32-bit version of Userspace RCU:
3831 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3832 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3833 cd userspace-rcu-0.9.* &&
3834 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3836 sudo make install &&
3841 . Using your distribution's package manager, or from source, install
3842 the following 32-bit versions of the following dependencies of
3843 LTTng-tools and LTTng-UST:
3846 * https://sourceforge.net/projects/libuuid/[libuuid]
3847 * http://directory.fsf.org/wiki/Popt[popt]
3848 * http://www.xmlsoft.org/[libxml2]
3851 . Download, build, and install a 32-bit version of the latest
3852 LTTng-UST{nbsp}{revision}:
3858 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.7.tar.bz2 &&
3859 tar -xf lttng-ust-latest-2.7.tar.bz2 &&
3860 cd lttng-ust-2.7.* &&
3861 ./configure --libdir=/usr/local/lib32 \
3862 CFLAGS=-m32 CXXFLAGS=-m32 \
3863 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3865 sudo make install &&
3872 Depending on your distribution,
3873 32-bit libraries could be installed at a different location than
3874 `/usr/lib32`. For example, Debian is known to install
3875 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3877 In this case, make sure to set `LDFLAGS` to all the
3878 relevant 32-bit library paths, for example:
3882 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3886 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3887 the 32-bit consumer daemon:
3893 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
3894 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
3895 cd lttng-tools-2.7.* &&
3896 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3897 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3899 cd src/bin/lttng-consumerd &&
3900 sudo make install &&
3905 . From your distribution or from source,
3906 <<installing-lttng,install>> the 64-bit versions of
3907 LTTng-UST and Userspace RCU.
3908 . Download, build, and install the 64-bit version of the
3909 latest LTTng-tools{nbsp}{revision}:
3915 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
3916 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
3917 cd lttng-tools-2.7.* &&
3918 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3919 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3921 sudo make install &&
3926 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3927 when linking your 32-bit application:
3930 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3931 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3934 For example, let's rebuild the quick start example in
3935 <<tracing-your-own-user-application,Trace a user application>> as an
3936 instrumented 32-bit application:
3941 gcc -m32 -c -I. hello-tp.c
3943 gcc -m32 -o hello hello.o hello-tp.o \
3944 -L/usr/lib32 -L/usr/local/lib32 \
3945 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3950 No special action is required to execute the 32-bit application and
3951 to trace it: use the command-line man:lttng(1) tool as usual.
3958 `tracef()` is a small LTTng-UST API designed for quick,
3959 man:printf(3)-like instrumentation without the burden of
3960 <<tracepoint-provider,creating>> and
3961 <<building-tracepoint-providers-and-user-application,building>>
3962 a tracepoint provider package.
3964 To use `tracef()` in your application:
3966 . In the C or C++ source files where you need to use `tracef()`,
3967 include `<lttng/tracef.h>`:
3972 #include <lttng/tracef.h>
3976 . In the application's source code, use `tracef()` like you would use
3984 tracef("my message: %d (%s)", my_integer, my_string);
3990 . Link your application with `liblttng-ust`:
3995 gcc -o app app.c -llttng-ust
3999 To trace the events that `tracef()` calls emit:
4001 * <<enabling-disabling-events,Create an event rule>> which matches the
4002 `lttng_ust_tracef:*` event name:
4007 lttng enable-event --userspace 'lttng_ust_tracef:*'
4012 .Limitations of `tracef()`
4014 The `tracef()` utility function was developed to make user space tracing
4015 super simple, albeit with notable disadvantages compared to
4016 <<defining-tracepoints,user-defined tracepoints>>:
4018 * All the emitted events have the same tracepoint provider and
4019 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4020 * There is no static type checking.
4021 * The only event record field you actually get, named `msg`, is a string
4022 potentially containing the values you passed to `tracef()`
4023 using your own format string. This also means that you cannot filter
4024 events with a custom expression at run time because there are no
4026 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4027 function behind the scenes to format the strings at run time, its
4028 expected performance is lower than with user-defined tracepoints,
4029 which do not require a conversion to a string.
4031 Taking this into consideration, `tracef()` is useful for some quick
4032 prototyping and debugging, but you should not consider it for any
4033 permanent and serious applicative instrumentation.
4039 ==== Use `tracelog()`
4041 The `tracelog()` API is very similar to <<tracef,`tracef()`>>, with
4042 the difference that it accepts an additional log level parameter.
4044 The goal of `tracelog()` is to ease the migration from logging to
4047 To use `tracelog()` in your application:
4049 . In the C or C++ source files where you need to use `tracelog()`,
4050 include `<lttng/tracelog.h>`:
4055 #include <lttng/tracelog.h>
4059 . In the application's source code, use `tracelog()` like you would use
4060 man:printf(3), except for the first parameter which is the log
4068 tracelog(TRACE_WARNING, "my message: %d (%s)",
4069 my_integer, my_string);
4075 See <<liblttng-ust-tracepoint-loglevel,Tracepoint log levels>> for
4076 a list of available log level names.
4078 . Link your application with `liblttng-ust`:
4083 gcc -o app app.c -llttng-ust
4087 To trace the events that `tracelog()` calls emit with a log level
4088 _as severe as_ a specific log level:
4090 * <<enabling-disabling-events,Create an event rule>> which matches the
4091 `lttng_ust_tracelog:*` event name and a minimum level
4097 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4098 --loglevel=TRACE_WARNING
4102 To trace the events that `tracelog()` calls emit with a
4103 _specific log level_:
4105 * Create an event rule which matches the `lttng_ust_tracelog:*`
4106 event name and a specific log level:
4111 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4112 --loglevel-only=TRACE_INFO
4117 [[prebuilt-ust-helpers]]
4118 === Prebuilt user space tracing helpers
4120 The LTTng-UST package provides a few helpers in the form or preloadable
4121 shared objects which automatically instrument system functions and
4124 The helper shared objects are normally found in dir:{/usr/lib}. If you
4125 built LTTng-UST <<building-from-source,from source>>, they are probably
4126 located in dir:{/usr/local/lib}.
4128 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4131 path:{liblttng-ust-libc-wrapper.so}::
4132 path:{liblttng-ust-pthread-wrapper.so}::
4133 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4134 memory and POSIX threads function tracing>>.
4136 path:{liblttng-ust-cyg-profile.so}::
4137 path:{liblttng-ust-cyg-profile-fast.so}::
4138 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4140 path:{liblttng-ust-dl.so}::
4141 <<liblttng-ust-dl,Dynamic linker tracing>>.
4143 To use a user space tracing helper with any user application:
4145 * Preload the helper shared object when you start the application:
4150 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4154 You can preload more than one helper:
4159 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4165 [[liblttng-ust-libc-pthread-wrapper]]
4166 ==== Instrument C standard library memory and POSIX threads functions
4168 The path:{liblttng-ust-libc-wrapper.so} and
4169 path:{liblttng-ust-pthread-wrapper.so} helpers
4170 add instrumentation to some C standard library and POSIX
4174 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4176 |TP provider name |TP name |Instrumented function
4178 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4179 |`calloc` |man:calloc(3)
4180 |`realloc` |man:realloc(3)
4181 |`free` |man:free(3)
4182 |`memalign` |man:memalign(3)
4183 |`posix_memalign` |man:posix_memalign(3)
4187 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4189 |TP provider name |TP name |Instrumented function
4191 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4192 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4193 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4194 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4197 When you preload the shared object, it replaces the functions listed
4198 in the previous tables by wrappers which contain tracepoints and call
4199 the replaced functions.
4202 [[liblttng-ust-cyg-profile]]
4203 ==== Instrument function entry and exit
4205 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4206 to the entry and exit points of functions.
4208 man:gcc(1) and man:clang(1) have an option named
4209 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4210 which generates instrumentation calls for entry and exit to functions.
4211 The LTTng-UST function tracing helpers,
4212 path:{liblttng-ust-cyg-profile.so} and
4213 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4214 to add tracepoints to the two generated functions (which contain
4215 `cyg_profile` in their names, hence the helper's name).
4217 To use the LTTng-UST function tracing helper, the source files to
4218 instrument must be built using the `-finstrument-functions` compiler
4221 There are two versions of the LTTng-UST function tracing helper:
4223 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4224 that you should only use when it can be _guaranteed_ that the
4225 complete event stream is recorded without any lost event record.
4226 Any kind of duplicate information is left out.
4228 This version contains the following tracepoints:
4232 .Points instrumented by preloading path:{liblttng-ust-cyg-profile-fast.so}.
4234 |TP provider name |TP name |Instrumented points
4236 .2+|`lttng_ust_cyg_profile_fast`
4242 Address of called function.
4249 Assuming no event record is lost, having only the function addresses on
4250 entry is enough to create a call graph, since an event record always
4251 contains the ID of the CPU that generated it.
4253 You can use a tool like man:addr2line(1) to convert function addresses
4254 back to source file names and line numbers.
4256 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4257 which also works in use cases where event records might get discarded or
4258 not recorded from application startup.
4259 In these cases, the trace analyzer needs more information to be
4260 able to reconstruct the program flow.
4262 This version contains the following tracepoints:
4266 .Points instrumented by preloading path:{liblttng-ust-cyg-profile.so}.
4268 |TP provider name |TP name |Instrumented point
4270 .2+|`lttng_ust_cyg_profile`
4276 Address of called function.
4285 Address of called function.
4292 TIP: It's sometimes a good idea to limit the number of source files that
4293 you compile with the `-finstrument-functions` option to prevent LTTng
4294 from writing an excessive amount of trace data at run time. When using
4295 man:gcc(1), you can use the
4296 `-finstrument-functions-exclude-function-list` option to avoid
4297 instrument entries and exits of specific function names.
4299 All the tracepoints that this helper contains have the
4300 <<liblttng-ust-tracepoint-loglevel,log level>> `TRACE_DEBUG_FUNCTION`.
4305 ==== Instrument the dynamic linker
4307 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4308 man:dlopen(3) and man:dlclose(3) function calls.
4311 .Functions instrumented by preloading path:{liblttng-ust-dl.so}.
4313 |TP provider name |TP name |Instrumented function
4321 Memory base address (where the dynamic linker placed the shared
4325 File system path to the loaded shared object.
4328 File size of the the loaded shared object.
4331 Last modification time (seconds since Epoch time) of the loaded shared
4338 Memory base address (where the dynamic linker placed the shared
4344 [[java-application]]
4345 === User space Java agent
4347 You can instrument a Java application which uses one of the following
4350 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4351 (JUL) core logging facilities.
4352 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4353 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4355 Each log statement emits an LTTng event once the
4356 application initializes the <<lttng-ust-agents,LTTng-UST Java agent>>
4360 .LTTng-UST Java agent imported by a Java application.
4361 image::java-app.png[]
4363 NOTE: We use http://openjdk.java.net/[OpenJDK] 7 for development and
4364 https://ci.lttng.org/[continuous integration], thus this version is
4365 directly supported. However, the LTTng-UST Java agent is also
4366 tested with OpenJDK 6.
4368 To use the LTTng-UST Java agent:
4370 . In the Java application's source code, import the LTTng-UST Java
4376 import org.lttng.ust.agent.LTTngAgent;
4380 . As soon as possible after the entry point of the application,
4381 initialize the LTTng-UST Java agent:
4386 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4390 Any log statement that the application executes before this
4391 initialization does not emit an LTTng event.
4393 . Use `java.util.logging` and/or log4j log statements and configuration
4394 as usual. Since the LTTng-UST Java agent adds a handler to the _root_
4395 loggers, you can trace any log statement from any logger.
4397 . Before exiting the application, dispose the LTTng-UST Java agent:
4402 lttngAgent.dispose();
4406 This is not strictly necessary, but it is recommended for a clean
4407 disposal of the agent's resources.
4409 Any log statement that the application executes after this disposal does
4410 not emit an LTTng event.
4412 . Include the LTTng-UST Java agent's JAR file, path:{liblttng-ust-agent.jar},
4414 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class path]
4415 when building the Java application.
4417 path:{liblttng-ust-agent.jar} is typically located in
4418 dir:{/usr/share/java}.
4420 IMPORTANT: The LTTng-UST Java agent must be
4421 <<installing-lttng,installed>> for the logging framework your
4424 .[[jul]]Use the LTTng-UST Java agent with `java.util.logging`.
4429 import java.util.logging.Logger;
4430 import org.lttng.ust.agent.LTTngAgent;
4434 private static final int answer = 42;
4436 public static void main(String[] argv) throws Exception
4439 Logger logger = Logger.getLogger("jello");
4441 // Call this as soon as possible (before logging)
4442 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4445 logger.info("some info");
4446 logger.warning("some warning");
4448 logger.finer("finer information; the answer is " + answer);
4450 logger.severe("error!");
4452 // Not mandatory, but cleaner
4453 lttngAgent.dispose();
4458 You can build this example like this:
4462 javac -cp /usr/share/java/liblttng-ust-agent.jar Test.java
4465 You can run the compiled class like this:
4469 java -cp /usr/share/java/liblttng-ust-agent.jar:. Test
4473 .[[log4j]]Use the LTTng-UST Java agent with Apache log4j 1.2.
4478 import org.apache.log4j.Logger;
4479 import org.apache.log4j.BasicConfigurator;
4480 import org.lttng.ust.agent.LTTngAgent;
4484 private static final int answer = 42;
4486 public static void main(String[] argv) throws Exception
4488 // Create and configure a logger
4489 Logger logger = Logger.getLogger(Test.class);
4490 BasicConfigurator.configure();
4492 // Call this as soon as possible (before logging)
4493 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4496 logger.info("some info");
4497 logger.warn("some warning");
4499 logger.debug("debug information; the answer is " + answer);
4501 logger.error("error!");
4502 logger.fatal("fatal error!");
4504 // Not mandatory, but cleaner
4505 lttngAgent.dispose();
4510 You can build this example like this:
4514 javac -cp /usr/share/java/liblttng-ust-agent.jar:$LOG4JCP Test.java
4517 where `$LOG4JCP` is the path to log4j's JAR file.
4519 You can run the compiled class like this:
4523 java -cp /usr/share/java/liblttng-ust-agent.jar:$LOG4JCP:. Test
4527 When you <<enabling-disabling-events,create an event rule>>, use the
4528 `--jul` (`java.util.logging`) or `--log4j` (log4j) option to target
4530 <<domain,tracing domain>>. You can also use the `--loglevel` or
4531 `--loglevel-only` option to target a range of JUL/log4j log levels or a
4532 specific JUL/log4j log level.
4536 [[python-application]]
4537 === User space Python agent
4539 You can instrument a Python 2 or Python 3 application which uses the
4540 standard https://docs.python.org/3/library/logging.html[`logging`]
4543 Each log statement emits an LTTng event once the
4544 application module imports the
4545 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4548 .A Python application importing the LTTng-UST Python agent.
4549 image::python-app.png[]
4551 To use the LTTng-UST Python agent:
4553 . In the Python application's source code, import the LTTng-UST Python
4563 The LTTng-UST Python agent automatically adds its logging handler to the
4564 root logger at import time.
4566 Any log statement that the application executes before this import does
4567 not emit an LTTng event.
4569 IMPORTANT: The LTTng-UST Python agent must be
4570 <<installing-lttng,installed>>.
4572 . Use log statements and logging configuration as usual.
4573 Since the LTTng-UST Python agent adds a handler to the _root_
4574 logger, you can trace any log statement from any logger.
4576 .Use the LTTng-UST Python agent.
4586 logging.basicConfig()
4587 logger = logging.getLogger('my-logger')
4590 logger.debug('debug message')
4591 logger.info('info message')
4592 logger.warn('warn message')
4593 logger.error('error message')
4594 logger.critical('critical message')
4598 if __name__ == '__main__':
4602 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4603 logging handler which prints to the standard error stream, is not
4604 strictly required for LTTng-UST tracing to work, but in versions of
4605 Python preceding 3.2, you could see a warning message which indicates
4606 that no handler exists for the logger `my-logger`.
4609 When you <<enabling-disabling-events,create an event rule>>, use the
4610 `--python` option to target the Python
4611 <<domain,tracing domain>>. You can also use
4612 the `--loglevel` or `--loglevel-only` option to target a range of
4613 Python log levels or a specific Python log level.
4615 When an application imports the LTTng-UST Python agent, the agent tries
4616 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4617 start the session daemon _before_ you start the Python application.
4618 If a session daemon is found, the agent tries to register to it
4619 during 5{nbsp}seconds, after which the application continues without
4620 LTTng tracing support. You can override this timeout value with the
4621 env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4624 If the session daemon stops while a Python application with an imported
4625 LTTng-UST Python agent runs, the agent retries to connect and to
4626 register to a session daemon every 3{nbsp}seconds. You can override this
4627 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4632 [[proc-lttng-logger-abi]]
4635 The `lttng-tracer` Linux kernel module, part of
4636 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4637 path:{/proc/lttng-logger} when it's loaded. Any application can write
4638 text data to this file to emit an LTTng event.
4641 .An application writes to the LTTng logger file to emit an LTTng event.
4642 image::lttng-logger.png[]
4644 The LTTng logger is the quickest method--not the most efficient,
4645 however--to add instrumentation to an application. It is designed
4646 mostly to instrument shell scripts:
4650 echo "Some message, some $variable" > /proc/lttng-logger
4653 Any event that the LTTng logger emits is named `lttng_logger` and
4654 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4655 other instrumentation points in the kernel tracing domain, **any Unix
4656 user** can <<enabling-disabling-events,create an event rule>> which
4657 matches its event name, not only the root user or users in the tracing
4660 To use the LTTng logger:
4662 * From any application, write text data to the path:{/proc/lttng-logger}
4665 The `msg` field of `lttng_logger` event records contains the
4668 NOTE: The maximum message length of an LTTng logger event is
4669 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4670 than one event to contain the remaining data.
4672 You should not use the LTTng logger to trace a user application which
4673 can be instrumented in a more efficient way, namely:
4675 * <<c-application,C and $$C++$$ applications>>.
4676 * <<java-application,Java applications>>.
4677 * <<python-application,Python applications>>.
4680 [[instrumenting-linux-kernel]]
4681 === LTTng kernel tracepoints
4683 NOTE: This section shows how to _add_ instrumentation points to the
4684 Linux kernel. The kernel's subsystems are already thoroughly
4685 instrumented at strategic places for LTTng when you
4686 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
4690 There are two methods to instrument the Linux kernel:
4692 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
4693 tracepoint which uses the `TRACE_EVENT()` API.
4695 Choose this if you want to instrumentation a Linux kernel tree with an
4696 instrumentation point compatible with ftrace, perf, and SystemTap.
4698 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
4699 instrument an out-of-tree kernel module.
4701 Choose this if you don't need ftrace, perf, or SystemTap support.
4705 [[linux-add-lttng-layer]]
4706 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
4708 This section shows how to add an LTTng layer to existing ftrace
4709 instrumentation using the `TRACE_EVENT()` API.
4711 This section does not document the `TRACE_EVENT()` macro. You can
4712 read the following articles to learn more about this API:
4714 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
4715 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
4716 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
4718 The following procedure assumes that your ftrace tracepoints are
4719 correctly defined in their own header and that they are created in
4720 one source file using the `CREATE_TRACE_POINTS` definition.
4722 To add an LTTng layer over an existing ftrace tracepoint:
4724 . Make sure the following kernel configuration options are
4730 * `CONFIG_HIGH_RES_TIMERS`
4731 * `CONFIG_TRACEPOINTS`
4734 . Build the Linux source tree with your custom ftrace tracepoints.
4735 . Boot the resulting Linux image on your target system.
4737 Confirm that the tracepoints exist by looking for their names in the
4738 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
4739 is your subsystem's name.
4741 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
4747 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
4748 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
4749 cd lttng-modules-2.8.*
4753 . In dir:{instrumentation/events/lttng-module}, relative to the root
4754 of the LTTng-modules source tree, create a header file named
4755 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
4756 LTTng-modules tracepoint definitions using the LTTng-modules
4759 Start with this template:
4763 .path:{instrumentation/events/lttng-module/my_subsys.h}
4766 #define TRACE_SYSTEM my_subsys
4768 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
4769 #define _LTTNG_MY_SUBSYS_H
4771 #include "../../../probes/lttng-tracepoint-event.h"
4772 #include <linux/tracepoint.h>
4774 LTTNG_TRACEPOINT_EVENT(
4776 * Format is identical to TRACE_EVENT()'s version for the three
4777 * following macro parameters:
4780 TP_PROTO(int my_int, const char *my_string),
4781 TP_ARGS(my_int, my_string),
4783 /* LTTng-modules specific macros */
4785 ctf_integer(int, my_int_field, my_int)
4786 ctf_string(my_bar_field, my_bar)
4790 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
4792 #include "../../../probes/define_trace.h"
4796 The entries in the `TP_FIELDS()` section are the list of fields for the
4797 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
4798 ftrace's `TRACE_EVENT()` macro.
4800 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
4801 complete description of the available `ctf_*()` macros.
4803 . Create the LTTng-modules probe's kernel module C source file,
4804 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
4809 .path:{probes/lttng-probe-my-subsys.c}
4811 #include <linux/module.h>
4812 #include "../lttng-tracer.h"
4815 * Build-time verification of mismatch between mainline
4816 * TRACE_EVENT() arguments and the LTTng-modules adaptation
4817 * layer LTTNG_TRACEPOINT_EVENT() arguments.
4819 #include <trace/events/my_subsys.h>
4821 /* Create LTTng tracepoint probes */
4822 #define LTTNG_PACKAGE_BUILD
4823 #define CREATE_TRACE_POINTS
4824 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
4826 #include "../instrumentation/events/lttng-module/my_subsys.h"
4828 MODULE_LICENSE("GPL and additional rights");
4829 MODULE_AUTHOR("Your name <your-email>");
4830 MODULE_DESCRIPTION("LTTng my_subsys probes");
4831 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
4832 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
4833 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
4834 LTTNG_MODULES_EXTRAVERSION);
4838 . Edit path:{probes/KBuild} and add your new kernel module object
4839 next to the existing ones:
4843 .path:{probes/KBuild}
4847 obj-m += lttng-probe-module.o
4848 obj-m += lttng-probe-power.o
4850 obj-m += lttng-probe-my-subsys.o
4856 . Build and install the LTTng kernel modules:
4861 make KERNELDIR=/path/to/linux
4862 sudo make modules_install
4866 Replace `/path/to/linux` with the path to the Linux source tree where
4867 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
4869 Note that you can also use the
4870 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
4871 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
4872 C code that need to be executed before the event fields are recorded.
4874 The best way to learn how to use the previous LTTng-modules macros is to
4875 inspect the existing LTTng-modules tracepoint definitions in the
4876 dir:{instrumentation/events/lttng-module} header files. Compare them
4877 with the Linux kernel mainline versions in the
4878 dir:{include/trace/events} directory of the Linux source tree.
4882 [[lttng-tracepoint-event-code]]
4883 ===== Use custom C code to access the data for tracepoint fields
4885 Although we recommended to always use the
4886 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
4887 the arguments and fields of an LTTng-modules tracepoint when possible,
4888 sometimes you need a more complex process to access the data that the
4889 tracer records as event record fields. In other words, you need local
4890 variables and multiple C{nbsp}statements instead of simple
4891 argument-based expressions that you pass to the
4892 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
4894 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
4895 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
4896 a block of C{nbsp}code to be executed before LTTng records the fields.
4897 The structure of this macro is:
4900 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
4902 LTTNG_TRACEPOINT_EVENT_CODE(
4904 * Format identical to the LTTNG_TRACEPOINT_EVENT()
4905 * version for the following three macro parameters:
4908 TP_PROTO(int my_int, const char *my_string),
4909 TP_ARGS(my_int, my_string),
4911 /* Declarations of custom local variables */
4914 unsigned long b = 0;
4915 const char *name = "(undefined)";
4916 struct my_struct *my_struct;
4920 * Custom code which uses both tracepoint arguments
4921 * (in TP_ARGS()) and local variables (in TP_locvar()).
4923 * Local variables are actually members of a structure pointed
4924 * to by the special variable tp_locvar.
4928 tp_locvar->a = my_int + 17;
4929 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
4930 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
4931 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
4932 put_my_struct(tp_locvar->my_struct);
4941 * Format identical to the LTTNG_TRACEPOINT_EVENT()
4942 * version for this, except that tp_locvar members can be
4943 * used in the argument expression parameters of
4944 * the ctf_*() macros.
4947 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
4948 ctf_integer(int, my_struct_a, tp_locvar->a)
4949 ctf_string(my_string_field, my_string)
4950 ctf_string(my_struct_name, tp_locvar->name)
4955 IMPORTANT: The C code defined in `TP_code()` must not have any side
4956 effects when executed. In particular, the code must not allocate
4957 memory or get resources without deallocating this memory or putting
4958 those resources afterwards.
4961 [[instrumenting-linux-kernel-tracing]]
4962 ==== Load and unload a custom probe kernel module
4964 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
4965 kernel module>> in the kernel before it can emit LTTng events.
4967 To load the default probe kernel modules and a custom probe kernel
4970 * Use the `--extra-kmod-probes` option to give extra probe modules
4971 to load when starting a root <<lttng-sessiond,session daemon>>:
4974 .Load the `my_subsys`, `usb`, and the default probe modules.
4978 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
4983 You only need to pass the subsystem name, not the whole kernel module
4986 To load _only_ a given custom probe kernel module:
4988 * Use the `--kmod-probes` option to give the probe modules
4989 to load when starting a root session daemon:
4992 .Load only the `my_subsys` and `usb` probe modules.
4996 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5001 To confirm that a probe module is loaded:
5008 lsmod | grep lttng_probe_usb
5012 To unload the loaded probe modules:
5014 * Kill the session daemon with `SIGTERM`:
5019 sudo pkill lttng-sessiond
5023 You can also use man:modprobe(8)'s `--remove` option if the session
5024 daemon terminates abnormally.
5027 [[controlling-tracing]]
5030 Once an application or a Linux kernel is
5031 <<instrumenting,instrumented>> for LTTng tracing,
5034 This section is divided in topics on how to use the various
5035 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5036 command-line tool>>, to _control_ the LTTng daemons and tracers.
5038 Note that the <<online-lttng-manpages,Online LTTng man pages>> are
5039 more comprehensive than the guides of this section. Refer to them if
5040 your use case is not included in this section.
5044 === Start a session daemon
5046 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5047 _before_ you can use the cmd:lttng command-line tool.
5049 You will see the following error when you run a command while no session
5053 Error: No session daemon is available
5056 The only command that automatically runs a session daemon is `create`,
5057 which you use to <<creating-destroying-tracing-sessions,create a tracing
5058 session>>. While this is most of the time the first operation that you
5059 do, sometimes it's not. Some examples are:
5061 * <<list-instrumentation-points,List the available instrumentation points>>.
5062 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5064 [[tracing-group]] Each Unix user must have its own running session
5065 daemon to trace user applications. The session daemon that the root user
5066 starts is the only one allowed to control the LTTng kernel tracer. Users
5067 that are part of the _tracing group_ can control the root session
5068 daemon. The default tracing group name is `tracing`; you can set it to
5069 something else with the `--group` option when you start the root session
5072 To start a user session daemon:
5074 * Run cmd:lttng-sessiond:
5079 lttng-sessiond --daemonize
5083 To start the root session daemon:
5085 * Run cmd:lttng-sessiond as the root user:
5090 sudo lttng-sessiond --daemonize
5094 In both cases, remove the `--daemonize` option to start the session
5095 daemon in foreground.
5097 To stop a session daemon, use cmd:kill on its process ID (standard
5100 Note that some Linux distributions could manage the LTTng session daemon
5101 as a service. In this case, you should use the service manager to
5102 start, restart, and stop session daemons.
5105 [[creating-destroying-tracing-sessions]]
5106 === Create and destroy a tracing session
5108 Almost all the LTTng control operations happen in the scope of
5109 a <<tracing-session,tracing session>>, which is the dialogue between the
5110 <<lttng-sessiond,session daemon>> and you.
5112 To create a tracing session with a generated name:
5114 * Use the `create` command:
5123 The created tracing session's name is `auto` followed by the
5126 To create a tracing session with a specific name:
5128 * Use the optional argument of the `create` command:
5133 lttng create my-session
5137 Replace `my-session` with the specific tracing session name.
5139 LTTng appends the creation date to the created tracing session's name.
5141 LTTng writes the traces of a tracing session in
5142 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5143 name of the tracing session. Note that the env:LTTNG_HOME environment
5144 variable defaults to `$HOME` if not set.
5146 To output LTTng traces to a non-default location:
5148 * Use the `--output` option of the `create` command:
5153 lttng create my-session --output=/tmp/some-directory
5157 You may create as many tracing sessions as you wish.
5159 To list all the existing tracing sessions for your Unix user:
5161 * Use the `list` command:
5170 When you create a tracing session, it is set as the _current tracing
5171 session_. The following man:lttng(1) commands operate on the current
5172 tracing session when you don't specify one:
5174 [role="list-3-cols"]
5190 To change the current tracing session:
5192 * Use the `set-session` command:
5197 lttng set-session new-session
5201 Replace `new-session` by the name of the new current tracing session.
5203 When you are done tracing in a given tracing session, you can destroy
5204 it. This operation frees the resources taken by the tracing session
5205 to destroy; it does not destroy the trace data that LTTng wrote for
5206 this tracing session.
5208 To destroy the current tracing session:
5210 * Use the `destroy` command:
5220 [[list-instrumentation-points]]
5221 === List the available instrumentation points
5223 The <<lttng-sessiond,session daemon>> can query the running instrumented
5224 user applications and the Linux kernel to get a list of available
5225 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5226 they are tracepoints and system calls. For the user space tracing
5227 domain, they are tracepoints. For the other tracing domains, they are
5230 To list the available instrumentation points:
5232 * Use the `list` command with the requested tracing domain's option
5236 * `--kernel`: Linux kernel tracepoints (your Unix user must be a root
5237 user, or it must be a member of the tracing group).
5238 * `--kernel --syscall`: Linux kernel system calls (your Unix user must
5239 be a root user, or it must be a member of the tracing group).
5240 * `--userspace`: user space tracepoints.
5241 * `--jul`: `java.util.logging` loggers.
5242 * `--log4j`: Apache log4j loggers.
5243 * `--python`: Python loggers.
5246 .List the available user space tracepoints.
5250 lttng list --userspace
5254 .List the available Linux kernel system call tracepoints.
5258 lttng list --kernel --syscall
5263 [[enabling-disabling-events]]
5264 === Create and enable an event rule
5266 Once you <<creating-destroying-tracing-sessions,create a tracing
5267 session>>, you can create <<event,event rules>> with the
5268 `enable-event` command.
5270 You specify each condition with a command-line option. The available
5271 condition options are shown in the following table.
5273 [role="growable",cols="asciidoc,asciidoc,default"]
5274 .Condition command-line options for the `enable-event` command.
5276 |Option |Description |Applicable tracing domains
5282 . +--probe=__ADDR__+
5283 . +--function=__ADDR__+
5286 Instead of using the default _tracepoint_ instrumentation type, use:
5288 . A Linux system call.
5289 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5290 . The entry and return points of a Linux function (symbol or address).
5294 |First positional argument.
5297 Tracepoint or system call name. In the case of a Linux KProbe or
5298 function, this is a custom name given to the event rule. With the
5299 JUL, log4j, and Python domains, this is a logger name.
5301 With a tracepoint, logger, or system call name, the last character
5302 can be `*` to match anything that remains.
5309 . +--loglevel=__LEVEL__+
5310 . +--loglevel-only=__LEVEL__+
5313 . Match only tracepoints or log statements with a logging level at
5314 least as severe as +__LEVEL__+.
5315 . Match only tracepoints or log statements with a logging level
5316 equal to +__LEVEL__+.
5318 You can get the list of available logging level names with
5319 `lttng enable-event --help`.
5321 |User space, JUL, log4j, and Python.
5323 |+--exclude=__EXCLUSIONS__+
5326 When you use a `*` character at the end of the tracepoint or logger
5327 name (first positional argument), exclude the specific names in the
5328 comma-delimited list +__EXCLUSIONS__+.
5331 User space, JUL, log4j, and Python.
5333 |+--filter=__EXPR__+
5336 Match only events which satisfy the expression +__EXPR__+.
5338 +__EXPR__+ is a C-like logical expression where identifiers are event
5339 fields (preceded with `$ctx.` for context fields). Nested expressions
5340 with `(` and `)`, and all the logical and comparison operators of the C
5341 language are supported. The precedence rules of those operators are the
5342 same as in the C language.
5344 When a comparison includes a non-existent event field, the whole filter
5345 expression evaluates to false.
5347 C integer and floating point number constants are supported, as well as
5348 literal strings between double quotes (`"`). Literal strings can
5349 contain a wildcard character (`*`) at the end to match anything that
5350 remains. This wildcard can be escaped using `\*`.
5352 Note that, although it is possible to use this option with the JUL,
5353 log4j, and Python tracing domains, the tracer evalutes the expression
5354 against the equivalent user space event.
5361 for more details about those command-line options.
5363 You attach an event rule to a <<channel,channel>> on creation. If you
5364 do not specify the channel with the `--channel` option, and if the event
5365 rule to create is the first in its <<domain,tracing domain>> for a given
5366 tracing session, then LTTng creates a _default channel_ for you. This
5367 default channel is reused in subsequent invocations of the
5368 `enable-event` command for the same tracing domain.
5370 An event rule is always enabled at creation time.
5372 The following examples show how you can combine the previous
5373 command-line options to create simple to more complex event rules.
5375 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5379 lttng enable-event --kernel sched_switch
5383 .Create an event rule matching four Linux kernel system calls (default channel).
5387 lttng enable-event --kernel --syscall open,write,read,close
5391 .Create an event rule matching a Linux kernel tracepoint with a filter expression (default channel).
5395 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5398 IMPORTANT: Make sure to always quote the filter string when you
5399 use man:lttng(1) from a shell.
5402 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5406 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5409 IMPORTANT: Make sure to always quote the wildcard character when you
5410 use man:lttng(1) from a shell.
5413 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5417 lttng enable-event --python my-app.'*' \
5418 --exclude='my-app.module,my-app.hello'
5422 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5426 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5430 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5434 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5438 The event rules of a given channel form a whitelist: as soon as an
5439 emitted event passes one of them, LTTng can record the event. For
5440 example, an event named `my_app:my_tracepoint` emitted from a user space
5441 tracepoint with a `TRACE_ERROR` log level passes both of the following
5446 lttng enable-event --userspace my_app:my_tracepoint
5447 lttng enable-event --userspace my_app:my_tracepoint \
5448 --loglevel=TRACE_INFO
5451 The second event rule is redundant: the first one includes
5455 [[disable-event-rule]]
5456 === Disable an event rule
5458 To disable an event rule that you <<enabling-disabling-events,created>>
5459 previously, use the `disable-event` command. This command disables _all_
5460 the event rules (of a given tracing domain and channel) which match an
5461 instrumentation point. The other conditions are not supported as of
5462 LTTng{nbsp}{revision}.
5464 The LTTng tracer does not record an emitted event which passes
5465 a _disabled_ event rule.
5467 .Disable an event rule matching a Python logger (default channel).
5471 lttng disable-event --python my-logger
5475 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5479 lttng disable-event --jul '*'
5483 .Disable _all_ the event rules of the default channel.
5485 The `--all-events` option is not, like the `--all` option of
5486 `enable-event`, the equivalent of the event name `*` (wildcard): it
5487 disables _all_ the event rules of a given channel.
5491 lttng disable-event --jul --all-events
5495 NOTE: You cannot delete an event rule once you create it.
5499 === Get the status of a tracing session
5501 To get the status of a tracing session, that is, its channels, event
5502 rules, and their attributes:
5504 * Use the `list` command with the tracing session's name:
5509 lttng list my-session
5513 Replace `my-session` with your tracing session's name.
5516 [[basic-tracing-session-control]]
5517 === Start and stop a tracing session
5519 Once you <<creating-destroying-tracing-sessions,create a tracing
5521 <<enabling-disabling-events,create one or more event rules>>,
5522 you can start and stop the tracers for this tracing session.
5524 To start tracing in the current tracing session:
5526 * Use the `start` command:
5535 To stop tracing in the current tracing session:
5537 * Use the `stop` command:
5546 LTTng is very flexible: you can launch user applications before
5547 or after the you start the tracers. The tracers only record the events
5548 if they pass enabled event rules and if they occur while the tracers are
5552 [[enabling-disabling-channels]]
5553 === Create a channel
5555 Once you create a tracing session, you can create a <<channel,channel>>
5556 with the `enable-channel` command.
5558 Note that LTTng automatically creates a default channel when, for a
5559 given <<domain,tracing domain>>, no channels exist and you
5560 <<enabling-disabling-events,create>> the first event rule. This default
5561 channel is named `channel0` and its attributes are set to reasonable
5562 values. Therefore, you only need to create a channel when you need
5563 non-default attributes.
5565 You specify each non-default channel attribute with a command-line
5566 option when you use the `enable-channel` command. The available
5567 command-line options are:
5569 [role="growable",cols="asciidoc,asciidoc"]
5570 .Command-line options for the `enable-channel` command.
5572 |Option |Description
5578 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5579 the default _discard_ mode.
5581 |`--buffers-pid` (user space tracing domain only)
5584 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5585 instead of the default per-user buffering scheme.
5587 |+--subbuf-size=__SIZE__+
5590 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5591 either for each Unix user (default), or for each instrumented process.
5593 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5595 |+--num-subbuf=__COUNT__+
5598 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5599 for each Unix user (default), or for each instrumented process.
5601 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5603 |+--tracefile-size=__SIZE__+
5606 Set the maximum size of each trace file that this channel writes within
5607 a stream to +__SIZE__+ bytes instead of no maximum.
5609 See <<tracefile-rotation,Trace file count and size>>.
5611 |+--tracefile-count=__COUNT__+
5614 Limit the number of trace files that this channel creates to
5615 +__COUNT__+ channels instead of no limit.
5617 See <<tracefile-rotation,Trace file count and size>>.
5619 |+--switch-timer=__PERIODUS__+
5622 Set the <<channel-switch-timer,switch timer period>>
5623 to +__PERIODUS__+{nbsp}µs.
5625 |+--read-timer=__PERIODUS__+
5628 Set the <<channel-read-timer,read timer period>>
5629 to +__PERIODUS__+{nbsp}µs.
5631 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5634 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5639 for more details about those command-line options.
5641 You can only create a channel in the Linux kernel and user space
5642 <<domain,tracing domains>>: other tracing domains have their own
5643 channel created on the fly when
5644 <<enabling-disabling-events,creating event rules>>.
5648 Because of a current LTTng limitation, you must create all channels
5649 _before_ you <<basic-tracing-session-control,start tracing>> in a given
5650 tracing session, that is, before the first time you run `lttng start`.
5652 Since LTTng automatically creates a default channel when you use the
5653 `enable-event` command with a specific tracing domain, you cannot, for
5654 example, create a Linux kernel event rule, start tracing, and then
5655 create a user space event rule, because no user space channel exists yet
5656 and it's too late to create one.
5658 For this reason, make sure to configure your channels properly
5659 before starting the tracers for the first time!
5662 The following examples show how you can combine the previous
5663 command-line options to create simple to more complex channels.
5665 .Create a Linux kernel channel with default attributes.
5669 lttng enable-channel --kernel my-channel
5673 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
5677 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
5678 --buffers-pid my-channel
5682 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
5686 lttng enable-channel --kernel --tracefile-count=8 \
5687 --tracefile-size=4194304 my-channel
5691 .Create a user space channel in overwrite (or _flight recorder_) mode.
5695 lttng enable-channel --userspace --overwrite my-channel
5699 You can <<enabling-disabling-events,create>> the same event rule in
5700 two different channels:
5704 lttng enable-event --userspace --channel=my-channel app:tp
5705 lttng enable-event --userspace --channel=other-channel app:tp
5708 If both channels are enabled, when a tracepoint named `app:tp` is
5709 reached, LTTng records two events, one for each channel.
5713 === Disable a channel
5715 To disable a specific channel that you <<enabling-disabling-channels,created>>
5716 previously, use the `disable-channel` command.
5718 .Disable a specific Linux kernel channel.
5722 lttng disable-channel --kernel my-channel
5726 The state of a channel precedes the individual states of event rules
5727 attached to it: event rules which belong to a disabled channel, even if
5728 they are enabled, are also considered disabled.
5732 === Add context fields to a channel
5734 Event record fields in trace files provide important information about
5735 events that occured previously, but sometimes some external context may
5736 help you solve a problem faster. Examples of context fields are:
5738 * The **process ID**, **thread ID**, **process name**, and
5739 **process priority** of the thread in which the event occurs.
5740 * The **hostname** of the system on which the event occurs.
5741 * The current values of many possible **performance counters** using
5743 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
5745 ** Branch instructions, misses, and loads.
5748 To get the full list of available context fields, see
5749 `lttng add-context --help`. Some context fields are reserved for a
5750 specific <<domain,tracing domain>> (Linux kernel or user space).
5752 You add context fields to <<channel,channels>>. All the events
5753 that a channel with added context fields records contain those fields.
5755 To add context fields to one or all the channels of a given tracing
5756 session, use the `add-context` command.
5758 .Add context fields to all the channels of the current tracing session.
5760 The following command line adds the virtual process identifier and
5761 the per-thread CPU cycles count fields to all the user space channels
5762 of the current tracing session.
5766 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
5770 .Add a context field to a specific channel.
5772 The following command line adds the thread identifier context field
5773 to the Linux kernel channel named `my-channel` in the current
5778 lttng add-context --kernel --channel=my-channel --type=tid
5782 NOTE: You cannot remove context fields from a channel once you add it.
5787 === Track process IDs
5789 It's often useful to allow only specific process IDs (PIDs) to emit
5790 events. For example, you may wish to record all the system calls made by
5791 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
5793 The `track` and `untrack` commands serve this purpose. Both commands
5794 operate on a whitelist of process IDs. You _add_ entries to this
5795 whitelist with the `track` command and remove entries with the `untrack`
5796 command. Any process which has one of the PIDs in the whitelist is
5797 allowed to emit LTTng events which pass an enabled <<event,event rule>>.
5799 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
5800 process with a given tracked ID exit and another process be given this
5801 ID, then the latter would also be allowed to emit events.
5803 .Track and untrack process IDs.
5805 For the sake of the following example, assume the target system has 16
5809 <<creating-destroying-tracing-sessions,create a tracing session>>,
5810 the whitelist contains all the possible PIDs:
5813 .All PIDs are tracked.
5814 image::track-all.png[]
5816 When the whitelist is full and you use the `track` command to specify
5817 some PIDs to track, LTTng first clears the whitelist, then it tracks
5818 the specific PIDs. After:
5822 lttng track --pid=3,4,7,10,13
5828 .PIDs 3, 4, 7, 10, and 13 are tracked.
5829 image::track-3-4-7-10-13.png[]
5831 You can add more PIDs to the whitelist afterwards:
5835 lttng track --pid=1,15,16
5841 .PIDs 1, 15, and 16 are added to the whitelist.
5842 image::track-1-3-4-7-10-13-15-16.png[]
5844 The `untrack` command removes entries from the PID tracker's whitelist.
5845 Given the previous example, the following command:
5849 lttng untrack --pid=3,7,10,13
5852 leads to this whitelist:
5855 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
5856 image::track-1-4-15-16.png[]
5858 LTTng can track all possible PIDs again using the `--all` option:
5862 lttng track --pid --all
5865 The result is, again:
5868 .All PIDs are tracked.
5869 image::track-all.png[]
5872 .Track only specific PIDs
5874 A very typical use case with PID tracking is to start with an empty
5875 whitelist, then <<basic-tracing-session-control,start the tracers>>,
5876 and then add PIDs manually while tracers are active. You can accomplish
5877 this by using the `--all` option of the `untrack` command to clear the
5878 whitelist after you create a tracing session:
5882 lttng untrack --pid --all
5888 .No PIDs are tracked.
5889 image::untrack-all.png[]
5891 If you trace with this whitelist configuration, the tracer records no
5892 events for this <<domain,tracing domain>> because no processes are
5893 tracked. You can use the `track` command as usual to track specific
5898 lttng track --pid=6,11
5904 .PIDs 6 and 11 are tracked.
5905 image::track-6-11.png[]
5910 [[saving-loading-tracing-session]]
5911 === Save and load tracing session configurations
5913 Configuring a <<tracing-session,tracing session>> can be long. Some of
5914 the tasks involved are:
5916 * <<enabling-disabling-channels,Create channels>> with
5917 specific attributes.
5918 * <<adding-context,Add context fields>> to specific channels.
5919 * <<enabling-disabling-events,Create event rules>> with specific log
5920 level and filter conditions.
5922 If you use LTTng to solve real world problems, chances are you have to
5923 record events using the same tracing session setup over and over,
5924 modifying a few variables each time in your instrumented program
5925 or environment. To avoid constant tracing session reconfiguration,
5926 the cmd:lttng command-line tool can save and load tracing session
5927 configurations to/from XML files.
5929 To save a given tracing session configuration:
5931 * Use the `save` command:
5936 lttng save my-session
5940 Replace `my-session` with the name of the tracing session to save.
5942 LTTng saves tracing session configurations to
5943 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
5944 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
5945 the `--output-path` option to change this destination directory.
5947 LTTng saves all configuration parameters, for example:
5949 * The tracing session name.
5950 * The trace data output path.
5951 * The channels with their state and all their attributes.
5952 * The context fields you added to channels.
5953 * The event rules with their state, log level and filter conditions.
5955 To load a tracing session:
5957 * Use the `load` command:
5962 lttng load my-session
5966 Replace `my-session` with the name of the tracing session to load.
5968 When LTTng loads a configuration, it restores your saved tracing session
5969 as if you just configured it manually.
5971 See man:lttng(1) for the complete list of command-line options. You
5972 can also save and load all many sessions at a time, and decide in which
5973 directory to output the XML files.
5976 [[sending-trace-data-over-the-network]]
5977 === Send trace data over the network
5979 LTTng can send the recorded trace data to a remote system over the
5980 network instead of writing it to the local file system.
5982 To send the trace data over the network:
5984 . On the _remote_ system (which can also be the target system),
5985 start an LTTng <<lttng-relayd,relay daemon>>:
5994 . On the _target_ system, create a tracing session configured to
5995 send trace data over the network:
6000 lttng create my-session --set-url=net://remote-system
6004 Replace `remote-system` by the host name or IP address of the
6005 remote system. See `lttng create --help` for the exact URL format.
6007 . On the target system, use the cmd:lttng command-line tool as usual.
6008 When tracing is active, the target's consumer daemon sends sub-buffers
6009 to the relay daemon running on the remote system intead of flushing
6010 them to the local file system. The relay daemon writes the received
6011 packets to the local file system.
6013 The relay daemon writes trace files to
6014 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6015 +__hostname__+ is the host name of the target system and +__session__+
6016 is the tracing session name. Note that the env:LTTNG_HOME environment
6017 variable defaults to `$HOME` if not set. Use the `--output` option of
6018 cmd:lttng-relayd to write trace files to another base directory.
6023 === View events as LTTng emits them (noch:{LTTng} live)
6025 LTTng live is a network protocol implemented by the
6026 <<lttng-relayd,relay daemon>> to allow compatible trace viewers to
6027 display events as LTTng emits them on the target system while tracing
6030 The relay daemon creates a _tee_: it forwards the trace data to both
6031 the local file system and to connected live viewers:
6034 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6039 . On the _target system_, create a <<tracing-session,tracing session>>
6045 lttng create my-session --live
6049 This spawns a local relay daemon.
6051 . Start the live viewer and configure it to connect to the relay
6052 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6057 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6064 * `hostname` with the host name of the target system.
6065 * `my-session` with the name of the tracing session to view.
6068 . Configure the tracing session as usual with the cmd:lttng
6069 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6071 You can list the available live tracing sessions with Babeltrace:
6075 babeltrace --input-format=lttng-live net://localhost
6078 You can start the relay daemon on another system. In this case, you need
6079 to specify the relay daemon's URL when you create the tracing session
6080 with the `--set-url` option. You also need to replace `localhost`
6081 in the procedure above with the host name of the system on which the
6082 relay daemon is running.
6084 See man:lttng(1) and man:lttng-relayd(8) for the complete list of
6085 command-line options.
6089 [[taking-a-snapshot]]
6090 === Take a snapshot of the current sub-buffers of a tracing session
6092 The normal behavior of LTTng is to append full sub-buffers to growing
6093 trace data files. This is ideal to keep a full history of the events
6094 that occurred on the target system, but it can
6095 represent too much data in some situations. For example, you may wish
6096 to trace your application continuously until some critical situation
6097 happens, in which case you only need the latest few recorded
6098 events to perform the desired analysis, not multi-gigabyte trace files.
6100 With the `snapshot` command, you can take a snapshot of the current
6101 sub-buffers of a given <<tracing-session,tracing session>>. LTTng can
6102 write the snapshot to the local file system or send it over the network.
6106 . Create a tracing session in _snapshot mode_:
6111 lttng create my-session --snapshot
6115 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6116 <<channel,channels>> created in this mode is automatically set to
6117 _overwrite_ (flight recorder mode).
6119 . Configure the tracing session as usual with the cmd:lttng
6120 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6122 . **Optional**: When you need to take a snapshot, stop tracing.
6124 You can take a snapshot when the tracers are active, but if you stop
6125 them first, you are sure that the data in the sub-buffers does not
6126 change before you actually take the snapshot.
6133 lttng snapshot record --name=my-first-snapshot
6137 LTTng writes the current sub-buffers of all the current tracing
6138 session's channels to trace files on the local file system. Those trace
6139 files have `my-first-snapshot` in their name.
6141 There is no difference between the format of a normal trace file and the
6142 format of a snapshot: viewers of LTTng traces also support LTTng
6145 By default, LTTng writes snapshot files to the path shown by
6146 `lttng snapshot list-output`. You can change this path or decide to send
6147 snapshots over the network using either:
6149 . An output path or URL that you specify when you create the
6151 . An snapshot output path or URL that you add using
6152 `lttng snapshot add-output`
6153 . An output path or URL that you provide directly to the
6154 `lttng snapshot record` command.
6156 Method 3 overrides method 2, which overrides method 1. When you
6157 specify a URL, a relay daemon must listen on a remote system (see
6158 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6163 === Use the machine interface
6165 With any command of the cmd:lttng command-line tool, you can use the
6166 `--mi=xml` argument (before the command name) to get an XML machine
6167 interface output, for example:
6171 lttng --mi=xml enable-event --kernel --syscall open
6174 A schema definition (XSD) is
6175 https://github.com/lttng/lttng-tools/blob/stable-{revision}/src/common/mi_lttng.xsd[available]
6176 to ease the integration with external tools as much as possible.
6180 [[persistent-memory-file-systems]]
6181 === Record trace data on persistent memory file systems
6183 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6184 (NVRAM) is random-access memory that retains its information when power
6185 is turned off (non-volatile). Systems with such memory can store data
6186 structures in RAM and retrieve them after a reboot, without flushing
6187 to typical _storage_.
6189 Linux supports NVRAM file systems thanks to either
6190 http://pramfs.sourceforge.net/[PRAMFS] or
6191 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6192 (requires Linux 4.1+).
6194 This section does not describe how to operate such file systems;
6195 we assume that you have a working persistent memory file system.
6197 When you create a <<tracing-session,tracing session>>, you can specify
6198 the path of the shared memory holding the sub-buffers. If you specify a
6199 location on an NVRAM file system, then you can retrieve the latest
6200 recorded trace data when the system reboots after a crash.
6202 To record trace data on a persistent memory file system and retrieve the
6203 trace data after a system crash:
6205 . Create a tracing session with a sub-buffer shared memory path located
6206 on an NVRAM file system:
6211 lttng create my-session -shm-path=/path/to/shm
6215 . Configure the tracing session as usual with the cmd:lttng
6216 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6218 . After a system crash, use the cmd:lttng-crash command-line tool to
6219 view the trace data recorded on the NVRAM file system:
6224 lttng-crash /path/to/shm
6228 The binary layout of the ring buffer files is not exactly the same as
6229 the trace files layout. This is why you need to use the cmd:lttng-crash
6230 utility instead of your preferred trace viewer directly.
6232 To convert the ring buffer files to LTTng trace files:
6234 * Use the `--extract` option of cmd:lttng-crash:
6239 lttng-crash --extract=/path/to/trace /path/to/shm
6243 See man:lttng-crash(1) for the complete list of command-line options.
6249 This section presents various references for LTTng packages such as
6250 links to online manpages, tables that the rest of the text needs,
6251 descriptions of library functions, and more.
6254 [[online-lttng-manpages]]
6255 === Online noch:{LTTng} manpages
6257 LTTng packages currently install the following link:/man[man pages],
6258 available online using the links below:
6262 ** man:lttng-crash(1)
6263 ** man:lttng-sessiond(8)
6264 ** man:lttng-relayd(8)
6266 ** man:lttng-gen-tp(1)
6268 ** man:lttng-ust-cyg-profile(3)
6269 ** man:lttng-ust-dl(3)
6273 === noch:{LTTng-UST}
6275 This section presents references of the LTTng-UST package.
6279 ==== noch:{LTTng-UST} library (+liblttng‑ust+)
6281 The LTTng-UST library, or `liblttng-ust`, is the main shared object
6282 against which user applications are linked to make LTTng user space
6285 The <<c-application,C application>> guide shows the complete
6286 process to instrument, build and run a C/$$C++$$ application using
6287 LTTng-UST, while this section contains a few important tables.
6290 [[liblttng-ust-tp-fields]]
6291 ===== Tracepoint fields macros (for `TP_FIELDS()`)
6293 The available macros to define tracepoint fields, which you must use
6294 within `TP_FIELDS()` in `TRACEPOINT_EVENT()`, are:
6296 [role="func-desc growable",cols="asciidoc,asciidoc"]
6297 .Available macros to define LTTng-UST tracepoint fields
6299 |Macro |Description and parameters
6302 +ctf_integer(__t__, __n__, __e__)+
6304 +ctf_integer_nowrite(__t__, __n__, __e__)+
6306 Standard integer, displayed in base 10.
6309 Integer C type (`int`, `long`, `size_t`, ...).
6315 Argument expression.
6317 |+ctf_integer_hex(__t__, __n__, __e__)+
6319 Standard integer, displayed in base 16.
6328 Argument expression.
6330 |+ctf_integer_network(__t__, __n__, __e__)+
6332 Integer in network byte order (big-endian), displayed in base 10.
6341 Argument expression.
6343 |+ctf_integer_network_hex(__t__, __n__, __e__)+
6345 Integer in network byte order, displayed in base 16.
6354 Argument expression.
6357 +ctf_float(__t__, __n__, __e__)+
6359 +ctf_float_nowrite(__t__, __n__, __e__)+
6361 Floating point number.
6364 Floating point number C type (`float` or `double`).
6370 Argument expression.
6373 +ctf_string(__n__, __e__)+
6375 +ctf_string_nowrite(__n__, __e__)+
6377 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6383 Argument expression.
6386 +ctf_array(__t__, __n__, __e__, __s__)+
6388 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6390 Statically-sized array of integers
6393 Array element C type.
6399 Argument expression.
6405 +ctf_array_text(__t__, __n__, __e__, __s__)+
6407 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6409 Statically-sized array, printed as text.
6411 The string does not need to be null-terminated.
6414 Array element C type (always `char`).
6420 Argument expression.
6426 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6428 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6430 Dynamically-sized array of integers.
6432 The type of +__E__+ must be unsigned.
6435 Array element C type.
6441 Argument expression.
6444 Length expression C type.
6450 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6452 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6454 Dynamically-sized array, displayed as text.
6456 The string does not need to be null-terminated.
6458 The type of +__E__+ must be unsigned.
6460 The behaviour is undefined if +__e__+ is `NULL`.
6463 Sequence element C type (always `char`).
6469 Argument expression.
6472 Length expression C type.
6478 The `_nowrite` versions omit themselves from the session trace, but are
6479 otherwise identical. This means the tracer does not write the `_nowrite`
6480 fields to the trace. Their primary purpose is to make some of the event
6481 context available to the <<enabling-disabling-events,event filters>>
6482 without having to commit the data to sub-buffers.
6485 [[liblttng-ust-tracepoint-loglevel]]
6486 ===== Tracepoint log levels (for `TRACEPOINT_LOGLEVEL()`)
6488 The following table shows the available log level values for the
6489 `TRACEPOINT_LOGLEVEL()` macro:
6495 Action must be taken immediately.
6498 Critical conditions.
6507 Normal, but significant, condition.
6510 Informational message.
6512 `TRACE_DEBUG_SYSTEM`::
6513 Debug information with system-level scope (set of programs).
6515 `TRACE_DEBUG_PROGRAM`::
6516 Debug information with program-level scope (set of processes).
6518 `TRACE_DEBUG_PROCESS`::
6519 Debug information with process-level scope (set of modules).
6521 `TRACE_DEBUG_MODULE`::
6522 Debug information with module (executable/library) scope (set of units).
6524 `TRACE_DEBUG_UNIT`::
6525 Debug information with compilation unit scope (set of functions).
6527 `TRACE_DEBUG_FUNCTION`::
6528 Debug information with function-level scope.
6530 `TRACE_DEBUG_LINE`::
6531 Debug information with line-level scope (TRACEPOINT_EVENT default).
6534 Debug-level message.
6536 Log levels `TRACE_EMERG` through `TRACE_INFO` and `TRACE_DEBUG` match
6537 http://man7.org/linux/man-pages/man3/syslog.3.html[syslog]
6538 level semantics. Log levels `TRACE_DEBUG_SYSTEM` through `TRACE_DEBUG`
6539 offer more fine-grained selection of debug information.
6542 [[lttng-modules-ref]]
6543 === noch:{LTTng-modules}
6545 This section presents references of the LTTng-modules package.
6549 [[lttng-modules-tp-fields]]
6550 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6552 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6553 tracepoint fields, which must be listed within `TP_FIELDS()` in
6554 `LTTNG_TRACEPOINT_EVENT()`, are:
6556 [role="func-desc growable",cols="asciidoc,asciidoc"]
6557 .Available macros to define LTTng-modules tracepoint fields
6559 |Macro |Description and parameters
6562 +ctf_integer(__t__, __n__, __e__)+
6564 +ctf_integer_nowrite(__t__, __n__, __e__)+
6566 +ctf_user_integer(__t__, __n__, __e__)+
6568 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6570 Standard integer, displayed in base 10.
6573 Integer C type (`int`, `long`, `size_t`, ...).
6579 Argument expression.
6582 +ctf_integer_hex(__t__, __n__, __e__)+
6584 +ctf_user_integer_hex(__t__, __n__, __e__)+
6586 Standard integer, displayed in base 16.
6595 Argument expression.
6597 |+ctf_integer_oct(__t__, __n__, __e__)+
6599 Standard integer, displayed in base 8.
6608 Argument expression.
6611 +ctf_integer_network(__t__, __n__, __e__)+
6613 +ctf_user_integer_network(__t__, __n__, __e__)+
6615 Integer in network byte order (big-endian), displayed in base 10.
6624 Argument expression.
6627 +ctf_integer_network_hex(__t__, __n__, __e__)+
6629 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6631 Integer in network byte order, displayed in base 16.
6640 Argument expression.
6643 +ctf_string(__n__, __e__)+
6645 +ctf_string_nowrite(__n__, __e__)+
6647 +ctf_user_string(__n__, __e__)+
6649 +ctf_user_string_nowrite(__n__, __e__)+
6651 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6657 Argument expression.
6660 +ctf_array(__t__, __n__, __e__, __s__)+
6662 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6664 +ctf_user_array(__t__, __n__, __e__, __s__)+
6666 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6668 Statically-sized array of integers
6671 Array element C type.
6677 Argument expression.
6683 +ctf_array_text(__t__, __n__, __e__, __s__)+
6685 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6687 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6689 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6691 Statically-sized array, printed as text.
6693 The string does not need to be null-terminated.
6696 Array element C type (always `char`).
6702 Argument expression.
6708 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6710 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6712 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6714 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6716 Dynamically-sized array of integers.
6718 The type of +__E__+ must be unsigned.
6721 Array element C type.
6727 Argument expression.
6730 Length expression C type.
6735 |+ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6737 Dynamically-sized array of integers, displayed in base 16.
6739 The type of +__E__+ must be unsigned.
6742 Array element C type.
6748 Argument expression.
6751 Length expression C type.
6756 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
6758 Dynamically-sized array of integers in network byte order (big-endian),
6759 displayed in base 10.
6761 The type of +__E__+ must be unsigned.
6764 Array element C type.
6770 Argument expression.
6773 Length expression C type.
6779 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6781 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6783 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6785 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6787 Dynamically-sized array, displayed as text.
6789 The string does not need to be null-terminated.
6791 The type of +__E__+ must be unsigned.
6793 The behaviour is undefined if +__e__+ is `NULL`.
6796 Sequence element C type (always `char`).
6802 Argument expression.
6805 Length expression C type.
6811 Use the `_user` versions when the argument expression, `e`, is
6812 a user space address. In the cases of `ctf_user_integer*()` and
6813 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
6816 The `_nowrite` versions omit themselves from the session trace, but are
6817 otherwise identical. This means the `_nowrite` fields won't be written
6818 in the recorded trace. Their primary purpose is to make some
6819 of the event context available to the
6820 <<enabling-disabling-events,event filters>> without having to
6821 commit the data to sub-buffers.
6827 Terms related to LTTng and to tracing in general:
6830 The http://diamon.org/babeltrace[Babeltrace] project, which includes
6831 the cmd:babeltrace command, some libraries, and Python bindings.
6833 <<channel-buffering-schemes,buffering scheme>>::
6834 A layout of sub-buffers applied to a given channel.
6836 <<channel,channel>>::
6837 An entity which is responsible for a set of ring buffers.
6839 <<event,Event rules>> are always attached to a specific channel.
6842 A reference of time for a tracer.
6844 <<lttng-consumerd,consumer daemon>>::
6845 A process which is responsible for consuming the full sub-buffers
6846 and write them to a file system or send them over the network.
6848 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
6849 mode in which the tracer _discards_ new event records when there's no
6850 sub-buffer space left to store them.
6853 The consequence of the execution of an instrumentation
6854 point, like a tracepoint that you manually place in some source code,
6855 or a Linux kernel KProbe.
6857 An event is said to _occur_ at a specific time. Different actions can
6858 be taken upon the occurrence of an event, like record the event's payload
6861 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
6862 The mechanism by which event records of a given channel are lost
6863 (not recorded) when there is no sub-buffer space left to store them.
6865 [[def-event-name]]event name::
6866 The name of an event, which is also the name of the event record.
6867 This is also called the _instrumentation point name_.
6870 A record, in a trace, of the payload of an event which occured.
6872 <<event,event rule>>::
6873 Set of conditions which must be satisfied for one or more occuring
6874 events to be recorded.
6876 `java.util.logging`::
6878 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
6880 <<instrumenting,instrumentation>>::
6881 The use of LTTng probes to make a piece of software traceable.
6883 instrumentation point::
6884 A point in the execution path of a piece of software that, when
6885 reached by this execution, can emit an event.
6887 instrumentation point name::
6888 See _<<def-event-name,event name>>_.
6891 A http://logging.apache.org/log4j/1.2/[logging library] for Java
6892 developed by the Apache Software Foundation.
6895 Level of severity of a log statement or user space
6896 instrumentation point.
6899 The _Linux Trace Toolkit: next generation_ project.
6901 <<lttng-cli,cmd:lttng>>::
6902 A command-line tool provided by the LTTng-tools project which you
6903 can use to send and receive control messages to and from a
6907 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
6908 which is a set of analyzing programs that are used to obtain a
6909 higher level view of an LTTng trace.
6911 cmd:lttng-consumerd::
6912 The name of the consumer daemon program.
6915 A utility provided by the LTTng-tools project which can convert
6916 ring buffer files (usually
6917 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
6920 LTTng Documentation::
6923 <<lttng-live,LTTng live>>::
6924 A communication protocol between the relay daemon and live viewers
6925 which makes it possible to see events "live", as they are received by
6928 <<lttng-modules,LTTng-modules>>::
6929 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
6930 which contains the Linux kernel modules to make the Linux kernel
6931 instrumentation points available for LTTng tracing.
6934 The name of the relay daemon program.
6936 cmd:lttng-sessiond::
6937 The name of the session daemon program.
6940 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
6941 contains the various programs and libraries used to
6942 <<controlling-tracing,control tracing>>.
6944 <<lttng-ust,LTTng-UST>>::
6945 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
6946 contains libraries to instrument user applications.
6948 <<lttng-ust-agents,LTTng-UST Java agent>>::
6949 A Java package provided by the LTTng-UST project to allow the
6950 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
6953 <<lttng-ust-agents,LTTng-UST Python agent>>::
6954 A Python package provided by the LTTng-UST project to allow the
6955 LTTng instrumentation of Python logging statements.
6957 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
6958 The event loss mode in which new event records overwrite older
6959 event records when there's no sub-buffer space left to store them.
6961 <<channel-buffering-schemes,per-process buffering>>::
6962 A buffering scheme in which each instrumented process has its own
6963 sub-buffers for a given user space channel.
6965 <<channel-buffering-schemes,per-user buffering>>::
6966 A buffering scheme in which all the processes of a Unix user share the
6967 same sub-buffer for a given user space channel.
6969 <<lttng-relayd,relay daemon>>::
6970 A process which is responsible for receiving the trace data sent by
6971 a distant consumer daemon.
6974 A set of sub-buffers.
6976 <<lttng-sessiond,session daemon>>::
6977 A process which receives control commands from you and orchestrates
6978 the tracers and various LTTng daemons.
6980 <<taking-a-snapshot,snapshot>>::
6981 A copy of the current data of all the sub-buffers of a given tracing
6982 session, saved as trace files.
6985 One part of an LTTng ring buffer which contains event records.
6988 The time information attached to an event when it is emitted.
6991 A set of files which are the concatenations of one or more
6992 flushed sub-buffers.
6995 The action of recording the events emitted by an application
6996 or by a system, or to initiate such recording by controlling
7000 The http://tracecompass.org[Trace Compass] project and application.
7003 An instrumentation point using the tracepoint mechanism of the Linux
7004 kernel or of LTTng-UST.
7006 tracepoint definition::
7007 The definition of a single tracepoint.
7010 The name of a tracepoint.
7012 tracepoint provider::
7013 A set of functions providing tracepoints to an instrumented user
7016 Not to be confused with a _tracepoint provider package_: many tracepoint
7017 providers can exist within a tracepoint provider package.
7019 tracepoint provider package::
7020 One or more tracepoint providers compiled as an object file or as
7024 A software which records emitted events.
7026 <<domain,tracing domain>>::
7027 A namespace for event sources.
7030 The Unix group in which a Unix user can be to be allowed to trace the
7033 <<tracing-session,tracing session>>::
7034 A stateful dialogue between you and a <<lttng-sessiond,session
7038 An application running in user space, as opposed to a Linux kernel
7039 module, for example.