1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 LTTng{nbsp}{revision} bears the name _Joannès_. A Berliner Weisse style
78 beer from the http://letreflenoir.com/[Trèfle Noir] microbrewery in
79 https://en.wikipedia.org/wiki/Rouyn-Noranda[Rouyn-Noranda], the
80 https://www.beeradvocate.com/beer/profile/20537/238967/[_**Joannès**_]
81 is a tangy beer with a distinct pink dress and intense fruit flavor,
82 thanks to the presence of fresh blackcurrant grown in Témiscamingue.
84 New features and changes in LTTng{nbsp}{revision}:
86 * **Tracing control**:
87 ** You can override the name or the URL of a tracing session
88 configuration when you use man:lttng-load(1) thanks to the new
89 opt:lttng-load(1):--override-name and
90 opt:lttng-load(1):--override-url options.
91 ** The new `lttng regenerate` command replaces the now deprecated
92 `lttng metadata` command of LTTng 2.8. man:lttng-regenerate(1) can
93 also <<regenerate-statedump,generate the state dump event records>>
94 of a given tracing session on demand, a handy feature when
95 <<taking-a-snapshot,taking a snapshot>>.
96 ** You can add PMU counters by raw ID with man:lttng-add-context(1):
101 lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
105 The format of the raw ID is the same as used with man:perf-record(1).
106 See <<adding-context,Add context fields to a channel>> for more
109 ** The LTTng <<lttng-relayd,relay daemon>> is now supported on
110 OS{nbsp}X and macOS for a smoother integration within a trace
111 analysis workflow, regardless of the platform used.
113 * **User space tracing**:
114 ** Improved performance (tested on x86-64 and ARMv7-A
115 (https://en.wikipedia.org/wiki/Cubieboard[Cubieboard])
117 ** New helper library (`liblttng-ust-fd`) to help with
118 <<liblttng-ust-fd,applications which close file descriptors that
119 don't belong to them>>, for example, in a loop which closes file
120 descriptors after man:fork(2), or BSD's `closeall()`.
121 ** More accurate <<liblttng-ust-dl,dynamic linker instrumentation>> and
122 state dump event records, especially when a dynamically loaded
123 library manually loads its own dependencies.
124 ** New `ctf_*()` field definition macros (see man:lttng-ust(3)):
125 *** `ctf_array_hex()`
126 *** `ctf_array_network()`
127 *** `ctf_array_network_hex()`
128 *** `ctf_sequence_hex()`
129 *** `ctf_sequence_network()`
130 *** `ctf_sequence_network_hex()`
131 ** New `lttng_ust_loaded` weak symbol defined by `liblttng-ust` for
132 an application to know if the LTTng-UST shared library is loaded
140 int lttng_ust_loaded __attribute__((weak));
144 if (lttng_ust_loaded) {
145 puts("LTTng-UST is loaded!");
147 puts("LTTng-UST is not loaded!");
155 ** LTTng-UST thread names have the `-ust` suffix.
157 * **Linux kernel tracing**:
158 ** Improved performance (tested on x86-64 and ARMv7-A
159 (https://en.wikipedia.org/wiki/Cubieboard[Cubieboard])
161 ** New enumeration <<lttng-modules-tp-fields,field definition macros>>:
162 `ctf_enum()` and `ctf_user_enum()`.
163 ** IPv4, IPv6, and TCP header data is recorded in the event records
164 produced by tracepoints starting with `net_`.
165 ** Detailed system call event records: `select`, `pselect6`, `poll`,
166 `ppoll`, `epoll_wait`, `epoll_pwait`, and `epoll_ctl` on all
167 architectures supported by LTTng-modules, and `accept4` on x86-64.
168 ** New I²C instrumentation: the `extract_sensitive_payload` parameter
169 of the new `lttng-probe-i2c` LTTng module controls whether or not
170 the payloads of I²C messages are recorded in I²C event records, since
171 they may contain sensitive data (for example, keystrokes).
172 ** When the LTTng kernel modules are built into the Linux kernel image,
173 the `CONFIG_TRACEPOINTS` configuration option is automatically
180 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
181 generation_ is a modern toolkit for tracing Linux systems and
182 applications. So your first question might be:
189 As the history of software engineering progressed and led to what
190 we now take for granted--complex, numerous and
191 interdependent software applications running in parallel on
192 sophisticated operating systems like Linux--the authors of such
193 components, software developers, began feeling a natural
194 urge to have tools that would ensure the robustness and good performance
195 of their masterpieces.
197 One major achievement in this field is, inarguably, the
198 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
199 an essential tool for developers to find and fix bugs. But even the best
200 debugger won't help make your software run faster, and nowadays, faster
201 software means either more work done by the same hardware, or cheaper
202 hardware for the same work.
204 A _profiler_ is often the tool of choice to identify performance
205 bottlenecks. Profiling is suitable to identify _where_ performance is
206 lost in a given software. The profiler outputs a profile, a statistical
207 summary of observed events, which you may use to discover which
208 functions took the most time to execute. However, a profiler won't
209 report _why_ some identified functions are the bottleneck. Bottlenecks
210 might only occur when specific conditions are met, conditions that are
211 sometimes impossible to capture by a statistical profiler, or impossible
212 to reproduce with an application altered by the overhead of an
213 event-based profiler. For a thorough investigation of software
214 performance issues, a history of execution is essential, with the
215 recorded values of variables and context fields you choose, and
216 with as little influence as possible on the instrumented software. This
217 is where tracing comes in handy.
219 _Tracing_ is a technique used to understand what goes on in a running
220 software system. The software used for tracing is called a _tracer_,
221 which is conceptually similar to a tape recorder. When recording,
222 specific instrumentation points placed in the software source code
223 generate events that are saved on a giant tape: a _trace_ file. You
224 can trace user applications and the operating system at the same time,
225 opening the possibility of resolving a wide range of problems that would
226 otherwise be extremely challenging.
228 Tracing is often compared to _logging_. However, tracers and loggers are
229 two different tools, serving two different purposes. Tracers are
230 designed to record much lower-level events that occur much more
231 frequently than log messages, often in the range of thousands per
232 second, with very little execution overhead. Logging is more appropriate
233 for a very high-level analysis of less frequent events: user accesses,
234 exceptional conditions (errors and warnings, for example), database
235 transactions, instant messaging communications, and such. Simply put,
236 logging is one of the many use cases that can be satisfied with tracing.
238 The list of recorded events inside a trace file can be read manually
239 like a log file for the maximum level of detail, but it is generally
240 much more interesting to perform application-specific analyses to
241 produce reduced statistics and graphs that are useful to resolve a
242 given problem. Trace viewers and analyzers are specialized tools
245 In the end, this is what LTTng is: a powerful, open source set of
246 tools to trace the Linux kernel and user applications at the same time.
247 LTTng is composed of several components actively maintained and
248 developed by its link:/community/#where[community].
251 [[lttng-alternatives]]
252 === Alternatives to noch:{LTTng}
254 Excluding proprietary solutions, a few competing software tracers
257 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
258 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
259 user scripts and is responsible for loading code into the
260 Linux kernel for further execution and collecting the outputted data.
261 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
262 subsystem in the Linux kernel in which a virtual machine can execute
263 programs passed from the user space to the kernel. You can attach
264 such programs to tracepoints and KProbes thanks to a system call, and
265 they can output data to the user space when executed thanks to
266 different mechanisms (pipe, VM register values, and eBPF maps, to name
268 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
269 is the de facto function tracer of the Linux kernel. Its user
270 interface is a set of special files in sysfs.
271 * https://perf.wiki.kernel.org/[perf] is
272 a performance analyzing tool for Linux which supports hardware
273 performance counters, tracepoints, as well as other counters and
274 types of probes. perf's controlling utility is the cmd:perf command
276 * http://linux.die.net/man/1/strace[strace]
277 is a command-line utility which records system calls made by a
278 user process, as well as signal deliveries and changes of process
279 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
280 to fulfill its function.
281 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
282 analyze Linux kernel events. You write scripts, or _chisels_ in
283 sysdig's jargon, in Lua and sysdig executes them while the system is
284 being traced or afterwards. sysdig's interface is the cmd:sysdig
285 command-line tool as well as the curses-based cmd:csysdig tool.
286 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
287 user space tracer which uses custom user scripts to produce plain text
288 traces. SystemTap converts the scripts to the C language, and then
289 compiles them as Linux kernel modules which are loaded to produce
290 trace data. SystemTap's primary user interface is the cmd:stap
293 The main distinctive features of LTTng is that it produces correlated
294 kernel and user space traces, as well as doing so with the lowest
295 overhead amongst other solutions. It produces trace files in the
296 http://diamon.org/ctf[CTF] format, a file format optimized
297 for the production and analyses of multi-gigabyte data.
299 LTTng is the result of more than 10 years of active open source
300 development by a community of passionate developers.
301 LTTng{nbsp}{revision} is currently available on major desktop and server
304 The main interface for tracing control is a single command-line tool
305 named cmd:lttng. The latter can create several tracing sessions, enable
306 and disable events on the fly, filter events efficiently with custom
307 user expressions, start and stop tracing, and much more. LTTng can
308 record the traces on the file system or send them over the network, and
309 keep them totally or partially. You can view the traces once tracing
310 becomes inactive or in real-time.
312 <<installing-lttng,Install LTTng now>> and
313 <<getting-started,start tracing>>!
319 **LTTng** is a set of software <<plumbing,components>> which interact to
320 <<instrumenting,instrument>> the Linux kernel and user applications, and
321 to <<controlling-tracing,control tracing>> (start and stop
322 tracing, enable and disable event rules, and the rest). Those
323 components are bundled into the following packages:
325 * **LTTng-tools**: Libraries and command-line interface to
327 * **LTTng-modules**: Linux kernel modules to instrument and
329 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
330 trace user applications.
332 Most distributions mark the LTTng-modules and LTTng-UST packages as
333 optional when installing LTTng-tools (which is always required). In the
334 following sections, we always provide the steps to install all three,
337 * You only need to install LTTng-modules if you intend to trace the
339 * You only need to install LTTng-UST if you intend to trace user
343 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 2 December 2016.
345 |Distribution |Available in releases |Alternatives
347 |https://www.ubuntu.com/[Ubuntu]
348 |Ubuntu{nbsp}14.04 _Trusty Tahr_ and Ubuntu{nbsp}16.04 _Xenial Xerus_:
349 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
350 |link:/docs/v2.8#doc-ubuntu[LTTng{nbsp}2.8 for Ubuntu{nbsp}16.10 _Yakkety Yak_].
352 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
353 other Ubuntu releases.
355 |https://getfedora.org/[Fedora]
357 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
360 |https://www.debian.org/[Debian]
362 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
365 |https://www.opensuse.org/[openSUSE]
367 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
370 |https://www.archlinux.org/[Arch Linux]
371 |<<arch-linux,Latest AUR packages>>.
374 |https://alpinelinux.org/[Alpine Linux]
376 |link:/docs/v2.8#doc-alpine-linux[LTTng{nbsp}2.8 for Alpine Linux{nbsp}"edge"].
378 LTTng{nbsp}2.8 for Alpine Linux{nbsp}3.5 (not released yet).
380 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
381 other Alpine Linux releases.
383 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
384 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
387 |https://buildroot.org/[Buildroot]
389 |link:/docs/v2.8#doc-buildroot[LTTng{nbsp}2.8 for Buildroot{nbsp}2016.11].
391 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
392 other Buildroot releases.
394 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
395 https://www.yoctoproject.org/[Yocto]
397 |link:/docs/v2.8#doc-oe-yocto[LTTng{nbsp}2.8 for Yocto Project{nbsp}2.2 _Morty_]
398 (`openembedded-core` layer).
400 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
401 other OpenEmbedded releases.
406 === [[ubuntu-official-repositories]]Ubuntu
409 ==== noch:{LTTng} Stable {revision} PPA
411 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
412 Stable{nbsp}{revision} PPA] offers the latest stable
413 LTTng{nbsp}{revision} packages for:
415 * Ubuntu{nbsp}14.04 _Trusty Tahr_
416 * Ubuntu{nbsp}16.04 _Xenial Xerus_
418 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
420 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
426 sudo apt-add-repository ppa:lttng/stable-2.9
431 . Install the main LTTng{nbsp}{revision} packages:
436 sudo apt-get install lttng-tools
437 sudo apt-get install lttng-modules-dkms
438 sudo apt-get install liblttng-ust-dev
442 . **If you need to instrument and trace
443 <<java-application,Java applications>>**, install the LTTng-UST
449 sudo apt-get install liblttng-ust-agent-java
453 . **If you need to instrument and trace
454 <<python-application,Python{nbsp}3 applications>>**, install the
455 LTTng-UST Python agent:
460 sudo apt-get install python3-lttngust
468 To install LTTng{nbsp}{revision} on Arch Linux using
469 https://archlinux.fr/yaourt-en[Yaourt]:
471 . Install the main LTTng{nbsp}{revision} packages:
476 yaourt -S lttng-tools
478 yaourt -S lttng-modules
482 . **If you need to instrument and trace <<python-application,Python
483 applications>>**, install the LTTng-UST Python agent:
488 yaourt -S python-lttngust
489 yaourt -S python2-lttngust
494 [[enterprise-distributions]]
495 === RHEL, SUSE, and other enterprise distributions
497 To install LTTng on enterprise Linux distributions, such as Red Hat
498 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
499 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
502 [[building-from-source]]
503 === Build from source
505 To build and install LTTng{nbsp}{revision} from source:
507 . Using your distribution's package manager, or from source, install
508 the following dependencies of LTTng-tools and LTTng-UST:
511 * https://sourceforge.net/projects/libuuid/[libuuid]
512 * http://directory.fsf.org/wiki/Popt[popt]
513 * http://liburcu.org/[Userspace RCU]
514 * http://www.xmlsoft.org/[libxml2]
517 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
523 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.9.tar.bz2 &&
524 tar -xf lttng-modules-latest-2.9.tar.bz2 &&
525 cd lttng-modules-2.9.* &&
527 sudo make modules_install &&
532 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
538 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.9.tar.bz2 &&
539 tar -xf lttng-ust-latest-2.9.tar.bz2 &&
540 cd lttng-ust-2.9.* &&
550 .Java and Python application tracing
552 If you need to instrument and trace <<java-application,Java
553 applications>>, pass the `--enable-java-agent-jul`,
554 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
555 `configure` script, depending on which Java logging framework you use.
557 If you need to instrument and trace <<python-application,Python
558 applications>>, pass the `--enable-python-agent` option to the
559 `configure` script. You can set the `PYTHON` environment variable to the
560 path to the Python interpreter for which to install the LTTng-UST Python
568 By default, LTTng-UST libraries are installed to
569 dir:{/usr/local/lib}, which is the de facto directory in which to
570 keep self-compiled and third-party libraries.
572 When <<building-tracepoint-providers-and-user-application,linking an
573 instrumented user application with `liblttng-ust`>>:
575 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
577 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
578 man:gcc(1), man:g++(1), or man:clang(1).
582 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
588 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
589 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
590 cd lttng-tools-2.9.* &&
598 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
599 previous steps automatically for a given version of LTTng and confine
600 the installed files in a specific directory. This can be useful to test
601 LTTng without installing it on your system.
607 This is a short guide to get started quickly with LTTng kernel and user
610 Before you follow this guide, make sure to <<installing-lttng,install>>
613 This tutorial walks you through the steps to:
615 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
616 . <<tracing-your-own-user-application,Trace a user application>> written
618 . <<viewing-and-analyzing-your-traces,View and analyze the
622 [[tracing-the-linux-kernel]]
623 === Trace the Linux kernel
625 The following command lines start with cmd:sudo because you need root
626 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
627 your Unix user is a member of the <<tracing-group,tracing group>>.
629 . Create a <<tracing-session,tracing session>> which writes its traces
630 to dir:{/tmp/my-kernel-trace}:
635 sudo lttng create my-kernel-session --output=/tmp/my-kernel-trace
639 . List the available kernel tracepoints and system calls:
645 lttng list --kernel --syscall
649 . Create <<event,event rules>> which match the desired instrumentation
650 point names, for example the `sched_switch` and `sched_process_fork`
651 tracepoints, and the man:open(2) and man:close(2) system calls:
656 sudo lttng enable-event --kernel sched_switch,sched_process_fork
657 sudo lttng enable-event --kernel --syscall open,close
661 You can also create an event rule which matches _all_ the Linux kernel
662 tracepoints (this will generate a lot of data when tracing):
667 sudo lttng enable-event --kernel --all
671 . <<basic-tracing-session-control,Start tracing>>:
680 . Do some operation on your system for a few seconds. For example,
681 load a website, or list the files of a directory.
682 . <<basic-tracing-session-control,Stop tracing>> and destroy the
693 The man:lttng-destroy(1) command does not destroy the trace data; it
694 only destroys the state of the tracing session.
696 . For the sake of this example, make the recorded trace accessible to
702 sudo chown -R $(whoami) /tmp/my-kernel-trace
706 See <<viewing-and-analyzing-your-traces,View and analyze the
707 recorded events>> to view the recorded events.
710 [[tracing-your-own-user-application]]
711 === Trace a user application
713 This section steps you through a simple example to trace a
714 _Hello world_ program written in C.
716 To create the traceable user application:
718 . Create the tracepoint provider header file, which defines the
719 tracepoints and the events they can generate:
725 #undef TRACEPOINT_PROVIDER
726 #define TRACEPOINT_PROVIDER hello_world
728 #undef TRACEPOINT_INCLUDE
729 #define TRACEPOINT_INCLUDE "./hello-tp.h"
731 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
734 #include <lttng/tracepoint.h>
744 ctf_string(my_string_field, my_string_arg)
745 ctf_integer(int, my_integer_field, my_integer_arg)
749 #endif /* _HELLO_TP_H */
751 #include <lttng/tracepoint-event.h>
755 . Create the tracepoint provider package source file:
761 #define TRACEPOINT_CREATE_PROBES
762 #define TRACEPOINT_DEFINE
764 #include "hello-tp.h"
768 . Build the tracepoint provider package:
773 gcc -c -I. hello-tp.c
777 . Create the _Hello World_ application source file:
784 #include "hello-tp.h"
786 int main(int argc, char *argv[])
790 puts("Hello, World!\nPress Enter to continue...");
793 * The following getchar() call is only placed here for the purpose
794 * of this demonstration, to pause the application in order for
795 * you to have time to list its tracepoints. It is not
801 * A tracepoint() call.
803 * Arguments, as defined in hello-tp.h:
805 * 1. Tracepoint provider name (required)
806 * 2. Tracepoint name (required)
807 * 3. my_integer_arg (first user-defined argument)
808 * 4. my_string_arg (second user-defined argument)
810 * Notice the tracepoint provider and tracepoint names are
811 * NOT strings: they are in fact parts of variables that the
812 * macros in hello-tp.h create.
814 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
816 for (x = 0; x < argc; ++x) {
817 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
820 puts("Quitting now!");
821 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
828 . Build the application:
837 . Link the application with the tracepoint provider package,
838 `liblttng-ust`, and `libdl`:
843 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
847 Here's the whole build process:
850 .User space tracing tutorial's build steps.
851 image::ust-flow.png[]
853 To trace the user application:
855 . Run the application with a few arguments:
860 ./hello world and beyond
869 Press Enter to continue...
873 . Start an LTTng <<lttng-sessiond,session daemon>>:
878 lttng-sessiond --daemonize
882 Note that a session daemon might already be running, for example as
883 a service that the distribution's service manager started.
885 . List the available user space tracepoints:
890 lttng list --userspace
894 You see the `hello_world:my_first_tracepoint` tracepoint listed
895 under the `./hello` process.
897 . Create a <<tracing-session,tracing session>>:
902 lttng create my-user-space-session
906 . Create an <<event,event rule>> which matches the
907 `hello_world:my_first_tracepoint` event name:
912 lttng enable-event --userspace hello_world:my_first_tracepoint
916 . <<basic-tracing-session-control,Start tracing>>:
925 . Go back to the running `hello` application and press Enter. The
926 program executes all `tracepoint()` instrumentation points and exits.
927 . <<basic-tracing-session-control,Stop tracing>> and destroy the
938 The man:lttng-destroy(1) command does not destroy the trace data; it
939 only destroys the state of the tracing session.
941 By default, LTTng saves the traces in
942 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
943 where +__name__+ is the tracing session name. The
944 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
946 See <<viewing-and-analyzing-your-traces,View and analyze the
947 recorded events>> to view the recorded events.
950 [[viewing-and-analyzing-your-traces]]
951 === View and analyze the recorded events
953 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
954 kernel>> and <<tracing-your-own-user-application,Trace a user
955 application>> tutorials, you can inspect the recorded events.
957 Many tools are available to read LTTng traces:
959 * **cmd:babeltrace** is a command-line utility which converts trace
960 formats; it supports the format that LTTng produces, CTF, as well as a
961 basic text output which can be ++grep++ed. The cmd:babeltrace command
962 is part of the http://diamon.org/babeltrace[Babeltrace] project.
963 * Babeltrace also includes
964 **https://www.python.org/[Python] bindings** so
965 that you can easily open and read an LTTng trace with your own script,
966 benefiting from the power of Python.
967 * http://tracecompass.org/[**Trace Compass**]
968 is a graphical user interface for viewing and analyzing any type of
969 logs or traces, including LTTng's.
970 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
971 project which includes many high-level analyses of LTTng kernel
972 traces, like scheduling statistics, interrupt frequency distribution,
973 top CPU usage, and more.
975 NOTE: This section assumes that the traces recorded during the previous
976 tutorials were saved to their default location, in the
977 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
978 environment variable defaults to `$HOME` if not set.
981 [[viewing-and-analyzing-your-traces-bt]]
982 ==== Use the cmd:babeltrace command-line tool
984 The simplest way to list all the recorded events of a trace is to pass
985 its path to cmd:babeltrace with no options:
989 babeltrace ~/lttng-traces/my-user-space-session*
992 cmd:babeltrace finds all traces recursively within the given path and
993 prints all their events, merging them in chronological order.
995 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1000 babeltrace /tmp/my-kernel-trace | grep _switch
1003 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1004 count the recorded events:
1008 babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1012 [[viewing-and-analyzing-your-traces-bt-python]]
1013 ==== Use the Babeltrace Python bindings
1015 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1016 is useful to isolate events by simple matching using man:grep(1) and
1017 similar utilities. However, more elaborate filters, such as keeping only
1018 event records with a field value falling within a specific range, are
1019 not trivial to write using a shell. Moreover, reductions and even the
1020 most basic computations involving multiple event records are virtually
1021 impossible to implement.
1023 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1024 to read the event records of an LTTng trace sequentially and compute the
1025 desired information.
1027 The following script accepts an LTTng Linux kernel trace path as its
1028 first argument and prints the short names of the top 5 running processes
1029 on CPU 0 during the whole trace:
1034 from collections import Counter
1040 if len(sys.argv) != 2:
1041 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1042 print(msg, file=sys.stderr)
1045 # A trace collection contains one or more traces
1046 col = babeltrace.TraceCollection()
1048 # Add the trace provided by the user (LTTng traces always have
1050 if col.add_trace(sys.argv[1], 'ctf') is None:
1051 raise RuntimeError('Cannot add trace')
1053 # This counter dict contains execution times:
1055 # task command name -> total execution time (ns)
1056 exec_times = Counter()
1058 # This contains the last `sched_switch` timestamp
1062 for event in col.events:
1063 # Keep only `sched_switch` events
1064 if event.name != 'sched_switch':
1067 # Keep only events which happened on CPU 0
1068 if event['cpu_id'] != 0:
1072 cur_ts = event.timestamp
1078 # Previous task command (short) name
1079 prev_comm = event['prev_comm']
1081 # Initialize entry in our dict if not yet done
1082 if prev_comm not in exec_times:
1083 exec_times[prev_comm] = 0
1085 # Compute previous command execution time
1086 diff = cur_ts - last_ts
1088 # Update execution time of this command
1089 exec_times[prev_comm] += diff
1091 # Update last timestamp
1095 for name, ns in exec_times.most_common(5):
1097 print('{:20}{} s'.format(name, s))
1102 if __name__ == '__main__':
1103 sys.exit(0 if top5proc() else 1)
1110 python3 top5proc.py /tmp/my-kernel-trace/kernel
1116 swapper/0 48.607245889 s
1117 chromium 7.192738188 s
1118 pavucontrol 0.709894415 s
1119 Compositor 0.660867933 s
1120 Xorg.bin 0.616753786 s
1123 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1124 weren't using the CPU that much when tracing, its first position in the
1129 == [[understanding-lttng]]Core concepts
1131 From a user's perspective, the LTTng system is built on a few concepts,
1132 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1133 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1134 Understanding how those objects relate to eachother is key in mastering
1137 The core concepts are:
1139 * <<tracing-session,Tracing session>>
1140 * <<domain,Tracing domain>>
1141 * <<channel,Channel and ring buffer>>
1142 * <<"event","Instrumentation point, event rule, event, and event record">>
1148 A _tracing session_ is a stateful dialogue between you and
1149 a <<lttng-sessiond,session daemon>>. You can
1150 <<creating-destroying-tracing-sessions,create a new tracing
1151 session>> with the `lttng create` command.
1153 Anything that you do when you control LTTng tracers happens within a
1154 tracing session. In particular, a tracing session:
1157 * Has its own set of trace files.
1158 * Has its own state of activity (started or stopped).
1159 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1161 * Has its own <<channel,channels>> which have their own
1162 <<event,event rules>>.
1165 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1166 image::concepts.png[]
1168 Those attributes and objects are completely isolated between different
1171 A tracing session is analogous to a cash machine session:
1172 the operations you do on the banking system through the cash machine do
1173 not alter the data of other users of the same system. In the case of
1174 the cash machine, a session lasts as long as your bank card is inside.
1175 In the case of LTTng, a tracing session lasts from the `lttng create`
1176 command to the `lttng destroy` command.
1179 .Each Unix user has its own set of tracing sessions.
1180 image::many-sessions.png[]
1183 [[tracing-session-mode]]
1184 ==== Tracing session mode
1186 LTTng can send the generated trace data to different locations. The
1187 _tracing session mode_ dictates where to send it. The following modes
1188 are available in LTTng{nbsp}{revision}:
1191 LTTng writes the traces to the file system of the machine being traced
1194 Network streaming mode::
1195 LTTng sends the traces over the network to a
1196 <<lttng-relayd,relay daemon>> running on a remote system.
1199 LTTng does not write the traces by default. Instead, you can request
1200 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1201 current tracing buffers, and to write it to the target's file system
1202 or to send it over the network to a <<lttng-relayd,relay daemon>>
1203 running on a remote system.
1206 This mode is similar to the network streaming mode, but a live
1207 trace viewer can connect to the distant relay daemon to
1208 <<lttng-live,view event records as LTTng generates them>> by
1215 A _tracing domain_ is a namespace for event sources. A tracing domain
1216 has its own properties and features.
1218 There are currently five available tracing domains:
1222 * `java.util.logging` (JUL)
1226 You must specify a tracing domain when using some commands to avoid
1227 ambiguity. For example, since all the domains support named tracepoints
1228 as event sources (instrumentation points that you manually insert in the
1229 source code), you need to specify a tracing domain when
1230 <<enabling-disabling-events,creating an event rule>> because all the
1231 tracing domains could have tracepoints with the same names.
1233 Some features are reserved to specific tracing domains. Dynamic function
1234 entry and return instrumentation points, for example, are currently only
1235 supported in the Linux kernel tracing domain, but support for other
1236 tracing domains could be added in the future.
1238 You can create <<channel,channels>> in the Linux kernel and user space
1239 tracing domains. The other tracing domains have a single default
1244 === Channel and ring buffer
1246 A _channel_ is an object which is responsible for a set of ring buffers.
1247 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1248 tracer emits an event, it can record it to one or more
1249 sub-buffers. The attributes of a channel determine what to do when
1250 there's no space left for a new event record because all sub-buffers
1251 are full, where to send a full sub-buffer, and other behaviours.
1253 A channel is always associated to a <<domain,tracing domain>>. The
1254 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1255 a default channel which you cannot configure.
1257 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1258 an event, it records it to the sub-buffers of all
1259 the enabled channels with a satisfied event rule, as long as those
1260 channels are part of active <<tracing-session,tracing sessions>>.
1263 [[channel-buffering-schemes]]
1264 ==== Per-user vs. per-process buffering schemes
1266 A channel has at least one ring buffer _per CPU_. LTTng always
1267 records an event to the ring buffer associated to the CPU on which it
1270 Two _buffering schemes_ are available when you
1271 <<enabling-disabling-channels,create a channel>> in the
1272 user space <<domain,tracing domain>>:
1274 Per-user buffering::
1275 Allocate one set of ring buffers--one per CPU--shared by all the
1276 instrumented processes of each Unix user.
1280 .Per-user buffering scheme.
1281 image::per-user-buffering.png[]
1284 Per-process buffering::
1285 Allocate one set of ring buffers--one per CPU--for each
1286 instrumented process.
1290 .Per-process buffering scheme.
1291 image::per-process-buffering.png[]
1294 The per-process buffering scheme tends to consume more memory than the
1295 per-user option because systems generally have more instrumented
1296 processes than Unix users running instrumented processes. However, the
1297 per-process buffering scheme ensures that one process having a high
1298 event throughput won't fill all the shared sub-buffers of the same
1301 The Linux kernel tracing domain has only one available buffering scheme
1302 which is to allocate a single set of ring buffers for the whole system.
1303 This scheme is similar to the per-user option, but with a single, global
1304 user "running" the kernel.
1307 [[channel-overwrite-mode-vs-discard-mode]]
1308 ==== Overwrite vs. discard event loss modes
1310 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1311 arc in the following animation) of a specific channel's ring buffer.
1312 When there's no space left in a sub-buffer, the tracer marks it as
1313 consumable (red) and another, empty sub-buffer starts receiving the
1314 following event records. A <<lttng-consumerd,consumer daemon>>
1315 eventually consumes the marked sub-buffer (returns to white).
1318 [role="docsvg-channel-subbuf-anim"]
1323 In an ideal world, sub-buffers are consumed faster than they are filled,
1324 as is the case in the previous animation. In the real world,
1325 however, all sub-buffers can be full at some point, leaving no space to
1326 record the following events.
1328 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1329 available, it is acceptable to lose event records when the alternative
1330 would be to cause substantial delays in the instrumented application's
1331 execution. LTTng privileges performance over integrity; it aims at
1332 perturbing the traced system as little as possible in order to make
1333 tracing of subtle race conditions and rare interrupt cascades possible.
1335 When it comes to losing event records because no empty sub-buffer is
1336 available, the channel's _event loss mode_ determines what to do. The
1337 available event loss modes are:
1340 Drop the newest event records until a the tracer
1341 releases a sub-buffer.
1344 Clear the sub-buffer containing the oldest event records and start
1345 writing the newest event records there.
1347 This mode is sometimes called _flight recorder mode_ because it's
1349 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1350 always keep a fixed amount of the latest data.
1352 Which mechanism you should choose depends on your context: prioritize
1353 the newest or the oldest event records in the ring buffer?
1355 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1356 as soon as a there's no space left for a new event record, whereas in
1357 discard mode, the tracer only discards the event record that doesn't
1360 In discard mode, LTTng increments a count of lost event records when
1361 an event record is lost and saves this count to the trace. In
1362 overwrite mode, LTTng keeps no information when it overwrites a
1363 sub-buffer before consuming it.
1365 There are a few ways to decrease your probability of losing event
1367 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1368 how you can fine-une the sub-buffer count and size of a channel to
1369 virtually stop losing event records, though at the cost of greater
1373 [[channel-subbuf-size-vs-subbuf-count]]
1374 ==== Sub-buffer count and size
1376 When you <<enabling-disabling-channels,create a channel>>, you can
1377 set its number of sub-buffers and their size.
1379 Note that there is noticeable CPU overhead introduced when
1380 switching sub-buffers (marking a full one as consumable and switching
1381 to an empty one for the following events to be recorded). Knowing this,
1382 the following list presents a few practical situations along with how
1383 to configure the sub-buffer count and size for them:
1385 * **High event throughput**: In general, prefer bigger sub-buffers to
1386 lower the risk of losing event records.
1388 Having bigger sub-buffers also ensures a lower sub-buffer switching
1391 The number of sub-buffers is only meaningful if you create the channel
1392 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1393 other sub-buffers are left unaltered.
1395 * **Low event throughput**: In general, prefer smaller sub-buffers
1396 since the risk of losing event records is low.
1398 Because events occur less frequently, the sub-buffer switching frequency
1399 should remain low and thus the tracer's overhead should not be a
1402 * **Low memory system**: If your target system has a low memory
1403 limit, prefer fewer first, then smaller sub-buffers.
1405 Even if the system is limited in memory, you want to keep the
1406 sub-buffers as big as possible to avoid a high sub-buffer switching
1409 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1410 which means event data is very compact. For example, the average
1411 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1412 sub-buffer size of 1{nbsp}MiB is considered big.
1414 The previous situations highlight the major trade-off between a few big
1415 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1416 frequency vs. how much data is lost in overwrite mode. Assuming a
1417 constant event throughput and using the overwrite mode, the two
1418 following configurations have the same ring buffer total size:
1421 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1426 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1427 switching frequency, but if a sub-buffer overwrite happens, half of
1428 the event records so far (4{nbsp}MiB) are definitely lost.
1429 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1430 overhead as the previous configuration, but if a sub-buffer
1431 overwrite happens, only the eighth of event records so far are
1434 In discard mode, the sub-buffers count parameter is pointless: use two
1435 sub-buffers and set their size according to the requirements of your
1439 [[channel-switch-timer]]
1440 ==== Switch timer period
1442 The _switch timer period_ is an important configurable attribute of
1443 a channel to ensure periodic sub-buffer flushing.
1445 When the _switch timer_ expires, a sub-buffer switch happens. You can
1446 set the switch timer period attribute when you
1447 <<enabling-disabling-channels,create a channel>> to ensure that event
1448 data is consumed and committed to trace files or to a distant relay
1449 daemon periodically in case of a low event throughput.
1452 [role="docsvg-channel-switch-timer"]
1457 This attribute is also convenient when you use big sub-buffers to cope
1458 with a sporadic high event throughput, even if the throughput is
1462 [[channel-read-timer]]
1463 ==== Read timer period
1465 By default, the LTTng tracers use a notification mechanism to signal a
1466 full sub-buffer so that a consumer daemon can consume it. When such
1467 notifications must be avoided, for example in real-time applications,
1468 you can use the channel's _read timer_ instead. When the read timer
1469 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1470 consumable sub-buffers.
1473 [[tracefile-rotation]]
1474 ==== Trace file count and size
1476 By default, trace files can grow as large as needed. You can set the
1477 maximum size of each trace file that a channel writes when you
1478 <<enabling-disabling-channels,create a channel>>. When the size of
1479 a trace file reaches the channel's fixed maximum size, LTTng creates
1480 another file to contain the next event records. LTTng appends a file
1481 count to each trace file name in this case.
1483 If you set the trace file size attribute when you create a channel, the
1484 maximum number of trace files that LTTng creates is _unlimited_ by
1485 default. To limit them, you can also set a maximum number of trace
1486 files. When the number of trace files reaches the channel's fixed
1487 maximum count, the oldest trace file is overwritten. This mechanism is
1488 called _trace file rotation_.
1492 === Instrumentation point, event rule, event, and event record
1494 An _event rule_ is a set of conditions which must be **all** satisfied
1495 for LTTng to record an occuring event.
1497 You set the conditions when you <<enabling-disabling-events,create
1500 You always attach an event rule to <<channel,channel>> when you create
1503 When an event passes the conditions of an event rule, LTTng records it
1504 in one of the attached channel's sub-buffers.
1506 The available conditions, as of LTTng{nbsp}{revision}, are:
1508 * The event rule _is enabled_.
1509 * The instrumentation point's type _is{nbsp}T_.
1510 * The instrumentation point's name (sometimes called _event name_)
1511 _matches{nbsp}N_, but _is not{nbsp}E_.
1512 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1513 _is exactly{nbsp}L_.
1514 * The fields of the event's payload _satisfy_ a filter
1515 expression{nbsp}__F__.
1517 As you can see, all the conditions but the dynamic filter are related to
1518 the event rule's status or to the instrumentation point, not to the
1519 occurring events. This is why, without a filter, checking if an event
1520 passes an event rule is not a dynamic task: when you create or modify an
1521 event rule, all the tracers of its tracing domain enable or disable the
1522 instrumentation points themselves once. This is possible because the
1523 attributes of an instrumentation point (type, name, and log level) are
1524 defined statically. In other words, without a dynamic filter, the tracer
1525 _does not evaluate_ the arguments of an instrumentation point unless it
1526 matches an enabled event rule.
1528 Note that, for LTTng to record an event, the <<channel,channel>> to
1529 which a matching event rule is attached must also be enabled, and the
1530 tracing session owning this channel must be active.
1533 .Logical path from an instrumentation point to an event record.
1534 image::event-rule.png[]
1536 .Event, event record, or event rule?
1538 With so many similar terms, it's easy to get confused.
1540 An **event** is the consequence of the execution of an _instrumentation
1541 point_, like a tracepoint that you manually place in some source code,
1542 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1543 time. Different actions can be taken upon the occurance of an event,
1544 like record the event's payload to a buffer.
1546 An **event record** is the representation of an event in a sub-buffer. A
1547 tracer is responsible for capturing the payload of an event, current
1548 context variables, the event's ID, and the event's timestamp. LTTng
1549 can append this sub-buffer to a trace file.
1551 An **event rule** is a set of conditions which must all be satisfied for
1552 LTTng to record an occuring event. Events still occur without
1553 satisfying event rules, but LTTng does not record them.
1558 == Components of noch:{LTTng}
1560 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1561 to call LTTng a simple _tool_ since it is composed of multiple
1562 interacting components. This section describes those components,
1563 explains their respective roles, and shows how they connect together to
1564 form the LTTng ecosystem.
1566 The following diagram shows how the most important components of LTTng
1567 interact with user applications, the Linux kernel, and you:
1570 .Control and trace data paths between LTTng components.
1571 image::plumbing.png[]
1573 The LTTng project incorporates:
1575 * **LTTng-tools**: Libraries and command-line interface to
1576 control tracing sessions.
1577 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1578 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1579 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1580 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1581 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1582 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1584 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1585 headers to instrument and trace any native user application.
1586 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1587 *** `liblttng-ust-libc-wrapper`
1588 *** `liblttng-ust-pthread-wrapper`
1589 *** `liblttng-ust-cyg-profile`
1590 *** `liblttng-ust-cyg-profile-fast`
1591 *** `liblttng-ust-dl`
1592 ** User space tracepoint provider source files generator command-line
1593 tool (man:lttng-gen-tp(1)).
1594 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1595 Java applications using `java.util.logging` or
1596 Apache log4j 1.2 logging.
1597 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1598 Python applications using the standard `logging` package.
1599 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1601 ** LTTng kernel tracer module.
1602 ** Tracing ring buffer kernel modules.
1603 ** Probe kernel modules.
1604 ** LTTng logger kernel module.
1608 === Tracing control command-line interface
1611 .The tracing control command-line interface.
1612 image::plumbing-lttng-cli.png[]
1614 The _man:lttng(1) command-line tool_ is the standard user interface to
1615 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1616 is part of LTTng-tools.
1618 The cmd:lttng tool is linked with
1619 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1620 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1622 The cmd:lttng tool has a Git-like interface:
1626 lttng <general options> <command> <command options>
1629 The <<controlling-tracing,Tracing control>> section explores the
1630 available features of LTTng using the cmd:lttng tool.
1633 [[liblttng-ctl-lttng]]
1634 === Tracing control library
1637 .The tracing control library.
1638 image::plumbing-liblttng-ctl.png[]
1640 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1641 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1642 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1644 The <<lttng-cli,cmd:lttng command-line tool>>
1645 is linked with `liblttng-ctl`.
1647 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1652 #include <lttng/lttng.h>
1655 Some objects are referenced by name (C string), such as tracing
1656 sessions, but most of them require to create a handle first using
1657 `lttng_create_handle()`.
1659 The best available developer documentation for `liblttng-ctl` is, as of
1660 LTTng{nbsp}{revision}, its installed header files. Every function and
1661 structure is thoroughly documented.
1665 === User space tracing library
1668 .The user space tracing library.
1669 image::plumbing-liblttng-ust.png[]
1671 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1672 is the LTTng user space tracer. It receives commands from a
1673 <<lttng-sessiond,session daemon>>, for example to
1674 enable and disable specific instrumentation points, and writes event
1675 records to ring buffers shared with a
1676 <<lttng-consumerd,consumer daemon>>.
1677 `liblttng-ust` is part of LTTng-UST.
1679 Public C header files are installed beside `liblttng-ust` to
1680 instrument any <<c-application,C or $$C++$$ application>>.
1682 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1683 packages, use their own library providing tracepoints which is
1684 linked with `liblttng-ust`.
1686 An application or library does not have to initialize `liblttng-ust`
1687 manually: its constructor does the necessary tasks to properly register
1688 to a session daemon. The initialization phase also enables the
1689 instrumentation points matching the <<event,event rules>> that you
1693 [[lttng-ust-agents]]
1694 === User space tracing agents
1697 .The user space tracing agents.
1698 image::plumbing-lttng-ust-agents.png[]
1700 The _LTTng-UST Java and Python agents_ are regular Java and Python
1701 packages which add LTTng tracing capabilities to the
1702 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1704 In the case of Java, the
1705 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1706 core logging facilities] and
1707 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1708 Note that Apache Log4{nbsp}2 is not supported.
1710 In the case of Python, the standard
1711 https://docs.python.org/3/library/logging.html[`logging`] package
1712 is supported. Both Python 2 and Python 3 modules can import the
1713 LTTng-UST Python agent package.
1715 The applications using the LTTng-UST agents are in the
1716 `java.util.logging` (JUL),
1717 log4j, and Python <<domain,tracing domains>>.
1719 Both agents use the same mechanism to trace the log statements. When an
1720 agent is initialized, it creates a log handler that attaches to the root
1721 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1722 When the application executes a log statement, it is passed to the
1723 agent's log handler by the root logger. The agent's log handler calls a
1724 native function in a tracepoint provider package shared library linked
1725 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1726 other fields, like its logger name and its log level. This native
1727 function contains a user space instrumentation point, hence tracing the
1730 The log level condition of an
1731 <<event,event rule>> is considered when tracing
1732 a Java or a Python application, and it's compatible with the standard
1733 JUL, log4j, and Python log levels.
1737 === LTTng kernel modules
1740 .The LTTng kernel modules.
1741 image::plumbing-lttng-modules.png[]
1743 The _LTTng kernel modules_ are a set of Linux kernel modules
1744 which implement the kernel tracer of the LTTng project. The LTTng
1745 kernel modules are part of LTTng-modules.
1747 The LTTng kernel modules include:
1749 * A set of _probe_ modules.
1751 Each module attaches to a specific subsystem
1752 of the Linux kernel using its tracepoint instrument points. There are
1753 also modules to attach to the entry and return points of the Linux
1754 system call functions.
1756 * _Ring buffer_ modules.
1758 A ring buffer implementation is provided as kernel modules. The LTTng
1759 kernel tracer writes to the ring buffer; a
1760 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1762 * The _LTTng kernel tracer_ module.
1763 * The _LTTng logger_ module.
1765 The LTTng logger module implements the special path:{/proc/lttng-logger}
1766 file so that any executable can generate LTTng events by opening and
1767 writing to this file.
1769 See <<proc-lttng-logger-abi,LTTng logger>>.
1771 Generally, you do not have to load the LTTng kernel modules manually
1772 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1773 daemon>> loads the necessary modules when starting. If you have extra
1774 probe modules, you can specify to load them to the session daemon on
1777 The LTTng kernel modules are installed in
1778 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1779 the kernel release (see `uname --kernel-release`).
1786 .The session daemon.
1787 image::plumbing-sessiond.png[]
1789 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1790 managing tracing sessions and for controlling the various components of
1791 LTTng. The session daemon is part of LTTng-tools.
1793 The session daemon sends control requests to and receives control
1796 * The <<lttng-ust,user space tracing library>>.
1798 Any instance of the user space tracing library first registers to
1799 a session daemon. Then, the session daemon can send requests to
1800 this instance, such as:
1803 ** Get the list of tracepoints.
1804 ** Share an <<event,event rule>> so that the user space tracing library
1805 can enable or disable tracepoints. Amongst the possible conditions
1806 of an event rule is a filter expression which `liblttng-ust` evalutes
1807 when an event occurs.
1808 ** Share <<channel,channel>> attributes and ring buffer locations.
1811 The session daemon and the user space tracing library use a Unix
1812 domain socket for their communication.
1814 * The <<lttng-ust-agents,user space tracing agents>>.
1816 Any instance of a user space tracing agent first registers to
1817 a session daemon. Then, the session daemon can send requests to
1818 this instance, such as:
1821 ** Get the list of loggers.
1822 ** Enable or disable a specific logger.
1825 The session daemon and the user space tracing agent use a TCP connection
1826 for their communication.
1828 * The <<lttng-modules,LTTng kernel tracer>>.
1829 * The <<lttng-consumerd,consumer daemon>>.
1831 The session daemon sends requests to the consumer daemon to instruct
1832 it where to send the trace data streams, amongst other information.
1834 * The <<lttng-relayd,relay daemon>>.
1836 The session daemon receives commands from the
1837 <<liblttng-ctl-lttng,tracing control library>>.
1839 The root session daemon loads the appropriate
1840 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1841 a <<lttng-consumerd,consumer daemon>> as soon as you create
1842 an <<event,event rule>>.
1844 The session daemon does not send and receive trace data: this is the
1845 role of the <<lttng-consumerd,consumer daemon>> and
1846 <<lttng-relayd,relay daemon>>. It does, however, generate the
1847 http://diamon.org/ctf/[CTF] metadata stream.
1849 Each Unix user can have its own session daemon instance. The
1850 tracing sessions managed by different session daemons are completely
1853 The root user's session daemon is the only one which is
1854 allowed to control the LTTng kernel tracer, and its spawned consumer
1855 daemon is the only one which is allowed to consume trace data from the
1856 LTTng kernel tracer. Note, however, that any Unix user which is a member
1857 of the <<tracing-group,tracing group>> is allowed
1858 to create <<channel,channels>> in the
1859 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1862 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1863 session daemon when using its `create` command if none is currently
1864 running. You can also start the session daemon manually.
1871 .The consumer daemon.
1872 image::plumbing-consumerd.png[]
1874 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
1875 ring buffers with user applications or with the LTTng kernel modules to
1876 collect trace data and send it to some location (on disk or to a
1877 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1878 is part of LTTng-tools.
1880 You do not start a consumer daemon manually: a consumer daemon is always
1881 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1882 <<event,event rule>>, that is, before you start tracing. When you kill
1883 its owner session daemon, the consumer daemon also exits because it is
1884 the session daemon's child process. Command-line options of
1885 man:lttng-sessiond(8) target the consumer daemon process.
1887 There are up to two running consumer daemons per Unix user, whereas only
1888 one session daemon can run per user. This is because each process can be
1889 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1890 and 64-bit processes, it is more efficient to have separate
1891 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1892 exception: it can have up to _three_ running consumer daemons: 32-bit
1893 and 64-bit instances for its user applications, and one more
1894 reserved for collecting kernel trace data.
1902 image::plumbing-relayd.png[]
1904 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1905 between remote session and consumer daemons, local trace files, and a
1906 remote live trace viewer. The relay daemon is part of LTTng-tools.
1908 The main purpose of the relay daemon is to implement a receiver of
1909 <<sending-trace-data-over-the-network,trace data over the network>>.
1910 This is useful when the target system does not have much file system
1911 space to record trace files locally.
1913 The relay daemon is also a server to which a
1914 <<lttng-live,live trace viewer>> can
1915 connect. The live trace viewer sends requests to the relay daemon to
1916 receive trace data as the target system emits events. The
1917 communication protocol is named _LTTng live_; it is used over TCP
1920 Note that you can start the relay daemon on the target system directly.
1921 This is the setup of choice when the use case is to view events as
1922 the target system emits them without the need of a remote system.
1926 == [[using-lttng]]Instrumentation
1928 There are many examples of tracing and monitoring in our everyday life:
1930 * You have access to real-time and historical weather reports and
1931 forecasts thanks to weather stations installed around the country.
1932 * You know your heart is safe thanks to an electrocardiogram.
1933 * You make sure not to drive your car too fast and to have enough fuel
1934 to reach your destination thanks to gauges visible on your dashboard.
1936 All the previous examples have something in common: they rely on
1937 **instruments**. Without the electrodes attached to the surface of your
1938 body's skin, cardiac monitoring is futile.
1940 LTTng, as a tracer, is no different from those real life examples. If
1941 you're about to trace a software system or, in other words, record its
1942 history of execution, you better have **instrumentation points** in the
1943 subject you're tracing, that is, the actual software.
1945 Various ways were developed to instrument a piece of software for LTTng
1946 tracing. The most straightforward one is to manually place
1947 instrumentation points, called _tracepoints_, in the software's source
1948 code. It is also possible to add instrumentation points dynamically in
1949 the Linux kernel <<domain,tracing domain>>.
1951 If you're only interested in tracing the Linux kernel, your
1952 instrumentation needs are probably already covered by LTTng's built-in
1953 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
1954 user application which is already instrumented for LTTng tracing.
1955 In such cases, you can skip this whole section and read the topics of
1956 the <<controlling-tracing,Tracing control>> section.
1958 Many methods are available to instrument a piece of software for LTTng
1961 * <<c-application,User space instrumentation for C and $$C++$$
1963 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
1964 * <<java-application,User space Java agent>>.
1965 * <<python-application,User space Python agent>>.
1966 * <<proc-lttng-logger-abi,LTTng logger>>.
1967 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
1971 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
1973 The procedure to instrument a C or $$C++$$ user application with
1974 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
1976 . <<tracepoint-provider,Create the source files of a tracepoint provider
1978 . <<probing-the-application-source-code,Add tracepoints to
1979 the application's source code>>.
1980 . <<building-tracepoint-providers-and-user-application,Build and link
1981 a tracepoint provider package and the user application>>.
1983 If you need quick, man:printf(3)-like instrumentation, you can skip
1984 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
1987 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
1988 instrument a user application with `liblttng-ust`.
1991 [[tracepoint-provider]]
1992 ==== Create the source files of a tracepoint provider package
1994 A _tracepoint provider_ is a set of compiled functions which provide
1995 **tracepoints** to an application, the type of instrumentation point
1996 supported by LTTng-UST. Those functions can emit events with
1997 user-defined fields and serialize those events as event records to one
1998 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
1999 macro, which you <<probing-the-application-source-code,insert in a user
2000 application's source code>>, calls those functions.
2002 A _tracepoint provider package_ is an object file (`.o`) or a shared
2003 library (`.so`) which contains one or more tracepoint providers.
2004 Its source files are:
2006 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2007 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2009 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2010 the LTTng user space tracer, at run time.
2013 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2014 image::ust-app.png[]
2016 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2017 skip creating and using a tracepoint provider and use
2018 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2022 ===== Create a tracepoint provider header file template
2024 A _tracepoint provider header file_ contains the tracepoint
2025 definitions of a tracepoint provider.
2027 To create a tracepoint provider header file:
2029 . Start from this template:
2033 .Tracepoint provider header file template (`.h` file extension).
2035 #undef TRACEPOINT_PROVIDER
2036 #define TRACEPOINT_PROVIDER provider_name
2038 #undef TRACEPOINT_INCLUDE
2039 #define TRACEPOINT_INCLUDE "./tp.h"
2041 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2044 #include <lttng/tracepoint.h>
2047 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2048 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2053 #include <lttng/tracepoint-event.h>
2059 * `provider_name` with the name of your tracepoint provider.
2060 * `"tp.h"` with the name of your tracepoint provider header file.
2062 . Below the `#include <lttng/tracepoint.h>` line, put your
2063 <<defining-tracepoints,tracepoint definitions>>.
2065 Your tracepoint provider name must be unique amongst all the possible
2066 tracepoint provider names used on the same target system. We
2067 suggest to include the name of your project or company in the name,
2068 for example, `org_lttng_my_project_tpp`.
2070 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2071 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2072 write are the <<defining-tracepoints,tracepoint definitions>>.
2075 [[defining-tracepoints]]
2076 ===== Create a tracepoint definition
2078 A _tracepoint definition_ defines, for a given tracepoint:
2080 * Its **input arguments**. They are the macro parameters that the
2081 `tracepoint()` macro accepts for this particular tracepoint
2082 in the user application's source code.
2083 * Its **output event fields**. They are the sources of event fields
2084 that form the payload of any event that the execution of the
2085 `tracepoint()` macro emits for this particular tracepoint.
2087 You can create a tracepoint definition by using the
2088 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2090 <<tpp-header,tracepoint provider header file template>>.
2092 The syntax of the `TRACEPOINT_EVENT()` macro is:
2095 .`TRACEPOINT_EVENT()` macro syntax.
2098 /* Tracepoint provider name */
2101 /* Tracepoint name */
2104 /* Input arguments */
2109 /* Output event fields */
2118 * `provider_name` with your tracepoint provider name.
2119 * `tracepoint_name` with your tracepoint name.
2120 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2121 * `fields` with the <<tpp-def-output-fields,output event field>>
2124 This tracepoint emits events named `provider_name:tracepoint_name`.
2127 .Event name's length limitation
2129 The concatenation of the tracepoint provider name and the
2130 tracepoint name must not exceed **254 characters**. If it does, the
2131 instrumented application compiles and runs, but LTTng throws multiple
2132 warnings and you could experience serious issues.
2135 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2138 .`TP_ARGS()` macro syntax.
2147 * `type` with the C type of the argument.
2148 * `arg_name` with the argument name.
2150 You can repeat `type` and `arg_name` up to 10 times to have
2151 more than one argument.
2153 .`TP_ARGS()` usage with three arguments.
2165 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2166 tracepoint definition with no input arguments.
2168 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2169 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2170 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2171 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2174 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2175 C expression that the tracer evalutes at the `tracepoint()` macro site
2176 in the application's source code. This expression provides a field's
2177 source of data. The argument expression can include input argument names
2178 listed in the `TP_ARGS()` macro.
2180 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2181 must be unique within a given tracepoint definition.
2183 Here's a complete tracepoint definition example:
2185 .Tracepoint definition.
2187 The following tracepoint definition defines a tracepoint which takes
2188 three input arguments and has four output event fields.
2192 #include "my-custom-structure.h"
2198 const struct my_custom_structure*, my_custom_structure,
2203 ctf_string(query_field, query)
2204 ctf_float(double, ratio_field, ratio)
2205 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2206 ctf_integer(int, send_size, my_custom_structure->send_size)
2211 You can refer to this tracepoint definition with the `tracepoint()`
2212 macro in your application's source code like this:
2216 tracepoint(my_provider, my_tracepoint,
2217 my_structure, some_ratio, the_query);
2221 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2222 if they satisfy an enabled <<event,event rule>>.
2225 [[using-tracepoint-classes]]
2226 ===== Use a tracepoint class
2228 A _tracepoint class_ is a class of tracepoints which share the same
2229 output event field definitions. A _tracepoint instance_ is one
2230 instance of such a defined tracepoint class, with its own tracepoint
2233 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2234 shorthand which defines both a tracepoint class and a tracepoint
2235 instance at the same time.
2237 When you build a tracepoint provider package, the C or $$C++$$ compiler
2238 creates one serialization function for each **tracepoint class**. A
2239 serialization function is responsible for serializing the event fields
2240 of a tracepoint to a sub-buffer when tracing.
2242 For various performance reasons, when your situation requires multiple
2243 tracepoint definitions with different names, but with the same event
2244 fields, we recommend that you manually create a tracepoint class
2245 and instantiate as many tracepoint instances as needed. One positive
2246 effect of such a design, amongst other advantages, is that all
2247 tracepoint instances of the same tracepoint class reuse the same
2248 serialization function, thus reducing
2249 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2251 .Use a tracepoint class and tracepoint instances.
2253 Consider the following three tracepoint definitions:
2265 ctf_integer(int, userid, userid)
2266 ctf_integer(size_t, len, len)
2278 ctf_integer(int, userid, userid)
2279 ctf_integer(size_t, len, len)
2291 ctf_integer(int, userid, userid)
2292 ctf_integer(size_t, len, len)
2297 In this case, we create three tracepoint classes, with one implicit
2298 tracepoint instance for each of them: `get_account`, `get_settings`, and
2299 `get_transaction`. However, they all share the same event field names
2300 and types. Hence three identical, yet independent serialization
2301 functions are created when you build the tracepoint provider package.
2303 A better design choice is to define a single tracepoint class and three
2304 tracepoint instances:
2308 /* The tracepoint class */
2309 TRACEPOINT_EVENT_CLASS(
2310 /* Tracepoint provider name */
2313 /* Tracepoint class name */
2316 /* Input arguments */
2322 /* Output event fields */
2324 ctf_integer(int, userid, userid)
2325 ctf_integer(size_t, len, len)
2329 /* The tracepoint instances */
2330 TRACEPOINT_EVENT_INSTANCE(
2331 /* Tracepoint provider name */
2334 /* Tracepoint class name */
2337 /* Tracepoint name */
2340 /* Input arguments */
2346 TRACEPOINT_EVENT_INSTANCE(
2355 TRACEPOINT_EVENT_INSTANCE(
2368 [[assigning-log-levels]]
2369 ===== Assign a log level to a tracepoint definition
2371 You can assign an optional _log level_ to a
2372 <<defining-tracepoints,tracepoint definition>>.
2374 Assigning different levels of severity to tracepoint definitions can
2375 be useful: when you <<enabling-disabling-events,create an event rule>>,
2376 you can target tracepoints having a log level as severe as a specific
2379 The concept of LTTng-UST log levels is similar to the levels found
2380 in typical logging frameworks:
2382 * In a logging framework, the log level is given by the function
2383 or method name you use at the log statement site: `debug()`,
2384 `info()`, `warn()`, `error()`, and so on.
2385 * In LTTng-UST, you statically assign the log level to a tracepoint
2386 definition; any `tracepoint()` macro invocation which refers to
2387 this definition has this log level.
2389 You can assign a log level to a tracepoint definition with the
2390 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2391 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2392 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2395 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2398 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2400 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2405 * `provider_name` with the tracepoint provider name.
2406 * `tracepoint_name` with the tracepoint name.
2407 * `log_level` with the log level to assign to the tracepoint
2408 definition named `tracepoint_name` in the `provider_name`
2409 tracepoint provider.
2411 See man:lttng-ust(3) for a list of available log level names.
2413 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2417 /* Tracepoint definition */
2426 ctf_integer(int, userid, userid)
2427 ctf_integer(size_t, len, len)
2431 /* Log level assignment */
2432 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2438 ===== Create a tracepoint provider package source file
2440 A _tracepoint provider package source file_ is a C source file which
2441 includes a <<tpp-header,tracepoint provider header file>> to expand its
2442 macros into event serialization and other functions.
2444 You can always use the following tracepoint provider package source
2448 .Tracepoint provider package source file template.
2450 #define TRACEPOINT_CREATE_PROBES
2455 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2456 header file>> name. You may also include more than one tracepoint
2457 provider header file here to create a tracepoint provider package
2458 holding more than one tracepoint providers.
2461 [[probing-the-application-source-code]]
2462 ==== Add tracepoints to an application's source code
2464 Once you <<tpp-header,create a tracepoint provider header file>>, you
2465 can use the `tracepoint()` macro in your application's
2466 source code to insert the tracepoints that this header
2467 <<defining-tracepoints,defines>>.
2469 The `tracepoint()` macro takes at least two parameters: the tracepoint
2470 provider name and the tracepoint name. The corresponding tracepoint
2471 definition defines the other parameters.
2473 .`tracepoint()` usage.
2475 The following <<defining-tracepoints,tracepoint definition>> defines a
2476 tracepoint which takes two input arguments and has two output event
2480 .Tracepoint provider header file.
2482 #include "my-custom-structure.h"
2489 const char*, cmd_name
2492 ctf_string(cmd_name, cmd_name)
2493 ctf_integer(int, number_of_args, argc)
2498 You can refer to this tracepoint definition with the `tracepoint()`
2499 macro in your application's source code like this:
2502 .Application's source file.
2506 int main(int argc, char* argv[])
2508 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2514 Note how the application's source code includes
2515 the tracepoint provider header file containing the tracepoint
2516 definitions to use, path:{tp.h}.
2519 .`tracepoint()` usage with a complex tracepoint definition.
2521 Consider this complex tracepoint definition, where multiple event
2522 fields refer to the same input arguments in their argument expression
2526 .Tracepoint provider header file.
2528 /* For `struct stat` */
2529 #include <sys/types.h>
2530 #include <sys/stat.h>
2542 ctf_integer(int, my_constant_field, 23 + 17)
2543 ctf_integer(int, my_int_arg_field, my_int_arg)
2544 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2545 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2546 my_str_arg[2] + my_str_arg[3])
2547 ctf_string(my_str_arg_field, my_str_arg)
2548 ctf_integer_hex(off_t, size_field, st->st_size)
2549 ctf_float(double, size_dbl_field, (double) st->st_size)
2550 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2551 size_t, strlen(my_str_arg) / 2)
2556 You can refer to this tracepoint definition with the `tracepoint()`
2557 macro in your application's source code like this:
2560 .Application's source file.
2562 #define TRACEPOINT_DEFINE
2569 stat("/etc/fstab", &s);
2570 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2576 If you look at the event record that LTTng writes when tracing this
2577 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2578 it should look like this:
2580 .Event record fields
2582 |Field's name |Field's value
2583 |`my_constant_field` |40
2584 |`my_int_arg_field` |23
2585 |`my_int_arg_field2` |529
2587 |`my_str_arg_field` |`Hello, World!`
2588 |`size_field` |0x12d
2589 |`size_dbl_field` |301.0
2590 |`half_my_str_arg_field` |`Hello,`
2594 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2595 compute--they use the call stack, for example. To avoid this
2596 computation when the tracepoint is disabled, you can use the
2597 `tracepoint_enabled()` and `do_tracepoint()` macros.
2599 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2603 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2605 tracepoint_enabled(provider_name, tracepoint_name)
2606 do_tracepoint(provider_name, tracepoint_name, ...)
2611 * `provider_name` with the tracepoint provider name.
2612 * `tracepoint_name` with the tracepoint name.
2614 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2615 `tracepoint_name` from the provider named `provider_name` is enabled
2618 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2619 if the tracepoint is enabled. Using `tracepoint()` with
2620 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2621 the `tracepoint_enabled()` check, thus a race condition is
2622 possible in this situation:
2625 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2627 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2628 stuff = prepare_stuff();
2631 tracepoint(my_provider, my_tracepoint, stuff);
2634 If the tracepoint is enabled after the condition, then `stuff` is not
2635 prepared: the emitted event will either contain wrong data, or the whole
2636 application could crash (segmentation fault, for example).
2638 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2639 `STAP_PROBEV()` call. If you need it, you must emit
2643 [[building-tracepoint-providers-and-user-application]]
2644 ==== Build and link a tracepoint provider package and an application
2646 Once you have one or more <<tpp-header,tracepoint provider header
2647 files>> and a <<tpp-source,tracepoint provider package source file>>,
2648 you can create the tracepoint provider package by compiling its source
2649 file. From here, multiple build and run scenarios are possible. The
2650 following table shows common application and library configurations
2651 along with the required command lines to achieve them.
2653 In the following diagrams, we use the following file names:
2656 Executable application.
2659 Application's object file.
2662 Tracepoint provider package object file.
2665 Tracepoint provider package archive file.
2668 Tracepoint provider package shared object file.
2671 User library object file.
2674 User library shared object file.
2676 We use the following symbols in the diagrams of table below:
2679 .Symbols used in the build scenario diagrams.
2680 image::ust-sit-symbols.png[]
2682 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2683 variable in the following instructions.
2685 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2686 .Common tracepoint provider package scenarios.
2688 |Scenario |Instructions
2691 The instrumented application is statically linked with
2692 the tracepoint provider package object.
2694 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2697 include::../common/ust-sit-step-tp-o.txt[]
2699 To build the instrumented application:
2701 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2706 #define TRACEPOINT_DEFINE
2710 . Compile the application source file:
2719 . Build the application:
2724 gcc -o app app.o tpp.o -llttng-ust -ldl
2728 To run the instrumented application:
2730 * Start the application:
2740 The instrumented application is statically linked with the
2741 tracepoint provider package archive file.
2743 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2746 To create the tracepoint provider package archive file:
2748 . Compile the <<tpp-source,tracepoint provider package source file>>:
2757 . Create the tracepoint provider package archive file:
2766 To build the instrumented application:
2768 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2773 #define TRACEPOINT_DEFINE
2777 . Compile the application source file:
2786 . Build the application:
2791 gcc -o app app.o tpp.a -llttng-ust -ldl
2795 To run the instrumented application:
2797 * Start the application:
2807 The instrumented application is linked with the tracepoint provider
2808 package shared object.
2810 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2813 include::../common/ust-sit-step-tp-so.txt[]
2815 To build the instrumented application:
2817 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2822 #define TRACEPOINT_DEFINE
2826 . Compile the application source file:
2835 . Build the application:
2840 gcc -o app app.o -ldl -L. -ltpp
2844 To run the instrumented application:
2846 * Start the application:
2856 The tracepoint provider package shared object is preloaded before the
2857 instrumented application starts.
2859 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2862 include::../common/ust-sit-step-tp-so.txt[]
2864 To build the instrumented application:
2866 . In path:{app.c}, before including path:{tpp.h}, add the
2872 #define TRACEPOINT_DEFINE
2873 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2877 . Compile the application source file:
2886 . Build the application:
2891 gcc -o app app.o -ldl
2895 To run the instrumented application with tracing support:
2897 * Preload the tracepoint provider package shared object and
2898 start the application:
2903 LD_PRELOAD=./libtpp.so ./app
2907 To run the instrumented application without tracing support:
2909 * Start the application:
2919 The instrumented application dynamically loads the tracepoint provider
2920 package shared object.
2922 See the <<dlclose-warning,warning about `dlclose()`>>.
2924 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
2927 include::../common/ust-sit-step-tp-so.txt[]
2929 To build the instrumented application:
2931 . In path:{app.c}, before including path:{tpp.h}, add the
2937 #define TRACEPOINT_DEFINE
2938 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2942 . Compile the application source file:
2951 . Build the application:
2956 gcc -o app app.o -ldl
2960 To run the instrumented application:
2962 * Start the application:
2972 The application is linked with the instrumented user library.
2974 The instrumented user library is statically linked with the tracepoint
2975 provider package object file.
2977 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
2980 include::../common/ust-sit-step-tp-o-fpic.txt[]
2982 To build the instrumented user library:
2984 . In path:{emon.c}, before including path:{tpp.h}, add the
2990 #define TRACEPOINT_DEFINE
2994 . Compile the user library source file:
2999 gcc -I. -fpic -c emon.c
3003 . Build the user library shared object:
3008 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3012 To build the application:
3014 . Compile the application source file:
3023 . Build the application:
3028 gcc -o app app.o -L. -lemon
3032 To run the application:
3034 * Start the application:
3044 The application is linked with the instrumented user library.
3046 The instrumented user library is linked with the tracepoint provider
3047 package shared object.
3049 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3052 include::../common/ust-sit-step-tp-so.txt[]
3054 To build the instrumented user library:
3056 . In path:{emon.c}, before including path:{tpp.h}, add the
3062 #define TRACEPOINT_DEFINE
3066 . Compile the user library source file:
3071 gcc -I. -fpic -c emon.c
3075 . Build the user library shared object:
3080 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3084 To build the application:
3086 . Compile the application source file:
3095 . Build the application:
3100 gcc -o app app.o -L. -lemon
3104 To run the application:
3106 * Start the application:
3116 The tracepoint provider package shared object is preloaded before the
3119 The application is linked with the instrumented user library.
3121 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3124 include::../common/ust-sit-step-tp-so.txt[]
3126 To build the instrumented user library:
3128 . In path:{emon.c}, before including path:{tpp.h}, add the
3134 #define TRACEPOINT_DEFINE
3135 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3139 . Compile the user library source file:
3144 gcc -I. -fpic -c emon.c
3148 . Build the user library shared object:
3153 gcc -shared -o libemon.so emon.o -ldl
3157 To build the application:
3159 . Compile the application source file:
3168 . Build the application:
3173 gcc -o app app.o -L. -lemon
3177 To run the application with tracing support:
3179 * Preload the tracepoint provider package shared object and
3180 start the application:
3185 LD_PRELOAD=./libtpp.so ./app
3189 To run the application without tracing support:
3191 * Start the application:
3201 The application is linked with the instrumented user library.
3203 The instrumented user library dynamically loads the tracepoint provider
3204 package shared object.
3206 See the <<dlclose-warning,warning about `dlclose()`>>.
3208 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3211 include::../common/ust-sit-step-tp-so.txt[]
3213 To build the instrumented user library:
3215 . In path:{emon.c}, before including path:{tpp.h}, add the
3221 #define TRACEPOINT_DEFINE
3222 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3226 . Compile the user library source file:
3231 gcc -I. -fpic -c emon.c
3235 . Build the user library shared object:
3240 gcc -shared -o libemon.so emon.o -ldl
3244 To build the application:
3246 . Compile the application source file:
3255 . Build the application:
3260 gcc -o app app.o -L. -lemon
3264 To run the application:
3266 * Start the application:
3276 The application dynamically loads the instrumented user library.
3278 The instrumented user library is linked with the tracepoint provider
3279 package shared object.
3281 See the <<dlclose-warning,warning about `dlclose()`>>.
3283 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3286 include::../common/ust-sit-step-tp-so.txt[]
3288 To build the instrumented user library:
3290 . In path:{emon.c}, before including path:{tpp.h}, add the
3296 #define TRACEPOINT_DEFINE
3300 . Compile the user library source file:
3305 gcc -I. -fpic -c emon.c
3309 . Build the user library shared object:
3314 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3318 To build the application:
3320 . Compile the application source file:
3329 . Build the application:
3334 gcc -o app app.o -ldl -L. -lemon
3338 To run the application:
3340 * Start the application:
3350 The application dynamically loads the instrumented user library.
3352 The instrumented user library dynamically loads the tracepoint provider
3353 package shared object.
3355 See the <<dlclose-warning,warning about `dlclose()`>>.
3357 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3360 include::../common/ust-sit-step-tp-so.txt[]
3362 To build the instrumented user library:
3364 . In path:{emon.c}, before including path:{tpp.h}, add the
3370 #define TRACEPOINT_DEFINE
3371 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3375 . Compile the user library source file:
3380 gcc -I. -fpic -c emon.c
3384 . Build the user library shared object:
3389 gcc -shared -o libemon.so emon.o -ldl
3393 To build the application:
3395 . Compile the application source file:
3404 . Build the application:
3409 gcc -o app app.o -ldl -L. -lemon
3413 To run the application:
3415 * Start the application:
3425 The tracepoint provider package shared object is preloaded before the
3428 The application dynamically loads the instrumented user library.
3430 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3433 include::../common/ust-sit-step-tp-so.txt[]
3435 To build the instrumented user library:
3437 . In path:{emon.c}, before including path:{tpp.h}, add the
3443 #define TRACEPOINT_DEFINE
3444 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3448 . Compile the user library source file:
3453 gcc -I. -fpic -c emon.c
3457 . Build the user library shared object:
3462 gcc -shared -o libemon.so emon.o -ldl
3466 To build the application:
3468 . Compile the application source file:
3477 . Build the application:
3482 gcc -o app app.o -L. -lemon
3486 To run the application with tracing support:
3488 * Preload the tracepoint provider package shared object and
3489 start the application:
3494 LD_PRELOAD=./libtpp.so ./app
3498 To run the application without tracing support:
3500 * Start the application:
3510 The application is statically linked with the tracepoint provider
3511 package object file.
3513 The application is linked with the instrumented user library.
3515 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3518 include::../common/ust-sit-step-tp-o.txt[]
3520 To build the instrumented user library:
3522 . In path:{emon.c}, before including path:{tpp.h}, add the
3528 #define TRACEPOINT_DEFINE
3532 . Compile the user library source file:
3537 gcc -I. -fpic -c emon.c
3541 . Build the user library shared object:
3546 gcc -shared -o libemon.so emon.o
3550 To build the application:
3552 . Compile the application source file:
3561 . Build the application:
3566 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3570 To run the instrumented application:
3572 * Start the application:
3582 The application is statically linked with the tracepoint provider
3583 package object file.
3585 The application dynamically loads the instrumented user library.
3587 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3590 include::../common/ust-sit-step-tp-o.txt[]
3592 To build the application:
3594 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3599 #define TRACEPOINT_DEFINE
3603 . Compile the application source file:
3612 . Build the application:
3617 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3622 The `--export-dynamic` option passed to the linker is necessary for the
3623 dynamically loaded library to ``see'' the tracepoint symbols defined in
3626 To build the instrumented user library:
3628 . Compile the user library source file:
3633 gcc -I. -fpic -c emon.c
3637 . Build the user library shared object:
3642 gcc -shared -o libemon.so emon.o
3646 To run the application:
3648 * Start the application:
3660 .Do not use man:dlclose(3) on a tracepoint provider package
3662 Never use man:dlclose(3) on any shared object which:
3664 * Is linked with, statically or dynamically, a tracepoint provider
3666 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3667 package shared object.
3669 This is currently considered **unsafe** due to a lack of reference
3670 counting from LTTng-UST to the shared object.
3672 A known workaround (available since glibc 2.2) is to use the
3673 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3674 effect of not unloading the loaded shared object, even if man:dlclose(3)
3677 You can also preload the tracepoint provider package shared object with
3678 the env:LD_PRELOAD environment variable to overcome this limitation.
3682 [[using-lttng-ust-with-daemons]]
3683 ===== Use noch:{LTTng-UST} with daemons
3685 If your instrumented application calls man:fork(2), man:clone(2),
3686 or BSD's man:rfork(2), without a following man:exec(3)-family
3687 system call, you must preload the path:{liblttng-ust-fork.so} shared
3688 object when you start the application.
3692 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3695 If your tracepoint provider package is
3696 a shared library which you also preload, you must put both
3697 shared objects in env:LD_PRELOAD:
3701 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3707 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3709 If your instrumented application closes one or more file descriptors
3710 which it did not open itself, you must preload the
3711 path:{liblttng-ust-fd.so} shared object when you start the application:
3715 LD_PRELOAD=liblttng-ust-fd.so ./my-app
3718 Typical use cases include closing all the file descriptors after
3719 man:fork(2) or man:rfork(2) and buggy applications doing
3723 [[lttng-ust-pkg-config]]
3724 ===== Use noch:{pkg-config}
3726 On some distributions, LTTng-UST ships with a
3727 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3728 metadata file. If this is your case, then you can use cmd:pkg-config to
3729 build an application on the command line:
3733 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3737 [[instrumenting-32-bit-app-on-64-bit-system]]
3738 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3740 In order to trace a 32-bit application running on a 64-bit system,
3741 LTTng must use a dedicated 32-bit
3742 <<lttng-consumerd,consumer daemon>>.
3744 The following steps show how to build and install a 32-bit consumer
3745 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3746 build and install the 32-bit LTTng-UST libraries, and how to build and
3747 link an instrumented 32-bit application in that context.
3749 To build a 32-bit instrumented application for a 64-bit target system,
3750 assuming you have a fresh target system with no installed Userspace RCU
3753 . Download, build, and install a 32-bit version of Userspace RCU:
3759 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3760 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3761 cd userspace-rcu-0.9.* &&
3762 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3764 sudo make install &&
3769 . Using your distribution's package manager, or from source, install
3770 the following 32-bit versions of the following dependencies of
3771 LTTng-tools and LTTng-UST:
3774 * https://sourceforge.net/projects/libuuid/[libuuid]
3775 * http://directory.fsf.org/wiki/Popt[popt]
3776 * http://www.xmlsoft.org/[libxml2]
3779 . Download, build, and install a 32-bit version of the latest
3780 LTTng-UST{nbsp}{revision}:
3786 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.9.tar.bz2 &&
3787 tar -xf lttng-ust-latest-2.9.tar.bz2 &&
3788 cd lttng-ust-2.9.* &&
3789 ./configure --libdir=/usr/local/lib32 \
3790 CFLAGS=-m32 CXXFLAGS=-m32 \
3791 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3793 sudo make install &&
3800 Depending on your distribution,
3801 32-bit libraries could be installed at a different location than
3802 `/usr/lib32`. For example, Debian is known to install
3803 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3805 In this case, make sure to set `LDFLAGS` to all the
3806 relevant 32-bit library paths, for example:
3810 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3814 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3815 the 32-bit consumer daemon:
3821 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
3822 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
3823 cd lttng-tools-2.9.* &&
3824 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3825 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3826 --disable-bin-lttng --disable-bin-lttng-crash \
3827 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3829 cd src/bin/lttng-consumerd &&
3830 sudo make install &&
3835 . From your distribution or from source,
3836 <<installing-lttng,install>> the 64-bit versions of
3837 LTTng-UST and Userspace RCU.
3838 . Download, build, and install the 64-bit version of the
3839 latest LTTng-tools{nbsp}{revision}:
3845 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
3846 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
3847 cd lttng-tools-2.9.* &&
3848 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3849 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3851 sudo make install &&
3856 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3857 when linking your 32-bit application:
3860 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3861 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3864 For example, let's rebuild the quick start example in
3865 <<tracing-your-own-user-application,Trace a user application>> as an
3866 instrumented 32-bit application:
3871 gcc -m32 -c -I. hello-tp.c
3873 gcc -m32 -o hello hello.o hello-tp.o \
3874 -L/usr/lib32 -L/usr/local/lib32 \
3875 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3880 No special action is required to execute the 32-bit application and
3881 to trace it: use the command-line man:lttng(1) tool as usual.
3888 man:tracef(3) is a small LTTng-UST API designed for quick,
3889 man:printf(3)-like instrumentation without the burden of
3890 <<tracepoint-provider,creating>> and
3891 <<building-tracepoint-providers-and-user-application,building>>
3892 a tracepoint provider package.
3894 To use `tracef()` in your application:
3896 . In the C or C++ source files where you need to use `tracef()`,
3897 include `<lttng/tracef.h>`:
3902 #include <lttng/tracef.h>
3906 . In the application's source code, use `tracef()` like you would use
3914 tracef("my message: %d (%s)", my_integer, my_string);
3920 . Link your application with `liblttng-ust`:
3925 gcc -o app app.c -llttng-ust
3929 To trace the events that `tracef()` calls emit:
3931 * <<enabling-disabling-events,Create an event rule>> which matches the
3932 `lttng_ust_tracef:*` event name:
3937 lttng enable-event --userspace 'lttng_ust_tracef:*'
3942 .Limitations of `tracef()`
3944 The `tracef()` utility function was developed to make user space tracing
3945 super simple, albeit with notable disadvantages compared to
3946 <<defining-tracepoints,user-defined tracepoints>>:
3948 * All the emitted events have the same tracepoint provider and
3949 tracepoint names, respectively `lttng_ust_tracef` and `event`.
3950 * There is no static type checking.
3951 * The only event record field you actually get, named `msg`, is a string
3952 potentially containing the values you passed to `tracef()`
3953 using your own format string. This also means that you cannot filter
3954 events with a custom expression at run time because there are no
3956 * Since `tracef()` uses the C standard library's man:vasprintf(3)
3957 function behind the scenes to format the strings at run time, its
3958 expected performance is lower than with user-defined tracepoints,
3959 which do not require a conversion to a string.
3961 Taking this into consideration, `tracef()` is useful for some quick
3962 prototyping and debugging, but you should not consider it for any
3963 permanent and serious applicative instrumentation.
3969 ==== Use `tracelog()`
3971 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
3972 the difference that it accepts an additional log level parameter.
3974 The goal of `tracelog()` is to ease the migration from logging to
3977 To use `tracelog()` in your application:
3979 . In the C or C++ source files where you need to use `tracelog()`,
3980 include `<lttng/tracelog.h>`:
3985 #include <lttng/tracelog.h>
3989 . In the application's source code, use `tracelog()` like you would use
3990 man:printf(3), except for the first parameter which is the log
3998 tracelog(TRACE_WARNING, "my message: %d (%s)",
3999 my_integer, my_string);
4005 See man:lttng-ust(3) for a list of available log level names.
4007 . Link your application with `liblttng-ust`:
4012 gcc -o app app.c -llttng-ust
4016 To trace the events that `tracelog()` calls emit with a log level
4017 _as severe as_ a specific log level:
4019 * <<enabling-disabling-events,Create an event rule>> which matches the
4020 `lttng_ust_tracelog:*` event name and a minimum level
4026 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4027 --loglevel=TRACE_WARNING
4031 To trace the events that `tracelog()` calls emit with a
4032 _specific log level_:
4034 * Create an event rule which matches the `lttng_ust_tracelog:*`
4035 event name and a specific log level:
4040 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4041 --loglevel-only=TRACE_INFO
4046 [[prebuilt-ust-helpers]]
4047 === Prebuilt user space tracing helpers
4049 The LTTng-UST package provides a few helpers in the form or preloadable
4050 shared objects which automatically instrument system functions and
4053 The helper shared objects are normally found in dir:{/usr/lib}. If you
4054 built LTTng-UST <<building-from-source,from source>>, they are probably
4055 located in dir:{/usr/local/lib}.
4057 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4060 path:{liblttng-ust-libc-wrapper.so}::
4061 path:{liblttng-ust-pthread-wrapper.so}::
4062 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4063 memory and POSIX threads function tracing>>.
4065 path:{liblttng-ust-cyg-profile.so}::
4066 path:{liblttng-ust-cyg-profile-fast.so}::
4067 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4069 path:{liblttng-ust-dl.so}::
4070 <<liblttng-ust-dl,Dynamic linker tracing>>.
4072 To use a user space tracing helper with any user application:
4074 * Preload the helper shared object when you start the application:
4079 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4083 You can preload more than one helper:
4088 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4094 [[liblttng-ust-libc-pthread-wrapper]]
4095 ==== Instrument C standard library memory and POSIX threads functions
4097 The path:{liblttng-ust-libc-wrapper.so} and
4098 path:{liblttng-ust-pthread-wrapper.so} helpers
4099 add instrumentation to some C standard library and POSIX
4103 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4105 |TP provider name |TP name |Instrumented function
4107 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4108 |`calloc` |man:calloc(3)
4109 |`realloc` |man:realloc(3)
4110 |`free` |man:free(3)
4111 |`memalign` |man:memalign(3)
4112 |`posix_memalign` |man:posix_memalign(3)
4116 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4118 |TP provider name |TP name |Instrumented function
4120 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4121 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4122 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4123 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4126 When you preload the shared object, it replaces the functions listed
4127 in the previous tables by wrappers which contain tracepoints and call
4128 the replaced functions.
4131 [[liblttng-ust-cyg-profile]]
4132 ==== Instrument function entry and exit
4134 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4135 to the entry and exit points of functions.
4137 man:gcc(1) and man:clang(1) have an option named
4138 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4139 which generates instrumentation calls for entry and exit to functions.
4140 The LTTng-UST function tracing helpers,
4141 path:{liblttng-ust-cyg-profile.so} and
4142 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4143 to add tracepoints to the two generated functions (which contain
4144 `cyg_profile` in their names, hence the helper's name).
4146 To use the LTTng-UST function tracing helper, the source files to
4147 instrument must be built using the `-finstrument-functions` compiler
4150 There are two versions of the LTTng-UST function tracing helper:
4152 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4153 that you should only use when it can be _guaranteed_ that the
4154 complete event stream is recorded without any lost event record.
4155 Any kind of duplicate information is left out.
4157 Assuming no event record is lost, having only the function addresses on
4158 entry is enough to create a call graph, since an event record always
4159 contains the ID of the CPU that generated it.
4161 You can use a tool like man:addr2line(1) to convert function addresses
4162 back to source file names and line numbers.
4164 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4165 which also works in use cases where event records might get discarded or
4166 not recorded from application startup.
4167 In these cases, the trace analyzer needs more information to be
4168 able to reconstruct the program flow.
4170 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4171 points of this helper.
4173 All the tracepoints that this helper provides have the
4174 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4176 TIP: It's sometimes a good idea to limit the number of source files that
4177 you compile with the `-finstrument-functions` option to prevent LTTng
4178 from writing an excessive amount of trace data at run time. When using
4179 man:gcc(1), you can use the
4180 `-finstrument-functions-exclude-function-list` option to avoid
4181 instrument entries and exits of specific function names.
4186 ==== Instrument the dynamic linker
4188 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4189 man:dlopen(3) and man:dlclose(3) function calls.
4191 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4196 [[java-application]]
4197 === User space Java agent
4199 You can instrument any Java application which uses one of the following
4202 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4203 (JUL) core logging facilities.
4204 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4205 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4208 .LTTng-UST Java agent imported by a Java application.
4209 image::java-app.png[]
4211 Note that the methods described below are new in LTTng{nbsp}{revision}.
4212 Previous LTTng versions use another technique.
4214 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4215 and https://ci.lttng.org/[continuous integration], thus this version is
4216 directly supported. However, the LTTng-UST Java agent is also tested
4217 with OpenJDK{nbsp}7.
4222 ==== Use the LTTng-UST Java agent for `java.util.logging`
4224 To use the LTTng-UST Java agent in a Java application which uses
4225 `java.util.logging` (JUL):
4227 . In the Java application's source code, import the LTTng-UST
4228 log handler package for `java.util.logging`:
4233 import org.lttng.ust.agent.jul.LttngLogHandler;
4237 . Create an LTTng-UST JUL log handler:
4242 Handler lttngUstLogHandler = new LttngLogHandler();
4246 . Add this handler to the JUL loggers which should emit LTTng events:
4251 Logger myLogger = Logger.getLogger("some-logger");
4253 myLogger.addHandler(lttngUstLogHandler);
4257 . Use `java.util.logging` log statements and configuration as usual.
4258 The loggers with an attached LTTng-UST log handler can emit
4261 . Before exiting the application, remove the LTTng-UST log handler from
4262 the loggers attached to it and call its `close()` method:
4267 myLogger.removeHandler(lttngUstLogHandler);
4268 lttngUstLogHandler.close();
4272 This is not strictly necessary, but it is recommended for a clean
4273 disposal of the handler's resources.
4275 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4276 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4278 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4279 path] when you build the Java application.
4281 The JAR files are typically located in dir:{/usr/share/java}.
4283 IMPORTANT: The LTTng-UST Java agent must be
4284 <<installing-lttng,installed>> for the logging framework your
4287 .Use the LTTng-UST Java agent for `java.util.logging`.
4292 import java.io.IOException;
4293 import java.util.logging.Handler;
4294 import java.util.logging.Logger;
4295 import org.lttng.ust.agent.jul.LttngLogHandler;
4299 private static final int answer = 42;
4301 public static void main(String[] argv) throws Exception
4304 Logger logger = Logger.getLogger("jello");
4306 // Create an LTTng-UST log handler
4307 Handler lttngUstLogHandler = new LttngLogHandler();
4309 // Add the LTTng-UST log handler to our logger
4310 logger.addHandler(lttngUstLogHandler);
4313 logger.info("some info");
4314 logger.warning("some warning");
4316 logger.finer("finer information; the answer is " + answer);
4318 logger.severe("error!");
4320 // Not mandatory, but cleaner
4321 logger.removeHandler(lttngUstLogHandler);
4322 lttngUstLogHandler.close();
4331 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4334 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4335 <<enabling-disabling-events,create an event rule>> matching the
4336 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4341 lttng enable-event --jul jello
4345 Run the compiled class:
4349 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4352 <<basic-tracing-session-control,Stop tracing>> and inspect the
4362 In the resulting trace, an <<event,event record>> generated by a Java
4363 application using `java.util.logging` is named `lttng_jul:event` and
4364 has the following fields:
4367 Log record's message.
4373 Name of the class in which the log statement was executed.
4376 Name of the method in which the log statement was executed.
4379 Logging time (timestamp in milliseconds).
4382 Log level integer value.
4385 ID of the thread in which the log statement was executed.
4387 You can use the opt:lttng-enable-event(1):--loglevel or
4388 opt:lttng-enable-event(1):--loglevel-only option of the
4389 man:lttng-enable-event(1) command to target a range of JUL log levels
4390 or a specific JUL log level.
4395 ==== Use the LTTng-UST Java agent for Apache log4j
4397 To use the LTTng-UST Java agent in a Java application which uses
4400 . In the Java application's source code, import the LTTng-UST
4401 log appender package for Apache log4j:
4406 import org.lttng.ust.agent.log4j.LttngLogAppender;
4410 . Create an LTTng-UST log4j log appender:
4415 Appender lttngUstLogAppender = new LttngLogAppender();
4419 . Add this appender to the log4j loggers which should emit LTTng events:
4424 Logger myLogger = Logger.getLogger("some-logger");
4426 myLogger.addAppender(lttngUstLogAppender);
4430 . Use Apache log4j log statements and configuration as usual. The
4431 loggers with an attached LTTng-UST log appender can emit LTTng events.
4433 . Before exiting the application, remove the LTTng-UST log appender from
4434 the loggers attached to it and call its `close()` method:
4439 myLogger.removeAppender(lttngUstLogAppender);
4440 lttngUstLogAppender.close();
4444 This is not strictly necessary, but it is recommended for a clean
4445 disposal of the appender's resources.
4447 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4448 files, path:{lttng-ust-agent-common.jar} and
4449 path:{lttng-ust-agent-log4j.jar}, in the
4450 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4451 path] when you build the Java application.
4453 The JAR files are typically located in dir:{/usr/share/java}.
4455 IMPORTANT: The LTTng-UST Java agent must be
4456 <<installing-lttng,installed>> for the logging framework your
4459 .Use the LTTng-UST Java agent for Apache log4j.
4464 import org.apache.log4j.Appender;
4465 import org.apache.log4j.Logger;
4466 import org.lttng.ust.agent.log4j.LttngLogAppender;
4470 private static final int answer = 42;
4472 public static void main(String[] argv) throws Exception
4475 Logger logger = Logger.getLogger("jello");
4477 // Create an LTTng-UST log appender
4478 Appender lttngUstLogAppender = new LttngLogAppender();
4480 // Add the LTTng-UST log appender to our logger
4481 logger.addAppender(lttngUstLogAppender);
4484 logger.info("some info");
4485 logger.warn("some warning");
4487 logger.debug("debug information; the answer is " + answer);
4489 logger.fatal("error!");
4491 // Not mandatory, but cleaner
4492 logger.removeAppender(lttngUstLogAppender);
4493 lttngUstLogAppender.close();
4499 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4504 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4507 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4508 <<enabling-disabling-events,create an event rule>> matching the
4509 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4514 lttng enable-event --log4j jello
4518 Run the compiled class:
4522 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4525 <<basic-tracing-session-control,Stop tracing>> and inspect the
4535 In the resulting trace, an <<event,event record>> generated by a Java
4536 application using log4j is named `lttng_log4j:event` and
4537 has the following fields:
4540 Log record's message.
4546 Name of the class in which the log statement was executed.
4549 Name of the method in which the log statement was executed.
4552 Name of the file in which the executed log statement is located.
4555 Line number at which the log statement was executed.
4561 Log level integer value.
4564 Name of the Java thread in which the log statement was executed.
4566 You can use the opt:lttng-enable-event(1):--loglevel or
4567 opt:lttng-enable-event(1):--loglevel-only option of the
4568 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4569 or a specific log4j log level.
4573 [[java-application-context]]
4574 ==== Provide application-specific context fields in a Java application
4576 A Java application-specific context field is a piece of state provided
4577 by the application which <<adding-context,you can add>>, using the
4578 man:lttng-add-context(1) command, to each <<event,event record>>
4579 produced by the log statements of this application.
4581 For example, a given object might have a current request ID variable.
4582 You can create a context information retriever for this object and
4583 assign a name to this current request ID. You can then, using the
4584 man:lttng-add-context(1) command, add this context field by name to
4585 the JUL or log4j <<channel,channel>>.
4587 To provide application-specific context fields in a Java application:
4589 . In the Java application's source code, import the LTTng-UST
4590 Java agent context classes and interfaces:
4595 import org.lttng.ust.agent.context.ContextInfoManager;
4596 import org.lttng.ust.agent.context.IContextInfoRetriever;
4600 . Create a context information retriever class, that is, a class which
4601 implements the `IContextInfoRetriever` interface:
4606 class MyContextInfoRetriever implements IContextInfoRetriever
4609 public Object retrieveContextInfo(String key)
4611 if (key.equals("intCtx")) {
4613 } else if (key.equals("strContext")) {
4614 return "context value!";
4623 This `retrieveContextInfo()` method is the only member of the
4624 `IContextInfoRetriever` interface. Its role is to return the current
4625 value of a state by name to create a context field. The names of the
4626 context fields and which state variables they return depends on your
4629 All primitive types and objects are supported as context fields.
4630 When `retrieveContextInfo()` returns an object, the context field
4631 serializer calls its `toString()` method to add a string field to
4632 event records. The method can also return `null`, which means that
4633 no context field is available for the required name.
4635 . Register an instance of your context information retriever class to
4636 the context information manager singleton:
4641 IContextInfoRetriever cir = new MyContextInfoRetriever();
4642 ContextInfoManager cim = ContextInfoManager.getInstance();
4643 cim.registerContextInfoRetriever("retrieverName", cir);
4647 . Before exiting the application, remove your context information
4648 retriever from the context information manager singleton:
4653 ContextInfoManager cim = ContextInfoManager.getInstance();
4654 cim.unregisterContextInfoRetriever("retrieverName");
4658 This is not strictly necessary, but it is recommended for a clean
4659 disposal of some manager's resources.
4661 . Build your Java application with LTTng-UST Java agent support as
4662 usual, following the procedure for either the <<jul,JUL>> or
4663 <<log4j,Apache log4j>> framework.
4666 .Provide application-specific context fields in a Java application.
4671 import java.util.logging.Handler;
4672 import java.util.logging.Logger;
4673 import org.lttng.ust.agent.jul.LttngLogHandler;
4674 import org.lttng.ust.agent.context.ContextInfoManager;
4675 import org.lttng.ust.agent.context.IContextInfoRetriever;
4679 // Our context information retriever class
4680 private static class MyContextInfoRetriever
4681 implements IContextInfoRetriever
4684 public Object retrieveContextInfo(String key) {
4685 if (key.equals("intCtx")) {
4687 } else if (key.equals("strContext")) {
4688 return "context value!";
4695 private static final int answer = 42;
4697 public static void main(String args[]) throws Exception
4699 // Get the context information manager instance
4700 ContextInfoManager cim = ContextInfoManager.getInstance();
4702 // Create and register our context information retriever
4703 IContextInfoRetriever cir = new MyContextInfoRetriever();
4704 cim.registerContextInfoRetriever("myRetriever", cir);
4707 Logger logger = Logger.getLogger("jello");
4709 // Create an LTTng-UST log handler
4710 Handler lttngUstLogHandler = new LttngLogHandler();
4712 // Add the LTTng-UST log handler to our logger
4713 logger.addHandler(lttngUstLogHandler);
4716 logger.info("some info");
4717 logger.warning("some warning");
4719 logger.finer("finer information; the answer is " + answer);
4721 logger.severe("error!");
4723 // Not mandatory, but cleaner
4724 logger.removeHandler(lttngUstLogHandler);
4725 lttngUstLogHandler.close();
4726 cim.unregisterContextInfoRetriever("myRetriever");
4735 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4738 <<creating-destroying-tracing-sessions,Create a tracing session>>
4739 and <<enabling-disabling-events,create an event rule>> matching the
4745 lttng enable-event --jul jello
4748 <<adding-context,Add the application-specific context fields>> to the
4753 lttng add-context --jul --type='$app.myRetriever:intCtx'
4754 lttng add-context --jul --type='$app.myRetriever:strContext'
4757 <<basic-tracing-session-control,Start tracing>>:
4764 Run the compiled class:
4768 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4771 <<basic-tracing-session-control,Stop tracing>> and inspect the
4783 [[python-application]]
4784 === User space Python agent
4786 You can instrument a Python 2 or Python 3 application which uses the
4787 standard https://docs.python.org/3/library/logging.html[`logging`]
4790 Each log statement emits an LTTng event once the
4791 application module imports the
4792 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4795 .A Python application importing the LTTng-UST Python agent.
4796 image::python-app.png[]
4798 To use the LTTng-UST Python agent:
4800 . In the Python application's source code, import the LTTng-UST Python
4810 The LTTng-UST Python agent automatically adds its logging handler to the
4811 root logger at import time.
4813 Any log statement that the application executes before this import does
4814 not emit an LTTng event.
4816 IMPORTANT: The LTTng-UST Python agent must be
4817 <<installing-lttng,installed>>.
4819 . Use log statements and logging configuration as usual.
4820 Since the LTTng-UST Python agent adds a handler to the _root_
4821 logger, you can trace any log statement from any logger.
4823 .Use the LTTng-UST Python agent.
4834 logging.basicConfig()
4835 logger = logging.getLogger('my-logger')
4838 logger.debug('debug message')
4839 logger.info('info message')
4840 logger.warn('warn message')
4841 logger.error('error message')
4842 logger.critical('critical message')
4846 if __name__ == '__main__':
4850 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4851 logging handler which prints to the standard error stream, is not
4852 strictly required for LTTng-UST tracing to work, but in versions of
4853 Python preceding 3.2, you could see a warning message which indicates
4854 that no handler exists for the logger `my-logger`.
4856 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4857 <<enabling-disabling-events,create an event rule>> matching the
4858 `my-logger` Python logger, and <<basic-tracing-session-control,start
4864 lttng enable-event --python my-logger
4868 Run the Python script:
4875 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4885 In the resulting trace, an <<event,event record>> generated by a Python
4886 application is named `lttng_python:event` and has the following fields:
4889 Logging time (string).
4892 Log record's message.
4898 Name of the function in which the log statement was executed.
4901 Line number at which the log statement was executed.
4904 Log level integer value.
4907 ID of the Python thread in which the log statement was executed.
4910 Name of the Python thread in which the log statement was executed.
4912 You can use the opt:lttng-enable-event(1):--loglevel or
4913 opt:lttng-enable-event(1):--loglevel-only option of the
4914 man:lttng-enable-event(1) command to target a range of Python log levels
4915 or a specific Python log level.
4917 When an application imports the LTTng-UST Python agent, the agent tries
4918 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4919 <<start-sessiond,start the session daemon>> _before_ you run the Python
4920 application. If a session daemon is found, the agent tries to register
4921 to it during 5{nbsp}seconds, after which the application continues
4922 without LTTng tracing support. You can override this timeout value with
4923 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4926 If the session daemon stops while a Python application with an imported
4927 LTTng-UST Python agent runs, the agent retries to connect and to
4928 register to a session daemon every 3{nbsp}seconds. You can override this
4929 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4934 [[proc-lttng-logger-abi]]
4937 The `lttng-tracer` Linux kernel module, part of
4938 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4939 path:{/proc/lttng-logger} when it's loaded. Any application can write
4940 text data to this file to emit an LTTng event.
4943 .An application writes to the LTTng logger file to emit an LTTng event.
4944 image::lttng-logger.png[]
4946 The LTTng logger is the quickest method--not the most efficient,
4947 however--to add instrumentation to an application. It is designed
4948 mostly to instrument shell scripts:
4952 echo "Some message, some $variable" > /proc/lttng-logger
4955 Any event that the LTTng logger emits is named `lttng_logger` and
4956 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4957 other instrumentation points in the kernel tracing domain, **any Unix
4958 user** can <<enabling-disabling-events,create an event rule>> which
4959 matches its event name, not only the root user or users in the
4960 <<tracing-group,tracing group>>.
4962 To use the LTTng logger:
4964 * From any application, write text data to the path:{/proc/lttng-logger}
4967 The `msg` field of `lttng_logger` event records contains the
4970 NOTE: The maximum message length of an LTTng logger event is
4971 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4972 than one event to contain the remaining data.
4974 You should not use the LTTng logger to trace a user application which
4975 can be instrumented in a more efficient way, namely:
4977 * <<c-application,C and $$C++$$ applications>>.
4978 * <<java-application,Java applications>>.
4979 * <<python-application,Python applications>>.
4981 .Use the LTTng logger.
4986 echo 'Hello, World!' > /proc/lttng-logger
4988 df --human-readable --print-type / > /proc/lttng-logger
4991 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4992 <<enabling-disabling-events,create an event rule>> matching the
4993 `lttng_logger` Linux kernel tracepoint, and
4994 <<basic-tracing-session-control,start tracing>>:
4999 lttng enable-event --kernel lttng_logger
5003 Run the Bash script:
5010 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5021 [[instrumenting-linux-kernel]]
5022 === LTTng kernel tracepoints
5024 NOTE: This section shows how to _add_ instrumentation points to the
5025 Linux kernel. The kernel's subsystems are already thoroughly
5026 instrumented at strategic places for LTTng when you
5027 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5031 There are two methods to instrument the Linux kernel:
5033 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5034 tracepoint which uses the `TRACE_EVENT()` API.
5036 Choose this if you want to instrumentation a Linux kernel tree with an
5037 instrumentation point compatible with ftrace, perf, and SystemTap.
5039 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5040 instrument an out-of-tree kernel module.
5042 Choose this if you don't need ftrace, perf, or SystemTap support.
5046 [[linux-add-lttng-layer]]
5047 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5049 This section shows how to add an LTTng layer to existing ftrace
5050 instrumentation using the `TRACE_EVENT()` API.
5052 This section does not document the `TRACE_EVENT()` macro. You can
5053 read the following articles to learn more about this API:
5055 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
5056 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
5057 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
5059 The following procedure assumes that your ftrace tracepoints are
5060 correctly defined in their own header and that they are created in
5061 one source file using the `CREATE_TRACE_POINTS` definition.
5063 To add an LTTng layer over an existing ftrace tracepoint:
5065 . Make sure the following kernel configuration options are
5071 * `CONFIG_HIGH_RES_TIMERS`
5072 * `CONFIG_TRACEPOINTS`
5075 . Build the Linux source tree with your custom ftrace tracepoints.
5076 . Boot the resulting Linux image on your target system.
5078 Confirm that the tracepoints exist by looking for their names in the
5079 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5080 is your subsystem's name.
5082 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5088 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.9.tar.bz2 &&
5089 tar -xf lttng-modules-latest-2.9.tar.bz2 &&
5090 cd lttng-modules-2.9.*
5094 . In dir:{instrumentation/events/lttng-module}, relative to the root
5095 of the LTTng-modules source tree, create a header file named
5096 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5097 LTTng-modules tracepoint definitions using the LTTng-modules
5100 Start with this template:
5104 .path:{instrumentation/events/lttng-module/my_subsys.h}
5107 #define TRACE_SYSTEM my_subsys
5109 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5110 #define _LTTNG_MY_SUBSYS_H
5112 #include "../../../probes/lttng-tracepoint-event.h"
5113 #include <linux/tracepoint.h>
5115 LTTNG_TRACEPOINT_EVENT(
5117 * Format is identical to TRACE_EVENT()'s version for the three
5118 * following macro parameters:
5121 TP_PROTO(int my_int, const char *my_string),
5122 TP_ARGS(my_int, my_string),
5124 /* LTTng-modules specific macros */
5126 ctf_integer(int, my_int_field, my_int)
5127 ctf_string(my_bar_field, my_bar)
5131 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5133 #include "../../../probes/define_trace.h"
5137 The entries in the `TP_FIELDS()` section are the list of fields for the
5138 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5139 ftrace's `TRACE_EVENT()` macro.
5141 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5142 complete description of the available `ctf_*()` macros.
5144 . Create the LTTng-modules probe's kernel module C source file,
5145 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5150 .path:{probes/lttng-probe-my-subsys.c}
5152 #include <linux/module.h>
5153 #include "../lttng-tracer.h"
5156 * Build-time verification of mismatch between mainline
5157 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5158 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5160 #include <trace/events/my_subsys.h>
5162 /* Create LTTng tracepoint probes */
5163 #define LTTNG_PACKAGE_BUILD
5164 #define CREATE_TRACE_POINTS
5165 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5167 #include "../instrumentation/events/lttng-module/my_subsys.h"
5169 MODULE_LICENSE("GPL and additional rights");
5170 MODULE_AUTHOR("Your name <your-email>");
5171 MODULE_DESCRIPTION("LTTng my_subsys probes");
5172 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5173 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5174 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5175 LTTNG_MODULES_EXTRAVERSION);
5179 . Edit path:{probes/Makefile} and add your new kernel module object
5180 next to the existing ones:
5184 .path:{probes/Makefile}
5188 obj-m += lttng-probe-module.o
5189 obj-m += lttng-probe-power.o
5191 obj-m += lttng-probe-my-subsys.o
5197 . Build and install the LTTng kernel modules:
5202 make KERNELDIR=/path/to/linux
5203 sudo make modules_install
5207 Replace `/path/to/linux` with the path to the Linux source tree where
5208 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5210 Note that you can also use the
5211 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5212 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5213 C code that need to be executed before the event fields are recorded.
5215 The best way to learn how to use the previous LTTng-modules macros is to
5216 inspect the existing LTTng-modules tracepoint definitions in the
5217 dir:{instrumentation/events/lttng-module} header files. Compare them
5218 with the Linux kernel mainline versions in the
5219 dir:{include/trace/events} directory of the Linux source tree.
5223 [[lttng-tracepoint-event-code]]
5224 ===== Use custom C code to access the data for tracepoint fields
5226 Although we recommended to always use the
5227 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5228 the arguments and fields of an LTTng-modules tracepoint when possible,
5229 sometimes you need a more complex process to access the data that the
5230 tracer records as event record fields. In other words, you need local
5231 variables and multiple C{nbsp}statements instead of simple
5232 argument-based expressions that you pass to the
5233 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5235 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5236 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5237 a block of C{nbsp}code to be executed before LTTng records the fields.
5238 The structure of this macro is:
5241 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5243 LTTNG_TRACEPOINT_EVENT_CODE(
5245 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5246 * version for the following three macro parameters:
5249 TP_PROTO(int my_int, const char *my_string),
5250 TP_ARGS(my_int, my_string),
5252 /* Declarations of custom local variables */
5255 unsigned long b = 0;
5256 const char *name = "(undefined)";
5257 struct my_struct *my_struct;
5261 * Custom code which uses both tracepoint arguments
5262 * (in TP_ARGS()) and local variables (in TP_locvar()).
5264 * Local variables are actually members of a structure pointed
5265 * to by the special variable tp_locvar.
5269 tp_locvar->a = my_int + 17;
5270 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5271 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5272 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5273 put_my_struct(tp_locvar->my_struct);
5282 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5283 * version for this, except that tp_locvar members can be
5284 * used in the argument expression parameters of
5285 * the ctf_*() macros.
5288 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5289 ctf_integer(int, my_struct_a, tp_locvar->a)
5290 ctf_string(my_string_field, my_string)
5291 ctf_string(my_struct_name, tp_locvar->name)
5296 IMPORTANT: The C code defined in `TP_code()` must not have any side
5297 effects when executed. In particular, the code must not allocate
5298 memory or get resources without deallocating this memory or putting
5299 those resources afterwards.
5302 [[instrumenting-linux-kernel-tracing]]
5303 ==== Load and unload a custom probe kernel module
5305 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5306 kernel module>> in the kernel before it can emit LTTng events.
5308 To load the default probe kernel modules and a custom probe kernel
5311 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5312 probe modules to load when starting a root <<lttng-sessiond,session
5316 .Load the `my_subsys`, `usb`, and the default probe modules.
5320 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5325 You only need to pass the subsystem name, not the whole kernel module
5328 To load _only_ a given custom probe kernel module:
5330 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5331 modules to load when starting a root session daemon:
5334 .Load only the `my_subsys` and `usb` probe modules.
5338 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5343 To confirm that a probe module is loaded:
5350 lsmod | grep lttng_probe_usb
5354 To unload the loaded probe modules:
5356 * Kill the session daemon with `SIGTERM`:
5361 sudo pkill lttng-sessiond
5365 You can also use man:modprobe(8)'s `--remove` option if the session
5366 daemon terminates abnormally.
5369 [[controlling-tracing]]
5372 Once an application or a Linux kernel is
5373 <<instrumenting,instrumented>> for LTTng tracing,
5376 This section is divided in topics on how to use the various
5377 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5378 command-line tool>>, to _control_ the LTTng daemons and tracers.
5380 NOTE: In the following subsections, we refer to an man:lttng(1) command
5381 using its man page name. For example, instead of _Run the `create`
5382 command to..._, we use _Run the man:lttng-create(1) command to..._.
5386 === Start a session daemon
5388 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5389 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5392 You will see the following error when you run a command while no session
5396 Error: No session daemon is available
5399 The only command that automatically runs a session daemon is
5400 man:lttng-create(1), which you use to
5401 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5402 this is most of the time the first operation that you do, sometimes it's
5403 not. Some examples are:
5405 * <<list-instrumentation-points,List the available instrumentation points>>.
5406 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5408 [[tracing-group]] Each Unix user must have its own running session
5409 daemon to trace user applications. The session daemon that the root user
5410 starts is the only one allowed to control the LTTng kernel tracer. Users
5411 that are part of the _tracing group_ can control the root session
5412 daemon. The default tracing group name is `tracing`; you can set it to
5413 something else with the opt:lttng-sessiond(8):--group option when you
5414 start the root session daemon.
5416 To start a user session daemon:
5418 * Run man:lttng-sessiond(8):
5423 lttng-sessiond --daemonize
5427 To start the root session daemon:
5429 * Run man:lttng-sessiond(8) as the root user:
5434 sudo lttng-sessiond --daemonize
5438 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5439 start the session daemon in foreground.
5441 To stop a session daemon, use man:kill(1) on its process ID (standard
5444 Note that some Linux distributions could manage the LTTng session daemon
5445 as a service. In this case, you should use the service manager to
5446 start, restart, and stop session daemons.
5449 [[creating-destroying-tracing-sessions]]
5450 === Create and destroy a tracing session
5452 Almost all the LTTng control operations happen in the scope of
5453 a <<tracing-session,tracing session>>, which is the dialogue between the
5454 <<lttng-sessiond,session daemon>> and you.
5456 To create a tracing session with a generated name:
5458 * Use the man:lttng-create(1) command:
5467 The created tracing session's name is `auto` followed by the
5470 To create a tracing session with a specific name:
5472 * Use the optional argument of the man:lttng-create(1) command:
5477 lttng create my-session
5481 Replace `my-session` with the specific tracing session name.
5483 LTTng appends the creation date to the created tracing session's name.
5485 LTTng writes the traces of a tracing session in
5486 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5487 name of the tracing session. Note that the env:LTTNG_HOME environment
5488 variable defaults to `$HOME` if not set.
5490 To output LTTng traces to a non-default location:
5492 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5497 lttng create --output=/tmp/some-directory my-session
5501 You may create as many tracing sessions as you wish.
5503 To list all the existing tracing sessions for your Unix user:
5505 * Use the man:lttng-list(1) command:
5514 When you create a tracing session, it is set as the _current tracing
5515 session_. The following man:lttng(1) commands operate on the current
5516 tracing session when you don't specify one:
5518 [role="list-3-cols"]
5535 To change the current tracing session:
5537 * Use the man:lttng-set-session(1) command:
5542 lttng set-session new-session
5546 Replace `new-session` by the name of the new current tracing session.
5548 When you are done tracing in a given tracing session, you can destroy
5549 it. This operation frees the resources taken by the tracing session
5550 to destroy; it does not destroy the trace data that LTTng wrote for
5551 this tracing session.
5553 To destroy the current tracing session:
5555 * Use the man:lttng-destroy(1) command:
5565 [[list-instrumentation-points]]
5566 === List the available instrumentation points
5568 The <<lttng-sessiond,session daemon>> can query the running instrumented
5569 user applications and the Linux kernel to get a list of available
5570 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5571 they are tracepoints and system calls. For the user space tracing
5572 domain, they are tracepoints. For the other tracing domains, they are
5575 To list the available instrumentation points:
5577 * Use the man:lttng-list(1) command with the requested tracing domain's
5581 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5582 must be a root user, or it must be a member of the
5583 <<tracing-group,tracing group>>).
5584 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5585 kernel system calls (your Unix user must be a root user, or it must be
5586 a member of the tracing group).
5587 * opt:lttng-list(1):--userspace: user space tracepoints.
5588 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5589 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5590 * opt:lttng-list(1):--python: Python loggers.
5593 .List the available user space tracepoints.
5597 lttng list --userspace
5601 .List the available Linux kernel system call tracepoints.
5605 lttng list --kernel --syscall
5610 [[enabling-disabling-events]]
5611 === Create and enable an event rule
5613 Once you <<creating-destroying-tracing-sessions,create a tracing
5614 session>>, you can create <<event,event rules>> with the
5615 man:lttng-enable-event(1) command.
5617 You specify each condition with a command-line option. The available
5618 condition options are shown in the following table.
5620 [role="growable",cols="asciidoc,asciidoc,default"]
5621 .Condition command-line options for the man:lttng-enable-event(1) command.
5623 |Option |Description |Applicable tracing domains
5629 . +--probe=__ADDR__+
5630 . +--function=__ADDR__+
5633 Instead of using the default _tracepoint_ instrumentation type, use:
5635 . A Linux system call.
5636 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5637 . The entry and return points of a Linux function (symbol or address).
5641 |First positional argument.
5644 Tracepoint or system call name. In the case of a Linux KProbe or
5645 function, this is a custom name given to the event rule. With the
5646 JUL, log4j, and Python domains, this is a logger name.
5648 With a tracepoint, logger, or system call name, the last character
5649 can be `*` to match anything that remains.
5656 . +--loglevel=__LEVEL__+
5657 . +--loglevel-only=__LEVEL__+
5660 . Match only tracepoints or log statements with a logging level at
5661 least as severe as +__LEVEL__+.
5662 . Match only tracepoints or log statements with a logging level
5663 equal to +__LEVEL__+.
5665 See man:lttng-enable-event(1) for the list of available logging level
5668 |User space, JUL, log4j, and Python.
5670 |+--exclude=__EXCLUSIONS__+
5673 When you use a `*` character at the end of the tracepoint or logger
5674 name (first positional argument), exclude the specific names in the
5675 comma-delimited list +__EXCLUSIONS__+.
5678 User space, JUL, log4j, and Python.
5680 |+--filter=__EXPR__+
5683 Match only events which satisfy the expression +__EXPR__+.
5685 See man:lttng-enable-event(1) to learn more about the syntax of a
5692 You attach an event rule to a <<channel,channel>> on creation. If you do
5693 not specify the channel with the opt:lttng-enable-event(1):--channel
5694 option, and if the event rule to create is the first in its
5695 <<domain,tracing domain>> for a given tracing session, then LTTng
5696 creates a _default channel_ for you. This default channel is reused in
5697 subsequent invocations of the man:lttng-enable-event(1) command for the
5698 same tracing domain.
5700 An event rule is always enabled at creation time.
5702 The following examples show how you can combine the previous
5703 command-line options to create simple to more complex event rules.
5705 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5709 lttng enable-event --kernel sched_switch
5713 .Create an event rule matching four Linux kernel system calls (default channel).
5717 lttng enable-event --kernel --syscall open,write,read,close
5721 .Create event rules matching tracepoints with filter expressions (default channel).
5725 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5730 lttng enable-event --kernel --all \
5731 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5736 lttng enable-event --jul my_logger \
5737 --filter='$app.retriever:cur_msg_id > 3'
5740 IMPORTANT: Make sure to always quote the filter string when you
5741 use man:lttng(1) from a shell.
5744 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5748 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5751 IMPORTANT: Make sure to always quote the wildcard character when you
5752 use man:lttng(1) from a shell.
5755 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5759 lttng enable-event --python my-app.'*' \
5760 --exclude='my-app.module,my-app.hello'
5764 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5768 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5772 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5776 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5780 The event rules of a given channel form a whitelist: as soon as an
5781 emitted event passes one of them, LTTng can record the event. For
5782 example, an event named `my_app:my_tracepoint` emitted from a user space
5783 tracepoint with a `TRACE_ERROR` log level passes both of the following
5788 lttng enable-event --userspace my_app:my_tracepoint
5789 lttng enable-event --userspace my_app:my_tracepoint \
5790 --loglevel=TRACE_INFO
5793 The second event rule is redundant: the first one includes
5797 [[disable-event-rule]]
5798 === Disable an event rule
5800 To disable an event rule that you <<enabling-disabling-events,created>>
5801 previously, use the man:lttng-disable-event(1) command. This command
5802 disables _all_ the event rules (of a given tracing domain and channel)
5803 which match an instrumentation point. The other conditions are not
5804 supported as of LTTng{nbsp}{revision}.
5806 The LTTng tracer does not record an emitted event which passes
5807 a _disabled_ event rule.
5809 .Disable an event rule matching a Python logger (default channel).
5813 lttng disable-event --python my-logger
5817 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5821 lttng disable-event --jul '*'
5825 .Disable _all_ the event rules of the default channel.
5827 The opt:lttng-disable-event(1):--all-events option is not, like the
5828 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5829 equivalent of the event name `*` (wildcard): it disables _all_ the event
5830 rules of a given channel.
5834 lttng disable-event --jul --all-events
5838 NOTE: You cannot delete an event rule once you create it.
5842 === Get the status of a tracing session
5844 To get the status of the current tracing session, that is, its
5845 parameters, its channels, event rules, and their attributes:
5847 * Use the man:lttng-status(1) command:
5857 To get the status of any tracing session:
5859 * Use the man:lttng-list(1) command with the tracing session's name:
5864 lttng list my-session
5868 Replace `my-session` with the desired tracing session's name.
5871 [[basic-tracing-session-control]]
5872 === Start and stop a tracing session
5874 Once you <<creating-destroying-tracing-sessions,create a tracing
5876 <<enabling-disabling-events,create one or more event rules>>,
5877 you can start and stop the tracers for this tracing session.
5879 To start tracing in the current tracing session:
5881 * Use the man:lttng-start(1) command:
5890 LTTng is very flexible: you can launch user applications before
5891 or after the you start the tracers. The tracers only record the events
5892 if they pass enabled event rules and if they occur while the tracers are
5895 To stop tracing in the current tracing session:
5897 * Use the man:lttng-stop(1) command:
5906 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5907 records>> or lost sub-buffers since the last time you ran
5908 man:lttng-start(1), warnings are printed when you run the
5909 man:lttng-stop(1) command.
5912 [[enabling-disabling-channels]]
5913 === Create a channel
5915 Once you create a tracing session, you can create a <<channel,channel>>
5916 with the man:lttng-enable-channel(1) command.
5918 Note that LTTng automatically creates a default channel when, for a
5919 given <<domain,tracing domain>>, no channels exist and you
5920 <<enabling-disabling-events,create>> the first event rule. This default
5921 channel is named `channel0` and its attributes are set to reasonable
5922 values. Therefore, you only need to create a channel when you need
5923 non-default attributes.
5925 You specify each non-default channel attribute with a command-line
5926 option when you use the man:lttng-enable-channel(1) command. The
5927 available command-line options are:
5929 [role="growable",cols="asciidoc,asciidoc"]
5930 .Command-line options for the man:lttng-enable-channel(1) command.
5932 |Option |Description
5938 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5939 the default _discard_ mode.
5941 |`--buffers-pid` (user space tracing domain only)
5944 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5945 instead of the default per-user buffering scheme.
5947 |+--subbuf-size=__SIZE__+
5950 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5951 either for each Unix user (default), or for each instrumented process.
5953 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5955 |+--num-subbuf=__COUNT__+
5958 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5959 for each Unix user (default), or for each instrumented process.
5961 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5963 |+--tracefile-size=__SIZE__+
5966 Set the maximum size of each trace file that this channel writes within
5967 a stream to +__SIZE__+ bytes instead of no maximum.
5969 See <<tracefile-rotation,Trace file count and size>>.
5971 |+--tracefile-count=__COUNT__+
5974 Limit the number of trace files that this channel creates to
5975 +__COUNT__+ channels instead of no limit.
5977 See <<tracefile-rotation,Trace file count and size>>.
5979 |+--switch-timer=__PERIODUS__+
5982 Set the <<channel-switch-timer,switch timer period>>
5983 to +__PERIODUS__+{nbsp}µs.
5985 |+--read-timer=__PERIODUS__+
5988 Set the <<channel-read-timer,read timer period>>
5989 to +__PERIODUS__+{nbsp}µs.
5991 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5994 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5998 You can only create a channel in the Linux kernel and user space
5999 <<domain,tracing domains>>: other tracing domains have their own channel
6000 created on the fly when <<enabling-disabling-events,creating event
6005 Because of a current LTTng limitation, you must create all channels
6006 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6007 tracing session, that is, before the first time you run
6010 Since LTTng automatically creates a default channel when you use the
6011 man:lttng-enable-event(1) command with a specific tracing domain, you
6012 cannot, for example, create a Linux kernel event rule, start tracing,
6013 and then create a user space event rule, because no user space channel
6014 exists yet and it's too late to create one.
6016 For this reason, make sure to configure your channels properly
6017 before starting the tracers for the first time!
6020 The following examples show how you can combine the previous
6021 command-line options to create simple to more complex channels.
6023 .Create a Linux kernel channel with default attributes.
6027 lttng enable-channel --kernel my-channel
6031 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6035 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6036 --buffers-pid my-channel
6040 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
6044 lttng enable-channel --kernel --tracefile-count=8 \
6045 --tracefile-size=4194304 my-channel
6049 .Create a user space channel in overwrite (or _flight recorder_) mode.
6053 lttng enable-channel --userspace --overwrite my-channel
6057 You can <<enabling-disabling-events,create>> the same event rule in
6058 two different channels:
6062 lttng enable-event --userspace --channel=my-channel app:tp
6063 lttng enable-event --userspace --channel=other-channel app:tp
6066 If both channels are enabled, when a tracepoint named `app:tp` is
6067 reached, LTTng records two events, one for each channel.
6071 === Disable a channel
6073 To disable a specific channel that you <<enabling-disabling-channels,created>>
6074 previously, use the man:lttng-disable-channel(1) command.
6076 .Disable a specific Linux kernel channel.
6080 lttng disable-channel --kernel my-channel
6084 The state of a channel precedes the individual states of event rules
6085 attached to it: event rules which belong to a disabled channel, even if
6086 they are enabled, are also considered disabled.
6090 === Add context fields to a channel
6092 Event record fields in trace files provide important information about
6093 events that occured previously, but sometimes some external context may
6094 help you solve a problem faster. Examples of context fields are:
6096 * The **process ID**, **thread ID**, **process name**, and
6097 **process priority** of the thread in which the event occurs.
6098 * The **hostname** of the system on which the event occurs.
6099 * The current values of many possible **performance counters** using
6101 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6103 ** Branch instructions, misses, and loads.
6105 * Any context defined at the application level (supported for the
6106 JUL and log4j <<domain,tracing domains>>).
6108 To get the full list of available context fields, see
6109 `lttng add-context --list`. Some context fields are reserved for a
6110 specific <<domain,tracing domain>> (Linux kernel or user space).
6112 You add context fields to <<channel,channels>>. All the events
6113 that a channel with added context fields records contain those fields.
6115 To add context fields to one or all the channels of a given tracing
6118 * Use the man:lttng-add-context(1) command.
6120 .Add context fields to all the channels of the current tracing session.
6122 The following command line adds the virtual process identifier and
6123 the per-thread CPU cycles count fields to all the user space channels
6124 of the current tracing session.
6128 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6132 .Add performance counter context fields by raw ID
6134 See man:lttng-add-context(1) for the exact format of the context field
6135 type, which is partly compatible with the format used in
6140 lttng add-context --userspace --type=perf:thread:raw:r0110:test
6141 lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6145 .Add a context field to a specific channel.
6147 The following command line adds the thread identifier context field
6148 to the Linux kernel channel named `my-channel` in the current
6153 lttng add-context --kernel --channel=my-channel --type=tid
6157 .Add an application-specific context field to a specific channel.
6159 The following command line adds the `cur_msg_id` context field of the
6160 `retriever` context retriever for all the instrumented
6161 <<java-application,Java applications>> recording <<event,event records>>
6162 in the channel named `my-channel`:
6166 lttng add-context --kernel --channel=my-channel \
6167 --type='$app:retriever:cur_msg_id'
6170 IMPORTANT: Make sure to always quote the `$` character when you
6171 use man:lttng-add-context(1) from a shell.
6174 NOTE: You cannot remove context fields from a channel once you add it.
6179 === Track process IDs
6181 It's often useful to allow only specific process IDs (PIDs) to emit
6182 events. For example, you may wish to record all the system calls made by
6183 a given process (à la http://linux.die.net/man/1/strace[strace]).
6185 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6186 purpose. Both commands operate on a whitelist of process IDs. You _add_
6187 entries to this whitelist with the man:lttng-track(1) command and remove
6188 entries with the man:lttng-untrack(1) command. Any process which has one
6189 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6190 an enabled <<event,event rule>>.
6192 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6193 process with a given tracked ID exit and another process be given this
6194 ID, then the latter would also be allowed to emit events.
6196 .Track and untrack process IDs.
6198 For the sake of the following example, assume the target system has 16
6202 <<creating-destroying-tracing-sessions,create a tracing session>>,
6203 the whitelist contains all the possible PIDs:
6206 .All PIDs are tracked.
6207 image::track-all.png[]
6209 When the whitelist is full and you use the man:lttng-track(1) command to
6210 specify some PIDs to track, LTTng first clears the whitelist, then it
6211 tracks the specific PIDs. After:
6215 lttng track --pid=3,4,7,10,13
6221 .PIDs 3, 4, 7, 10, and 13 are tracked.
6222 image::track-3-4-7-10-13.png[]
6224 You can add more PIDs to the whitelist afterwards:
6228 lttng track --pid=1,15,16
6234 .PIDs 1, 15, and 16 are added to the whitelist.
6235 image::track-1-3-4-7-10-13-15-16.png[]
6237 The man:lttng-untrack(1) command removes entries from the PID tracker's
6238 whitelist. Given the previous example, the following command:
6242 lttng untrack --pid=3,7,10,13
6245 leads to this whitelist:
6248 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6249 image::track-1-4-15-16.png[]
6251 LTTng can track all possible PIDs again using the opt:track(1):--all
6256 lttng track --pid --all
6259 The result is, again:
6262 .All PIDs are tracked.
6263 image::track-all.png[]
6266 .Track only specific PIDs
6268 A very typical use case with PID tracking is to start with an empty
6269 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6270 then add PIDs manually while tracers are active. You can accomplish this
6271 by using the opt:lttng-untrack(1):--all option of the
6272 man:lttng-untrack(1) command to clear the whitelist after you
6273 <<creating-destroying-tracing-sessions,create a tracing session>>:
6277 lttng untrack --pid --all
6283 .No PIDs are tracked.
6284 image::untrack-all.png[]
6286 If you trace with this whitelist configuration, the tracer records no
6287 events for this <<domain,tracing domain>> because no processes are
6288 tracked. You can use the man:lttng-track(1) command as usual to track
6289 specific PIDs, for example:
6293 lttng track --pid=6,11
6299 .PIDs 6 and 11 are tracked.
6300 image::track-6-11.png[]
6305 [[saving-loading-tracing-session]]
6306 === Save and load tracing session configurations
6308 Configuring a <<tracing-session,tracing session>> can be long. Some of
6309 the tasks involved are:
6311 * <<enabling-disabling-channels,Create channels>> with
6312 specific attributes.
6313 * <<adding-context,Add context fields>> to specific channels.
6314 * <<enabling-disabling-events,Create event rules>> with specific log
6315 level and filter conditions.
6317 If you use LTTng to solve real world problems, chances are you have to
6318 record events using the same tracing session setup over and over,
6319 modifying a few variables each time in your instrumented program
6320 or environment. To avoid constant tracing session reconfiguration,
6321 the man:lttng(1) command-line tool can save and load tracing session
6322 configurations to/from XML files.
6324 To save a given tracing session configuration:
6326 * Use the man:lttng-save(1) command:
6331 lttng save my-session
6335 Replace `my-session` with the name of the tracing session to save.
6337 LTTng saves tracing session configurations to
6338 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6339 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6340 the opt:lttng-save(1):--output-path option to change this destination
6343 LTTng saves all configuration parameters, for example:
6345 * The tracing session name.
6346 * The trace data output path.
6347 * The channels with their state and all their attributes.
6348 * The context fields you added to channels.
6349 * The event rules with their state, log level and filter conditions.
6351 To load a tracing session:
6353 * Use the man:lttng-load(1) command:
6358 lttng load my-session
6362 Replace `my-session` with the name of the tracing session to load.
6364 When LTTng loads a configuration, it restores your saved tracing session
6365 as if you just configured it manually.
6367 See man:lttng(1) for the complete list of command-line options. You
6368 can also save and load all many sessions at a time, and decide in which
6369 directory to output the XML files.
6372 [[sending-trace-data-over-the-network]]
6373 === Send trace data over the network
6375 LTTng can send the recorded trace data to a remote system over the
6376 network instead of writing it to the local file system.
6378 To send the trace data over the network:
6380 . On the _remote_ system (which can also be the target system),
6381 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6390 . On the _target_ system, create a tracing session configured to
6391 send trace data over the network:
6396 lttng create my-session --set-url=net://remote-system
6400 Replace `remote-system` by the host name or IP address of the
6401 remote system. See man:lttng-create(1) for the exact URL format.
6403 . On the target system, use the man:lttng(1) command-line tool as usual.
6404 When tracing is active, the target's consumer daemon sends sub-buffers
6405 to the relay daemon running on the remote system intead of flushing
6406 them to the local file system. The relay daemon writes the received
6407 packets to the local file system.
6409 The relay daemon writes trace files to
6410 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6411 +__hostname__+ is the host name of the target system and +__session__+
6412 is the tracing session name. Note that the env:LTTNG_HOME environment
6413 variable defaults to `$HOME` if not set. Use the
6414 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6415 trace files to another base directory.
6420 === View events as LTTng emits them (noch:{LTTng} live)
6422 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6423 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6424 display events as LTTng emits them on the target system while tracing is
6427 The relay daemon creates a _tee_: it forwards the trace data to both
6428 the local file system and to connected live viewers:
6431 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6436 . On the _target system_, create a <<tracing-session,tracing session>>
6442 lttng create --live my-session
6446 This spawns a local relay daemon.
6448 . Start the live viewer and configure it to connect to the relay
6449 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6454 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6461 * `hostname` with the host name of the target system.
6462 * `my-session` with the name of the tracing session to view.
6465 . Configure the tracing session as usual with the man:lttng(1)
6466 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6468 You can list the available live tracing sessions with Babeltrace:
6472 babeltrace --input-format=lttng-live net://localhost
6475 You can start the relay daemon on another system. In this case, you need
6476 to specify the relay daemon's URL when you create the tracing session
6477 with the opt:lttng-create(1):--set-url option. You also need to replace
6478 `localhost` in the procedure above with the host name of the system on
6479 which the relay daemon is running.
6481 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6482 command-line options.
6486 [[taking-a-snapshot]]
6487 === Take a snapshot of the current sub-buffers of a tracing session
6489 The normal behavior of LTTng is to append full sub-buffers to growing
6490 trace data files. This is ideal to keep a full history of the events
6491 that occurred on the target system, but it can
6492 represent too much data in some situations. For example, you may wish
6493 to trace your application continuously until some critical situation
6494 happens, in which case you only need the latest few recorded
6495 events to perform the desired analysis, not multi-gigabyte trace files.
6497 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6498 current sub-buffers of a given <<tracing-session,tracing session>>.
6499 LTTng can write the snapshot to the local file system or send it over
6504 . Create a tracing session in _snapshot mode_:
6509 lttng create --snapshot my-session
6513 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6514 <<channel,channels>> created in this mode is automatically set to
6515 _overwrite_ (flight recorder mode).
6517 . Configure the tracing session as usual with the man:lttng(1)
6518 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6520 . **Optional**: When you need to take a snapshot,
6521 <<basic-tracing-session-control,stop tracing>>.
6523 You can take a snapshot when the tracers are active, but if you stop
6524 them first, you are sure that the data in the sub-buffers does not
6525 change before you actually take the snapshot.
6532 lttng snapshot record --name=my-first-snapshot
6536 LTTng writes the current sub-buffers of all the current tracing
6537 session's channels to trace files on the local file system. Those trace
6538 files have `my-first-snapshot` in their name.
6540 There is no difference between the format of a normal trace file and the
6541 format of a snapshot: viewers of LTTng traces also support LTTng
6544 By default, LTTng writes snapshot files to the path shown by
6545 `lttng snapshot list-output`. You can change this path or decide to send
6546 snapshots over the network using either:
6548 . An output path or URL that you specify when you create the
6550 . An snapshot output path or URL that you add using
6551 `lttng snapshot add-output`
6552 . An output path or URL that you provide directly to the
6553 `lttng snapshot record` command.
6555 Method 3 overrides method 2, which overrides method 1. When you
6556 specify a URL, a relay daemon must listen on a remote system (see
6557 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6562 === Use the machine interface
6564 With any command of the man:lttng(1) command-line tool, you can set the
6565 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6566 XML machine interface output, for example:
6570 lttng --mi=xml enable-event --kernel --syscall open
6573 A schema definition (XSD) is
6574 https://github.com/lttng/lttng-tools/blob/stable-2.9/src/common/mi-lttng-3.0.xsd[available]
6575 to ease the integration with external tools as much as possible.
6579 [[metadata-regenerate]]
6580 === Regenerate the metadata of an LTTng trace
6582 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6583 data stream files and a metadata file. This metadata file contains,
6584 amongst other things, information about the offset of the clock sources
6585 used to timestamp <<event,event records>> when tracing.
6587 If, once a <<tracing-session,tracing session>> is
6588 <<basic-tracing-session-control,started>>, a major
6589 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6590 happens, the trace's clock offset also needs to be updated. You
6591 can use the `metadata` item of the man:lttng-regenerate(1) command
6594 The main use case of this command is to allow a system to boot with
6595 an incorrect wall time and trace it with LTTng before its wall time
6596 is corrected. Once the system is known to be in a state where its
6597 wall time is correct, it can run `lttng regenerate metadata`.
6599 To regenerate the metadata of an LTTng trace:
6601 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6606 lttng regenerate metadata
6612 `lttng regenerate metadata` has the following limitations:
6614 * Tracing session <<creating-destroying-tracing-sessions,created>>
6616 * User space <<channel,channels>>, if any, are using
6617 <<channel-buffering-schemes,per-user buffering>>.
6622 [[regenerate-statedump]]
6623 === Regenerate the state dump of a tracing session
6625 The LTTng kernel and user space tracers generate state dump
6626 <<event,event records>> when the application starts or when you
6627 <<basic-tracing-session-control,start a tracing session>>. An analysis
6628 can use the state dump event records to set an initial state before it
6629 builds the rest of the state from the following event records.
6630 http://tracecompass.org/[Trace Compass] is a notable example of an
6631 application which uses the state dump of an LTTng trace.
6633 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6634 state dump event records are not included in the snapshot because they
6635 were recorded to a sub-buffer that has been consumed or overwritten
6638 You can use the `lttng regenerate statedump` command to emit the state
6639 dump event records again.
6641 To regenerate the state dump of the current tracing session, provided
6642 create it in snapshot mode, before you take a snapshot:
6644 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6649 lttng regenerate statedump
6653 . <<basic-tracing-session-control,Stop the tracing session>>:
6662 . <<taking-a-snapshot,Take a snapshot>>:
6667 lttng snapshot record --name=my-snapshot
6671 Depending on the event throughput, you should run steps 1 and 2
6672 as closely as possible.
6674 NOTE: To record the state dump events, you need to
6675 <<enabling-disabling-events,create event rules>> which enable them.
6676 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6677 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6681 [[persistent-memory-file-systems]]
6682 === Record trace data on persistent memory file systems
6684 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6685 (NVRAM) is random-access memory that retains its information when power
6686 is turned off (non-volatile). Systems with such memory can store data
6687 structures in RAM and retrieve them after a reboot, without flushing
6688 to typical _storage_.
6690 Linux supports NVRAM file systems thanks to either
6691 http://pramfs.sourceforge.net/[PRAMFS] or
6692 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6693 (requires Linux 4.1+).
6695 This section does not describe how to operate such file systems;
6696 we assume that you have a working persistent memory file system.
6698 When you create a <<tracing-session,tracing session>>, you can specify
6699 the path of the shared memory holding the sub-buffers. If you specify a
6700 location on an NVRAM file system, then you can retrieve the latest
6701 recorded trace data when the system reboots after a crash.
6703 To record trace data on a persistent memory file system and retrieve the
6704 trace data after a system crash:
6706 . Create a tracing session with a sub-buffer shared memory path located
6707 on an NVRAM file system:
6712 lttng create --shm-path=/path/to/shm
6716 . Configure the tracing session as usual with the man:lttng(1)
6717 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6719 . After a system crash, use the man:lttng-crash(1) command-line tool to
6720 view the trace data recorded on the NVRAM file system:
6725 lttng-crash /path/to/shm
6729 The binary layout of the ring buffer files is not exactly the same as
6730 the trace files layout. This is why you need to use man:lttng-crash(1)
6731 instead of your preferred trace viewer directly.
6733 To convert the ring buffer files to LTTng trace files:
6735 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6740 lttng-crash --extract=/path/to/trace /path/to/shm
6748 [[lttng-modules-ref]]
6749 === noch:{LTTng-modules}
6753 [[lttng-tracepoint-enum]]
6754 ==== `LTTNG_TRACEPOINT_ENUM()` usage
6756 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
6760 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
6765 * `name` with the name of the enumeration (C identifier, unique
6766 amongst all the defined enumerations).
6767 * `entries` with a list of enumeration entries.
6769 The available enumeration entry macros are:
6771 +ctf_enum_value(__name__, __value__)+::
6772 Entry named +__name__+ mapped to the integral value +__value__+.
6774 +ctf_enum_range(__name__, __begin__, __end__)+::
6775 Entry named +__name__+ mapped to the range of integral values between
6776 +__begin__+ (included) and +__end__+ (included).
6778 +ctf_enum_auto(__name__)+::
6779 Entry named +__name__+ mapped to the integral value following the
6780 last mapping's value.
6782 The last value of a `ctf_enum_value()` entry is its +__value__+
6785 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
6787 If `ctf_enum_auto()` is the first entry in the list, its integral
6790 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
6791 to use a defined enumeration as a tracepoint field.
6793 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
6797 LTTNG_TRACEPOINT_ENUM(
6800 ctf_enum_auto("AUTO: EXPECT 0")
6801 ctf_enum_value("VALUE: 23", 23)
6802 ctf_enum_value("VALUE: 27", 27)
6803 ctf_enum_auto("AUTO: EXPECT 28")
6804 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
6805 ctf_enum_auto("AUTO: EXPECT 304")
6813 [[lttng-modules-tp-fields]]
6814 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6816 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6817 tracepoint fields, which must be listed within `TP_FIELDS()` in
6818 `LTTNG_TRACEPOINT_EVENT()`, are:
6820 [role="func-desc growable",cols="asciidoc,asciidoc"]
6821 .Available macros to define LTTng-modules tracepoint fields
6823 |Macro |Description and parameters
6826 +ctf_integer(__t__, __n__, __e__)+
6828 +ctf_integer_nowrite(__t__, __n__, __e__)+
6830 +ctf_user_integer(__t__, __n__, __e__)+
6832 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6834 Standard integer, displayed in base 10.
6837 Integer C type (`int`, `long`, `size_t`, ...).
6843 Argument expression.
6846 +ctf_integer_hex(__t__, __n__, __e__)+
6848 +ctf_user_integer_hex(__t__, __n__, __e__)+
6850 Standard integer, displayed in base 16.
6859 Argument expression.
6861 |+ctf_integer_oct(__t__, __n__, __e__)+
6863 Standard integer, displayed in base 8.
6872 Argument expression.
6875 +ctf_integer_network(__t__, __n__, __e__)+
6877 +ctf_user_integer_network(__t__, __n__, __e__)+
6879 Integer in network byte order (big-endian), displayed in base 10.
6888 Argument expression.
6891 +ctf_integer_network_hex(__t__, __n__, __e__)+
6893 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6895 Integer in network byte order, displayed in base 16.
6904 Argument expression.
6907 +ctf_enum(__N__, __t__, __n__, __e__)+
6909 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
6911 +ctf_user_enum(__N__, __t__, __n__, __e__)+
6913 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
6918 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
6921 Integer C type (`int`, `long`, `size_t`, ...).
6927 Argument expression.
6930 +ctf_string(__n__, __e__)+
6932 +ctf_string_nowrite(__n__, __e__)+
6934 +ctf_user_string(__n__, __e__)+
6936 +ctf_user_string_nowrite(__n__, __e__)+
6938 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6944 Argument expression.
6947 +ctf_array(__t__, __n__, __e__, __s__)+
6949 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6951 +ctf_user_array(__t__, __n__, __e__, __s__)+
6953 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6955 Statically-sized array of integers.
6958 Array element C type.
6964 Argument expression.
6970 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
6972 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6974 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
6976 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6978 Statically-sized array of bits.
6980 The type of +__e__+ must be an integer type. +__s__+ is the number
6981 of elements of such type in +__e__+, not the number of bits.
6984 Array element C type.
6990 Argument expression.
6996 +ctf_array_text(__t__, __n__, __e__, __s__)+
6998 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7000 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7002 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7004 Statically-sized array, printed as text.
7006 The string does not need to be null-terminated.
7009 Array element C type (always `char`).
7015 Argument expression.
7021 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7023 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7025 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7027 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7029 Dynamically-sized array of integers.
7031 The type of +__E__+ must be unsigned.
7034 Array element C type.
7040 Argument expression.
7043 Length expression C type.
7049 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7051 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7053 Dynamically-sized array of integers, displayed in base 16.
7055 The type of +__E__+ must be unsigned.
7058 Array element C type.
7064 Argument expression.
7067 Length expression C type.
7072 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7074 Dynamically-sized array of integers in network byte order (big-endian),
7075 displayed in base 10.
7077 The type of +__E__+ must be unsigned.
7080 Array element C type.
7086 Argument expression.
7089 Length expression C type.
7095 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7097 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7099 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7101 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7103 Dynamically-sized array of bits.
7105 The type of +__e__+ must be an integer type. +__s__+ is the number
7106 of elements of such type in +__e__+, not the number of bits.
7108 The type of +__E__+ must be unsigned.
7111 Array element C type.
7117 Argument expression.
7120 Length expression C type.
7126 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7128 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7130 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7132 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7134 Dynamically-sized array, displayed as text.
7136 The string does not need to be null-terminated.
7138 The type of +__E__+ must be unsigned.
7140 The behaviour is undefined if +__e__+ is `NULL`.
7143 Sequence element C type (always `char`).
7149 Argument expression.
7152 Length expression C type.
7158 Use the `_user` versions when the argument expression, `e`, is
7159 a user space address. In the cases of `ctf_user_integer*()` and
7160 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7163 The `_nowrite` versions omit themselves from the session trace, but are
7164 otherwise identical. This means the `_nowrite` fields won't be written
7165 in the recorded trace. Their primary purpose is to make some
7166 of the event context available to the
7167 <<enabling-disabling-events,event filters>> without having to
7168 commit the data to sub-buffers.
7174 Terms related to LTTng and to tracing in general:
7177 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7178 the cmd:babeltrace command, some libraries, and Python bindings.
7180 <<channel-buffering-schemes,buffering scheme>>::
7181 A layout of sub-buffers applied to a given channel.
7183 <<channel,channel>>::
7184 An entity which is responsible for a set of ring buffers.
7186 <<event,Event rules>> are always attached to a specific channel.
7189 A reference of time for a tracer.
7191 <<lttng-consumerd,consumer daemon>>::
7192 A process which is responsible for consuming the full sub-buffers
7193 and write them to a file system or send them over the network.
7195 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7196 mode in which the tracer _discards_ new event records when there's no
7197 sub-buffer space left to store them.
7200 The consequence of the execution of an instrumentation
7201 point, like a tracepoint that you manually place in some source code,
7202 or a Linux kernel KProbe.
7204 An event is said to _occur_ at a specific time. Different actions can
7205 be taken upon the occurance of an event, like record the event's payload
7208 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7209 The mechanism by which event records of a given channel are lost
7210 (not recorded) when there is no sub-buffer space left to store them.
7212 [[def-event-name]]event name::
7213 The name of an event, which is also the name of the event record.
7214 This is also called the _instrumentation point name_.
7217 A record, in a trace, of the payload of an event which occured.
7219 <<event,event rule>>::
7220 Set of conditions which must be satisfied for one or more occuring
7221 events to be recorded.
7223 `java.util.logging`::
7225 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7227 <<instrumenting,instrumentation>>::
7228 The use of LTTng probes to make a piece of software traceable.
7230 instrumentation point::
7231 A point in the execution path of a piece of software that, when
7232 reached by this execution, can emit an event.
7234 instrumentation point name::
7235 See _<<def-event-name,event name>>_.
7238 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7239 developed by the Apache Software Foundation.
7242 Level of severity of a log statement or user space
7243 instrumentation point.
7246 The _Linux Trace Toolkit: next generation_ project.
7248 <<lttng-cli,cmd:lttng>>::
7249 A command-line tool provided by the LTTng-tools project which you
7250 can use to send and receive control messages to and from a
7254 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7255 which is a set of analyzing programs that are used to obtain a
7256 higher level view of an LTTng trace.
7258 cmd:lttng-consumerd::
7259 The name of the consumer daemon program.
7262 A utility provided by the LTTng-tools project which can convert
7263 ring buffer files (usually
7264 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7267 LTTng Documentation::
7270 <<lttng-live,LTTng live>>::
7271 A communication protocol between the relay daemon and live viewers
7272 which makes it possible to see events "live", as they are received by
7275 <<lttng-modules,LTTng-modules>>::
7276 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7277 which contains the Linux kernel modules to make the Linux kernel
7278 instrumentation points available for LTTng tracing.
7281 The name of the relay daemon program.
7283 cmd:lttng-sessiond::
7284 The name of the session daemon program.
7287 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7288 contains the various programs and libraries used to
7289 <<controlling-tracing,control tracing>>.
7291 <<lttng-ust,LTTng-UST>>::
7292 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7293 contains libraries to instrument user applications.
7295 <<lttng-ust-agents,LTTng-UST Java agent>>::
7296 A Java package provided by the LTTng-UST project to allow the
7297 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7300 <<lttng-ust-agents,LTTng-UST Python agent>>::
7301 A Python package provided by the LTTng-UST project to allow the
7302 LTTng instrumentation of Python logging statements.
7304 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7305 The event loss mode in which new event records overwrite older
7306 event records when there's no sub-buffer space left to store them.
7308 <<channel-buffering-schemes,per-process buffering>>::
7309 A buffering scheme in which each instrumented process has its own
7310 sub-buffers for a given user space channel.
7312 <<channel-buffering-schemes,per-user buffering>>::
7313 A buffering scheme in which all the processes of a Unix user share the
7314 same sub-buffer for a given user space channel.
7316 <<lttng-relayd,relay daemon>>::
7317 A process which is responsible for receiving the trace data sent by
7318 a distant consumer daemon.
7321 A set of sub-buffers.
7323 <<lttng-sessiond,session daemon>>::
7324 A process which receives control commands from you and orchestrates
7325 the tracers and various LTTng daemons.
7327 <<taking-a-snapshot,snapshot>>::
7328 A copy of the current data of all the sub-buffers of a given tracing
7329 session, saved as trace files.
7332 One part of an LTTng ring buffer which contains event records.
7335 The time information attached to an event when it is emitted.
7338 A set of files which are the concatenations of one or more
7339 flushed sub-buffers.
7342 The action of recording the events emitted by an application
7343 or by a system, or to initiate such recording by controlling
7347 The http://tracecompass.org[Trace Compass] project and application.
7350 An instrumentation point using the tracepoint mechanism of the Linux
7351 kernel or of LTTng-UST.
7353 tracepoint definition::
7354 The definition of a single tracepoint.
7357 The name of a tracepoint.
7359 tracepoint provider::
7360 A set of functions providing tracepoints to an instrumented user
7363 Not to be confused with a _tracepoint provider package_: many tracepoint
7364 providers can exist within a tracepoint provider package.
7366 tracepoint provider package::
7367 One or more tracepoint providers compiled as an object file or as
7371 A software which records emitted events.
7373 <<domain,tracing domain>>::
7374 A namespace for event sources.
7376 <<tracing-group,tracing group>>::
7377 The Unix group in which a Unix user can be to be allowed to trace the
7380 <<tracing-session,tracing session>>::
7381 A stateful dialogue between you and a <<lttng-sessiond,session
7385 An application running in user space, as opposed to a Linux kernel
7386 module, for example.