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 Skip this section if you’re familiar with software tracing and with the
28 * ``**<<installing-lttng,Installation>>**'' describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 Skip this section if you already properly installed LTTng on your target
35 * ``**<<getting-started,Quick start>>**'' is a concise guide to
36 get 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 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
50 components of the LTTng machinery, like the daemons, the libraries,
51 and the command-line interface.
53 * ``**<<instrumenting,Instrumentation>>**'' shows different ways to
54 instrument user applications and the Linux kernel for LTTng tracing.
56 Instrumenting source code is essential to provide a meaningful
59 Skip this section if you don't have a programming background.
61 * ``**<<controlling-tracing,Tracing control>>**'' is divided into topics
62 which demonstrate how to use the vast array of features that
63 LTTng{nbsp}{revision} offers.
65 * ``**<<reference,Reference>>**'' contains API reference tables.
67 * ``**<<glossary,Glossary>>**'' is a specialized dictionary of terms
68 related to LTTng or to the field of software tracing.
71 include::../common/convention.txt[]
74 include::../common/acknowledgements.txt[]
78 == What's new in LTTng{nbsp}{revision}?
80 LTTng{nbsp}{revision} bears the name _Nordicité_, the product of a
81 collaboration between https://champlibre.co/[Champ Libre] and
82 https://www.boreale.com/[Boréale]. This farmhouse IPA is brewed with
83 https://en.wikipedia.org/wiki/Kveik[Kveik] yeast and Québec-grown
84 barley, oats, and juniper branches. The result is a remarkable, fruity,
85 hazy golden IPA that offers a balanced touch of resinous and woodsy
88 New features and changes in LTTng{nbsp}{revision}:
92 * The LTTng trigger API of <<liblttng-ctl-lttng,`liblttng-ctl`>> now
93 offers the ``__event rule matches__'' condition (an <<event-rule,event
94 rule>> matches an event) as well as the following new actions:
97 * <<basic-tracing-session-control,Start or stop>> a recording session.
98 * <<session-rotation,Archive the current trace chunk>> of a
99 recording session (rotate).
100 * <<taking-a-snapshot,Take a snapshot>> of a recording session.
103 As a reminder, a <<trigger,trigger>> is a condition-actions pair. When
104 the condition of a trigger is satisfied, LTTng attempts to execute its
107 This feature is also available with the new man:lttng-add-trigger(1),
108 man:lttng-remove-trigger(1), and man:lttng-list-triggers(1)
109 <<lttng-cli,cmd:lttng>> commands.
111 Starting from LTTng{nbsp}{revision}, a trigger may have more than one
114 See “<<add-event-rule-matches-trigger,Add an ``event rule matches''
115 trigger to a session daemon>>” to learn more.
117 * The LTTng <<lttng-ust,user space>> and <<lttng-modules,kernel>>
118 tracers offer the new namespace context field `time_ns`, which is the
119 inode number, in the proc file system, of the current clock namespace.
121 See man:lttng-add-context(1), man:lttng-ust(3), and
122 man:time_namespaces(7).
124 * The link:/man[manual pages] of LTTng-tools now have a terminology and
125 style which match the LTTng Documentation, many fixes, more internal
126 and manual page links, clearer lists and procedures, superior
127 consistency, and usage examples.
129 The new man:lttng-event-rule(7) manual page explains the new, common
130 way to specify an event rule on the command line.
132 The new man:lttng-concepts(7) manual page explains the core concepts of
133 LTTng. Its contents is essentially the ``<<core-concepts,Core
134 concepts>>'' section of this documentation, but more adapted to the
141 The major version part of the `liblttng-ust`
142 https://en.wikipedia.org/wiki/Soname[soname] is bumped, which means you
143 **must recompile** your instrumented applications/libraries and
144 <<tracepoint-provider,tracepoint provider packages>> to use
145 LTTng-UST{nbsp}{revision}.
147 This change became a necessity to clean up the library and for
148 `liblttng-ust` to stop exporting private symbols.
150 Also, LTTng{nbsp}{revision} prepends the `lttng_ust_` and `LTTNG_UST_`
151 prefix to all public macro/definition/function names to offer a
152 consistent API namespace. The LTTng{nbsp}2.12 API is still available;
153 see the ``Compatibility with previous APIs'' section of
157 Other notable changes:
159 * The `liblttng-ust` C{nbsp}API offers the new man:lttng_ust_vtracef(3)
160 and man:lttng_ust_vtracelog(3) macros which are to
161 man:lttng_ust_tracef(3) and man:lttng_ust_tracelog(3) what
162 man:vprintf(3) is to man:printf(3).
164 * LTTng-UST now only depends on https://liburcu.org/[`liburcu`] at build
165 time, not at run time.
169 * The preferred display base of event record integer fields which
170 contain memory addresses is now hexadecimal instead of decimal.
172 * The `pid` field is removed from `lttng_statedump_file_descriptor`
173 event records and the `file_table_address` field is added.
175 This new field is the address of the `files_struct` structure which
176 contains the file descriptor.
179 ``https://github.com/lttng/lttng-modules/commit/e7a0ca7205fd4be7c829d171baa8823fe4784c90[statedump: introduce `file_table_address`]''
182 * The `flags` field of `syscall_entry_clone` event records is now a
183 structure containing two enumerations (exit signal and options).
185 This change makes the flag values more readable and meaningful.
188 ``https://github.com/lttng/lttng-modules/commit/d775625e2ba4825b73b5897e7701ad6e2bdba115[syscalls: Make `clone()`'s `flags` field a 2 enum struct]''
191 * The memory footprint of the kernel tracer is improved: the latter only
192 generates metadata for the specific system call recording event rules
193 that you <<enabling-disabling-events,create>>.
199 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
200 generation_ is a modern toolkit for tracing Linux systems and
201 applications. So your first question might be:
208 As the history of software engineering progressed and led to what
209 we now take for granted--complex, numerous and
210 interdependent software applications running in parallel on
211 sophisticated operating systems like Linux--the authors of such
212 components, software developers, began feeling a natural
213 urge to have tools that would ensure the robustness and good performance
214 of their masterpieces.
216 One major achievement in this field is, inarguably, the
217 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
218 an essential tool for developers to find and fix bugs. But even the best
219 debugger won't help make your software run faster, and nowadays, faster
220 software means either more work done by the same hardware, or cheaper
221 hardware for the same work.
223 A _profiler_ is often the tool of choice to identify performance
224 bottlenecks. Profiling is suitable to identify _where_ performance is
225 lost in a given piece of software. The profiler outputs a profile, a
226 statistical summary of observed events, which you may use to discover
227 which functions took the most time to execute. However, a profiler won't
228 report _why_ some identified functions are the bottleneck. Bottlenecks
229 might only occur when specific conditions are met, conditions that are
230 sometimes impossible to capture by a statistical profiler, or impossible
231 to reproduce with an application altered by the overhead of an
232 event-based profiler. For a thorough investigation of software
233 performance issues, a history of execution is essential, with the
234 recorded values of variables and context fields you choose, and with as
235 little influence as possible on the instrumented application. This is
236 where tracing comes in handy.
238 _Tracing_ is a technique used to understand what goes on in a running
239 software system. The piece of software used for tracing is called a
240 _tracer_, which is conceptually similar to a tape recorder. When
241 recording, specific instrumentation points placed in the software source
242 code generate events that are saved on a giant tape: a _trace_ file. You
243 can record user application and operating system events at the same
244 time, opening the possibility of resolving a wide range of problems that
245 would otherwise be extremely challenging.
247 Tracing is often compared to _logging_. However, tracers and loggers are
248 two different tools, serving two different purposes. Tracers are
249 designed to record much lower-level events that occur much more
250 frequently than log messages, often in the range of thousands per
251 second, with very little execution overhead. Logging is more appropriate
252 for a very high-level analysis of less frequent events: user accesses,
253 exceptional conditions (errors and warnings, for example), database
254 transactions, instant messaging communications, and such. Simply put,
255 logging is one of the many use cases that can be satisfied with tracing.
257 The list of recorded events inside a trace file can be read manually
258 like a log file for the maximum level of detail, but it's generally
259 much more interesting to perform application-specific analyses to
260 produce reduced statistics and graphs that are useful to resolve a
261 given problem. Trace viewers and analyzers are specialized tools
264 In the end, this is what LTTng is: a powerful, open source set of
265 tools to trace the Linux kernel and user applications at the same time.
266 LTTng is composed of several components actively maintained and
267 developed by its link:/community/#where[community].
270 [[lttng-alternatives]]
271 === Alternatives to noch:{LTTng}
273 Excluding proprietary solutions, a few competing software tracers
276 https://github.com/dtrace4linux/linux[dtrace4linux]::
277 A port of Sun Microsystems' DTrace to Linux.
279 The cmd:dtrace tool interprets user scripts and is responsible for
280 loading code into the Linux kernel for further execution and collecting
283 https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF]::
284 A subsystem in the Linux kernel in which a virtual machine can
285 execute programs passed from the user space to the kernel.
287 You can attach such programs to tracepoints and kprobes thanks to a
288 system call, and they can output data to the user space when executed
289 thanks to different mechanisms (pipe, VM register values, and eBPF maps,
292 https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]::
293 The de facto function tracer of the Linux kernel.
295 Its user interface is a set of special files in sysfs.
297 https://perf.wiki.kernel.org/[perf]::
298 A performance analysis tool for Linux which supports hardware
299 performance counters, tracepoints, as well as other counters and
302 The controlling utility of perf is the cmd:perf command line/text UI
305 https://linux.die.net/man/1/strace[strace]::
306 A command-line utility which records system calls made by a
307 user process, as well as signal deliveries and changes of process
310 strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace] to
311 fulfill its function.
313 https://www.sysdig.org/[sysdig]::
314 Like SystemTap, uses scripts to analyze Linux kernel events.
316 You write scripts, or _chisels_ in the jargon of sysdig, in Lua and
317 sysdig executes them while it traces the system or afterwards. The
318 interface of sysdig is the cmd:sysdig command-line tool as well as the
319 text UI-based cmd:csysdig tool.
321 https://sourceware.org/systemtap/[SystemTap]::
322 A Linux kernel and user space tracer which uses custom user scripts
323 to produce plain text traces.
325 SystemTap converts the scripts to the C language, and then compiles them
326 as Linux kernel modules which are loaded to produce trace data. The
327 primary user interface of SystemTap is the cmd:stap command-line tool.
329 The main distinctive features of LTTng is that it produces correlated
330 kernel and user space traces, as well as doing so with the lowest
331 overhead amongst other solutions. It produces trace files in the
332 https://diamon.org/ctf[CTF] format, a file format optimized
333 for the production and analyses of multi-gigabyte data.
335 LTTng is the result of more than 10{nbsp}years of active open source
336 development by a community of passionate developers. LTTng is currently
337 available on major desktop and server Linux distributions.
339 The main interface for tracing control is a single command-line tool
340 named cmd:lttng. The latter can create several recording sessions, enable
341 and disable recording event rules on the fly, filter events efficiently
342 with custom user expressions, start and stop tracing, and much more.
343 LTTng can write the traces on the file system or send them over the
344 network, and keep them totally or partially. You can make LTTng execute
345 user-defined actions when LTTng emits an event. You can view the traces
346 once tracing becomes inactive or as LTTng records events.
348 <<installing-lttng,Install LTTng now>> and
349 <<getting-started,start tracing>>!
355 **LTTng** is a set of software <<plumbing,components>> which interact to
356 <<instrumenting,instrument>> the Linux kernel and user applications, and
357 to <<controlling-tracing,control tracing>> (start and stop
358 recording, create recording event rules, and the rest). Those
359 components are bundled into the following packages:
362 Libraries and command-line interface to control tracing.
365 Linux kernel modules to instrument and trace the kernel.
368 Libraries and Java/Python packages to instrument and trace user
371 Most distributions mark the LTTng-modules and LTTng-UST packages as
372 optional when installing LTTng-tools (which is always required). In the
373 following sections, we always provide the steps to install all three,
376 * You only need to install LTTng-modules if you intend to use
377 the Linux kernel LTTng tracer.
379 * You only need to install LTTng-UST if you intend to use the user
383 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 17{nbsp}October{nbsp}2023.
386 |Distribution |Available in releases
388 |https://www.ubuntu.com/[Ubuntu]
389 |xref:ubuntu[Ubuntu 22.04 LTS _Jammy Jellyfish_, Ubuntu 23.04 _Lunar Lobster_, and Ubuntu 23.10 _Mantic Minotaur_].
391 Ubuntu{nbsp}18.04 LTS _Bionic Beaver_ and Ubuntu{nbsp}20.04 LTS _Focal Fossa_:
392 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
394 |https://www.debian.org/[Debian]
395 |<<debian,Debian{nbsp}12 _bookworm_>>.
397 |https://getfedora.org/[Fedora]
398 |xref:fedora[Fedora{nbsp}37, Fedora{nbsp}38, and Fedora{nbsp}39].
400 |https://www.archlinux.org/[Arch Linux]
401 |<<arch-linux,_extra_ repository and AUR>>.
403 |https://alpinelinux.org/[Alpine Linux]
404 |xref:alpine-linux[Alpine Linux{nbsp}3.16, Alpine Linux{nbsp}3.17, and Alpine Linux{nbsp}3.18].
406 |https://buildroot.org/[Buildroot]
407 |xref:buildroot[Buildroot{nbsp}2022.02, Buildroot{nbsp}2022.05,
408 Buildroot{nbsp}2022.08, Buildroot{nbsp}2022.11, Buildroot{nbsp}2023.02,
409 Buildroot{nbsp}2023.05, and Buildroot{nbsp}2023.08].
411 |https://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
412 https://www.yoctoproject.org/[Yocto]
413 |xref:oe-yocto[Yocto Project{nbsp}3.3 _Honister_, Yocto Project{nbsp}4.0 _Kirkstone_,
414 Yocto Project{nbsp}4.1 _Langdale_, Yocto Project{nbsp}4.2 _Mickledore_, and
415 Yocto Project{nbsp}4.3 _Nanbield_].
421 For https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
422 packages, see https://packages.efficios.com/[EfficiOS Enterprise
425 For other distributions, <<building-from-source,build LTTng from
430 === [[ubuntu-official-repository]]Ubuntu
432 LTTng{nbsp}{revision} is available on Ubuntu 22.04 LTS _Jammy Jellyfish_, Ubuntu 23.04 _Lunar Lobster_, and Ubuntu 23.10 _Mantic Minotaur_. For previous supported releases of Ubuntu, <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
434 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}22.04 LTS _Jammy Jellyfish_:
436 . Install the main LTTng{nbsp}{revision} packages:
441 # apt-get install lttng-tools
442 # apt-get install lttng-modules-dkms
443 # apt-get install liblttng-ust-dev
447 . **If you need to instrument and trace <<java-application,Java applications>>**,
448 install the LTTng-UST Java agent:
453 # apt-get install liblttng-ust-agent-java
457 . **If you need to instrument and trace <<python-application,Python{nbsp}3
458 applications>>**, install the LTTng-UST Python agent:
463 # apt-get install python3-lttngust
468 === Ubuntu: noch:{LTTng} Stable {revision} PPA
470 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
471 Stable{nbsp}{revision} PPA] offers the latest stable LTTng{nbsp}{revision}
472 packages for Ubuntu{nbsp}18.04 LTS _Bionic Beaver_, Ubuntu{nbsp}20.04 LTS _Focal Fossa_,
473 and Ubuntu{nbsp}22.04 LTS _Jammy Jellyfish_.
475 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
478 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
482 [role="term",subs="attributes"]
484 # apt-add-repository ppa:lttng/stable-{revision}
489 . Install the main LTTng{nbsp}{revision} packages:
494 # apt-get install lttng-tools
495 # apt-get install lttng-modules-dkms
496 # apt-get install liblttng-ust-dev
500 . **If you need to instrument and trace
501 <<java-application,Java applications>>**, install the LTTng-UST
507 # apt-get install liblttng-ust-agent-java
511 . **If you need to instrument and trace
512 <<python-application,Python{nbsp}3 applications>>**, install the
513 LTTng-UST Python agent:
518 # apt-get install python3-lttngust
525 To install LTTng{nbsp}{revision} on Debian{nbsp}12 _bookworm_:
527 . Install the main LTTng{nbsp}{revision} packages:
532 # apt install lttng-modules-dkms
533 # apt install liblttng-ust-dev
534 # apt install lttng-tools
538 . **If you need to instrument and trace <<java-application,Java
539 applications>>**, install the LTTng-UST Java agent:
544 # apt install liblttng-ust-agent-java
548 . **If you need to instrument and trace <<python-application,Python
549 applications>>**, install the LTTng-UST Python agent:
554 # apt install python3-lttngust
561 To install LTTng{nbsp}{revision} on Fedora{nbsp}37, Fedora{nbsp}38, or
564 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
570 # yum install lttng-tools
571 # yum install lttng-ust
575 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
578 [role="term",subs="attributes,specialcharacters"]
581 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-{revision}.tar.bz2 &&
582 tar -xf lttng-modules-latest-{revision}.tar.bz2 &&
583 cd lttng-modules-{revision}.* &&
585 sudo make modules_install &&
591 .Java and Python application instrumentation and tracing
593 If you need to instrument and trace <<java-application,Java
594 applications>> on Fedora, you need to build and install
595 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
596 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
597 `--enable-java-agent-all` options to the `configure` script, depending
598 on which Java logging framework you use.
600 If you need to instrument and trace <<python-application,Python
601 applications>> on Fedora, you need to build and install
602 LTTng-UST{nbsp}{revision} from source and pass the
603 `--enable-python-agent` option to the `configure` script.
610 LTTng-UST{nbsp}{revision} is available in the _extra_
611 repository of Arch Linux, while LTTng-tools{nbsp}{revision} and
612 LTTng-modules{nbsp}{revision} are available in the
613 https://aur.archlinux.org/[AUR].
615 To install LTTng{nbsp}{revision} on Arch Linux, using
616 https://github.com/Jguer/yay[yay] for the AUR packages:
618 . Install the main LTTng{nbsp}{revision} packages:
623 # pacman -Sy lttng-ust
624 $ yay -Sy lttng-tools
625 $ yay -Sy lttng-modules
629 . **If you need to instrument and trace <<python-application,Python
630 applications>>**, install the LTTng-UST Python agent:
635 # pacman -Sy python-lttngust
643 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
644 Alpine Linux{nbsp}3.16, Alpine Linux{nbsp}3.17, or Alpine Linux{nbsp}3.18:
646 . Add the LTTng packages:
651 # apk add lttng-tools
652 # apk add lttng-ust-dev
656 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
659 [role="term",subs="attributes,specialcharacters"]
662 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-{revision}.tar.bz2 &&
663 tar -xf lttng-modules-latest-{revision}.tar.bz2 &&
664 cd lttng-modules-{revision}.* &&
666 sudo make modules_install &&
675 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2022.02, Buildroot{nbsp}2022.05,
676 Buildroot{nbsp}2022.08, Buildroot{nbsp}2022.11, Buildroot{nbsp}2023.02,
677 Buildroot{nbsp}2023.05, or Buildroot{nbsp}2023.08:
679 . Launch the Buildroot configuration tool:
688 . In **Kernel**, check **Linux kernel**.
689 . In **Toolchain**, check **Enable WCHAR support**.
690 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
691 check **lttng-modules** and **lttng-tools**.
692 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
693 **Other**, check **lttng-libust**.
697 === OpenEmbedded and Yocto
699 LTTng{nbsp}{revision} recipes are available in the
700 https://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
701 layer for Yocto Project{nbsp}3.3 _Honister_, Yocto Project{nbsp}4.0 _Kirkstone_,
702 Yocto Project{nbsp}4.1 _Langdale_, Yocto Project{nbsp}4.2 _Mickledore_, and
703 Yocto Project{nbsp}4.3 _Nanbield_ under the following names:
709 With BitBake, the simplest way to include LTTng recipes in your target
710 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
713 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
718 . Select a machine and an image recipe.
719 . Click **Edit image recipe**.
720 . Under the **All recipes** tab, search for **lttng**.
721 . Check the desired LTTng recipes.
724 [[building-from-source]]
725 === Build from source
727 To build and install LTTng{nbsp}{revision} from source:
729 . Using the package manager of your distribution, or from source,
730 install the following dependencies of LTTng-tools and LTTng-UST:
733 * https://sourceforge.net/projects/libuuid/[libuuid]
734 * https://directory.fsf.org/wiki/Popt[popt]
735 * https://liburcu.org/[Userspace RCU]
736 * http://www.xmlsoft.org/[libxml2]
737 * **Optional**: https://github.com/numactl/numactl[numactl]
740 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
746 wget https://lttng.org/files/lttng-modules/lttng-modules-latest-2.13.tar.bz2 &&
747 tar -xf lttng-modules-latest-2.13.tar.bz2 &&
748 cd lttng-modules-2.13.* &&
750 sudo make modules_install &&
755 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
761 wget https://lttng.org/files/lttng-ust/lttng-ust-latest-2.13.tar.bz2 &&
762 tar -xf lttng-ust-latest-2.13.tar.bz2 &&
763 cd lttng-ust-2.13.* &&
771 Add `--disable-numa` to `./configure` if you don't have
772 https://github.com/numactl/numactl[numactl].
776 .Java and Python application tracing
778 If you need to instrument and have LTTng trace <<java-application,Java
779 applications>>, pass the `--enable-java-agent-jul`,
780 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
781 `configure` script, depending on which Java logging framework you use.
783 If you need to instrument and have LTTng trace
784 <<python-application,Python applications>>, pass the
785 `--enable-python-agent` option to the `configure` script. You can set
786 the env:PYTHON environment variable to the path to the Python interpreter
787 for which to install the LTTng-UST Python agent package.
794 By default, LTTng-UST libraries are installed to
795 dir:{/usr/local/lib}, which is the de facto directory in which to
796 keep self-compiled and third-party libraries.
798 When <<building-tracepoint-providers-and-user-application,linking an
799 instrumented user application with `liblttng-ust`>>:
801 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
804 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
805 man:gcc(1), man:g++(1), or man:clang(1).
809 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
815 wget https://lttng.org/files/lttng-tools/lttng-tools-latest-2.13.tar.bz2 &&
816 tar -xf lttng-tools-latest-2.13.tar.bz2 &&
817 cd lttng-tools-2.13.* &&
825 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
826 previous steps automatically for a given version of LTTng and confine
827 the installed files to a specific directory. This can be useful to try
828 LTTng without installing it on your system.
834 This is a short guide to get started quickly with LTTng kernel and user
837 Before you follow this guide, make sure to <<installing-lttng,install>>
840 This tutorial walks you through the steps to:
842 . <<tracing-the-linux-kernel,Record Linux kernel events>>.
844 . <<tracing-your-own-user-application,Record the events of a user
845 application>> written in C.
847 . <<viewing-and-analyzing-your-traces,View and analyze the
851 [[tracing-the-linux-kernel]]
852 === Record Linux kernel events
854 NOTE: The following command lines start with the `#` prompt because you
855 need root privileges to control the Linux kernel LTTng tracer. You can
856 also control the kernel tracer as a regular user if your Unix user is a
857 member of the <<tracing-group,tracing group>>.
859 . Create a <<tracing-session,recording session>> to write LTTng traces
860 to dir:{/tmp/my-kernel-trace}:
865 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
869 . List the available kernel tracepoints and system calls:
874 # lttng list --kernel
875 # lttng list --kernel --syscall
879 . Create <<event,recording event rules>> which match events having
880 the desired names, for example the `sched_switch` and
881 `sched_process_fork` tracepoints, and the man:open(2) and man:close(2)
887 # lttng enable-event --kernel sched_switch,sched_process_fork
888 # lttng enable-event --kernel --syscall open,close
892 Create a recording event rule which matches _all_ the Linux kernel
893 tracepoint events with the opt:lttng-enable-event(1):--all option
894 (recording with such a recording event rule generates a lot of data):
899 # lttng enable-event --kernel --all
903 . <<basic-tracing-session-control,Start recording>>:
912 . Do some operation on your system for a few seconds. For example,
913 load a website, or list the files of a directory.
915 . <<creating-destroying-tracing-sessions,Destroy>> the current
925 The man:lttng-destroy(1) command doesn't destroy the trace data; it
926 only destroys the state of the recording session.
928 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
929 implicitly (see ``<<basic-tracing-session-control,Start and stop a
930 recording session>>''). You need to stop recording to make LTTng flush
931 the remaining trace data and make the trace readable.
933 . For the sake of this example, make the recorded trace accessible to
939 # chown -R $(whoami) /tmp/my-kernel-trace
943 See ``<<viewing-and-analyzing-your-traces,View and analyze the
944 recorded events>>'' to view the recorded events.
947 [[tracing-your-own-user-application]]
948 === Record user application events
950 This section walks you through a simple example to record the events of
951 a _Hello world_ program written in{nbsp}C.
953 To create the traceable user application:
955 . Create the tracepoint provider header file, which defines the
956 tracepoints and the events they can generate:
962 #undef LTTNG_UST_TRACEPOINT_PROVIDER
963 #define LTTNG_UST_TRACEPOINT_PROVIDER hello_world
965 #undef LTTNG_UST_TRACEPOINT_INCLUDE
966 #define LTTNG_UST_TRACEPOINT_INCLUDE "./hello-tp.h"
968 #if !defined(_HELLO_TP_H) || defined(LTTNG_UST_TRACEPOINT_HEADER_MULTI_READ)
971 #include <lttng/tracepoint.h>
973 LTTNG_UST_TRACEPOINT_EVENT(
978 char *, my_string_arg
981 lttng_ust_field_string(my_string_field, my_string_arg)
982 lttng_ust_field_integer(int, my_integer_field, my_integer_arg)
986 #endif /* _HELLO_TP_H */
988 #include <lttng/tracepoint-event.h>
992 . Create the tracepoint provider package source file:
998 #define LTTNG_UST_TRACEPOINT_CREATE_PROBES
999 #define LTTNG_UST_TRACEPOINT_DEFINE
1001 #include "hello-tp.h"
1005 . Build the tracepoint provider package:
1010 $ gcc -c -I. hello-tp.c
1014 . Create the _Hello World_ application source file:
1021 #include "hello-tp.h"
1023 int main(int argc, char *argv[])
1027 puts("Hello, World!\nPress Enter to continue...");
1030 * The following getchar() call only exists for the purpose of this
1031 * demonstration, to pause the application in order for you to have
1032 * time to list its tracepoints. You don't need it otherwise.
1037 * An lttng_ust_tracepoint() call.
1039 * Arguments, as defined in `hello-tp.h`:
1041 * 1. Tracepoint provider name (required)
1042 * 2. Tracepoint name (required)
1043 * 3. `my_integer_arg` (first user-defined argument)
1044 * 4. `my_string_arg` (second user-defined argument)
1046 * Notice the tracepoint provider and tracepoint names are
1047 * C identifiers, NOT strings: they're in fact parts of variables
1048 * that the macros in `hello-tp.h` create.
1050 lttng_ust_tracepoint(hello_world, my_first_tracepoint, 23,
1053 for (i = 0; i < argc; i++) {
1054 lttng_ust_tracepoint(hello_world, my_first_tracepoint,
1058 puts("Quitting now!");
1059 lttng_ust_tracepoint(hello_world, my_first_tracepoint,
1066 . Build the application:
1075 . Link the application with the tracepoint provider package,
1076 `liblttng-ust` and `libdl`:
1081 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
1085 Here's the whole build process:
1088 .Build steps of the user space tracing tutorial.
1089 image::ust-flow.png[]
1091 To record the events of the user application:
1093 . Run the application with a few arguments:
1098 $ ./hello world and beyond
1107 Press Enter to continue...
1111 . Start an LTTng <<lttng-sessiond,session daemon>>:
1116 $ lttng-sessiond --daemonize
1120 NOTE: A session daemon might already be running, for example as a
1121 service that the service manager of your distribution started.
1123 . List the available user space tracepoints:
1128 $ lttng list --userspace
1132 You see the `hello_world:my_first_tracepoint` tracepoint listed
1133 under the `./hello` process.
1135 . Create a <<tracing-session,recording session>>:
1140 $ lttng create my-user-space-session
1144 . Create a <<event,recording event rule>> which matches user space
1145 tracepoint events named `hello_world:my_first_tracepoint`:
1150 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1154 . <<basic-tracing-session-control,Start recording>>:
1163 . Go back to the running `hello` application and press **Enter**.
1165 The program executes all `lttng_ust_tracepoint()` instrumentation
1166 points, emitting events as the event rule you created in step{nbsp}5
1170 . <<creating-destroying-tracing-sessions,Destroy>> the current
1180 The man:lttng-destroy(1) command doesn't destroy the trace data; it
1181 only destroys the state of the recording session.
1183 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1184 implicitly (see ``<<basic-tracing-session-control,Start and stop a
1185 recording session>>''). You need to stop recording to make LTTng flush
1186 the remaining trace data and make the trace readable.
1188 By default, LTTng saves the traces to the
1189 +$LTTNG_HOME/lttng-traces/__NAME__-__DATE__-__TIME__+ directory, where
1190 +__NAME__+ is the recording session name. The env:LTTNG_HOME environment
1191 variable defaults to `$HOME` if not set.
1194 [[viewing-and-analyzing-your-traces]]
1195 === View and analyze the recorded events
1197 Once you have completed the <<tracing-the-linux-kernel,Record Linux
1198 kernel events>> and <<tracing-your-own-user-application,Record user
1199 application events>> tutorials, you can inspect the recorded events.
1201 There are many tools you can use to read LTTng traces:
1203 https://babeltrace.org/[Babeltrace{nbsp}2]::
1204 A rich, flexible trace manipulation toolkit which includes
1205 a versatile command-line interface
1206 (man:babeltrace2(1)),
1207 a https://babeltrace.org/docs/v2.0/libbabeltrace2/[C{nbsp}library],
1208 and https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1209 so that you can easily process or convert an LTTng trace with
1212 The Babeltrace{nbsp}2 project ships with a plugin
1213 (man:babeltrace2-plugin-ctf(7)) which supports the format of the traces
1214 which LTTng produces, https://diamon.org/ctf/[CTF].
1216 http://tracecompass.org/[Trace Compass]::
1217 A graphical user interface for viewing and analyzing any type of
1218 logs or traces, including those of LTTng.
1220 https://github.com/lttng/lttng-analyses[LTTng analyses]::
1221 An experimental project which includes many high-level analyses of
1222 LTTng kernel traces, like scheduling statistics, interrupt
1223 frequency distribution, top CPU usage, and more.
1225 NOTE: This section assumes that LTTng wrote the traces it recorded
1226 during the previous tutorials to their default location, in the
1227 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1228 environment variable defaults to `$HOME` if not set.
1231 [[viewing-and-analyzing-your-traces-bt]]
1232 ==== Use the cmd:babeltrace2 command-line tool
1234 The simplest way to list all the recorded events of an LTTng trace is to
1235 pass its path to man:babeltrace2(1), without options:
1239 $ babeltrace2 ~/lttng-traces/my-user-space-session*
1242 The cmd:babeltrace2 command finds all traces recursively within the
1243 given path and prints all their events, sorting them chronologically.
1245 Pipe the output of cmd:babeltrace2 into a tool like man:grep(1) for
1250 $ babeltrace2 /tmp/my-kernel-trace | grep _switch
1253 Pipe the output of cmd:babeltrace2 into a tool like man:wc(1) to count
1254 the recorded events:
1258 $ babeltrace2 /tmp/my-kernel-trace | grep _open | wc --lines
1262 [[viewing-and-analyzing-your-traces-bt-python]]
1263 ==== Use the Babeltrace{nbsp}2 Python bindings
1265 The <<viewing-and-analyzing-your-traces-bt,text output of
1266 cmd:babeltrace2>> is useful to isolate event records by simple matching
1267 using man:grep(1) and similar utilities. However, more elaborate
1268 filters, such as keeping only event records with a field value falling
1269 within a specific range, are not trivial to write using a shell.
1270 Moreover, reductions and even the most basic computations involving
1271 multiple event records are virtually impossible to implement.
1273 Fortunately, Babeltrace{nbsp}2 ships with
1274 https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1275 which make it easy to read the event records of an LTTng trace
1276 sequentially and compute the desired information.
1278 The following script accepts an LTTng Linux kernel trace path as its
1279 first argument and prints the short names of the top five running
1280 processes on CPU{nbsp}0 during the whole trace:
1291 # Get the trace path from the first command-line argument
1292 it = bt2.TraceCollectionMessageIterator(sys.argv[1])
1294 # This counter dictionary will hold execution times:
1296 # Task command name -> Total execution time (ns)
1297 exec_times = collections.Counter()
1299 # This holds the last `sched_switch` timestamp
1303 # We only care about event messages
1304 if type(msg) is not bt2._EventMessageConst:
1307 # Event of the event message
1310 # Keep only `sched_switch` events
1311 if event.cls.name != 'sched_switch':
1314 # Keep only records of events which LTTng emitted from CPU 0
1315 if event.packet.context_field['cpu_id'] != 0:
1318 # Event timestamp (ns)
1319 cur_ts = msg.default_clock_snapshot.ns_from_origin
1325 # (Short) name of the previous task command
1326 prev_comm = str(event.payload_field['prev_comm'])
1328 # Initialize an entry in our dictionary if not done yet
1329 if prev_comm not in exec_times:
1330 exec_times[prev_comm] = 0
1332 # Compute previous command execution time
1333 diff = cur_ts - last_ts
1335 # Update execution time of this command
1336 exec_times[prev_comm] += diff
1338 # Update last timestamp
1342 for name, ns in exec_times.most_common(5):
1343 print('{:20}{} s'.format(name, ns / 1e9))
1346 if __name__ == '__main__':
1354 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1360 swapper/0 48.607245889 s
1361 chromium 7.192738188 s
1362 pavucontrol 0.709894415 s
1363 Compositor 0.660867933 s
1364 Xorg.bin 0.616753786 s
1367 Note that `swapper/0` is the ``idle'' process of CPU{nbsp}0 on Linux;
1368 since we weren't using the CPU that much when recording, its first
1369 position in the list makes sense.
1373 == [[understanding-lttng]]Core concepts
1375 From a user's perspective, the LTTng system is built on a few concepts,
1376 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1377 operates by sending commands to the <<lttng-sessiond,session daemon>>
1378 (through <<liblttng-ctl-lttng,`liblttng-ctl`>>).
1380 Understanding how those objects relate to each other is key to master
1383 The core concepts of LTTng are:
1385 * <<"event-rule","Instrumentation point, event rule, and event">>
1386 * <<trigger,Trigger>>
1387 * <<tracing-session,Recording session>>
1388 * <<domain,Tracing domain>>
1389 * <<channel,Channel and ring buffer>>
1390 * <<event,Recording event rule and event record>>
1392 NOTE: The man:lttng-concepts(7) manual page also documents the core
1393 concepts of LTTng, with more links to other LTTng-tools manual pages.
1397 === Instrumentation point, event rule, and event
1399 An _instrumentation point_ is a point, within a piece of software,
1400 which, when executed, creates an LTTng _event_.
1402 LTTng offers various <<instrumentation-point-types,types of
1405 An _event rule_ is a set of conditions to match a set of events.
1407 When LTTng creates an event{nbsp}__E__, an event rule{nbsp}__ER__ is
1408 said to __match__{nbsp}__E__ when{nbsp}__E__ satisfies _all_ the
1409 conditions of{nbsp}__ER__. This concept is similar to a
1410 https://en.wikipedia.org/wiki/Regular_expression[regular expression]
1411 which matches a set of strings.
1413 When an event rule matches an event, LTTng _emits_ the event, therefore
1414 attempting to execute one or more actions.
1418 [[event-creation-emission-opti]]The event creation and emission
1419 processes are documentation concepts to help understand the journey from
1420 an instrumentation point to the execution of actions.
1422 The actual creation of an event can be costly because LTTng needs to
1423 evaluate the arguments of the instrumentation point.
1425 In practice, LTTng implements various optimizations for the Linux kernel
1426 and user space <<domain,tracing domains>> to avoid actually creating an
1427 event when the tracer knows, thanks to properties which are independent
1428 from the event payload and current context, that it would never emit
1429 such an event. Those properties are:
1431 * The <<instrumentation-point-types,instrumentation point type>>.
1433 * The instrumentation point name.
1435 * The instrumentation point log level.
1437 * For a <<event,recording event rule>>:
1438 ** The status of the rule itself.
1439 ** The status of the <<channel,channel>>.
1440 ** The activity of the <<tracing-session,recording session>>.
1441 ** Whether or not the process for which LTTng would create the event is
1442 <<pid-tracking,allowed to record events>>.
1444 In other words: if, for a given instrumentation point{nbsp}__IP__, the
1445 LTTng tracer knows that it would never emit an event,
1446 executing{nbsp}__IP__ represents a simple boolean variable check and,
1447 for a Linux kernel recording event rule, a few process attribute checks.
1450 As of LTTng{nbsp}{revision}, there are two places where you can find an
1453 <<event,Recording event rule>>::
1454 A specific type of event rule of which the action is to record the
1455 matched event as an event record.
1457 See ``<<enabling-disabling-events,Create and enable a recording event
1458 rule>>'' to learn more.
1460 ``Event rule matches'' <<trigger,trigger>> condition (since LTTng{nbsp}2.13)::
1461 When the event rule of the trigger condition matches an event, LTTng
1462 can execute user-defined actions such as sending an LTTng
1464 <<basic-tracing-session-control,starting a recording session>>,
1467 See “<<add-event-rule-matches-trigger,Add an ``event rule matches''
1468 trigger to a session daemon>>” to learn more.
1470 For LTTng to emit an event{nbsp}__E__,{nbsp}__E__ must satisfy _all_ the
1471 basic conditions of an event rule{nbsp}__ER__, that is:
1473 * The instrumentation point from which LTTng
1474 creates{nbsp}__E__ has a specific
1475 <<instrumentation-point-types,type>>.
1477 * A pattern matches the name of{nbsp}__E__ while another pattern
1480 * The log level of the instrumentation point from which LTTng
1481 creates{nbsp}__E__ is at least as severe as some value, or is exactly
1484 * The fields of the payload of{nbsp}__E__ and the current context fields
1485 satisfy a filter expression.
1487 A <<event,recording event rule>> has additional, implicit conditions to
1491 [[instrumentation-point-types]]
1492 ==== Instrumentation point types
1494 As of LTTng{nbsp}{revision}, the available instrumentation point
1495 types are, depending on the <<domain,tracing domain>>:
1499 A statically defined point in the source code of the kernel
1500 image or of a kernel module using the
1501 <<lttng-modules,LTTng-modules>> macros.
1503 Linux kernel system call:::
1504 Entry, exit, or both of a Linux kernel system call.
1506 Linux https://www.kernel.org/doc/html/latest/trace/kprobes.html[kprobe]:::
1507 A single probe dynamically placed in the compiled kernel code.
1509 When you create such an instrumentation point, you set its memory
1510 address or symbol name.
1512 Linux user space probe:::
1513 A single probe dynamically placed at the entry of a compiled
1514 user space application/library function through the kernel.
1516 When you create such an instrumentation point, you set:
1519 With the ELF method::
1520 Its application/library path and its symbol name.
1522 With the USDT method::
1523 Its application/library path, its provider name, and its probe name.
1525 ``USDT'' stands for _SystemTap User-level Statically Defined Tracing_,
1526 a http://dtrace.org/blogs/about/[DTrace]-style marker.
1529 As of LTTng{nbsp}{revision}, LTTng only supports USDT probes which
1530 are _not_ reference-counted.
1532 Linux https://www.kernel.org/doc/html/latest/trace/kprobes.html[kretprobe]:::
1533 Entry, exit, or both of a Linux kernel function.
1535 When you create such an instrumentation point, you set the memory
1536 address or symbol name of its function.
1540 A statically defined point in the source code of a C/$$C++$$
1541 application/library using the
1542 <<lttng-ust,LTTng-UST>> macros.
1544 `java.util.logging`, Apache log4j, and Python::
1545 Java or Python logging statement:::
1546 A method call on a Java or Python logger attached to an
1549 See ``<<list-instrumentation-points,List the available instrumentation
1550 points>>'' to learn how to list available Linux kernel, user space, and
1551 logging instrumentation points.
1557 A _trigger_ associates a condition to one or more actions.
1559 When the condition of a trigger is satisfied, LTTng attempts to execute
1562 As of LTTng{nbsp}{revision}, the available trigger conditions and
1567 * The consumed buffer size of a given <<tracing-session,recording
1568 session>> becomes greater than some value.
1570 * The buffer usage of a given <<channel,channel>> becomes greater than
1573 * The buffer usage of a given channel becomes less than some value.
1575 * There's an ongoing <<session-rotation,recording session rotation>>.
1577 * A recording session rotation becomes completed.
1579 * An <<add-event-rule-matches-trigger,event rule matches>> an event.
1583 * <<trigger-event-notif,Send a notification>> to a user application.
1584 * <<basic-tracing-session-control,Start>> a given recording session.
1585 * <<basic-tracing-session-control,Stop>> a given recording session.
1586 * <<session-rotation,Archive the current trace chunk>> of a given
1587 recording session (rotate).
1588 * <<taking-a-snapshot,Take a snapshot>> of a given recording session.
1590 A trigger belongs to a <<lttng-sessiond,session daemon>>, not to a
1591 specific recording session. For a given session daemon, each Unix user has
1592 its own, private triggers. Note, however, that the `root` Unix user may,
1593 for the root session daemon:
1595 * Add a trigger as another Unix user.
1597 * List all the triggers, regardless of their owner.
1599 * Remove a trigger which belongs to another Unix user.
1601 For a given session daemon and Unix user, a trigger has a unique name.
1605 === Recording session
1607 A _recording session_ (named ``tracing session'' prior to
1608 LTTng{nbsp}2.13) is a stateful dialogue between you and a
1609 <<lttng-sessiond,session daemon>> for everything related to
1610 <<event,event recording>>.
1612 Everything that you do when you control LTTng tracers to record events
1613 happens within a recording session. In particular, a recording session:
1615 * Has its own name, unique for a given session daemon.
1617 * Has its own set of trace files, if any.
1619 * Has its own state of activity (started or stopped).
1621 An active recording session is an implicit <<event,recording event rule>>
1624 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1627 * Has its own <<channel,channels>> to which are attached their own
1628 recording event rules.
1630 * Has its own <<pid-tracking,process attribute inclusion sets>>.
1633 .A _recording session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,recording event rules>>.
1634 image::concepts.png[]
1636 Those attributes and objects are completely isolated between different
1639 A recording session is like an
1640 https://en.wikipedia.org/wiki/Automated_teller_machine[ATM] session: the
1641 operations you do on the banking system through the ATM don't alter the
1642 data of other users of the same system. In the case of the ATM, a
1643 session lasts as long as your bank card is inside. In the case of LTTng,
1644 a recording session lasts from the man:lttng-create(1) command to the
1645 man:lttng-destroy(1) command.
1648 .Each Unix user has its own set of recording sessions.
1649 image::many-sessions.png[]
1651 A recording session belongs to a <<lttng-sessiond,session daemon>>. For a
1652 given session daemon, each Unix user has its own, private recording
1653 sessions. Note, however, that the `root` Unix user may operate on or
1654 destroy another user's recording session.
1657 [[tracing-session-mode]]
1658 ==== Recording session mode
1660 LTTng offers four recording session modes:
1662 [[local-mode]]Local mode::
1663 Write the trace data to the local file system.
1665 [[net-streaming-mode]]Network streaming mode::
1666 Send the trace data over the network to a listening
1667 <<lttng-relayd,relay daemon>>.
1669 [[snapshot-mode]]Snapshot mode::
1670 Only write the trace data to the local file system or send it to a
1671 listening relay daemon when LTTng <<taking-a-snapshot,takes a
1674 LTTng forces all the <<channel,channels>>
1675 to be created to be configured to be snapshot-ready.
1677 LTTng takes a snapshot of such a recording session when:
1680 * You run the man:lttng-snapshot(1) command.
1682 * LTTng executes a `snapshot-session` <<trigger,trigger>> action.
1685 [[live-mode]]Live mode::
1686 Send the trace data over the network to a listening relay daemon
1687 for <<lttng-live,live reading>>.
1689 An LTTng live reader (for example, man:babeltrace2(1)) can connect to
1690 the same relay daemon to receive trace data while the recording session is
1697 A _tracing domain_ identifies a type of LTTng tracer.
1699 A tracing domain has its own properties and features.
1701 There are currently five available tracing domains:
1705 * `java.util.logging` (JUL)
1709 You must specify a tracing domain to target a type of LTTng tracer when
1710 using some <<lttng-cli,cmd:lttng>> commands to avoid ambiguity. For
1711 example, because the Linux kernel and user space tracing domains support
1712 named tracepoints as <<event-rule,instrumentation points>>, you need to
1713 specify a tracing domain when you <<enabling-disabling-events,create
1714 an event rule>> because both tracing domains could have tracepoints
1715 sharing the same name.
1717 You can create <<channel,channels>> in the Linux kernel and user space
1718 tracing domains. The other tracing domains have a single, default
1723 === Channel and ring buffer
1725 A _channel_ is an object which is responsible for a set of
1728 Each ring buffer is divided into multiple _sub-buffers_. When a
1729 <<event,recording event rule>>
1730 matches an event, LTTng can record it to one or more sub-buffers of one
1733 When you <<enabling-disabling-channels,create a channel>>, you set its
1734 final attributes, that is:
1736 * Its <<channel-buffering-schemes,buffering scheme>>.
1738 * What to do <<channel-overwrite-mode-vs-discard-mode,when there's no
1739 space left>> for a new event record because all sub-buffers are full.
1741 * The <<channel-subbuf-size-vs-subbuf-count,size of each ring buffer and
1742 how many sub-buffers>> a ring buffer has.
1744 * The <<tracefile-rotation,size of each trace file LTTng writes for this
1745 channel and the maximum count>> of trace files.
1747 * The periods of its <<channel-read-timer,read>>,
1748 <<channel-switch-timer,switch>>, and <<channel-monitor-timer,monitor>>
1751 * For a Linux kernel channel: its output type.
1753 See the opt:lttng-enable-channel(1):--output option of the
1754 man:lttng-enable-channel(1) command.
1756 * For a user space channel: the value of its
1757 <<blocking-timeout-example,blocking timeout>>.
1759 A channel is always associated to a <<domain,tracing domain>>. The
1760 `java.util.logging` (JUL), log4j, and Python tracing domains each have a
1761 default channel which you can't configure.
1763 A channel owns <<event,recording event rules>>.
1766 [[channel-buffering-schemes]]
1767 ==== Buffering scheme
1769 A channel has at least one ring buffer _per CPU_. LTTng always records
1770 an event to the ring buffer dedicated to the CPU which emits it.
1772 The buffering scheme of a user space channel determines what has its own
1773 set of per-CPU ring buffers:
1775 Per-user buffering::
1776 Allocate one set of ring buffers--one per CPU--shared by all the
1777 instrumented processes of:
1778 If your Unix user is `root`:::
1783 .Per-user buffering scheme (recording session belongs to the `root` Unix user).
1784 image::per-user-buffering-root.png[]
1792 .Per-user buffering scheme (recording session belongs to the `Bob` Unix user).
1793 image::per-user-buffering.png[]
1796 Per-process buffering::
1797 Allocate one set of ring buffers--one per CPU--for each
1798 instrumented process of:
1799 If your Unix user is `root`:::
1804 .Per-process buffering scheme (recording session belongs to the `root` Unix user).
1805 image::per-process-buffering-root.png[]
1813 .Per-process buffering scheme (recording session belongs to the `Bob` Unix user).
1814 image::per-process-buffering.png[]
1817 The per-process buffering scheme tends to consume more memory than the
1818 per-user option because systems generally have more instrumented
1819 processes than Unix users running instrumented processes. However, the
1820 per-process buffering scheme ensures that one process having a high
1821 event throughput won't fill all the shared sub-buffers of the same Unix
1824 The buffering scheme of a Linux kernel channel is always to allocate a
1825 single set of ring buffers for the whole system. This scheme is similar
1826 to the per-user option, but with a single, global user ``running'' the
1830 [[channel-overwrite-mode-vs-discard-mode]]
1831 ==== Event record loss mode
1833 When LTTng emits an event, LTTng can record it to a specific, available
1834 sub-buffer within the ring buffers of specific channels. When there's no
1835 space left in a sub-buffer, the tracer marks it as consumable and
1836 another, available sub-buffer starts receiving the following event
1837 records. An LTTng <<lttng-consumerd,consumer daemon>> eventually
1838 consumes the marked sub-buffer, which returns to the available state.
1841 [role="docsvg-channel-subbuf-anim"]
1846 In an ideal world, sub-buffers are consumed faster than they're filled.
1847 In the real world, however, all sub-buffers can be full at some point,
1848 leaving no space to record the following events.
1850 In an ideal world, sub-buffers are consumed faster than they're filled,
1851 as it's the case in the previous animation. In the real world,
1852 however, all sub-buffers can be full at some point, leaving no space to
1853 record the following events.
1855 By default, <<lttng-modules,LTTng-modules>> and <<lttng-ust,LTTng-UST>>
1856 are _non-blocking_ tracers: when there's no available sub-buffer to
1857 record an event, it's acceptable to lose event records when the
1858 alternative would be to cause substantial delays in the execution of the
1859 instrumented application. LTTng privileges performance over integrity;
1860 it aims at perturbing the instrumented application as little as possible
1861 in order to make the detection of subtle race conditions and rare
1862 interrupt cascades possible.
1864 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1865 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1866 example>> to learn how to use the blocking mode.
1868 When it comes to losing event records because there's no available
1869 sub-buffer, or because the blocking timeout of
1870 the channel is reached, the _event record loss mode_ of the channel
1871 determines what to do. The available event record loss modes are:
1873 [[discard-mode]]Discard mode::
1874 Drop the newest event records until a sub-buffer becomes available.
1876 This is the only available mode when you specify a blocking timeout.
1878 With this mode, LTTng increments a count of lost event records when an
1879 event record is lost and saves this count to the trace. A trace reader
1880 can use the saved discarded event record count of the trace to decide
1881 whether or not to perform some analysis even if trace data is known to
1884 [[overwrite-mode]]Overwrite mode::
1885 Clear the sub-buffer containing the oldest event records and start
1886 writing the newest event records there.
1888 This mode is sometimes called _flight recorder mode_ because it's
1889 similar to a https://en.wikipedia.org/wiki/Flight_recorder[flight
1890 recorder]: always keep a fixed amount of the latest data. It's also
1891 similar to the roll mode of an oscilloscope.
1893 Since LTTng{nbsp}2.8, with this mode, LTTng writes to a given sub-buffer
1894 its sequence number within its data stream. With a <<local-mode,local>>,
1895 <<net-streaming-mode,network streaming>>, or <<live-mode,live>> recording
1896 session, a trace reader can use such sequence numbers to report lost
1897 packets. A trace reader can use the saved discarded sub-buffer (packet)
1898 count of the trace to decide whether or not to perform some analysis
1899 even if trace data is known to be missing.
1901 With this mode, LTTng doesn't write to the trace the exact number of
1902 lost event records in the lost sub-buffers.
1904 Which mechanism you should choose depends on your context: prioritize
1905 the newest or the oldest event records in the ring buffer?
1907 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1908 as soon as a there's no space left for a new event record, whereas in
1909 discard mode, the tracer only discards the event record that doesn't
1912 There are a few ways to decrease your probability of losing event
1913 records. The ``<<channel-subbuf-size-vs-subbuf-count,Sub-buffer size and
1914 count>>'' section shows how to fine-tune the sub-buffer size and count
1915 of a channel to virtually stop losing event records, though at the cost
1916 of greater memory usage.
1919 [[channel-subbuf-size-vs-subbuf-count]]
1920 ==== Sub-buffer size and count
1922 A channel has one or more ring buffer for each CPU of the target system.
1924 See the ``<<channel-buffering-schemes,Buffering scheme>>'' section to
1925 learn how many ring buffers of a given channel are dedicated to each CPU
1926 depending on its buffering scheme.
1928 Set the size of each sub-buffer the ring buffers of a channel contain
1929 and how many there are
1930 when you <<enabling-disabling-channels,create it>>.
1932 Note that LTTng switching the current sub-buffer of a ring buffer
1933 (marking a full one as consumable and switching to an available one for
1934 LTTng to record the next events) introduces noticeable CPU overhead.
1935 Knowing this, the following list presents a few practical situations
1936 along with how to configure the sub-buffer size and count for them:
1938 High event throughput::
1939 In general, prefer large sub-buffers to lower the risk of losing
1942 Having larger sub-buffers also ensures a lower sub-buffer switching
1945 The sub-buffer count is only meaningful if you create the channel in
1946 <<overwrite-mode,overwrite mode>>: in this case, if LTTng overwrites a
1947 sub-buffer, then the other sub-buffers are left unaltered.
1949 Low event throughput::
1950 In general, prefer smaller sub-buffers since the risk of losing
1951 event records is low.
1953 Because LTTng emits events less frequently, the sub-buffer switching
1954 frequency should remain low and therefore the overhead of the tracer
1955 shouldn't be a problem.
1958 If your target system has a low memory limit, prefer fewer first,
1959 then smaller sub-buffers.
1961 Even if the system is limited in memory, you want to keep the
1962 sub-buffers as large as possible to avoid a high sub-buffer switching
1965 Note that LTTng uses https://diamon.org/ctf/[CTF] as its trace format,
1966 which means event record data is very compact. For example, the average
1967 LTTng kernel event record weights about 32{nbsp}bytes. Therefore, a
1968 sub-buffer size of 1{nbsp}MiB is considered large.
1970 The previous scenarios highlight the major trade-off between a few large
1971 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1972 frequency vs. how many event records are lost in overwrite mode.
1973 Assuming a constant event throughput and using the overwrite mode, the
1974 two following configurations have the same ring buffer total size:
1977 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1982 Two sub-buffers of 4{nbsp}MiB each::
1983 Expect a very low sub-buffer switching frequency, but if LTTng
1984 ever needs to overwrite a sub-buffer, half of the event records so
1985 far (4{nbsp}MiB) are definitely lost.
1987 Eight sub-buffers of 1{nbsp}MiB each::
1988 Expect four times the tracer overhead of the configuration above,
1989 but if LTTng needs to overwrite a sub-buffer, only the eighth of
1990 event records so far (1{nbsp}MiB) are definitely lost.
1992 In <<discard-mode,discard mode>>, the sub-buffer count parameter is
1993 pointless: use two sub-buffers and set their size according to your
1997 [[tracefile-rotation]]
1998 ==== Maximum trace file size and count (trace file rotation)
2000 By default, trace files can grow as large as needed.
2002 Set the maximum size of each trace file that LTTng writes of a given
2003 channel when you <<enabling-disabling-channels,create it>>.
2005 When the size of a trace file reaches the fixed maximum size of the
2006 channel, LTTng creates another file to contain the next event records.
2007 LTTng appends a file count to each trace file name in this case.
2009 If you set the trace file size attribute when you create a channel, the
2010 maximum number of trace files that LTTng creates is _unlimited_ by
2011 default. To limit them, set a maximum number of trace files. When the
2012 number of trace files reaches the fixed maximum count of the channel,
2013 LTTng overwrites the oldest trace file. This mechanism is called _trace
2018 Even if you don't limit the trace file count, always assume that LTTng
2019 manages all the trace files of the recording session.
2021 In other words, there's no safe way to know if LTTng still holds a given
2022 trace file open with the trace file rotation feature.
2024 The only way to obtain an unmanaged, self-contained LTTng trace before
2025 you <<creating-destroying-tracing-sessions,destroy the recording session>>
2026 is with the <<session-rotation,recording session rotation>> feature, which
2027 is available since LTTng{nbsp}2.11.
2034 Each channel can have up to three optional timers:
2036 [[channel-switch-timer]]Switch timer::
2037 When this timer expires, a sub-buffer switch happens: for each ring
2038 buffer of the channel, LTTng marks the current sub-buffer as
2039 consumable and _switches_ to an available one to record the next
2043 [role="docsvg-channel-switch-timer"]
2048 A switch timer is useful to ensure that LTTng consumes and commits trace
2049 data to trace files or to a distant <<lttng-relayd,relay daemon>>
2050 periodically in case of a low event throughput.
2052 Such a timer is also convenient when you use large
2053 <<channel-subbuf-size-vs-subbuf-count,sub-buffers>> to cope with a
2054 sporadic high event throughput, even if the throughput is otherwise low.
2056 Set the period of the switch timer of a channel when you
2057 <<enabling-disabling-channels,create it>> with
2058 the opt:lttng-enable-channel(1):--switch-timer option.
2060 [[channel-read-timer]]Read timer::
2061 When this timer expires, LTTng checks for full, consumable
2064 By default, the LTTng tracers use an asynchronous message mechanism to
2065 signal a full sub-buffer so that a <<lttng-consumerd,consumer daemon>>
2068 When such messages must be avoided, for example in real-time
2069 applications, use this timer instead.
2071 Set the period of the read timer of a channel when you
2072 <<enabling-disabling-channels,create it>> with the
2073 opt:lttng-enable-channel(1):--read-timer option.
2075 [[channel-monitor-timer]]Monitor timer::
2076 When this timer expires, the consumer daemon samples some channel
2077 statistics to evaluate the following <<trigger,trigger>>
2081 . The consumed buffer size of a given <<tracing-session,recording
2082 session>> becomes greater than some value.
2083 . The buffer usage of a given channel becomes greater than some value.
2084 . The buffer usage of a given channel becomes less than some value.
2087 If you disable the monitor timer of a channel{nbsp}__C__:
2090 * The consumed buffer size value of the recording session of{nbsp}__C__
2091 could be wrong for trigger condition type{nbsp}1: the consumed buffer
2092 size of{nbsp}__C__ won't be part of the grand total.
2094 * The buffer usage trigger conditions (types{nbsp}2 and{nbsp}3)
2095 for{nbsp}__C__ will never be satisfied.
2098 Set the period of the monitor timer of a channel when you
2099 <<enabling-disabling-channels,create it>> with the
2100 opt:lttng-enable-channel(1):--monitor-timer option.
2104 === Recording event rule and event record
2106 A _recording event rule_ is a specific type of <<event-rule,event rule>>
2107 of which the action is to serialize and record the matched event as an
2110 Set the explicit conditions of a recording event rule when you
2111 <<enabling-disabling-events,create it>>. A recording event rule also has
2112 the following implicit conditions:
2114 * The recording event rule itself is enabled.
2116 A recording event rule is enabled on creation.
2118 * The <<channel,channel>> to which the recording event rule is attached
2121 A channel is enabled on creation.
2123 * The <<tracing-session,recording session>> of the recording event rule is
2124 <<basic-tracing-session-control,active>> (started).
2126 A recording session is inactive (stopped) on creation.
2128 * The process for which LTTng creates an event to match is
2129 <<pid-tracking,allowed to record events>>.
2131 All processes are allowed to record events on recording session
2134 You always attach a recording event rule to a channel, which belongs to
2135 a recording session, when you create it.
2137 When a recording event rule{nbsp}__ER__ matches an event{nbsp}__E__,
2138 LTTng attempts to serialize and record{nbsp}__E__ to one of the
2139 available sub-buffers of the channel to which{nbsp}__E__ is attached.
2141 When multiple matching recording event rules are attached to the same
2142 channel, LTTng attempts to serialize and record the matched event
2143 _once_. In the following example, the second recording event rule is
2144 redundant when both are enabled:
2148 $ lttng enable-event --userspace hello:world
2149 $ lttng enable-event --userspace hello:world --loglevel=INFO
2153 .Logical path from an instrumentation point to an event record.
2154 image::event-rule.png[]
2156 As of LTTng{nbsp}{revision}, you cannot remove a recording event
2157 rule: it exists as long as its recording session exists.
2161 == Components of noch:{LTTng}
2163 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
2164 to call LTTng a simple _tool_ since it's composed of multiple
2165 interacting components.
2167 This section describes those components, explains their respective
2168 roles, and shows how they connect together to form the LTTng ecosystem.
2170 The following diagram shows how the most important components of LTTng
2171 interact with user applications, the Linux kernel, and you:
2174 .Control and trace data paths between LTTng components.
2175 image::plumbing.png[]
2177 The LTTng project integrates:
2180 Libraries and command-line interface to control recording sessions:
2182 * <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
2183 * <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
2184 * <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
2185 * <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
2186 * <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
2187 * <<persistent-memory-file-systems,`lttng-crash` command-line tool>>
2188 (man:lttng-crash(1)).
2191 Libraries and Java/Python packages to instrument and trace user
2194 * <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
2195 headers to instrument and trace any native user application.
2196 * <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
2197 ** `liblttng-ust-libc-wrapper`
2198 ** `liblttng-ust-pthread-wrapper`
2199 ** `liblttng-ust-cyg-profile`
2200 ** `liblttng-ust-cyg-profile-fast`
2201 ** `liblttng-ust-dl`
2202 * <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
2203 Java applications using `java.util.logging` or
2204 Apache log4j{nbsp}1.2 logging.
2205 * <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
2206 Python applications using the standard `logging` package.
2209 <<lttng-modules,Linux kernel modules>> to instrument and trace the
2212 * LTTng kernel tracer module.
2213 * Recording ring buffer kernel modules.
2214 * Probe kernel modules.
2215 * LTTng logger kernel module.
2219 === Tracing control command-line interface
2221 The _man:lttng(1) command-line tool_ is the standard user interface to
2222 control LTTng <<tracing-session,recording sessions>>.
2224 The cmd:lttng tool is part of LTTng-tools.
2226 The cmd:lttng tool is linked with
2227 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
2228 one or more <<lttng-sessiond,session daemons>> behind the scenes.
2230 The cmd:lttng tool has a Git-like interface:
2234 $ lttng [GENERAL OPTIONS] <COMMAND> [COMMAND OPTIONS]
2237 The ``<<controlling-tracing,Tracing control>>'' section explores the
2238 available features of LTTng through its cmd:lttng tool.
2241 [[liblttng-ctl-lttng]]
2242 === Tracing control library
2245 .The tracing control library.
2246 image::plumbing-liblttng-ctl.png[]
2248 The _LTTng control library_, `liblttng-ctl`, is used to communicate with
2249 a <<lttng-sessiond,session daemon>> using a C{nbsp}API that hides the
2250 underlying details of the protocol.
2252 `liblttng-ctl` is part of LTTng-tools.
2254 The <<lttng-cli,cmd:lttng command-line tool>> is linked with
2257 Use `liblttng-ctl` in C or $$C++$$ source code by including its
2262 #include <lttng/lttng.h>
2265 As of LTTng{nbsp}{revision}, the best available developer documentation
2266 for `liblttng-ctl` is its installed header files. Functions and
2267 structures are documented with header comments.
2271 === User space tracing library
2274 .The user space tracing library.
2275 image::plumbing-liblttng-ust.png[]
2277 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
2278 is the LTTng user space tracer.
2280 `liblttng-ust` receives commands from a <<lttng-sessiond,session
2281 daemon>>, for example to allow specific instrumentation points to emit
2282 LTTng <<event-rule,events>>, and writes event records to <<channel,ring
2283 buffers>> shared with a <<lttng-consumerd,consumer daemon>>.
2285 `liblttng-ust` is part of LTTng-UST.
2287 `liblttng-ust` can also send asynchronous messages to the session daemon
2288 when it emits an event. This supports the ``event rule matches''
2289 <<trigger,trigger>> condition feature (see
2290 “<<add-event-rule-matches-trigger,Add an ``event rule matches'' trigger
2291 to a session daemon>>”).
2293 Public C{nbsp}header files are installed beside `liblttng-ust` to
2294 instrument any <<c-application,C or $$C++$$ application>>.
2296 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
2297 packages, use their own <<tracepoint-provider,tracepoint provider
2298 package>> which is linked with `liblttng-ust`.
2300 An application or library doesn't have to initialize `liblttng-ust`
2301 manually: its constructor does the necessary tasks to register the
2302 application to a session daemon. The initialization phase also
2303 configures instrumentation points depending on the <<event-rule,event
2304 rules>> that you already created.
2307 [[lttng-ust-agents]]
2308 === User space tracing agents
2311 .The user space tracing agents.
2312 image::plumbing-lttng-ust-agents.png[]
2314 The _LTTng-UST Java and Python agents_ are regular Java and Python
2315 packages which add LTTng tracing capabilities to the
2316 native logging frameworks.
2318 The LTTng-UST agents are part of LTTng-UST.
2320 In the case of Java, the
2321 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
2322 core logging facilities] and
2323 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
2324 Note that Apache Log4j{nbsp}2 isn't supported.
2326 In the case of Python, the standard
2327 https://docs.python.org/3/library/logging.html[`logging`] package
2328 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
2329 LTTng-UST Python agent package.
2331 The applications using the LTTng-UST agents are in the
2332 `java.util.logging` (JUL), log4j, and Python <<domain,tracing domains>>.
2334 Both agents use the same mechanism to convert log statements to LTTng
2335 events. When an agent initializes, it creates a log handler that
2336 attaches to the root logger. The agent also registers to a
2337 <<lttng-sessiond,session daemon>>. When the user application executes a
2338 log statement, the root logger passes it to the log handler of the
2339 agent. The custom log handler of the agent calls a native function in a
2340 tracepoint provider package shared library linked with
2341 <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
2342 other fields, like its logger name and its log level. This native
2343 function contains a user space instrumentation point, therefore tracing
2346 The log level condition of a <<event,recording event rule>> is
2347 considered when tracing a Java or a Python application, and it's
2348 compatible with the standard `java.util.logging`, log4j, and Python log
2353 === LTTng kernel modules
2356 .The LTTng kernel modules.
2357 image::plumbing-lttng-modules.png[]
2359 The _LTTng kernel modules_ are a set of Linux kernel modules
2360 which implement the kernel tracer of the LTTng project.
2362 The LTTng kernel modules are part of LTTng-modules.
2364 The LTTng kernel modules include:
2366 * A set of _probe_ modules.
2368 Each module attaches to a specific subsystem
2369 of the Linux kernel using its tracepoint instrument points.
2371 There are also modules to attach to the entry and return points of the
2372 Linux system call functions.
2374 * _Ring buffer_ modules.
2376 A ring buffer implementation is provided as kernel modules. The LTTng
2377 kernel tracer writes to ring buffers; a
2378 <<lttng-consumerd,consumer daemon>> reads from ring buffers.
2380 * The _LTTng kernel tracer_ module.
2381 * The <<proc-lttng-logger-abi,_LTTng logger_>> module.
2383 The LTTng logger module implements the special path:{/proc/lttng-logger}
2384 (and path:{/dev/lttng-logger}, since LTTng{nbsp}2.11) files so that any
2385 executable can generate LTTng events by opening those files and
2388 The LTTng kernel tracer can also send asynchronous messages to the
2389 <<lttng-sessiond,session daemon>> when it emits an event.
2390 This supports the ``event rule matches''
2391 <<trigger,trigger>> condition feature (see
2392 “<<add-event-rule-matches-trigger,Add an ``event rule matches'' trigger
2393 to a session daemon>>”).
2395 Generally, you don't have to load the LTTng kernel modules manually
2396 (using man:modprobe(8), for example): a root session daemon loads the
2397 necessary modules when starting. If you have extra probe modules, you
2398 can specify to load them to the session daemon on the command line
2399 (see the opt:lttng-sessiond(8):--extra-kmod-probes option).
2401 The LTTng kernel modules are installed in
2402 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
2403 the kernel release (output of `uname --kernel-release`).
2410 .The session daemon.
2411 image::plumbing-sessiond.png[]
2413 The _session daemon_, man:lttng-sessiond(8), is a
2414 https://en.wikipedia.org/wiki/Daemon_(computing)[daemon] which:
2416 * Manages <<tracing-session,recording sessions>>.
2418 * Controls the various components (like tracers and
2419 <<lttng-consumerd,consumer daemons>>) of LTTng.
2421 * Sends <<notif-trigger-api,asynchronous notifications>> to user
2424 The session daemon is part of LTTng-tools.
2426 The session daemon sends control requests to and receives control
2429 * The <<lttng-ust,user space tracing library>>.
2431 Any instance of the user space tracing library first registers to
2432 a session daemon. Then, the session daemon can send requests to
2433 this instance, such as:
2436 ** Get the list of tracepoints.
2437 ** Share a <<event,recording event rule>> so that the user space tracing
2438 library can decide whether or not a given tracepoint can emit events.
2439 Amongst the possible conditions of a recording event rule is a filter
2440 expression which `liblttng-ust` evaluates before it emits an event.
2441 ** Share <<channel,channel>> attributes and ring buffer locations.
2444 The session daemon and the user space tracing library use a Unix
2445 domain socket to communicate.
2447 * The <<lttng-ust-agents,user space tracing agents>>.
2449 Any instance of a user space tracing agent first registers to
2450 a session daemon. Then, the session daemon can send requests to
2451 this instance, such as:
2454 ** Get the list of loggers.
2455 ** Enable or disable a specific logger.
2458 The session daemon and the user space tracing agent use a TCP connection
2461 * The <<lttng-modules,LTTng kernel tracer>>.
2462 * The <<lttng-consumerd,consumer daemon>>.
2464 The session daemon sends requests to the consumer daemon to instruct
2465 it where to send the trace data streams, amongst other information.
2467 * The <<lttng-relayd,relay daemon>>.
2469 The session daemon receives commands from the
2470 <<liblttng-ctl-lttng,tracing control library>>.
2472 The session daemon can receive asynchronous messages from the
2473 <<lttng-ust,user space>> and <<lttng-modules,kernel>> tracers
2474 when they emit events. This supports the ``event rule matches''
2475 <<trigger,trigger>> condition feature (see
2476 “<<add-event-rule-matches-trigger,Add an ``event rule matches'' trigger
2477 to a session daemon>>”).
2479 The root session daemon loads the appropriate
2480 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
2481 one or more <<lttng-consumerd,consumer daemons>> as soon as you create
2482 a <<event,recording event rule>>.
2484 The session daemon doesn't send and receive trace data: this is the
2485 role of the <<lttng-consumerd,consumer daemon>> and
2486 <<lttng-relayd,relay daemon>>. It does, however, generate the
2487 https://diamon.org/ctf/[CTF] metadata stream.
2489 Each Unix user can have its own session daemon instance. The
2490 recording sessions which different session daemons manage are completely
2493 The root user's session daemon is the only one which is
2494 allowed to control the LTTng kernel tracer, and its spawned consumer
2495 daemon is the only one which is allowed to consume trace data from the
2496 LTTng kernel tracer. Note, however, that any Unix user which is a member
2497 of the <<tracing-group,tracing group>> is allowed
2498 to create <<channel,channels>> in the
2499 Linux kernel <<domain,tracing domain>>, and therefore to use the Linux
2500 kernel LTTng tracer.
2502 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2503 session daemon when using its `create` command if none is currently
2504 running. You can also start the session daemon manually.
2511 .The consumer daemon.
2512 image::plumbing-consumerd.png[]
2514 The _consumer daemon_, cmd:lttng-consumerd, is a
2515 https://en.wikipedia.org/wiki/Daemon_(computing)[daemon] which shares
2516 ring buffers with user applications or with the LTTng kernel modules to
2517 collect trace data and send it to some location (file system or to a
2518 <<lttng-relayd,relay daemon>> over the network).
2520 The consumer daemon is part of LTTng-tools.
2522 You don't start a consumer daemon manually: a consumer daemon is always
2523 spawned by a <<lttng-sessiond,session daemon>> as soon as you create a
2524 <<event,recording event rule>>, that is, before you start recording. When
2525 you kill its owner session daemon, the consumer daemon also exits
2526 because it's the child process of the session daemon. Command-line
2527 options of man:lttng-sessiond(8) target the consumer daemon process.
2529 There are up to two running consumer daemons per Unix user, whereas only
2530 one session daemon can run per user. This is because each process can be
2531 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2532 and 64-bit processes, it's more efficient to have separate
2533 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2534 exception: it can have up to _three_ running consumer daemons: 32-bit
2535 and 64-bit instances for its user applications, and one more
2536 reserved for collecting kernel trace data.
2544 image::plumbing-relayd.png[]
2546 The _relay daemon_, man:lttng-relayd(8), is a
2547 https://en.wikipedia.org/wiki/Daemon_(computing)[daemon] acting as a bridge
2548 between remote session and consumer daemons, local trace files, and a
2549 remote live trace reader.
2551 The relay daemon is part of LTTng-tools.
2553 The main purpose of the relay daemon is to implement a receiver of
2554 <<sending-trace-data-over-the-network,trace data over the network>>.
2555 This is useful when the target system doesn't have much file system
2556 space to write trace files locally.
2558 The relay daemon is also a server to which a
2559 <<lttng-live,live trace reader>> can
2560 connect. The live trace reader sends requests to the relay daemon to
2561 receive trace data as the target system records events. The
2562 communication protocol is named _LTTng live_; it's used over TCP
2565 Note that you can start the relay daemon on the target system directly.
2566 This is the setup of choice when the use case is to view/analyze events
2567 as the target system records them without the need of a remote system.
2571 == [[using-lttng]]Instrumentation
2573 There are many examples of tracing and monitoring in our everyday life:
2575 * You have access to real-time and historical weather reports and
2576 forecasts thanks to weather stations installed around the country.
2577 * You know your heart is safe thanks to an electrocardiogram.
2578 * You make sure not to drive your car too fast and to have enough fuel
2579 to reach your destination thanks to gauges visible on your dashboard.
2581 All the previous examples have something in common: they rely on
2582 **instruments**. Without the electrodes attached to the surface of your
2583 body skin, cardiac monitoring is futile.
2585 LTTng, as a tracer, is no different from those real life examples. If
2586 you're about to trace a software system or, in other words, record its
2587 history of execution, you better have **instrumentation points** in the
2588 subject you're tracing, that is, the actual software system.
2590 <<instrumentation-point-types,Various ways>> were developed to
2591 instrument a piece of software for LTTng tracing. The most
2592 straightforward one is to manually place static instrumentation points,
2593 called _tracepoints_, in the source code of the application. The Linux
2594 kernel <<domain,tracing domain>> also makes it possible to dynamically
2595 add instrumentation points.
2597 If you're only interested in tracing the Linux kernel, your
2598 instrumentation needs are probably already covered by the built-in
2599 <<lttng-modules,Linux kernel instrumentation points>> of LTTng. You may
2600 also wish to have LTTng trace a user application which is already
2601 instrumented for LTTng tracing. In such cases, skip this whole section
2602 and read the topics of the ``<<controlling-tracing,Tracing control>>''
2605 Many methods are available to instrument a piece of software for LTTng
2608 * <<c-application,Instrument a C/$$C++$$ user application>>.
2609 * <<prebuilt-ust-helpers,Load a prebuilt user space tracing helper>>.
2610 * <<java-application,Instrument a Java application>>.
2611 * <<python-application,Instrument a Python application>>.
2612 * <<proc-lttng-logger-abi,Use the LTTng logger>>.
2613 * <<instrumenting-linux-kernel,Instrument a Linux kernel image or module>>.
2617 === [[cxx-application]]Instrument a C/$$C++$$ user application
2619 The high level procedure to instrument a C or $$C++$$ user application
2620 with the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`,
2623 . <<tracepoint-provider,Create the source files of a tracepoint provider
2626 . <<probing-the-application-source-code,Add tracepoints to
2627 the source code of the application>>.
2629 . <<building-tracepoint-providers-and-user-application,Build and link
2630 a tracepoint provider package and the user application>>.
2632 If you need quick, man:printf(3)-like instrumentation, skip those steps
2633 and use <<tracef,`lttng_ust_tracef()`>> or
2634 <<tracelog,`lttng_ust_tracelog()`>> instead.
2636 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2637 instrument a user application with `liblttng-ust`.
2640 [[tracepoint-provider]]
2641 ==== Create the source files of a tracepoint provider package
2643 A _tracepoint provider_ is a set of compiled functions which provide
2644 **tracepoints** to an application, the type of instrumentation point
2645 which LTTng-UST provides.
2647 Those functions can make LTTng emit events with user-defined fields and
2648 serialize those events as event records to one or more LTTng-UST
2649 <<channel,channel>> sub-buffers. The `lttng_ust_tracepoint()` macro,
2650 which you <<probing-the-application-source-code,insert in the source
2651 code of a user application>>, calls those functions.
2653 A _tracepoint provider package_ is an object file (`.o`) or a shared
2654 library (`.so`) which contains one or more tracepoint providers. Its
2657 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2658 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2660 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2661 the LTTng user space tracer, at run time.
2664 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2665 image::ust-app.png[]
2667 NOTE: If you need quick, man:printf(3)-like instrumentation, skip
2668 creating and using a tracepoint provider and use
2669 <<tracef,`lttng_ust_tracef()`>> or <<tracelog,`lttng_ust_tracelog()`>>
2674 ===== Create a tracepoint provider header file template
2676 A _tracepoint provider header file_ contains the tracepoint definitions
2677 of a tracepoint provider.
2679 To create a tracepoint provider header file:
2681 . Start from this template:
2685 .Tracepoint provider header file template (`.h` file extension).
2687 #undef LTTNG_UST_TRACEPOINT_PROVIDER
2688 #define LTTNG_UST_TRACEPOINT_PROVIDER provider_name
2690 #undef LTTNG_UST_TRACEPOINT_INCLUDE
2691 #define LTTNG_UST_TRACEPOINT_INCLUDE "./tp.h"
2693 #if !defined(_TP_H) || defined(LTTNG_UST_TRACEPOINT_HEADER_MULTI_READ)
2696 #include <lttng/tracepoint.h>
2699 * Use LTTNG_UST_TRACEPOINT_EVENT(), LTTNG_UST_TRACEPOINT_EVENT_CLASS(),
2700 * LTTNG_UST_TRACEPOINT_EVENT_INSTANCE(), and
2701 * LTTNG_UST_TRACEPOINT_LOGLEVEL() here.
2706 #include <lttng/tracepoint-event.h>
2712 * +__provider_name__+ with the name of your tracepoint provider.
2713 * `"tp.h"` with the name of your tracepoint provider header file.
2715 . Below the `#include <lttng/tracepoint.h>` line, put your
2716 <<defining-tracepoints,tracepoint definitions>>.
2718 Your tracepoint provider name must be unique amongst all the possible
2719 tracepoint provider names used on the same target system. We suggest to
2720 include the name of your project or company in the name, for example,
2721 `org_lttng_my_project_tpp`.
2724 [[defining-tracepoints]]
2725 ===== Create a tracepoint definition
2727 A _tracepoint definition_ defines, for a given tracepoint:
2729 * Its **input arguments**.
2731 They're the macro parameters that the `lttng_ust_tracepoint()` macro
2732 accepts for this particular tracepoint in the source code of the user
2735 * Its **output event fields**.
2737 They're the sources of event fields that form the payload of any event
2738 that the execution of the `lttng_ust_tracepoint()` macro emits for this
2739 particular tracepoint.
2741 Create a tracepoint definition with the
2742 `LTTNG_UST_TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2744 <<tpp-header,tracepoint provider header file template>>.
2746 The syntax of the `LTTNG_UST_TRACEPOINT_EVENT()` macro is:
2749 .`LTTNG_UST_TRACEPOINT_EVENT()` macro syntax.
2751 LTTNG_UST_TRACEPOINT_EVENT(
2752 /* Tracepoint provider name */
2755 /* Tracepoint name */
2758 /* Input arguments */
2763 /* Output event fields */
2764 LTTNG_UST_TP_FIELDS(
2772 * +__provider_name__+ with your tracepoint provider name.
2773 * +__tracepoint_name__+ with your tracepoint name.
2774 * +__arguments__+ with the <<tpp-def-input-args,input arguments>>.
2775 * +__fields__+ with the <<tpp-def-output-fields,output event field>>
2778 The full name of this tracepoint is `provider_name:tracepoint_name`.
2781 .Event name length limitation
2783 The concatenation of the tracepoint provider name and the tracepoint
2784 name must not exceed **254{nbsp}characters**. If it does, the
2785 instrumented application compiles and runs, but LTTng throws multiple
2786 warnings and you could experience serious issues.
2789 [[tpp-def-input-args]]The syntax of the `LTTNG_UST_TP_ARGS()` macro is:
2792 .`LTTNG_UST_TP_ARGS()` macro syntax.
2801 * +__type__+ with the C{nbsp}type of the argument.
2802 * +__arg_name__+ with the argument name.
2804 You can repeat +__type__+ and +__arg_name__+ up to 10{nbsp}times to have
2805 more than one argument.
2807 .`LTTNG_UST_TP_ARGS()` usage with three arguments.
2819 The `LTTNG_UST_TP_ARGS()` and `LTTNG_UST_TP_ARGS(void)` forms are valid
2820 to create a tracepoint definition with no input arguments.
2822 [[tpp-def-output-fields]]The `LTTNG_UST_TP_FIELDS()` macro contains a
2823 list of `lttng_ust_field_*()` macros. Each `lttng_ust_field_*()` macro
2824 defines one event field. See man:lttng-ust(3) for a complete description
2825 of the available `lttng_ust_field_*()` macros. A `lttng_ust_field_*()`
2826 macro specifies the type, size, and byte order of one event field.
2828 Each `lttng_ust_field_*()` macro takes an _argument expression_
2829 parameter. This is a C{nbsp}expression that the tracer evaluates at the
2830 `lttng_ust_tracepoint()` macro site in the source code of the
2831 application. This expression provides the source of data of a field. The
2832 argument expression can include input argument names listed in the
2833 `LTTNG_UST_TP_ARGS()` macro.
2835 Each `lttng_ust_field_*()` macro also takes a _field name_ parameter.
2836 Field names must be unique within a given tracepoint definition.
2838 Here's a complete tracepoint definition example:
2840 .Tracepoint definition.
2842 The following tracepoint definition defines a tracepoint which takes
2843 three input arguments and has four output event fields.
2847 #include "my-custom-structure.h"
2849 LTTNG_UST_TRACEPOINT_EVENT(
2853 const struct my_custom_structure *, my_custom_structure,
2857 LTTNG_UST_TP_FIELDS(
2858 lttng_ust_field_string(query_field, query)
2859 lttng_ust_field_float(double, ratio_field, ratio)
2860 lttng_ust_field_integer(int, recv_size,
2861 my_custom_structure->recv_size)
2862 lttng_ust_field_integer(int, send_size,
2863 my_custom_structure->send_size)
2868 Refer to this tracepoint definition with the `lttng_ust_tracepoint()`
2869 macro in the source code of your application like this:
2873 lttng_ust_tracepoint(my_provider, my_tracepoint,
2874 my_structure, some_ratio, the_query);
2878 NOTE: The LTTng-UST tracer only evaluates the arguments of a tracepoint
2879 at run time when such a tracepoint _could_ emit an event. See
2880 <<event-creation-emission-opti,this note>> to learn more.
2883 [[using-tracepoint-classes]]
2884 ===== Use a tracepoint class
2886 A _tracepoint class_ is a class of tracepoints which share the same
2887 output event field definitions. A _tracepoint instance_ is one
2888 instance of such a defined tracepoint class, with its own tracepoint
2891 The <<defining-tracepoints,`LTTNG_UST_TRACEPOINT_EVENT()` macro>> is
2892 actually a shorthand which defines both a tracepoint class and a
2893 tracepoint instance at the same time.
2895 When you build a tracepoint provider package, the C or $$C++$$ compiler
2896 creates one serialization function for each **tracepoint class**. A
2897 serialization function is responsible for serializing the event fields
2898 of a tracepoint to a sub-buffer when recording.
2900 For various performance reasons, when your situation requires multiple
2901 tracepoint definitions with different names, but with the same event
2902 fields, we recommend that you manually create a tracepoint class and
2903 instantiate as many tracepoint instances as needed. One positive effect
2904 of such a design, amongst other advantages, is that all tracepoint
2905 instances of the same tracepoint class reuse the same serialization
2906 function, thus reducing
2907 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2909 .Use a tracepoint class and tracepoint instances.
2911 Consider the following three tracepoint definitions:
2915 LTTNG_UST_TRACEPOINT_EVENT(
2922 LTTNG_UST_TP_FIELDS(
2923 lttng_ust_field_integer(int, userid, userid)
2924 lttng_ust_field_integer(size_t, len, len)
2928 LTTNG_UST_TRACEPOINT_EVENT(
2935 LTTNG_UST_TP_FIELDS(
2936 lttng_ust_field_integer(int, userid, userid)
2937 lttng_ust_field_integer(size_t, len, len)
2941 LTTNG_UST_TRACEPOINT_EVENT(
2948 LTTNG_UST_TP_FIELDS(
2949 lttng_ust_field_integer(int, userid, userid)
2950 lttng_ust_field_integer(size_t, len, len)
2955 In this case, we create three tracepoint classes, with one implicit
2956 tracepoint instance for each of them: `get_account`, `get_settings`, and
2957 `get_transaction`. However, they all share the same event field names
2958 and types. Hence three identical, yet independent serialization
2959 functions are created when you build the tracepoint provider package.
2961 A better design choice is to define a single tracepoint class and three
2962 tracepoint instances:
2966 /* The tracepoint class */
2967 LTTNG_UST_TRACEPOINT_EVENT_CLASS(
2968 /* Tracepoint class provider name */
2971 /* Tracepoint class name */
2974 /* Input arguments */
2980 /* Output event fields */
2981 LTTNG_UST_TP_FIELDS(
2982 lttng_ust_field_integer(int, userid, userid)
2983 lttng_ust_field_integer(size_t, len, len)
2987 /* The tracepoint instances */
2988 LTTNG_UST_TRACEPOINT_EVENT_INSTANCE(
2989 /* Tracepoint class provider name */
2992 /* Tracepoint class name */
2995 /* Instance provider name */
2998 /* Tracepoint name */
3001 /* Input arguments */
3007 LTTNG_UST_TRACEPOINT_EVENT_INSTANCE(
3016 LTTNG_UST_TRACEPOINT_EVENT_INSTANCE(
3028 The tracepoint class and instance provider names must be the same if the
3029 `LTTNG_UST_TRACEPOINT_EVENT_CLASS()` and
3030 `LTTNG_UST_TRACEPOINT_EVENT_INSTANCE()` expansions are part of the same
3031 translation unit. See man:lttng-ust(3) to learn more.
3034 [[assigning-log-levels]]
3035 ===== Assign a log level to a tracepoint definition
3037 Assign a _log level_ to a <<defining-tracepoints,tracepoint definition>>
3038 with the `LTTNG_UST_TRACEPOINT_LOGLEVEL()` macro.
3040 Assigning different levels of severity to tracepoint definitions can be
3041 useful: when you <<enabling-disabling-events,create a recording event
3042 rule>>, you can target tracepoints having a log level at least as severe
3043 as a specific value.
3045 The concept of LTTng-UST log levels is similar to the levels found
3046 in typical logging frameworks:
3048 * In a logging framework, the log level is given by the function
3049 or method name you use at the log statement site: `debug()`,
3050 `info()`, `warn()`, `error()`, and so on.
3052 * In LTTng-UST, you statically assign the log level to a tracepoint
3053 definition; any `lttng_ust_tracepoint()` macro invocation which refers
3054 to this definition has this log level.
3056 You must use `LTTNG_UST_TRACEPOINT_LOGLEVEL()` _after_ the
3057 <<defining-tracepoints,`LTTNG_UST_TRACEPOINT_EVENT()`>> or
3058 <<using-tracepoint-classes,`LTTNG_UST_TRACEPOINT_INSTANCE()`>> macro for
3061 The syntax of the `LTTNG_UST_TRACEPOINT_LOGLEVEL()` macro is:
3064 .`LTTNG_UST_TRACEPOINT_LOGLEVEL()` macro syntax.
3066 LTTNG_UST_TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
3071 * +__provider_name__+ with the tracepoint provider name.
3072 * +__tracepoint_name__+ with the tracepoint name.
3073 * +__log_level__+ with the log level to assign to the tracepoint
3074 definition named +__tracepoint_name__+ in the +__provider_name__+
3075 tracepoint provider.
3077 See man:lttng-ust(3) for a list of available log level names.
3079 .Assign the `LTTNG_UST_TRACEPOINT_LOGLEVEL_DEBUG_UNIT` log level to a tracepoint definition.
3083 /* Tracepoint definition */
3084 LTTNG_UST_TRACEPOINT_EVENT(
3091 LTTNG_UST_TP_FIELDS(
3092 lttng_ust_field_integer(int, userid, userid)
3093 lttng_ust_field_integer(size_t, len, len)
3097 /* Log level assignment */
3098 LTTNG_UST_TRACEPOINT_LOGLEVEL(my_app, get_transaction,
3099 LTTNG_UST_TRACEPOINT_LOGLEVEL_DEBUG_UNIT)
3105 ===== Create a tracepoint provider package source file
3107 A _tracepoint provider package source file_ is a C source file which
3108 includes a <<tpp-header,tracepoint provider header file>> to expand its
3109 macros into event serialization and other functions.
3111 Use the following tracepoint provider package source file template:
3114 .Tracepoint provider package source file template.
3116 #define LTTNG_UST_TRACEPOINT_CREATE_PROBES
3121 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
3122 header file>> name. You may also include more than one tracepoint
3123 provider header file here to create a tracepoint provider package
3124 holding more than one tracepoint providers.
3127 [[probing-the-application-source-code]]
3128 ==== Add tracepoints to the source code of an application
3130 Once you <<tpp-header,create a tracepoint provider header file>>, use
3131 the `lttng_ust_tracepoint()` macro in the source code of your
3132 application to insert the tracepoints that this header
3133 <<defining-tracepoints,defines>>.
3135 The `lttng_ust_tracepoint()` macro takes at least two parameters: the
3136 tracepoint provider name and the tracepoint name. The corresponding
3137 tracepoint definition defines the other parameters.
3139 .`lttng_ust_tracepoint()` usage.
3141 The following <<defining-tracepoints,tracepoint definition>> defines a
3142 tracepoint which takes two input arguments and has two output event
3146 .Tracepoint provider header file.
3148 #include "my-custom-structure.h"
3150 LTTNG_UST_TRACEPOINT_EVENT(
3155 const char *, cmd_name
3157 LTTNG_UST_TP_FIELDS(
3158 lttng_ust_field_string(cmd_name, cmd_name)
3159 lttng_ust_field_integer(int, number_of_args, argc)
3164 Refer to this tracepoint definition with the `lttng_ust_tracepoint()`
3165 macro in the source code of your application like this:
3168 .Application source file.
3172 int main(int argc, char* argv[])
3174 lttng_ust_tracepoint(my_provider, my_tracepoint, argc, argv[0]);
3179 Note how the source code of the application includes
3180 the tracepoint provider header file containing the tracepoint
3181 definitions to use, path:{tp.h}.
3184 .`lttng_ust_tracepoint()` usage with a complex tracepoint definition.
3186 Consider this complex tracepoint definition, where multiple event
3187 fields refer to the same input arguments in their argument expression
3191 .Tracepoint provider header file.
3193 /* For `struct stat` */
3194 #include <sys/types.h>
3195 #include <sys/stat.h>
3198 LTTNG_UST_TRACEPOINT_EVENT(
3206 LTTNG_UST_TP_FIELDS(
3207 lttng_ust_field_integer(int, my_constant_field, 23 + 17)
3208 lttng_ust_field_integer(int, my_int_arg_field, my_int_arg)
3209 lttng_ust_field_integer(int, my_int_arg_field2,
3210 my_int_arg * my_int_arg)
3211 lttng_ust_field_integer(int, sum4_field,
3212 my_str_arg[0] + my_str_arg[1] +
3213 my_str_arg[2] + my_str_arg[3])
3214 lttng_ust_field_string(my_str_arg_field, my_str_arg)
3215 lttng_ust_field_integer_hex(off_t, size_field, st->st_size)
3216 lttng_ust_field_float(double, size_dbl_field, (double) st->st_size)
3217 lttng_ust_field_sequence_text(char, half_my_str_arg_field,
3219 strlen(my_str_arg) / 2)
3224 Refer to this tracepoint definition with the `lttng_ust_tracepoint()`
3225 macro in the source code of your application like this:
3228 .Application source file.
3230 #define LTTNG_UST_TRACEPOINT_DEFINE
3237 stat("/etc/fstab", &s);
3238 lttng_ust_tracepoint(my_provider, my_tracepoint, 23,
3239 "Hello, World!", &s);
3245 If you look at the event record that LTTng writes when recording this
3246 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
3247 it should look like this:
3249 .Event record fields
3251 |Field name |Field value
3252 |`my_constant_field` |40
3253 |`my_int_arg_field` |23
3254 |`my_int_arg_field2` |529
3256 |`my_str_arg_field` |`Hello, World!`
3257 |`size_field` |0x12d
3258 |`size_dbl_field` |301.0
3259 |`half_my_str_arg_field` |`Hello,`
3263 Sometimes, the arguments you pass to `lttng_ust_tracepoint()` are
3264 expensive to evaluate--they use the call stack, for example. To avoid
3265 this computation when LTTng wouldn't emit any event anyway, use the
3266 `lttng_ust_tracepoint_enabled()` and `lttng_ust_do_tracepoint()` macros.
3268 The syntax of the `lttng_ust_tracepoint_enabled()` and
3269 `lttng_ust_do_tracepoint()` macros is:
3272 .`lttng_ust_tracepoint_enabled()` and `lttng_ust_do_tracepoint()` macros syntax.
3274 lttng_ust_tracepoint_enabled(provider_name, tracepoint_name)
3276 lttng_ust_do_tracepoint(provider_name, tracepoint_name, ...)
3281 * +__provider_name__+ with the tracepoint provider name.
3282 * +__tracepoint_name__+ with the tracepoint name.
3284 `lttng_ust_tracepoint_enabled()` returns a non-zero value if executing
3285 the tracepoint named `tracepoint_name` from the provider named
3286 `provider_name` _could_ make LTTng emit an event, depending on the
3287 payload of said event.
3289 `lttng_ust_do_tracepoint()` is like `lttng_ust_tracepoint()`, except
3290 that it doesn't check what `lttng_ust_tracepoint_enabled()` checks.
3291 Using `lttng_ust_tracepoint()` with `lttng_ust_tracepoint_enabled()` is
3292 dangerous because `lttng_ust_tracepoint()` also contains the
3293 `lttng_ust_tracepoint_enabled()` check; therefore, a race condition is
3294 possible in this situation:
3297 .Possible race condition when using `lttng_ust_tracepoint_enabled()` with `lttng_ust_tracepoint()`.
3299 if (lttng_ust_tracepoint_enabled(my_provider, my_tracepoint)) {
3300 stuff = prepare_stuff();
3303 lttng_ust_tracepoint(my_provider, my_tracepoint, stuff);
3306 If `lttng_ust_tracepoint_enabled()` is false, but would be true after
3307 the conditional block, then `stuff` isn't prepared: the emitted event
3308 will either contain wrong data, or the whole application could crash
3309 (with a segmentation fault, for example).
3311 NOTE: Neither `lttng_ust_tracepoint_enabled()` nor
3312 `lttng_ust_do_tracepoint()` have an `STAP_PROBEV()` call. If you need
3313 it, you must emit this call yourself.
3316 [[building-tracepoint-providers-and-user-application]]
3317 ==== Build and link a tracepoint provider package and an application
3319 Once you have one or more <<tpp-header,tracepoint provider header
3320 files>> and a <<tpp-source,tracepoint provider package source file>>,
3321 create the tracepoint provider package by compiling its source
3322 file. From here, multiple build and run scenarios are possible. The
3323 following table shows common application and library configurations
3324 along with the required command lines to achieve them.
3326 In the following diagrams, we use the following file names:
3329 Executable application.
3332 Application object file.
3335 Tracepoint provider package object file.
3338 Tracepoint provider package archive file.
3341 Tracepoint provider package shared object file.
3344 User library object file.
3347 User library shared object file.
3349 We use the following symbols in the diagrams of table below:
3352 .Symbols used in the build scenario diagrams.
3353 image::ust-sit-symbols.png[]
3355 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
3356 variable in the following instructions.
3358 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
3359 .Common tracepoint provider package scenarios.
3361 |Scenario |Instructions
3364 The instrumented application is statically linked with
3365 the tracepoint provider package object.
3367 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
3370 include::../common/ust-sit-step-tp-o.txt[]
3372 To build the instrumented application:
3374 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3379 #define LTTNG_UST_TRACEPOINT_DEFINE
3383 . Compile the application source file:
3392 . Build the application:
3397 $ gcc -o app app.o tpp.o -llttng-ust -ldl
3401 To run the instrumented application:
3403 * Start the application:
3413 The instrumented application is statically linked with the
3414 tracepoint provider package archive file.
3416 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
3419 To create the tracepoint provider package archive file:
3421 . Compile the <<tpp-source,tracepoint provider package source file>>:
3430 . Create the tracepoint provider package archive file:
3435 $ ar rcs tpp.a tpp.o
3439 To build the instrumented application:
3441 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3446 #define LTTNG_UST_TRACEPOINT_DEFINE
3450 . Compile the application source file:
3459 . Build the application:
3464 $ gcc -o app app.o tpp.a -llttng-ust -ldl
3468 To run the instrumented application:
3470 * Start the application:
3480 The instrumented application is linked with the tracepoint provider
3481 package shared object.
3483 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
3486 include::../common/ust-sit-step-tp-so.txt[]
3488 To build the instrumented application:
3490 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3495 #define LTTNG_UST_TRACEPOINT_DEFINE
3499 . Compile the application source file:
3508 . Build the application:
3513 $ gcc -o app app.o -ldl -L. -ltpp
3517 To run the instrumented application:
3519 * Start the application:
3529 The tracepoint provider package shared object is preloaded before the
3530 instrumented application starts.
3532 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3535 include::../common/ust-sit-step-tp-so.txt[]
3537 To build the instrumented application:
3539 . In path:{app.c}, before including path:{tpp.h}, add the
3545 #define LTTNG_UST_TRACEPOINT_DEFINE
3546 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3550 . Compile the application source file:
3559 . Build the application:
3564 $ gcc -o app app.o -ldl
3568 To run the instrumented application with tracing support:
3570 * Preload the tracepoint provider package shared object and
3571 start the application:
3576 $ LD_PRELOAD=./libtpp.so ./app
3580 To run the instrumented application without tracing support:
3582 * Start the application:
3592 The instrumented application dynamically loads the tracepoint provider
3593 package shared object.
3595 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3598 include::../common/ust-sit-step-tp-so.txt[]
3600 To build the instrumented application:
3602 . In path:{app.c}, before including path:{tpp.h}, add the
3608 #define LTTNG_UST_TRACEPOINT_DEFINE
3609 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3613 . Compile the application source file:
3622 . Build the application:
3627 $ gcc -o app app.o -ldl
3631 To run the instrumented application:
3633 * Start the application:
3643 The application is linked with the instrumented user library.
3645 The instrumented user library is statically linked with the tracepoint
3646 provider package object file.
3648 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3651 include::../common/ust-sit-step-tp-o-fpic.txt[]
3653 To build the instrumented user library:
3655 . In path:{emon.c}, before including path:{tpp.h}, add the
3661 #define LTTNG_UST_TRACEPOINT_DEFINE
3665 . Compile the user library source file:
3670 $ gcc -I. -fpic -c emon.c
3674 . Build the user library shared object:
3679 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3683 To build the application:
3685 . Compile the application source file:
3694 . Build the application:
3699 $ gcc -o app app.o -L. -lemon
3703 To run the application:
3705 * Start the application:
3715 The application is linked with the instrumented user library.
3717 The instrumented user library is linked with the tracepoint provider
3718 package shared object.
3720 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3723 include::../common/ust-sit-step-tp-so.txt[]
3725 To build the instrumented user library:
3727 . In path:{emon.c}, before including path:{tpp.h}, add the
3733 #define LTTNG_UST_TRACEPOINT_DEFINE
3737 . Compile the user library source file:
3742 $ gcc -I. -fpic -c emon.c
3746 . Build the user library shared object:
3751 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3755 To build the application:
3757 . Compile the application source file:
3766 . Build the application:
3771 $ gcc -o app app.o -L. -lemon
3775 To run the application:
3777 * Start the application:
3787 The tracepoint provider package shared object is preloaded before the
3790 The application is linked with the instrumented user library.
3792 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3795 include::../common/ust-sit-step-tp-so.txt[]
3797 To build the instrumented user library:
3799 . In path:{emon.c}, before including path:{tpp.h}, add the
3805 #define LTTNG_UST_TRACEPOINT_DEFINE
3806 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3810 . Compile the user library source file:
3815 $ gcc -I. -fpic -c emon.c
3819 . Build the user library shared object:
3824 $ gcc -shared -o libemon.so emon.o -ldl
3828 To build the application:
3830 . Compile the application source file:
3839 . Build the application:
3844 $ gcc -o app app.o -L. -lemon
3848 To run the application with tracing support:
3850 * Preload the tracepoint provider package shared object and
3851 start the application:
3856 $ LD_PRELOAD=./libtpp.so ./app
3860 To run the application without tracing support:
3862 * Start the application:
3872 The application is linked with the instrumented user library.
3874 The instrumented user library dynamically loads the tracepoint provider
3875 package shared object.
3877 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3880 include::../common/ust-sit-step-tp-so.txt[]
3882 To build the instrumented user library:
3884 . In path:{emon.c}, before including path:{tpp.h}, add the
3890 #define LTTNG_UST_TRACEPOINT_DEFINE
3891 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3895 . Compile the user library source file:
3900 $ gcc -I. -fpic -c emon.c
3904 . Build the user library shared object:
3909 $ gcc -shared -o libemon.so emon.o -ldl
3913 To build the application:
3915 . Compile the application source file:
3924 . Build the application:
3929 $ gcc -o app app.o -L. -lemon
3933 To run the application:
3935 * Start the application:
3945 The application dynamically loads the instrumented user library.
3947 The instrumented user library is linked with the tracepoint provider
3948 package shared object.
3950 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3953 include::../common/ust-sit-step-tp-so.txt[]
3955 To build the instrumented user library:
3957 . In path:{emon.c}, before including path:{tpp.h}, add the
3963 #define LTTNG_UST_TRACEPOINT_DEFINE
3967 . Compile the user library source file:
3972 $ gcc -I. -fpic -c emon.c
3976 . Build the user library shared object:
3981 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3985 To build the application:
3987 . Compile the application source file:
3996 . Build the application:
4001 $ gcc -o app app.o -ldl -L. -lemon
4005 To run the application:
4007 * Start the application:
4017 The application dynamically loads the instrumented user library.
4019 The instrumented user library dynamically loads the tracepoint provider
4020 package shared object.
4022 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
4025 include::../common/ust-sit-step-tp-so.txt[]
4027 To build the instrumented user library:
4029 . In path:{emon.c}, before including path:{tpp.h}, add the
4035 #define LTTNG_UST_TRACEPOINT_DEFINE
4036 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
4040 . Compile the user library source file:
4045 $ gcc -I. -fpic -c emon.c
4049 . Build the user library shared object:
4054 $ gcc -shared -o libemon.so emon.o -ldl
4058 To build the application:
4060 . Compile the application source file:
4069 . Build the application:
4074 $ gcc -o app app.o -ldl -L. -lemon
4078 To run the application:
4080 * Start the application:
4090 The tracepoint provider package shared object is preloaded before the
4093 The application dynamically loads the instrumented user library.
4095 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
4098 include::../common/ust-sit-step-tp-so.txt[]
4100 To build the instrumented user library:
4102 . In path:{emon.c}, before including path:{tpp.h}, add the
4108 #define LTTNG_UST_TRACEPOINT_DEFINE
4109 #define LTTNG_UST_TRACEPOINT_PROBE_DYNAMIC_LINKAGE
4113 . Compile the user library source file:
4118 $ gcc -I. -fpic -c emon.c
4122 . Build the user library shared object:
4127 $ gcc -shared -o libemon.so emon.o -ldl
4131 To build the application:
4133 . Compile the application source file:
4142 . Build the application:
4147 $ gcc -o app app.o -L. -lemon
4151 To run the application with tracing support:
4153 * Preload the tracepoint provider package shared object and
4154 start the application:
4159 $ LD_PRELOAD=./libtpp.so ./app
4163 To run the application without tracing support:
4165 * Start the application:
4175 The application is statically linked with the tracepoint provider
4176 package object file.
4178 The application is linked with the instrumented user library.
4180 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
4183 include::../common/ust-sit-step-tp-o.txt[]
4185 To build the instrumented user library:
4187 . In path:{emon.c}, before including path:{tpp.h}, add the
4193 #define LTTNG_UST_TRACEPOINT_DEFINE
4197 . Compile the user library source file:
4202 $ gcc -I. -fpic -c emon.c
4206 . Build the user library shared object:
4211 $ gcc -shared -o libemon.so emon.o
4215 To build the application:
4217 . Compile the application source file:
4226 . Build the application:
4231 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
4235 To run the instrumented application:
4237 * Start the application:
4247 The application is statically linked with the tracepoint provider
4248 package object file.
4250 The application dynamically loads the instrumented user library.
4252 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
4255 include::../common/ust-sit-step-tp-o.txt[]
4257 To build the application:
4259 . In path:{app.c}, before including path:{tpp.h}, add the following line:
4264 #define LTTNG_UST_TRACEPOINT_DEFINE
4268 . Compile the application source file:
4277 . Build the application:
4282 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
4287 The `--export-dynamic` option passed to the linker is necessary for the
4288 dynamically loaded library to ``see'' the tracepoint symbols defined in
4291 To build the instrumented user library:
4293 . Compile the user library source file:
4298 $ gcc -I. -fpic -c emon.c
4302 . Build the user library shared object:
4307 $ gcc -shared -o libemon.so emon.o
4311 To run the application:
4313 * Start the application:
4324 [[using-lttng-ust-with-daemons]]
4325 ===== Use noch:{LTTng-UST} with daemons
4327 If your instrumented application calls man:fork(2), man:clone(2),
4328 or BSD's man:rfork(2), without a following man:exec(3)-family
4329 system call, you must preload the path:{liblttng-ust-fork.so} shared
4330 object when you start the application.
4334 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
4337 If your tracepoint provider package is
4338 a shared library which you also preload, you must put both
4339 shared objects in env:LD_PRELOAD:
4343 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
4349 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
4351 If your instrumented application closes one or more file descriptors
4352 which it did not open itself, you must preload the
4353 path:{liblttng-ust-fd.so} shared object when you start the application:
4357 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
4360 Typical use cases include closing all the file descriptors after
4361 man:fork(2) or man:rfork(2) and buggy applications doing
4365 [[lttng-ust-pkg-config]]
4366 ===== Use noch:{pkg-config}
4368 On some distributions, LTTng-UST ships with a
4369 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
4370 metadata file. If this is your case, then use cmd:pkg-config to
4371 build an application on the command line:
4375 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
4379 [[instrumenting-32-bit-app-on-64-bit-system]]
4380 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
4382 In order to trace a 32-bit application running on a 64-bit system,
4383 LTTng must use a dedicated 32-bit
4384 <<lttng-consumerd,consumer daemon>>.
4386 The following steps show how to build and install a 32-bit consumer
4387 daemon, which is _not_ part of the default 64-bit LTTng build, how to
4388 build and install the 32-bit LTTng-UST libraries, and how to build and
4389 link an instrumented 32-bit application in that context.
4391 To build a 32-bit instrumented application for a 64-bit target system,
4392 assuming you have a fresh target system with no installed Userspace RCU
4395 . Download, build, and install a 32-bit version of Userspace RCU:
4400 $ cd $(mktemp -d) &&
4401 wget https://lttng.org/files/urcu/userspace-rcu-latest-0.13.tar.bz2 &&
4402 tar -xf userspace-rcu-latest-0.13.tar.bz2 &&
4403 cd userspace-rcu-0.13.* &&
4404 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
4406 sudo make install &&
4411 . Using the package manager of your distribution, or from source,
4412 install the 32-bit versions of the following dependencies of
4413 LTTng-tools and LTTng-UST:
4416 * https://sourceforge.net/projects/libuuid/[libuuid]
4417 * https://directory.fsf.org/wiki/Popt[popt]
4418 * https://www.xmlsoft.org/[libxml2]
4419 * **Optional**: https://github.com/numactl/numactl[numactl]
4422 . Download, build, and install a 32-bit version of the latest
4423 LTTng-UST{nbsp}{revision}:
4428 $ cd $(mktemp -d) &&
4429 wget https://lttng.org/files/lttng-ust/lttng-ust-latest-2.13.tar.bz2 &&
4430 tar -xf lttng-ust-latest-2.13.tar.bz2 &&
4431 cd lttng-ust-2.13.* &&
4432 ./configure --libdir=/usr/local/lib32 \
4433 CFLAGS=-m32 CXXFLAGS=-m32 \
4434 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
4436 sudo make install &&
4441 Add `--disable-numa` to `./configure` if you don't have
4442 https://github.com/numactl/numactl[numactl].
4446 Depending on your distribution, 32-bit libraries could be installed at a
4447 different location than `/usr/lib32`. For example, Debian is known to
4448 install some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
4450 In this case, make sure to set `LDFLAGS` to all the
4451 relevant 32-bit library paths, for example:
4455 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
4459 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
4460 the 32-bit consumer daemon:
4465 $ cd $(mktemp -d) &&
4466 wget https://lttng.org/files/lttng-tools/lttng-tools-latest-2.13.tar.bz2 &&
4467 tar -xf lttng-tools-latest-2.13.tar.bz2 &&
4468 cd lttng-tools-2.13.* &&
4469 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
4470 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
4471 --disable-bin-lttng --disable-bin-lttng-crash \
4472 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
4474 cd src/bin/lttng-consumerd &&
4475 sudo make install &&
4480 . From your distribution or from source, <<installing-lttng,install>>
4481 the 64-bit versions of LTTng-UST and Userspace RCU.
4483 . Download, build, and install the 64-bit version of the
4484 latest LTTng-tools{nbsp}{revision}:
4489 $ cd $(mktemp -d) &&
4490 wget https://lttng.org/files/lttng-tools/lttng-tools-latest-2.13.tar.bz2 &&
4491 tar -xf lttng-tools-latest-2.13.tar.bz2 &&
4492 cd lttng-tools-2.13.* &&
4493 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
4494 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
4496 sudo make install &&
4501 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
4502 when linking your 32-bit application:
4505 -m32 -L/usr/lib32 -L/usr/local/lib32 \
4506 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
4509 For example, let's rebuild the quick start example in
4510 ``<<tracing-your-own-user-application,Record user application events>>''
4511 as an instrumented 32-bit application:
4516 $ gcc -m32 -c -I. hello-tp.c
4517 $ gcc -m32 -c hello.c
4518 $ gcc -m32 -o hello hello.o hello-tp.o \
4519 -L/usr/lib32 -L/usr/local/lib32 \
4520 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
4525 No special action is required to execute the 32-bit application and
4526 for LTTng to trace it: use the command-line man:lttng(1) tool as usual.
4531 ==== Use `lttng_ust_tracef()`
4533 man:lttng_ust_tracef(3) is a small LTTng-UST API designed for quick,
4534 man:printf(3)-like instrumentation without the burden of
4535 <<tracepoint-provider,creating>> and
4536 <<building-tracepoint-providers-and-user-application,building>>
4537 a tracepoint provider package.
4539 To use `lttng_ust_tracef()` in your application:
4541 . In the C or $$C++$$ source files where you need to use
4542 `lttng_ust_tracef()`, include `<lttng/tracef.h>`:
4547 #include <lttng/tracef.h>
4551 . In the source code of the application, use `lttng_ust_tracef()` like
4552 you would use man:printf(3):
4559 lttng_ust_tracef("my message: %d (%s)", my_integer, my_string);
4565 . Link your application with `liblttng-ust`:
4570 $ gcc -o app app.c -llttng-ust
4574 To record the events that `lttng_ust_tracef()` calls emit:
4576 * <<enabling-disabling-events,Create a recording event rule>> which
4577 matches user space events named `lttng_ust_tracef:*`:
4582 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4587 .Limitations of `lttng_ust_tracef()`
4589 The `lttng_ust_tracef()` utility function was developed to make user
4590 space tracing super simple, albeit with notable disadvantages compared
4591 to <<defining-tracepoints,user-defined tracepoints>>:
4593 * All the created events have the same tracepoint provider and
4594 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4595 * There's no static type checking.
4596 * The only event record field you actually get, named `msg`, is a string
4597 potentially containing the values you passed to `lttng_ust_tracef()`
4598 using your own format string. This also means that you can't filter
4599 events with a custom expression at run time because there are no
4601 * Since `lttng_ust_tracef()` uses the man:vasprintf(3) function of the
4602 C{nbsp}standard library behind the scenes to format the strings at run
4603 time, its expected performance is lower than with user-defined
4604 tracepoints, which don't require a conversion to a string.
4606 Taking this into consideration, `lttng_ust_tracef()` is useful for some
4607 quick prototyping and debugging, but you shouldn't consider it for any
4608 permanent and serious applicative instrumentation.
4614 ==== Use `lttng_ust_tracelog()`
4616 The man:tracelog(3) API is very similar to
4617 <<tracef,`lttng_ust_tracef()`>>, with the difference that it accepts an
4618 additional log level parameter.
4620 The goal of `lttng_ust_tracelog()` is to ease the migration from logging
4623 To use `lttng_ust_tracelog()` in your application:
4625 . In the C or $$C++$$ source files where you need to use `tracelog()`,
4626 include `<lttng/tracelog.h>`:
4631 #include <lttng/tracelog.h>
4635 . In the source code of the application, use `lttng_ust_tracelog()` like
4636 you would use man:printf(3), except for the first parameter which is
4644 tracelog(LTTNG_UST_TRACEPOINT_LOGLEVEL_WARNING,
4645 "my message: %d (%s)", my_integer, my_string);
4651 See man:lttng-ust(3) for a list of available log level names.
4653 . Link your application with `liblttng-ust`:
4658 $ gcc -o app app.c -llttng-ust
4662 To record the events that `lttng_ust_tracelog()` calls emit with a log
4663 level _at least as severe as_ a specific log level:
4665 * <<enabling-disabling-events,Create a recording event rule>> which
4666 matches user space tracepoint events named `lttng_ust_tracelog:*` and
4667 with some minimum level of severity:
4672 $ lttng enable-event --userspace 'lttng_ust_tracelog:*' \
4677 To record the events that `lttng_ust_tracelog()` calls emit with a
4678 _specific log level_:
4680 * Create a recording event rule which matches tracepoint events named
4681 `lttng_ust_tracelog:*` and with a specific log level:
4686 $ lttng enable-event --userspace 'lttng_ust_tracelog:*' \
4687 --loglevel-only=INFO
4692 [[prebuilt-ust-helpers]]
4693 === Load a prebuilt user space tracing helper
4695 The LTTng-UST package provides a few helpers in the form of preloadable
4696 shared objects which automatically instrument system functions and
4699 The helper shared objects are normally found in dir:{/usr/lib}. If you
4700 built LTTng-UST <<building-from-source,from source>>, they're probably
4701 located in dir:{/usr/local/lib}.
4703 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4706 path:{liblttng-ust-libc-wrapper.so}::
4707 path:{liblttng-ust-pthread-wrapper.so}::
4708 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4709 memory and POSIX threads function tracing>>.
4711 path:{liblttng-ust-cyg-profile.so}::
4712 path:{liblttng-ust-cyg-profile-fast.so}::
4713 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4715 path:{liblttng-ust-dl.so}::
4716 <<liblttng-ust-dl,Dynamic linker tracing>>.
4718 To use a user space tracing helper with any user application:
4720 * Preload the helper shared object when you start the application:
4725 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4729 You can preload more than one helper:
4734 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4740 [[liblttng-ust-libc-pthread-wrapper]]
4741 ==== Instrument C standard library memory and POSIX threads functions
4743 The path:{liblttng-ust-libc-wrapper.so} and
4744 path:{liblttng-ust-pthread-wrapper.so} helpers
4745 add instrumentation to some C standard library and POSIX
4749 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4751 |TP provider name |TP name |Instrumented function
4753 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4754 |`calloc` |man:calloc(3)
4755 |`realloc` |man:realloc(3)
4756 |`free` |man:free(3)
4757 |`memalign` |man:memalign(3)
4758 |`posix_memalign` |man:posix_memalign(3)
4762 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4764 |TP provider name |TP name |Instrumented function
4766 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4767 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4768 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4769 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4772 When you preload the shared object, it replaces the functions listed
4773 in the previous tables by wrappers which contain tracepoints and call
4774 the replaced functions.
4777 [[liblttng-ust-cyg-profile]]
4778 ==== Instrument function entry and exit
4780 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4781 to the entry and exit points of functions.
4783 man:gcc(1) and man:clang(1) have an option named
4784 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4785 which generates instrumentation calls for entry and exit to functions.
4786 The LTTng-UST function tracing helpers,
4787 path:{liblttng-ust-cyg-profile.so} and
4788 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4789 to add tracepoints to the two generated functions (which contain
4790 `cyg_profile` in their names, hence the name of the helper).
4792 To use the LTTng-UST function tracing helper, the source files to
4793 instrument must be built using the `-finstrument-functions` compiler
4796 There are two versions of the LTTng-UST function tracing helper:
4798 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4799 that you should only use when it can be _guaranteed_ that the
4800 complete event stream is recorded without any lost event record.
4801 Any kind of duplicate information is left out.
4803 Assuming no event record is lost, having only the function addresses on
4804 entry is enough to create a call graph, since an event record always
4805 contains the ID of the CPU that generated it.
4807 Use a tool like man:addr2line(1) to convert function addresses back to
4808 source file names and line numbers.
4810 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4811 which also works in use cases where event records might get discarded or
4812 not recorded from application startup.
4813 In these cases, the trace analyzer needs more information to be
4814 able to reconstruct the program flow.
4816 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4817 points of this helper.
4819 All the tracepoints that this helper provides have the log level
4820 `LTTNG_UST_TRACEPOINT_LOGLEVEL_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4822 TIP: It's sometimes a good idea to limit the number of source files that
4823 you compile with the `-finstrument-functions` option to prevent LTTng
4824 from writing an excessive amount of trace data at run time. When using
4826 `-finstrument-functions-exclude-function-list` option to avoid
4827 instrument entries and exits of specific function names.
4832 ==== Instrument the dynamic linker
4834 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4835 man:dlopen(3) and man:dlclose(3) function calls.
4837 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4842 [[java-application]]
4843 === Instrument a Java application
4845 You can instrument any Java application which uses one of the following
4848 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4849 (JUL) core logging facilities.
4851 * https://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4852 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 isn't supported.
4855 .LTTng-UST Java agent imported by a Java application.
4856 image::java-app.png[]
4858 Note that the methods described below are new in LTTng{nbsp}2.8.
4859 Previous LTTng versions use another technique.
4861 NOTE: We use https://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4862 and https://ci.lttng.org/[continuous integration], thus this version is
4863 directly supported. However, the LTTng-UST Java agent is also tested
4864 with OpenJDK{nbsp}7.
4869 ==== Use the LTTng-UST Java agent for `java.util.logging`
4871 To use the LTTng-UST Java agent in a Java application which uses
4872 `java.util.logging` (JUL):
4874 . In the source code of the Java application, import the LTTng-UST log
4875 handler package for `java.util.logging`:
4880 import org.lttng.ust.agent.jul.LttngLogHandler;
4884 . Create an LTTng-UST `java.util.logging` log handler:
4889 Handler lttngUstLogHandler = new LttngLogHandler();
4893 . Add this handler to the `java.util.logging` loggers which should emit
4899 Logger myLogger = Logger.getLogger("some-logger");
4901 myLogger.addHandler(lttngUstLogHandler);
4905 . Use `java.util.logging` log statements and configuration as usual.
4906 The loggers with an attached LTTng-UST log handler can emit
4909 . Before exiting the application, remove the LTTng-UST log handler from
4910 the loggers attached to it and call its `close()` method:
4915 myLogger.removeHandler(lttngUstLogHandler);
4916 lttngUstLogHandler.close();
4920 This isn't strictly necessary, but it's recommended for a clean
4921 disposal of the resources of the handler.
4923 . Include the common and JUL-specific JAR files of the LTTng-UST Java agent,
4924 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4926 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4927 path] when you build the Java application.
4929 The JAR files are typically located in dir:{/usr/share/java}.
4931 IMPORTANT: The LTTng-UST Java agent must be
4932 <<installing-lttng,installed>> for the logging framework your
4935 .Use the LTTng-UST Java agent for `java.util.logging`.
4940 import java.io.IOException;
4941 import java.util.logging.Handler;
4942 import java.util.logging.Logger;
4943 import org.lttng.ust.agent.jul.LttngLogHandler;
4947 private static final int answer = 42;
4949 public static void main(String[] argv) throws Exception
4952 Logger logger = Logger.getLogger("jello");
4954 // Create an LTTng-UST log handler
4955 Handler lttngUstLogHandler = new LttngLogHandler();
4957 // Add the LTTng-UST log handler to our logger
4958 logger.addHandler(lttngUstLogHandler);
4961 logger.info("some info");
4962 logger.warning("some warning");
4964 logger.finer("finer information; the answer is " + answer);
4966 logger.severe("error!");
4968 // Not mandatory, but cleaner
4969 logger.removeHandler(lttngUstLogHandler);
4970 lttngUstLogHandler.close();
4979 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4982 <<creating-destroying-tracing-sessions,Create a recording session>>,
4983 <<enabling-disabling-events,create a recording event rule>> matching JUL
4984 events named `jello`, and <<basic-tracing-session-control,start
4990 $ lttng enable-event --jul jello
4994 Run the compiled class:
4998 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
5001 <<basic-tracing-session-control,Stop recording>> and inspect the
5011 In the resulting trace, an <<event,event record>> which a Java
5012 application using `java.util.logging` generated is named
5013 `lttng_jul:event` and has the following fields:
5022 Name of the class in which the log statement was executed.
5025 Name of the method in which the log statement was executed.
5028 Logging time (timestamp in milliseconds).
5031 Log level integer value.
5034 ID of the thread in which the log statement was executed.
5036 Use the opt:lttng-enable-event(1):--loglevel or
5037 opt:lttng-enable-event(1):--loglevel-only option of the
5038 man:lttng-enable-event(1) command to target a range of
5039 `java.util.logging` log levels or a specific `java.util.logging` log
5045 ==== Use the LTTng-UST Java agent for Apache log4j
5047 To use the LTTng-UST Java agent in a Java application which uses
5048 Apache log4j{nbsp}1.2:
5050 . In the source code of the Java application, import the LTTng-UST log
5051 appender package for Apache log4j:
5056 import org.lttng.ust.agent.log4j.LttngLogAppender;
5060 . Create an LTTng-UST log4j log appender:
5065 Appender lttngUstLogAppender = new LttngLogAppender();
5069 . Add this appender to the log4j loggers which should emit LTTng events:
5074 Logger myLogger = Logger.getLogger("some-logger");
5076 myLogger.addAppender(lttngUstLogAppender);
5080 . Use Apache log4j log statements and configuration as usual. The
5081 loggers with an attached LTTng-UST log appender can emit LTTng events.
5083 . Before exiting the application, remove the LTTng-UST log appender from
5084 the loggers attached to it and call its `close()` method:
5089 myLogger.removeAppender(lttngUstLogAppender);
5090 lttngUstLogAppender.close();
5094 This isn't strictly necessary, but it's recommended for a clean
5095 disposal of the resources of the appender.
5097 . Include the common and log4j-specific JAR
5098 files of the LTTng-UST Java agent, path:{lttng-ust-agent-common.jar} and
5099 path:{lttng-ust-agent-log4j.jar}, in the
5100 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
5101 path] when you build the Java application.
5103 The JAR files are typically located in dir:{/usr/share/java}.
5105 IMPORTANT: The LTTng-UST Java agent must be
5106 <<installing-lttng,installed>> for the logging framework your
5109 .Use the LTTng-UST Java agent for Apache log4j.
5114 import org.apache.log4j.Appender;
5115 import org.apache.log4j.Logger;
5116 import org.lttng.ust.agent.log4j.LttngLogAppender;
5120 private static final int answer = 42;
5122 public static void main(String[] argv) throws Exception
5125 Logger logger = Logger.getLogger("jello");
5127 // Create an LTTng-UST log appender
5128 Appender lttngUstLogAppender = new LttngLogAppender();
5130 // Add the LTTng-UST log appender to our logger
5131 logger.addAppender(lttngUstLogAppender);
5134 logger.info("some info");
5135 logger.warn("some warning");
5137 logger.debug("debug information; the answer is " + answer);
5139 logger.fatal("error!");
5141 // Not mandatory, but cleaner
5142 logger.removeAppender(lttngUstLogAppender);
5143 lttngUstLogAppender.close();
5149 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
5154 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
5157 <<creating-destroying-tracing-sessions,Create a recording session>>,
5158 <<enabling-disabling-events,create a recording event rule>> matching
5159 log4j events named `jello`, and <<basic-tracing-session-control,start
5165 $ lttng enable-event --log4j jello
5169 Run the compiled class:
5173 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
5176 <<basic-tracing-session-control,Stop recording>> and inspect the
5186 In the resulting trace, an <<event,event record>> which a Java
5187 application using log4j generated is named `lttng_log4j:event` and
5188 has the following fields:
5197 Name of the class in which the log statement was executed.
5200 Name of the method in which the log statement was executed.
5203 Name of the file in which the executed log statement is located.
5206 Line number at which the log statement was executed.
5212 Log level integer value.
5215 Name of the Java thread in which the log statement was executed.
5217 Use the opt:lttng-enable-event(1):--loglevel or
5218 opt:lttng-enable-event(1):--loglevel-only option of the
5219 man:lttng-enable-event(1) command to target a range of Apache log4j
5220 log levels or a specific log4j log level.
5224 [[java-application-context]]
5225 ==== Provide application-specific context fields in a Java application
5227 A Java application-specific context field is a piece of state which
5228 the Java application provides. You can <<adding-context,add>> such
5229 a context field to be recorded, using the
5230 man:lttng-add-context(1) command, to each <<event,event record>>
5231 which the log statements of this application produce.
5233 For example, a given object might have a current request ID variable.
5234 You can create a context information retriever for this object and
5235 assign a name to this current request ID. You can then, using the
5236 man:lttng-add-context(1) command, add this context field by name so that
5237 LTTng writes it to the event records of a given `java.util.logging` or
5238 log4j <<channel,channel>>.
5240 To provide application-specific context fields in a Java application:
5242 . In the source code of the Java application, import the LTTng-UST
5243 Java agent context classes and interfaces:
5248 import org.lttng.ust.agent.context.ContextInfoManager;
5249 import org.lttng.ust.agent.context.IContextInfoRetriever;
5253 . Create a context information retriever class, that is, a class which
5254 implements the `IContextInfoRetriever` interface:
5259 class MyContextInfoRetriever implements IContextInfoRetriever
5262 public Object retrieveContextInfo(String key)
5264 if (key.equals("intCtx")) {
5266 } else if (key.equals("strContext")) {
5267 return "context value!";
5276 This `retrieveContextInfo()` method is the only member of the
5277 `IContextInfoRetriever` interface. Its role is to return the current
5278 value of a state by name to create a context field. The names of the
5279 context fields and which state variables they return depends on your
5282 All primitive types and objects are supported as context fields.
5283 When `retrieveContextInfo()` returns an object, the context field
5284 serializer calls its `toString()` method to add a string field to
5285 event records. The method can also return `null`, which means that
5286 no context field is available for the required name.
5288 . Register an instance of your context information retriever class to
5289 the context information manager singleton:
5294 IContextInfoRetriever cir = new MyContextInfoRetriever();
5295 ContextInfoManager cim = ContextInfoManager.getInstance();
5296 cim.registerContextInfoRetriever("retrieverName", cir);
5300 . Before exiting the application, remove your context information
5301 retriever from the context information manager singleton:
5306 ContextInfoManager cim = ContextInfoManager.getInstance();
5307 cim.unregisterContextInfoRetriever("retrieverName");
5311 This isn't strictly necessary, but it's recommended for a clean
5312 disposal of some resources of the manager.
5314 . Build your Java application with LTTng-UST Java agent support as
5315 usual, following the procedure for either the
5316 <<jul,`java.util.logging`>> or <<log4j,Apache log4j>> framework.
5318 .Provide application-specific context fields in a Java application.
5323 import java.util.logging.Handler;
5324 import java.util.logging.Logger;
5325 import org.lttng.ust.agent.jul.LttngLogHandler;
5326 import org.lttng.ust.agent.context.ContextInfoManager;
5327 import org.lttng.ust.agent.context.IContextInfoRetriever;
5331 // Our context information retriever class
5332 private static class MyContextInfoRetriever
5333 implements IContextInfoRetriever
5336 public Object retrieveContextInfo(String key) {
5337 if (key.equals("intCtx")) {
5339 } else if (key.equals("strContext")) {
5340 return "context value!";
5347 private static final int answer = 42;
5349 public static void main(String args[]) throws Exception
5351 // Get the context information manager instance
5352 ContextInfoManager cim = ContextInfoManager.getInstance();
5354 // Create and register our context information retriever
5355 IContextInfoRetriever cir = new MyContextInfoRetriever();
5356 cim.registerContextInfoRetriever("myRetriever", cir);
5359 Logger logger = Logger.getLogger("jello");
5361 // Create an LTTng-UST log handler
5362 Handler lttngUstLogHandler = new LttngLogHandler();
5364 // Add the LTTng-UST log handler to our logger
5365 logger.addHandler(lttngUstLogHandler);
5368 logger.info("some info");
5369 logger.warning("some warning");
5371 logger.finer("finer information; the answer is " + answer);
5373 logger.severe("error!");
5375 // Not mandatory, but cleaner
5376 logger.removeHandler(lttngUstLogHandler);
5377 lttngUstLogHandler.close();
5378 cim.unregisterContextInfoRetriever("myRetriever");
5387 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
5390 <<creating-destroying-tracing-sessions,Create a recording session>> and
5391 <<enabling-disabling-events,create a recording event rule>> matching
5392 `java.util.logging` events named `jello`:
5397 $ lttng enable-event --jul jello
5400 <<adding-context,Add the application-specific context fields>> to be
5401 recorded to the event records of the `java.util.logging` channel:
5405 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
5406 $ lttng add-context --jul --type='$app.myRetriever:strContext'
5409 <<basic-tracing-session-control,Start recording>>:
5416 Run the compiled class:
5420 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
5423 <<basic-tracing-session-control,Stop recording>> and inspect the
5435 [[python-application]]
5436 === Instrument a Python application
5438 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
5440 https://docs.python.org/3/library/logging.html[`logging`] package.
5442 Each log statement creates an LTTng event once the application module
5443 imports the <<lttng-ust-agents,LTTng-UST Python agent>> package.
5446 .A Python application importing the LTTng-UST Python agent.
5447 image::python-app.png[]
5449 To use the LTTng-UST Python agent:
5451 . In the source code of the Python application, import the LTTng-UST
5461 The LTTng-UST Python agent automatically adds its logging handler to the
5462 root logger at import time.
5464 A log statement that the application executes before this import doesn't
5465 create an LTTng event.
5467 IMPORTANT: The LTTng-UST Python agent must be
5468 <<installing-lttng,installed>>.
5470 . Use log statements and logging configuration as usual.
5471 Since the LTTng-UST Python agent adds a handler to the _root_
5472 logger, any log statement from any logger can emit an LTTng event.
5474 .Use the LTTng-UST Python agent.
5485 logging.basicConfig()
5486 logger = logging.getLogger('my-logger')
5489 logger.debug('debug message')
5490 logger.info('info message')
5491 logger.warn('warn message')
5492 logger.error('error message')
5493 logger.critical('critical message')
5497 if __name__ == '__main__':
5501 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
5502 logging handler which prints to the standard error stream, isn't
5503 strictly required for LTTng-UST tracing to work, but in versions of
5504 Python preceding{nbsp}3.2, you could see a warning message which
5505 indicates that no handler exists for the logger `my-logger`.
5507 <<creating-destroying-tracing-sessions,Create a recording session>>,
5508 <<enabling-disabling-events,create a recording event rule>> matching
5509 Python logging events named `my-logger`, and
5510 <<basic-tracing-session-control,start recording>>:
5515 $ lttng enable-event --python my-logger
5519 Run the Python script:
5526 <<basic-tracing-session-control,Stop recording>> and inspect the
5536 In the resulting trace, an <<event,event record>> which a Python
5537 application generated is named `lttng_python:event` and has the
5541 Logging time (string).
5550 Name of the function in which the log statement was executed.
5553 Line number at which the log statement was executed.
5556 Log level integer value.
5559 ID of the Python thread in which the log statement was executed.
5562 Name of the Python thread in which the log statement was executed.
5564 Use the opt:lttng-enable-event(1):--loglevel or
5565 opt:lttng-enable-event(1):--loglevel-only option of the
5566 man:lttng-enable-event(1) command to target a range of Python log levels
5567 or a specific Python log level.
5569 When an application imports the LTTng-UST Python agent, the agent tries
5570 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5571 <<start-sessiond,start the session daemon>> _before_ you run the Python
5572 application. If a session daemon is found, the agent tries to register
5573 to it during five seconds, after which the application continues
5574 without LTTng tracing support. Override this timeout value with
5575 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5578 If the session daemon stops while a Python application with an imported
5579 LTTng-UST Python agent runs, the agent retries to connect and to
5580 register to a session daemon every three seconds. Override this
5581 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5586 [[proc-lttng-logger-abi]]
5587 === Use the LTTng logger
5589 The `lttng-tracer` Linux kernel module, part of
5590 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger files
5591 path:{/proc/lttng-logger} and path:{/dev/lttng-logger} (since
5592 LTTng{nbsp}2.11) when it's loaded. Any application can write text data
5593 to any of those files to create one or more LTTng events.
5596 .An application writes to the LTTng logger file to create one or more LTTng events.
5597 image::lttng-logger.png[]
5599 The LTTng logger is the quickest method--not the most efficient,
5600 however--to add instrumentation to an application. It's designed
5601 mostly to instrument shell scripts:
5605 $ echo "Some message, some $variable" > /dev/lttng-logger
5608 Any event that the LTTng logger creates is named `lttng_logger` and
5609 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5610 other instrumentation points in the kernel tracing domain, **any Unix
5611 user** can <<enabling-disabling-events,create a recording event rule>>
5612 which matches events named `lttng_logger`, not only the root user or
5613 users in the <<tracing-group,tracing group>>.
5615 To use the LTTng logger:
5617 * From any application, write text data to the path:{/dev/lttng-logger}
5620 The `msg` field of `lttng_logger` event records contains the
5623 NOTE: The maximum message length of an LTTng logger event is
5624 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5625 than one event to contain the remaining data.
5627 You shouldn't use the LTTng logger to trace a user application which you
5628 can instrument in a more efficient way, namely:
5630 * <<c-application,C and $$C++$$ applications>>.
5631 * <<java-application,Java applications>>.
5632 * <<python-application,Python applications>>.
5634 .Use the LTTng logger.
5639 echo 'Hello, World!' > /dev/lttng-logger
5641 df --human-readable --print-type / > /dev/lttng-logger
5644 <<creating-destroying-tracing-sessions,Create a recording session>>,
5645 <<enabling-disabling-events,create a recording event rule>> matching
5646 Linux kernel tracepoint events named `lttng_logger`, and
5647 <<basic-tracing-session-control,start recording>>:
5652 $ lttng enable-event --kernel lttng_logger
5656 Run the Bash script:
5663 <<basic-tracing-session-control,Stop recording>> and inspect the recorded
5674 [[instrumenting-linux-kernel]]
5675 === Instrument a Linux kernel image or module
5677 NOTE: This section shows how to _add_ instrumentation points to the
5678 Linux kernel. The subsystems of the kernel are already thoroughly
5679 instrumented at strategic points for LTTng when you
5680 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5684 [[linux-add-lttng-layer]]
5685 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5687 This section shows how to add an LTTng layer to existing ftrace
5688 instrumentation using the `TRACE_EVENT()` API.
5690 This section doesn't document the `TRACE_EVENT()` macro. Read the
5691 following articles to learn more about this API:
5693 * https://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5694 * https://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5695 * https://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5697 The following procedure assumes that your ftrace tracepoints are
5698 correctly defined in their own header and that they're created in
5699 one source file using the `CREATE_TRACE_POINTS` definition.
5701 To add an LTTng layer over an existing ftrace tracepoint:
5703 . Make sure the following kernel configuration options are
5709 * `CONFIG_HIGH_RES_TIMERS`
5710 * `CONFIG_TRACEPOINTS`
5713 . Build the Linux source tree with your custom ftrace tracepoints.
5714 . Boot the resulting Linux image on your target system.
5716 Confirm that the tracepoints exist by looking for their names in the
5717 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5718 is your subsystem name.
5720 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5725 $ cd $(mktemp -d) &&
5726 wget https://lttng.org/files/lttng-modules/lttng-modules-latest-2.13.tar.bz2 &&
5727 tar -xf lttng-modules-latest-2.13.tar.bz2 &&
5728 cd lttng-modules-2.13.*
5732 . In dir:{instrumentation/events/lttng-module}, relative to the root
5733 of the LTTng-modules source tree, create a header file named
5734 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5735 LTTng-modules tracepoint definitions using the LTTng-modules
5738 Start with this template:
5742 .path:{instrumentation/events/lttng-module/my_subsys.h}
5745 #define TRACE_SYSTEM my_subsys
5747 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5748 #define _LTTNG_MY_SUBSYS_H
5750 #include "../../../probes/lttng-tracepoint-event.h"
5751 #include <linux/tracepoint.h>
5753 LTTNG_TRACEPOINT_EVENT(
5755 * Format is identical to the TRACE_EVENT() version for the three
5756 * following macro parameters:
5759 TP_PROTO(int my_int, const char *my_string),
5760 TP_ARGS(my_int, my_string),
5762 /* LTTng-modules specific macros */
5764 ctf_integer(int, my_int_field, my_int)
5765 ctf_string(my_bar_field, my_bar)
5769 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5771 #include "../../../probes/define_trace.h"
5775 The entries in the `TP_FIELDS()` section are the list of fields for the
5776 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5777 the `TRACE_EVENT()` ftrace macro.
5779 See ``<<lttng-modules-tp-fields,Tracepoint fields macros>>'' for a
5780 complete description of the available `ctf_*()` macros.
5782 . Create the kernel module C{nbsp}source file of the LTTng-modules
5783 probe, +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5788 .path:{probes/lttng-probe-my-subsys.c}
5790 #include <linux/module.h>
5791 #include "../lttng-tracer.h"
5794 * Build-time verification of mismatch between mainline
5795 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5796 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5798 #include <trace/events/my_subsys.h>
5800 /* Create LTTng tracepoint probes */
5801 #define LTTNG_PACKAGE_BUILD
5802 #define CREATE_TRACE_POINTS
5803 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5805 #include "../instrumentation/events/lttng-module/my_subsys.h"
5807 MODULE_LICENSE("GPL and additional rights");
5808 MODULE_AUTHOR("Your name <your-email>");
5809 MODULE_DESCRIPTION("LTTng my_subsys probes");
5810 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5811 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5812 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5813 LTTNG_MODULES_EXTRAVERSION);
5817 . Edit path:{probes/KBuild} and add your new kernel module object
5818 next to the existing ones:
5822 .path:{probes/KBuild}
5826 obj-m += lttng-probe-module.o
5827 obj-m += lttng-probe-power.o
5829 obj-m += lttng-probe-my-subsys.o
5835 . Build and install the LTTng kernel modules:
5840 $ make KERNELDIR=/path/to/linux
5841 # make modules_install && depmod -a
5845 Replace `/path/to/linux` with the path to the Linux source tree where
5846 you defined and used tracepoints with the `TRACE_EVENT()` ftrace macro.
5848 Note that you can also use the
5849 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5850 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5851 C{nbsp}code that need to be executed before LTTng records the event
5854 The best way to learn how to use the previous LTTng-modules macros is to
5855 inspect the existing LTTng-modules tracepoint definitions in the
5856 dir:{instrumentation/events/lttng-module} header files. Compare them
5857 with the Linux kernel mainline versions in the
5858 dir:{include/trace/events} directory of the Linux source tree.
5862 [[lttng-tracepoint-event-code]]
5863 ===== Use custom C code to access the data for tracepoint fields
5865 Although we recommended to always use the
5866 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5867 the arguments and fields of an LTTng-modules tracepoint when possible,
5868 sometimes you need a more complex process to access the data that the
5869 tracer records as event record fields. In other words, you need local
5870 variables and multiple C{nbsp}statements instead of simple
5871 argument-based expressions that you pass to the
5872 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5874 Use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5875 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5876 a block of C{nbsp}code to be executed before LTTng records the fields.
5877 The structure of this macro is:
5880 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5882 LTTNG_TRACEPOINT_EVENT_CODE(
5884 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5885 * version for the following three macro parameters:
5888 TP_PROTO(int my_int, const char *my_string),
5889 TP_ARGS(my_int, my_string),
5891 /* Declarations of custom local variables */
5894 unsigned long b = 0;
5895 const char *name = "(undefined)";
5896 struct my_struct *my_struct;
5900 * Custom code which uses both tracepoint arguments
5901 * (in TP_ARGS()) and local variables (in TP_locvar()).
5903 * Local variables are actually members of a structure pointed
5904 * to by the special variable tp_locvar.
5908 tp_locvar->a = my_int + 17;
5909 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5910 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5911 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5912 put_my_struct(tp_locvar->my_struct);
5921 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5922 * version for this, except that tp_locvar members can be
5923 * used in the argument expression parameters of
5924 * the ctf_*() macros.
5927 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5928 ctf_integer(int, my_struct_a, tp_locvar->a)
5929 ctf_string(my_string_field, my_string)
5930 ctf_string(my_struct_name, tp_locvar->name)
5935 IMPORTANT: The C code defined in `TP_code()` must not have any side
5936 effects when executed. In particular, the code must not allocate
5937 memory or get resources without deallocating this memory or putting
5938 those resources afterwards.
5941 [[instrumenting-linux-kernel-tracing]]
5942 ==== Load and unload a custom probe kernel module
5944 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5945 kernel module>> in the kernel before it can emit LTTng events.
5947 To load the default probe kernel modules and a custom probe kernel
5950 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5951 probe modules to load when starting a root <<lttng-sessiond,session
5955 .Load the `my_subsys`, `usb`, and the default probe modules.
5959 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5964 You only need to pass the subsystem name, not the whole kernel module
5967 To load _only_ a given custom probe kernel module:
5969 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5970 modules to load when starting a root session daemon:
5973 .Load only the `my_subsys` and `usb` probe modules.
5977 # lttng-sessiond --kmod-probes=my_subsys,usb
5982 To confirm that a probe module is loaded:
5989 $ lsmod | grep lttng_probe_usb
5993 To unload the loaded probe modules:
5995 * Kill the session daemon with `SIGTERM`:
6000 # pkill lttng-sessiond
6004 You can also use the `--remove` option of man:modprobe(8) if the session
6005 daemon terminates abnormally.
6008 [[controlling-tracing]]
6011 Once an application or a Linux kernel is <<instrumenting,instrumented>>
6012 for LTTng tracing, you can _trace_ it.
6014 In the LTTng context, _tracing_ means making sure that LTTng attempts to
6015 execute some action(s) when a CPU executes an instrumentation point.
6017 This section is divided in topics on how to use the various
6018 <<plumbing,components of LTTng>>, in particular the
6019 <<lttng-cli,cmd:lttng command-line tool>>, to _control_ the LTTng
6020 daemons and tracers.
6022 NOTE: In the following subsections, we refer to an man:lttng(1) command
6023 using its man page name. For example, instead of ``Run the `create`
6024 command to'', we write ``Run the man:lttng-create(1) command to''.
6028 === Start a session daemon
6030 In some situations, you need to run a <<lttng-sessiond,session daemon>>
6031 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
6034 You will see the following error when you run a command while no session
6038 Error: No session daemon is available
6041 The only command that automatically runs a session daemon is
6042 man:lttng-create(1), which you use to
6043 <<creating-destroying-tracing-sessions,create a recording session>>. While
6044 this could be your most used first operation, sometimes it's not. Some
6047 * <<list-instrumentation-points,List the available instrumentation points>>.
6048 * <<saving-loading-tracing-session,Load a recording session configuration>>.
6049 * <<add-event-rule-matches-trigger,Add a trigger>>.
6051 All the examples above don't require a recording session to operate on.
6053 [[tracing-group]] Each Unix user can have its own running session daemon
6054 to use the user space LTTng tracer. The session daemon that the `root`
6055 user starts is the only one allowed to control the LTTng kernel tracer.
6056 Members of the Unix _tracing group_ may connect to and control the root
6057 session daemon, even for user space tracing. See the ``Session daemon
6058 connection'' section of man:lttng(1) to learn more about the Unix
6061 To start a user session daemon:
6063 * Run man:lttng-sessiond(8):
6068 $ lttng-sessiond --daemonize
6072 To start the root session daemon:
6074 * Run man:lttng-sessiond(8) as the `root` user:
6079 # lttng-sessiond --daemonize
6083 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
6084 start the session daemon in foreground.
6086 To stop a session daemon, kill its process (see man:kill(1)) with the
6087 standard `TERM` signal.
6089 Note that some Linux distributions could manage the LTTng session daemon
6090 as a service. In this case, we suggest that you use the service manager
6091 to start, restart, and stop session daemons.
6094 [[creating-destroying-tracing-sessions]]
6095 === Create and destroy a recording session
6097 Many LTTng control operations happen in the scope of a
6098 <<tracing-session,recording session>>, which is the dialogue between the
6099 <<lttng-sessiond,session daemon>> and you for everything related to
6100 <<event,event recording>>.
6102 To create a recording session with a generated name:
6104 * Use the man:lttng-create(1) command:
6113 The name of the created recording session is `auto` followed by the
6116 To create a recording session with a specific name:
6118 * Use the optional argument of the man:lttng-create(1) command:
6123 $ lttng create SESSION
6127 Replace +__SESSION__+ with your specific recording session name.
6129 In <<local-mode,local mode>>, LTTng writes the traces of a recording
6130 session to the +$LTTNG_HOME/lttng-traces/__NAME__-__DATE__-__TIME__+
6131 directory by default, where +__NAME__+ is the name of the recording
6132 session. Note that the env:LTTNG_HOME environment variable defaults to
6135 To output LTTng traces to a non-default location:
6137 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1)
6143 $ lttng create my-session --output=/tmp/some-directory
6147 You may create as many recording sessions as you wish.
6149 To list all the existing recording sessions for your Unix user, or for
6150 all users if your Unix user is `root`:
6152 * Use the man:lttng-list(1) command:
6161 [[cur-tracing-session]]When you create a recording session, the
6162 man:lttng-create(1) command sets it as the _current recording session_.
6163 The following man:lttng(1) commands operate on the current recording
6164 session when you don't specify one:
6166 [role="list-3-cols"]
6167 * man:lttng-add-context(1)
6168 * man:lttng-clear(1)
6169 * man:lttng-destroy(1)
6170 * man:lttng-disable-channel(1)
6171 * man:lttng-disable-event(1)
6172 * man:lttng-disable-rotation(1)
6173 * man:lttng-enable-channel(1)
6174 * man:lttng-enable-event(1)
6175 * man:lttng-enable-rotation(1)
6177 * man:lttng-regenerate(1)
6178 * man:lttng-rotate(1)
6180 * man:lttng-snapshot(1)
6181 * man:lttng-start(1)
6182 * man:lttng-status(1)
6184 * man:lttng-track(1)
6185 * man:lttng-untrack(1)
6188 To change the current recording session:
6190 * Use the man:lttng-set-session(1) command:
6195 $ lttng set-session SESSION
6199 Replace +__SESSION__+ with the name of the new current recording session.
6201 When you're done recording in a given recording session, destroy it.
6202 This operation frees the resources taken by the recording session to
6203 destroy; it doesn't destroy the trace data that LTTng wrote for this
6204 recording session (see ``<<clear,Clear a recording session>>'' for one
6207 To destroy the current recording session:
6209 * Use the man:lttng-destroy(1) command:
6218 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
6219 implicitly (see ``<<basic-tracing-session-control,Start and stop a
6220 recording session>>''). You need to stop recording to make LTTng flush the
6221 remaining trace data and make the trace readable.
6224 [[list-instrumentation-points]]
6225 === List the available instrumentation points
6227 The <<lttng-sessiond,session daemon>> can query the running instrumented
6228 user applications and the Linux kernel to get a list of available
6229 instrumentation points:
6231 * LTTng tracepoints and system calls for the Linux kernel
6232 <<domain,tracing domain>>.
6234 * LTTng tracepoints for the user space tracing domain.
6236 To list the available instrumentation points:
6238 . <<start-sessiond,Make sure>> there's a running
6239 <<lttng-sessiond,session daemon>> to which your Unix user can
6242 . Use the man:lttng-list(1) command with the option of the requested
6243 tracing domain amongst:
6246 opt:lttng-list(1):--kernel::
6247 Linux kernel tracepoints.
6249 Your Unix user must be `root`, or it must be a member of the Unix
6250 <<tracing-group,tracing group>>.
6252 opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall::
6253 Linux kernel system calls.
6255 Your Unix user must be `root`, or it must be a member of the Unix
6256 <<tracing-group,tracing group>>.
6258 opt:lttng-list(1):--userspace::
6259 User space tracepoints.
6261 opt:lttng-list(1):--jul::
6262 `java.util.logging` loggers.
6264 opt:lttng-list(1):--log4j::
6265 Apache log4j loggers.
6267 opt:lttng-list(1):--python::
6271 .List the available user space tracepoints.
6275 $ lttng list --userspace
6279 .List the available Linux kernel system calls.
6283 $ lttng list --kernel --syscall
6288 [[enabling-disabling-events]]
6289 === Create and enable a recording event rule
6291 Once you <<creating-destroying-tracing-sessions,create a recording
6292 session>>, you can create <<event,recording event rules>> with the
6293 man:lttng-enable-event(1) command.
6295 The man:lttng-enable-event(1) command always attaches an event rule to a
6296 <<channel,channel>> on creation. The command can create a _default
6297 channel_, named `channel0`, for you. The man:lttng-enable-event(1)
6298 command reuses the default channel each time you run it for the same
6299 tracing domain and session.
6301 A recording event rule is always enabled at creation time.
6303 The following examples show how to combine the command-line arguments of
6304 the man:lttng-enable-event(1) command to create simple to more complex
6305 recording event rules within the <<cur-tracing-session,current recording
6308 .Create a recording event rule matching specific Linux kernel tracepoint events (default channel).
6312 # lttng enable-event --kernel sched_switch
6316 .Create a recording event rule matching Linux kernel system call events with four specific names (default channel).
6320 # lttng enable-event --kernel --syscall open,write,read,close
6324 .Create recording event rules matching tracepoint events which satisfy a filter expressions (default channel).
6328 # lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
6333 # lttng enable-event --kernel --all \
6334 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
6339 $ lttng enable-event --jul my_logger \
6340 --filter='$app.retriever:cur_msg_id > 3'
6343 IMPORTANT: Make sure to always single-quote the filter string when you
6344 run man:lttng(1) from a shell.
6346 See also ``<<pid-tracking,Allow specific processes to record events>>''
6347 which offers another, more efficient filtering mechanism for process ID,
6348 user ID, and group ID attributes.
6351 .Create a recording event rule matching any user space event from the `my_app` tracepoint provider and with a log level range (default channel).
6355 $ lttng enable-event --userspace my_app:'*' --loglevel=INFO
6358 IMPORTANT: Make sure to always single-quote the wildcard character when
6359 you run man:lttng(1) from a shell.
6362 .Create a recording event rule matching user space events named specifically, but with name exclusions (default channel).
6366 $ lttng enable-event --userspace my_app:'*' \
6367 --exclude=my_app:set_user,my_app:handle_sig
6371 .Create a recording event rule matching any Apache log4j event with a specific log level (default channel).
6375 $ lttng enable-event --log4j --all --loglevel-only=WARN
6379 .Create a recording event rule, attached to a specific channel, and matching user space tracepoint events named `my_app:my_tracepoint`.
6383 $ lttng enable-event --userspace my_app:my_tracepoint \
6384 --channel=my-channel
6388 .Create a recording event rule matching user space probe events for the `malloc` function entry in path:{/usr/lib/libc.so.6}:
6392 # lttng enable-event --kernel \
6393 --userspace-probe=/usr/lib/libc.so.6:malloc \
6398 .Create a recording event rule matching user space probe events for the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
6402 # lttng enable-event --kernel \
6403 --userspace-probe=sdt:serv:server:accept_request \
6404 server_accept_request
6408 The recording event rules of a given channel form a whitelist: as soon
6409 as an event rule matches an event, LTTng emits it _once_ and therefore
6410 <<channel-overwrite-mode-vs-discard-mode,can>> record it. For example,
6411 the following rules both match user space tracepoint events named
6412 `my_app:my_tracepoint` with an `INFO` log level:
6416 $ lttng enable-event --userspace my_app:my_tracepoint
6417 $ lttng enable-event --userspace my_app:my_tracepoint \
6421 The second recording event rule is redundant: the first one includes the
6425 [[disable-event-rule]]
6426 === Disable a recording event rule
6428 To disable a <<event,recording event rule>> that you
6429 <<enabling-disabling-events,created>> previously, use the
6430 man:lttng-disable-event(1) command.
6432 man:lttng-disable-event(1) can only find recording event rules to
6433 disable by their <<instrumentation-point-types,instrumentation point
6434 type>> and event name conditions. Therefore, you cannot disable
6435 recording event rules having a specific instrumentation point log level
6436 condition, for example.
6438 LTTng doesn't emit (and, therefore, won't record) an event which only
6439 _disabled_ recording event rules match.
6441 .Disable event rules matching Python logging events from the `my-logger` logger (default <<channel,channel>>, <<cur-tracing-session,current recording session>>).
6445 $ lttng disable-event --python my-logger
6449 .Disable event rules matching all `java.util.logging` events (default channel, recording session `my-session`).
6453 $ lttng disable-event --jul --session=my-session '*'
6457 .Disable _all_ the Linux kernel recording event rules (channel `my-chan`, current recording session).
6459 The opt:lttng-disable-event(1):--all-events option isn't, like the
6460 opt:lttng-enable-event(1):--all option of the man:lttng-enable-event(1)
6461 command, an alias for the event name globbing pattern `*`: it disables
6462 _all_ the recording event rules of a given channel.
6466 # lttng disable-event --kernel --channel=my-chan --all-events
6470 NOTE: You can't _remove_ a recording event rule once you create it.
6474 === Get the status of a recording session
6476 To get the status of the <<cur-tracing-session,current recording
6477 session>>, that is, its parameters, its channels, recording event rules,
6478 and their attributes:
6480 * Use the man:lttng-status(1) command:
6489 To get the status of any recording session:
6491 * Use the man:lttng-list(1) command with the name of the recording
6497 $ lttng list SESSION
6501 Replace +__SESSION__+ with the recording session name.
6504 [[basic-tracing-session-control]]
6505 === Start and stop a recording session
6507 Once you <<creating-destroying-tracing-sessions,create a recording
6508 session>> and <<enabling-disabling-events,create one or more recording
6509 event rules>>, you can start and stop the tracers for this recording
6512 To start the <<cur-tracing-session,current recording session>>:
6514 * Use the man:lttng-start(1) command:
6523 LTTng is flexible: you can launch user applications before or after you
6524 start the tracers. An LTTng tracer only <<event,records an event>> if a
6525 recording event rule matches it, which means the tracer is active.
6527 The `start-session` <<trigger,trigger>> action can also start a recording
6530 To stop the current recording session:
6532 * Use the man:lttng-stop(1) command:
6541 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6542 records>> or lost sub-buffers since the last time you ran
6543 man:lttng-start(1), the man:lttng-stop(1) command prints corresponding
6546 IMPORTANT: You need to stop recording to make LTTng flush the remaining
6547 trace data and make the trace readable. Note that the
6548 man:lttng-destroy(1) command (see
6549 ``<<creating-destroying-tracing-sessions,Create and destroy a recording
6550 session>>'') also runs the man:lttng-stop(1) command implicitly.
6552 The `stop-session` <<trigger,trigger>> action can also stop a recording
6557 === Clear a recording session
6559 You might need to remove all the current tracing data of one or more
6560 <<tracing-session,recording sessions>> between multiple attempts to
6561 reproduce a problem without interrupting the LTTng recording activity.
6563 To clear the tracing data of the
6564 <<cur-tracing-session,current recording session>>:
6566 * Use the man:lttng-clear(1) command:
6575 To clear the tracing data of all the recording sessions:
6577 * Use the `lttng clear` command with its opt:lttng-clear(1):--all
6588 [[enabling-disabling-channels]]
6589 === Create a channel
6591 Once you <<creating-destroying-tracing-sessions,create a recording
6592 session>>, you can create a <<channel,channel>> with the
6593 man:lttng-enable-channel(1) command.
6595 Note that LTTng can automatically create a default channel when you
6596 <<enabling-disabling-events,create a recording event rule>>.
6597 Therefore, you only need to create a channel when you need non-default
6600 Specify each non-default channel attribute with a command-line
6601 option when you run the man:lttng-enable-channel(1) command.
6603 You can only create a custom channel in the Linux kernel and user space
6604 <<domain,tracing domains>>: the Java/Python logging tracing domains have
6605 their own default channel which LTTng automatically creates when you
6606 <<enabling-disabling-events,create a recording event rule>>.
6610 As of LTTng{nbsp}{revision}, you may _not_ perform the
6611 following operations with the man:lttng-enable-channel(1) command:
6613 * Change an attribute of an existing channel.
6615 * Enable a disabled channel once its recording session has been
6616 <<basic-tracing-session-control,active>> at least once.
6618 * Create a channel once its recording session has been active at
6621 * Create a user space channel with a given
6622 <<channel-buffering-schemes,buffering scheme>> and create a second
6623 user space channel with a different buffering scheme in the same
6627 The following examples show how to combine the command-line options of
6628 the man:lttng-enable-channel(1) command to create simple to more complex
6629 channels within the <<cur-tracing-session,current recording session>>.
6631 .Create a Linux kernel channel with default attributes.
6635 # lttng enable-channel --kernel my-channel
6639 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6643 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6644 --buffers-pid my-channel
6648 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6650 <<creating-destroying-tracing-sessions,Create a recording session>>,
6651 create the channel, <<enabling-disabling-events,create a recording event
6652 rule>>, and <<basic-tracing-session-control,start recording>>:
6657 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-chan
6658 $ lttng enable-event --userspace --channel=blocking-chan --all
6662 Run an application instrumented with LTTng-UST tracepoints and allow it
6667 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6671 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream.
6675 # lttng enable-channel --kernel --tracefile-count=8 \
6676 --tracefile-size=4194304 my-channel
6680 .Create a user space channel in <<overwrite-mode,overwrite>> (or ``flight recorder'') mode.
6684 $ lttng enable-channel --userspace --overwrite my-channel
6688 .<<enabling-disabling-events,Create>> the same <<event,recording event rule>> attached to two different channels.
6692 $ lttng enable-event --userspace --channel=my-channel app:tp
6693 $ lttng enable-event --userspace --channel=other-channel app:tp
6696 When a CPU executes the `app:tp` <<c-application,user space
6697 tracepoint>>, the two recording event rules above match the created
6698 event, making LTTng emit the event. Because the recording event rules
6699 are not attached to the same channel, LTTng records the event twice.
6704 === Disable a channel
6706 To disable a specific channel that you
6707 <<enabling-disabling-channels,created>> previously, use the
6708 man:lttng-disable-channel(1) command.
6710 .Disable a specific Linux kernel channel (<<cur-tracing-session,current recording session>>).
6714 # lttng disable-channel --kernel my-channel
6718 An enabled channel is an implicit <<event,recording event rule>>
6721 NOTE: As of LTTng{nbsp}{revision}, you may _not_ enable a disabled
6722 channel once its recording session has been
6723 <<basic-tracing-session-control,started>> at least once.
6727 === Add context fields to be recorded to the event records of a channel
6729 <<event,Event record>> fields in trace files provide important
6730 information about previously emitted events, but sometimes some external
6731 context may help you solve a problem faster.
6733 Examples of context fields are:
6735 * The **process ID**, **thread ID**, **process name**, and
6736 **process priority** of the thread from which LTTng emits the event.
6738 * The **hostname** of the system on which LTTng emits the event.
6740 * The Linux kernel and user call stacks (since LTTng{nbsp}2.11).
6742 * The current values of many possible **performance counters** using
6745 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6747 ** Branch instructions, misses, and loads.
6750 * Any state defined at the application level (supported for the
6751 `java.util.logging` and Apache log4j <<domain,tracing domains>>).
6753 To get the full list of available context fields:
6755 * Use the opt:lttng-add-context(1):--list option of the
6756 man:lttng-add-context(1) command:
6760 $ lttng add-context --list
6763 .Add context fields to be recorded to the event records of all the <<channel,channels>> of the <<cur-tracing-session,current recording session>>.
6765 The following command line adds the virtual process identifier and the
6766 per-thread CPU cycles count fields to all the user space channels of the
6767 current recording session.
6771 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6775 .Add performance counter context fields by raw ID
6777 See man:lttng-add-context(1) for the exact format of the context field
6778 type, which is partly compatible with the format used in
6783 # lttng add-context --userspace --type=perf:thread:raw:r0110:test
6784 # lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6788 .Add context fields to be recorded to the event records of a specific channel.
6790 The following command line adds the thread identifier and user call
6791 stack context fields to the Linux kernel channel named `my-channel` of
6792 the <<cur-tracing-session,current recording session>>.
6796 # lttng add-context --kernel --channel=my-channel \
6797 --type=tid --type=callstack-user
6801 .Add an <<java-application-context,application-specific context field>> to be recorded to the event records of a specific channel.
6803 The following command line makes sure LTTng writes the `cur_msg_id`
6804 context field of the `retriever` context retriever to all the Java
6805 logging <<event,event records>> of the channel named `my-channel`:
6809 # lttng add-context --kernel --channel=my-channel \
6810 --type='$app:retriever:cur_msg_id'
6813 IMPORTANT: Make sure to always single-quote the `$` character when you
6814 run man:lttng-add-context(1) from a shell.
6817 NOTE: You can't undo what the man:lttng-add-context(1) command does.
6822 === Allow specific processes to record events
6824 It's often useful to only allow processes with specific attributes to
6825 record events. For example, you may wish to record all the system calls
6826 which a given process makes (à la man:strace(1)).
6828 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6829 purpose. Both commands operate on _inclusion sets_ of process
6830 attributes. The available process attribute types are:
6832 Linux kernel <<domain,tracing domain>>::
6836 * Virtual process ID (VPID).
6838 This is the PID as seen by the application.
6840 * Unix user ID (UID).
6842 * Virtual Unix user ID (VUID).
6844 This is the UID as seen by the application.
6846 * Unix group ID (GID).
6848 * Virtual Unix group ID (VGID).
6850 This is the GID as seen by the application.
6852 User space tracing domain::
6858 A <<tracing-session,recording session>> has nine process
6859 attribute inclusion sets: six for the Linux kernel <<domain,tracing domain>>
6860 and three for the user space tracing domain.
6862 For a given recording session, a process{nbsp}__P__ is allowed to record
6863 LTTng events for a given <<domain,tracing domain>>{nbsp}__D__ if _all_
6864 the attributes of{nbsp}__P__ are part of the inclusion sets
6867 Whether a process is allowed or not to record LTTng events is an
6868 implicit condition of all <<event,recording event rules>>. Therefore, if
6869 LTTng creates an event{nbsp}__E__ for a given process, but this process
6870 may not record events, then no recording event rule matches{nbsp}__E__,
6871 which means LTTng won't emit and record{nbsp}__E__.
6873 When you <<creating-destroying-tracing-sessions,create a recording
6874 session>>, all its process attribute inclusion sets contain all the
6875 possible values. In other words, all processes are allowed to record
6878 Add values to an inclusion set with the man:lttng-track(1) command and
6879 remove values with the man:lttng-untrack(1) command.
6883 The process attribute values are _numeric_.
6885 Should a process with a given ID (part of an inclusion set), for
6886 example, exit, and then a new process be given this same ID, then the
6887 latter would also be allowed to record events.
6889 With the man:lttng-track(1) command, you can add Unix user and group
6890 _names_ to the user and group inclusion sets: the
6891 <<lttng-sessiond,session daemon>> finds the corresponding UID, VUID,
6892 GID, or VGID once on _addition_ to the inclusion set. This means that if
6893 you rename the user or group after you run the man:lttng-track(1)
6894 command, its user/group ID remains part of the inclusion sets.
6897 .Allow processes to record events based on their virtual process ID (VPID).
6899 For the sake of the following example, assume the target system has
6900 16{nbsp}possible VPIDs.
6903 <<creating-destroying-tracing-sessions,create a recording session>>,
6904 the user space VPID inclusion set contains _all_ the possible VPIDs:
6907 .The VPID inclusion set is full.
6908 image::track-all.png[]
6910 When the inclusion set is full and you run the man:lttng-track(1)
6911 command to specify some VPIDs, LTTng:
6913 . Clears the inclusion set.
6914 . Adds the specific VPIDs to the inclusion set.
6920 $ lttng track --userspace --vpid=3,4,7,10,13
6923 the VPID inclusion set is:
6926 .The VPID inclusion set contains the VPIDs 3, 4, 7, 10, and 13.
6927 image::track-3-4-7-10-13.png[]
6929 Add more VPIDs to the inclusion set afterwards:
6933 $ lttng track --userspace --vpid=1,15,16
6939 .VPIDs 1, 15, and 16 are added to the inclusion set.
6940 image::track-1-3-4-7-10-13-15-16.png[]
6942 The man:lttng-untrack(1) command removes entries from process attribute
6943 inclusion sets. Given the previous example, the following command:
6947 $ lttng untrack --userspace --vpid=3,7,10,13
6950 leads to this VPID inclusion set:
6953 .VPIDs 3, 7, 10, and 13 are removed from the inclusion set.
6954 image::track-1-4-15-16.png[]
6956 You can make the VPID inclusion set full again with the
6957 opt:lttng-track(1):--all option:
6961 $ lttng track --userspace --vpid --all
6964 The result is, again:
6967 .The VPID inclusion set is full.
6968 image::track-all.png[]
6971 .Allow specific processes to record events based on their user ID (UID).
6973 A typical use case with process attribute inclusion sets is to start
6974 with an empty inclusion set, then <<basic-tracing-session-control,start
6975 the tracers>>, and finally add values manually while the tracers are
6978 Use the opt:lttng-untrack(1):--all option of the
6979 man:lttng-untrack(1) command to clear the inclusion set after you
6980 <<creating-destroying-tracing-sessions,create a recording session>>, for
6981 example (with UIDs):
6985 # lttng untrack --kernel --uid --all
6991 .The UID inclusion set is empty.
6992 image::untrack-all.png[]
6994 If the LTTng tracer runs with this inclusion set configuration, it
6995 records no events within the <<cur-tracing-session,current recording
6996 session>> because no processes is allowed to do so. Use the
6997 man:lttng-track(1) command as usual to add specific values to the UID
6998 inclusion set when you need to, for example:
7002 # lttng track --kernel --uid=http,11
7008 .UIDs 6 (`http`) and 11 are part of the UID inclusion set.
7009 image::track-6-11.png[]
7014 [[saving-loading-tracing-session]]
7015 === Save and load recording session configurations
7017 Configuring a <<tracing-session,recording session>> can be long. Some of
7018 the tasks involved are:
7020 * <<enabling-disabling-channels,Create channels>> with
7021 specific attributes.
7023 * <<adding-context,Add context fields>> to be recorded to the
7024 <<event,event records>> of specific channels.
7026 * <<enabling-disabling-events,Create recording event rules>> with
7027 specific log level, filter, and other conditions.
7029 If you use LTTng to solve real world problems, chances are you have to
7030 record events using the same recording session setup over and over,
7031 modifying a few variables each time in your instrumented program or
7034 To avoid constant recording session reconfiguration, the man:lttng(1)
7035 command-line tool can save and load recording session configurations
7038 To save a given recording session configuration:
7040 * Use the man:lttng-save(1) command:
7045 $ lttng save SESSION
7049 Replace +__SESSION__+ with the name of the recording session to save.
7051 LTTng saves recording session configurations to
7052 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
7053 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. See
7054 man:lttng-save(1) to learn more about the recording session configuration
7057 LTTng saves all configuration parameters, for example:
7059 * The recording session name.
7060 * The trace data output path.
7061 * The <<channel,channels>>, with their state and all their attributes.
7062 * The context fields you added to channels.
7063 * The <<event,recording event rules>> with their state and conditions.
7065 To load a recording session:
7067 * Use the man:lttng-load(1) command:
7072 $ lttng load SESSION
7076 Replace +__SESSION__+ with the name of the recording session to load.
7078 When LTTng loads a configuration, it restores your saved recording session
7079 as if you just configured it manually.
7081 You can also save and load many sessions at a time; see
7082 man:lttng-save(1) and man:lttng-load(1) to learn more.
7085 [[sending-trace-data-over-the-network]]
7086 === Send trace data over the network
7088 LTTng can send the recorded trace data of a <<tracing-session,recording
7089 session>> to a remote system over the network instead of writing it to
7090 the local file system.
7092 To send the trace data over the network:
7094 . On the _remote_ system (which can also be the target system),
7095 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
7104 . On the _target_ system, create a recording session
7105 <<net-streaming-mode,configured>> to send trace data over the network:
7110 $ lttng create my-session --set-url=net://remote-system
7114 Replace +__remote-system__+ with the host name or IP address of the
7115 remote system. See man:lttng-create(1) for the exact URL format.
7117 . On the target system, use the man:lttng(1) command-line tool as usual.
7119 When recording is <<basic-tracing-session-control,active>>, the
7120 <<lttng-consumerd,consumer daemon>> of the target sends the contents of
7121 <<channel,sub-buffers>> to the remote relay daemon instead of flushing
7122 them to the local file system. The relay daemon writes the received
7123 packets to its local file system.
7125 See the ``Output directory'' section of man:lttng-relayd(8) to learn
7126 where a relay daemon writes its received trace data.
7131 === View events as LTTng records them (noch:{LTTng} live)
7133 _LTTng live_ is a network protocol implemented by the
7134 <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)) to allow compatible
7135 trace readers to display or analyze <<event,event records>> as LTTng
7136 records events on the target system while recording is
7137 <<basic-tracing-session-control,active>>.
7139 The relay daemon creates a _tee_: it forwards the trace data to both the
7140 local file system and to connected live readers:
7143 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live reader.
7148 . On the _target system_, create a <<tracing-session,recording session>>
7154 $ lttng create my-session --live
7158 This operation spawns a local relay daemon.
7160 . Start the live reader and configure it to connect to the relay daemon.
7162 For example, with man:babeltrace2(1):
7167 $ babeltrace2 net://localhost/host/HOSTNAME/my-session
7171 Replace +__HOSTNAME__+ with the host name of the target system.
7173 . Configure the recording session as usual with the man:lttng(1)
7174 command-line tool, and <<basic-tracing-session-control,start recording>>.
7176 List the available live recording sessions with man:babeltrace2(1):
7180 $ babeltrace2 net://localhost
7183 You can start the relay daemon on another system. In this case, you need
7184 to specify the URL of the relay daemon when you
7185 <<creating-destroying-tracing-sessions,create the recording session>> with
7186 the opt:lttng-create(1):--set-url option of the man:lttng-create(1)
7187 command. You also need to replace +__localhost__+ in the procedure above
7188 with the host name of the system on which the relay daemon runs.
7192 [[taking-a-snapshot]]
7193 === Take a snapshot of the current sub-buffers of a recording session
7195 The normal behavior of LTTng is to append full sub-buffers to growing
7196 trace data files. This is ideal to keep a full history of the events
7197 which the target system emitted, but it can represent too much data in
7200 For example, you may wish to have LTTng record your application
7201 continuously until some critical situation happens, in which case you
7202 only need the latest few recorded events to perform the desired
7203 analysis, not multi-gigabyte trace files.
7205 With the man:lttng-snapshot(1) command, you can take a _snapshot_ of the
7206 current <<channel,sub-buffers>> of a given <<tracing-session,recording
7207 session>>. LTTng can write the snapshot to the local file system or send
7208 it over the network.
7211 .A snapshot is a copy of the current sub-buffers, which LTTng does _not_ clear after the operation.
7212 image::snapshot.png[]
7214 The snapshot feature of LTTng is similar to how a
7215 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder] or the
7216 ``roll'' mode of an oscilloscope work.
7218 TIP: If you wish to create unmanaged, self-contained, non-overlapping
7219 trace chunk archives instead of a simple copy of the current
7220 sub-buffers, see the <<session-rotation,recording session rotation>>
7221 feature (available since LTTng{nbsp}2.11).
7223 To take a snapshot of the <<cur-tracing-session,current recording
7226 . Create a recording session in <<snapshot-mode,snapshot mode>>:
7231 $ lttng create my-session --snapshot
7235 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
7236 <<channel,channels>> created in this mode is automatically set to
7237 <<overwrite-mode,_overwrite_>>.
7239 . Configure the recording session as usual with the man:lttng(1)
7240 command-line tool, and <<basic-tracing-session-control,start
7243 . **Optional**: When you need to take a snapshot,
7244 <<basic-tracing-session-control,stop recording>>.
7246 You can take a snapshot when the tracers are active, but if you stop
7247 them first, you're guaranteed that the trace data in the sub-buffers
7248 doesn't change before you actually take the snapshot.
7255 $ lttng snapshot record --name=my-first-snapshot
7259 LTTng writes the current sub-buffers of all the channels of the
7260 <<cur-tracing-session,current recording session>> to
7261 trace files on the local file system. Those trace files have
7262 `my-first-snapshot` in their name.
7264 There's no difference between the format of a normal trace file and the
7265 format of a snapshot: LTTng trace readers also support LTTng snapshots.
7267 By default, LTTng writes snapshot files to the path shown by
7271 $ lttng snapshot list-output
7274 You can change this path or decide to send snapshots over the network
7277 . An output path or URL that you specify when you
7278 <<creating-destroying-tracing-sessions,create the recording session>>.
7280 . A snapshot output path or URL that you add using the
7281 `add-output` action of the man:lttng-snapshot(1) command.
7283 . An output path or URL that you provide directly to the
7284 `record` action of the man:lttng-snapshot(1) command.
7286 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
7287 you specify a URL, a <<lttng-relayd,relay daemon>> must listen on a
7288 remote system (see ``<<sending-trace-data-over-the-network,Send trace
7289 data over the network>>'').
7291 The `snapshot-session` <<trigger,trigger>> action can also take
7292 a recording session snapshot.
7296 [[session-rotation]]
7297 === Archive the current trace chunk (rotate a recording session)
7299 The <<taking-a-snapshot,snapshot user guide>> shows how to dump the
7300 current sub-buffers of a recording session to the file system or send them
7301 over the network. When you take a snapshot, LTTng doesn't clear the ring
7302 buffers of the recording session: if you take another snapshot immediately
7303 after, both snapshots could contain overlapping trace data.
7305 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
7306 _recording session rotation_ is a feature which appends the content of the
7307 ring buffers to what's already on the file system or sent over the
7308 network since the creation of the recording session or since the last
7309 rotation, and then clears those ring buffers to avoid trace data
7312 What LTTng is about to write when performing a recording session rotation
7313 is called the _current trace chunk_. When LTTng writes or sends over the
7314 network this current trace chunk, it becomes a _trace chunk archive_.
7315 Therefore, a recording session rotation operation _archives_ the current
7319 .A recording session rotation operation _archives_ the current trace chunk.
7320 image::rotation.png[]
7322 A trace chunk archive is a self-contained LTTng trace which LTTng
7323 doesn't manage anymore: you can read it, modify it, move it, or remove
7326 As of LTTng{nbsp}{revision}, there are three methods to perform a
7327 recording session rotation:
7329 * <<immediate-rotation,Immediately>>.
7331 * With a <<rotation-schedule,rotation schedule>>.
7333 * Through the execution of a `rotate-session` <<trigger,trigger>>
7336 [[immediate-rotation]]To perform an immediate rotation of the
7337 <<cur-tracing-session,current recording session>>:
7339 . <<creating-destroying-tracing-sessions,Create a recording session>> in
7340 <<local-mode,local mode>> or <<net-streaming-mode,network streaming
7341 mode>> (only those two recording session modes support recording session
7347 # lttng create my-session
7351 . <<enabling-disabling-events,Create one or more recording event rules>>
7352 and <<basic-tracing-session-control,start recording>>:
7357 # lttng enable-event --kernel sched_'*'
7362 . When needed, immediately rotate the current recording session:
7371 The man:lttng-rotate(1) command prints the path to the created trace
7372 chunk archive. See its manual page to learn about the format of trace
7373 chunk archive directory names.
7375 Perform other immediate rotations while the recording session is active.
7376 It's guaranteed that all the trace chunk archives don't contain
7377 overlapping trace data. You can also perform an immediate rotation once
7378 you have <<basic-tracing-session-control,stopped>> the recording session.
7380 . When you're done recording,
7381 <<creating-destroying-tracing-sessions,destroy the current recording
7391 The recording session destruction operation creates one last trace chunk
7392 archive from the current trace chunk.
7394 [[rotation-schedule]]A recording session rotation schedule is a planned
7395 rotation which LTTng performs automatically based on one of the
7396 following conditions:
7398 * A timer with a configured period expires.
7400 * The total size of the _flushed_ part of the current trace chunk
7401 becomes greater than or equal to a configured value.
7403 To schedule a rotation of the <<cur-tracing-session,current recording
7404 session>>, set a _rotation schedule_:
7406 . <<creating-destroying-tracing-sessions,Create a recording session>> in
7407 <<local-mode,local mode>> or <<net-streaming-mode,network streaming
7408 mode>> (only those two creation modes support recording session
7414 # lttng create my-session
7418 . <<enabling-disabling-events,Create one or more recording event rules>>:
7423 # lttng enable-event --kernel sched_'*'
7427 . Set a recording session rotation schedule:
7432 # lttng enable-rotation --timer=10s
7436 In this example, we set a rotation schedule so that LTTng performs a
7437 recording session rotation every ten seconds.
7439 See man:lttng-enable-rotation(1) to learn more about other ways to set a
7442 . <<basic-tracing-session-control,Start recording>>:
7451 LTTng performs recording session rotations automatically while the
7452 recording session is active thanks to the rotation schedule.
7454 . When you're done recording,
7455 <<creating-destroying-tracing-sessions,destroy the current recording
7465 The recording session destruction operation creates one last trace chunk
7466 archive from the current trace chunk.
7468 Unset a recording session rotation schedule with the
7469 man:lttng-disable-rotation(1) command.
7473 [[add-event-rule-matches-trigger]]
7474 === Add an ``event rule matches'' trigger to a session daemon
7476 With the man:lttng-add-trigger(1) command, you can add a
7477 <<trigger,trigger>> to a <<lttng-sessiond,session daemon>>.
7479 A trigger associates an LTTng tracing condition to one or more actions:
7480 when the condition is satisfied, LTTng attempts to execute the actions.
7482 A trigger doesn't need any <<tracing-session,recording session>> to exist:
7483 it belongs to a session daemon.
7485 As of LTTng{nbsp}{revision}, many condition types are available through
7486 the <<liblttng-ctl-lttng,`liblttng-ctl`>> C{nbsp}API, but the
7487 man:lttng-add-trigger(1) command only accepts the ``event rule matches''
7490 An ``event rule matches'' condition is satisfied when its event rule
7493 Unlike a <<event,recording event rule>>, the event rule of an
7494 ``event rule matches'' trigger condition has no implicit conditions,
7497 * It has no enabled/disabled state.
7498 * It has no attached <<channel,channel>>.
7499 * It doesn't belong to a <<tracing-session,recording session>>.
7501 Both the man:lttng-add-trigger(1) and man:lttng-enable-event(1) commands
7502 accept command-line arguments to specify an <<event-rule,event rule>>.
7503 That being said, the former is a more recent command and therefore
7504 follows the common event rule specification format (see
7505 man:lttng-event-rule(7)).
7507 .Start a <<tracing-session,recording session>> when an event rule matches.
7509 This example shows how to add the following trigger to the root
7510 <<lttng-sessiond,session daemon>>:
7513 An event rule matches a Linux kernel system call event of which the
7514 name starts with `exec` and `*/ls` matches the `filename` payload
7517 With such an event rule, LTTng emits an event when the cmd:ls program
7521 <<basic-tracing-session-control,Start the recording session>>
7524 To add such a trigger to the root session daemon:
7526 . **If there's no currently running LTTng root session daemon**, start
7531 # lttng-sessiond --daemonize
7534 . <<creating-destroying-tracing-sessions,Create a recording session>>
7536 <<enabling-disabling-events,create a recording event rule>> matching
7537 all the system call events:
7541 # lttng create pitou
7542 # lttng enable-event --kernel --syscall --all
7545 . Add the trigger to the root session daemon:
7549 # lttng add-trigger --condition=event-rule-matches \
7550 --type=syscall --name='exec*' \
7551 --filter='filename == "*/ls"' \
7552 --action=start-session pitou
7555 Confirm that the trigger exists with the man:lttng-list-triggers(1)
7560 # lttng list-triggers
7563 . Make sure the `pitou` recording session is still inactive (stopped):
7570 The first line should be something like:
7573 Recording session pitou: [inactive]
7576 Run the cmd:ls program to fire the LTTng trigger above:
7583 At this point, the `pitou` recording session should be active
7584 (started). Confirm this with the man:lttng-list(1) command again:
7591 The first line should now look like:
7594 Recording session pitou: [active]
7597 This line confirms that the LTTng trigger you added fired, therefore
7598 starting the `pitou` recording session.
7601 .[[trigger-event-notif]]Send a notification to a user application when an event rule matches.
7603 This example shows how to add the following trigger to the root
7604 <<lttng-sessiond,session daemon>>:
7607 An event rule matches a Linux kernel tracepoint event named
7608 `sched_switch` and of which the value of the `next_comm` payload
7611 With such an event rule, LTTng emits an event when Linux gives access to
7612 the processor to a process named `bash`.
7615 Send an LTTng notification to a user application.
7617 Moreover, we'll specify a _capture descriptor_ with the
7618 `event-rule-matches` trigger condition so that the user application can
7619 get the value of a specific `sched_switch` event payload field.
7621 First, write and build the user application:
7623 . Create the C{nbsp}source file of the application:
7631 #include <stdbool.h>
7634 #include <lttng/lttng.h>
7637 * Subscribes to notifications, through the notification channel
7638 * `notification_channel`, which match the condition of the trigger
7639 * named `trigger_name`.
7641 * Returns `true` on success.
7643 static bool subscribe(struct lttng_notification_channel *notification_channel,
7644 const char *trigger_name)
7646 const struct lttng_condition *condition = NULL;
7647 struct lttng_triggers *triggers = NULL;
7648 unsigned int trigger_count;
7650 enum lttng_error_code error_code;
7651 enum lttng_trigger_status trigger_status;
7654 /* Get all LTTng triggers */
7655 error_code = lttng_list_triggers(&triggers);
7656 assert(error_code == LTTNG_OK);
7658 /* Get the number of triggers */
7659 trigger_status = lttng_triggers_get_count(triggers, &trigger_count);
7660 assert(trigger_status == LTTNG_TRIGGER_STATUS_OK);
7662 /* Find the trigger named `trigger_name` */
7663 for (i = 0; i < trigger_count; i++) {
7664 const struct lttng_trigger *trigger;
7665 const char *this_trigger_name;
7667 trigger = lttng_triggers_get_at_index(triggers, i);
7668 trigger_status = lttng_trigger_get_name(trigger, &this_trigger_name);
7669 assert(trigger_status == LTTNG_TRIGGER_STATUS_OK);
7671 if (strcmp(this_trigger_name, trigger_name) == 0) {
7672 /* Trigger found: subscribe with its condition */
7673 enum lttng_notification_channel_status notification_channel_status;
7675 notification_channel_status = lttng_notification_channel_subscribe(
7676 notification_channel,
7677 lttng_trigger_get_const_condition(trigger));
7678 assert(notification_channel_status ==
7679 LTTNG_NOTIFICATION_CHANNEL_STATUS_OK);
7685 lttng_triggers_destroy(triggers);
7690 * Handles the evaluation `evaluation` of a single notification.
7692 static void handle_evaluation(const struct lttng_evaluation *evaluation)
7694 enum lttng_evaluation_status evaluation_status;
7695 const struct lttng_event_field_value *array_field_value;
7696 const struct lttng_event_field_value *string_field_value;
7697 enum lttng_event_field_value_status event_field_value_status;
7698 const char *string_field_string_value;
7700 /* Get the value of the first captured (string) field */
7701 evaluation_status = lttng_evaluation_event_rule_matches_get_captured_values(
7702 evaluation, &array_field_value);
7703 assert(evaluation_status == LTTNG_EVALUATION_STATUS_OK);
7704 event_field_value_status =
7705 lttng_event_field_value_array_get_element_at_index(
7706 array_field_value, 0, &string_field_value);
7707 assert(event_field_value_status == LTTNG_EVENT_FIELD_VALUE_STATUS_OK);
7708 assert(lttng_event_field_value_get_type(string_field_value) ==
7709 LTTNG_EVENT_FIELD_VALUE_TYPE_STRING);
7710 event_field_value_status = lttng_event_field_value_string_get_value(
7711 string_field_value, &string_field_string_value);
7712 assert(event_field_value_status == LTTNG_EVENT_FIELD_VALUE_STATUS_OK);
7714 /* Print the string value of the field */
7715 puts(string_field_string_value);
7718 int main(int argc, char *argv[])
7720 int exit_status = EXIT_SUCCESS;
7721 struct lttng_notification_channel *notification_channel;
7722 enum lttng_notification_channel_status notification_channel_status;
7723 const struct lttng_condition *condition;
7724 const char *trigger_name;
7728 trigger_name = argv[1];
7731 * Create a notification channel.
7733 * A notification channel connects the user application to the LTTng
7736 * You can use this notification channel to listen to various types
7739 notification_channel = lttng_notification_channel_create(
7740 lttng_session_daemon_notification_endpoint);
7741 assert(notification_channel);
7744 * Subscribe to notifications which match the condition of the
7745 * trigger named `trigger_name`.
7747 if (!subscribe(notification_channel, trigger_name)) {
7749 "Error: Failed to subscribe to notifications (trigger `%s`).\n",
7751 exit_status = EXIT_FAILURE;
7756 * Notification loop.
7758 * Put this in a dedicated thread to avoid blocking the main thread.
7761 struct lttng_notification *notification;
7762 enum lttng_notification_channel_status status;
7763 const struct lttng_evaluation *notification_evaluation;
7765 /* Receive the next notification */
7766 status = lttng_notification_channel_get_next_notification(
7767 notification_channel, ¬ification);
7770 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7772 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7774 * The session daemon can drop notifications if a receiving
7775 * application doesn't consume the notifications fast
7779 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7781 * The session daemon closed the notification channel.
7783 * This is typically caused by a session daemon shutting
7788 /* Unhandled conditions or errors */
7789 exit_status = EXIT_FAILURE;
7794 * Handle the condition evaluation.
7796 * A notification provides, amongst other things:
7798 * * The condition that caused LTTng to send this notification.
7800 * * The condition evaluation, which provides more specific
7801 * information on the evaluation of the condition.
7803 handle_evaluation(lttng_notification_get_evaluation(notification));
7805 /* Destroy the notification object */
7806 lttng_notification_destroy(notification);
7810 lttng_notification_channel_destroy(notification_channel);
7816 This application prints the first captured string field value of the
7817 condition evaluation of each LTTng notification it receives.
7819 . Build the `notif-app` application,
7820 using https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
7821 to provide the right compiler and linker flags:
7826 $ gcc -o notif-app notif-app.c $(pkg-config --cflags --libs lttng-ctl)
7830 Now, to add the trigger to the root session daemon:
7833 . **If there's no currently running LTTng root session daemon**, start
7838 # lttng-sessiond --daemonize
7841 . Add the trigger, naming it `sched-switch-notif`, to the root
7846 # lttng add-trigger --name=sched-switch-notif \
7847 --condition=event-rule-matches \
7848 --type=kernel --name=sched_switch \
7849 --filter='next_comm == "bash"' --capture=prev_comm \
7853 Confirm that the `sched-switch-notif` trigger exists with the
7854 man:lttng-list-triggers(1) command:
7858 # lttng list-triggers
7861 Run the cmd:notif-app application, passing the name of the trigger
7862 of which to watch the notifications:
7866 # ./notif-app sched-switch-notif
7869 Now, in an interactive Bash, type a few keys to fire the
7870 `sched-switch-notif` trigger. Watch the `notif-app` application print
7871 the previous process names.
7876 === Use the machine interface
7878 With any command of the man:lttng(1) command-line tool, set the
7879 opt:lttng(1):--mi option to `xml` (before the command name) to get an
7880 XML machine interface output, for example:
7884 $ lttng --mi=xml list my-session
7887 A schema definition (XSD) is
7888 https://github.com/lttng/lttng-tools/blob/stable-{revision}/src/common/mi-lttng-4.0.xsd[available]
7889 to ease the integration with external tools as much as possible.
7893 [[metadata-regenerate]]
7894 === Regenerate the metadata of an LTTng trace
7896 An LTTng trace, which is a https://diamon.org/ctf[CTF] trace, has both
7897 data stream files and a metadata stream file. This metadata file
7898 contains, amongst other things, information about the offset of the
7899 clock sources which LTTng uses to assign timestamps to <<event,event
7900 records>> when recording.
7902 If, once a <<tracing-session,recording session>> is
7903 <<basic-tracing-session-control,started>>, a major
7904 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
7905 happens, the clock offset of the trace also needs to be updated. Use
7906 the `metadata` item of the man:lttng-regenerate(1) command to do so.
7908 The main use case of this command is to allow a system to boot with
7909 an incorrect wall time and have LTTng trace it before its wall time
7910 is corrected. Once the system is known to be in a state where its
7911 wall time is correct, you can run `lttng regenerate metadata`.
7913 To regenerate the metadata stream files of the
7914 <<cur-tracing-session,current recording session>>:
7916 * Use the `metadata` item of the man:lttng-regenerate(1) command:
7921 $ lttng regenerate metadata
7927 [[regenerate-statedump]]
7928 === Regenerate the state dump event records of a recording session
7930 The LTTng kernel and user space tracers generate state dump
7931 <<event,event records>> when the application starts or when you
7932 <<basic-tracing-session-control,start a recording session>>.
7934 An analysis can use the state dump event records to set an initial state
7935 before it builds the rest of the state from the subsequent event
7936 records. http://tracecompass.org/[Trace Compass] and
7937 https://github.com/lttng/lttng-analyses[LTTng analyses] are notable
7938 examples of applications which use the state dump of an LTTng trace.
7940 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
7941 state dump event records aren't included in the snapshot trace files
7942 because they were recorded to a <<channel,sub-buffer>> that has been
7943 consumed or <<overwrite-mode,overwritten>> already.
7945 Use the `statedump` item of the man:lttng-regenerate(1) command to emit
7946 and record the state dump events again.
7948 To regenerate the state dump of the <<cur-tracing-session,current
7949 recording session>>, provided you created it in <<snapshot-mode,snapshot
7950 mode>>, before you take a snapshot:
7952 . Use the `statedump` item of the man:lttng-regenerate(1) command:
7957 $ lttng regenerate statedump
7961 . <<basic-tracing-session-control,Stop the recording session>>:
7970 . <<taking-a-snapshot,Take a snapshot>>:
7975 $ lttng snapshot record --name=my-snapshot
7979 Depending on the event throughput, you should run steps{nbsp}1
7980 and{nbsp}2 as closely as possible.
7984 To record the state dump events, you need to
7985 <<enabling-disabling-events,create recording event rules>> which enable
7988 * The names of LTTng-UST state dump tracepoints start with
7989 `lttng_ust_statedump:`.
7991 * The names of LTTng-modules state dump tracepoints start with
7997 [[persistent-memory-file-systems]]
7998 === Record trace data on persistent memory file systems
8000 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile
8001 random-access memory] (NVRAM) is random-access memory that retains its
8002 information when power is turned off (non-volatile). Systems with such
8003 memory can store data structures in RAM and retrieve them after a
8004 reboot, without flushing to typical _storage_.
8006 Linux supports NVRAM file systems thanks to either
8007 http://pramfs.sourceforge.net/[PRAMFS] or
8008 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
8009 (requires Linux{nbsp}4.1+).
8011 This section doesn't describe how to operate such file systems; we
8012 assume that you have a working persistent memory file system.
8014 When you <<creating-destroying-tracing-sessions,create a recording
8015 session>>, you can specify the path of the shared memory holding the
8016 sub-buffers. If you specify a location on an NVRAM file system, then you
8017 can retrieve the latest recorded trace data when the system reboots
8020 To record trace data on a persistent memory file system and retrieve the
8021 trace data after a system crash:
8023 . Create a recording session with a <<channel,sub-buffer>> shared memory
8024 path located on an NVRAM file system:
8029 $ lttng create my-session --shm-path=/path/to/shm/on/nvram
8033 . Configure the recording session as usual with the man:lttng(1)
8034 command-line tool, and <<basic-tracing-session-control,start
8037 . After a system crash, use the man:lttng-crash(1) command-line tool to
8038 read the trace data recorded on the NVRAM file system:
8043 $ lttng-crash /path/to/shm/on/nvram
8047 The binary layout of the ring buffer files isn't exactly the same as the
8048 trace files layout. This is why you need to use man:lttng-crash(1)
8049 instead of some standard LTTng trace reader.
8051 To convert the ring buffer files to LTTng trace files:
8053 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
8058 $ lttng-crash --extract=/path/to/trace /path/to/shm/on/nvram
8064 [[notif-trigger-api]]
8065 === Get notified when the buffer usage of a channel is too high or too low
8067 With the notification and <<trigger,trigger>> C{nbsp}API of
8068 <<liblttng-ctl-lttng,`liblttng-ctl`>>, LTTng can notify your user
8069 application when the buffer usage of one or more <<channel,channels>>
8070 becomes too low or too high.
8072 Use this API and enable or disable <<event,recording event rules>> while
8073 a recording session <<basic-tracing-session-control,is active>> to avoid
8074 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>, for
8077 .Send a notification to a user application when the buffer usage of an LTTng channel is too high.
8079 In this example, we create and build an application which gets notified
8080 when the buffer usage of a specific LTTng channel is higher than
8083 We only print that it's the case in this example, but we could as well
8084 use the `liblttng-ctl` C{nbsp}API to <<enabling-disabling-events,disable
8085 recording event rules>> when this happens, for example.
8087 . Create the C{nbsp}source file of the application:
8096 #include <lttng/lttng.h>
8098 int main(int argc, char *argv[])
8100 int exit_status = EXIT_SUCCESS;
8101 struct lttng_notification_channel *notification_channel;
8102 struct lttng_condition *condition;
8103 struct lttng_action *action;
8104 struct lttng_trigger *trigger;
8105 const char *recording_session_name;
8106 const char *channel_name;
8109 recording_session_name = argv[1];
8110 channel_name = argv[2];
8113 * Create a notification channel.
8115 * A notification channel connects the user application to the LTTng
8118 * You can use this notification channel to listen to various types
8121 notification_channel = lttng_notification_channel_create(
8122 lttng_session_daemon_notification_endpoint);
8125 * Create a "buffer usage becomes greater than" condition.
8127 * In this case, the condition is satisfied when the buffer usage
8128 * becomes greater than or equal to 75 %.
8130 * We create the condition for a specific recording session name,
8131 * channel name, and for the user space tracing domain.
8133 * The following condition types also exist:
8135 * * The buffer usage of a channel becomes less than a given value.
8137 * * The consumed data size of a recording session becomes greater
8138 * than a given value.
8140 * * A recording session rotation becomes ongoing.
8142 * * A recording session rotation becomes completed.
8144 * * A given event rule matches an event.
8146 condition = lttng_condition_buffer_usage_high_create();
8147 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
8148 lttng_condition_buffer_usage_set_session_name(condition,
8149 recording_session_name);
8150 lttng_condition_buffer_usage_set_channel_name(condition,
8152 lttng_condition_buffer_usage_set_domain_type(condition,
8156 * Create an action (receive a notification) to execute when the
8157 * condition created above is satisfied.
8159 action = lttng_action_notify_create();
8164 * A trigger associates a condition to an action: LTTng executes
8165 * the action when the condition is satisfied.
8167 trigger = lttng_trigger_create(condition, action);
8169 /* Register the trigger to the LTTng session daemon. */
8170 lttng_register_trigger(trigger);
8173 * Now that we have registered a trigger, LTTng will send a
8174 * notification every time its condition is met through a
8175 * notification channel.
8177 * To receive this notification, we must subscribe to notifications
8178 * which match the same condition.
8180 lttng_notification_channel_subscribe(notification_channel,
8184 * Notification loop.
8186 * Put this in a dedicated thread to avoid blocking the main thread.
8189 struct lttng_notification *notification;
8190 enum lttng_notification_channel_status status;
8191 const struct lttng_evaluation *notification_evaluation;
8192 const struct lttng_condition *notification_condition;
8193 double buffer_usage;
8195 /* Receive the next notification. */
8196 status = lttng_notification_channel_get_next_notification(
8197 notification_channel, ¬ification);
8200 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
8202 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
8204 * The session daemon can drop notifications if a monitoring
8205 * application isn't consuming the notifications fast
8209 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
8211 * The session daemon closed the notification channel.
8213 * This is typically caused by a session daemon shutting
8218 /* Unhandled conditions or errors. */
8219 exit_status = EXIT_FAILURE;
8224 * A notification provides, amongst other things:
8226 * * The condition that caused LTTng to send this notification.
8228 * * The condition evaluation, which provides more specific
8229 * information on the evaluation of the condition.
8231 * The condition evaluation provides the buffer usage
8232 * value at the moment the condition was satisfied.
8234 notification_condition = lttng_notification_get_condition(
8236 notification_evaluation = lttng_notification_get_evaluation(
8239 /* We're subscribed to only one condition. */
8240 assert(lttng_condition_get_type(notification_condition) ==
8241 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
8244 * Get the exact sampled buffer usage from the condition
8247 lttng_evaluation_buffer_usage_get_usage_ratio(
8248 notification_evaluation, &buffer_usage);
8251 * At this point, instead of printing a message, we could do
8252 * something to reduce the buffer usage of the channel, like
8253 * disable specific events, for example.
8255 printf("Buffer usage is %f %% in recording session \"%s\", "
8256 "user space channel \"%s\".\n", buffer_usage * 100,
8257 recording_session_name, channel_name);
8259 /* Destroy the notification object. */
8260 lttng_notification_destroy(notification);
8264 lttng_action_destroy(action);
8265 lttng_condition_destroy(condition);
8266 lttng_trigger_destroy(trigger);
8267 lttng_notification_channel_destroy(notification_channel);
8273 . Build the `notif-app` application, linking it with `liblttng-ctl`:
8278 $ gcc -o notif-app notif-app.c $(pkg-config --cflags --libs lttng-ctl)
8282 . <<creating-destroying-tracing-sessions,Create a recording session>>,
8283 <<enabling-disabling-events,create a recording event rule>> matching
8284 all the user space tracepoint events, and
8285 <<basic-tracing-session-control,start recording>>:
8290 $ lttng create my-session
8291 $ lttng enable-event --userspace --all
8296 If you create the channel manually with the man:lttng-enable-channel(1)
8297 command, you can set its <<channel-monitor-timer,monitor timer>> to
8298 control how frequently LTTng samples the current values of the channel
8299 properties to evaluate user conditions.
8301 . Run the `notif-app` application.
8303 This program accepts the <<tracing-session,recording session>> and
8304 user space channel names as its two first arguments. The channel
8305 which LTTng automatically creates with the man:lttng-enable-event(1)
8306 command above is named `channel0`:
8311 $ ./notif-app my-session channel0
8315 . In another terminal, run an application with a very high event
8316 throughput so that the 75{nbsp}% buffer usage condition is reached.
8318 In the first terminal, the application should print lines like this:
8321 Buffer usage is 81.45197 % in recording session "my-session", user space
8325 If you don't see anything, try to make the threshold of the condition in
8326 path:{notif-app.c} lower (0.1{nbsp}%, for example), and then rebuild the
8327 `notif-app` application (step{nbsp}2) and run it again (step{nbsp}4).
8334 [[lttng-modules-ref]]
8335 === noch:{LTTng-modules}
8339 [[lttng-tracepoint-enum]]
8340 ==== `LTTNG_TRACEPOINT_ENUM()` usage
8342 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
8346 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
8351 * `name` with the name of the enumeration (C identifier, unique
8352 amongst all the defined enumerations).
8353 * `entries` with a list of enumeration entries.
8355 The available enumeration entry macros are:
8357 +ctf_enum_value(__name__, __value__)+::
8358 Entry named +__name__+ mapped to the integral value +__value__+.
8360 +ctf_enum_range(__name__, __begin__, __end__)+::
8361 Entry named +__name__+ mapped to the range of integral values between
8362 +__begin__+ (included) and +__end__+ (included).
8364 +ctf_enum_auto(__name__)+::
8365 Entry named +__name__+ mapped to the integral value following the
8368 The last value of a `ctf_enum_value()` entry is its +__value__+
8371 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
8373 If `ctf_enum_auto()` is the first entry in the list, its integral
8376 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
8377 to use a defined enumeration as a tracepoint field.
8379 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
8383 LTTNG_TRACEPOINT_ENUM(
8386 ctf_enum_auto("AUTO: EXPECT 0")
8387 ctf_enum_value("VALUE: 23", 23)
8388 ctf_enum_value("VALUE: 27", 27)
8389 ctf_enum_auto("AUTO: EXPECT 28")
8390 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
8391 ctf_enum_auto("AUTO: EXPECT 304")
8399 [[lttng-modules-tp-fields]]
8400 ==== Tracepoint fields macros (for `TP_FIELDS()`)
8402 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
8403 tracepoint fields, which must be listed within `TP_FIELDS()` in
8404 `LTTNG_TRACEPOINT_EVENT()`, are:
8406 [role="func-desc growable",cols="asciidoc,asciidoc"]
8407 .Available macros to define LTTng-modules tracepoint fields
8409 |Macro |Description and parameters
8412 +ctf_integer(__t__, __n__, __e__)+
8414 +ctf_integer_nowrite(__t__, __n__, __e__)+
8416 +ctf_user_integer(__t__, __n__, __e__)+
8418 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
8420 Standard integer, displayed in base{nbsp}10.
8423 Integer C type (`int`, `long`, `size_t`, ...).
8429 Argument expression.
8432 +ctf_integer_hex(__t__, __n__, __e__)+
8434 +ctf_user_integer_hex(__t__, __n__, __e__)+
8436 Standard integer, displayed in base{nbsp}16.
8445 Argument expression.
8447 |+ctf_integer_oct(__t__, __n__, __e__)+
8449 Standard integer, displayed in base{nbsp}8.
8458 Argument expression.
8461 +ctf_integer_network(__t__, __n__, __e__)+
8463 +ctf_user_integer_network(__t__, __n__, __e__)+
8465 Integer in network byte order (big-endian), displayed in base{nbsp}10.
8474 Argument expression.
8477 +ctf_integer_network_hex(__t__, __n__, __e__)+
8479 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
8481 Integer in network byte order, displayed in base{nbsp}16.
8490 Argument expression.
8493 +ctf_enum(__N__, __t__, __n__, __e__)+
8495 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
8497 +ctf_user_enum(__N__, __t__, __n__, __e__)+
8499 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
8504 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
8507 Integer C type (`int`, `long`, `size_t`, ...).
8513 Argument expression.
8516 +ctf_string(__n__, __e__)+
8518 +ctf_string_nowrite(__n__, __e__)+
8520 +ctf_user_string(__n__, __e__)+
8522 +ctf_user_string_nowrite(__n__, __e__)+
8524 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
8530 Argument expression.
8533 +ctf_array(__t__, __n__, __e__, __s__)+
8535 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
8537 +ctf_user_array(__t__, __n__, __e__, __s__)+
8539 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
8541 Statically-sized array of integers.
8544 Array element C type.
8550 Argument expression.
8556 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
8558 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
8560 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
8562 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
8564 Statically-sized array of bits.
8566 The type of +__e__+ must be an integer type. +__s__+ is the number
8567 of elements of such type in +__e__+, not the number of bits.
8570 Array element C type.
8576 Argument expression.
8582 +ctf_array_text(__t__, __n__, __e__, __s__)+
8584 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
8586 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
8588 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
8590 Statically-sized array, printed as text.
8592 The string doesn't need to be null-terminated.
8595 Array element C type (always `char`).
8601 Argument expression.
8607 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
8609 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
8611 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
8613 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
8615 Dynamically-sized array of integers.
8617 The type of +__E__+ must be unsigned.
8620 Array element C type.
8626 Argument expression.
8629 Length expression C type.
8635 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
8637 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
8639 Dynamically-sized array of integers, displayed in base{nbsp}16.
8641 The type of +__E__+ must be unsigned.
8644 Array element C type.
8650 Argument expression.
8653 Length expression C type.
8658 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
8660 Dynamically-sized array of integers in network byte order (big-endian),
8661 displayed in base{nbsp}10.
8663 The type of +__E__+ must be unsigned.
8666 Array element C type.
8672 Argument expression.
8675 Length expression C type.
8681 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
8683 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
8685 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
8687 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
8689 Dynamically-sized array of bits.
8691 The type of +__e__+ must be an integer type. +__s__+ is the number
8692 of elements of such type in +__e__+, not the number of bits.
8694 The type of +__E__+ must be unsigned.
8697 Array element C type.
8703 Argument expression.
8706 Length expression C type.
8712 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
8714 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
8716 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
8718 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
8720 Dynamically-sized array, displayed as text.
8722 The string doesn't need to be null-terminated.
8724 The type of +__E__+ must be unsigned.
8726 The behaviour is undefined if +__e__+ is `NULL`.
8729 Sequence element C type (always `char`).
8735 Argument expression.
8738 Length expression C type.
8744 Use the `_user` versions when the argument expression, `e`, is
8745 a user space address. In the cases of `ctf_user_integer*()` and
8746 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
8749 The `_nowrite` versions omit themselves from the trace data, but are
8750 otherwise identical. This means LTTng won't write the `_nowrite` fields
8751 to the recorded trace. Their primary purpose is to make some of the
8752 event context available to the <<enabling-disabling-events,recording
8753 event rule filters>> without having to commit the data to
8754 <<channel,sub-buffers>>.
8760 Terms related to LTTng and to tracing in general:
8762 [[def-action]]action::
8763 The part of a <<def-trigger,trigger>> which LTTng executes when the
8764 trigger <<def-condition,condition>> is satisfied.
8767 The https://diamon.org/babeltrace[Babeltrace] project, which includes:
8770 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]
8771 command-line interface.
8772 * The libbabeltrace2 library which offers a
8773 https://babeltrace.org/docs/v2.0/libbabeltrace2/[C API].
8774 * https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings].
8777 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
8778 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
8780 [[def-channel]]<<channel,channel>>::
8781 An entity which is responsible for a set of
8782 <<def-ring-buffer,ring buffers>>.
8784 <<def-recording-event-rule,Recording event rules>> are always attached
8785 to a specific channel.
8788 A source of time for a <<def-tracer,tracer>>.
8790 [[def-condition]]condition::
8791 The part of a <<def-trigger,trigger>> which must be satisfied for
8792 LTTng to attempt to execute the trigger <<def-action,actions>>.
8794 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
8795 A program which is responsible for consuming the full
8796 <<def-sub-buffer,sub-buffers>> and write them to a file system or
8797 send them over the network.
8799 [[def-current-trace-chunk]]current trace chunk::
8800 A <<def-trace-chunk,trace chunk>> which includes the current content
8801 of all the <<def-sub-buffer,sub-buffers>> of the
8802 <<def-tracing-session,recording session>> and the stream files
8803 produced since the latest event amongst:
8805 * The creation of the recording session.
8806 * The last <<def-tracing-session-rotation,recording session rotation>>, if
8809 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
8810 The <<def-event-record-loss-mode,event record loss mode>> in which
8811 the <<def-tracer,tracer>> _discards_ new <<def-event-record,event
8812 records>> when there's no <<def-sub-buffer,sub-buffer>> space left to
8815 [[def-event]]event::
8816 The execution of an <<def-instrumentation-point,instrumentation
8817 point>>, like a <<def-tracepoint,tracepoint>> that you manually place
8818 in some source code, or a Linux kprobe.
8820 When an instrumentation point is executed, LTTng creates an event.
8822 When an <<def-event-rule,event rule>> matches the event,
8823 <<def-lttng,LTTng>> executes some action, for example:
8825 * Record its payload to a <<def-sub-buffer,sub-buffer>> as an
8826 <<def-event-record,event record>>.
8827 * Attempt to execute the user-defined actions of a
8828 <<def-trigger,trigger>> with an
8829 <<add-event-rule-matches-trigger,``event rule matches''>> condition.
8831 [[def-event-name]]event name::
8832 The name of an <<def-event,event>>, which is also the name of the
8833 <<def-event-record,event record>>.
8835 This is also called the _instrumentation point name_.
8837 [[def-event-record]]event record::
8838 A record (binary serialization), in a <<def-trace,trace>>, of the
8839 payload of an <<def-event,event>>.
8841 The payload of an event record has zero or more _fields_.
8843 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
8844 The mechanism by which event records of a given
8845 <<def-channel,channel>> are lost (not recorded) when there's no
8846 <<def-sub-buffer,sub-buffer>> space left to store them.
8848 [[def-event-rule]]<<event-rule,event rule>>::
8849 Set of conditions which an <<def-event,event>> must satisfy
8850 for LTTng to execute some action.
8852 An event rule is said to _match_ events, like a
8853 https://en.wikipedia.org/wiki/Regular_expression[regular expression]
8856 A <<def-recording-event-rule,recording event rule>> is a specific type
8857 of event rule of which the action is to <<def-record,record>> the event
8858 to a <<def-sub-buffer,sub-buffer>>.
8860 [[def-incl-set]]inclusion set::
8861 In the <<pid-tracking,process attribute inclusion set>> context: a
8862 set of <<def-proc-attr,process attributes>> of a given type.
8864 <<instrumenting,instrumentation>>::
8865 The use of <<def-lttng,LTTng>> probes to make a kernel or
8866 <<def-user-application,user application>> traceable.
8868 [[def-instrumentation-point]]instrumentation point::
8869 A point in the execution path of a kernel or
8870 <<def-user-application,user application>> which, when executed,
8871 create an <<def-event,event>>.
8873 instrumentation point name::
8874 See _<<def-event-name,event name>>_.
8876 `java.util.logging`::
8878 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities]
8879 of the Java platform.
8882 A https://logging.apache.org/log4j/1.2/[logging library] for Java
8883 developed by the Apache Software Foundation.
8886 Level of severity of a log statement or user space
8887 <<def-instrumentation-point,instrumentation point>>.
8889 [[def-lttng]]LTTng::
8890 The _Linux Trace Toolkit: next generation_ project.
8892 <<lttng-cli,cmd:lttng>>::
8893 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
8894 project which you can use to send and receive control messages to and
8895 from a <<def-session-daemon,session daemon>>.
8898 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
8899 which is a set of analyzing programs that you can use to obtain a
8900 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
8902 cmd:lttng-consumerd::
8903 The name of the <<def-consumer-daemon,consumer daemon>> program.
8906 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
8907 which can convert <<def-ring-buffer,ring buffer>> files (usually
8908 <<persistent-memory-file-systems,saved on a persistent memory file
8909 system>>) to <<def-trace,trace>> files.
8911 See man:lttng-crash(1).
8913 LTTng Documentation::
8916 <<lttng-live,LTTng live>>::
8917 A communication protocol between the <<lttng-relayd,relay daemon>> and
8918 live readers which makes it possible to show or analyze
8919 <<def-event-record,event records>> ``live'', as they're received by
8920 the <<def-relay-daemon,relay daemon>>.
8922 <<lttng-modules,LTTng-modules>>::
8923 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
8924 which contains the Linux kernel modules to make the Linux kernel
8925 <<def-instrumentation-point,instrumentation points>> available for
8926 <<def-lttng,LTTng>> tracing.
8929 The name of the <<def-relay-daemon,relay daemon>> program.
8931 cmd:lttng-sessiond::
8932 The name of the <<def-session-daemon,session daemon>> program.
8934 [[def-lttng-tools]]LTTng-tools::
8935 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
8936 contains the various programs and libraries used to
8937 <<controlling-tracing,control tracing>>.
8939 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
8940 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
8941 contains libraries to instrument
8942 <<def-user-application,user applications>>.
8944 <<lttng-ust-agents,LTTng-UST Java agent>>::
8945 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
8946 allow the LTTng instrumentation of `java.util.logging` and Apache
8947 log4j{nbsp}1.2 logging statements.
8949 <<lttng-ust-agents,LTTng-UST Python agent>>::
8950 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
8951 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
8954 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
8955 The <<def-event-record-loss-mode,event record loss mode>> in which new
8956 <<def-event-record,event records>> _overwrite_ older event records
8957 when there's no <<def-sub-buffer,sub-buffer>> space left to store
8960 <<channel-buffering-schemes,per-process buffering>>::
8961 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
8962 process has its own <<def-sub-buffer,sub-buffers>> for a given user
8963 space <<def-channel,channel>>.
8965 <<channel-buffering-schemes,per-user buffering>>::
8966 A <<def-buffering-scheme,buffering scheme>> in which all the processes
8967 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
8968 given user space <<def-channel,channel>>.
8970 [[def-proc-attr]]process attribute::
8971 In the <<pid-tracking,process attribute inclusion set>> context:
8974 * A virtual process ID.
8976 * A virtual Unix user ID.
8978 * A virtual Unix group ID.
8981 See <<def-event-record,_event record_>>.
8983 [[def-record]]record (_verb_)::
8984 Serialize the binary payload of an <<def-event,event>> to a
8985 <<def-sub-buffer,sub-buffer>>.
8987 [[def-recording-event-rule]]<<event,recording event rule>>::
8988 Specific type of <<def-event-rule,event rule>> of which the action is
8989 to <<def-record,record>> the matched event to a
8990 <<def-sub-buffer,sub-buffer>>.
8992 [[def-tracing-session]][[def-recording-session]]<<tracing-session,recording session>>::
8993 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
8995 [[def-tracing-session-rotation]]<<session-rotation,recording session rotation>>::
8996 The action of archiving the
8997 <<def-current-trace-chunk,current trace chunk>> of a
8998 <<def-tracing-session,recording session>>.
9000 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
9001 A process which is responsible for receiving the <<def-trace,trace>>
9002 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
9004 [[def-ring-buffer]]ring buffer::
9005 A set of <<def-sub-buffer,sub-buffers>>.
9008 See _<<def-tracing-session-rotation,recording session rotation>>_.
9010 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
9011 A process which receives control commands from you and orchestrates
9012 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
9014 <<taking-a-snapshot,snapshot>>::
9015 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
9016 of a given <<def-tracing-session,recording session>>, saved as
9017 <<def-trace,trace>> files.
9019 [[def-sub-buffer]]sub-buffer::
9020 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
9021 which contains <<def-event-record,event records>>.
9024 The time information attached to an <<def-event,event>> when LTTng
9027 [[def-trace]]trace (_noun_)::
9030 * One https://diamon.org/ctf/[CTF] metadata stream file.
9031 * One or more CTF data stream files which are the concatenations of one
9032 or more flushed <<def-sub-buffer,sub-buffers>>.
9034 [[def-trace-verb]]trace (_verb_)::
9035 From the perspective of a <<def-tracer,tracer>>: attempt to execute
9036 one or more actions when emitting an <<def-event,event>> in an
9037 application or in a system.
9039 [[def-trace-chunk]]trace chunk::
9040 A self-contained <<def-trace,trace>> which is part of a
9041 <<def-tracing-session,recording session>>. Each
9042 <<def-tracing-session-rotation, recording session rotation>> produces a
9043 <<def-trace-chunk-archive,trace chunk archive>>.
9045 [[def-trace-chunk-archive]]trace chunk archive::
9046 The result of a <<def-tracing-session-rotation, recording session
9049 <<def-lttng,LTTng>> doesn't manage any trace chunk archive, even if its
9050 containing <<def-tracing-session,recording session>> is still active: you
9051 are free to read it, modify it, move it, or remove it.
9054 The http://tracecompass.org[Trace Compass] project and application.
9056 [[def-tracepoint]]tracepoint::
9057 An instrumentation point using the tracepoint mechanism of the Linux
9058 kernel or of <<def-lttng-ust,LTTng-UST>>.
9060 tracepoint definition::
9061 The definition of a single <<def-tracepoint,tracepoint>>.
9064 The name of a <<def-tracepoint,tracepoint>>.
9066 [[def-tracepoint-provider]]tracepoint provider::
9067 A set of functions providing <<def-tracepoint,tracepoints>> to an
9068 instrumented <<def-user-application,user application>>.
9070 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
9071 provider package>>: many tracepoint providers can exist within a
9072 tracepoint provider package.
9074 [[def-tracepoint-provider-package]]tracepoint provider package::
9075 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
9076 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
9077 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
9080 [[def-tracer]]tracer::
9081 A piece of software which executes some action when it emits
9082 an <<def-event,event>>, like <<def-record,record>> it to some
9085 <<domain,tracing domain>>::
9086 A type of LTTng <<def-tracer,tracer>>.
9088 <<tracing-group,tracing group>>::
9089 The Unix group which a Unix user can be part of to be allowed to
9090 control the Linux kernel LTTng <<def-tracer,tracer>>.
9092 [[def-trigger]]<<trigger,trigger>>::
9093 A <<def-condition,condition>>-<<def-action,actions>> pair; when the
9094 condition of a trigger is satisfied, LTTng attempts to execute its
9097 [[def-user-application]]user application::
9098 An application (program or library) running in user space, as opposed
9099 to a Linux kernel module, for example.