1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference links and tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 * **Tracing control**:
78 ** Dynamic filter support for <<event,event rules>> in the Linux kernel
79 <<domain,tracing domain>>. For example:
84 lttng enable-event --kernel irq_handler_entry --filter='irq == 28'
88 ** Wildcard support in the instrumentation point name of an event rule
89 in the Linux kernel tracing domain. For example:
94 lttng enable-event --kernel 'sched_*'
98 ** New `lttng track` and `lttng untrack` commands to make
99 <<pid-tracking,PID tracking>> super-fast for both the Linux kernel
100 and the user space tracing domains.
102 When LTTng _tracks_ one or more PIDs, only the processes having those PIDs
103 can emit events for a given tracing session.
105 ** New `--shm-path` option of the `lttng create` command to specify the
106 path where LTTng creates the shared memory holding the ring buffers.
108 This feature is useful when used with persistent memory file systems to
109 extract the latest recorded trace data in the event of a crash requiring
112 The new man:lttng-crash(1) command-line utility can extract trace data
113 from such a file (see <<persistent-memory-file-systems,Record trace data
114 on persistent memory file systems>>).
116 * **User space tracing**:
117 ** New <<python-application,LTTng-UST Python agent>> which makes it easy
118 to trace existing Python applications that are using the standard
119 https://docs.python.org/3/howto/logging.html[`logging` package].
121 This agent is compatible with both the Python 2 and Python 3 languages.
123 ** New <<tracelog,`tracelog()`>> facility to ease the migration from
126 `tracelog()` is similar to <<tracef,`tracef()`>>,
127 but it accepts an additional log level parameter.
129 ** Plugin support in LTTng-UST to provide a custom clock source and to
130 retrieve the current CPU number.
132 This feature exists for very advanced use cases.
135 https://github.com/lttng/lttng-ust/tree/stable-{revision}/doc/examples/clock-override[clock-override]
137 https://github.com/lttng/lttng-ust/tree/stable-{revision}/doc/examples/getcpu-override[getcpu-override]
138 examples for more details.
140 Moreover, LTTng{nbsp}{revision} boasts great stability, benifiting from
141 piles of bug fixes and more-than-welcome internal refactorings.
143 To learn more about the new features of LTTng{nbsp}{revision}, see
144 https://lttng.org/blog/2015/10/14/lttng-2.7-released/[the release announcement].
150 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
151 generation_ is a modern toolkit for tracing Linux systems and
152 applications. So your first question might be:
159 As the history of software engineering progressed and led to what
160 we now take for granted--complex, numerous and
161 interdependent software applications running in parallel on
162 sophisticated operating systems like Linux--the authors of such
163 components, software developers, began feeling a natural
164 urge to have tools that would ensure the robustness and good performance
165 of their masterpieces.
167 One major achievement in this field is, inarguably, the
168 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
169 an essential tool for developers to find and fix bugs. But even the best
170 debugger won't help make your software run faster, and nowadays, faster
171 software means either more work done by the same hardware, or cheaper
172 hardware for the same work.
174 A _profiler_ is often the tool of choice to identify performance
175 bottlenecks. Profiling is suitable to identify _where_ performance is
176 lost in a given software. The profiler outputs a profile, a statistical
177 summary of observed events, which you may use to discover which
178 functions took the most time to execute. However, a profiler won't
179 report _why_ some identified functions are the bottleneck. Bottlenecks
180 might only occur when specific conditions are met, conditions that are
181 sometimes impossible to capture by a statistical profiler, or impossible
182 to reproduce with an application altered by the overhead of an
183 event-based profiler. For a thorough investigation of software
184 performance issues, a history of execution is essential, with the
185 recorded values of variables and context fields you choose, and
186 with as little influence as possible on the instrumented software. This
187 is where tracing comes in handy.
189 _Tracing_ is a technique used to understand what goes on in a running
190 software system. The software used for tracing is called a _tracer_,
191 which is conceptually similar to a tape recorder. When recording,
192 specific instrumentation points placed in the software source code
193 generate events that are saved on a giant tape: a _trace_ file. You
194 can trace user applications and the operating system at the same time,
195 opening the possibility of resolving a wide range of problems that would
196 otherwise be extremely challenging.
198 Tracing is often compared to _logging_. However, tracers and loggers are
199 two different tools, serving two different purposes. Tracers are
200 designed to record much lower-level events that occur much more
201 frequently than log messages, often in the range of thousands per
202 second, with very little execution overhead. Logging is more appropriate
203 for a very high-level analysis of less frequent events: user accesses,
204 exceptional conditions (errors and warnings, for example), database
205 transactions, instant messaging communications, and such. Simply put,
206 logging is one of the many use cases that can be satisfied with tracing.
208 The list of recorded events inside a trace file can be read manually
209 like a log file for the maximum level of detail, but it is generally
210 much more interesting to perform application-specific analyses to
211 produce reduced statistics and graphs that are useful to resolve a
212 given problem. Trace viewers and analyzers are specialized tools
215 In the end, this is what LTTng is: a powerful, open source set of
216 tools to trace the Linux kernel and user applications at the same time.
217 LTTng is composed of several components actively maintained and
218 developed by its link:/community/#where[community].
221 [[lttng-alternatives]]
222 === Alternatives to noch:{LTTng}
224 Excluding proprietary solutions, a few competing software tracers
227 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
228 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
229 user scripts and is responsible for loading code into the
230 Linux kernel for further execution and collecting the outputted data.
231 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
232 subsystem in the Linux kernel in which a virtual machine can execute
233 programs passed from the user space to the kernel. You can attach
234 such programs to tracepoints and KProbes thanks to a system call, and
235 they can output data to the user space when executed thanks to
236 different mechanisms (pipe, VM register values, and eBPF maps, to name
238 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
239 is the de facto function tracer of the Linux kernel. Its user
240 interface is a set of special files in sysfs.
241 * https://perf.wiki.kernel.org/[perf] is
242 a performance analyzing tool for Linux which supports hardware
243 performance counters, tracepoints, as well as other counters and
244 types of probes. perf's controlling utility is the cmd:perf command
246 * http://linux.die.net/man/1/strace[strace]
247 is a command-line utility which records system calls made by a
248 user process, as well as signal deliveries and changes of process
249 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
250 to fulfill its function.
251 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
252 analyze Linux kernel events. You write scripts, or _chisels_ in
253 sysdig's jargon, in Lua and sysdig executes them while the system is
254 being traced or afterwards. sysdig's interface is the cmd:sysdig
255 command-line tool as well as the curses-based cmd:csysdig tool.
256 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
257 user space tracer which uses custom user scripts to produce plain text
258 traces. SystemTap converts the scripts to the C language, and then
259 compiles them as Linux kernel modules which are loaded to produce
260 trace data. SystemTap's primary user interface is the cmd:stap
263 The main distinctive features of LTTng is that it produces correlated
264 kernel and user space traces, as well as doing so with the lowest
265 overhead amongst other solutions. It produces trace files in the
266 http://diamon.org/ctf[CTF] format, a file format optimized
267 for the production and analyses of multi-gigabyte data.
269 LTTng is the result of more than 10 years of active open source
270 development by a community of passionate developers.
271 LTTng{nbsp}{revision} is currently available on major desktop and server
274 The main interface for tracing control is a single command-line tool
275 named cmd:lttng. The latter can create several tracing sessions, enable
276 and disable events on the fly, filter events efficiently with custom
277 user expressions, start and stop tracing, and much more. LTTng can
278 record the traces on the file system or send them over the network, and
279 keep them totally or partially. You can view the traces once tracing
280 becomes inactive or in real-time.
282 <<installing-lttng,Install LTTng now>> and
283 <<getting-started,start tracing>>!
289 **LTTng** is a set of software <<plumbing,components>> which interact to
290 <<instrumenting,instrument>> the Linux kernel and user applications, and
291 to <<controlling-tracing,control tracing>> (start and stop
292 tracing, enable and disable event rules, and the rest). Those
293 components are bundled into the following packages:
295 * **LTTng-tools**: Libraries and command-line interface to
297 * **LTTng-modules**: Linux kernel modules to instrument and
299 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
300 trace user applications.
302 Most distributions mark the LTTng-modules and LTTng-UST packages as
303 optional when installing LTTng-tools (which is always required). In the
304 following sections, we always provide the steps to install all three,
307 * You only need to install LTTng-modules if you intend to trace the
309 * You only need to install LTTng-UST if you intend to trace user
313 .Availability of LTTng{nbsp}{revision} for major Linux distributions.
315 |Distribution |Available in releases |Alternatives
318 |<<ubuntu,Ubuntu{nbsp}16.04 _Xenial Xerus_>>
319 |LTTng{nbsp}2.8 for Ubuntu{nbsp}16.10 _Yakkety Yak_.
321 LTTng{nbsp}{revision} for Ubuntu{nbsp}12.04 _Precise Pangolin_,
322 Ubuntu{nbsp}14.04 _Trusty Tahr_, and Ubuntu{nbsp}16.04 _Xenial Xerus_:
323 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
325 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
326 other Ubuntu releases.
330 |LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} for
331 Fedora{nbsp}25 and Fedora{nbsp}26 (both are not released yet).
333 <<building-from-source,Build LTTng-modules{nbsp}{revision} from
336 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
337 other Fedora releases.
341 |LTTng{nbsp}2.8 for Debian "stretch" (testing).
343 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
344 other Debian releases.
347 |<<opensuse,openSUSE Leap{nbsp}42.1>>
348 |<<building-from-source,Build LTTng{nbsp}{revision} from source>> for
349 other openSUSE releases.
354 LTTng{nbsp}2.8 on the AUR.
356 <<building-from-source,Build LTTng{nbsp}{revision} from source>>.
360 |LTTng{nbsp}2.8 for Alpine Linux "edge".
362 LTTng{nbsp}2.8 for Alpine Linux{nbsp}3.5 (not released yet).
364 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
365 other Alpine Linux releases.
368 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
372 |<<"buildroot","Buildroot{nbsp}2016.02, Buildroot{nbsp}2016.05,
373 and Buildroot{nbsp}2016.08">>
374 |LTTng{nbsp}2.8 for Buildroot{nbsp}2016.11 (not released yet).
376 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
377 other Buildroot releases.
379 |OpenEmbedded and Yocto
380 |<<oe-yocto,Yocto Project{nbsp}2.1 _Krogoth_>> (`openembedded-core` layer)
381 |LTTng{nbsp}2.8 for Yocto Project{nbsp}2.2 _Morty_.
383 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
384 other Yocto releases.
389 === [[ubuntu-official-repositories]]Ubuntu
391 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}16.04 _Xenial Xerus_.
392 For previous releases of Ubuntu, <<ubuntu-ppa,use the LTTng
393 Stable{nbsp}{revision} PPA>>.
395 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}16.04 _Xenial Xerus_:
397 . Install the main LTTng{nbsp}{revision} packages:
402 sudo apt-get install lttng-tools
403 sudo apt-get install lttng-modules-dkms
404 sudo apt-get install liblttng-ust-dev
408 . **If you need to instrument and trace
409 <<java-application,Java applications>>**, install the LTTng-UST
415 sudo apt-get install liblttng-ust-agent-java
419 . **If you need to instrument and trace
420 <<python-application,Python{nbsp}3 applications>>**, install the
421 LTTng-UST Python agent:
426 sudo apt-get install python3-lttngust
432 ==== noch:{LTTng} Stable {revision} PPA
435 https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng Stable{nbsp}{revision} PPA]
436 offers the latest stable LTTng{nbsp}{revision} packages for:
438 * Ubuntu{nbsp}12.04 _Precise Pangolin_
439 * Ubuntu{nbsp}14.04 _Trusty Tahr_
440 * Ubuntu{nbsp}16.04 _Xenial Xerus_
442 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
445 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
451 sudo apt-add-repository ppa:lttng/stable-2.7
456 . Install the main LTTng{nbsp}{revision} packages:
461 sudo apt-get install lttng-tools
462 sudo apt-get install lttng-modules-dkms
463 sudo apt-get install liblttng-ust-dev
467 . **If you need to instrument and trace
468 <<java-application,Java applications>>**, install the LTTng-UST
474 sudo apt-get install liblttng-ust-agent-java
478 . **If you need to instrument and trace
479 <<python-application,Python{nbsp}3 applications>>**, install the
480 LTTng-UST Python agent:
485 sudo apt-get install python3-lttngust
491 === noch:{openSUSE}/RPM
493 To install LTTng{nbsp}{revision} on openSUSE Leap{nbsp}42.1:
495 * Install the main LTTng{nbsp}{revision} packages:
500 sudo zypper install lttng-tools
501 sudo zypper install lttng-modules
502 sudo zypper install lttng-ust-devel
507 .Java and Python application instrumentation and tracing
509 If you need to instrument and trace <<java-application,Java
510 applications>> on openSUSE, you need to build and install
511 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
512 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
513 `--enable-java-agent-all` options to the `configure` script, depending
514 on which Java logging framework you use.
516 If you need to instrument and trace <<python-application,Python
517 applications>> on openSUSE, you need to build and install
518 LTTng-UST{nbsp}{revision} from source and pass the
519 `--enable-python-agent` option to the `configure` script.
526 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2016.02,
527 Buildroot{nbsp}2016.05, or Buildroot{nbsp}2016.08:
529 . Launch the Buildroot configuration tool:
538 . In **Kernel**, check **Linux kernel**.
539 . In **Toolchain**, check **Enable WCHAR support**.
540 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
541 check **lttng-modules** and **lttng-tools**.
542 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
543 **Other**, check **lttng-libust**.
547 === OpenEmbedded and Yocto
549 LTTng{nbsp}{revision} recipes are available in the
550 http://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
551 layer for Yocto Project{nbsp}2.1 _Krogoth_ under the following names:
557 With BitBake, the simplest way to include LTTng recipes in your target
558 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
561 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
566 . Select a machine and an image recipe.
567 . Click **Edit image recipe**.
568 . Under the **All recipes** tab, search for **lttng**.
569 . Check the desired LTTng recipes.
572 .Java and Python application instrumentation and tracing
574 If you need to instrument and trace <<java-application,Java
575 applications>> on openSUSE, you need to build and install
576 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
577 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
578 `--enable-java-agent-all` options to the `configure` script, depending
579 on which Java logging framework you use.
581 If you need to instrument and trace <<python-application,Python
582 applications>> on openSUSE, you need to build and install
583 LTTng-UST{nbsp}{revision} from source and pass the
584 `--enable-python-agent` option to the `configure` script.
588 [[enterprise-distributions]]
589 === RHEL, SUSE, and other enterprise distributions
591 To install LTTng on enterprise Linux distributions, such as Red Hat
592 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
593 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
596 [[building-from-source]]
597 === Build from source
599 To build and install LTTng{nbsp}{revision} from source:
601 . Using your distribution's package manager, or from source, install
602 the following dependencies of LTTng-tools and LTTng-UST:
605 * https://sourceforge.net/projects/libuuid/[libuuid]
606 * http://directory.fsf.org/wiki/Popt[popt]
607 * http://liburcu.org/[Userspace RCU]
608 * http://www.xmlsoft.org/[libxml2]
611 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
617 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.7.tar.bz2 &&
618 tar -xf lttng-modules-latest-2.7.tar.bz2 &&
619 cd lttng-modules-2.7.* &&
621 sudo make modules_install &&
626 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
632 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.7.tar.bz2 &&
633 tar -xf lttng-ust-latest-2.7.tar.bz2 &&
634 cd lttng-ust-2.7.* &&
644 .Java and Python application tracing
646 If you need to instrument and trace <<java-application,Java
647 applications>>, pass the `--enable-java-agent-jul`,
648 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
649 `configure` script, depending on which Java logging framework you use.
651 If you need to instrument and trace <<python-application,Python
652 applications>>, pass the `--enable-python-agent` option to the
653 `configure` script. You can set the `PYTHON` environment variable to the
654 path to the Python interpreter for which to install the LTTng-UST Python
662 By default, LTTng-UST libraries are installed to
663 dir:{/usr/local/lib}, which is the de facto directory in which to
664 keep self-compiled and third-party libraries.
666 When <<building-tracepoint-providers-and-user-application,linking an
667 instrumented user application with `liblttng-ust`>>:
669 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
671 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
672 man:gcc(1), man:g++(1), or man:clang(1).
676 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
682 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
683 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
684 cd lttng-tools-2.7.* &&
692 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
693 previous steps automatically for a given version of LTTng and confine
694 the installed files in a specific directory. This can be useful to test
695 LTTng without installing it on your system.
701 This is a short guide to get started quickly with LTTng kernel and user
704 Before you follow this guide, make sure to <<installing-lttng,install>>
707 This tutorial walks you through the steps to:
709 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
710 . <<tracing-your-own-user-application,Trace a user application>> written
712 . <<viewing-and-analyzing-your-traces,View and analyze the
716 [[tracing-the-linux-kernel]]
717 === Trace the Linux kernel
719 The following command lines start with cmd:sudo because you need root
720 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
721 your Unix user is a member of the <<lttng-sessiond,tracing group>>.
723 . Create a <<tracing-session,tracing session>>:
728 sudo lttng create my-kernel-session
732 . List the available kernel tracepoints and system calls:
741 . Create an <<event,event rule>> which matches the desired event names,
742 for example `sched_switch` and `sched_process_fork`:
747 sudo lttng enable-event --kernel sched_switch,sched_process_fork
751 You can also create an event rule which _matches_ all the Linux kernel
752 tracepoints (this will generate a lot of data when tracing):
757 sudo lttng enable-event --kernel --all
770 . Do some operation on your system for a few seconds. For example,
771 load a website, or list the files of a directory.
772 . Stop tracing and destroy the tracing session:
782 The `destroy` command does not destroy the trace data; it only destroys
783 the state of the tracing session.
785 By default, LTTng saves the traces in
786 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
787 where +__name__+ is the tracing session name. Note that the
788 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
790 See <<viewing-and-analyzing-your-traces,View and analyze the
791 recorded events>> to view the recorded events.
794 [[tracing-your-own-user-application]]
795 === Trace a user application
797 This section steps you through a simple example to trace a
798 _Hello world_ program written in C.
800 To create the traceable user application:
802 . Create the tracepoint provider header file, which defines the
803 tracepoints and the events they can generate:
809 #undef TRACEPOINT_PROVIDER
810 #define TRACEPOINT_PROVIDER hello_world
812 #undef TRACEPOINT_INCLUDE
813 #define TRACEPOINT_INCLUDE "./hello-tp.h"
815 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
818 #include <lttng/tracepoint.h>
828 ctf_string(my_string_field, my_string_arg)
829 ctf_integer(int, my_integer_field, my_integer_arg)
833 #endif /* _HELLO_TP_H */
835 #include <lttng/tracepoint-event.h>
839 . Create the tracepoint provider package source file:
845 #define TRACEPOINT_CREATE_PROBES
846 #define TRACEPOINT_DEFINE
848 #include "hello-tp.h"
852 . Build the tracepoint provider package:
857 gcc -c -I. hello-tp.c
861 . Create the _Hello World_ application source file:
868 #include "hello-tp.h"
870 int main(int argc, char *argv[])
874 puts("Hello, World!\nPress Enter to continue...");
877 * The following getchar() call is only placed here for the purpose
878 * of this demonstration, to pause the application in order for
879 * you to have time to list its tracepoints. It is not
885 * A tracepoint() call.
887 * Arguments, as defined in hello-tp.h:
889 * 1. Tracepoint provider name (required)
890 * 2. Tracepoint name (required)
891 * 3. my_integer_arg (first user-defined argument)
892 * 4. my_string_arg (second user-defined argument)
894 * Notice the tracepoint provider and tracepoint names are
895 * NOT strings: they are in fact parts of variables that the
896 * macros in hello-tp.h create.
898 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
900 for (x = 0; x < argc; ++x) {
901 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
904 puts("Quitting now!");
905 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
912 . Build the application:
921 . Link the application with the tracepoint provider package,
922 `liblttng-ust`, and `libdl`:
927 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
931 Here's the whole build process:
934 .User space tracing tutorial's build steps.
935 image::ust-flow.png[]
937 To trace the user application:
939 . Run the application with a few arguments:
944 ./hello world and beyond
953 Press Enter to continue...
957 . Start an LTTng <<lttng-sessiond,session daemon>>:
962 lttng-sessiond --daemonize
966 Note that a session daemon might already be running, for example as
967 a service that the distribution's service manager started.
969 . List the available user space tracepoints:
974 lttng list --userspace
978 You see the `hello_world:my_first_tracepoint` tracepoint listed
979 under the `./hello` process.
981 . Create a <<tracing-session,tracing session>>:
986 lttng create my-user-space-session
990 . Create an <<event,event rule>> which matches the
991 `hello_world:my_first_tracepoint` event name:
996 lttng enable-event --userspace hello_world:my_first_tracepoint
1009 . Go back to the running `hello` application and press Enter. The
1010 program executes all `tracepoint()` instrumentation points and exits.
1011 . Stop tracing and destroy the tracing session:
1021 The `destroy` command does not destroy the trace data; it only destroys
1022 the state of the tracing session.
1024 By default, LTTng saves the traces in
1025 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1026 where +__name__+ is the tracing session name. Note that the
1027 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1029 See <<viewing-and-analyzing-your-traces,View and analyze the
1030 recorded events>> to view the recorded events.
1033 [[viewing-and-analyzing-your-traces]]
1034 === View and analyze the recorded events
1036 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1037 kernel>> and <<tracing-your-own-user-application,Trace a user
1038 application>> tutorials, you can inspect the recorded events.
1040 Many tools are available to read LTTng traces:
1042 * **cmd:babeltrace** is a command-line utility which converts trace
1043 formats; it supports the format that LTTng produces, CTF, as well as a
1044 basic text output which can be ++grep++ed. The cmd:babeltrace command
1045 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1046 * Babeltrace also includes
1047 **https://www.python.org/[Python] bindings** so
1048 that you can easily open and read an LTTng trace with your own script,
1049 benefiting from the power of Python.
1050 * http://tracecompass.org/[**Trace Compass**]
1051 is a graphical user interface for viewing and analyzing any type of
1052 logs or traces, including LTTng's.
1053 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1054 project which includes many high-level analyses of LTTng kernel
1055 traces, like scheduling statistics, interrupt frequency distribution,
1056 top CPU usage, and more.
1058 NOTE: This section assumes that the traces recorded during the previous
1059 tutorials were saved to their default location, in the
1060 dir:{$LTTNG_HOME/lttng-traces} directory. Note that the env:LTTNG_HOME
1061 environment variable defaults to `$HOME` if not set.
1064 [[viewing-and-analyzing-your-traces-bt]]
1065 ==== Use the cmd:babeltrace command-line tool
1067 The simplest way to list all the recorded events of a trace is to pass
1068 its path to cmd:babeltrace with no options:
1072 babeltrace ~/lttng-traces/my-user-space-session*
1075 cmd:babeltrace finds all traces recursively within the given path and
1076 prints all their events, merging them in chronological order.
1078 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1083 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_
1086 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1087 count the recorded events:
1091 babeltrace ~/lttng-traces/my-kernel-session* | grep sys_read | wc --lines
1095 [[viewing-and-analyzing-your-traces-bt-python]]
1096 ==== Use the Babeltrace Python bindings
1098 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1099 is useful to isolate events by simple matching using man:grep(1) and
1100 similar utilities. However, more elaborate filters, such as keeping only
1101 event records with a field value falling within a specific range, are
1102 not trivial to write using a shell. Moreover, reductions and even the
1103 most basic computations involving multiple event records are virtually
1104 impossible to implement.
1106 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1107 to read the event records of an LTTng trace sequentially and compute the
1108 desired information.
1110 The following script accepts an LTTng Linux kernel trace path as its
1111 first argument and prints the short names of the top 5 running processes
1112 on CPU 0 during the whole trace:
1117 from collections import Counter
1123 if len(sys.argv) != 2:
1124 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1125 print(msg, file=sys.stderr)
1128 # A trace collection contains one or more traces
1129 col = babeltrace.TraceCollection()
1131 # Add the trace provided by the user (LTTng traces always have
1133 if col.add_trace(sys.argv[1], 'ctf') is None:
1134 raise RuntimeError('Cannot add trace')
1136 # This counter dict contains execution times:
1138 # task command name -> total execution time (ns)
1139 exec_times = Counter()
1141 # This contains the last `sched_switch` timestamp
1145 for event in col.events:
1146 # Keep only `sched_switch` events
1147 if event.name != 'sched_switch':
1150 # Keep only events which happened on CPU 0
1151 if event['cpu_id'] != 0:
1155 cur_ts = event.timestamp
1161 # Previous task command (short) name
1162 prev_comm = event['prev_comm']
1164 # Initialize entry in our dict if not yet done
1165 if prev_comm not in exec_times:
1166 exec_times[prev_comm] = 0
1168 # Compute previous command execution time
1169 diff = cur_ts - last_ts
1171 # Update execution time of this command
1172 exec_times[prev_comm] += diff
1174 # Update last timestamp
1178 for name, ns in exec_times.most_common(5):
1180 print('{:20}{} s'.format(name, s))
1185 if __name__ == '__main__':
1186 sys.exit(0 if top5proc() else 1)
1193 python3 top5proc.py ~/lttng-traces/my-kernel-session*/kernel
1199 swapper/0 48.607245889 s
1200 chromium 7.192738188 s
1201 pavucontrol 0.709894415 s
1202 Compositor 0.660867933 s
1203 Xorg.bin 0.616753786 s
1206 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1207 weren't using the CPU that much when tracing, its first position in the
1212 == [[understanding-lttng]]Core concepts
1214 From a user's perspective, the LTTng system is built on a few concepts,
1215 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1216 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1217 Understanding how those objects relate to eachother is key in mastering
1220 The core concepts are:
1222 * <<tracing-session,Tracing session>>
1223 * <<domain,Tracing domain>>
1224 * <<channel,Channel and ring buffer>>
1225 * <<"event","Instrumentation point, event rule, event, and event record">>
1231 A _tracing session_ is a stateful dialogue between you and
1232 a <<lttng-sessiond,session daemon>>. You can
1233 <<creating-destroying-tracing-sessions,create a new tracing
1234 session>> with the `lttng create` command.
1236 Anything that you do when you control LTTng tracers happens within a
1237 tracing session. In particular, a tracing session:
1240 * Has its own set of trace files.
1241 * Has its own state of activity (started or stopped).
1242 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1244 * Has its own <<channel,channels>> which have their own
1245 <<event,event rules>>.
1248 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1249 image::concepts.png[]
1251 Those attributes and objects are completely isolated between different
1254 A tracing session is analogous to a cash machine session:
1255 the operations you do on the banking system through the cash machine do
1256 not alter the data of other users of the same system. In the case of
1257 the cash machine, a session lasts as long as your bank card is inside.
1258 In the case of LTTng, a tracing session lasts from the `lttng create`
1259 command to the `lttng destroy` command.
1262 .Each Unix user has its own set of tracing sessions.
1263 image::many-sessions.png[]
1266 [[tracing-session-mode]]
1267 ==== Tracing session mode
1269 LTTng can send the generated trace data to different locations. The
1270 _tracing session mode_ dictates where to send it. The following modes
1271 are available in LTTng{nbsp}{revision}:
1274 LTTng writes the traces to the file system of the machine being traced
1277 Network streaming mode::
1278 LTTng sends the traces over the network to a
1279 <<lttng-relayd,relay daemon>> running on a remote system.
1282 LTTng does not write the traces by default. Instead, you can request
1283 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1284 current tracing buffers, and to write it to the target's file system
1285 or to send it over the network to a <<lttng-relayd,relay daemon>>
1286 running on a remote system.
1289 This mode is similar to the network streaming mode, but a live
1290 trace viewer can connect to the distant relay daemon to
1291 <<lttng-live,view event records as LTTng generates them>> by
1298 A _tracing domain_ is a namespace for event sources. A tracing domain
1299 has its own properties and features.
1301 There are currently five available tracing domains:
1305 * `java.util.logging` (JUL)
1309 You must specify a tracing domain when using some commands to avoid
1310 ambiguity. For example, since all the domains support named tracepoints
1311 as event sources (instrumentation points that you manually insert in the
1312 source code), you need to specify a tracing domain when
1313 <<enabling-disabling-events,creating an event rule>> because all the
1314 tracing domains could have tracepoints with the same names.
1316 Some features are reserved to specific tracing domains. Dynamic function
1317 entry and return instrumentation points, for example, are currently only
1318 supported in the Linux kernel tracing domain, but support for other
1319 tracing domains could be added in the future.
1321 You can create <<channel,channels>> in the Linux kernel and user space
1322 tracing domains. The other tracing domains have a single default
1327 === Channel and ring buffer
1329 A _channel_ is an object which is responsible for a set of ring buffers.
1330 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1331 tracer emits an event, it can record it to one or more
1332 sub-buffers. The attributes of a channel determine what to do when
1333 there's no space left for a new event record because all sub-buffers
1334 are full, where to send a full sub-buffer, and other behaviours.
1336 A channel is always associated to a <<domain,tracing domain>>. The
1337 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1338 a default channel which you cannot configure.
1340 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1341 an event, it records it to the sub-buffers of all
1342 the enabled channels with a satisfied event rule, as long as those
1343 channels are part of active <<tracing-session,tracing sessions>>.
1346 [[channel-buffering-schemes]]
1347 ==== Per-user vs. per-process buffering schemes
1349 A channel has at least one ring buffer _per CPU_. LTTng always
1350 records an event to the ring buffer associated to the CPU on which it
1353 Two _buffering schemes_ are available when you
1354 <<enabling-disabling-channels,create a channel>> in the
1355 user space <<domain,tracing domain>>:
1357 Per-user buffering::
1358 Allocate one set of ring buffers--one per CPU--shared by all the
1359 instrumented processes of each Unix user.
1363 .Per-user buffering scheme.
1364 image::per-user-buffering.png[]
1367 Per-process buffering::
1368 Allocate one set of ring buffers--one per CPU--for each
1369 instrumented process.
1373 .Per-process buffering scheme.
1374 image::per-process-buffering.png[]
1377 The per-process buffering scheme tends to consume more memory than the
1378 per-user option because systems generally have more instrumented
1379 processes than Unix users running instrumented processes. However, the
1380 per-process buffering scheme ensures that one process having a high
1381 event throughput won't fill all the shared sub-buffers of the same
1384 The Linux kernel tracing domain has only one available buffering scheme
1385 which is to allocate a single set of ring buffers for the whole system.
1386 This scheme is similar to the per-user option, but with a single, global
1387 user "running" the kernel.
1390 [[channel-overwrite-mode-vs-discard-mode]]
1391 ==== Overwrite vs. discard event loss modes
1393 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1394 arc in the following animation) of a specific channel's ring buffer.
1395 When there's no space left in a sub-buffer, the tracer marks it as
1396 consumable (red) and another, empty sub-buffer starts receiving the
1397 following event records. A <<lttng-consumerd,consumer daemon>>
1398 eventually consumes the marked sub-buffer (returns to white).
1401 [role="docsvg-channel-subbuf-anim"]
1406 In an ideal world, sub-buffers are consumed faster than they are filled,
1407 as is the case in the previous animation. In the real world,
1408 however, all sub-buffers can be full at some point, leaving no space to
1409 record the following events.
1411 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1412 available, it is acceptable to lose event records when the alternative
1413 would be to cause substantial delays in the instrumented application's
1414 execution. LTTng privileges performance over integrity; it aims at
1415 perturbing the traced system as little as possible in order to make
1416 tracing of subtle race conditions and rare interrupt cascades possible.
1418 When it comes to losing event records because no empty sub-buffer is
1419 available, the channel's _event loss mode_ determines what to do. The
1420 available event loss modes are:
1423 Drop the newest event records until a the tracer
1424 releases a sub-buffer.
1427 Clear the sub-buffer containing the oldest event records and start
1428 writing the newest event records there.
1430 This mode is sometimes called _flight recorder mode_ because it's
1432 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1433 always keep a fixed amount of the latest data.
1435 Which mechanism you should choose depends on your context: prioritize
1436 the newest or the oldest event records in the ring buffer?
1438 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1439 as soon as a there's no space left for a new event record, whereas in
1440 discard mode, the tracer only discards the event record that doesn't
1443 In discard mode, LTTng increments a count of lost event records when
1444 an event record is lost and saves this count to the trace. In
1445 overwrite mode, LTTng keeps no information when it overwrites a
1446 sub-buffer before consuming it.
1448 There are a few ways to decrease your probability of losing event
1450 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1451 how you can fine-une the sub-buffer count and size of a channel to
1452 virtually stop losing event records, though at the cost of greater
1456 [[channel-subbuf-size-vs-subbuf-count]]
1457 ==== Sub-buffer count and size
1459 When you <<enabling-disabling-channels,create a channel>>, you can
1460 set its number of sub-buffers and their size.
1462 Note that there is noticeable CPU overhead introduced when
1463 switching sub-buffers (marking a full one as consumable and switching
1464 to an empty one for the following events to be recorded). Knowing this,
1465 the following list presents a few practical situations along with how
1466 to configure the sub-buffer count and size for them:
1468 * **High event throughput**: In general, prefer bigger sub-buffers to
1469 lower the risk of losing event records.
1471 Having bigger sub-buffers also ensures a lower sub-buffer switching
1474 The number of sub-buffers is only meaningful if you create the channel
1475 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1476 other sub-buffers are left unaltered.
1478 * **Low event throughput**: In general, prefer smaller sub-buffers
1479 since the risk of losing event records is low.
1481 Because events occur less frequently, the sub-buffer switching frequency
1482 should remain low and thus the tracer's overhead should not be a
1485 * **Low memory system**: If your target system has a low memory
1486 limit, prefer fewer first, then smaller sub-buffers.
1488 Even if the system is limited in memory, you want to keep the
1489 sub-buffers as big as possible to avoid a high sub-buffer switching
1492 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1493 which means event data is very compact. For example, the average
1494 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1495 sub-buffer size of 1{nbsp}MiB is considered big.
1497 The previous situations highlight the major trade-off between a few big
1498 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1499 frequency vs. how much data is lost in overwrite mode. Assuming a
1500 constant event throughput and using the overwrite mode, the two
1501 following configurations have the same ring buffer total size:
1504 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1509 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1510 switching frequency, but if a sub-buffer overwrite happens, half of
1511 the event records so far (4{nbsp}MiB) are definitely lost.
1512 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1513 overhead as the previous configuration, but if a sub-buffer
1514 overwrite happens, only the eighth of event records so far are
1517 In discard mode, the sub-buffers count parameter is pointless: use two
1518 sub-buffers and set their size according to the requirements of your
1522 [[channel-switch-timer]]
1523 ==== Switch timer period
1525 The _switch timer period_ is an important configurable attribute of
1526 a channel to ensure periodic sub-buffer flushing.
1528 When the _switch timer_ expires, a sub-buffer switch happens. You can
1529 set the switch timer period attribute when you
1530 <<enabling-disabling-channels,create a channel>> to ensure that event
1531 data is consumed and committed to trace files or to a distant relay
1532 daemon periodically in case of a low event throughput.
1535 [role="docsvg-channel-switch-timer"]
1540 This attribute is also convenient when you use big sub-buffers to cope
1541 with a sporadic high event throughput, even if the throughput is
1545 [[channel-read-timer]]
1546 ==== Read timer period
1548 By default, the LTTng tracers use a notification mechanism to signal a
1549 full sub-buffer so that a consumer daemon can consume it. When such
1550 notifications must be avoided, for example in real-time applications,
1551 you can use the channel's _read timer_ instead. When the read timer
1552 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1553 consumable sub-buffers.
1556 [[tracefile-rotation]]
1557 ==== Trace file count and size
1559 By default, trace files can grow as large as needed. You can set the
1560 maximum size of each trace file that a channel writes when you
1561 <<enabling-disabling-channels,create a channel>>. When the size of
1562 a trace file reaches the channel's fixed maximum size, LTTng creates
1563 another file to contain the next event records. LTTng appends a file
1564 count to each trace file name in this case.
1566 If you set the trace file size attribute when you create a channel, the
1567 maximum number of trace files that LTTng creates is _unlimited_ by
1568 default. To limit them, you can also set a maximum number of trace
1569 files. When the number of trace files reaches the channel's fixed
1570 maximum count, the oldest trace file is overwritten. This mechanism is
1571 called _trace file rotation_.
1575 === Instrumentation point, event rule, event, and event record
1577 An _event rule_ is a set of conditions which must be **all** satisfied
1578 for LTTng to record an occuring event.
1580 You set the conditions when you <<enabling-disabling-events,create
1583 You always attach an event rule to <<channel,channel>> when you create
1586 When an event passes the conditions of an event rule, LTTng records it
1587 in one of the attached channel's sub-buffers.
1589 The available conditions, as of LTTng{nbsp}{revision}, are:
1591 * The event rule _is enabled_.
1592 * The instrumentation point's type _is{nbsp}T_.
1593 * The instrumentation point's name (sometimes called _event name_)
1594 _matches{nbsp}N_, but _is not{nbsp}E_.
1595 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1596 _is exactly{nbsp}L_.
1597 * The fields of the event's payload _satisfy_ a filter
1598 expression{nbsp}__F__.
1600 As you can see, all the conditions but the dynamic filter are related to
1601 the event rule's status or to the instrumentation point, not to the
1602 occurring events. This is why, without a filter, checking if an event
1603 passes an event rule is not a dynamic task: when you create or modify an
1604 event rule, all the tracers of its tracing domain enable or disable the
1605 instrumentation points themselves once. This is possible because the
1606 attributes of an instrumentation point (type, name, and log level) are
1607 defined statically. In other words, without a dynamic filter, the tracer
1608 _does not evaluate_ the arguments of an instrumentation point unless it
1609 matches an enabled event rule.
1611 Note that, for LTTng to record an event, the <<channel,channel>> to
1612 which a matching event rule is attached must also be enabled, and the
1613 tracing session owning this channel must be active.
1616 .Logical path from an instrumentation point to an event record.
1617 image::event-rule.png[]
1619 .Event, event record, or event rule?
1621 With so many similar terms, it's easy to get confused.
1623 An **event** is the consequence of the execution of an _instrumentation
1624 point_, like a tracepoint that you manually place in some source code,
1625 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1626 time. Different actions can be taken upon the occurance of an event,
1627 like record the event's payload to a buffer.
1629 An **event record** is the representation of an event in a sub-buffer. A
1630 tracer is responsible for capturing the payload of an event, current
1631 context variables, the event's ID, and the event's timestamp. LTTng
1632 can append this sub-buffer to a trace file.
1634 An **event rule** is a set of conditions which must all be satisfied for
1635 LTTng to record an occuring event. Events still occur without
1636 satisfying event rules, but LTTng does not record them.
1641 == Components of noch:{LTTng}
1643 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1644 to call LTTng a simple _tool_ since it is composed of multiple
1645 interacting components. This section describes those components,
1646 explains their respective roles, and shows how they connect together to
1647 form the LTTng ecosystem.
1649 The following diagram shows how the most important components of LTTng
1650 interact with user applications, the Linux kernel, and you:
1653 .Control and trace data paths between LTTng components.
1654 image::plumbing.png[]
1656 The LTTng project incorporates:
1658 * **LTTng-tools**: Libraries and command-line interface to
1659 control tracing sessions.
1660 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1661 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1662 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1663 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1664 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1665 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1667 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1668 headers to instrument and trace any native user application.
1669 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1670 *** `liblttng-ust-libc-wrapper`
1671 *** `liblttng-ust-pthread-wrapper`
1672 *** `liblttng-ust-cyg-profile`
1673 *** `liblttng-ust-cyg-profile-fast`
1674 *** `liblttng-ust-dl`
1675 ** User space tracepoint provider source files generator command-line
1676 tool (man:lttng-gen-tp(1)).
1677 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1678 Java applications using `java.util.logging` or
1679 Apache log4j 1.2 logging.
1680 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1681 Python applications using the standard `logging` package.
1682 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1684 ** LTTng kernel tracer module.
1685 ** Tracing ring buffer kernel modules.
1686 ** Probe kernel modules.
1687 ** LTTng logger kernel module.
1691 === Tracing control command-line interface
1694 .The tracing control command-line interface.
1695 image::plumbing-lttng-cli.png[]
1697 The _man:lttng(1) command-line tool_ is the standard user interface to
1698 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1699 is part of LTTng-tools.
1701 The cmd:lttng tool is linked with
1702 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1703 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1705 The cmd:lttng tool has a Git-like interface:
1709 lttng <general options> <command> <command options>
1712 The <<controlling-tracing,Tracing control>> section explores the
1713 available features of LTTng using the cmd:lttng tool.
1716 [[liblttng-ctl-lttng]]
1717 === Tracing control library
1720 .The tracing control library.
1721 image::plumbing-liblttng-ctl.png[]
1723 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1724 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1725 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1727 The <<lttng-cli,cmd:lttng command-line tool>>
1728 is linked with `liblttng-ctl`.
1730 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1735 #include <lttng/lttng.h>
1738 Some objects are referenced by name (C string), such as tracing
1739 sessions, but most of them require to create a handle first using
1740 `lttng_create_handle()`.
1742 The best available developer documentation for `liblttng-ctl` is, as of
1743 LTTng{nbsp}{revision}, its installed header files. Every function and
1744 structure is thoroughly documented.
1748 === User space tracing library
1751 .The user space tracing library.
1752 image::plumbing-liblttng-ust.png[]
1754 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1755 is the LTTng user space tracer. It receives commands from a
1756 <<lttng-sessiond,session daemon>>, for example to
1757 enable and disable specific instrumentation points, and writes event
1758 records to ring buffers shared with a
1759 <<lttng-consumerd,consumer daemon>>.
1760 `liblttng-ust` is part of LTTng-UST.
1762 Public C header files are installed beside `liblttng-ust` to
1763 instrument any <<c-application,C or $$C++$$ application>>.
1765 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1766 packages, use their own library providing tracepoints which is
1767 linked with `liblttng-ust`.
1769 An application or library does not have to initialize `liblttng-ust`
1770 manually: its constructor does the necessary tasks to properly register
1771 to a session daemon. The initialization phase also enables the
1772 instrumentation points matching the <<event,event rules>> that you
1776 [[lttng-ust-agents]]
1777 === User space tracing agents
1780 .The user space tracing agents.
1781 image::plumbing-lttng-ust-agents.png[]
1783 The _LTTng-UST Java and Python agents_ are regular Java and Python
1784 packages which add LTTng tracing capabilities to the
1785 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1787 In the case of Java, the
1788 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1789 core logging facilities] and
1790 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1791 Note that Apache Log4{nbsp}2 is not supported.
1793 In the case of Python, the standard
1794 https://docs.python.org/3/library/logging.html[`logging`] package
1795 is supported. Both Python 2 and Python 3 modules can import the
1796 LTTng-UST Python agent package.
1798 The applications using the LTTng-UST agents are in the
1799 `java.util.logging` (JUL),
1800 log4j, and Python <<domain,tracing domains>>.
1802 Both agents use the same mechanism to trace the log statements. When an
1803 agent is initialized, it creates a log handler that attaches to the root
1804 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1805 When the application executes a log statement, it is passed to the
1806 agent's log handler by the root logger. The agent's log handler calls a
1807 native function in a tracepoint provider package shared library linked
1808 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1809 other fields, like its logger name and its log level. This native
1810 function contains a user space instrumentation point, hence tracing the
1813 The log level condition of an
1814 <<event,event rule>> is considered when tracing
1815 a Java or a Python application, and it's compatible with the standard
1816 JUL, log4j, and Python log levels.
1820 === LTTng kernel modules
1823 .The LTTng kernel modules.
1824 image::plumbing-lttng-modules.png[]
1826 The _LTTng kernel modules_ are a set of Linux kernel modules
1827 which implement the kernel tracer of the LTTng project. The LTTng
1828 kernel modules are part of LTTng-modules.
1830 The LTTng kernel modules include:
1832 * A set of _probe_ modules.
1834 Each module attaches to a specific subsystem
1835 of the Linux kernel using its tracepoint instrument points. There are
1836 also modules to attach to the entry and return points of the Linux
1837 system call functions.
1839 * _Ring buffer_ modules.
1841 A ring buffer implementation is provided as kernel modules. The LTTng
1842 kernel tracer writes to the ring buffer; a
1843 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1845 * The _LTTng kernel tracer_ module.
1846 * The _LTTng logger_ module.
1848 The LTTng logger module implements the special path:{/proc/lttng-logger}
1849 file so that any executable can generate LTTng events by opening and
1850 writing to this file.
1852 See <<proc-lttng-logger-abi,LTTng logger>>.
1854 Generally, you do not have to load the LTTng kernel modules manually
1855 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1856 daemon>> loads the necessary modules when starting. If you have extra
1857 probe modules, you can specify to load them to the session daemon on
1860 The LTTng kernel modules are installed in
1861 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1862 the kernel release (see `uname --kernel-release`).
1869 .The session daemon.
1870 image::plumbing-sessiond.png[]
1872 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1873 managing tracing sessions and for controlling the various components of
1874 LTTng. The session daemon is part of LTTng-tools.
1876 The session daemon sends control requests to and receives control
1879 * The <<lttng-ust,user space tracing library>>.
1881 Any instance of the user space tracing library first registers to
1882 a session daemon. Then, the session daemon can send requests to
1883 this instance, such as:
1886 ** Get the list of tracepoints.
1887 ** Share an <<event,event rule>> so that the user space tracing library
1888 can enable or disable tracepoints. Amongst the possible conditions
1889 of an event rule is a filter expression which `liblttng-ust` evalutes
1890 when an event occurs.
1891 ** Share <<channel,channel>> attributes and ring buffer locations.
1894 The session daemon and the user space tracing library use a Unix
1895 domain socket for their communication.
1897 * The <<lttng-ust-agents,user space tracing agents>>.
1899 Any instance of a user space tracing agent first registers to
1900 a session daemon. Then, the session daemon can send requests to
1901 this instance, such as:
1904 ** Get the list of loggers.
1905 ** Enable or disable a specific logger.
1908 The session daemon and the user space tracing agent use a TCP connection
1909 for their communication.
1911 * The <<lttng-modules,LTTng kernel tracer>>.
1912 * The <<lttng-consumerd,consumer daemon>>.
1914 The session daemon sends requests to the consumer daemon to instruct
1915 it where to send the trace data streams, amongst other information.
1917 * The <<lttng-relayd,relay daemon>>.
1919 The session daemon receives commands from the
1920 <<liblttng-ctl-lttng,tracing control library>>.
1922 The root session daemon loads the appropriate
1923 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1924 a <<lttng-consumerd,consumer daemon>> as soon as you create
1925 an <<event,event rule>>.
1927 The session daemon does not send and receive trace data: this is the
1928 role of the <<lttng-consumerd,consumer daemon>> and
1929 <<lttng-relayd,relay daemon>>. It does, however, generate the
1930 http://diamon.org/ctf/[CTF] metadata stream.
1932 Each Unix user can have its own session daemon instance. The
1933 tracing sessions managed by different session daemons are completely
1936 The root user's session daemon is the only one which is
1937 allowed to control the LTTng kernel tracer, and its spawned consumer
1938 daemon is the only one which is allowed to consume trace data from the
1939 LTTng kernel tracer. Note, however, that any Unix user which is a member
1940 of the <<tracing-group,tracing group>> is allowed
1941 to create <<channel,channels>> in the
1942 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1945 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1946 session daemon when using its `create` command if none is currently
1947 running. You can also start the session daemon manually.
1954 .The consumer daemon.
1955 image::plumbing-consumerd.png[]
1957 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
1958 ring buffers with user applications or with the LTTng kernel modules to
1959 collect trace data and send it to some location (on disk or to a
1960 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1961 is part of LTTng-tools.
1963 You do not start a consumer daemon manually: a consumer daemon is always
1964 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1965 <<event,event rule>>, that is, before you start tracing. When you kill
1966 its owner session daemon, the consumer daemon also exits because it is
1967 the session daemon's child process. Command-line options of
1968 man:lttng-sessiond(8) target the consumer daemon process.
1970 There are up to two running consumer daemons per Unix user, whereas only
1971 one session daemon can run per user. This is because each process can be
1972 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1973 and 64-bit processes, it is more efficient to have separate
1974 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1975 exception: it can have up to _three_ running consumer daemons: 32-bit
1976 and 64-bit instances for its user applications, and one more
1977 reserved for collecting kernel trace data.
1985 image::plumbing-relayd.png[]
1987 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1988 between remote session and consumer daemons, local trace files, and a
1989 remote live trace viewer. The relay daemon is part of LTTng-tools.
1991 The main purpose of the relay daemon is to implement a receiver of
1992 <<sending-trace-data-over-the-network,trace data over the network>>.
1993 This is useful when the target system does not have much file system
1994 space to record trace files locally.
1996 The relay daemon is also a server to which a
1997 <<lttng-live,live trace viewer>> can
1998 connect. The live trace viewer sends requests to the relay daemon to
1999 receive trace data as the target system emits events. The
2000 communication protocol is named _LTTng live_; it is used over TCP
2003 Note that you can start the relay daemon on the target system directly.
2004 This is the setup of choice when the use case is to view events as
2005 the target system emits them without the need of a remote system.
2009 == [[using-lttng]]Instrumentation
2011 There are many examples of tracing and monitoring in our everyday life:
2013 * You have access to real-time and historical weather reports and
2014 forecasts thanks to weather stations installed around the country.
2015 * You know your heart is safe thanks to an electrocardiogram.
2016 * You make sure not to drive your car too fast and to have enough fuel
2017 to reach your destination thanks to gauges visible on your dashboard.
2019 All the previous examples have something in common: they rely on
2020 **instruments**. Without the electrodes attached to the surface of your
2021 body's skin, cardiac monitoring is futile.
2023 LTTng, as a tracer, is no different from those real life examples. If
2024 you're about to trace a software system or, in other words, record its
2025 history of execution, you better have **instrumentation points** in the
2026 subject you're tracing, that is, the actual software.
2028 Various ways were developed to instrument a piece of software for LTTng
2029 tracing. The most straightforward one is to manually place
2030 instrumentation points, called _tracepoints_, in the software's source
2031 code. It is also possible to add instrumentation points dynamically in
2032 the Linux kernel <<domain,tracing domain>>.
2034 If you're only interested in tracing the Linux kernel, your
2035 instrumentation needs are probably already covered by LTTng's built-in
2036 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2037 user application which is already instrumented for LTTng tracing.
2038 In such cases, you can skip this whole section and read the topics of
2039 the <<controlling-tracing,Tracing control>> section.
2041 Many methods are available to instrument a piece of software for LTTng
2044 * <<c-application,User space instrumentation for C and $$C++$$
2046 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2047 * <<java-application,User space Java agent>>.
2048 * <<python-application,User space Python agent>>.
2049 * <<proc-lttng-logger-abi,LTTng logger>>.
2050 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2054 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2056 The procedure to instrument a C or $$C++$$ user application with
2057 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2059 . <<tracepoint-provider,Create the source files of a tracepoint provider
2061 . <<probing-the-application-source-code,Add tracepoints to
2062 the application's source code>>.
2063 . <<building-tracepoint-providers-and-user-application,Build and link
2064 a tracepoint provider package and the user application>>.
2066 If you need quick, man:printf(3)-like instrumentation, you can skip
2067 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2070 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2071 instrument a user application with `liblttng-ust`.
2074 [[tracepoint-provider]]
2075 ==== Create the source files of a tracepoint provider package
2077 A _tracepoint provider_ is a set of compiled functions which provide
2078 **tracepoints** to an application, the type of instrumentation point
2079 supported by LTTng-UST. Those functions can emit events with
2080 user-defined fields and serialize those events as event records to one
2081 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2082 macro, which you <<probing-the-application-source-code,insert in a user
2083 application's source code>>, calls those functions.
2085 A _tracepoint provider package_ is an object file (`.o`) or a shared
2086 library (`.so`) which contains one or more tracepoint providers.
2087 Its source files are:
2089 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2090 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2092 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2093 the LTTng user space tracer, at run time.
2096 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2097 image::ust-app.png[]
2099 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2100 skip creating and using a tracepoint provider and use
2101 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2105 ===== Create a tracepoint provider header file template
2107 A _tracepoint provider header file_ contains the tracepoint
2108 definitions of a tracepoint provider.
2110 To create a tracepoint provider header file:
2112 . Start from this template:
2116 .Tracepoint provider header file template (`.h` file extension).
2118 #undef TRACEPOINT_PROVIDER
2119 #define TRACEPOINT_PROVIDER provider_name
2121 #undef TRACEPOINT_INCLUDE
2122 #define TRACEPOINT_INCLUDE "./tp.h"
2124 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2127 #include <lttng/tracepoint.h>
2130 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2131 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2136 #include <lttng/tracepoint-event.h>
2142 * `provider_name` with the name of your tracepoint provider.
2143 * `"tp.h"` with the name of your tracepoint provider header file.
2145 . Below the `#include <lttng/tracepoint.h>` line, put your
2146 <<defining-tracepoints,tracepoint definitions>>.
2148 Your tracepoint provider name must be unique amongst all the possible
2149 tracepoint provider names used on the same target system. We
2150 suggest to include the name of your project or company in the name,
2151 for example, `org_lttng_my_project_tpp`.
2153 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2154 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2155 write are the <<defining-tracepoints,tracepoint definitions>>.
2158 [[defining-tracepoints]]
2159 ===== Create a tracepoint definition
2161 A _tracepoint definition_ defines, for a given tracepoint:
2163 * Its **input arguments**. They are the macro parameters that the
2164 `tracepoint()` macro accepts for this particular tracepoint
2165 in the user application's source code.
2166 * Its **output event fields**. They are the sources of event fields
2167 that form the payload of any event that the execution of the
2168 `tracepoint()` macro emits for this particular tracepoint.
2170 You can create a tracepoint definition by using the
2171 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2173 <<tpp-header,tracepoint provider header file template>>.
2175 The syntax of the `TRACEPOINT_EVENT()` macro is:
2178 .`TRACEPOINT_EVENT()` macro syntax.
2181 /* Tracepoint provider name */
2184 /* Tracepoint name */
2187 /* Input arguments */
2192 /* Output event fields */
2201 * `provider_name` with your tracepoint provider name.
2202 * `tracepoint_name` with your tracepoint name.
2203 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2204 * `fields` with the <<tpp-def-output-fields,output event field>>
2207 This tracepoint emits events named `provider_name:tracepoint_name`.
2210 .Event name's length limitation
2212 The concatenation of the tracepoint provider name and the
2213 tracepoint name must not exceed **254 characters**. If it does, the
2214 instrumented application compiles and runs, but LTTng throws multiple
2215 warnings and you could experience serious issues.
2218 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2221 .`TP_ARGS()` macro syntax.
2230 * `type` with the C type of the argument.
2231 * `arg_name` with the argument name.
2233 You can repeat `type` and `arg_name` up to 10 times to have
2234 more than one argument.
2236 .`TP_ARGS()` usage with three arguments.
2248 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2249 tracepoint definition with no input arguments.
2251 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2252 `ctf_*()` macros. Each `ctf_*()` macro defines one event field.
2253 See <<liblttng-ust-tp-fields,Tracepoint fields macros>> for a
2254 complete description of the available `ctf_*()` macros.
2255 A `ctf_*()` macro specifies the type, size, and byte order of
2258 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2259 C expression that the tracer evalutes at the `tracepoint()` macro site
2260 in the application's source code. This expression provides a field's
2261 source of data. The argument expression can include input argument names
2262 listed in the `TP_ARGS()` macro.
2264 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2265 must be unique within a given tracepoint definition.
2267 Here's a complete tracepoint definition example:
2269 .Tracepoint definition.
2271 The following tracepoint definition defines a tracepoint which takes
2272 three input arguments and has four output event fields.
2276 #include "my-custom-structure.h"
2282 const struct my_custom_structure*, my_custom_structure,
2287 ctf_string(query_field, query)
2288 ctf_float(double, ratio_field, ratio)
2289 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2290 ctf_integer(int, send_size, my_custom_structure->send_size)
2295 You can refer to this tracepoint definition with the `tracepoint()`
2296 macro in your application's source code like this:
2300 tracepoint(my_provider, my_tracepoint,
2301 my_structure, some_ratio, the_query);
2305 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2306 if they satisfy an enabled <<event,event rule>>.
2309 [[using-tracepoint-classes]]
2310 ===== Use a tracepoint class
2312 A _tracepoint class_ is a class of tracepoints which share the same
2313 output event field definitions. A _tracepoint instance_ is one
2314 instance of such a defined tracepoint class, with its own tracepoint
2317 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2318 shorthand which defines both a tracepoint class and a tracepoint
2319 instance at the same time.
2321 When you build a tracepoint provider package, the C or $$C++$$ compiler
2322 creates one serialization function for each **tracepoint class**. A
2323 serialization function is responsible for serializing the event fields
2324 of a tracepoint to a sub-buffer when tracing.
2326 For various performance reasons, when your situation requires multiple
2327 tracepoint definitions with different names, but with the same event
2328 fields, we recommend that you manually create a tracepoint class
2329 and instantiate as many tracepoint instances as needed. One positive
2330 effect of such a design, amongst other advantages, is that all
2331 tracepoint instances of the same tracepoint class reuse the same
2332 serialization function, thus reducing
2333 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2335 .Use a tracepoint class and tracepoint instances.
2337 Consider the following three tracepoint definitions:
2349 ctf_integer(int, userid, userid)
2350 ctf_integer(size_t, len, len)
2362 ctf_integer(int, userid, userid)
2363 ctf_integer(size_t, len, len)
2375 ctf_integer(int, userid, userid)
2376 ctf_integer(size_t, len, len)
2381 In this case, we create three tracepoint classes, with one implicit
2382 tracepoint instance for each of them: `get_account`, `get_settings`, and
2383 `get_transaction`. However, they all share the same event field names
2384 and types. Hence three identical, yet independent serialization
2385 functions are created when you build the tracepoint provider package.
2387 A better design choice is to define a single tracepoint class and three
2388 tracepoint instances:
2392 /* The tracepoint class */
2393 TRACEPOINT_EVENT_CLASS(
2394 /* Tracepoint provider name */
2397 /* Tracepoint class name */
2400 /* Input arguments */
2406 /* Output event fields */
2408 ctf_integer(int, userid, userid)
2409 ctf_integer(size_t, len, len)
2413 /* The tracepoint instances */
2414 TRACEPOINT_EVENT_INSTANCE(
2415 /* Tracepoint provider name */
2418 /* Tracepoint class name */
2421 /* Tracepoint name */
2424 /* Input arguments */
2430 TRACEPOINT_EVENT_INSTANCE(
2439 TRACEPOINT_EVENT_INSTANCE(
2452 [[assigning-log-levels]]
2453 ===== Assign a log level to a tracepoint definition
2455 You can assign an optional _log level_ to a
2456 <<defining-tracepoints,tracepoint definition>>.
2458 Assigning different levels of severity to tracepoint definitions can
2459 be useful: when you <<enabling-disabling-events,create an event rule>>,
2460 you can target tracepoints having a log level as severe as a specific
2463 The concept of LTTng-UST log levels is similar to the levels found
2464 in typical logging frameworks:
2466 * In a logging framework, the log level is given by the function
2467 or method name you use at the log statement site: `debug()`,
2468 `info()`, `warn()`, `error()`, and so on.
2469 * In LTTng-UST, you statically assign the log level to a tracepoint
2470 definition; any `tracepoint()` macro invocation which refers to
2471 this definition has this log level.
2473 You can assign a log level to a tracepoint definition with the
2474 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2475 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2476 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2479 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2482 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2484 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2489 * `provider_name` with the tracepoint provider name.
2490 * `tracepoint_name` with the tracepoint name.
2491 * `log_level` with the log level to assign to the tracepoint
2492 definition named `tracepoint_name` in the `provider_name`
2493 tracepoint provider.
2495 See <<liblttng-ust-tracepoint-loglevel,Tracepoint log levels>> for
2496 a list of available log level names.
2498 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2502 /* Tracepoint definition */
2511 ctf_integer(int, userid, userid)
2512 ctf_integer(size_t, len, len)
2516 /* Log level assignment */
2517 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2523 ===== Create a tracepoint provider package source file
2525 A _tracepoint provider package source file_ is a C source file which
2526 includes a <<tpp-header,tracepoint provider header file>> to expand its
2527 macros into event serialization and other functions.
2529 You can always use the following tracepoint provider package source
2533 .Tracepoint provider package source file template.
2535 #define TRACEPOINT_CREATE_PROBES
2540 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2541 header file>> name. You may also include more than one tracepoint
2542 provider header file here to create a tracepoint provider package
2543 holding more than one tracepoint providers.
2546 [[probing-the-application-source-code]]
2547 ==== Add tracepoints to an application's source code
2549 Once you <<tpp-header,create a tracepoint provider header file>>, you
2550 can use the `tracepoint()` macro in your application's
2551 source code to insert the tracepoints that this header
2552 <<defining-tracepoints,defines>>.
2554 The `tracepoint()` macro takes at least two parameters: the tracepoint
2555 provider name and the tracepoint name. The corresponding tracepoint
2556 definition defines the other parameters.
2558 .`tracepoint()` usage.
2560 The following <<defining-tracepoints,tracepoint definition>> defines a
2561 tracepoint which takes two input arguments and has two output event
2565 .Tracepoint provider header file.
2567 #include "my-custom-structure.h"
2574 const char*, cmd_name
2577 ctf_string(cmd_name, cmd_name)
2578 ctf_integer(int, number_of_args, argc)
2583 You can refer to this tracepoint definition with the `tracepoint()`
2584 macro in your application's source code like this:
2587 .Application's source file.
2591 int main(int argc, char* argv[])
2593 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2599 Note how the application's source code includes
2600 the tracepoint provider header file containing the tracepoint
2601 definitions to use, path:{tp.h}.
2604 .`tracepoint()` usage with a complex tracepoint definition.
2606 Consider this complex tracepoint definition, where multiple event
2607 fields refer to the same input arguments in their argument expression
2611 .Tracepoint provider header file.
2613 /* For `struct stat` */
2614 #include <sys/types.h>
2615 #include <sys/stat.h>
2627 ctf_integer(int, my_constant_field, 23 + 17)
2628 ctf_integer(int, my_int_arg_field, my_int_arg)
2629 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2630 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2631 my_str_arg[2] + my_str_arg[3])
2632 ctf_string(my_str_arg_field, my_str_arg)
2633 ctf_integer_hex(off_t, size_field, st->st_size)
2634 ctf_float(double, size_dbl_field, (double) st->st_size)
2635 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2636 size_t, strlen(my_str_arg) / 2)
2641 You can refer to this tracepoint definition with the `tracepoint()`
2642 macro in your application's source code like this:
2645 .Application's source file.
2647 #define TRACEPOINT_DEFINE
2654 stat("/etc/fstab", &s);
2655 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2661 If you look at the event record that LTTng writes when tracing this
2662 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2663 it should look like this:
2665 .Event record fields
2667 |Field's name |Field's value
2668 |`my_constant_field` |40
2669 |`my_int_arg_field` |23
2670 |`my_int_arg_field2` |529
2672 |`my_str_arg_field` |`Hello, World!`
2673 |`size_field` |0x12d
2674 |`size_dbl_field` |301.0
2675 |`half_my_str_arg_field` |`Hello,`
2679 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2680 compute--they use the call stack, for example. To avoid this
2681 computation when the tracepoint is disabled, you can use the
2682 `tracepoint_enabled()` and `do_tracepoint()` macros.
2684 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2688 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2690 tracepoint_enabled(provider_name, tracepoint_name)
2691 do_tracepoint(provider_name, tracepoint_name, ...)
2696 * `provider_name` with the tracepoint provider name.
2697 * `tracepoint_name` with the tracepoint name.
2699 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2700 `tracepoint_name` from the provider named `provider_name` is enabled
2703 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2704 if the tracepoint is enabled. Using `tracepoint()` with
2705 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2706 the `tracepoint_enabled()` check, thus a race condition is
2707 possible in this situation:
2710 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2712 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2713 stuff = prepare_stuff();
2716 tracepoint(my_provider, my_tracepoint, stuff);
2719 If the tracepoint is enabled after the condition, then `stuff` is not
2720 prepared: the emitted event will either contain wrong data, or the whole
2721 application could crash (segmentation fault, for example).
2723 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2724 `STAP_PROBEV()` call. If you need it, you must emit
2728 [[building-tracepoint-providers-and-user-application]]
2729 ==== Build and link a tracepoint provider package and an application
2731 Once you have one or more <<tpp-header,tracepoint provider header
2732 files>> and a <<tpp-source,tracepoint provider package source file>>,
2733 you can create the tracepoint provider package by compiling its source
2734 file. From here, multiple build and run scenarios are possible. The
2735 following table shows common application and library configurations
2736 along with the required command lines to achieve them.
2738 In the following diagrams, we use the following file names:
2741 Executable application.
2744 Application's object file.
2747 Tracepoint provider package object file.
2750 Tracepoint provider package archive file.
2753 Tracepoint provider package shared object file.
2756 User library object file.
2759 User library shared object file.
2761 We use the following symbols in the diagrams of table below:
2764 .Symbols used in the build scenario diagrams.
2765 image::ust-sit-symbols.png[]
2767 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2768 variable in the following instructions.
2770 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2771 .Common tracepoint provider package scenarios.
2773 |Scenario |Instructions
2776 The instrumented application is statically linked with
2777 the tracepoint provider package object.
2779 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2782 include::../common/ust-sit-step-tp-o.txt[]
2784 To build the instrumented application:
2786 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2791 #define TRACEPOINT_DEFINE
2795 . Compile the application source file:
2804 . Build the application:
2809 gcc -o app app.o tpp.o -llttng-ust -ldl
2813 To run the instrumented application:
2815 * Start the application:
2825 The instrumented application is statically linked with the
2826 tracepoint provider package archive file.
2828 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2831 To create the tracepoint provider package archive file:
2833 . Compile the <<tpp-source,tracepoint provider package source file>>:
2842 . Create the tracepoint provider package archive file:
2851 To build the instrumented application:
2853 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2858 #define TRACEPOINT_DEFINE
2862 . Compile the application source file:
2871 . Build the application:
2876 gcc -o app app.o tpp.a -llttng-ust -ldl
2880 To run the instrumented application:
2882 * Start the application:
2892 The instrumented application is linked with the tracepoint provider
2893 package shared object.
2895 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2898 include::../common/ust-sit-step-tp-so.txt[]
2900 To build the instrumented application:
2902 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2907 #define TRACEPOINT_DEFINE
2911 . Compile the application source file:
2920 . Build the application:
2925 gcc -o app app.o -ldl -L. -ltpp
2929 To run the instrumented application:
2931 * Start the application:
2941 The tracepoint provider package shared object is preloaded before the
2942 instrumented application starts.
2944 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2947 include::../common/ust-sit-step-tp-so.txt[]
2949 To build the instrumented application:
2951 . In path:{app.c}, before including path:{tpp.h}, add the
2957 #define TRACEPOINT_DEFINE
2958 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2962 . Compile the application source file:
2971 . Build the application:
2976 gcc -o app app.o -ldl
2980 To run the instrumented application with tracing support:
2982 * Preload the tracepoint provider package shared object and
2983 start the application:
2988 LD_PRELOAD=./libtpp.so ./app
2992 To run the instrumented application without tracing support:
2994 * Start the application:
3004 The instrumented application dynamically loads the tracepoint provider
3005 package shared object.
3007 See the <<dlclose-warning,warning about `dlclose()`>>.
3009 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3012 include::../common/ust-sit-step-tp-so.txt[]
3014 To build the instrumented application:
3016 . In path:{app.c}, before including path:{tpp.h}, add the
3022 #define TRACEPOINT_DEFINE
3023 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3027 . Compile the application source file:
3036 . Build the application:
3041 gcc -o app app.o -ldl
3045 To run the instrumented application:
3047 * Start the application:
3057 The application is linked with the instrumented user library.
3059 The instrumented user library is statically linked with the tracepoint
3060 provider package object file.
3062 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3065 include::../common/ust-sit-step-tp-o-fpic.txt[]
3067 To build the instrumented user library:
3069 . In path:{emon.c}, before including path:{tpp.h}, add the
3075 #define TRACEPOINT_DEFINE
3079 . Compile the user library source file:
3084 gcc -I. -fpic -c emon.c
3088 . Build the user library shared object:
3093 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3097 To build the application:
3099 . Compile the application source file:
3108 . Build the application:
3113 gcc -o app app.o -L. -lemon
3117 To run the application:
3119 * Start the application:
3129 The application is linked with the instrumented user library.
3131 The instrumented user library is linked with the tracepoint provider
3132 package shared object.
3134 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3137 include::../common/ust-sit-step-tp-so.txt[]
3139 To build the instrumented user library:
3141 . In path:{emon.c}, before including path:{tpp.h}, add the
3147 #define TRACEPOINT_DEFINE
3151 . Compile the user library source file:
3156 gcc -I. -fpic -c emon.c
3160 . Build the user library shared object:
3165 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3169 To build the application:
3171 . Compile the application source file:
3180 . Build the application:
3185 gcc -o app app.o -L. -lemon
3189 To run the application:
3191 * Start the application:
3201 The tracepoint provider package shared object is preloaded before the
3204 The application is linked with the instrumented user library.
3206 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3209 include::../common/ust-sit-step-tp-so.txt[]
3211 To build the instrumented user library:
3213 . In path:{emon.c}, before including path:{tpp.h}, add the
3219 #define TRACEPOINT_DEFINE
3220 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3224 . Compile the user library source file:
3229 gcc -I. -fpic -c emon.c
3233 . Build the user library shared object:
3238 gcc -shared -o libemon.so emon.o -ldl
3242 To build the application:
3244 . Compile the application source file:
3253 . Build the application:
3258 gcc -o app app.o -L. -lemon
3262 To run the application with tracing support:
3264 * Preload the tracepoint provider package shared object and
3265 start the application:
3270 LD_PRELOAD=./libtpp.so ./app
3274 To run the application without tracing support:
3276 * Start the application:
3286 The application is linked with the instrumented user library.
3288 The instrumented user library dynamically loads the tracepoint provider
3289 package shared object.
3291 See the <<dlclose-warning,warning about `dlclose()`>>.
3293 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3296 include::../common/ust-sit-step-tp-so.txt[]
3298 To build the instrumented user library:
3300 . In path:{emon.c}, before including path:{tpp.h}, add the
3306 #define TRACEPOINT_DEFINE
3307 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3311 . Compile the user library source file:
3316 gcc -I. -fpic -c emon.c
3320 . Build the user library shared object:
3325 gcc -shared -o libemon.so emon.o -ldl
3329 To build the application:
3331 . Compile the application source file:
3340 . Build the application:
3345 gcc -o app app.o -L. -lemon
3349 To run the application:
3351 * Start the application:
3361 The application dynamically loads the instrumented user library.
3363 The instrumented user library is linked with the tracepoint provider
3364 package shared object.
3366 See the <<dlclose-warning,warning about `dlclose()`>>.
3368 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3371 include::../common/ust-sit-step-tp-so.txt[]
3373 To build the instrumented user library:
3375 . In path:{emon.c}, before including path:{tpp.h}, add the
3381 #define TRACEPOINT_DEFINE
3385 . Compile the user library source file:
3390 gcc -I. -fpic -c emon.c
3394 . Build the user library shared object:
3399 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3403 To build the application:
3405 . Compile the application source file:
3414 . Build the application:
3419 gcc -o app app.o -ldl -L. -lemon
3423 To run the application:
3425 * Start the application:
3435 The application dynamically loads the instrumented user library.
3437 The instrumented user library dynamically loads the tracepoint provider
3438 package shared object.
3440 See the <<dlclose-warning,warning about `dlclose()`>>.
3442 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3445 include::../common/ust-sit-step-tp-so.txt[]
3447 To build the instrumented user library:
3449 . In path:{emon.c}, before including path:{tpp.h}, add the
3455 #define TRACEPOINT_DEFINE
3456 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3460 . Compile the user library source file:
3465 gcc -I. -fpic -c emon.c
3469 . Build the user library shared object:
3474 gcc -shared -o libemon.so emon.o -ldl
3478 To build the application:
3480 . Compile the application source file:
3489 . Build the application:
3494 gcc -o app app.o -ldl -L. -lemon
3498 To run the application:
3500 * Start the application:
3510 The tracepoint provider package shared object is preloaded before the
3513 The application dynamically loads the instrumented user library.
3515 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3518 include::../common/ust-sit-step-tp-so.txt[]
3520 To build the instrumented user library:
3522 . In path:{emon.c}, before including path:{tpp.h}, add the
3528 #define TRACEPOINT_DEFINE
3529 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3533 . Compile the user library source file:
3538 gcc -I. -fpic -c emon.c
3542 . Build the user library shared object:
3547 gcc -shared -o libemon.so emon.o -ldl
3551 To build the application:
3553 . Compile the application source file:
3562 . Build the application:
3567 gcc -o app app.o -L. -lemon
3571 To run the application with tracing support:
3573 * Preload the tracepoint provider package shared object and
3574 start the application:
3579 LD_PRELOAD=./libtpp.so ./app
3583 To run the application without tracing support:
3585 * Start the application:
3595 The application is statically linked with the tracepoint provider
3596 package object file.
3598 The application is linked with the instrumented user library.
3600 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3603 include::../common/ust-sit-step-tp-o.txt[]
3605 To build the instrumented user library:
3607 . In path:{emon.c}, before including path:{tpp.h}, add the
3613 #define TRACEPOINT_DEFINE
3617 . Compile the user library source file:
3622 gcc -I. -fpic -c emon.c
3626 . Build the user library shared object:
3631 gcc -shared -o libemon.so emon.o
3635 To build the application:
3637 . Compile the application source file:
3646 . Build the application:
3651 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3655 To run the instrumented application:
3657 * Start the application:
3667 The application is statically linked with the tracepoint provider
3668 package object file.
3670 The application dynamically loads the instrumented user library.
3672 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3675 include::../common/ust-sit-step-tp-o.txt[]
3677 To build the application:
3679 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3684 #define TRACEPOINT_DEFINE
3688 . Compile the application source file:
3697 . Build the application:
3702 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3707 The `--export-dynamic` option passed to the linker is necessary for the
3708 dynamically loaded library to ``see'' the tracepoint symbols defined in
3711 To build the instrumented user library:
3713 . Compile the user library source file:
3718 gcc -I. -fpic -c emon.c
3722 . Build the user library shared object:
3727 gcc -shared -o libemon.so emon.o
3731 To run the application:
3733 * Start the application:
3745 .Do not use man:dlclose(3) on a tracepoint provider package
3747 Never use man:dlclose(3) on any shared object which:
3749 * Is linked with, statically or dynamically, a tracepoint provider
3751 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3752 package shared object.
3754 This is currently considered **unsafe** due to a lack of reference
3755 counting from LTTng-UST to the shared object.
3757 A known workaround (available since glibc 2.2) is to use the
3758 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3759 effect of not unloading the loaded shared object, even if man:dlclose(3)
3762 You can also preload the tracepoint provider package shared object with
3763 the env:LD_PRELOAD environment variable to overcome this limitation.
3767 [[using-lttng-ust-with-daemons]]
3768 ===== Use noch:{LTTng-UST} with daemons
3770 If your instrumented application calls man:fork(2), man:clone(2),
3771 or BSD's man:rfork(2), without a following man:exec(3)-family
3772 system call, you must preload the path:{liblttng-ust-fork.so} shared
3773 object when starting the application.
3777 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3780 If your tracepoint provider package is
3781 a shared library which you also preload, you must put both
3782 shared objects in env:LD_PRELOAD:
3786 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3790 [[lttng-ust-pkg-config]]
3791 ===== Use noch:{pkg-config}
3793 On some distributions, LTTng-UST ships with a
3794 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3795 metadata file. If this is your case, then you can use cmd:pkg-config to
3796 build an application on the command line:
3800 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3804 [[instrumenting-32-bit-app-on-64-bit-system]]
3805 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3807 In order to trace a 32-bit application running on a 64-bit system,
3808 LTTng must use a dedicated 32-bit
3809 <<lttng-consumerd,consumer daemon>>.
3811 The following steps show how to build and install a 32-bit consumer
3812 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3813 build and install the 32-bit LTTng-UST libraries, and how to build and
3814 link an instrumented 32-bit application in that context.
3816 To build a 32-bit instrumented application for a 64-bit target system,
3817 assuming you have a fresh target system with no installed Userspace RCU
3820 . Download, build, and install a 32-bit version of Userspace RCU:
3826 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3827 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3828 cd userspace-rcu-0.9.* &&
3829 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3831 sudo make install &&
3836 . Using your distribution's package manager, or from source, install
3837 the following 32-bit versions of the following dependencies of
3838 LTTng-tools and LTTng-UST:
3841 * https://sourceforge.net/projects/libuuid/[libuuid]
3842 * http://directory.fsf.org/wiki/Popt[popt]
3843 * http://www.xmlsoft.org/[libxml2]
3846 . Download, build, and install a 32-bit version of the latest
3847 LTTng-UST{nbsp}{revision}:
3853 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.7.tar.bz2 &&
3854 tar -xf lttng-ust-latest-2.7.tar.bz2 &&
3855 cd lttng-ust-2.7.* &&
3856 ./configure --libdir=/usr/local/lib32 \
3857 CFLAGS=-m32 CXXFLAGS=-m32 \
3858 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3860 sudo make install &&
3867 Depending on your distribution,
3868 32-bit libraries could be installed at a different location than
3869 `/usr/lib32`. For example, Debian is known to install
3870 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3872 In this case, make sure to set `LDFLAGS` to all the
3873 relevant 32-bit library paths, for example:
3877 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3881 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3882 the 32-bit consumer daemon:
3888 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
3889 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
3890 cd lttng-tools-2.7.* &&
3891 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3892 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3894 cd src/bin/lttng-consumerd &&
3895 sudo make install &&
3900 . From your distribution or from source,
3901 <<installing-lttng,install>> the 64-bit versions of
3902 LTTng-UST and Userspace RCU.
3903 . Download, build, and install the 64-bit version of the
3904 latest LTTng-tools{nbsp}{revision}:
3910 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.7.tar.bz2 &&
3911 tar -xf lttng-tools-latest-2.7.tar.bz2 &&
3912 cd lttng-tools-2.7.* &&
3913 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3914 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3916 sudo make install &&
3921 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3922 when linking your 32-bit application:
3925 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3926 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3929 For example, let's rebuild the quick start example in
3930 <<tracing-your-own-user-application,Trace a user application>> as an
3931 instrumented 32-bit application:
3936 gcc -m32 -c -I. hello-tp.c
3938 gcc -m32 -o hello hello.o hello-tp.o \
3939 -L/usr/lib32 -L/usr/local/lib32 \
3940 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3945 No special action is required to execute the 32-bit application and
3946 to trace it: use the command-line man:lttng(1) tool as usual.
3953 `tracef()` is a small LTTng-UST API designed for quick,
3954 man:printf(3)-like instrumentation without the burden of
3955 <<tracepoint-provider,creating>> and
3956 <<building-tracepoint-providers-and-user-application,building>>
3957 a tracepoint provider package.
3959 To use `tracef()` in your application:
3961 . In the C or C++ source files where you need to use `tracef()`,
3962 include `<lttng/tracef.h>`:
3967 #include <lttng/tracef.h>
3971 . In the application's source code, use `tracef()` like you would use
3979 tracef("my message: %d (%s)", my_integer, my_string);
3985 . Link your application with `liblttng-ust`:
3990 gcc -o app app.c -llttng-ust
3994 To trace the events that `tracef()` calls emit:
3996 * <<enabling-disabling-events,Create an event rule>> which matches the
3997 `lttng_ust_tracef:*` event name:
4002 lttng enable-event --userspace 'lttng_ust_tracef:*'
4007 .Limitations of `tracef()`
4009 The `tracef()` utility function was developed to make user space tracing
4010 super simple, albeit with notable disadvantages compared to
4011 <<defining-tracepoints,user-defined tracepoints>>:
4013 * All the emitted events have the same tracepoint provider and
4014 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4015 * There is no static type checking.
4016 * The only event record field you actually get, named `msg`, is a string
4017 potentially containing the values you passed to `tracef()`
4018 using your own format string. This also means that you cannot filter
4019 events with a custom expression at run time because there are no
4021 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4022 function behind the scenes to format the strings at run time, its
4023 expected performance is lower than with user-defined tracepoints,
4024 which do not require a conversion to a string.
4026 Taking this into consideration, `tracef()` is useful for some quick
4027 prototyping and debugging, but you should not consider it for any
4028 permanent and serious applicative instrumentation.
4034 ==== Use `tracelog()`
4036 The `tracelog()` API is very similar to <<tracef,`tracef()`>>, with
4037 the difference that it accepts an additional log level parameter.
4039 The goal of `tracelog()` is to ease the migration from logging to
4042 To use `tracelog()` in your application:
4044 . In the C or C++ source files where you need to use `tracelog()`,
4045 include `<lttng/tracelog.h>`:
4050 #include <lttng/tracelog.h>
4054 . In the application's source code, use `tracelog()` like you would use
4055 man:printf(3), except for the first parameter which is the log
4063 tracelog(TRACE_WARNING, "my message: %d (%s)",
4064 my_integer, my_string);
4070 See <<liblttng-ust-tracepoint-loglevel,Tracepoint log levels>> for
4071 a list of available log level names.
4073 . Link your application with `liblttng-ust`:
4078 gcc -o app app.c -llttng-ust
4082 To trace the events that `tracelog()` calls emit with a log level
4083 _as severe as_ a specific log level:
4085 * <<enabling-disabling-events,Create an event rule>> which matches the
4086 `lttng_ust_tracelog:*` event name and a minimum level
4092 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4093 --loglevel=TRACE_WARNING
4097 To trace the events that `tracelog()` calls emit with a
4098 _specific log level_:
4100 * Create an event rule which matches the `lttng_ust_tracelog:*`
4101 event name and a specific log level:
4106 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4107 --loglevel-only=TRACE_INFO
4112 [[prebuilt-ust-helpers]]
4113 === Prebuilt user space tracing helpers
4115 The LTTng-UST package provides a few helpers in the form or preloadable
4116 shared objects which automatically instrument system functions and
4119 The helper shared objects are normally found in dir:{/usr/lib}. If you
4120 built LTTng-UST <<building-from-source,from source>>, they are probably
4121 located in dir:{/usr/local/lib}.
4123 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4126 path:{liblttng-ust-libc-wrapper.so}::
4127 path:{liblttng-ust-pthread-wrapper.so}::
4128 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4129 memory and POSIX threads function tracing>>.
4131 path:{liblttng-ust-cyg-profile.so}::
4132 path:{liblttng-ust-cyg-profile-fast.so}::
4133 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4135 path:{liblttng-ust-dl.so}::
4136 <<liblttng-ust-dl,Dynamic linker tracing>>.
4138 To use a user space tracing helper with any user application:
4140 * Preload the helper shared object when you start the application:
4145 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4149 You can preload more than one helper:
4154 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4160 [[liblttng-ust-libc-pthread-wrapper]]
4161 ==== Instrument C standard library memory and POSIX threads functions
4163 The path:{liblttng-ust-libc-wrapper.so} and
4164 path:{liblttng-ust-pthread-wrapper.so} helpers
4165 add instrumentation to some C standard library and POSIX
4169 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4171 |TP provider name |TP name |Instrumented function
4173 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4174 |`calloc` |man:calloc(3)
4175 |`realloc` |man:realloc(3)
4176 |`free` |man:free(3)
4177 |`memalign` |man:memalign(3)
4178 |`posix_memalign` |man:posix_memalign(3)
4182 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4184 |TP provider name |TP name |Instrumented function
4186 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4187 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4188 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4189 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4192 When you preload the shared object, it replaces the functions listed
4193 in the previous tables by wrappers which contain tracepoints and call
4194 the replaced functions.
4197 [[liblttng-ust-cyg-profile]]
4198 ==== Instrument function entry and exit
4200 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4201 to the entry and exit points of functions.
4203 man:gcc(1) and man:clang(1) have an option named
4204 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4205 which generates instrumentation calls for entry and exit to functions.
4206 The LTTng-UST function tracing helpers,
4207 path:{liblttng-ust-cyg-profile.so} and
4208 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4209 to add tracepoints to the two generated functions (which contain
4210 `cyg_profile` in their names, hence the helper's name).
4212 To use the LTTng-UST function tracing helper, the source files to
4213 instrument must be built using the `-finstrument-functions` compiler
4216 There are two versions of the LTTng-UST function tracing helper:
4218 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4219 that you should only use when it can be _guaranteed_ that the
4220 complete event stream is recorded without any lost event record.
4221 Any kind of duplicate information is left out.
4223 This version contains the following tracepoints:
4227 .Points instrumented by preloading path:{liblttng-ust-cyg-profile-fast.so}.
4229 |TP provider name |TP name |Instrumented points
4231 .2+|`lttng_ust_cyg_profile_fast`
4237 Address of called function.
4244 Assuming no event record is lost, having only the function addresses on
4245 entry is enough to create a call graph, since an event record always
4246 contains the ID of the CPU that generated it.
4248 You can use a tool like man:addr2line(1) to convert function addresses
4249 back to source file names and line numbers.
4251 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4252 which also works in use cases where event records might get discarded or
4253 not recorded from application startup.
4254 In these cases, the trace analyzer needs more information to be
4255 able to reconstruct the program flow.
4257 This version contains the following tracepoints:
4261 .Points instrumented by preloading path:{liblttng-ust-cyg-profile.so}.
4263 |TP provider name |TP name |Instrumented point
4265 .2+|`lttng_ust_cyg_profile`
4271 Address of called function.
4280 Address of called function.
4287 TIP: It's sometimes a good idea to limit the number of source files that
4288 you compile with the `-finstrument-functions` option to prevent LTTng
4289 from writing an excessive amount of trace data at run time. When using
4290 man:gcc(1), you can use the
4291 `-finstrument-functions-exclude-function-list` option to avoid
4292 instrument entries and exits of specific function names.
4294 All the tracepoints that this helper contains have the
4295 <<liblttng-ust-tracepoint-loglevel,log level>> `TRACE_DEBUG_FUNCTION`.
4300 ==== Instrument the dynamic linker
4302 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4303 man:dlopen(3) and man:dlclose(3) function calls.
4306 .Functions instrumented by preloading path:{liblttng-ust-dl.so}.
4308 |TP provider name |TP name |Instrumented function
4316 Memory base address (where the dynamic linker placed the shared
4320 File system path to the loaded shared object.
4323 File size of the the loaded shared object.
4326 Last modification time (seconds since Epoch time) of the loaded shared
4333 Memory base address (where the dynamic linker placed the shared
4339 [[java-application]]
4340 === User space Java agent
4342 You can instrument a Java application which uses one of the following
4345 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4346 (JUL) core logging facilities.
4347 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4348 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4350 Each log statement emits an LTTng event once the
4351 application initializes the <<lttng-ust-agents,LTTng-UST Java agent>>
4355 .LTTng-UST Java agent imported by a Java application.
4356 image::java-app.png[]
4358 NOTE: We use http://openjdk.java.net/[OpenJDK] 7 for development and
4359 https://ci.lttng.org/[continuous integration], thus this version is
4360 directly supported. However, the LTTng-UST Java agent is also
4361 tested with OpenJDK 6.
4363 To use the LTTng-UST Java agent:
4365 . In the Java application's source code, import the LTTng-UST Java
4371 import org.lttng.ust.agent.LTTngAgent;
4375 . As soon as possible after the entry point of the application,
4376 initialize the LTTng-UST Java agent:
4381 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4385 Any log statement that the application executes before this
4386 initialization does not emit an LTTng event.
4388 . Use `java.util.logging` and/or log4j log statements and configuration
4389 as usual. Since the LTTng-UST Java agent adds a handler to the _root_
4390 loggers, you can trace any log statement from any logger.
4392 . Before exiting the application, dispose the LTTng-UST Java agent:
4397 lttngAgent.dispose();
4401 This is not strictly necessary, but it is recommended for a clean
4402 disposal of the agent's resources.
4404 Any log statement that the application executes after this disposal does
4405 not emit an LTTng event.
4407 . Include the LTTng-UST Java agent's JAR file, path:{liblttng-ust-agent.jar},
4409 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class path]
4410 when building the Java application.
4412 path:{liblttng-ust-agent.jar} is typically located in
4413 dir:{/usr/share/java}.
4415 IMPORTANT: The LTTng-UST Java agent must be
4416 <<installing-lttng,installed>> for the logging framework your
4419 .[[jul]]Use the LTTng-UST Java agent with `java.util.logging`.
4424 import java.util.logging.Logger;
4425 import org.lttng.ust.agent.LTTngAgent;
4429 private static final int answer = 42;
4431 public static void main(String[] argv) throws Exception
4434 Logger logger = Logger.getLogger("jello");
4436 // Call this as soon as possible (before logging)
4437 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4440 logger.info("some info");
4441 logger.warning("some warning");
4443 logger.finer("finer information; the answer is " + answer);
4445 logger.severe("error!");
4447 // Not mandatory, but cleaner
4448 lttngAgent.dispose();
4453 You can build this example like this:
4457 javac -cp /usr/share/java/liblttng-ust-agent.jar Test.java
4460 You can run the compiled class like this:
4464 java -cp /usr/share/java/liblttng-ust-agent.jar:. Test
4468 .[[log4j]]Use the LTTng-UST Java agent with Apache log4j 1.2.
4473 import org.apache.log4j.Logger;
4474 import org.apache.log4j.BasicConfigurator;
4475 import org.lttng.ust.agent.LTTngAgent;
4479 private static final int answer = 42;
4481 public static void main(String[] argv) throws Exception
4483 // Create and configure a logger
4484 Logger logger = Logger.getLogger(Test.class);
4485 BasicConfigurator.configure();
4487 // Call this as soon as possible (before logging)
4488 LTTngAgent lttngAgent = LTTngAgent.getLTTngAgent();
4491 logger.info("some info");
4492 logger.warn("some warning");
4494 logger.debug("debug information; the answer is " + answer);
4496 logger.error("error!");
4497 logger.fatal("fatal error!");
4499 // Not mandatory, but cleaner
4500 lttngAgent.dispose();
4505 You can build this example like this:
4509 javac -cp /usr/share/java/liblttng-ust-agent.jar:$LOG4JCP Test.java
4512 where `$LOG4JCP` is the path to log4j's JAR file.
4514 You can run the compiled class like this:
4518 java -cp /usr/share/java/liblttng-ust-agent.jar:$LOG4JCP:. Test
4522 When you <<enabling-disabling-events,create an event rule>>, use the
4523 `--jul` (`java.util.logging`) or `--log4j` (log4j) option to target
4525 <<domain,tracing domain>>. You can also use the `--loglevel` or
4526 `--loglevel-only` option to target a range of JUL/log4j log levels or a
4527 specific JUL/log4j log level.
4531 [[python-application]]
4532 === User space Python agent
4534 You can instrument a Python 2 or Python 3 application which uses the
4535 standard https://docs.python.org/3/library/logging.html[`logging`]
4538 Each log statement emits an LTTng event once the
4539 application module imports the
4540 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4543 .A Python application importing the LTTng-UST Python agent.
4544 image::python-app.png[]
4546 To use the LTTng-UST Python agent:
4548 . In the Python application's source code, import the LTTng-UST Python
4558 The LTTng-UST Python agent automatically adds its logging handler to the
4559 root logger at import time.
4561 Any log statement that the application executes before this import does
4562 not emit an LTTng event.
4564 IMPORTANT: The LTTng-UST Python agent must be
4565 <<installing-lttng,installed>>.
4567 . Use log statements and logging configuration as usual.
4568 Since the LTTng-UST Python agent adds a handler to the _root_
4569 logger, you can trace any log statement from any logger.
4571 .Use the LTTng-UST Python agent.
4581 logging.basicConfig()
4582 logger = logging.getLogger('my-logger')
4585 logger.debug('debug message')
4586 logger.info('info message')
4587 logger.warn('warn message')
4588 logger.error('error message')
4589 logger.critical('critical message')
4593 if __name__ == '__main__':
4597 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4598 logging handler which prints to the standard error stream, is not
4599 strictly required for LTTng-UST tracing to work, but in versions of
4600 Python preceding 3.2, you could see a warning message which indicates
4601 that no handler exists for the logger `my-logger`.
4604 When you <<enabling-disabling-events,create an event rule>>, use the
4605 `--python` option to target the Python
4606 <<domain,tracing domain>>. You can also use
4607 the `--loglevel` or `--loglevel-only` option to target a range of
4608 Python log levels or a specific Python log level.
4610 When an application imports the LTTng-UST Python agent, the agent tries
4611 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4612 start the session daemon _before_ you start the Python application.
4613 If a session daemon is found, the agent tries to register to it
4614 during 5{nbsp}seconds, after which the application continues without
4615 LTTng tracing support. You can override this timeout value with the
4616 env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4619 If the session daemon stops while a Python application with an imported
4620 LTTng-UST Python agent runs, the agent retries to connect and to
4621 register to a session daemon every 3{nbsp}seconds. You can override this
4622 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4627 [[proc-lttng-logger-abi]]
4630 The `lttng-tracer` Linux kernel module, part of
4631 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4632 path:{/proc/lttng-logger} when it's loaded. Any application can write
4633 text data to this file to emit an LTTng event.
4636 .An application writes to the LTTng logger file to emit an LTTng event.
4637 image::lttng-logger.png[]
4639 The LTTng logger is the quickest method--not the most efficient,
4640 however--to add instrumentation to an application. It is designed
4641 mostly to instrument shell scripts:
4645 echo "Some message, some $variable" > /proc/lttng-logger
4648 Any event that the LTTng logger emits is named `lttng_logger` and
4649 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4650 other instrumentation points in the kernel tracing domain, **any Unix
4651 user** can <<enabling-disabling-events,create an event rule>> which
4652 matches its event name, not only the root user or users in the tracing
4655 To use the LTTng logger:
4657 * From any application, write text data to the path:{/proc/lttng-logger}
4660 The `msg` field of `lttng_logger` event records contains the
4663 NOTE: The maximum message length of an LTTng logger event is
4664 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4665 than one event to contain the remaining data.
4667 You should not use the LTTng logger to trace a user application which
4668 can be instrumented in a more efficient way, namely:
4670 * <<c-application,C and $$C++$$ applications>>.
4671 * <<java-application,Java applications>>.
4672 * <<python-application,Python applications>>.
4675 [[instrumenting-linux-kernel]]
4676 === LTTng kernel tracepoints
4678 NOTE: This section shows how to _add_ instrumentation points to the
4679 Linux kernel. The kernel's subsystems are already thoroughly
4680 instrumented at strategic places for LTTng when you
4681 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
4685 There are two methods to instrument the Linux kernel:
4687 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
4688 tracepoint which uses the `TRACE_EVENT()` API.
4690 Choose this if you want to instrumentation a Linux kernel tree with an
4691 instrumentation point compatible with ftrace, perf, and SystemTap.
4693 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
4694 instrument an out-of-tree kernel module.
4696 Choose this if you don't need ftrace, perf, or SystemTap support.
4700 [[linux-add-lttng-layer]]
4701 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
4703 This section shows how to add an LTTng layer to existing ftrace
4704 instrumentation using the `TRACE_EVENT()` API.
4706 This section does not document the `TRACE_EVENT()` macro. You can
4707 read the following articles to learn more about this API:
4709 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
4710 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
4711 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
4713 The following procedure assumes that your ftrace tracepoints are
4714 correctly defined in their own header and that they are created in
4715 one source file using the `CREATE_TRACE_POINTS` definition.
4717 To add an LTTng layer over an existing ftrace tracepoint:
4719 . Make sure the following kernel configuration options are
4725 * `CONFIG_HIGH_RES_TIMERS`
4726 * `CONFIG_TRACEPOINTS`
4729 . Build the Linux source tree with your custom ftrace tracepoints.
4730 . Boot the resulting Linux image on your target system.
4732 Confirm that the tracepoints exist by looking for their names in the
4733 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
4734 is your subsystem's name.
4736 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
4742 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.8.tar.bz2 &&
4743 tar -xf lttng-modules-latest-2.8.tar.bz2 &&
4744 cd lttng-modules-2.8.*
4748 . In dir:{instrumentation/events/lttng-module}, relative to the root
4749 of the LTTng-modules source tree, create a header file named
4750 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
4751 LTTng-modules tracepoint definitions using the LTTng-modules
4754 Start with this template:
4758 .path:{instrumentation/events/lttng-module/my_subsys.h}
4761 #define TRACE_SYSTEM my_subsys
4763 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
4764 #define _LTTNG_MY_SUBSYS_H
4766 #include "../../../probes/lttng-tracepoint-event.h"
4767 #include <linux/tracepoint.h>
4769 LTTNG_TRACEPOINT_EVENT(
4771 * Format is identical to TRACE_EVENT()'s version for the three
4772 * following macro parameters:
4775 TP_PROTO(int my_int, const char *my_string),
4776 TP_ARGS(my_int, my_string),
4778 /* LTTng-modules specific macros */
4780 ctf_integer(int, my_int_field, my_int)
4781 ctf_string(my_bar_field, my_bar)
4785 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
4787 #include "../../../probes/define_trace.h"
4791 The entries in the `TP_FIELDS()` section are the list of fields for the
4792 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
4793 ftrace's `TRACE_EVENT()` macro.
4795 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
4796 complete description of the available `ctf_*()` macros.
4798 . Create the LTTng-modules probe's kernel module C source file,
4799 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
4804 .path:{probes/lttng-probe-my-subsys.c}
4806 #include <linux/module.h>
4807 #include "../lttng-tracer.h"
4810 * Build-time verification of mismatch between mainline
4811 * TRACE_EVENT() arguments and the LTTng-modules adaptation
4812 * layer LTTNG_TRACEPOINT_EVENT() arguments.
4814 #include <trace/events/my_subsys.h>
4816 /* Create LTTng tracepoint probes */
4817 #define LTTNG_PACKAGE_BUILD
4818 #define CREATE_TRACE_POINTS
4819 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
4821 #include "../instrumentation/events/lttng-module/my_subsys.h"
4823 MODULE_LICENSE("GPL and additional rights");
4824 MODULE_AUTHOR("Your name <your-email>");
4825 MODULE_DESCRIPTION("LTTng my_subsys probes");
4826 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
4827 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
4828 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
4829 LTTNG_MODULES_EXTRAVERSION);
4833 . Edit path:{probes/Makefile} and add your new kernel module object
4834 next to the existing ones:
4838 .path:{probes/Makefile}
4842 obj-m += lttng-probe-module.o
4843 obj-m += lttng-probe-power.o
4845 obj-m += lttng-probe-my-subsys.o
4851 . Build and install the LTTng kernel modules:
4856 make KERNELDIR=/path/to/linux
4857 sudo make modules_install
4861 Replace `/path/to/linux` with the path to the Linux source tree where
4862 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
4864 Note that you can also use the
4865 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
4866 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
4867 C code that need to be executed before the event fields are recorded.
4869 The best way to learn how to use the previous LTTng-modules macros is to
4870 inspect the existing LTTng-modules tracepoint definitions in the
4871 dir:{instrumentation/events/lttng-module} header files. Compare them
4872 with the Linux kernel mainline versions in the
4873 dir:{include/trace/events} directory of the Linux source tree.
4877 [[lttng-tracepoint-event-code]]
4878 ===== Use custom C code to access the data for tracepoint fields
4880 Although we recommended to always use the
4881 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
4882 the arguments and fields of an LTTng-modules tracepoint when possible,
4883 sometimes you need a more complex process to access the data that the
4884 tracer records as event record fields. In other words, you need local
4885 variables and multiple C{nbsp}statements instead of simple
4886 argument-based expressions that you pass to the
4887 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
4889 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
4890 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
4891 a block of C{nbsp}code to be executed before LTTng records the fields.
4892 The structure of this macro is:
4895 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
4897 LTTNG_TRACEPOINT_EVENT_CODE(
4899 * Format identical to the LTTNG_TRACEPOINT_EVENT()
4900 * version for the following three macro parameters:
4903 TP_PROTO(int my_int, const char *my_string),
4904 TP_ARGS(my_int, my_string),
4906 /* Declarations of custom local variables */
4909 unsigned long b = 0;
4910 const char *name = "(undefined)";
4911 struct my_struct *my_struct;
4915 * Custom code which uses both tracepoint arguments
4916 * (in TP_ARGS()) and local variables (in TP_locvar()).
4918 * Local variables are actually members of a structure pointed
4919 * to by the special variable tp_locvar.
4923 tp_locvar->a = my_int + 17;
4924 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
4925 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
4926 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
4927 put_my_struct(tp_locvar->my_struct);
4936 * Format identical to the LTTNG_TRACEPOINT_EVENT()
4937 * version for this, except that tp_locvar members can be
4938 * used in the argument expression parameters of
4939 * the ctf_*() macros.
4942 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
4943 ctf_integer(int, my_struct_a, tp_locvar->a)
4944 ctf_string(my_string_field, my_string)
4945 ctf_string(my_struct_name, tp_locvar->name)
4950 IMPORTANT: The C code defined in `TP_code()` must not have any side
4951 effects when executed. In particular, the code must not allocate
4952 memory or get resources without deallocating this memory or putting
4953 those resources afterwards.
4956 [[instrumenting-linux-kernel-tracing]]
4957 ==== Load and unload a custom probe kernel module
4959 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
4960 kernel module>> in the kernel before it can emit LTTng events.
4962 To load the default probe kernel modules and a custom probe kernel
4965 * Use the `--extra-kmod-probes` option to give extra probe modules
4966 to load when starting a root <<lttng-sessiond,session daemon>>:
4969 .Load the `my_subsys`, `usb`, and the default probe modules.
4973 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
4978 You only need to pass the subsystem name, not the whole kernel module
4981 To load _only_ a given custom probe kernel module:
4983 * Use the `--kmod-probes` option to give the probe modules
4984 to load when starting a root session daemon:
4987 .Load only the `my_subsys` and `usb` probe modules.
4991 sudo lttng-sessiond --kmod-probes=my_subsys,usb
4996 To confirm that a probe module is loaded:
5003 lsmod | grep lttng_probe_usb
5007 To unload the loaded probe modules:
5009 * Kill the session daemon with `SIGTERM`:
5014 sudo pkill lttng-sessiond
5018 You can also use man:modprobe(8)'s `--remove` option if the session
5019 daemon terminates abnormally.
5022 [[controlling-tracing]]
5025 Once an application or a Linux kernel is
5026 <<instrumenting,instrumented>> for LTTng tracing,
5029 This section is divided in topics on how to use the various
5030 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5031 command-line tool>>, to _control_ the LTTng daemons and tracers.
5033 Note that the <<online-lttng-manpages,Online LTTng man pages>> are
5034 more comprehensive than the guides of this section. Refer to them if
5035 your use case is not included in this section.
5039 === Start a session daemon
5041 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5042 _before_ you can use the cmd:lttng command-line tool.
5044 You will see the following error when you run a command while no session
5048 Error: No session daemon is available
5051 The only command that automatically runs a session daemon is `create`,
5052 which you use to <<creating-destroying-tracing-sessions,create a tracing
5053 session>>. While this is most of the time the first operation that you
5054 do, sometimes it's not. Some examples are:
5056 * <<list-instrumentation-points,List the available instrumentation points>>.
5057 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5059 [[tracing-group]] Each Unix user must have its own running session
5060 daemon to trace user applications. The session daemon that the root user
5061 starts is the only one allowed to control the LTTng kernel tracer. Users
5062 that are part of the _tracing group_ can control the root session
5063 daemon. The default tracing group name is `tracing`; you can set it to
5064 something else with the `--group` option when you start the root session
5067 To start a user session daemon:
5069 * Run cmd:lttng-sessiond:
5074 lttng-sessiond --daemonize
5078 To start the root session daemon:
5080 * Run cmd:lttng-sessiond as the root user:
5085 sudo lttng-sessiond --daemonize
5089 In both cases, remove the `--daemonize` option to start the session
5090 daemon in foreground.
5092 To stop a session daemon, use cmd:kill on its process ID (standard
5095 Note that some Linux distributions could manage the LTTng session daemon
5096 as a service. In this case, you should use the service manager to
5097 start, restart, and stop session daemons.
5100 [[creating-destroying-tracing-sessions]]
5101 === Create and destroy a tracing session
5103 Almost all the LTTng control operations happen in the scope of
5104 a <<tracing-session,tracing session>>, which is the dialogue between the
5105 <<lttng-sessiond,session daemon>> and you.
5107 To create a tracing session with a generated name:
5109 * Use the `create` command:
5118 The created tracing session's name is `auto` followed by the
5121 To create a tracing session with a specific name:
5123 * Use the optional argument of the `create` command:
5128 lttng create my-session
5132 Replace `my-session` with the specific tracing session name.
5134 LTTng appends the creation date to the created tracing session's name.
5136 LTTng writes the traces of a tracing session in
5137 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5138 name of the tracing session. Note that the env:LTTNG_HOME environment
5139 variable defaults to `$HOME` if not set.
5141 To output LTTng traces to a non-default location:
5143 * Use the `--output` option of the `create` command:
5148 lttng create --output=/tmp/some-directory my-session
5152 You may create as many tracing sessions as you wish.
5154 To list all the existing tracing sessions for your Unix user:
5156 * Use the `list` command:
5165 When you create a tracing session, it is set as the _current tracing
5166 session_. The following man:lttng(1) commands operate on the current
5167 tracing session when you don't specify one:
5169 [role="list-3-cols"]
5185 To change the current tracing session:
5187 * Use the `set-session` command:
5192 lttng set-session new-session
5196 Replace `new-session` by the name of the new current tracing session.
5198 When you are done tracing in a given tracing session, you can destroy
5199 it. This operation frees the resources taken by the tracing session
5200 to destroy; it does not destroy the trace data that LTTng wrote for
5201 this tracing session.
5203 To destroy the current tracing session:
5205 * Use the `destroy` command:
5215 [[list-instrumentation-points]]
5216 === List the available instrumentation points
5218 The <<lttng-sessiond,session daemon>> can query the running instrumented
5219 user applications and the Linux kernel to get a list of available
5220 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5221 they are tracepoints and system calls. For the user space tracing
5222 domain, they are tracepoints. For the other tracing domains, they are
5225 To list the available instrumentation points:
5227 * Use the `list` command with the requested tracing domain's option
5231 * `--kernel`: Linux kernel tracepoints (your Unix user must be a root
5232 user, or it must be a member of the tracing group).
5233 * `--kernel --syscall`: Linux kernel system calls (your Unix user must
5234 be a root user, or it must be a member of the tracing group).
5235 * `--userspace`: user space tracepoints.
5236 * `--jul`: `java.util.logging` loggers.
5237 * `--log4j`: Apache log4j loggers.
5238 * `--python`: Python loggers.
5241 .List the available user space tracepoints.
5245 lttng list --userspace
5249 .List the available Linux kernel system call tracepoints.
5253 lttng list --kernel --syscall
5258 [[enabling-disabling-events]]
5259 === Create and enable an event rule
5261 Once you <<creating-destroying-tracing-sessions,create a tracing
5262 session>>, you can create <<event,event rules>> with the
5263 `enable-event` command.
5265 You specify each condition with a command-line option. The available
5266 condition options are shown in the following table.
5268 [role="growable",cols="asciidoc,asciidoc,default"]
5269 .Condition command-line options for the `enable-event` command.
5271 |Option |Description |Applicable tracing domains
5277 . +--probe=__ADDR__+
5278 . +--function=__ADDR__+
5281 Instead of using the default _tracepoint_ instrumentation type, use:
5283 . A Linux system call.
5284 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5285 . The entry and return points of a Linux function (symbol or address).
5289 |First positional argument.
5292 Tracepoint or system call name. In the case of a Linux KProbe or
5293 function, this is a custom name given to the event rule. With the
5294 JUL, log4j, and Python domains, this is a logger name.
5296 With a tracepoint, logger, or system call name, the last character
5297 can be `*` to match anything that remains.
5304 . +--loglevel=__LEVEL__+
5305 . +--loglevel-only=__LEVEL__+
5308 . Match only tracepoints or log statements with a logging level at
5309 least as severe as +__LEVEL__+.
5310 . Match only tracepoints or log statements with a logging level
5311 equal to +__LEVEL__+.
5313 You can get the list of available logging level names with
5314 `lttng enable-event --help`.
5316 |User space, JUL, log4j, and Python.
5318 |+--exclude=__EXCLUSIONS__+
5321 When you use a `*` character at the end of the tracepoint or logger
5322 name (first positional argument), exclude the specific names in the
5323 comma-delimited list +__EXCLUSIONS__+.
5326 User space, JUL, log4j, and Python.
5328 |+--filter=__EXPR__+
5331 Match only events which satisfy the expression +__EXPR__+.
5333 +__EXPR__+ is a C-like logical expression where identifiers are event
5334 fields (preceded with `$ctx.` for context fields). Nested expressions
5335 with `(` and `)`, and all the logical and comparison operators of the C
5336 language are supported. The precedence rules of those operators are the
5337 same as in the C language.
5339 When a comparison includes a non-existent event field, the whole filter
5340 expression evaluates to false.
5342 C integer and floating point number constants are supported, as well as
5343 literal strings between double quotes (`"`). Literal strings can
5344 contain a wildcard character (`*`) at the end to match anything that
5345 remains. This wildcard can be escaped using `\*`.
5347 Note that, although it is possible to use this option with the JUL,
5348 log4j, and Python tracing domains, the tracer evalutes the expression
5349 against the equivalent user space event.
5356 for more details about those command-line options.
5358 You attach an event rule to a <<channel,channel>> on creation. If you
5359 do not specify the channel with the `--channel` option, and if the event
5360 rule to create is the first in its <<domain,tracing domain>> for a given
5361 tracing session, then LTTng creates a _default channel_ for you. This
5362 default channel is reused in subsequent invocations of the
5363 `enable-event` command for the same tracing domain.
5365 An event rule is always enabled at creation time.
5367 The following examples show how you can combine the previous
5368 command-line options to create simple to more complex event rules.
5370 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5374 lttng enable-event --kernel sched_switch
5378 .Create an event rule matching four Linux kernel system calls (default channel).
5382 lttng enable-event --kernel --syscall open,write,read,close
5386 .Create an event rule matching a Linux kernel tracepoint with a filter expression (default channel).
5390 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5393 IMPORTANT: Make sure to always quote the filter string when you
5394 use man:lttng(1) from a shell.
5397 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5401 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5404 IMPORTANT: Make sure to always quote the wildcard character when you
5405 use man:lttng(1) from a shell.
5408 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5412 lttng enable-event --python my-app.'*' \
5413 --exclude='my-app.module,my-app.hello'
5417 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5421 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5425 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5429 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5433 The event rules of a given channel form a whitelist: as soon as an
5434 emitted event passes one of them, LTTng can record the event. For
5435 example, an event named `my_app:my_tracepoint` emitted from a user space
5436 tracepoint with a `TRACE_ERROR` log level passes both of the following
5441 lttng enable-event --userspace my_app:my_tracepoint
5442 lttng enable-event --userspace my_app:my_tracepoint \
5443 --loglevel=TRACE_INFO
5446 The second event rule is redundant: the first one includes
5450 [[disable-event-rule]]
5451 === Disable an event rule
5453 To disable an event rule that you <<enabling-disabling-events,created>>
5454 previously, use the `disable-event` command. This command disables _all_
5455 the event rules (of a given tracing domain and channel) which match an
5456 instrumentation point. The other conditions are not supported as of
5457 LTTng{nbsp}{revision}.
5459 The LTTng tracer does not record an emitted event which passes
5460 a _disabled_ event rule.
5462 .Disable an event rule matching a Python logger (default channel).
5466 lttng disable-event --python my-logger
5470 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5474 lttng disable-event --jul '*'
5478 .Disable _all_ the event rules of the default channel.
5480 The `--all-events` option is not, like the `--all` option of
5481 `enable-event`, the equivalent of the event name `*` (wildcard): it
5482 disables _all_ the event rules of a given channel.
5486 lttng disable-event --jul --all-events
5490 NOTE: You cannot delete an event rule once you create it.
5494 === Get the status of a tracing session
5496 To get the status of a tracing session, that is, its channels, event
5497 rules, and their attributes:
5499 * Use the `list` command with the tracing session's name:
5504 lttng list my-session
5508 Replace `my-session` with your tracing session's name.
5511 [[basic-tracing-session-control]]
5512 === Start and stop a tracing session
5514 Once you <<creating-destroying-tracing-sessions,create a tracing
5516 <<enabling-disabling-events,create one or more event rules>>,
5517 you can start and stop the tracers for this tracing session.
5519 To start tracing in the current tracing session:
5521 * Use the `start` command:
5530 To stop tracing in the current tracing session:
5532 * Use the `stop` command:
5541 LTTng is very flexible: you can launch user applications before
5542 or after the you start the tracers. The tracers only record the events
5543 if they pass enabled event rules and if they occur while the tracers are
5547 [[enabling-disabling-channels]]
5548 === Create a channel
5550 Once you create a tracing session, you can create a <<channel,channel>>
5551 with the `enable-channel` command.
5553 Note that LTTng automatically creates a default channel when, for a
5554 given <<domain,tracing domain>>, no channels exist and you
5555 <<enabling-disabling-events,create>> the first event rule. This default
5556 channel is named `channel0` and its attributes are set to reasonable
5557 values. Therefore, you only need to create a channel when you need
5558 non-default attributes.
5560 You specify each non-default channel attribute with a command-line
5561 option when you use the `enable-channel` command. The available
5562 command-line options are:
5564 [role="growable",cols="asciidoc,asciidoc"]
5565 .Command-line options for the `enable-channel` command.
5567 |Option |Description
5573 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5574 the default _discard_ mode.
5576 |`--buffers-pid` (user space tracing domain only)
5579 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5580 instead of the default per-user buffering scheme.
5582 |+--subbuf-size=__SIZE__+
5585 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5586 either for each Unix user (default), or for each instrumented process.
5588 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5590 |+--num-subbuf=__COUNT__+
5593 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5594 for each Unix user (default), or for each instrumented process.
5596 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5598 |+--tracefile-size=__SIZE__+
5601 Set the maximum size of each trace file that this channel writes within
5602 a stream to +__SIZE__+ bytes instead of no maximum.
5604 See <<tracefile-rotation,Trace file count and size>>.
5606 |+--tracefile-count=__COUNT__+
5609 Limit the number of trace files that this channel creates to
5610 +__COUNT__+ channels instead of no limit.
5612 See <<tracefile-rotation,Trace file count and size>>.
5614 |+--switch-timer=__PERIODUS__+
5617 Set the <<channel-switch-timer,switch timer period>>
5618 to +__PERIODUS__+{nbsp}µs.
5620 |+--read-timer=__PERIODUS__+
5623 Set the <<channel-read-timer,read timer period>>
5624 to +__PERIODUS__+{nbsp}µs.
5626 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5629 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5634 for more details about those command-line options.
5636 You can only create a channel in the Linux kernel and user space
5637 <<domain,tracing domains>>: other tracing domains have their own
5638 channel created on the fly when
5639 <<enabling-disabling-events,creating event rules>>.
5643 Because of a current LTTng limitation, you must create all channels
5644 _before_ you <<basic-tracing-session-control,start tracing>> in a given
5645 tracing session, that is, before the first time you run `lttng start`.
5647 Since LTTng automatically creates a default channel when you use the
5648 `enable-event` command with a specific tracing domain, you cannot, for
5649 example, create a Linux kernel event rule, start tracing, and then
5650 create a user space event rule, because no user space channel exists yet
5651 and it's too late to create one.
5653 For this reason, make sure to configure your channels properly
5654 before starting the tracers for the first time!
5657 The following examples show how you can combine the previous
5658 command-line options to create simple to more complex channels.
5660 .Create a Linux kernel channel with default attributes.
5664 lttng enable-channel --kernel my-channel
5668 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
5672 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
5673 --buffers-pid my-channel
5677 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
5681 lttng enable-channel --kernel --tracefile-count=8 \
5682 --tracefile-size=4194304 my-channel
5686 .Create a user space channel in overwrite (or _flight recorder_) mode.
5690 lttng enable-channel --userspace --overwrite my-channel
5694 You can <<enabling-disabling-events,create>> the same event rule in
5695 two different channels:
5699 lttng enable-event --userspace --channel=my-channel app:tp
5700 lttng enable-event --userspace --channel=other-channel app:tp
5703 If both channels are enabled, when a tracepoint named `app:tp` is
5704 reached, LTTng records two events, one for each channel.
5708 === Disable a channel
5710 To disable a specific channel that you <<enabling-disabling-channels,created>>
5711 previously, use the `disable-channel` command.
5713 .Disable a specific Linux kernel channel.
5717 lttng disable-channel --kernel my-channel
5721 The state of a channel precedes the individual states of event rules
5722 attached to it: event rules which belong to a disabled channel, even if
5723 they are enabled, are also considered disabled.
5727 === Add context fields to a channel
5729 Event record fields in trace files provide important information about
5730 events that occured previously, but sometimes some external context may
5731 help you solve a problem faster. Examples of context fields are:
5733 * The **process ID**, **thread ID**, **process name**, and
5734 **process priority** of the thread in which the event occurs.
5735 * The **hostname** of the system on which the event occurs.
5736 * The current values of many possible **performance counters** using
5738 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
5740 ** Branch instructions, misses, and loads.
5743 To get the full list of available context fields, see
5744 `lttng add-context --help`. Some context fields are reserved for a
5745 specific <<domain,tracing domain>> (Linux kernel or user space).
5747 You add context fields to <<channel,channels>>. All the events
5748 that a channel with added context fields records contain those fields.
5750 To add context fields to one or all the channels of a given tracing
5751 session, use the `add-context` command.
5753 .Add context fields to all the channels of the current tracing session.
5755 The following command line adds the virtual process identifier and
5756 the per-thread CPU cycles count fields to all the user space channels
5757 of the current tracing session.
5761 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
5765 .Add a context field to a specific channel.
5767 The following command line adds the thread identifier context field
5768 to the Linux kernel channel named `my-channel` in the current
5773 lttng add-context --kernel --channel=my-channel --type=tid
5777 NOTE: You cannot remove context fields from a channel once you add it.
5782 === Track process IDs
5784 It's often useful to allow only specific process IDs (PIDs) to emit
5785 events. For example, you may wish to record all the system calls made by
5786 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
5788 The `track` and `untrack` commands serve this purpose. Both commands
5789 operate on a whitelist of process IDs. You _add_ entries to this
5790 whitelist with the `track` command and remove entries with the `untrack`
5791 command. Any process which has one of the PIDs in the whitelist is
5792 allowed to emit LTTng events which pass an enabled <<event,event rule>>.
5794 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
5795 process with a given tracked ID exit and another process be given this
5796 ID, then the latter would also be allowed to emit events.
5798 .Track and untrack process IDs.
5800 For the sake of the following example, assume the target system has 16
5804 <<creating-destroying-tracing-sessions,create a tracing session>>,
5805 the whitelist contains all the possible PIDs:
5808 .All PIDs are tracked.
5809 image::track-all.png[]
5811 When the whitelist is full and you use the `track` command to specify
5812 some PIDs to track, LTTng first clears the whitelist, then it tracks
5813 the specific PIDs. After:
5817 lttng track --pid=3,4,7,10,13
5823 .PIDs 3, 4, 7, 10, and 13 are tracked.
5824 image::track-3-4-7-10-13.png[]
5826 You can add more PIDs to the whitelist afterwards:
5830 lttng track --pid=1,15,16
5836 .PIDs 1, 15, and 16 are added to the whitelist.
5837 image::track-1-3-4-7-10-13-15-16.png[]
5839 The `untrack` command removes entries from the PID tracker's whitelist.
5840 Given the previous example, the following command:
5844 lttng untrack --pid=3,7,10,13
5847 leads to this whitelist:
5850 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
5851 image::track-1-4-15-16.png[]
5853 LTTng can track all possible PIDs again using the `--all` option:
5857 lttng track --pid --all
5860 The result is, again:
5863 .All PIDs are tracked.
5864 image::track-all.png[]
5867 .Track only specific PIDs
5869 A very typical use case with PID tracking is to start with an empty
5870 whitelist, then <<basic-tracing-session-control,start the tracers>>,
5871 and then add PIDs manually while tracers are active. You can accomplish
5872 this by using the `--all` option of the `untrack` command to clear the
5873 whitelist after you create a tracing session:
5877 lttng untrack --pid --all
5883 .No PIDs are tracked.
5884 image::untrack-all.png[]
5886 If you trace with this whitelist configuration, the tracer records no
5887 events for this <<domain,tracing domain>> because no processes are
5888 tracked. You can use the `track` command as usual to track specific
5893 lttng track --pid=6,11
5899 .PIDs 6 and 11 are tracked.
5900 image::track-6-11.png[]
5905 [[saving-loading-tracing-session]]
5906 === Save and load tracing session configurations
5908 Configuring a <<tracing-session,tracing session>> can be long. Some of
5909 the tasks involved are:
5911 * <<enabling-disabling-channels,Create channels>> with
5912 specific attributes.
5913 * <<adding-context,Add context fields>> to specific channels.
5914 * <<enabling-disabling-events,Create event rules>> with specific log
5915 level and filter conditions.
5917 If you use LTTng to solve real world problems, chances are you have to
5918 record events using the same tracing session setup over and over,
5919 modifying a few variables each time in your instrumented program
5920 or environment. To avoid constant tracing session reconfiguration,
5921 the cmd:lttng command-line tool can save and load tracing session
5922 configurations to/from XML files.
5924 To save a given tracing session configuration:
5926 * Use the `save` command:
5931 lttng save my-session
5935 Replace `my-session` with the name of the tracing session to save.
5937 LTTng saves tracing session configurations to
5938 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
5939 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
5940 the `--output-path` option to change this destination directory.
5942 LTTng saves all configuration parameters, for example:
5944 * The tracing session name.
5945 * The trace data output path.
5946 * The channels with their state and all their attributes.
5947 * The context fields you added to channels.
5948 * The event rules with their state, log level and filter conditions.
5950 To load a tracing session:
5952 * Use the `load` command:
5957 lttng load my-session
5961 Replace `my-session` with the name of the tracing session to load.
5963 When LTTng loads a configuration, it restores your saved tracing session
5964 as if you just configured it manually.
5966 See man:lttng(1) for the complete list of command-line options. You
5967 can also save and load all many sessions at a time, and decide in which
5968 directory to output the XML files.
5971 [[sending-trace-data-over-the-network]]
5972 === Send trace data over the network
5974 LTTng can send the recorded trace data to a remote system over the
5975 network instead of writing it to the local file system.
5977 To send the trace data over the network:
5979 . On the _remote_ system (which can also be the target system),
5980 start an LTTng <<lttng-relayd,relay daemon>>:
5989 . On the _target_ system, create a tracing session configured to
5990 send trace data over the network:
5995 lttng create my-session --set-url=net://remote-system
5999 Replace `remote-system` by the host name or IP address of the
6000 remote system. See `lttng create --help` for the exact URL format.
6002 . On the target system, use the cmd:lttng command-line tool as usual.
6003 When tracing is active, the target's consumer daemon sends sub-buffers
6004 to the relay daemon running on the remote system intead of flushing
6005 them to the local file system. The relay daemon writes the received
6006 packets to the local file system.
6008 The relay daemon writes trace files to
6009 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6010 +__hostname__+ is the host name of the target system and +__session__+
6011 is the tracing session name. Note that the env:LTTNG_HOME environment
6012 variable defaults to `$HOME` if not set. Use the `--output` option of
6013 cmd:lttng-relayd to write trace files to another base directory.
6018 === View events as LTTng emits them (noch:{LTTng} live)
6020 LTTng live is a network protocol implemented by the
6021 <<lttng-relayd,relay daemon>> to allow compatible trace viewers to
6022 display events as LTTng emits them on the target system while tracing
6025 The relay daemon creates a _tee_: it forwards the trace data to both
6026 the local file system and to connected live viewers:
6029 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6034 . On the _target system_, create a <<tracing-session,tracing session>>
6040 lttng create --live my-session
6044 This spawns a local relay daemon.
6046 . Start the live viewer and configure it to connect to the relay
6047 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6052 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6059 * `hostname` with the host name of the target system.
6060 * `my-session` with the name of the tracing session to view.
6063 . Configure the tracing session as usual with the cmd:lttng
6064 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6066 You can list the available live tracing sessions with Babeltrace:
6070 babeltrace --input-format=lttng-live net://localhost
6073 You can start the relay daemon on another system. In this case, you need
6074 to specify the relay daemon's URL when you create the tracing session
6075 with the `--set-url` option. You also need to replace `localhost`
6076 in the procedure above with the host name of the system on which the
6077 relay daemon is running.
6079 See man:lttng(1) and man:lttng-relayd(8) for the complete list of
6080 command-line options.
6084 [[taking-a-snapshot]]
6085 === Take a snapshot of the current sub-buffers of a tracing session
6087 The normal behavior of LTTng is to append full sub-buffers to growing
6088 trace data files. This is ideal to keep a full history of the events
6089 that occurred on the target system, but it can
6090 represent too much data in some situations. For example, you may wish
6091 to trace your application continuously until some critical situation
6092 happens, in which case you only need the latest few recorded
6093 events to perform the desired analysis, not multi-gigabyte trace files.
6095 With the `snapshot` command, you can take a snapshot of the current
6096 sub-buffers of a given <<tracing-session,tracing session>>. LTTng can
6097 write the snapshot to the local file system or send it over the network.
6101 . Create a tracing session in _snapshot mode_:
6106 lttng create --snapshot my-session
6110 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6111 <<channel,channels>> created in this mode is automatically set to
6112 _overwrite_ (flight recorder mode).
6114 . Configure the tracing session as usual with the cmd:lttng
6115 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6117 . **Optional**: When you need to take a snapshot, stop tracing.
6119 You can take a snapshot when the tracers are active, but if you stop
6120 them first, you are sure that the data in the sub-buffers does not
6121 change before you actually take the snapshot.
6128 lttng snapshot record --name=my-first-snapshot
6132 LTTng writes the current sub-buffers of all the current tracing
6133 session's channels to trace files on the local file system. Those trace
6134 files have `my-first-snapshot` in their name.
6136 There is no difference between the format of a normal trace file and the
6137 format of a snapshot: viewers of LTTng traces also support LTTng
6140 By default, LTTng writes snapshot files to the path shown by
6141 `lttng snapshot list-output`. You can change this path or decide to send
6142 snapshots over the network using either:
6144 . An output path or URL that you specify when you create the
6146 . An snapshot output path or URL that you add using
6147 `lttng snapshot add-output`
6148 . An output path or URL that you provide directly to the
6149 `lttng snapshot record` command.
6151 Method 3 overrides method 2, which overrides method 1. When you
6152 specify a URL, a relay daemon must listen on a remote system (see
6153 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6158 === Use the machine interface
6160 With any command of the cmd:lttng command-line tool, you can use the
6161 `--mi=xml` argument (before the command name) to get an XML machine
6162 interface output, for example:
6166 lttng --mi=xml enable-event --kernel --syscall open
6169 A schema definition (XSD) is
6170 https://github.com/lttng/lttng-tools/blob/stable-{revision}/src/common/mi_lttng.xsd[available]
6171 to ease the integration with external tools as much as possible.
6175 [[persistent-memory-file-systems]]
6176 === Record trace data on persistent memory file systems
6178 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6179 (NVRAM) is random-access memory that retains its information when power
6180 is turned off (non-volatile). Systems with such memory can store data
6181 structures in RAM and retrieve them after a reboot, without flushing
6182 to typical _storage_.
6184 Linux supports NVRAM file systems thanks to either
6185 http://pramfs.sourceforge.net/[PRAMFS] or
6186 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6187 (requires Linux 4.1+).
6189 This section does not describe how to operate such file systems;
6190 we assume that you have a working persistent memory file system.
6192 When you create a <<tracing-session,tracing session>>, you can specify
6193 the path of the shared memory holding the sub-buffers. If you specify a
6194 location on an NVRAM file system, then you can retrieve the latest
6195 recorded trace data when the system reboots after a crash.
6197 To record trace data on a persistent memory file system and retrieve the
6198 trace data after a system crash:
6200 . Create a tracing session with a sub-buffer shared memory path located
6201 on an NVRAM file system:
6206 lttng create --shm-path=/path/to/shm
6210 . Configure the tracing session as usual with the cmd:lttng
6211 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6213 . After a system crash, use the cmd:lttng-crash command-line tool to
6214 view the trace data recorded on the NVRAM file system:
6219 lttng-crash /path/to/shm
6223 The binary layout of the ring buffer files is not exactly the same as
6224 the trace files layout. This is why you need to use the cmd:lttng-crash
6225 utility instead of your preferred trace viewer directly.
6227 To convert the ring buffer files to LTTng trace files:
6229 * Use the `--extract` option of cmd:lttng-crash:
6234 lttng-crash --extract=/path/to/trace /path/to/shm
6238 See man:lttng-crash(1) for the complete list of command-line options.
6244 This section presents various references for LTTng packages such as
6245 links to online manpages, tables that the rest of the text needs,
6246 descriptions of library functions, and more.
6249 [[online-lttng-manpages]]
6250 === Online noch:{LTTng} manpages
6252 LTTng packages currently install the following link:/man[man pages],
6253 available online using the links below:
6257 ** man:lttng-crash(1)
6258 ** man:lttng-sessiond(8)
6259 ** man:lttng-relayd(8)
6261 ** man:lttng-gen-tp(1)
6263 ** man:lttng-ust-cyg-profile(3)
6264 ** man:lttng-ust-dl(3)
6268 === noch:{LTTng-UST}
6270 This section presents references of the LTTng-UST package.
6274 ==== noch:{LTTng-UST} library (+liblttng‑ust+)
6276 The LTTng-UST library, or `liblttng-ust`, is the main shared object
6277 against which user applications are linked to make LTTng user space
6280 The <<c-application,C application>> guide shows the complete
6281 process to instrument, build and run a C/$$C++$$ application using
6282 LTTng-UST, while this section contains a few important tables.
6285 [[liblttng-ust-tp-fields]]
6286 ===== Tracepoint fields macros (for `TP_FIELDS()`)
6288 The available macros to define tracepoint fields, which you must use
6289 within `TP_FIELDS()` in `TRACEPOINT_EVENT()`, are:
6291 [role="func-desc growable",cols="asciidoc,asciidoc"]
6292 .Available macros to define LTTng-UST tracepoint fields
6294 |Macro |Description and parameters
6297 +ctf_integer(__t__, __n__, __e__)+
6299 +ctf_integer_nowrite(__t__, __n__, __e__)+
6301 Standard integer, displayed in base 10.
6304 Integer C type (`int`, `long`, `size_t`, ...).
6310 Argument expression.
6312 |+ctf_integer_hex(__t__, __n__, __e__)+
6314 Standard integer, displayed in base 16.
6323 Argument expression.
6325 |+ctf_integer_network(__t__, __n__, __e__)+
6327 Integer in network byte order (big-endian), displayed in base 10.
6336 Argument expression.
6338 |+ctf_integer_network_hex(__t__, __n__, __e__)+
6340 Integer in network byte order, displayed in base 16.
6349 Argument expression.
6352 +ctf_float(__t__, __n__, __e__)+
6354 +ctf_float_nowrite(__t__, __n__, __e__)+
6356 Floating point number.
6359 Floating point number C type (`float` or `double`).
6365 Argument expression.
6368 +ctf_string(__n__, __e__)+
6370 +ctf_string_nowrite(__n__, __e__)+
6372 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6378 Argument expression.
6381 +ctf_array(__t__, __n__, __e__, __s__)+
6383 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6385 Statically-sized array of integers
6388 Array element C type.
6394 Argument expression.
6400 +ctf_array_text(__t__, __n__, __e__, __s__)+
6402 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6404 Statically-sized array, printed as text.
6406 The string does not need to be null-terminated.
6409 Array element C type (always `char`).
6415 Argument expression.
6421 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6423 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6425 Dynamically-sized array of integers.
6427 The type of +__E__+ must be unsigned.
6430 Array element C type.
6436 Argument expression.
6439 Length expression C type.
6445 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6447 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6449 Dynamically-sized array, displayed as text.
6451 The string does not need to be null-terminated.
6453 The type of +__E__+ must be unsigned.
6455 The behaviour is undefined if +__e__+ is `NULL`.
6458 Sequence element C type (always `char`).
6464 Argument expression.
6467 Length expression C type.
6473 The `_nowrite` versions omit themselves from the session trace, but are
6474 otherwise identical. This means the tracer does not write the `_nowrite`
6475 fields to the trace. Their primary purpose is to make some of the event
6476 context available to the <<enabling-disabling-events,event filters>>
6477 without having to commit the data to sub-buffers.
6480 [[liblttng-ust-tracepoint-loglevel]]
6481 ===== Tracepoint log levels (for `TRACEPOINT_LOGLEVEL()`)
6483 The following table shows the available log level values for the
6484 `TRACEPOINT_LOGLEVEL()` macro:
6490 Action must be taken immediately.
6493 Critical conditions.
6502 Normal, but significant, condition.
6505 Informational message.
6507 `TRACE_DEBUG_SYSTEM`::
6508 Debug information with system-level scope (set of programs).
6510 `TRACE_DEBUG_PROGRAM`::
6511 Debug information with program-level scope (set of processes).
6513 `TRACE_DEBUG_PROCESS`::
6514 Debug information with process-level scope (set of modules).
6516 `TRACE_DEBUG_MODULE`::
6517 Debug information with module (executable/library) scope (set of units).
6519 `TRACE_DEBUG_UNIT`::
6520 Debug information with compilation unit scope (set of functions).
6522 `TRACE_DEBUG_FUNCTION`::
6523 Debug information with function-level scope.
6525 `TRACE_DEBUG_LINE`::
6526 Debug information with line-level scope (TRACEPOINT_EVENT default).
6529 Debug-level message.
6531 Log levels `TRACE_EMERG` through `TRACE_INFO` and `TRACE_DEBUG` match
6532 http://man7.org/linux/man-pages/man3/syslog.3.html[syslog]
6533 level semantics. Log levels `TRACE_DEBUG_SYSTEM` through `TRACE_DEBUG`
6534 offer more fine-grained selection of debug information.
6537 [[lttng-modules-ref]]
6538 === noch:{LTTng-modules}
6540 This section presents references of the LTTng-modules package.
6544 [[lttng-modules-tp-fields]]
6545 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6547 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6548 tracepoint fields, which must be listed within `TP_FIELDS()` in
6549 `LTTNG_TRACEPOINT_EVENT()`, are:
6551 [role="func-desc growable",cols="asciidoc,asciidoc"]
6552 .Available macros to define LTTng-modules tracepoint fields
6554 |Macro |Description and parameters
6557 +ctf_integer(__t__, __n__, __e__)+
6559 +ctf_integer_nowrite(__t__, __n__, __e__)+
6561 +ctf_user_integer(__t__, __n__, __e__)+
6563 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6565 Standard integer, displayed in base 10.
6568 Integer C type (`int`, `long`, `size_t`, ...).
6574 Argument expression.
6577 +ctf_integer_hex(__t__, __n__, __e__)+
6579 +ctf_user_integer_hex(__t__, __n__, __e__)+
6581 Standard integer, displayed in base 16.
6590 Argument expression.
6592 |+ctf_integer_oct(__t__, __n__, __e__)+
6594 Standard integer, displayed in base 8.
6603 Argument expression.
6606 +ctf_integer_network(__t__, __n__, __e__)+
6608 +ctf_user_integer_network(__t__, __n__, __e__)+
6610 Integer in network byte order (big-endian), displayed in base 10.
6619 Argument expression.
6622 +ctf_integer_network_hex(__t__, __n__, __e__)+
6624 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6626 Integer in network byte order, displayed in base 16.
6635 Argument expression.
6638 +ctf_string(__n__, __e__)+
6640 +ctf_string_nowrite(__n__, __e__)+
6642 +ctf_user_string(__n__, __e__)+
6644 +ctf_user_string_nowrite(__n__, __e__)+
6646 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6652 Argument expression.
6655 +ctf_array(__t__, __n__, __e__, __s__)+
6657 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6659 +ctf_user_array(__t__, __n__, __e__, __s__)+
6661 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6663 Statically-sized array of integers
6666 Array element C type.
6672 Argument expression.
6678 +ctf_array_text(__t__, __n__, __e__, __s__)+
6680 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6682 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6684 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6686 Statically-sized array, printed as text.
6688 The string does not need to be null-terminated.
6691 Array element C type (always `char`).
6697 Argument expression.
6703 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6705 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6707 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6709 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6711 Dynamically-sized array of integers.
6713 The type of +__E__+ must be unsigned.
6716 Array element C type.
6722 Argument expression.
6725 Length expression C type.
6730 |+ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6732 Dynamically-sized array of integers, displayed in base 16.
6734 The type of +__E__+ must be unsigned.
6737 Array element C type.
6743 Argument expression.
6746 Length expression C type.
6751 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
6753 Dynamically-sized array of integers in network byte order (big-endian),
6754 displayed in base 10.
6756 The type of +__E__+ must be unsigned.
6759 Array element C type.
6765 Argument expression.
6768 Length expression C type.
6774 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6776 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6778 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
6780 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
6782 Dynamically-sized array, displayed as text.
6784 The string does not need to be null-terminated.
6786 The type of +__E__+ must be unsigned.
6788 The behaviour is undefined if +__e__+ is `NULL`.
6791 Sequence element C type (always `char`).
6797 Argument expression.
6800 Length expression C type.
6806 Use the `_user` versions when the argument expression, `e`, is
6807 a user space address. In the cases of `ctf_user_integer*()` and
6808 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
6811 The `_nowrite` versions omit themselves from the session trace, but are
6812 otherwise identical. This means the `_nowrite` fields won't be written
6813 in the recorded trace. Their primary purpose is to make some
6814 of the event context available to the
6815 <<enabling-disabling-events,event filters>> without having to
6816 commit the data to sub-buffers.
6822 Terms related to LTTng and to tracing in general:
6825 The http://diamon.org/babeltrace[Babeltrace] project, which includes
6826 the cmd:babeltrace command, some libraries, and Python bindings.
6828 <<channel-buffering-schemes,buffering scheme>>::
6829 A layout of sub-buffers applied to a given channel.
6831 <<channel,channel>>::
6832 An entity which is responsible for a set of ring buffers.
6834 <<event,Event rules>> are always attached to a specific channel.
6837 A reference of time for a tracer.
6839 <<lttng-consumerd,consumer daemon>>::
6840 A process which is responsible for consuming the full sub-buffers
6841 and write them to a file system or send them over the network.
6843 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
6844 mode in which the tracer _discards_ new event records when there's no
6845 sub-buffer space left to store them.
6848 The consequence of the execution of an instrumentation
6849 point, like a tracepoint that you manually place in some source code,
6850 or a Linux kernel KProbe.
6852 An event is said to _occur_ at a specific time. Different actions can
6853 be taken upon the occurance of an event, like record the event's payload
6856 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
6857 The mechanism by which event records of a given channel are lost
6858 (not recorded) when there is no sub-buffer space left to store them.
6860 [[def-event-name]]event name::
6861 The name of an event, which is also the name of the event record.
6862 This is also called the _instrumentation point name_.
6865 A record, in a trace, of the payload of an event which occured.
6867 <<event,event rule>>::
6868 Set of conditions which must be satisfied for one or more occuring
6869 events to be recorded.
6871 `java.util.logging`::
6873 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
6875 <<instrumenting,instrumentation>>::
6876 The use of LTTng probes to make a piece of software traceable.
6878 instrumentation point::
6879 A point in the execution path of a piece of software that, when
6880 reached by this execution, can emit an event.
6882 instrumentation point name::
6883 See _<<def-event-name,event name>>_.
6886 A http://logging.apache.org/log4j/1.2/[logging library] for Java
6887 developed by the Apache Software Foundation.
6890 Level of severity of a log statement or user space
6891 instrumentation point.
6894 The _Linux Trace Toolkit: next generation_ project.
6896 <<lttng-cli,cmd:lttng>>::
6897 A command-line tool provided by the LTTng-tools project which you
6898 can use to send and receive control messages to and from a
6902 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
6903 which is a set of analyzing programs that are used to obtain a
6904 higher level view of an LTTng trace.
6906 cmd:lttng-consumerd::
6907 The name of the consumer daemon program.
6910 A utility provided by the LTTng-tools project which can convert
6911 ring buffer files (usually
6912 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
6915 LTTng Documentation::
6918 <<lttng-live,LTTng live>>::
6919 A communication protocol between the relay daemon and live viewers
6920 which makes it possible to see events "live", as they are received by
6923 <<lttng-modules,LTTng-modules>>::
6924 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
6925 which contains the Linux kernel modules to make the Linux kernel
6926 instrumentation points available for LTTng tracing.
6929 The name of the relay daemon program.
6931 cmd:lttng-sessiond::
6932 The name of the session daemon program.
6935 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
6936 contains the various programs and libraries used to
6937 <<controlling-tracing,control tracing>>.
6939 <<lttng-ust,LTTng-UST>>::
6940 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
6941 contains libraries to instrument user applications.
6943 <<lttng-ust-agents,LTTng-UST Java agent>>::
6944 A Java package provided by the LTTng-UST project to allow the
6945 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
6948 <<lttng-ust-agents,LTTng-UST Python agent>>::
6949 A Python package provided by the LTTng-UST project to allow the
6950 LTTng instrumentation of Python logging statements.
6952 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
6953 The event loss mode in which new event records overwrite older
6954 event records when there's no sub-buffer space left to store them.
6956 <<channel-buffering-schemes,per-process buffering>>::
6957 A buffering scheme in which each instrumented process has its own
6958 sub-buffers for a given user space channel.
6960 <<channel-buffering-schemes,per-user buffering>>::
6961 A buffering scheme in which all the processes of a Unix user share the
6962 same sub-buffer for a given user space channel.
6964 <<lttng-relayd,relay daemon>>::
6965 A process which is responsible for receiving the trace data sent by
6966 a distant consumer daemon.
6969 A set of sub-buffers.
6971 <<lttng-sessiond,session daemon>>::
6972 A process which receives control commands from you and orchestrates
6973 the tracers and various LTTng daemons.
6975 <<taking-a-snapshot,snapshot>>::
6976 A copy of the current data of all the sub-buffers of a given tracing
6977 session, saved as trace files.
6980 One part of an LTTng ring buffer which contains event records.
6983 The time information attached to an event when it is emitted.
6986 A set of files which are the concatenations of one or more
6987 flushed sub-buffers.
6990 The action of recording the events emitted by an application
6991 or by a system, or to initiate such recording by controlling
6995 The http://tracecompass.org[Trace Compass] project and application.
6998 An instrumentation point using the tracepoint mechanism of the Linux
6999 kernel or of LTTng-UST.
7001 tracepoint definition::
7002 The definition of a single tracepoint.
7005 The name of a tracepoint.
7007 tracepoint provider::
7008 A set of functions providing tracepoints to an instrumented user
7011 Not to be confused with a _tracepoint provider package_: many tracepoint
7012 providers can exist within a tracepoint provider package.
7014 tracepoint provider package::
7015 One or more tracepoint providers compiled as an object file or as
7019 A software which records emitted events.
7021 <<domain,tracing domain>>::
7022 A namespace for event sources.
7025 The Unix group in which a Unix user can be to be allowed to trace the
7028 <<tracing-session,tracing session>>::
7029 A stateful dialogue between you and a <<lttng-sessiond,session
7033 An application running in user space, as opposed to a Linux kernel
7034 module, for example.