1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng {revision}?
77 * **Tracing control**:
78 ** You can override the name or the URL of a tracing session
79 configuration when you use man:lttng-load(1) thanks to the new
80 opt:lttng-load(1):--override-name and
81 opt:lttng-load(1):--override-url options.
82 ** The new `lttng regenerate` command replaces the now deprecated
83 `lttng metadata` command of LTTng 2.8. man:lttng-regenerate(1) can
84 also <<regenerate-statedump,generate the state dump event records>>
85 of a given tracing session on demand, a handy feature when
86 <<taking-a-snapshot,taking a snapshot>>.
87 ** You can add PMU counters by raw ID with man:lttng-add-context(1):
92 lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
96 The format of the raw ID is the same as used with man:perf-record(1).
97 See <<adding-context,Add context fields to a channel>> for more
100 ** The LTTng <<lttng-relayd,relay daemon>> is now supported on
101 OS{nbsp}X and macOS for a smoother integration within a trace
102 analysis workflow, regardless of the platform used.
104 * **User space tracing**:
105 ** Improved performance (tested on x86-64 and ARMv7-A
106 (https://en.wikipedia.org/wiki/Cubieboard[Cubieboard])
108 ** New helper library (`liblttng-ust-fd`) to help with
109 <<liblttng-ust-fd,applications which close file descriptors that
110 don't belong to them>>, for example, in a loop which closes file
111 descriptors after man:fork(2), or BSD's `closeall()`.
112 ** More accurate <<liblttng-ust-dl,dynamic linker instrumentation>> and
113 state dump event records, especially when a dynamically loaded
114 library manually loads its own dependencies.
115 ** New `ctf_*()` field definition macros (see man:lttng-ust(3)):
116 *** `ctf_array_hex()`
117 *** `ctf_array_network()`
118 *** `ctf_array_network_hex()`
119 *** `ctf_sequence_hex()`
120 *** `ctf_sequence_network()`
121 *** `ctf_sequence_network_hex()`
122 ** New `lttng_ust_loaded` weak symbol defined by `liblttng-ust` for
123 an application to know if the LTTng-UST shared library is loaded
131 int lttng_ust_loaded __attribute__((weak));
135 if (lttng_ust_loaded) {
136 puts("LTTng-UST is loaded!");
138 puts("LTTng-UST is not loaded!");
146 ** LTTng-UST thread names have the `-ust` suffix.
148 * **Linux kernel tracing**:
149 ** Improved performance (tested on x86-64 and ARMv7-A
150 (https://en.wikipedia.org/wiki/Cubieboard[Cubieboard])
152 ** New enumeration <<lttng-modules-tp-fields,field definition macros>>:
153 `ctf_enum()` and `ctf_user_enum()`.
154 ** IPv4, IPv6, and TCP header data is recorded in the event records
155 produced by tracepoints starting with `net_`.
156 ** Detailed system call event records: `select`, `pselect6`, `poll`,
157 `ppoll`, `epoll_wait`, `epoll_pwait`, and `epoll_ctl` on all
158 architectures supported by LTTng-modules, and `accept4` on x86-64.
159 ** New I²C instrumentation: the `extract_sensitive_payload` parameter
160 of the new `lttng-probe-i2c` LTTng module controls whether or not
161 the payloads of I²C messages are recorded in I²C event records, since
162 they may contain sensitive data (for example, keystrokes).
163 ** When the LTTng kernel modules are built into the Linux kernel image,
164 the `CONFIG_TRACEPOINTS` configuration option is automatically
171 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
172 generation_ is a modern toolkit for tracing Linux systems and
173 applications. So your first question might be:
180 As the history of software engineering progressed and led to what
181 we now take for granted--complex, numerous and
182 interdependent software applications running in parallel on
183 sophisticated operating systems like Linux--the authors of such
184 components, software developers, began feeling a natural
185 urge to have tools that would ensure the robustness and good performance
186 of their masterpieces.
188 One major achievement in this field is, inarguably, the
189 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
190 an essential tool for developers to find and fix bugs. But even the best
191 debugger won't help make your software run faster, and nowadays, faster
192 software means either more work done by the same hardware, or cheaper
193 hardware for the same work.
195 A _profiler_ is often the tool of choice to identify performance
196 bottlenecks. Profiling is suitable to identify _where_ performance is
197 lost in a given software. The profiler outputs a profile, a statistical
198 summary of observed events, which you may use to discover which
199 functions took the most time to execute. However, a profiler won't
200 report _why_ some identified functions are the bottleneck. Bottlenecks
201 might only occur when specific conditions are met, conditions that are
202 sometimes impossible to capture by a statistical profiler, or impossible
203 to reproduce with an application altered by the overhead of an
204 event-based profiler. For a thorough investigation of software
205 performance issues, a history of execution is essential, with the
206 recorded values of variables and context fields you choose, and
207 with as little influence as possible on the instrumented software. This
208 is where tracing comes in handy.
210 _Tracing_ is a technique used to understand what goes on in a running
211 software system. The software used for tracing is called a _tracer_,
212 which is conceptually similar to a tape recorder. When recording,
213 specific instrumentation points placed in the software source code
214 generate events that are saved on a giant tape: a _trace_ file. You
215 can trace user applications and the operating system at the same time,
216 opening the possibility of resolving a wide range of problems that would
217 otherwise be extremely challenging.
219 Tracing is often compared to _logging_. However, tracers and loggers are
220 two different tools, serving two different purposes. Tracers are
221 designed to record much lower-level events that occur much more
222 frequently than log messages, often in the range of thousands per
223 second, with very little execution overhead. Logging is more appropriate
224 for a very high-level analysis of less frequent events: user accesses,
225 exceptional conditions (errors and warnings, for example), database
226 transactions, instant messaging communications, and such. Simply put,
227 logging is one of the many use cases that can be satisfied with tracing.
229 The list of recorded events inside a trace file can be read manually
230 like a log file for the maximum level of detail, but it is generally
231 much more interesting to perform application-specific analyses to
232 produce reduced statistics and graphs that are useful to resolve a
233 given problem. Trace viewers and analyzers are specialized tools
236 In the end, this is what LTTng is: a powerful, open source set of
237 tools to trace the Linux kernel and user applications at the same time.
238 LTTng is composed of several components actively maintained and
239 developed by its link:/community/#where[community].
242 [[lttng-alternatives]]
243 === Alternatives to noch:{LTTng}
245 Excluding proprietary solutions, a few competing software tracers
248 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
249 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
250 user scripts and is responsible for loading code into the
251 Linux kernel for further execution and collecting the outputted data.
252 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
253 subsystem in the Linux kernel in which a virtual machine can execute
254 programs passed from the user space to the kernel. You can attach
255 such programs to tracepoints and KProbes thanks to a system call, and
256 they can output data to the user space when executed thanks to
257 different mechanisms (pipe, VM register values, and eBPF maps, to name
259 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
260 is the de facto function tracer of the Linux kernel. Its user
261 interface is a set of special files in sysfs.
262 * https://perf.wiki.kernel.org/[perf] is
263 a performance analyzing tool for Linux which supports hardware
264 performance counters, tracepoints, as well as other counters and
265 types of probes. perf's controlling utility is the cmd:perf command
267 * http://linux.die.net/man/1/strace[strace]
268 is a command-line utility which records system calls made by a
269 user process, as well as signal deliveries and changes of process
270 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
271 to fulfill its function.
272 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
273 analyze Linux kernel events. You write scripts, or _chisels_ in
274 sysdig's jargon, in Lua and sysdig executes them while the system is
275 being traced or afterwards. sysdig's interface is the cmd:sysdig
276 command-line tool as well as the curses-based cmd:csysdig tool.
277 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
278 user space tracer which uses custom user scripts to produce plain text
279 traces. SystemTap converts the scripts to the C language, and then
280 compiles them as Linux kernel modules which are loaded to produce
281 trace data. SystemTap's primary user interface is the cmd:stap
284 The main distinctive features of LTTng is that it produces correlated
285 kernel and user space traces, as well as doing so with the lowest
286 overhead amongst other solutions. It produces trace files in the
287 http://diamon.org/ctf[CTF] format, a file format optimized
288 for the production and analyses of multi-gigabyte data.
290 LTTng is the result of more than 10 years of active open source
291 development by a community of passionate developers.
292 LTTng{nbsp}{revision} is currently available on major desktop and server
295 The main interface for tracing control is a single command-line tool
296 named cmd:lttng. The latter can create several tracing sessions, enable
297 and disable events on the fly, filter events efficiently with custom
298 user expressions, start and stop tracing, and much more. LTTng can
299 record the traces on the file system or send them over the network, and
300 keep them totally or partially. You can view the traces once tracing
301 becomes inactive or in real-time.
303 <<installing-lttng,Install LTTng now>> and
304 <<getting-started,start tracing>>!
310 **LTTng** is a set of software <<plumbing,components>> which interact to
311 <<instrumenting,instrument>> the Linux kernel and user applications, and
312 to <<controlling-tracing,control tracing>> (start and stop
313 tracing, enable and disable event rules, and the rest). Those
314 components are bundled into the following packages:
316 * **LTTng-tools**: Libraries and command-line interface to
318 * **LTTng-modules**: Linux kernel modules to instrument and
320 * **LTTng-UST**: Libraries and Java/Python packages to instrument and
321 trace user applications.
323 Most distributions mark the LTTng-modules and LTTng-UST packages as
324 optional when installing LTTng-tools (which is always required). In the
325 following sections, we always provide the steps to install all three,
328 * You only need to install LTTng-modules if you intend to trace the
330 * You only need to install LTTng-UST if you intend to trace user
334 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 2 December 2016.
336 |Distribution |Available in releases |Alternatives
338 |https://www.ubuntu.com/[Ubuntu]
339 |Ubuntu{nbsp}14.04 _Trusty Tahr_ and Ubuntu{nbsp}16.04 _Xenial Xerus_:
340 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
341 |link:/docs/v2.8#doc-ubuntu[LTTng{nbsp}2.8 for Ubuntu{nbsp}16.10 _Yakkety Yak_].
343 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
344 other Ubuntu releases.
346 |https://getfedora.org/[Fedora]
348 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
351 |https://www.debian.org/[Debian]
353 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
356 |https://www.opensuse.org/[openSUSE]
358 |<<building-from-source,Build LTTng-modules{nbsp}{revision} from
361 |https://www.archlinux.org/[Arch Linux]
362 |<<arch-linux,Latest AUR packages>>.
365 |https://alpinelinux.org/[Alpine Linux]
367 |link:/docs/v2.8#doc-alpine-linux[LTTng{nbsp}2.8 for Alpine Linux{nbsp}"edge"].
369 LTTng{nbsp}2.8 for Alpine Linux{nbsp}3.5 (not released yet).
371 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
372 other Alpine Linux releases.
374 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
375 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
378 |https://buildroot.org/[Buildroot]
380 |link:/docs/v2.8#doc-buildroot[LTTng{nbsp}2.8 for Buildroot{nbsp}2016.11].
382 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
383 other Buildroot releases.
385 |http://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
386 https://www.yoctoproject.org/[Yocto]
388 |link:/docs/v2.8#doc-oe-yocto[LTTng{nbsp}2.8 for Yocto Project{nbsp}2.2 _Morty_]
389 (`openembedded-core` layer).
391 <<building-from-source,Build LTTng{nbsp}{revision} from source>> for
392 other OpenEmbedded releases.
397 === [[ubuntu-official-repositories]]Ubuntu
400 ==== noch:{LTTng} Stable {revision} PPA
402 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
403 Stable{nbsp}{revision} PPA] offers the latest stable
404 LTTng{nbsp}{revision} packages for:
406 * Ubuntu{nbsp}14.04 _Trusty Tahr_
407 * Ubuntu{nbsp}16.04 _Xenial Xerus_
409 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision} PPA:
411 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
417 sudo apt-add-repository ppa:lttng/stable-2.9
422 . Install the main LTTng{nbsp}{revision} packages:
427 sudo apt-get install lttng-tools
428 sudo apt-get install lttng-modules-dkms
429 sudo apt-get install liblttng-ust-dev
433 . **If you need to instrument and trace
434 <<java-application,Java applications>>**, install the LTTng-UST
440 sudo apt-get install liblttng-ust-agent-java
444 . **If you need to instrument and trace
445 <<python-application,Python{nbsp}3 applications>>**, install the
446 LTTng-UST Python agent:
451 sudo apt-get install python3-lttngust
459 To install LTTng{nbsp}{revision} on Arch Linux using
460 https://archlinux.fr/yaourt-en[Yaourt]:
462 . Install the main LTTng{nbsp}{revision} packages:
467 yaourt -S lttng-tools
469 yaourt -S lttng-modules
473 . **If you need to instrument and trace <<python-application,Python
474 applications>>**, install the LTTng-UST Python agent:
479 yaourt -S python-lttngust
480 yaourt -S python2-lttngust
485 [[enterprise-distributions]]
486 === RHEL, SUSE, and other enterprise distributions
488 To install LTTng on enterprise Linux distributions, such as Red Hat
489 Enterprise Linux (RHEL) and SUSE Linux Enterprise Server (SUSE), please
490 see http://packages.efficios.com/[EfficiOS Enterprise Packages].
493 [[building-from-source]]
494 === Build from source
496 To build and install LTTng{nbsp}{revision} from source:
498 . Using your distribution's package manager, or from source, install
499 the following dependencies of LTTng-tools and LTTng-UST:
502 * https://sourceforge.net/projects/libuuid/[libuuid]
503 * http://directory.fsf.org/wiki/Popt[popt]
504 * http://liburcu.org/[Userspace RCU]
505 * http://www.xmlsoft.org/[libxml2]
508 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
514 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.9.tar.bz2 &&
515 tar -xf lttng-modules-latest-2.9.tar.bz2 &&
516 cd lttng-modules-2.9.* &&
518 sudo make modules_install &&
523 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
529 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.9.tar.bz2 &&
530 tar -xf lttng-ust-latest-2.9.tar.bz2 &&
531 cd lttng-ust-2.9.* &&
541 .Java and Python application tracing
543 If you need to instrument and trace <<java-application,Java
544 applications>>, pass the `--enable-java-agent-jul`,
545 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
546 `configure` script, depending on which Java logging framework you use.
548 If you need to instrument and trace <<python-application,Python
549 applications>>, pass the `--enable-python-agent` option to the
550 `configure` script. You can set the `PYTHON` environment variable to the
551 path to the Python interpreter for which to install the LTTng-UST Python
559 By default, LTTng-UST libraries are installed to
560 dir:{/usr/local/lib}, which is the de facto directory in which to
561 keep self-compiled and third-party libraries.
563 When <<building-tracepoint-providers-and-user-application,linking an
564 instrumented user application with `liblttng-ust`>>:
566 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
568 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
569 man:gcc(1), man:g++(1), or man:clang(1).
573 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
579 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
580 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
581 cd lttng-tools-2.9.* &&
589 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
590 previous steps automatically for a given version of LTTng and confine
591 the installed files in a specific directory. This can be useful to test
592 LTTng without installing it on your system.
598 This is a short guide to get started quickly with LTTng kernel and user
601 Before you follow this guide, make sure to <<installing-lttng,install>>
604 This tutorial walks you through the steps to:
606 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
607 . <<tracing-your-own-user-application,Trace a user application>> written
609 . <<viewing-and-analyzing-your-traces,View and analyze the
613 [[tracing-the-linux-kernel]]
614 === Trace the Linux kernel
616 The following command lines start with cmd:sudo because you need root
617 privileges to trace the Linux kernel. You can avoid using cmd:sudo if
618 your Unix user is a member of the <<tracing-group,tracing group>>.
620 . Create a <<tracing-session,tracing session>> which writes its traces
621 to dir:{/tmp/my-kernel-trace}:
626 sudo lttng create my-kernel-session --output=/tmp/my-kernel-trace
630 . List the available kernel tracepoints and system calls:
636 lttng list --kernel --syscall
640 . Create <<event,event rules>> which match the desired instrumentation
641 point names, for example the `sched_switch` and `sched_process_fork`
642 tracepoints, and the man:open(2) and man:close(2) system calls:
647 sudo lttng enable-event --kernel sched_switch,sched_process_fork
648 sudo lttng enable-event --kernel --syscall open,close
652 You can also create an event rule which matches _all_ the Linux kernel
653 tracepoints (this will generate a lot of data when tracing):
658 sudo lttng enable-event --kernel --all
662 . <<basic-tracing-session-control,Start tracing>>:
671 . Do some operation on your system for a few seconds. For example,
672 load a website, or list the files of a directory.
673 . <<basic-tracing-session-control,Stop tracing>> and destroy the
684 The man:lttng-destroy(1) command does not destroy the trace data; it
685 only destroys the state of the tracing session.
687 . For the sake of this example, make the recorded trace accessible to
693 sudo chown -R $(whoami) /tmp/my-kernel-trace
697 See <<viewing-and-analyzing-your-traces,View and analyze the
698 recorded events>> to view the recorded events.
701 [[tracing-your-own-user-application]]
702 === Trace a user application
704 This section steps you through a simple example to trace a
705 _Hello world_ program written in C.
707 To create the traceable user application:
709 . Create the tracepoint provider header file, which defines the
710 tracepoints and the events they can generate:
716 #undef TRACEPOINT_PROVIDER
717 #define TRACEPOINT_PROVIDER hello_world
719 #undef TRACEPOINT_INCLUDE
720 #define TRACEPOINT_INCLUDE "./hello-tp.h"
722 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
725 #include <lttng/tracepoint.h>
735 ctf_string(my_string_field, my_string_arg)
736 ctf_integer(int, my_integer_field, my_integer_arg)
740 #endif /* _HELLO_TP_H */
742 #include <lttng/tracepoint-event.h>
746 . Create the tracepoint provider package source file:
752 #define TRACEPOINT_CREATE_PROBES
753 #define TRACEPOINT_DEFINE
755 #include "hello-tp.h"
759 . Build the tracepoint provider package:
764 gcc -c -I. hello-tp.c
768 . Create the _Hello World_ application source file:
775 #include "hello-tp.h"
777 int main(int argc, char *argv[])
781 puts("Hello, World!\nPress Enter to continue...");
784 * The following getchar() call is only placed here for the purpose
785 * of this demonstration, to pause the application in order for
786 * you to have time to list its tracepoints. It is not
792 * A tracepoint() call.
794 * Arguments, as defined in hello-tp.h:
796 * 1. Tracepoint provider name (required)
797 * 2. Tracepoint name (required)
798 * 3. my_integer_arg (first user-defined argument)
799 * 4. my_string_arg (second user-defined argument)
801 * Notice the tracepoint provider and tracepoint names are
802 * NOT strings: they are in fact parts of variables that the
803 * macros in hello-tp.h create.
805 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
807 for (x = 0; x < argc; ++x) {
808 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
811 puts("Quitting now!");
812 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
819 . Build the application:
828 . Link the application with the tracepoint provider package,
829 `liblttng-ust`, and `libdl`:
834 gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
838 Here's the whole build process:
841 .User space tracing tutorial's build steps.
842 image::ust-flow.png[]
844 To trace the user application:
846 . Run the application with a few arguments:
851 ./hello world and beyond
860 Press Enter to continue...
864 . Start an LTTng <<lttng-sessiond,session daemon>>:
869 lttng-sessiond --daemonize
873 Note that a session daemon might already be running, for example as
874 a service that the distribution's service manager started.
876 . List the available user space tracepoints:
881 lttng list --userspace
885 You see the `hello_world:my_first_tracepoint` tracepoint listed
886 under the `./hello` process.
888 . Create a <<tracing-session,tracing session>>:
893 lttng create my-user-space-session
897 . Create an <<event,event rule>> which matches the
898 `hello_world:my_first_tracepoint` event name:
903 lttng enable-event --userspace hello_world:my_first_tracepoint
907 . <<basic-tracing-session-control,Start tracing>>:
916 . Go back to the running `hello` application and press Enter. The
917 program executes all `tracepoint()` instrumentation points and exits.
918 . <<basic-tracing-session-control,Stop tracing>> and destroy the
929 The man:lttng-destroy(1) command does not destroy the trace data; it
930 only destroys the state of the tracing session.
932 By default, LTTng saves the traces in
933 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
934 where +__name__+ is the tracing session name. The
935 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
937 See <<viewing-and-analyzing-your-traces,View and analyze the
938 recorded events>> to view the recorded events.
941 [[viewing-and-analyzing-your-traces]]
942 === View and analyze the recorded events
944 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
945 kernel>> and <<tracing-your-own-user-application,Trace a user
946 application>> tutorials, you can inspect the recorded events.
948 Many tools are available to read LTTng traces:
950 * **cmd:babeltrace** is a command-line utility which converts trace
951 formats; it supports the format that LTTng produces, CTF, as well as a
952 basic text output which can be ++grep++ed. The cmd:babeltrace command
953 is part of the http://diamon.org/babeltrace[Babeltrace] project.
954 * Babeltrace also includes
955 **https://www.python.org/[Python] bindings** so
956 that you can easily open and read an LTTng trace with your own script,
957 benefiting from the power of Python.
958 * http://tracecompass.org/[**Trace Compass**]
959 is a graphical user interface for viewing and analyzing any type of
960 logs or traces, including LTTng's.
961 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
962 project which includes many high-level analyses of LTTng kernel
963 traces, like scheduling statistics, interrupt frequency distribution,
964 top CPU usage, and more.
966 NOTE: This section assumes that the traces recorded during the previous
967 tutorials were saved to their default location, in the
968 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
969 environment variable defaults to `$HOME` if not set.
972 [[viewing-and-analyzing-your-traces-bt]]
973 ==== Use the cmd:babeltrace command-line tool
975 The simplest way to list all the recorded events of a trace is to pass
976 its path to cmd:babeltrace with no options:
980 babeltrace ~/lttng-traces/my-user-space-session*
983 cmd:babeltrace finds all traces recursively within the given path and
984 prints all their events, merging them in chronological order.
986 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
991 babeltrace /tmp/my-kernel-trace | grep _switch
994 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
995 count the recorded events:
999 babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1003 [[viewing-and-analyzing-your-traces-bt-python]]
1004 ==== Use the Babeltrace Python bindings
1006 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1007 is useful to isolate events by simple matching using man:grep(1) and
1008 similar utilities. However, more elaborate filters, such as keeping only
1009 event records with a field value falling within a specific range, are
1010 not trivial to write using a shell. Moreover, reductions and even the
1011 most basic computations involving multiple event records are virtually
1012 impossible to implement.
1014 Fortunately, Babeltrace ships with Python 3 bindings which makes it easy
1015 to read the event records of an LTTng trace sequentially and compute the
1016 desired information.
1018 The following script accepts an LTTng Linux kernel trace path as its
1019 first argument and prints the short names of the top 5 running processes
1020 on CPU 0 during the whole trace:
1025 from collections import Counter
1031 if len(sys.argv) != 2:
1032 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1033 print(msg, file=sys.stderr)
1036 # A trace collection contains one or more traces
1037 col = babeltrace.TraceCollection()
1039 # Add the trace provided by the user (LTTng traces always have
1041 if col.add_trace(sys.argv[1], 'ctf') is None:
1042 raise RuntimeError('Cannot add trace')
1044 # This counter dict contains execution times:
1046 # task command name -> total execution time (ns)
1047 exec_times = Counter()
1049 # This contains the last `sched_switch` timestamp
1053 for event in col.events:
1054 # Keep only `sched_switch` events
1055 if event.name != 'sched_switch':
1058 # Keep only events which happened on CPU 0
1059 if event['cpu_id'] != 0:
1063 cur_ts = event.timestamp
1069 # Previous task command (short) name
1070 prev_comm = event['prev_comm']
1072 # Initialize entry in our dict if not yet done
1073 if prev_comm not in exec_times:
1074 exec_times[prev_comm] = 0
1076 # Compute previous command execution time
1077 diff = cur_ts - last_ts
1079 # Update execution time of this command
1080 exec_times[prev_comm] += diff
1082 # Update last timestamp
1086 for name, ns in exec_times.most_common(5):
1088 print('{:20}{} s'.format(name, s))
1093 if __name__ == '__main__':
1094 sys.exit(0 if top5proc() else 1)
1101 python3 top5proc.py /tmp/my-kernel-trace/kernel
1107 swapper/0 48.607245889 s
1108 chromium 7.192738188 s
1109 pavucontrol 0.709894415 s
1110 Compositor 0.660867933 s
1111 Xorg.bin 0.616753786 s
1114 Note that `swapper/0` is the "idle" process of CPU 0 on Linux; since we
1115 weren't using the CPU that much when tracing, its first position in the
1120 == [[understanding-lttng]]Core concepts
1122 From a user's perspective, the LTTng system is built on a few concepts,
1123 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1124 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1125 Understanding how those objects relate to eachother is key in mastering
1128 The core concepts are:
1130 * <<tracing-session,Tracing session>>
1131 * <<domain,Tracing domain>>
1132 * <<channel,Channel and ring buffer>>
1133 * <<"event","Instrumentation point, event rule, event, and event record">>
1139 A _tracing session_ is a stateful dialogue between you and
1140 a <<lttng-sessiond,session daemon>>. You can
1141 <<creating-destroying-tracing-sessions,create a new tracing
1142 session>> with the `lttng create` command.
1144 Anything that you do when you control LTTng tracers happens within a
1145 tracing session. In particular, a tracing session:
1148 * Has its own set of trace files.
1149 * Has its own state of activity (started or stopped).
1150 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1152 * Has its own <<channel,channels>> which have their own
1153 <<event,event rules>>.
1156 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1157 image::concepts.png[]
1159 Those attributes and objects are completely isolated between different
1162 A tracing session is analogous to a cash machine session:
1163 the operations you do on the banking system through the cash machine do
1164 not alter the data of other users of the same system. In the case of
1165 the cash machine, a session lasts as long as your bank card is inside.
1166 In the case of LTTng, a tracing session lasts from the `lttng create`
1167 command to the `lttng destroy` command.
1170 .Each Unix user has its own set of tracing sessions.
1171 image::many-sessions.png[]
1174 [[tracing-session-mode]]
1175 ==== Tracing session mode
1177 LTTng can send the generated trace data to different locations. The
1178 _tracing session mode_ dictates where to send it. The following modes
1179 are available in LTTng{nbsp}{revision}:
1182 LTTng writes the traces to the file system of the machine being traced
1185 Network streaming mode::
1186 LTTng sends the traces over the network to a
1187 <<lttng-relayd,relay daemon>> running on a remote system.
1190 LTTng does not write the traces by default. Instead, you can request
1191 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1192 current tracing buffers, and to write it to the target's file system
1193 or to send it over the network to a <<lttng-relayd,relay daemon>>
1194 running on a remote system.
1197 This mode is similar to the network streaming mode, but a live
1198 trace viewer can connect to the distant relay daemon to
1199 <<lttng-live,view event records as LTTng generates them>> by
1206 A _tracing domain_ is a namespace for event sources. A tracing domain
1207 has its own properties and features.
1209 There are currently five available tracing domains:
1213 * `java.util.logging` (JUL)
1217 You must specify a tracing domain when using some commands to avoid
1218 ambiguity. For example, since all the domains support named tracepoints
1219 as event sources (instrumentation points that you manually insert in the
1220 source code), you need to specify a tracing domain when
1221 <<enabling-disabling-events,creating an event rule>> because all the
1222 tracing domains could have tracepoints with the same names.
1224 Some features are reserved to specific tracing domains. Dynamic function
1225 entry and return instrumentation points, for example, are currently only
1226 supported in the Linux kernel tracing domain, but support for other
1227 tracing domains could be added in the future.
1229 You can create <<channel,channels>> in the Linux kernel and user space
1230 tracing domains. The other tracing domains have a single default
1235 === Channel and ring buffer
1237 A _channel_ is an object which is responsible for a set of ring buffers.
1238 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1239 tracer emits an event, it can record it to one or more
1240 sub-buffers. The attributes of a channel determine what to do when
1241 there's no space left for a new event record because all sub-buffers
1242 are full, where to send a full sub-buffer, and other behaviours.
1244 A channel is always associated to a <<domain,tracing domain>>. The
1245 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1246 a default channel which you cannot configure.
1248 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1249 an event, it records it to the sub-buffers of all
1250 the enabled channels with a satisfied event rule, as long as those
1251 channels are part of active <<tracing-session,tracing sessions>>.
1254 [[channel-buffering-schemes]]
1255 ==== Per-user vs. per-process buffering schemes
1257 A channel has at least one ring buffer _per CPU_. LTTng always
1258 records an event to the ring buffer associated to the CPU on which it
1261 Two _buffering schemes_ are available when you
1262 <<enabling-disabling-channels,create a channel>> in the
1263 user space <<domain,tracing domain>>:
1265 Per-user buffering::
1266 Allocate one set of ring buffers--one per CPU--shared by all the
1267 instrumented processes of each Unix user.
1271 .Per-user buffering scheme.
1272 image::per-user-buffering.png[]
1275 Per-process buffering::
1276 Allocate one set of ring buffers--one per CPU--for each
1277 instrumented process.
1281 .Per-process buffering scheme.
1282 image::per-process-buffering.png[]
1285 The per-process buffering scheme tends to consume more memory than the
1286 per-user option because systems generally have more instrumented
1287 processes than Unix users running instrumented processes. However, the
1288 per-process buffering scheme ensures that one process having a high
1289 event throughput won't fill all the shared sub-buffers of the same
1292 The Linux kernel tracing domain has only one available buffering scheme
1293 which is to allocate a single set of ring buffers for the whole system.
1294 This scheme is similar to the per-user option, but with a single, global
1295 user "running" the kernel.
1298 [[channel-overwrite-mode-vs-discard-mode]]
1299 ==== Overwrite vs. discard event loss modes
1301 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1302 arc in the following animation) of a specific channel's ring buffer.
1303 When there's no space left in a sub-buffer, the tracer marks it as
1304 consumable (red) and another, empty sub-buffer starts receiving the
1305 following event records. A <<lttng-consumerd,consumer daemon>>
1306 eventually consumes the marked sub-buffer (returns to white).
1309 [role="docsvg-channel-subbuf-anim"]
1314 In an ideal world, sub-buffers are consumed faster than they are filled,
1315 as is the case in the previous animation. In the real world,
1316 however, all sub-buffers can be full at some point, leaving no space to
1317 record the following events.
1319 By design, LTTng is a _non-blocking_ tracer: when no empty sub-buffer is
1320 available, it is acceptable to lose event records when the alternative
1321 would be to cause substantial delays in the instrumented application's
1322 execution. LTTng privileges performance over integrity; it aims at
1323 perturbing the traced system as little as possible in order to make
1324 tracing of subtle race conditions and rare interrupt cascades possible.
1326 When it comes to losing event records because no empty sub-buffer is
1327 available, the channel's _event loss mode_ determines what to do. The
1328 available event loss modes are:
1331 Drop the newest event records until a the tracer
1332 releases a sub-buffer.
1335 Clear the sub-buffer containing the oldest event records and start
1336 writing the newest event records there.
1338 This mode is sometimes called _flight recorder mode_ because it's
1340 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1341 always keep a fixed amount of the latest data.
1343 Which mechanism you should choose depends on your context: prioritize
1344 the newest or the oldest event records in the ring buffer?
1346 Beware that, in overwrite mode, the tracer abandons a whole sub-buffer
1347 as soon as a there's no space left for a new event record, whereas in
1348 discard mode, the tracer only discards the event record that doesn't
1351 In discard mode, LTTng increments a count of lost event records when
1352 an event record is lost and saves this count to the trace. In
1353 overwrite mode, LTTng keeps no information when it overwrites a
1354 sub-buffer before consuming it.
1356 There are a few ways to decrease your probability of losing event
1358 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1359 how you can fine-une the sub-buffer count and size of a channel to
1360 virtually stop losing event records, though at the cost of greater
1364 [[channel-subbuf-size-vs-subbuf-count]]
1365 ==== Sub-buffer count and size
1367 When you <<enabling-disabling-channels,create a channel>>, you can
1368 set its number of sub-buffers and their size.
1370 Note that there is noticeable CPU overhead introduced when
1371 switching sub-buffers (marking a full one as consumable and switching
1372 to an empty one for the following events to be recorded). Knowing this,
1373 the following list presents a few practical situations along with how
1374 to configure the sub-buffer count and size for them:
1376 * **High event throughput**: In general, prefer bigger sub-buffers to
1377 lower the risk of losing event records.
1379 Having bigger sub-buffers also ensures a lower sub-buffer switching
1382 The number of sub-buffers is only meaningful if you create the channel
1383 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1384 other sub-buffers are left unaltered.
1386 * **Low event throughput**: In general, prefer smaller sub-buffers
1387 since the risk of losing event records is low.
1389 Because events occur less frequently, the sub-buffer switching frequency
1390 should remain low and thus the tracer's overhead should not be a
1393 * **Low memory system**: If your target system has a low memory
1394 limit, prefer fewer first, then smaller sub-buffers.
1396 Even if the system is limited in memory, you want to keep the
1397 sub-buffers as big as possible to avoid a high sub-buffer switching
1400 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1401 which means event data is very compact. For example, the average
1402 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1403 sub-buffer size of 1{nbsp}MiB is considered big.
1405 The previous situations highlight the major trade-off between a few big
1406 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1407 frequency vs. how much data is lost in overwrite mode. Assuming a
1408 constant event throughput and using the overwrite mode, the two
1409 following configurations have the same ring buffer total size:
1412 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1417 * **2 sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1418 switching frequency, but if a sub-buffer overwrite happens, half of
1419 the event records so far (4{nbsp}MiB) are definitely lost.
1420 * **8 sub-buffers of 1{nbsp}MiB each**: Expect 4{nbsp}times the tracer's
1421 overhead as the previous configuration, but if a sub-buffer
1422 overwrite happens, only the eighth of event records so far are
1425 In discard mode, the sub-buffers count parameter is pointless: use two
1426 sub-buffers and set their size according to the requirements of your
1430 [[channel-switch-timer]]
1431 ==== Switch timer period
1433 The _switch timer period_ is an important configurable attribute of
1434 a channel to ensure periodic sub-buffer flushing.
1436 When the _switch timer_ expires, a sub-buffer switch happens. You can
1437 set the switch timer period attribute when you
1438 <<enabling-disabling-channels,create a channel>> to ensure that event
1439 data is consumed and committed to trace files or to a distant relay
1440 daemon periodically in case of a low event throughput.
1443 [role="docsvg-channel-switch-timer"]
1448 This attribute is also convenient when you use big sub-buffers to cope
1449 with a sporadic high event throughput, even if the throughput is
1453 [[channel-read-timer]]
1454 ==== Read timer period
1456 By default, the LTTng tracers use a notification mechanism to signal a
1457 full sub-buffer so that a consumer daemon can consume it. When such
1458 notifications must be avoided, for example in real-time applications,
1459 you can use the channel's _read timer_ instead. When the read timer
1460 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1461 consumable sub-buffers.
1464 [[tracefile-rotation]]
1465 ==== Trace file count and size
1467 By default, trace files can grow as large as needed. You can set the
1468 maximum size of each trace file that a channel writes when you
1469 <<enabling-disabling-channels,create a channel>>. When the size of
1470 a trace file reaches the channel's fixed maximum size, LTTng creates
1471 another file to contain the next event records. LTTng appends a file
1472 count to each trace file name in this case.
1474 If you set the trace file size attribute when you create a channel, the
1475 maximum number of trace files that LTTng creates is _unlimited_ by
1476 default. To limit them, you can also set a maximum number of trace
1477 files. When the number of trace files reaches the channel's fixed
1478 maximum count, the oldest trace file is overwritten. This mechanism is
1479 called _trace file rotation_.
1483 === Instrumentation point, event rule, event, and event record
1485 An _event rule_ is a set of conditions which must be **all** satisfied
1486 for LTTng to record an occuring event.
1488 You set the conditions when you <<enabling-disabling-events,create
1491 You always attach an event rule to <<channel,channel>> when you create
1494 When an event passes the conditions of an event rule, LTTng records it
1495 in one of the attached channel's sub-buffers.
1497 The available conditions, as of LTTng{nbsp}{revision}, are:
1499 * The event rule _is enabled_.
1500 * The instrumentation point's type _is{nbsp}T_.
1501 * The instrumentation point's name (sometimes called _event name_)
1502 _matches{nbsp}N_, but _is not{nbsp}E_.
1503 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1504 _is exactly{nbsp}L_.
1505 * The fields of the event's payload _satisfy_ a filter
1506 expression{nbsp}__F__.
1508 As you can see, all the conditions but the dynamic filter are related to
1509 the event rule's status or to the instrumentation point, not to the
1510 occurring events. This is why, without a filter, checking if an event
1511 passes an event rule is not a dynamic task: when you create or modify an
1512 event rule, all the tracers of its tracing domain enable or disable the
1513 instrumentation points themselves once. This is possible because the
1514 attributes of an instrumentation point (type, name, and log level) are
1515 defined statically. In other words, without a dynamic filter, the tracer
1516 _does not evaluate_ the arguments of an instrumentation point unless it
1517 matches an enabled event rule.
1519 Note that, for LTTng to record an event, the <<channel,channel>> to
1520 which a matching event rule is attached must also be enabled, and the
1521 tracing session owning this channel must be active.
1524 .Logical path from an instrumentation point to an event record.
1525 image::event-rule.png[]
1527 .Event, event record, or event rule?
1529 With so many similar terms, it's easy to get confused.
1531 An **event** is the consequence of the execution of an _instrumentation
1532 point_, like a tracepoint that you manually place in some source code,
1533 or a Linux kernel KProbe. An event is said to _occur_ at a specific
1534 time. Different actions can be taken upon the occurance of an event,
1535 like record the event's payload to a buffer.
1537 An **event record** is the representation of an event in a sub-buffer. A
1538 tracer is responsible for capturing the payload of an event, current
1539 context variables, the event's ID, and the event's timestamp. LTTng
1540 can append this sub-buffer to a trace file.
1542 An **event rule** is a set of conditions which must all be satisfied for
1543 LTTng to record an occuring event. Events still occur without
1544 satisfying event rules, but LTTng does not record them.
1549 == Components of noch:{LTTng}
1551 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1552 to call LTTng a simple _tool_ since it is composed of multiple
1553 interacting components. This section describes those components,
1554 explains their respective roles, and shows how they connect together to
1555 form the LTTng ecosystem.
1557 The following diagram shows how the most important components of LTTng
1558 interact with user applications, the Linux kernel, and you:
1561 .Control and trace data paths between LTTng components.
1562 image::plumbing.png[]
1564 The LTTng project incorporates:
1566 * **LTTng-tools**: Libraries and command-line interface to
1567 control tracing sessions.
1568 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1569 ** <<lttng-consumerd,Consumer daemon>> (man:lttng-consumerd(8)).
1570 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1571 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1572 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1573 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1575 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1576 headers to instrument and trace any native user application.
1577 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1578 *** `liblttng-ust-libc-wrapper`
1579 *** `liblttng-ust-pthread-wrapper`
1580 *** `liblttng-ust-cyg-profile`
1581 *** `liblttng-ust-cyg-profile-fast`
1582 *** `liblttng-ust-dl`
1583 ** User space tracepoint provider source files generator command-line
1584 tool (man:lttng-gen-tp(1)).
1585 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1586 Java applications using `java.util.logging` or
1587 Apache log4j 1.2 logging.
1588 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1589 Python applications using the standard `logging` package.
1590 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1592 ** LTTng kernel tracer module.
1593 ** Tracing ring buffer kernel modules.
1594 ** Probe kernel modules.
1595 ** LTTng logger kernel module.
1599 === Tracing control command-line interface
1602 .The tracing control command-line interface.
1603 image::plumbing-lttng-cli.png[]
1605 The _man:lttng(1) command-line tool_ is the standard user interface to
1606 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1607 is part of LTTng-tools.
1609 The cmd:lttng tool is linked with
1610 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1611 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1613 The cmd:lttng tool has a Git-like interface:
1617 lttng <general options> <command> <command options>
1620 The <<controlling-tracing,Tracing control>> section explores the
1621 available features of LTTng using the cmd:lttng tool.
1624 [[liblttng-ctl-lttng]]
1625 === Tracing control library
1628 .The tracing control library.
1629 image::plumbing-liblttng-ctl.png[]
1631 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1632 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1633 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1635 The <<lttng-cli,cmd:lttng command-line tool>>
1636 is linked with `liblttng-ctl`.
1638 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1643 #include <lttng/lttng.h>
1646 Some objects are referenced by name (C string), such as tracing
1647 sessions, but most of them require to create a handle first using
1648 `lttng_create_handle()`.
1650 The best available developer documentation for `liblttng-ctl` is, as of
1651 LTTng{nbsp}{revision}, its installed header files. Every function and
1652 structure is thoroughly documented.
1656 === User space tracing library
1659 .The user space tracing library.
1660 image::plumbing-liblttng-ust.png[]
1662 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1663 is the LTTng user space tracer. It receives commands from a
1664 <<lttng-sessiond,session daemon>>, for example to
1665 enable and disable specific instrumentation points, and writes event
1666 records to ring buffers shared with a
1667 <<lttng-consumerd,consumer daemon>>.
1668 `liblttng-ust` is part of LTTng-UST.
1670 Public C header files are installed beside `liblttng-ust` to
1671 instrument any <<c-application,C or $$C++$$ application>>.
1673 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1674 packages, use their own library providing tracepoints which is
1675 linked with `liblttng-ust`.
1677 An application or library does not have to initialize `liblttng-ust`
1678 manually: its constructor does the necessary tasks to properly register
1679 to a session daemon. The initialization phase also enables the
1680 instrumentation points matching the <<event,event rules>> that you
1684 [[lttng-ust-agents]]
1685 === User space tracing agents
1688 .The user space tracing agents.
1689 image::plumbing-lttng-ust-agents.png[]
1691 The _LTTng-UST Java and Python agents_ are regular Java and Python
1692 packages which add LTTng tracing capabilities to the
1693 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1695 In the case of Java, the
1696 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1697 core logging facilities] and
1698 https://logging.apache.org/log4j/1.2/[Apache log4j 1.2] are supported.
1699 Note that Apache Log4{nbsp}2 is not supported.
1701 In the case of Python, the standard
1702 https://docs.python.org/3/library/logging.html[`logging`] package
1703 is supported. Both Python 2 and Python 3 modules can import the
1704 LTTng-UST Python agent package.
1706 The applications using the LTTng-UST agents are in the
1707 `java.util.logging` (JUL),
1708 log4j, and Python <<domain,tracing domains>>.
1710 Both agents use the same mechanism to trace the log statements. When an
1711 agent is initialized, it creates a log handler that attaches to the root
1712 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1713 When the application executes a log statement, it is passed to the
1714 agent's log handler by the root logger. The agent's log handler calls a
1715 native function in a tracepoint provider package shared library linked
1716 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1717 other fields, like its logger name and its log level. This native
1718 function contains a user space instrumentation point, hence tracing the
1721 The log level condition of an
1722 <<event,event rule>> is considered when tracing
1723 a Java or a Python application, and it's compatible with the standard
1724 JUL, log4j, and Python log levels.
1728 === LTTng kernel modules
1731 .The LTTng kernel modules.
1732 image::plumbing-lttng-modules.png[]
1734 The _LTTng kernel modules_ are a set of Linux kernel modules
1735 which implement the kernel tracer of the LTTng project. The LTTng
1736 kernel modules are part of LTTng-modules.
1738 The LTTng kernel modules include:
1740 * A set of _probe_ modules.
1742 Each module attaches to a specific subsystem
1743 of the Linux kernel using its tracepoint instrument points. There are
1744 also modules to attach to the entry and return points of the Linux
1745 system call functions.
1747 * _Ring buffer_ modules.
1749 A ring buffer implementation is provided as kernel modules. The LTTng
1750 kernel tracer writes to the ring buffer; a
1751 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1753 * The _LTTng kernel tracer_ module.
1754 * The _LTTng logger_ module.
1756 The LTTng logger module implements the special path:{/proc/lttng-logger}
1757 file so that any executable can generate LTTng events by opening and
1758 writing to this file.
1760 See <<proc-lttng-logger-abi,LTTng logger>>.
1762 Generally, you do not have to load the LTTng kernel modules manually
1763 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1764 daemon>> loads the necessary modules when starting. If you have extra
1765 probe modules, you can specify to load them to the session daemon on
1768 The LTTng kernel modules are installed in
1769 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1770 the kernel release (see `uname --kernel-release`).
1777 .The session daemon.
1778 image::plumbing-sessiond.png[]
1780 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1781 managing tracing sessions and for controlling the various components of
1782 LTTng. The session daemon is part of LTTng-tools.
1784 The session daemon sends control requests to and receives control
1787 * The <<lttng-ust,user space tracing library>>.
1789 Any instance of the user space tracing library first registers to
1790 a session daemon. Then, the session daemon can send requests to
1791 this instance, such as:
1794 ** Get the list of tracepoints.
1795 ** Share an <<event,event rule>> so that the user space tracing library
1796 can enable or disable tracepoints. Amongst the possible conditions
1797 of an event rule is a filter expression which `liblttng-ust` evalutes
1798 when an event occurs.
1799 ** Share <<channel,channel>> attributes and ring buffer locations.
1802 The session daemon and the user space tracing library use a Unix
1803 domain socket for their communication.
1805 * The <<lttng-ust-agents,user space tracing agents>>.
1807 Any instance of a user space tracing agent first registers to
1808 a session daemon. Then, the session daemon can send requests to
1809 this instance, such as:
1812 ** Get the list of loggers.
1813 ** Enable or disable a specific logger.
1816 The session daemon and the user space tracing agent use a TCP connection
1817 for their communication.
1819 * The <<lttng-modules,LTTng kernel tracer>>.
1820 * The <<lttng-consumerd,consumer daemon>>.
1822 The session daemon sends requests to the consumer daemon to instruct
1823 it where to send the trace data streams, amongst other information.
1825 * The <<lttng-relayd,relay daemon>>.
1827 The session daemon receives commands from the
1828 <<liblttng-ctl-lttng,tracing control library>>.
1830 The root session daemon loads the appropriate
1831 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1832 a <<lttng-consumerd,consumer daemon>> as soon as you create
1833 an <<event,event rule>>.
1835 The session daemon does not send and receive trace data: this is the
1836 role of the <<lttng-consumerd,consumer daemon>> and
1837 <<lttng-relayd,relay daemon>>. It does, however, generate the
1838 http://diamon.org/ctf/[CTF] metadata stream.
1840 Each Unix user can have its own session daemon instance. The
1841 tracing sessions managed by different session daemons are completely
1844 The root user's session daemon is the only one which is
1845 allowed to control the LTTng kernel tracer, and its spawned consumer
1846 daemon is the only one which is allowed to consume trace data from the
1847 LTTng kernel tracer. Note, however, that any Unix user which is a member
1848 of the <<tracing-group,tracing group>> is allowed
1849 to create <<channel,channels>> in the
1850 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1853 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1854 session daemon when using its `create` command if none is currently
1855 running. You can also start the session daemon manually.
1862 .The consumer daemon.
1863 image::plumbing-consumerd.png[]
1865 The _consumer daemon_, man:lttng-consumerd(8), is a daemon which shares
1866 ring buffers with user applications or with the LTTng kernel modules to
1867 collect trace data and send it to some location (on disk or to a
1868 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
1869 is part of LTTng-tools.
1871 You do not start a consumer daemon manually: a consumer daemon is always
1872 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
1873 <<event,event rule>>, that is, before you start tracing. When you kill
1874 its owner session daemon, the consumer daemon also exits because it is
1875 the session daemon's child process. Command-line options of
1876 man:lttng-sessiond(8) target the consumer daemon process.
1878 There are up to two running consumer daemons per Unix user, whereas only
1879 one session daemon can run per user. This is because each process can be
1880 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
1881 and 64-bit processes, it is more efficient to have separate
1882 corresponding 32-bit and 64-bit consumer daemons. The root user is an
1883 exception: it can have up to _three_ running consumer daemons: 32-bit
1884 and 64-bit instances for its user applications, and one more
1885 reserved for collecting kernel trace data.
1893 image::plumbing-relayd.png[]
1895 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
1896 between remote session and consumer daemons, local trace files, and a
1897 remote live trace viewer. The relay daemon is part of LTTng-tools.
1899 The main purpose of the relay daemon is to implement a receiver of
1900 <<sending-trace-data-over-the-network,trace data over the network>>.
1901 This is useful when the target system does not have much file system
1902 space to record trace files locally.
1904 The relay daemon is also a server to which a
1905 <<lttng-live,live trace viewer>> can
1906 connect. The live trace viewer sends requests to the relay daemon to
1907 receive trace data as the target system emits events. The
1908 communication protocol is named _LTTng live_; it is used over TCP
1911 Note that you can start the relay daemon on the target system directly.
1912 This is the setup of choice when the use case is to view events as
1913 the target system emits them without the need of a remote system.
1917 == [[using-lttng]]Instrumentation
1919 There are many examples of tracing and monitoring in our everyday life:
1921 * You have access to real-time and historical weather reports and
1922 forecasts thanks to weather stations installed around the country.
1923 * You know your heart is safe thanks to an electrocardiogram.
1924 * You make sure not to drive your car too fast and to have enough fuel
1925 to reach your destination thanks to gauges visible on your dashboard.
1927 All the previous examples have something in common: they rely on
1928 **instruments**. Without the electrodes attached to the surface of your
1929 body's skin, cardiac monitoring is futile.
1931 LTTng, as a tracer, is no different from those real life examples. If
1932 you're about to trace a software system or, in other words, record its
1933 history of execution, you better have **instrumentation points** in the
1934 subject you're tracing, that is, the actual software.
1936 Various ways were developed to instrument a piece of software for LTTng
1937 tracing. The most straightforward one is to manually place
1938 instrumentation points, called _tracepoints_, in the software's source
1939 code. It is also possible to add instrumentation points dynamically in
1940 the Linux kernel <<domain,tracing domain>>.
1942 If you're only interested in tracing the Linux kernel, your
1943 instrumentation needs are probably already covered by LTTng's built-in
1944 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
1945 user application which is already instrumented for LTTng tracing.
1946 In such cases, you can skip this whole section and read the topics of
1947 the <<controlling-tracing,Tracing control>> section.
1949 Many methods are available to instrument a piece of software for LTTng
1952 * <<c-application,User space instrumentation for C and $$C++$$
1954 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
1955 * <<java-application,User space Java agent>>.
1956 * <<python-application,User space Python agent>>.
1957 * <<proc-lttng-logger-abi,LTTng logger>>.
1958 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
1962 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
1964 The procedure to instrument a C or $$C++$$ user application with
1965 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
1967 . <<tracepoint-provider,Create the source files of a tracepoint provider
1969 . <<probing-the-application-source-code,Add tracepoints to
1970 the application's source code>>.
1971 . <<building-tracepoint-providers-and-user-application,Build and link
1972 a tracepoint provider package and the user application>>.
1974 If you need quick, man:printf(3)-like instrumentation, you can skip
1975 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
1978 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
1979 instrument a user application with `liblttng-ust`.
1982 [[tracepoint-provider]]
1983 ==== Create the source files of a tracepoint provider package
1985 A _tracepoint provider_ is a set of compiled functions which provide
1986 **tracepoints** to an application, the type of instrumentation point
1987 supported by LTTng-UST. Those functions can emit events with
1988 user-defined fields and serialize those events as event records to one
1989 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
1990 macro, which you <<probing-the-application-source-code,insert in a user
1991 application's source code>>, calls those functions.
1993 A _tracepoint provider package_ is an object file (`.o`) or a shared
1994 library (`.so`) which contains one or more tracepoint providers.
1995 Its source files are:
1997 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
1998 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2000 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2001 the LTTng user space tracer, at run time.
2004 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2005 image::ust-app.png[]
2007 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2008 skip creating and using a tracepoint provider and use
2009 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2013 ===== Create a tracepoint provider header file template
2015 A _tracepoint provider header file_ contains the tracepoint
2016 definitions of a tracepoint provider.
2018 To create a tracepoint provider header file:
2020 . Start from this template:
2024 .Tracepoint provider header file template (`.h` file extension).
2026 #undef TRACEPOINT_PROVIDER
2027 #define TRACEPOINT_PROVIDER provider_name
2029 #undef TRACEPOINT_INCLUDE
2030 #define TRACEPOINT_INCLUDE "./tp.h"
2032 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2035 #include <lttng/tracepoint.h>
2038 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2039 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2044 #include <lttng/tracepoint-event.h>
2050 * `provider_name` with the name of your tracepoint provider.
2051 * `"tp.h"` with the name of your tracepoint provider header file.
2053 . Below the `#include <lttng/tracepoint.h>` line, put your
2054 <<defining-tracepoints,tracepoint definitions>>.
2056 Your tracepoint provider name must be unique amongst all the possible
2057 tracepoint provider names used on the same target system. We
2058 suggest to include the name of your project or company in the name,
2059 for example, `org_lttng_my_project_tpp`.
2061 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2062 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2063 write are the <<defining-tracepoints,tracepoint definitions>>.
2066 [[defining-tracepoints]]
2067 ===== Create a tracepoint definition
2069 A _tracepoint definition_ defines, for a given tracepoint:
2071 * Its **input arguments**. They are the macro parameters that the
2072 `tracepoint()` macro accepts for this particular tracepoint
2073 in the user application's source code.
2074 * Its **output event fields**. They are the sources of event fields
2075 that form the payload of any event that the execution of the
2076 `tracepoint()` macro emits for this particular tracepoint.
2078 You can create a tracepoint definition by using the
2079 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2081 <<tpp-header,tracepoint provider header file template>>.
2083 The syntax of the `TRACEPOINT_EVENT()` macro is:
2086 .`TRACEPOINT_EVENT()` macro syntax.
2089 /* Tracepoint provider name */
2092 /* Tracepoint name */
2095 /* Input arguments */
2100 /* Output event fields */
2109 * `provider_name` with your tracepoint provider name.
2110 * `tracepoint_name` with your tracepoint name.
2111 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2112 * `fields` with the <<tpp-def-output-fields,output event field>>
2115 This tracepoint emits events named `provider_name:tracepoint_name`.
2118 .Event name's length limitation
2120 The concatenation of the tracepoint provider name and the
2121 tracepoint name must not exceed **254 characters**. If it does, the
2122 instrumented application compiles and runs, but LTTng throws multiple
2123 warnings and you could experience serious issues.
2126 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2129 .`TP_ARGS()` macro syntax.
2138 * `type` with the C type of the argument.
2139 * `arg_name` with the argument name.
2141 You can repeat `type` and `arg_name` up to 10 times to have
2142 more than one argument.
2144 .`TP_ARGS()` usage with three arguments.
2156 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2157 tracepoint definition with no input arguments.
2159 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2160 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2161 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2162 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2165 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2166 C expression that the tracer evalutes at the `tracepoint()` macro site
2167 in the application's source code. This expression provides a field's
2168 source of data. The argument expression can include input argument names
2169 listed in the `TP_ARGS()` macro.
2171 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2172 must be unique within a given tracepoint definition.
2174 Here's a complete tracepoint definition example:
2176 .Tracepoint definition.
2178 The following tracepoint definition defines a tracepoint which takes
2179 three input arguments and has four output event fields.
2183 #include "my-custom-structure.h"
2189 const struct my_custom_structure*, my_custom_structure,
2194 ctf_string(query_field, query)
2195 ctf_float(double, ratio_field, ratio)
2196 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2197 ctf_integer(int, send_size, my_custom_structure->send_size)
2202 You can refer to this tracepoint definition with the `tracepoint()`
2203 macro in your application's source code like this:
2207 tracepoint(my_provider, my_tracepoint,
2208 my_structure, some_ratio, the_query);
2212 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2213 if they satisfy an enabled <<event,event rule>>.
2216 [[using-tracepoint-classes]]
2217 ===== Use a tracepoint class
2219 A _tracepoint class_ is a class of tracepoints which share the same
2220 output event field definitions. A _tracepoint instance_ is one
2221 instance of such a defined tracepoint class, with its own tracepoint
2224 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2225 shorthand which defines both a tracepoint class and a tracepoint
2226 instance at the same time.
2228 When you build a tracepoint provider package, the C or $$C++$$ compiler
2229 creates one serialization function for each **tracepoint class**. A
2230 serialization function is responsible for serializing the event fields
2231 of a tracepoint to a sub-buffer when tracing.
2233 For various performance reasons, when your situation requires multiple
2234 tracepoint definitions with different names, but with the same event
2235 fields, we recommend that you manually create a tracepoint class
2236 and instantiate as many tracepoint instances as needed. One positive
2237 effect of such a design, amongst other advantages, is that all
2238 tracepoint instances of the same tracepoint class reuse the same
2239 serialization function, thus reducing
2240 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2242 .Use a tracepoint class and tracepoint instances.
2244 Consider the following three tracepoint definitions:
2256 ctf_integer(int, userid, userid)
2257 ctf_integer(size_t, len, len)
2269 ctf_integer(int, userid, userid)
2270 ctf_integer(size_t, len, len)
2282 ctf_integer(int, userid, userid)
2283 ctf_integer(size_t, len, len)
2288 In this case, we create three tracepoint classes, with one implicit
2289 tracepoint instance for each of them: `get_account`, `get_settings`, and
2290 `get_transaction`. However, they all share the same event field names
2291 and types. Hence three identical, yet independent serialization
2292 functions are created when you build the tracepoint provider package.
2294 A better design choice is to define a single tracepoint class and three
2295 tracepoint instances:
2299 /* The tracepoint class */
2300 TRACEPOINT_EVENT_CLASS(
2301 /* Tracepoint provider name */
2304 /* Tracepoint class name */
2307 /* Input arguments */
2313 /* Output event fields */
2315 ctf_integer(int, userid, userid)
2316 ctf_integer(size_t, len, len)
2320 /* The tracepoint instances */
2321 TRACEPOINT_EVENT_INSTANCE(
2322 /* Tracepoint provider name */
2325 /* Tracepoint class name */
2328 /* Tracepoint name */
2331 /* Input arguments */
2337 TRACEPOINT_EVENT_INSTANCE(
2346 TRACEPOINT_EVENT_INSTANCE(
2359 [[assigning-log-levels]]
2360 ===== Assign a log level to a tracepoint definition
2362 You can assign an optional _log level_ to a
2363 <<defining-tracepoints,tracepoint definition>>.
2365 Assigning different levels of severity to tracepoint definitions can
2366 be useful: when you <<enabling-disabling-events,create an event rule>>,
2367 you can target tracepoints having a log level as severe as a specific
2370 The concept of LTTng-UST log levels is similar to the levels found
2371 in typical logging frameworks:
2373 * In a logging framework, the log level is given by the function
2374 or method name you use at the log statement site: `debug()`,
2375 `info()`, `warn()`, `error()`, and so on.
2376 * In LTTng-UST, you statically assign the log level to a tracepoint
2377 definition; any `tracepoint()` macro invocation which refers to
2378 this definition has this log level.
2380 You can assign a log level to a tracepoint definition with the
2381 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2382 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2383 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2386 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2389 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2391 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2396 * `provider_name` with the tracepoint provider name.
2397 * `tracepoint_name` with the tracepoint name.
2398 * `log_level` with the log level to assign to the tracepoint
2399 definition named `tracepoint_name` in the `provider_name`
2400 tracepoint provider.
2402 See man:lttng-ust(3) for a list of available log level names.
2404 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2408 /* Tracepoint definition */
2417 ctf_integer(int, userid, userid)
2418 ctf_integer(size_t, len, len)
2422 /* Log level assignment */
2423 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2429 ===== Create a tracepoint provider package source file
2431 A _tracepoint provider package source file_ is a C source file which
2432 includes a <<tpp-header,tracepoint provider header file>> to expand its
2433 macros into event serialization and other functions.
2435 You can always use the following tracepoint provider package source
2439 .Tracepoint provider package source file template.
2441 #define TRACEPOINT_CREATE_PROBES
2446 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2447 header file>> name. You may also include more than one tracepoint
2448 provider header file here to create a tracepoint provider package
2449 holding more than one tracepoint providers.
2452 [[probing-the-application-source-code]]
2453 ==== Add tracepoints to an application's source code
2455 Once you <<tpp-header,create a tracepoint provider header file>>, you
2456 can use the `tracepoint()` macro in your application's
2457 source code to insert the tracepoints that this header
2458 <<defining-tracepoints,defines>>.
2460 The `tracepoint()` macro takes at least two parameters: the tracepoint
2461 provider name and the tracepoint name. The corresponding tracepoint
2462 definition defines the other parameters.
2464 .`tracepoint()` usage.
2466 The following <<defining-tracepoints,tracepoint definition>> defines a
2467 tracepoint which takes two input arguments and has two output event
2471 .Tracepoint provider header file.
2473 #include "my-custom-structure.h"
2480 const char*, cmd_name
2483 ctf_string(cmd_name, cmd_name)
2484 ctf_integer(int, number_of_args, argc)
2489 You can refer to this tracepoint definition with the `tracepoint()`
2490 macro in your application's source code like this:
2493 .Application's source file.
2497 int main(int argc, char* argv[])
2499 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2505 Note how the application's source code includes
2506 the tracepoint provider header file containing the tracepoint
2507 definitions to use, path:{tp.h}.
2510 .`tracepoint()` usage with a complex tracepoint definition.
2512 Consider this complex tracepoint definition, where multiple event
2513 fields refer to the same input arguments in their argument expression
2517 .Tracepoint provider header file.
2519 /* For `struct stat` */
2520 #include <sys/types.h>
2521 #include <sys/stat.h>
2533 ctf_integer(int, my_constant_field, 23 + 17)
2534 ctf_integer(int, my_int_arg_field, my_int_arg)
2535 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2536 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2537 my_str_arg[2] + my_str_arg[3])
2538 ctf_string(my_str_arg_field, my_str_arg)
2539 ctf_integer_hex(off_t, size_field, st->st_size)
2540 ctf_float(double, size_dbl_field, (double) st->st_size)
2541 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2542 size_t, strlen(my_str_arg) / 2)
2547 You can refer to this tracepoint definition with the `tracepoint()`
2548 macro in your application's source code like this:
2551 .Application's source file.
2553 #define TRACEPOINT_DEFINE
2560 stat("/etc/fstab", &s);
2561 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2567 If you look at the event record that LTTng writes when tracing this
2568 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2569 it should look like this:
2571 .Event record fields
2573 |Field's name |Field's value
2574 |`my_constant_field` |40
2575 |`my_int_arg_field` |23
2576 |`my_int_arg_field2` |529
2578 |`my_str_arg_field` |`Hello, World!`
2579 |`size_field` |0x12d
2580 |`size_dbl_field` |301.0
2581 |`half_my_str_arg_field` |`Hello,`
2585 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2586 compute--they use the call stack, for example. To avoid this
2587 computation when the tracepoint is disabled, you can use the
2588 `tracepoint_enabled()` and `do_tracepoint()` macros.
2590 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2594 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2596 tracepoint_enabled(provider_name, tracepoint_name)
2597 do_tracepoint(provider_name, tracepoint_name, ...)
2602 * `provider_name` with the tracepoint provider name.
2603 * `tracepoint_name` with the tracepoint name.
2605 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2606 `tracepoint_name` from the provider named `provider_name` is enabled
2609 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2610 if the tracepoint is enabled. Using `tracepoint()` with
2611 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2612 the `tracepoint_enabled()` check, thus a race condition is
2613 possible in this situation:
2616 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2618 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2619 stuff = prepare_stuff();
2622 tracepoint(my_provider, my_tracepoint, stuff);
2625 If the tracepoint is enabled after the condition, then `stuff` is not
2626 prepared: the emitted event will either contain wrong data, or the whole
2627 application could crash (segmentation fault, for example).
2629 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2630 `STAP_PROBEV()` call. If you need it, you must emit
2634 [[building-tracepoint-providers-and-user-application]]
2635 ==== Build and link a tracepoint provider package and an application
2637 Once you have one or more <<tpp-header,tracepoint provider header
2638 files>> and a <<tpp-source,tracepoint provider package source file>>,
2639 you can create the tracepoint provider package by compiling its source
2640 file. From here, multiple build and run scenarios are possible. The
2641 following table shows common application and library configurations
2642 along with the required command lines to achieve them.
2644 In the following diagrams, we use the following file names:
2647 Executable application.
2650 Application's object file.
2653 Tracepoint provider package object file.
2656 Tracepoint provider package archive file.
2659 Tracepoint provider package shared object file.
2662 User library object file.
2665 User library shared object file.
2667 We use the following symbols in the diagrams of table below:
2670 .Symbols used in the build scenario diagrams.
2671 image::ust-sit-symbols.png[]
2673 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2674 variable in the following instructions.
2676 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2677 .Common tracepoint provider package scenarios.
2679 |Scenario |Instructions
2682 The instrumented application is statically linked with
2683 the tracepoint provider package object.
2685 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2688 include::../common/ust-sit-step-tp-o.txt[]
2690 To build the instrumented application:
2692 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2697 #define TRACEPOINT_DEFINE
2701 . Compile the application source file:
2710 . Build the application:
2715 gcc -o app app.o tpp.o -llttng-ust -ldl
2719 To run the instrumented application:
2721 * Start the application:
2731 The instrumented application is statically linked with the
2732 tracepoint provider package archive file.
2734 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2737 To create the tracepoint provider package archive file:
2739 . Compile the <<tpp-source,tracepoint provider package source file>>:
2748 . Create the tracepoint provider package archive file:
2757 To build the instrumented application:
2759 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2764 #define TRACEPOINT_DEFINE
2768 . Compile the application source file:
2777 . Build the application:
2782 gcc -o app app.o tpp.a -llttng-ust -ldl
2786 To run the instrumented application:
2788 * Start the application:
2798 The instrumented application is linked with the tracepoint provider
2799 package shared object.
2801 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2804 include::../common/ust-sit-step-tp-so.txt[]
2806 To build the instrumented application:
2808 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2813 #define TRACEPOINT_DEFINE
2817 . Compile the application source file:
2826 . Build the application:
2831 gcc -o app app.o -ldl -L. -ltpp
2835 To run the instrumented application:
2837 * Start the application:
2847 The tracepoint provider package shared object is preloaded before the
2848 instrumented application starts.
2850 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2853 include::../common/ust-sit-step-tp-so.txt[]
2855 To build the instrumented application:
2857 . In path:{app.c}, before including path:{tpp.h}, add the
2863 #define TRACEPOINT_DEFINE
2864 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2868 . Compile the application source file:
2877 . Build the application:
2882 gcc -o app app.o -ldl
2886 To run the instrumented application with tracing support:
2888 * Preload the tracepoint provider package shared object and
2889 start the application:
2894 LD_PRELOAD=./libtpp.so ./app
2898 To run the instrumented application without tracing support:
2900 * Start the application:
2910 The instrumented application dynamically loads the tracepoint provider
2911 package shared object.
2913 See the <<dlclose-warning,warning about `dlclose()`>>.
2915 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
2918 include::../common/ust-sit-step-tp-so.txt[]
2920 To build the instrumented application:
2922 . In path:{app.c}, before including path:{tpp.h}, add the
2928 #define TRACEPOINT_DEFINE
2929 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2933 . Compile the application source file:
2942 . Build the application:
2947 gcc -o app app.o -ldl
2951 To run the instrumented application:
2953 * Start the application:
2963 The application is linked with the instrumented user library.
2965 The instrumented user library is statically linked with the tracepoint
2966 provider package object file.
2968 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
2971 include::../common/ust-sit-step-tp-o-fpic.txt[]
2973 To build the instrumented user library:
2975 . In path:{emon.c}, before including path:{tpp.h}, add the
2981 #define TRACEPOINT_DEFINE
2985 . Compile the user library source file:
2990 gcc -I. -fpic -c emon.c
2994 . Build the user library shared object:
2999 gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3003 To build the application:
3005 . Compile the application source file:
3014 . Build the application:
3019 gcc -o app app.o -L. -lemon
3023 To run the application:
3025 * Start the application:
3035 The application is linked with the instrumented user library.
3037 The instrumented user library is linked with the tracepoint provider
3038 package shared object.
3040 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3043 include::../common/ust-sit-step-tp-so.txt[]
3045 To build the instrumented user library:
3047 . In path:{emon.c}, before including path:{tpp.h}, add the
3053 #define TRACEPOINT_DEFINE
3057 . Compile the user library source file:
3062 gcc -I. -fpic -c emon.c
3066 . Build the user library shared object:
3071 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3075 To build the application:
3077 . Compile the application source file:
3086 . Build the application:
3091 gcc -o app app.o -L. -lemon
3095 To run the application:
3097 * Start the application:
3107 The tracepoint provider package shared object is preloaded before the
3110 The application is linked with the instrumented user library.
3112 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3115 include::../common/ust-sit-step-tp-so.txt[]
3117 To build the instrumented user library:
3119 . In path:{emon.c}, before including path:{tpp.h}, add the
3125 #define TRACEPOINT_DEFINE
3126 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3130 . Compile the user library source file:
3135 gcc -I. -fpic -c emon.c
3139 . Build the user library shared object:
3144 gcc -shared -o libemon.so emon.o -ldl
3148 To build the application:
3150 . Compile the application source file:
3159 . Build the application:
3164 gcc -o app app.o -L. -lemon
3168 To run the application with tracing support:
3170 * Preload the tracepoint provider package shared object and
3171 start the application:
3176 LD_PRELOAD=./libtpp.so ./app
3180 To run the application without tracing support:
3182 * Start the application:
3192 The application is linked with the instrumented user library.
3194 The instrumented user library dynamically loads the tracepoint provider
3195 package shared object.
3197 See the <<dlclose-warning,warning about `dlclose()`>>.
3199 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3202 include::../common/ust-sit-step-tp-so.txt[]
3204 To build the instrumented user library:
3206 . In path:{emon.c}, before including path:{tpp.h}, add the
3212 #define TRACEPOINT_DEFINE
3213 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3217 . Compile the user library source file:
3222 gcc -I. -fpic -c emon.c
3226 . Build the user library shared object:
3231 gcc -shared -o libemon.so emon.o -ldl
3235 To build the application:
3237 . Compile the application source file:
3246 . Build the application:
3251 gcc -o app app.o -L. -lemon
3255 To run the application:
3257 * Start the application:
3267 The application dynamically loads the instrumented user library.
3269 The instrumented user library is linked with the tracepoint provider
3270 package shared object.
3272 See the <<dlclose-warning,warning about `dlclose()`>>.
3274 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3277 include::../common/ust-sit-step-tp-so.txt[]
3279 To build the instrumented user library:
3281 . In path:{emon.c}, before including path:{tpp.h}, add the
3287 #define TRACEPOINT_DEFINE
3291 . Compile the user library source file:
3296 gcc -I. -fpic -c emon.c
3300 . Build the user library shared object:
3305 gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3309 To build the application:
3311 . Compile the application source file:
3320 . Build the application:
3325 gcc -o app app.o -ldl -L. -lemon
3329 To run the application:
3331 * Start the application:
3341 The application dynamically loads the instrumented user library.
3343 The instrumented user library dynamically loads the tracepoint provider
3344 package shared object.
3346 See the <<dlclose-warning,warning about `dlclose()`>>.
3348 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3351 include::../common/ust-sit-step-tp-so.txt[]
3353 To build the instrumented user library:
3355 . In path:{emon.c}, before including path:{tpp.h}, add the
3361 #define TRACEPOINT_DEFINE
3362 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3366 . Compile the user library source file:
3371 gcc -I. -fpic -c emon.c
3375 . Build the user library shared object:
3380 gcc -shared -o libemon.so emon.o -ldl
3384 To build the application:
3386 . Compile the application source file:
3395 . Build the application:
3400 gcc -o app app.o -ldl -L. -lemon
3404 To run the application:
3406 * Start the application:
3416 The tracepoint provider package shared object is preloaded before the
3419 The application dynamically loads the instrumented user library.
3421 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3424 include::../common/ust-sit-step-tp-so.txt[]
3426 To build the instrumented user library:
3428 . In path:{emon.c}, before including path:{tpp.h}, add the
3434 #define TRACEPOINT_DEFINE
3435 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3439 . Compile the user library source file:
3444 gcc -I. -fpic -c emon.c
3448 . Build the user library shared object:
3453 gcc -shared -o libemon.so emon.o -ldl
3457 To build the application:
3459 . Compile the application source file:
3468 . Build the application:
3473 gcc -o app app.o -L. -lemon
3477 To run the application with tracing support:
3479 * Preload the tracepoint provider package shared object and
3480 start the application:
3485 LD_PRELOAD=./libtpp.so ./app
3489 To run the application without tracing support:
3491 * Start the application:
3501 The application is statically linked with the tracepoint provider
3502 package object file.
3504 The application is linked with the instrumented user library.
3506 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3509 include::../common/ust-sit-step-tp-o.txt[]
3511 To build the instrumented user library:
3513 . In path:{emon.c}, before including path:{tpp.h}, add the
3519 #define TRACEPOINT_DEFINE
3523 . Compile the user library source file:
3528 gcc -I. -fpic -c emon.c
3532 . Build the user library shared object:
3537 gcc -shared -o libemon.so emon.o
3541 To build the application:
3543 . Compile the application source file:
3552 . Build the application:
3557 gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3561 To run the instrumented application:
3563 * Start the application:
3573 The application is statically linked with the tracepoint provider
3574 package object file.
3576 The application dynamically loads the instrumented user library.
3578 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3581 include::../common/ust-sit-step-tp-o.txt[]
3583 To build the application:
3585 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3590 #define TRACEPOINT_DEFINE
3594 . Compile the application source file:
3603 . Build the application:
3608 gcc -Wl,--export-dynamic -o app app.o tpp.o \
3613 The `--export-dynamic` option passed to the linker is necessary for the
3614 dynamically loaded library to ``see'' the tracepoint symbols defined in
3617 To build the instrumented user library:
3619 . Compile the user library source file:
3624 gcc -I. -fpic -c emon.c
3628 . Build the user library shared object:
3633 gcc -shared -o libemon.so emon.o
3637 To run the application:
3639 * Start the application:
3651 .Do not use man:dlclose(3) on a tracepoint provider package
3653 Never use man:dlclose(3) on any shared object which:
3655 * Is linked with, statically or dynamically, a tracepoint provider
3657 * Calls man:dlopen(3) itself to dynamically open a tracepoint provider
3658 package shared object.
3660 This is currently considered **unsafe** due to a lack of reference
3661 counting from LTTng-UST to the shared object.
3663 A known workaround (available since glibc 2.2) is to use the
3664 `RTLD_NODELETE` flag when calling man:dlopen(3) initially. This has the
3665 effect of not unloading the loaded shared object, even if man:dlclose(3)
3668 You can also preload the tracepoint provider package shared object with
3669 the env:LD_PRELOAD environment variable to overcome this limitation.
3673 [[using-lttng-ust-with-daemons]]
3674 ===== Use noch:{LTTng-UST} with daemons
3676 If your instrumented application calls man:fork(2), man:clone(2),
3677 or BSD's man:rfork(2), without a following man:exec(3)-family
3678 system call, you must preload the path:{liblttng-ust-fork.so} shared
3679 object when you start the application.
3683 LD_PRELOAD=liblttng-ust-fork.so ./my-app
3686 If your tracepoint provider package is
3687 a shared library which you also preload, you must put both
3688 shared objects in env:LD_PRELOAD:
3692 LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3698 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3700 If your instrumented application closes one or more file descriptors
3701 which it did not open itself, you must preload the
3702 path:{liblttng-ust-fd.so} shared object when you start the application:
3706 LD_PRELOAD=liblttng-ust-fd.so ./my-app
3709 Typical use cases include closing all the file descriptors after
3710 man:fork(2) or man:rfork(2) and buggy applications doing
3714 [[lttng-ust-pkg-config]]
3715 ===== Use noch:{pkg-config}
3717 On some distributions, LTTng-UST ships with a
3718 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3719 metadata file. If this is your case, then you can use cmd:pkg-config to
3720 build an application on the command line:
3724 gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3728 [[instrumenting-32-bit-app-on-64-bit-system]]
3729 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3731 In order to trace a 32-bit application running on a 64-bit system,
3732 LTTng must use a dedicated 32-bit
3733 <<lttng-consumerd,consumer daemon>>.
3735 The following steps show how to build and install a 32-bit consumer
3736 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3737 build and install the 32-bit LTTng-UST libraries, and how to build and
3738 link an instrumented 32-bit application in that context.
3740 To build a 32-bit instrumented application for a 64-bit target system,
3741 assuming you have a fresh target system with no installed Userspace RCU
3744 . Download, build, and install a 32-bit version of Userspace RCU:
3750 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3751 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3752 cd userspace-rcu-0.9.* &&
3753 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3755 sudo make install &&
3760 . Using your distribution's package manager, or from source, install
3761 the following 32-bit versions of the following dependencies of
3762 LTTng-tools and LTTng-UST:
3765 * https://sourceforge.net/projects/libuuid/[libuuid]
3766 * http://directory.fsf.org/wiki/Popt[popt]
3767 * http://www.xmlsoft.org/[libxml2]
3770 . Download, build, and install a 32-bit version of the latest
3771 LTTng-UST{nbsp}{revision}:
3777 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.9.tar.bz2 &&
3778 tar -xf lttng-ust-latest-2.9.tar.bz2 &&
3779 cd lttng-ust-2.9.* &&
3780 ./configure --libdir=/usr/local/lib32 \
3781 CFLAGS=-m32 CXXFLAGS=-m32 \
3782 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3784 sudo make install &&
3791 Depending on your distribution,
3792 32-bit libraries could be installed at a different location than
3793 `/usr/lib32`. For example, Debian is known to install
3794 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3796 In this case, make sure to set `LDFLAGS` to all the
3797 relevant 32-bit library paths, for example:
3801 LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3805 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3806 the 32-bit consumer daemon:
3812 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
3813 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
3814 cd lttng-tools-2.9.* &&
3815 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3816 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3818 cd src/bin/lttng-consumerd &&
3819 sudo make install &&
3824 . From your distribution or from source,
3825 <<installing-lttng,install>> the 64-bit versions of
3826 LTTng-UST and Userspace RCU.
3827 . Download, build, and install the 64-bit version of the
3828 latest LTTng-tools{nbsp}{revision}:
3834 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.9.tar.bz2 &&
3835 tar -xf lttng-tools-latest-2.9.tar.bz2 &&
3836 cd lttng-tools-2.9.* &&
3837 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3838 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3840 sudo make install &&
3845 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3846 when linking your 32-bit application:
3849 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3850 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3853 For example, let's rebuild the quick start example in
3854 <<tracing-your-own-user-application,Trace a user application>> as an
3855 instrumented 32-bit application:
3860 gcc -m32 -c -I. hello-tp.c
3862 gcc -m32 -o hello hello.o hello-tp.o \
3863 -L/usr/lib32 -L/usr/local/lib32 \
3864 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3869 No special action is required to execute the 32-bit application and
3870 to trace it: use the command-line man:lttng(1) tool as usual.
3877 man:tracef(3) is a small LTTng-UST API designed for quick,
3878 man:printf(3)-like instrumentation without the burden of
3879 <<tracepoint-provider,creating>> and
3880 <<building-tracepoint-providers-and-user-application,building>>
3881 a tracepoint provider package.
3883 To use `tracef()` in your application:
3885 . In the C or C++ source files where you need to use `tracef()`,
3886 include `<lttng/tracef.h>`:
3891 #include <lttng/tracef.h>
3895 . In the application's source code, use `tracef()` like you would use
3903 tracef("my message: %d (%s)", my_integer, my_string);
3909 . Link your application with `liblttng-ust`:
3914 gcc -o app app.c -llttng-ust
3918 To trace the events that `tracef()` calls emit:
3920 * <<enabling-disabling-events,Create an event rule>> which matches the
3921 `lttng_ust_tracef:*` event name:
3926 lttng enable-event --userspace 'lttng_ust_tracef:*'
3931 .Limitations of `tracef()`
3933 The `tracef()` utility function was developed to make user space tracing
3934 super simple, albeit with notable disadvantages compared to
3935 <<defining-tracepoints,user-defined tracepoints>>:
3937 * All the emitted events have the same tracepoint provider and
3938 tracepoint names, respectively `lttng_ust_tracef` and `event`.
3939 * There is no static type checking.
3940 * The only event record field you actually get, named `msg`, is a string
3941 potentially containing the values you passed to `tracef()`
3942 using your own format string. This also means that you cannot filter
3943 events with a custom expression at run time because there are no
3945 * Since `tracef()` uses the C standard library's man:vasprintf(3)
3946 function behind the scenes to format the strings at run time, its
3947 expected performance is lower than with user-defined tracepoints,
3948 which do not require a conversion to a string.
3950 Taking this into consideration, `tracef()` is useful for some quick
3951 prototyping and debugging, but you should not consider it for any
3952 permanent and serious applicative instrumentation.
3958 ==== Use `tracelog()`
3960 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
3961 the difference that it accepts an additional log level parameter.
3963 The goal of `tracelog()` is to ease the migration from logging to
3966 To use `tracelog()` in your application:
3968 . In the C or C++ source files where you need to use `tracelog()`,
3969 include `<lttng/tracelog.h>`:
3974 #include <lttng/tracelog.h>
3978 . In the application's source code, use `tracelog()` like you would use
3979 man:printf(3), except for the first parameter which is the log
3987 tracelog(TRACE_WARNING, "my message: %d (%s)",
3988 my_integer, my_string);
3994 See man:lttng-ust(3) for a list of available log level names.
3996 . Link your application with `liblttng-ust`:
4001 gcc -o app app.c -llttng-ust
4005 To trace the events that `tracelog()` calls emit with a log level
4006 _as severe as_ a specific log level:
4008 * <<enabling-disabling-events,Create an event rule>> which matches the
4009 `lttng_ust_tracelog:*` event name and a minimum level
4015 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4016 --loglevel=TRACE_WARNING
4020 To trace the events that `tracelog()` calls emit with a
4021 _specific log level_:
4023 * Create an event rule which matches the `lttng_ust_tracelog:*`
4024 event name and a specific log level:
4029 lttng enable-event --userspace 'lttng_ust_tracelog:*'
4030 --loglevel-only=TRACE_INFO
4035 [[prebuilt-ust-helpers]]
4036 === Prebuilt user space tracing helpers
4038 The LTTng-UST package provides a few helpers in the form or preloadable
4039 shared objects which automatically instrument system functions and
4042 The helper shared objects are normally found in dir:{/usr/lib}. If you
4043 built LTTng-UST <<building-from-source,from source>>, they are probably
4044 located in dir:{/usr/local/lib}.
4046 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4049 path:{liblttng-ust-libc-wrapper.so}::
4050 path:{liblttng-ust-pthread-wrapper.so}::
4051 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4052 memory and POSIX threads function tracing>>.
4054 path:{liblttng-ust-cyg-profile.so}::
4055 path:{liblttng-ust-cyg-profile-fast.so}::
4056 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4058 path:{liblttng-ust-dl.so}::
4059 <<liblttng-ust-dl,Dynamic linker tracing>>.
4061 To use a user space tracing helper with any user application:
4063 * Preload the helper shared object when you start the application:
4068 LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4072 You can preload more than one helper:
4077 LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4083 [[liblttng-ust-libc-pthread-wrapper]]
4084 ==== Instrument C standard library memory and POSIX threads functions
4086 The path:{liblttng-ust-libc-wrapper.so} and
4087 path:{liblttng-ust-pthread-wrapper.so} helpers
4088 add instrumentation to some C standard library and POSIX
4092 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4094 |TP provider name |TP name |Instrumented function
4096 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4097 |`calloc` |man:calloc(3)
4098 |`realloc` |man:realloc(3)
4099 |`free` |man:free(3)
4100 |`memalign` |man:memalign(3)
4101 |`posix_memalign` |man:posix_memalign(3)
4105 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4107 |TP provider name |TP name |Instrumented function
4109 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4110 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4111 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4112 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4115 When you preload the shared object, it replaces the functions listed
4116 in the previous tables by wrappers which contain tracepoints and call
4117 the replaced functions.
4120 [[liblttng-ust-cyg-profile]]
4121 ==== Instrument function entry and exit
4123 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4124 to the entry and exit points of functions.
4126 man:gcc(1) and man:clang(1) have an option named
4127 https://gcc.gnu.org/onlinedocs/gcc/Code-Gen-Options.html[`-finstrument-functions`]
4128 which generates instrumentation calls for entry and exit to functions.
4129 The LTTng-UST function tracing helpers,
4130 path:{liblttng-ust-cyg-profile.so} and
4131 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4132 to add tracepoints to the two generated functions (which contain
4133 `cyg_profile` in their names, hence the helper's name).
4135 To use the LTTng-UST function tracing helper, the source files to
4136 instrument must be built using the `-finstrument-functions` compiler
4139 There are two versions of the LTTng-UST function tracing helper:
4141 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4142 that you should only use when it can be _guaranteed_ that the
4143 complete event stream is recorded without any lost event record.
4144 Any kind of duplicate information is left out.
4146 Assuming no event record is lost, having only the function addresses on
4147 entry is enough to create a call graph, since an event record always
4148 contains the ID of the CPU that generated it.
4150 You can use a tool like man:addr2line(1) to convert function addresses
4151 back to source file names and line numbers.
4153 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4154 which also works in use cases where event records might get discarded or
4155 not recorded from application startup.
4156 In these cases, the trace analyzer needs more information to be
4157 able to reconstruct the program flow.
4159 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4160 points of this helper.
4162 All the tracepoints that this helper provides have the
4163 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4165 TIP: It's sometimes a good idea to limit the number of source files that
4166 you compile with the `-finstrument-functions` option to prevent LTTng
4167 from writing an excessive amount of trace data at run time. When using
4168 man:gcc(1), you can use the
4169 `-finstrument-functions-exclude-function-list` option to avoid
4170 instrument entries and exits of specific function names.
4175 ==== Instrument the dynamic linker
4177 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4178 man:dlopen(3) and man:dlclose(3) function calls.
4180 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4185 [[java-application]]
4186 === User space Java agent
4188 You can instrument any Java application which uses one of the following
4191 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4192 (JUL) core logging facilities.
4193 * http://logging.apache.org/log4j/1.2/[**Apache log4j 1.2**], since
4194 LTTng 2.6. Note that Apache Log4j{nbsp}2 is not supported.
4197 .LTTng-UST Java agent imported by a Java application.
4198 image::java-app.png[]
4200 Note that the methods described below are new in LTTng{nbsp}{revision}.
4201 Previous LTTng versions use another technique.
4203 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4204 and https://ci.lttng.org/[continuous integration], thus this version is
4205 directly supported. However, the LTTng-UST Java agent is also tested
4206 with OpenJDK{nbsp}7.
4211 ==== Use the LTTng-UST Java agent for `java.util.logging`
4213 To use the LTTng-UST Java agent in a Java application which uses
4214 `java.util.logging` (JUL):
4216 . In the Java application's source code, import the LTTng-UST
4217 log handler package for `java.util.logging`:
4222 import org.lttng.ust.agent.jul.LttngLogHandler;
4226 . Create an LTTng-UST JUL log handler:
4231 Handler lttngUstLogHandler = new LttngLogHandler();
4235 . Add this handler to the JUL loggers which should emit LTTng events:
4240 Logger myLogger = Logger.getLogger("some-logger");
4242 myLogger.addHandler(lttngUstLogHandler);
4246 . Use `java.util.logging` log statements and configuration as usual.
4247 The loggers with an attached LTTng-UST log handler can emit
4250 . Before exiting the application, remove the LTTng-UST log handler from
4251 the loggers attached to it and call its `close()` method:
4256 myLogger.removeHandler(lttngUstLogHandler);
4257 lttngUstLogHandler.close();
4261 This is not strictly necessary, but it is recommended for a clean
4262 disposal of the handler's resources.
4264 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4265 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4267 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4268 path] when you build the Java application.
4270 The JAR files are typically located in dir:{/usr/share/java}.
4272 IMPORTANT: The LTTng-UST Java agent must be
4273 <<installing-lttng,installed>> for the logging framework your
4276 .Use the LTTng-UST Java agent for `java.util.logging`.
4281 import java.io.IOException;
4282 import java.util.logging.Handler;
4283 import java.util.logging.Logger;
4284 import org.lttng.ust.agent.jul.LttngLogHandler;
4288 private static final int answer = 42;
4290 public static void main(String[] argv) throws Exception
4293 Logger logger = Logger.getLogger("jello");
4295 // Create an LTTng-UST log handler
4296 Handler lttngUstLogHandler = new LttngLogHandler();
4298 // Add the LTTng-UST log handler to our logger
4299 logger.addHandler(lttngUstLogHandler);
4302 logger.info("some info");
4303 logger.warning("some warning");
4305 logger.finer("finer information; the answer is " + answer);
4307 logger.severe("error!");
4309 // Not mandatory, but cleaner
4310 logger.removeHandler(lttngUstLogHandler);
4311 lttngUstLogHandler.close();
4320 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4323 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4324 <<enabling-disabling-events,create an event rule>> matching the
4325 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4330 lttng enable-event --jul jello
4334 Run the compiled class:
4338 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4341 <<basic-tracing-session-control,Stop tracing>> and inspect the
4351 You can use the opt:lttng-enable-event(1):--loglevel or
4352 opt:lttng-enable-event(1):--loglevel-only option of the
4353 man:lttng-enable-event(1) command to target a range of JUL log levels
4354 or a specific JUL log level.
4359 ==== Use the LTTng-UST Java agent for Apache log4j
4361 To use the LTTng-UST Java agent in a Java application which uses
4364 . In the Java application's source code, import the LTTng-UST
4365 log appender package for Apache log4j:
4370 import org.lttng.ust.agent.log4j.LttngLogAppender;
4374 . Create an LTTng-UST log4j log appender:
4379 Appender lttngUstLogAppender = new LttngLogAppender();
4383 . Add this appender to the log4j loggers which should emit LTTng events:
4388 Logger myLogger = Logger.getLogger("some-logger");
4390 myLogger.addAppender(lttngUstLogAppender);
4394 . Use Apache log4j log statements and configuration as usual. The
4395 loggers with an attached LTTng-UST log appender can emit LTTng events.
4397 . Before exiting the application, remove the LTTng-UST log appender from
4398 the loggers attached to it and call its `close()` method:
4403 myLogger.removeAppender(lttngUstLogAppender);
4404 lttngUstLogAppender.close();
4408 This is not strictly necessary, but it is recommended for a clean
4409 disposal of the appender's resources.
4411 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4412 files, path:{lttng-ust-agent-common.jar} and
4413 path:{lttng-ust-agent-log4j.jar}, in the
4414 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4415 path] when you build the Java application.
4417 The JAR files are typically located in dir:{/usr/share/java}.
4419 IMPORTANT: The LTTng-UST Java agent must be
4420 <<installing-lttng,installed>> for the logging framework your
4423 .Use the LTTng-UST Java agent for Apache log4j.
4428 import org.apache.log4j.Appender;
4429 import org.apache.log4j.Logger;
4430 import org.lttng.ust.agent.log4j.LttngLogAppender;
4434 private static final int answer = 42;
4436 public static void main(String[] argv) throws Exception
4439 Logger logger = Logger.getLogger("jello");
4441 // Create an LTTng-UST log appender
4442 Appender lttngUstLogAppender = new LttngLogAppender();
4444 // Add the LTTng-UST log appender to our logger
4445 logger.addAppender(lttngUstLogAppender);
4448 logger.info("some info");
4449 logger.warn("some warning");
4451 logger.debug("debug information; the answer is " + answer);
4453 logger.fatal("error!");
4455 // Not mandatory, but cleaner
4456 logger.removeAppender(lttngUstLogAppender);
4457 lttngUstLogAppender.close();
4463 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4468 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4471 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4472 <<enabling-disabling-events,create an event rule>> matching the
4473 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4478 lttng enable-event --log4j jello
4482 Run the compiled class:
4486 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4489 <<basic-tracing-session-control,Stop tracing>> and inspect the
4499 You can use the opt:lttng-enable-event(1):--loglevel or
4500 opt:lttng-enable-event(1):--loglevel-only option of the
4501 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4502 or a specific log4j log level.
4506 [[java-application-context]]
4507 ==== Provide application-specific context fields in a Java application
4509 A Java application-specific context field is a piece of state provided
4510 by the application which <<adding-context,you can add>>, using the
4511 man:lttng-add-context(1) command, to each <<event,event record>>
4512 produced by the log statements of this application.
4514 For example, a given object might have a current request ID variable.
4515 You can create a context information retriever for this object and
4516 assign a name to this current request ID. You can then, using the
4517 man:lttng-add-context(1) command, add this context field by name to
4518 the JUL or log4j <<channel,channel>>.
4520 To provide application-specific context fields in a Java application:
4522 . In the Java application's source code, import the LTTng-UST
4523 Java agent context classes and interfaces:
4528 import org.lttng.ust.agent.context.ContextInfoManager;
4529 import org.lttng.ust.agent.context.IContextInfoRetriever;
4533 . Create a context information retriever class, that is, a class which
4534 implements the `IContextInfoRetriever` interface:
4539 class MyContextInfoRetriever implements IContextInfoRetriever
4542 public Object retrieveContextInfo(String key)
4544 if (key.equals("intCtx")) {
4546 } else if (key.equals("strContext")) {
4547 return "context value!";
4556 This `retrieveContextInfo()` method is the only member of the
4557 `IContextInfoRetriever` interface. Its role is to return the current
4558 value of a state by name to create a context field. The names of the
4559 context fields and which state variables they return depends on your
4562 All primitive types and objects are supported as context fields.
4563 When `retrieveContextInfo()` returns an object, the context field
4564 serializer calls its `toString()` method to add a string field to
4565 event records. The method can also return `null`, which means that
4566 no context field is available for the required name.
4568 . Register an instance of your context information retriever class to
4569 the context information manager singleton:
4574 IContextInfoRetriever cir = new MyContextInfoRetriever();
4575 ContextInfoManager cim = ContextInfoManager.getInstance();
4576 cim.registerContextInfoRetriever("retrieverName", cir);
4580 . Before exiting the application, remove your context information
4581 retriever from the context information manager singleton:
4586 ContextInfoManager cim = ContextInfoManager.getInstance();
4587 cim.unregisterContextInfoRetriever("retrieverName");
4591 This is not strictly necessary, but it is recommended for a clean
4592 disposal of some manager's resources.
4594 . Build your Java application with LTTng-UST Java agent support as
4595 usual, following the procedure for either the <<jul,JUL>> or
4596 <<log4j,Apache log4j>> framework.
4599 .Provide application-specific context fields in a Java application.
4604 import java.util.logging.Handler;
4605 import java.util.logging.Logger;
4606 import org.lttng.ust.agent.jul.LttngLogHandler;
4607 import org.lttng.ust.agent.context.ContextInfoManager;
4608 import org.lttng.ust.agent.context.IContextInfoRetriever;
4612 // Our context information retriever class
4613 private static class MyContextInfoRetriever
4614 implements IContextInfoRetriever
4617 public Object retrieveContextInfo(String key) {
4618 if (key.equals("intCtx")) {
4620 } else if (key.equals("strContext")) {
4621 return "context value!";
4628 private static final int answer = 42;
4630 public static void main(String args[]) throws Exception
4632 // Get the context information manager instance
4633 ContextInfoManager cim = ContextInfoManager.getInstance();
4635 // Create and register our context information retriever
4636 IContextInfoRetriever cir = new MyContextInfoRetriever();
4637 cim.registerContextInfoRetriever("myRetriever", cir);
4640 Logger logger = Logger.getLogger("jello");
4642 // Create an LTTng-UST log handler
4643 Handler lttngUstLogHandler = new LttngLogHandler();
4645 // Add the LTTng-UST log handler to our logger
4646 logger.addHandler(lttngUstLogHandler);
4649 logger.info("some info");
4650 logger.warning("some warning");
4652 logger.finer("finer information; the answer is " + answer);
4654 logger.severe("error!");
4656 // Not mandatory, but cleaner
4657 logger.removeHandler(lttngUstLogHandler);
4658 lttngUstLogHandler.close();
4659 cim.unregisterContextInfoRetriever("myRetriever");
4668 javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4671 <<creating-destroying-tracing-sessions,Create a tracing session>>
4672 and <<enabling-disabling-events,create an event rule>> matching the
4678 lttng enable-event --jul jello
4681 <<adding-context,Add the application-specific context fields>> to the
4686 lttng add-context --jul --type='$app.myRetriever:intCtx'
4687 lttng add-context --jul --type='$app.myRetriever:strContext'
4690 <<basic-tracing-session-control,Start tracing>>:
4697 Run the compiled class:
4701 java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4704 <<basic-tracing-session-control,Stop tracing>> and inspect the
4716 [[python-application]]
4717 === User space Python agent
4719 You can instrument a Python 2 or Python 3 application which uses the
4720 standard https://docs.python.org/3/library/logging.html[`logging`]
4723 Each log statement emits an LTTng event once the
4724 application module imports the
4725 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4728 .A Python application importing the LTTng-UST Python agent.
4729 image::python-app.png[]
4731 To use the LTTng-UST Python agent:
4733 . In the Python application's source code, import the LTTng-UST Python
4743 The LTTng-UST Python agent automatically adds its logging handler to the
4744 root logger at import time.
4746 Any log statement that the application executes before this import does
4747 not emit an LTTng event.
4749 IMPORTANT: The LTTng-UST Python agent must be
4750 <<installing-lttng,installed>>.
4752 . Use log statements and logging configuration as usual.
4753 Since the LTTng-UST Python agent adds a handler to the _root_
4754 logger, you can trace any log statement from any logger.
4756 .Use the LTTng-UST Python agent.
4767 logging.basicConfig()
4768 logger = logging.getLogger('my-logger')
4771 logger.debug('debug message')
4772 logger.info('info message')
4773 logger.warn('warn message')
4774 logger.error('error message')
4775 logger.critical('critical message')
4779 if __name__ == '__main__':
4783 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4784 logging handler which prints to the standard error stream, is not
4785 strictly required for LTTng-UST tracing to work, but in versions of
4786 Python preceding 3.2, you could see a warning message which indicates
4787 that no handler exists for the logger `my-logger`.
4789 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4790 <<enabling-disabling-events,create an event rule>> matching the
4791 `my-logger` Python logger, and <<basic-tracing-session-control,start
4797 lttng enable-event --python my-logger
4801 Run the Python script:
4808 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4818 You can use the opt:lttng-enable-event(1):--loglevel or
4819 opt:lttng-enable-event(1):--loglevel-only option of the
4820 man:lttng-enable-event(1) command to target a range of Python log levels
4821 or a specific Python log level.
4823 When an application imports the LTTng-UST Python agent, the agent tries
4824 to register to a <<lttng-sessiond,session daemon>>. Note that you must
4825 <<start-sessiond,start the session daemon>> _before_ you run the Python
4826 application. If a session daemon is found, the agent tries to register
4827 to it during 5{nbsp}seconds, after which the application continues
4828 without LTTng tracing support. You can override this timeout value with
4829 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
4832 If the session daemon stops while a Python application with an imported
4833 LTTng-UST Python agent runs, the agent retries to connect and to
4834 register to a session daemon every 3{nbsp}seconds. You can override this
4835 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
4840 [[proc-lttng-logger-abi]]
4843 The `lttng-tracer` Linux kernel module, part of
4844 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger file
4845 path:{/proc/lttng-logger} when it's loaded. Any application can write
4846 text data to this file to emit an LTTng event.
4849 .An application writes to the LTTng logger file to emit an LTTng event.
4850 image::lttng-logger.png[]
4852 The LTTng logger is the quickest method--not the most efficient,
4853 however--to add instrumentation to an application. It is designed
4854 mostly to instrument shell scripts:
4858 echo "Some message, some $variable" > /proc/lttng-logger
4861 Any event that the LTTng logger emits is named `lttng_logger` and
4862 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
4863 other instrumentation points in the kernel tracing domain, **any Unix
4864 user** can <<enabling-disabling-events,create an event rule>> which
4865 matches its event name, not only the root user or users in the
4866 <<tracing-group,tracing group>>.
4868 To use the LTTng logger:
4870 * From any application, write text data to the path:{/proc/lttng-logger}
4873 The `msg` field of `lttng_logger` event records contains the
4876 NOTE: The maximum message length of an LTTng logger event is
4877 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
4878 than one event to contain the remaining data.
4880 You should not use the LTTng logger to trace a user application which
4881 can be instrumented in a more efficient way, namely:
4883 * <<c-application,C and $$C++$$ applications>>.
4884 * <<java-application,Java applications>>.
4885 * <<python-application,Python applications>>.
4887 .Use the LTTng logger.
4892 echo 'Hello, World!' > /proc/lttng-logger
4894 df --human-readable --print-type / > /proc/lttng-logger
4897 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4898 <<enabling-disabling-events,create an event rule>> matching the
4899 `lttng_logger` Linux kernel tracepoint, and
4900 <<basic-tracing-session-control,start tracing>>:
4905 lttng enable-event --kernel lttng_logger
4909 Run the Bash script:
4916 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4927 [[instrumenting-linux-kernel]]
4928 === LTTng kernel tracepoints
4930 NOTE: This section shows how to _add_ instrumentation points to the
4931 Linux kernel. The kernel's subsystems are already thoroughly
4932 instrumented at strategic places for LTTng when you
4933 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
4937 There are two methods to instrument the Linux kernel:
4939 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
4940 tracepoint which uses the `TRACE_EVENT()` API.
4942 Choose this if you want to instrumentation a Linux kernel tree with an
4943 instrumentation point compatible with ftrace, perf, and SystemTap.
4945 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
4946 instrument an out-of-tree kernel module.
4948 Choose this if you don't need ftrace, perf, or SystemTap support.
4952 [[linux-add-lttng-layer]]
4953 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
4955 This section shows how to add an LTTng layer to existing ftrace
4956 instrumentation using the `TRACE_EVENT()` API.
4958 This section does not document the `TRACE_EVENT()` macro. You can
4959 read the following articles to learn more about this API:
4961 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part 1)]
4962 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part 2)]
4963 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part 3)]
4965 The following procedure assumes that your ftrace tracepoints are
4966 correctly defined in their own header and that they are created in
4967 one source file using the `CREATE_TRACE_POINTS` definition.
4969 To add an LTTng layer over an existing ftrace tracepoint:
4971 . Make sure the following kernel configuration options are
4977 * `CONFIG_HIGH_RES_TIMERS`
4978 * `CONFIG_TRACEPOINTS`
4981 . Build the Linux source tree with your custom ftrace tracepoints.
4982 . Boot the resulting Linux image on your target system.
4984 Confirm that the tracepoints exist by looking for their names in the
4985 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
4986 is your subsystem's name.
4988 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
4994 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.9.tar.bz2 &&
4995 tar -xf lttng-modules-latest-2.9.tar.bz2 &&
4996 cd lttng-modules-2.9.*
5000 . In dir:{instrumentation/events/lttng-module}, relative to the root
5001 of the LTTng-modules source tree, create a header file named
5002 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5003 LTTng-modules tracepoint definitions using the LTTng-modules
5006 Start with this template:
5010 .path:{instrumentation/events/lttng-module/my_subsys.h}
5013 #define TRACE_SYSTEM my_subsys
5015 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5016 #define _LTTNG_MY_SUBSYS_H
5018 #include "../../../probes/lttng-tracepoint-event.h"
5019 #include <linux/tracepoint.h>
5021 LTTNG_TRACEPOINT_EVENT(
5023 * Format is identical to TRACE_EVENT()'s version for the three
5024 * following macro parameters:
5027 TP_PROTO(int my_int, const char *my_string),
5028 TP_ARGS(my_int, my_string),
5030 /* LTTng-modules specific macros */
5032 ctf_integer(int, my_int_field, my_int)
5033 ctf_string(my_bar_field, my_bar)
5037 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5039 #include "../../../probes/define_trace.h"
5043 The entries in the `TP_FIELDS()` section are the list of fields for the
5044 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5045 ftrace's `TRACE_EVENT()` macro.
5047 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5048 complete description of the available `ctf_*()` macros.
5050 . Create the LTTng-modules probe's kernel module C source file,
5051 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5056 .path:{probes/lttng-probe-my-subsys.c}
5058 #include <linux/module.h>
5059 #include "../lttng-tracer.h"
5062 * Build-time verification of mismatch between mainline
5063 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5064 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5066 #include <trace/events/my_subsys.h>
5068 /* Create LTTng tracepoint probes */
5069 #define LTTNG_PACKAGE_BUILD
5070 #define CREATE_TRACE_POINTS
5071 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5073 #include "../instrumentation/events/lttng-module/my_subsys.h"
5075 MODULE_LICENSE("GPL and additional rights");
5076 MODULE_AUTHOR("Your name <your-email>");
5077 MODULE_DESCRIPTION("LTTng my_subsys probes");
5078 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5079 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5080 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5081 LTTNG_MODULES_EXTRAVERSION);
5085 . Edit path:{probes/Makefile} and add your new kernel module object
5086 next to the existing ones:
5090 .path:{probes/Makefile}
5094 obj-m += lttng-probe-module.o
5095 obj-m += lttng-probe-power.o
5097 obj-m += lttng-probe-my-subsys.o
5103 . Build and install the LTTng kernel modules:
5108 make KERNELDIR=/path/to/linux
5109 sudo make modules_install
5113 Replace `/path/to/linux` with the path to the Linux source tree where
5114 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5116 Note that you can also use the
5117 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5118 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5119 C code that need to be executed before the event fields are recorded.
5121 The best way to learn how to use the previous LTTng-modules macros is to
5122 inspect the existing LTTng-modules tracepoint definitions in the
5123 dir:{instrumentation/events/lttng-module} header files. Compare them
5124 with the Linux kernel mainline versions in the
5125 dir:{include/trace/events} directory of the Linux source tree.
5129 [[lttng-tracepoint-event-code]]
5130 ===== Use custom C code to access the data for tracepoint fields
5132 Although we recommended to always use the
5133 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5134 the arguments and fields of an LTTng-modules tracepoint when possible,
5135 sometimes you need a more complex process to access the data that the
5136 tracer records as event record fields. In other words, you need local
5137 variables and multiple C{nbsp}statements instead of simple
5138 argument-based expressions that you pass to the
5139 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5141 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5142 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5143 a block of C{nbsp}code to be executed before LTTng records the fields.
5144 The structure of this macro is:
5147 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5149 LTTNG_TRACEPOINT_EVENT_CODE(
5151 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5152 * version for the following three macro parameters:
5155 TP_PROTO(int my_int, const char *my_string),
5156 TP_ARGS(my_int, my_string),
5158 /* Declarations of custom local variables */
5161 unsigned long b = 0;
5162 const char *name = "(undefined)";
5163 struct my_struct *my_struct;
5167 * Custom code which uses both tracepoint arguments
5168 * (in TP_ARGS()) and local variables (in TP_locvar()).
5170 * Local variables are actually members of a structure pointed
5171 * to by the special variable tp_locvar.
5175 tp_locvar->a = my_int + 17;
5176 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5177 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5178 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5179 put_my_struct(tp_locvar->my_struct);
5188 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5189 * version for this, except that tp_locvar members can be
5190 * used in the argument expression parameters of
5191 * the ctf_*() macros.
5194 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5195 ctf_integer(int, my_struct_a, tp_locvar->a)
5196 ctf_string(my_string_field, my_string)
5197 ctf_string(my_struct_name, tp_locvar->name)
5202 IMPORTANT: The C code defined in `TP_code()` must not have any side
5203 effects when executed. In particular, the code must not allocate
5204 memory or get resources without deallocating this memory or putting
5205 those resources afterwards.
5208 [[instrumenting-linux-kernel-tracing]]
5209 ==== Load and unload a custom probe kernel module
5211 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5212 kernel module>> in the kernel before it can emit LTTng events.
5214 To load the default probe kernel modules and a custom probe kernel
5217 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5218 probe modules to load when starting a root <<lttng-sessiond,session
5222 .Load the `my_subsys`, `usb`, and the default probe modules.
5226 sudo lttng-sessiond --extra-kmod-probes=my_subsys,usb
5231 You only need to pass the subsystem name, not the whole kernel module
5234 To load _only_ a given custom probe kernel module:
5236 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5237 modules to load when starting a root session daemon:
5240 .Load only the `my_subsys` and `usb` probe modules.
5244 sudo lttng-sessiond --kmod-probes=my_subsys,usb
5249 To confirm that a probe module is loaded:
5256 lsmod | grep lttng_probe_usb
5260 To unload the loaded probe modules:
5262 * Kill the session daemon with `SIGTERM`:
5267 sudo pkill lttng-sessiond
5271 You can also use man:modprobe(8)'s `--remove` option if the session
5272 daemon terminates abnormally.
5275 [[controlling-tracing]]
5278 Once an application or a Linux kernel is
5279 <<instrumenting,instrumented>> for LTTng tracing,
5282 This section is divided in topics on how to use the various
5283 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5284 command-line tool>>, to _control_ the LTTng daemons and tracers.
5286 NOTE: In the following subsections, we refer to an man:lttng(1) command
5287 using its man page name. For example, instead of _Run the `create`
5288 command to..._, we use _Run the man:lttng-create(1) command to..._.
5292 === Start a session daemon
5294 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5295 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5298 You will see the following error when you run a command while no session
5302 Error: No session daemon is available
5305 The only command that automatically runs a session daemon is
5306 man:lttng-create(1), which you use to
5307 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5308 this is most of the time the first operation that you do, sometimes it's
5309 not. Some examples are:
5311 * <<list-instrumentation-points,List the available instrumentation points>>.
5312 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5314 [[tracing-group]] Each Unix user must have its own running session
5315 daemon to trace user applications. The session daemon that the root user
5316 starts is the only one allowed to control the LTTng kernel tracer. Users
5317 that are part of the _tracing group_ can control the root session
5318 daemon. The default tracing group name is `tracing`; you can set it to
5319 something else with the opt:lttng-sessiond(8):--group option when you
5320 start the root session daemon.
5322 To start a user session daemon:
5324 * Run man:lttng-sessiond(8):
5329 lttng-sessiond --daemonize
5333 To start the root session daemon:
5335 * Run man:lttng-sessiond(8) as the root user:
5340 sudo lttng-sessiond --daemonize
5344 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5345 start the session daemon in foreground.
5347 To stop a session daemon, use man:kill(1) on its process ID (standard
5350 Note that some Linux distributions could manage the LTTng session daemon
5351 as a service. In this case, you should use the service manager to
5352 start, restart, and stop session daemons.
5355 [[creating-destroying-tracing-sessions]]
5356 === Create and destroy a tracing session
5358 Almost all the LTTng control operations happen in the scope of
5359 a <<tracing-session,tracing session>>, which is the dialogue between the
5360 <<lttng-sessiond,session daemon>> and you.
5362 To create a tracing session with a generated name:
5364 * Use the man:lttng-create(1) command:
5373 The created tracing session's name is `auto` followed by the
5376 To create a tracing session with a specific name:
5378 * Use the optional argument of the man:lttng-create(1) command:
5383 lttng create my-session
5387 Replace `my-session` with the specific tracing session name.
5389 LTTng appends the creation date to the created tracing session's name.
5391 LTTng writes the traces of a tracing session in
5392 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5393 name of the tracing session. Note that the env:LTTNG_HOME environment
5394 variable defaults to `$HOME` if not set.
5396 To output LTTng traces to a non-default location:
5398 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5403 lttng create --output=/tmp/some-directory my-session
5407 You may create as many tracing sessions as you wish.
5409 To list all the existing tracing sessions for your Unix user:
5411 * Use the man:lttng-list(1) command:
5420 When you create a tracing session, it is set as the _current tracing
5421 session_. The following man:lttng(1) commands operate on the current
5422 tracing session when you don't specify one:
5424 [role="list-3-cols"]
5441 To change the current tracing session:
5443 * Use the man:lttng-set-session(1) command:
5448 lttng set-session new-session
5452 Replace `new-session` by the name of the new current tracing session.
5454 When you are done tracing in a given tracing session, you can destroy
5455 it. This operation frees the resources taken by the tracing session
5456 to destroy; it does not destroy the trace data that LTTng wrote for
5457 this tracing session.
5459 To destroy the current tracing session:
5461 * Use the man:lttng-destroy(1) command:
5471 [[list-instrumentation-points]]
5472 === List the available instrumentation points
5474 The <<lttng-sessiond,session daemon>> can query the running instrumented
5475 user applications and the Linux kernel to get a list of available
5476 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5477 they are tracepoints and system calls. For the user space tracing
5478 domain, they are tracepoints. For the other tracing domains, they are
5481 To list the available instrumentation points:
5483 * Use the man:lttng-list(1) command with the requested tracing domain's
5487 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5488 must be a root user, or it must be a member of the
5489 <<tracing-group,tracing group>>).
5490 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5491 kernel system calls (your Unix user must be a root user, or it must be
5492 a member of the tracing group).
5493 * opt:lttng-list(1):--userspace: user space tracepoints.
5494 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5495 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5496 * opt:lttng-list(1):--python: Python loggers.
5499 .List the available user space tracepoints.
5503 lttng list --userspace
5507 .List the available Linux kernel system call tracepoints.
5511 lttng list --kernel --syscall
5516 [[enabling-disabling-events]]
5517 === Create and enable an event rule
5519 Once you <<creating-destroying-tracing-sessions,create a tracing
5520 session>>, you can create <<event,event rules>> with the
5521 man:lttng-enable-event(1) command.
5523 You specify each condition with a command-line option. The available
5524 condition options are shown in the following table.
5526 [role="growable",cols="asciidoc,asciidoc,default"]
5527 .Condition command-line options for the man:lttng-enable-event(1) command.
5529 |Option |Description |Applicable tracing domains
5535 . +--probe=__ADDR__+
5536 . +--function=__ADDR__+
5539 Instead of using the default _tracepoint_ instrumentation type, use:
5541 . A Linux system call.
5542 . A Linux https://lwn.net/Articles/132196/[KProbe] (symbol or address).
5543 . The entry and return points of a Linux function (symbol or address).
5547 |First positional argument.
5550 Tracepoint or system call name. In the case of a Linux KProbe or
5551 function, this is a custom name given to the event rule. With the
5552 JUL, log4j, and Python domains, this is a logger name.
5554 With a tracepoint, logger, or system call name, the last character
5555 can be `*` to match anything that remains.
5562 . +--loglevel=__LEVEL__+
5563 . +--loglevel-only=__LEVEL__+
5566 . Match only tracepoints or log statements with a logging level at
5567 least as severe as +__LEVEL__+.
5568 . Match only tracepoints or log statements with a logging level
5569 equal to +__LEVEL__+.
5571 See man:lttng-enable-event(1) for the list of available logging level
5574 |User space, JUL, log4j, and Python.
5576 |+--exclude=__EXCLUSIONS__+
5579 When you use a `*` character at the end of the tracepoint or logger
5580 name (first positional argument), exclude the specific names in the
5581 comma-delimited list +__EXCLUSIONS__+.
5584 User space, JUL, log4j, and Python.
5586 |+--filter=__EXPR__+
5589 Match only events which satisfy the expression +__EXPR__+.
5591 See man:lttng-enable-event(1) to learn more about the syntax of a
5598 You attach an event rule to a <<channel,channel>> on creation. If you do
5599 not specify the channel with the opt:lttng-enable-event(1):--channel
5600 option, and if the event rule to create is the first in its
5601 <<domain,tracing domain>> for a given tracing session, then LTTng
5602 creates a _default channel_ for you. This default channel is reused in
5603 subsequent invocations of the man:lttng-enable-event(1) command for the
5604 same tracing domain.
5606 An event rule is always enabled at creation time.
5608 The following examples show how you can combine the previous
5609 command-line options to create simple to more complex event rules.
5611 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5615 lttng enable-event --kernel sched_switch
5619 .Create an event rule matching four Linux kernel system calls (default channel).
5623 lttng enable-event --kernel --syscall open,write,read,close
5627 .Create event rules matching tracepoints with filter expressions (default channel).
5631 lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5636 lttng enable-event --kernel --all \
5637 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5642 lttng enable-event --jul my_logger \
5643 --filter='$app.retriever:cur_msg_id > 3'
5646 IMPORTANT: Make sure to always quote the filter string when you
5647 use man:lttng(1) from a shell.
5650 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5654 lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5657 IMPORTANT: Make sure to always quote the wildcard character when you
5658 use man:lttng(1) from a shell.
5661 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5665 lttng enable-event --python my-app.'*' \
5666 --exclude='my-app.module,my-app.hello'
5670 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5674 lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5678 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5682 lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5686 The event rules of a given channel form a whitelist: as soon as an
5687 emitted event passes one of them, LTTng can record the event. For
5688 example, an event named `my_app:my_tracepoint` emitted from a user space
5689 tracepoint with a `TRACE_ERROR` log level passes both of the following
5694 lttng enable-event --userspace my_app:my_tracepoint
5695 lttng enable-event --userspace my_app:my_tracepoint \
5696 --loglevel=TRACE_INFO
5699 The second event rule is redundant: the first one includes
5703 [[disable-event-rule]]
5704 === Disable an event rule
5706 To disable an event rule that you <<enabling-disabling-events,created>>
5707 previously, use the man:lttng-disable-event(1) command. This command
5708 disables _all_ the event rules (of a given tracing domain and channel)
5709 which match an instrumentation point. The other conditions are not
5710 supported as of LTTng{nbsp}{revision}.
5712 The LTTng tracer does not record an emitted event which passes
5713 a _disabled_ event rule.
5715 .Disable an event rule matching a Python logger (default channel).
5719 lttng disable-event --python my-logger
5723 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5727 lttng disable-event --jul '*'
5731 .Disable _all_ the event rules of the default channel.
5733 The opt:lttng-disable-event(1):--all-events option is not, like the
5734 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5735 equivalent of the event name `*` (wildcard): it disables _all_ the event
5736 rules of a given channel.
5740 lttng disable-event --jul --all-events
5744 NOTE: You cannot delete an event rule once you create it.
5748 === Get the status of a tracing session
5750 To get the status of the current tracing session, that is, its
5751 parameters, its channels, event rules, and their attributes:
5753 * Use the man:lttng-status(1) command:
5763 To get the status of any tracing session:
5765 * Use the man:lttng-list(1) command with the tracing session's name:
5770 lttng list my-session
5774 Replace `my-session` with the desired tracing session's name.
5777 [[basic-tracing-session-control]]
5778 === Start and stop a tracing session
5780 Once you <<creating-destroying-tracing-sessions,create a tracing
5782 <<enabling-disabling-events,create one or more event rules>>,
5783 you can start and stop the tracers for this tracing session.
5785 To start tracing in the current tracing session:
5787 * Use the man:lttng-start(1) command:
5796 LTTng is very flexible: you can launch user applications before
5797 or after the you start the tracers. The tracers only record the events
5798 if they pass enabled event rules and if they occur while the tracers are
5801 To stop tracing in the current tracing session:
5803 * Use the man:lttng-stop(1) command:
5812 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
5813 records>> or lost sub-buffers since the last time you ran
5814 man:lttng-start(1), warnings are printed when you run the
5815 man:lttng-stop(1) command.
5818 [[enabling-disabling-channels]]
5819 === Create a channel
5821 Once you create a tracing session, you can create a <<channel,channel>>
5822 with the man:lttng-enable-channel(1) command.
5824 Note that LTTng automatically creates a default channel when, for a
5825 given <<domain,tracing domain>>, no channels exist and you
5826 <<enabling-disabling-events,create>> the first event rule. This default
5827 channel is named `channel0` and its attributes are set to reasonable
5828 values. Therefore, you only need to create a channel when you need
5829 non-default attributes.
5831 You specify each non-default channel attribute with a command-line
5832 option when you use the man:lttng-enable-channel(1) command. The
5833 available command-line options are:
5835 [role="growable",cols="asciidoc,asciidoc"]
5836 .Command-line options for the man:lttng-enable-channel(1) command.
5838 |Option |Description
5844 <<channel-overwrite-mode-vs-discard-mode,event loss mode>> instead of
5845 the default _discard_ mode.
5847 |`--buffers-pid` (user space tracing domain only)
5850 Use the per-process <<channel-buffering-schemes,buffering scheme>>
5851 instead of the default per-user buffering scheme.
5853 |+--subbuf-size=__SIZE__+
5856 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
5857 either for each Unix user (default), or for each instrumented process.
5859 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5861 |+--num-subbuf=__COUNT__+
5864 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
5865 for each Unix user (default), or for each instrumented process.
5867 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
5869 |+--tracefile-size=__SIZE__+
5872 Set the maximum size of each trace file that this channel writes within
5873 a stream to +__SIZE__+ bytes instead of no maximum.
5875 See <<tracefile-rotation,Trace file count and size>>.
5877 |+--tracefile-count=__COUNT__+
5880 Limit the number of trace files that this channel creates to
5881 +__COUNT__+ channels instead of no limit.
5883 See <<tracefile-rotation,Trace file count and size>>.
5885 |+--switch-timer=__PERIODUS__+
5888 Set the <<channel-switch-timer,switch timer period>>
5889 to +__PERIODUS__+{nbsp}µs.
5891 |+--read-timer=__PERIODUS__+
5894 Set the <<channel-read-timer,read timer period>>
5895 to +__PERIODUS__+{nbsp}µs.
5897 |+--output=__TYPE__+ (Linux kernel tracing domain only)
5900 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
5904 You can only create a channel in the Linux kernel and user space
5905 <<domain,tracing domains>>: other tracing domains have their own channel
5906 created on the fly when <<enabling-disabling-events,creating event
5911 Because of a current LTTng limitation, you must create all channels
5912 _before_ you <<basic-tracing-session-control,start tracing>> in a given
5913 tracing session, that is, before the first time you run
5916 Since LTTng automatically creates a default channel when you use the
5917 man:lttng-enable-event(1) command with a specific tracing domain, you
5918 cannot, for example, create a Linux kernel event rule, start tracing,
5919 and then create a user space event rule, because no user space channel
5920 exists yet and it's too late to create one.
5922 For this reason, make sure to configure your channels properly
5923 before starting the tracers for the first time!
5926 The following examples show how you can combine the previous
5927 command-line options to create simple to more complex channels.
5929 .Create a Linux kernel channel with default attributes.
5933 lttng enable-channel --kernel my-channel
5937 .Create a user space channel with 4 sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
5941 lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
5942 --buffers-pid my-channel
5946 .Create a Linux kernel channel which rotates 8 trace files of 4{nbsp}MiB each for each stream
5950 lttng enable-channel --kernel --tracefile-count=8 \
5951 --tracefile-size=4194304 my-channel
5955 .Create a user space channel in overwrite (or _flight recorder_) mode.
5959 lttng enable-channel --userspace --overwrite my-channel
5963 You can <<enabling-disabling-events,create>> the same event rule in
5964 two different channels:
5968 lttng enable-event --userspace --channel=my-channel app:tp
5969 lttng enable-event --userspace --channel=other-channel app:tp
5972 If both channels are enabled, when a tracepoint named `app:tp` is
5973 reached, LTTng records two events, one for each channel.
5977 === Disable a channel
5979 To disable a specific channel that you <<enabling-disabling-channels,created>>
5980 previously, use the man:lttng-disable-channel(1) command.
5982 .Disable a specific Linux kernel channel.
5986 lttng disable-channel --kernel my-channel
5990 The state of a channel precedes the individual states of event rules
5991 attached to it: event rules which belong to a disabled channel, even if
5992 they are enabled, are also considered disabled.
5996 === Add context fields to a channel
5998 Event record fields in trace files provide important information about
5999 events that occured previously, but sometimes some external context may
6000 help you solve a problem faster. Examples of context fields are:
6002 * The **process ID**, **thread ID**, **process name**, and
6003 **process priority** of the thread in which the event occurs.
6004 * The **hostname** of the system on which the event occurs.
6005 * The current values of many possible **performance counters** using
6007 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6009 ** Branch instructions, misses, and loads.
6011 * Any context defined at the application level (supported for the
6012 JUL and log4j <<domain,tracing domains>>).
6014 To get the full list of available context fields, see
6015 `lttng add-context --list`. Some context fields are reserved for a
6016 specific <<domain,tracing domain>> (Linux kernel or user space).
6018 You add context fields to <<channel,channels>>. All the events
6019 that a channel with added context fields records contain those fields.
6021 To add context fields to one or all the channels of a given tracing
6024 * Use the man:lttng-add-context(1) command.
6026 .Add context fields to all the channels of the current tracing session.
6028 The following command line adds the virtual process identifier and
6029 the per-thread CPU cycles count fields to all the user space channels
6030 of the current tracing session.
6034 lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6038 .Add performance counter context fields by raw ID
6040 See man:lttng-add-context(1) for the exact format of the context field
6041 type, which is partly compatible with the format used in
6046 lttng add-context --userspace --type=perf:thread:raw:r0110:test
6047 lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6051 .Add a context field to a specific channel.
6053 The following command line adds the thread identifier context field
6054 to the Linux kernel channel named `my-channel` in the current
6059 lttng add-context --kernel --channel=my-channel --type=tid
6063 .Add an application-specific context field to a specific channel.
6065 The following command line adds the `cur_msg_id` context field of the
6066 `retriever` context retriever for all the instrumented
6067 <<java-application,Java applications>> recording <<event,event records>>
6068 in the channel named `my-channel`:
6072 lttng add-context --kernel --channel=my-channel \
6073 --type='$app:retriever:cur_msg_id'
6076 IMPORTANT: Make sure to always quote the `$` character when you
6077 use man:lttng-add-context(1) from a shell.
6080 NOTE: You cannot remove context fields from a channel once you add it.
6085 === Track process IDs
6087 It's often useful to allow only specific process IDs (PIDs) to emit
6088 events. For example, you may wish to record all the system calls made by
6089 a given process (à la http://linux.die.net/man/1/strace[strace]).
6091 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6092 purpose. Both commands operate on a whitelist of process IDs. You _add_
6093 entries to this whitelist with the man:lttng-track(1) command and remove
6094 entries with the man:lttng-untrack(1) command. Any process which has one
6095 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6096 an enabled <<event,event rule>>.
6098 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6099 process with a given tracked ID exit and another process be given this
6100 ID, then the latter would also be allowed to emit events.
6102 .Track and untrack process IDs.
6104 For the sake of the following example, assume the target system has 16
6108 <<creating-destroying-tracing-sessions,create a tracing session>>,
6109 the whitelist contains all the possible PIDs:
6112 .All PIDs are tracked.
6113 image::track-all.png[]
6115 When the whitelist is full and you use the man:lttng-track(1) command to
6116 specify some PIDs to track, LTTng first clears the whitelist, then it
6117 tracks the specific PIDs. After:
6121 lttng track --pid=3,4,7,10,13
6127 .PIDs 3, 4, 7, 10, and 13 are tracked.
6128 image::track-3-4-7-10-13.png[]
6130 You can add more PIDs to the whitelist afterwards:
6134 lttng track --pid=1,15,16
6140 .PIDs 1, 15, and 16 are added to the whitelist.
6141 image::track-1-3-4-7-10-13-15-16.png[]
6143 The man:lttng-untrack(1) command removes entries from the PID tracker's
6144 whitelist. Given the previous example, the following command:
6148 lttng untrack --pid=3,7,10,13
6151 leads to this whitelist:
6154 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6155 image::track-1-4-15-16.png[]
6157 LTTng can track all possible PIDs again using the opt:track(1):--all
6162 lttng track --pid --all
6165 The result is, again:
6168 .All PIDs are tracked.
6169 image::track-all.png[]
6172 .Track only specific PIDs
6174 A very typical use case with PID tracking is to start with an empty
6175 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6176 then add PIDs manually while tracers are active. You can accomplish this
6177 by using the opt:lttng-untrack(1):--all option of the
6178 man:lttng-untrack(1) command to clear the whitelist after you
6179 <<creating-destroying-tracing-sessions,create a tracing session>>:
6183 lttng untrack --pid --all
6189 .No PIDs are tracked.
6190 image::untrack-all.png[]
6192 If you trace with this whitelist configuration, the tracer records no
6193 events for this <<domain,tracing domain>> because no processes are
6194 tracked. You can use the man:lttng-track(1) command as usual to track
6195 specific PIDs, for example:
6199 lttng track --pid=6,11
6205 .PIDs 6 and 11 are tracked.
6206 image::track-6-11.png[]
6211 [[saving-loading-tracing-session]]
6212 === Save and load tracing session configurations
6214 Configuring a <<tracing-session,tracing session>> can be long. Some of
6215 the tasks involved are:
6217 * <<enabling-disabling-channels,Create channels>> with
6218 specific attributes.
6219 * <<adding-context,Add context fields>> to specific channels.
6220 * <<enabling-disabling-events,Create event rules>> with specific log
6221 level and filter conditions.
6223 If you use LTTng to solve real world problems, chances are you have to
6224 record events using the same tracing session setup over and over,
6225 modifying a few variables each time in your instrumented program
6226 or environment. To avoid constant tracing session reconfiguration,
6227 the man:lttng(1) command-line tool can save and load tracing session
6228 configurations to/from XML files.
6230 To save a given tracing session configuration:
6232 * Use the man:lttng-save(1) command:
6237 lttng save my-session
6241 Replace `my-session` with the name of the tracing session to save.
6243 LTTng saves tracing session configurations to
6244 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6245 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6246 the opt:lttng-save(1):--output-path option to change this destination
6249 LTTng saves all configuration parameters, for example:
6251 * The tracing session name.
6252 * The trace data output path.
6253 * The channels with their state and all their attributes.
6254 * The context fields you added to channels.
6255 * The event rules with their state, log level and filter conditions.
6257 To load a tracing session:
6259 * Use the man:lttng-load(1) command:
6264 lttng load my-session
6268 Replace `my-session` with the name of the tracing session to load.
6270 When LTTng loads a configuration, it restores your saved tracing session
6271 as if you just configured it manually.
6273 See man:lttng(1) for the complete list of command-line options. You
6274 can also save and load all many sessions at a time, and decide in which
6275 directory to output the XML files.
6278 [[sending-trace-data-over-the-network]]
6279 === Send trace data over the network
6281 LTTng can send the recorded trace data to a remote system over the
6282 network instead of writing it to the local file system.
6284 To send the trace data over the network:
6286 . On the _remote_ system (which can also be the target system),
6287 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6296 . On the _target_ system, create a tracing session configured to
6297 send trace data over the network:
6302 lttng create my-session --set-url=net://remote-system
6306 Replace `remote-system` by the host name or IP address of the
6307 remote system. See man:lttng-create(1) for the exact URL format.
6309 . On the target system, use the man:lttng(1) command-line tool as usual.
6310 When tracing is active, the target's consumer daemon sends sub-buffers
6311 to the relay daemon running on the remote system intead of flushing
6312 them to the local file system. The relay daemon writes the received
6313 packets to the local file system.
6315 The relay daemon writes trace files to
6316 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6317 +__hostname__+ is the host name of the target system and +__session__+
6318 is the tracing session name. Note that the env:LTTNG_HOME environment
6319 variable defaults to `$HOME` if not set. Use the
6320 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6321 trace files to another base directory.
6326 === View events as LTTng emits them (noch:{LTTng} live)
6328 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6329 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6330 display events as LTTng emits them on the target system while tracing is
6333 The relay daemon creates a _tee_: it forwards the trace data to both
6334 the local file system and to connected live viewers:
6337 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6342 . On the _target system_, create a <<tracing-session,tracing session>>
6348 lttng create --live my-session
6352 This spawns a local relay daemon.
6354 . Start the live viewer and configure it to connect to the relay
6355 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6360 babeltrace --input-format=lttng-live net://localhost/host/hostname/my-session
6367 * `hostname` with the host name of the target system.
6368 * `my-session` with the name of the tracing session to view.
6371 . Configure the tracing session as usual with the man:lttng(1)
6372 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6374 You can list the available live tracing sessions with Babeltrace:
6378 babeltrace --input-format=lttng-live net://localhost
6381 You can start the relay daemon on another system. In this case, you need
6382 to specify the relay daemon's URL when you create the tracing session
6383 with the opt:lttng-create(1):--set-url option. You also need to replace
6384 `localhost` in the procedure above with the host name of the system on
6385 which the relay daemon is running.
6387 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6388 command-line options.
6392 [[taking-a-snapshot]]
6393 === Take a snapshot of the current sub-buffers of a tracing session
6395 The normal behavior of LTTng is to append full sub-buffers to growing
6396 trace data files. This is ideal to keep a full history of the events
6397 that occurred on the target system, but it can
6398 represent too much data in some situations. For example, you may wish
6399 to trace your application continuously until some critical situation
6400 happens, in which case you only need the latest few recorded
6401 events to perform the desired analysis, not multi-gigabyte trace files.
6403 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6404 current sub-buffers of a given <<tracing-session,tracing session>>.
6405 LTTng can write the snapshot to the local file system or send it over
6410 . Create a tracing session in _snapshot mode_:
6415 lttng create --snapshot my-session
6419 The <<channel-overwrite-mode-vs-discard-mode,event loss mode>> of
6420 <<channel,channels>> created in this mode is automatically set to
6421 _overwrite_ (flight recorder mode).
6423 . Configure the tracing session as usual with the man:lttng(1)
6424 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6426 . **Optional**: When you need to take a snapshot,
6427 <<basic-tracing-session-control,stop tracing>>.
6429 You can take a snapshot when the tracers are active, but if you stop
6430 them first, you are sure that the data in the sub-buffers does not
6431 change before you actually take the snapshot.
6438 lttng snapshot record --name=my-first-snapshot
6442 LTTng writes the current sub-buffers of all the current tracing
6443 session's channels to trace files on the local file system. Those trace
6444 files have `my-first-snapshot` in their name.
6446 There is no difference between the format of a normal trace file and the
6447 format of a snapshot: viewers of LTTng traces also support LTTng
6450 By default, LTTng writes snapshot files to the path shown by
6451 `lttng snapshot list-output`. You can change this path or decide to send
6452 snapshots over the network using either:
6454 . An output path or URL that you specify when you create the
6456 . An snapshot output path or URL that you add using
6457 `lttng snapshot add-output`
6458 . An output path or URL that you provide directly to the
6459 `lttng snapshot record` command.
6461 Method 3 overrides method 2, which overrides method 1. When you
6462 specify a URL, a relay daemon must listen on a remote system (see
6463 <<sending-trace-data-over-the-network,Send trace data over the network>>).
6468 === Use the machine interface
6470 With any command of the man:lttng(1) command-line tool, you can set the
6471 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6472 XML machine interface output, for example:
6476 lttng --mi=xml enable-event --kernel --syscall open
6479 A schema definition (XSD) is
6480 https://github.com/lttng/lttng-tools/blob/stable-2.9/src/common/mi-lttng-3.0.xsd[available]
6481 to ease the integration with external tools as much as possible.
6485 [[metadata-regenerate]]
6486 === Regenerate the metadata of an LTTng trace
6488 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6489 data stream files and a metadata file. This metadata file contains,
6490 amongst other things, information about the offset of the clock sources
6491 used to timestamp <<event,event records>> when tracing.
6493 If, once a <<tracing-session,tracing session>> is
6494 <<basic-tracing-session-control,started>>, a major
6495 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6496 happens, the trace's clock offset also needs to be updated. You
6497 can use the `metadata` item of the man:lttng-regenerate(1) command
6500 The main use case of this command is to allow a system to boot with
6501 an incorrect wall time and trace it with LTTng before its wall time
6502 is corrected. Once the system is known to be in a state where its
6503 wall time is correct, it can run `lttng regenerate metadata`.
6505 To regenerate the metadata of an LTTng trace:
6507 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6512 lttng regenerate metadata
6518 `lttng regenerate metadata` has the following limitations:
6520 * Tracing session <<creating-destroying-tracing-sessions,created>>
6522 * User space <<channel,channels>>, if any, are using
6523 <<channel-buffering-schemes,per-user buffering>>.
6528 [[regenerate-statedump]]
6529 === Regenerate the state dump of a tracing session
6531 The LTTng kernel and user space tracers generate state dump
6532 <<event,event records>> when the application starts or when you
6533 <<basic-tracing-session-control,start a tracing session>>. An analysis
6534 can use the state dump event records to set an initial state before it
6535 builds the rest of the state from the following event records.
6536 http://tracecompass.org/[Trace Compass] is a notable example of an
6537 application which uses the state dump of an LTTng trace.
6539 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
6540 state dump event records are not included in the snapshot because they
6541 were recorded to a sub-buffer that has been consumed or overwritten
6544 You can use the `lttng regenerate statedump` command to emit the state
6545 dump event records again.
6547 To regenerate the state dump of the current tracing session, provided
6548 create it in snapshot mode, before you take a snapshot:
6550 . Use the `statedump` item of the man:lttng-regenerate(1) command:
6555 lttng regenerate statedump
6559 . <<basic-tracing-session-control,Stop the tracing session>>:
6568 . <<taking-a-snapshot,Take a snapshot>>:
6573 lttng snapshot record --name=my-snapshot
6577 Depending on the event throughput, you should run steps 1 and 2
6578 as closely as possible.
6580 NOTE: To record the state dump events, you need to
6581 <<enabling-disabling-events,create event rules>> which enable them.
6582 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
6583 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
6587 [[persistent-memory-file-systems]]
6588 === Record trace data on persistent memory file systems
6590 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
6591 (NVRAM) is random-access memory that retains its information when power
6592 is turned off (non-volatile). Systems with such memory can store data
6593 structures in RAM and retrieve them after a reboot, without flushing
6594 to typical _storage_.
6596 Linux supports NVRAM file systems thanks to either
6597 http://pramfs.sourceforge.net/[PRAMFS] or
6598 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
6599 (requires Linux 4.1+).
6601 This section does not describe how to operate such file systems;
6602 we assume that you have a working persistent memory file system.
6604 When you create a <<tracing-session,tracing session>>, you can specify
6605 the path of the shared memory holding the sub-buffers. If you specify a
6606 location on an NVRAM file system, then you can retrieve the latest
6607 recorded trace data when the system reboots after a crash.
6609 To record trace data on a persistent memory file system and retrieve the
6610 trace data after a system crash:
6612 . Create a tracing session with a sub-buffer shared memory path located
6613 on an NVRAM file system:
6618 lttng create --shm-path=/path/to/shm
6622 . Configure the tracing session as usual with the man:lttng(1)
6623 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6625 . After a system crash, use the man:lttng-crash(1) command-line tool to
6626 view the trace data recorded on the NVRAM file system:
6631 lttng-crash /path/to/shm
6635 The binary layout of the ring buffer files is not exactly the same as
6636 the trace files layout. This is why you need to use man:lttng-crash(1)
6637 instead of your preferred trace viewer directly.
6639 To convert the ring buffer files to LTTng trace files:
6641 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
6646 lttng-crash --extract=/path/to/trace /path/to/shm
6654 [[lttng-modules-ref]]
6655 === noch:{LTTng-modules}
6659 [[lttng-tracepoint-enum]]
6660 ==== `LTTNG_TRACEPOINT_ENUM()` usage
6662 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
6666 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
6671 * `name` with the name of the enumeration (C identifier, unique
6672 amongst all the defined enumerations).
6673 * `entries` with a list of enumeration entries.
6675 The available enumeration entry macros are:
6677 +ctf_enum_value(__name__, __value__)+::
6678 Entry named +__name__+ mapped to the integral value +__value__+.
6680 +ctf_enum_range(__name__, __begin__, __end__)+::
6681 Entry named +__name__+ mapped to the range of integral values between
6682 +__begin__+ (included) and +__end__+ (included).
6684 +ctf_enum_auto(__name__)+::
6685 Entry named +__name__+ mapped to the integral value following the
6686 last mapping's value.
6688 The last value of a `ctf_enum_value()` entry is its +__value__+
6691 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
6693 If `ctf_enum_auto()` is the first entry in the list, its integral
6696 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
6697 to use a defined enumeration as a tracepoint field.
6699 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
6703 LTTNG_TRACEPOINT_ENUM(
6706 ctf_enum_auto("AUTO: EXPECT 0")
6707 ctf_enum_value("VALUE: 23", 23)
6708 ctf_enum_value("VALUE: 27", 27)
6709 ctf_enum_auto("AUTO: EXPECT 28")
6710 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
6711 ctf_enum_auto("AUTO: EXPECT 304")
6719 [[lttng-modules-tp-fields]]
6720 ==== Tracepoint fields macros (for `TP_FIELDS()`)
6722 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
6723 tracepoint fields, which must be listed within `TP_FIELDS()` in
6724 `LTTNG_TRACEPOINT_EVENT()`, are:
6726 [role="func-desc growable",cols="asciidoc,asciidoc"]
6727 .Available macros to define LTTng-modules tracepoint fields
6729 |Macro |Description and parameters
6732 +ctf_integer(__t__, __n__, __e__)+
6734 +ctf_integer_nowrite(__t__, __n__, __e__)+
6736 +ctf_user_integer(__t__, __n__, __e__)+
6738 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
6740 Standard integer, displayed in base 10.
6743 Integer C type (`int`, `long`, `size_t`, ...).
6749 Argument expression.
6752 +ctf_integer_hex(__t__, __n__, __e__)+
6754 +ctf_user_integer_hex(__t__, __n__, __e__)+
6756 Standard integer, displayed in base 16.
6765 Argument expression.
6767 |+ctf_integer_oct(__t__, __n__, __e__)+
6769 Standard integer, displayed in base 8.
6778 Argument expression.
6781 +ctf_integer_network(__t__, __n__, __e__)+
6783 +ctf_user_integer_network(__t__, __n__, __e__)+
6785 Integer in network byte order (big-endian), displayed in base 10.
6794 Argument expression.
6797 +ctf_integer_network_hex(__t__, __n__, __e__)+
6799 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
6801 Integer in network byte order, displayed in base 16.
6810 Argument expression.
6813 +ctf_enum(__N__, __t__, __n__, __e__)+
6815 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
6817 +ctf_user_enum(__N__, __t__, __n__, __e__)+
6819 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
6824 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
6827 Integer C type (`int`, `long`, `size_t`, ...).
6833 Argument expression.
6836 +ctf_string(__n__, __e__)+
6838 +ctf_string_nowrite(__n__, __e__)+
6840 +ctf_user_string(__n__, __e__)+
6842 +ctf_user_string_nowrite(__n__, __e__)+
6844 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
6850 Argument expression.
6853 +ctf_array(__t__, __n__, __e__, __s__)+
6855 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
6857 +ctf_user_array(__t__, __n__, __e__, __s__)+
6859 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
6861 Statically-sized array of integers.
6864 Array element C type.
6870 Argument expression.
6876 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
6878 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6880 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
6882 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
6884 Statically-sized array of bits.
6886 The type of +__e__+ must be an integer type. +__s__+ is the number
6887 of elements of such type in +__e__+, not the number of bits.
6890 Array element C type.
6896 Argument expression.
6902 +ctf_array_text(__t__, __n__, __e__, __s__)+
6904 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
6906 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
6908 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
6910 Statically-sized array, printed as text.
6912 The string does not need to be null-terminated.
6915 Array element C type (always `char`).
6921 Argument expression.
6927 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
6929 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6931 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
6933 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
6935 Dynamically-sized array of integers.
6937 The type of +__E__+ must be unsigned.
6940 Array element C type.
6946 Argument expression.
6949 Length expression C type.
6955 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6957 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
6959 Dynamically-sized array of integers, displayed in base 16.
6961 The type of +__E__+ must be unsigned.
6964 Array element C type.
6970 Argument expression.
6973 Length expression C type.
6978 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
6980 Dynamically-sized array of integers in network byte order (big-endian),
6981 displayed in base 10.
6983 The type of +__E__+ must be unsigned.
6986 Array element C type.
6992 Argument expression.
6995 Length expression C type.
7001 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7003 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7005 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7007 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7009 Dynamically-sized array of bits.
7011 The type of +__e__+ must be an integer type. +__s__+ is the number
7012 of elements of such type in +__e__+, not the number of bits.
7014 The type of +__E__+ must be unsigned.
7017 Array element C type.
7023 Argument expression.
7026 Length expression C type.
7032 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7034 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7036 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7038 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7040 Dynamically-sized array, displayed as text.
7042 The string does not need to be null-terminated.
7044 The type of +__E__+ must be unsigned.
7046 The behaviour is undefined if +__e__+ is `NULL`.
7049 Sequence element C type (always `char`).
7055 Argument expression.
7058 Length expression C type.
7064 Use the `_user` versions when the argument expression, `e`, is
7065 a user space address. In the cases of `ctf_user_integer*()` and
7066 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7069 The `_nowrite` versions omit themselves from the session trace, but are
7070 otherwise identical. This means the `_nowrite` fields won't be written
7071 in the recorded trace. Their primary purpose is to make some
7072 of the event context available to the
7073 <<enabling-disabling-events,event filters>> without having to
7074 commit the data to sub-buffers.
7080 Terms related to LTTng and to tracing in general:
7083 The http://diamon.org/babeltrace[Babeltrace] project, which includes
7084 the cmd:babeltrace command, some libraries, and Python bindings.
7086 <<channel-buffering-schemes,buffering scheme>>::
7087 A layout of sub-buffers applied to a given channel.
7089 <<channel,channel>>::
7090 An entity which is responsible for a set of ring buffers.
7092 <<event,Event rules>> are always attached to a specific channel.
7095 A reference of time for a tracer.
7097 <<lttng-consumerd,consumer daemon>>::
7098 A process which is responsible for consuming the full sub-buffers
7099 and write them to a file system or send them over the network.
7101 <<channel-overwrite-mode-vs-discard-mode,discard mode>>:: The event loss
7102 mode in which the tracer _discards_ new event records when there's no
7103 sub-buffer space left to store them.
7106 The consequence of the execution of an instrumentation
7107 point, like a tracepoint that you manually place in some source code,
7108 or a Linux kernel KProbe.
7110 An event is said to _occur_ at a specific time. Different actions can
7111 be taken upon the occurance of an event, like record the event's payload
7114 <<channel-overwrite-mode-vs-discard-mode,event loss mode>>::
7115 The mechanism by which event records of a given channel are lost
7116 (not recorded) when there is no sub-buffer space left to store them.
7118 [[def-event-name]]event name::
7119 The name of an event, which is also the name of the event record.
7120 This is also called the _instrumentation point name_.
7123 A record, in a trace, of the payload of an event which occured.
7125 <<event,event rule>>::
7126 Set of conditions which must be satisfied for one or more occuring
7127 events to be recorded.
7129 `java.util.logging`::
7131 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7133 <<instrumenting,instrumentation>>::
7134 The use of LTTng probes to make a piece of software traceable.
7136 instrumentation point::
7137 A point in the execution path of a piece of software that, when
7138 reached by this execution, can emit an event.
7140 instrumentation point name::
7141 See _<<def-event-name,event name>>_.
7144 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7145 developed by the Apache Software Foundation.
7148 Level of severity of a log statement or user space
7149 instrumentation point.
7152 The _Linux Trace Toolkit: next generation_ project.
7154 <<lttng-cli,cmd:lttng>>::
7155 A command-line tool provided by the LTTng-tools project which you
7156 can use to send and receive control messages to and from a
7160 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7161 which is a set of analyzing programs that are used to obtain a
7162 higher level view of an LTTng trace.
7164 cmd:lttng-consumerd::
7165 The name of the consumer daemon program.
7168 A utility provided by the LTTng-tools project which can convert
7169 ring buffer files (usually
7170 <<persistent-memory-file-systems,saved on a persistent memory file system>>)
7173 LTTng Documentation::
7176 <<lttng-live,LTTng live>>::
7177 A communication protocol between the relay daemon and live viewers
7178 which makes it possible to see events "live", as they are received by
7181 <<lttng-modules,LTTng-modules>>::
7182 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7183 which contains the Linux kernel modules to make the Linux kernel
7184 instrumentation points available for LTTng tracing.
7187 The name of the relay daemon program.
7189 cmd:lttng-sessiond::
7190 The name of the session daemon program.
7193 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7194 contains the various programs and libraries used to
7195 <<controlling-tracing,control tracing>>.
7197 <<lttng-ust,LTTng-UST>>::
7198 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7199 contains libraries to instrument user applications.
7201 <<lttng-ust-agents,LTTng-UST Java agent>>::
7202 A Java package provided by the LTTng-UST project to allow the
7203 LTTng instrumentation of `java.util.logging` and Apache log4j 1.2
7206 <<lttng-ust-agents,LTTng-UST Python agent>>::
7207 A Python package provided by the LTTng-UST project to allow the
7208 LTTng instrumentation of Python logging statements.
7210 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7211 The event loss mode in which new event records overwrite older
7212 event records when there's no sub-buffer space left to store them.
7214 <<channel-buffering-schemes,per-process buffering>>::
7215 A buffering scheme in which each instrumented process has its own
7216 sub-buffers for a given user space channel.
7218 <<channel-buffering-schemes,per-user buffering>>::
7219 A buffering scheme in which all the processes of a Unix user share the
7220 same sub-buffer for a given user space channel.
7222 <<lttng-relayd,relay daemon>>::
7223 A process which is responsible for receiving the trace data sent by
7224 a distant consumer daemon.
7227 A set of sub-buffers.
7229 <<lttng-sessiond,session daemon>>::
7230 A process which receives control commands from you and orchestrates
7231 the tracers and various LTTng daemons.
7233 <<taking-a-snapshot,snapshot>>::
7234 A copy of the current data of all the sub-buffers of a given tracing
7235 session, saved as trace files.
7238 One part of an LTTng ring buffer which contains event records.
7241 The time information attached to an event when it is emitted.
7244 A set of files which are the concatenations of one or more
7245 flushed sub-buffers.
7248 The action of recording the events emitted by an application
7249 or by a system, or to initiate such recording by controlling
7253 The http://tracecompass.org[Trace Compass] project and application.
7256 An instrumentation point using the tracepoint mechanism of the Linux
7257 kernel or of LTTng-UST.
7259 tracepoint definition::
7260 The definition of a single tracepoint.
7263 The name of a tracepoint.
7265 tracepoint provider::
7266 A set of functions providing tracepoints to an instrumented user
7269 Not to be confused with a _tracepoint provider package_: many tracepoint
7270 providers can exist within a tracepoint provider package.
7272 tracepoint provider package::
7273 One or more tracepoint providers compiled as an object file or as
7277 A software which records emitted events.
7279 <<domain,tracing domain>>::
7280 A namespace for event sources.
7282 <<tracing-group,tracing group>>::
7283 The Unix group in which a Unix user can be to be allowed to trace the
7286 <<tracing-session,tracing session>>::
7287 A stateful dialogue between you and a <<lttng-sessiond,session
7291 An application running in user space, as opposed to a Linux kernel
7292 module, for example.