1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 Skip this section if you’re familiar with software tracing and with the
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 Skip this section if you already properly installed LTTng on your target
35 * **<<getting-started,Quick start>>** is a concise guide to
36 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 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 Skip this section if you don't 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{nbsp}{revision}?
77 LTTng{nbsp}{revision} bears the name _Ta Meilleure_, a Northeast IPA
78 beer brewed by https://lagabiere.com/[Lagabière]. Translating to ``Your
79 best one'', this beer gives out strong aromas of passion fruit, lemon,
80 and peaches. Tastewise, expect a lot of fruit, a creamy texture, and a
81 smooth lingering hop bitterness.
83 New features and changes in LTTng{nbsp}{revision}:
87 * Clear the contents of one or more <<tracing-session,tracing sessions>>
88 without having to destroy and reconfigure them
89 with the new man:lttng-clear(1) command.
91 This is especially useful to clear the tracing data of a tracing session
92 between attempts to reproduce a problem.
94 See <<clear,Clear a tracing session>>.
96 * Before LTTng{nbsp}{revision}, the man:lttng-track(1) and
97 man:lttng-untrack(1) commands used to add and remove process IDs
98 (PIDs) to a whitelist so that LTTng would only trace processes with
101 LTTng{nbsp}{revision} adds Unix user IDs (UIDs) and Unix group IDs
102 (GIDs) to the available <<pid-tracking,process attributes to track>>.
103 You can specify numeric user/group IDs and user/group names to track,
108 $ lttng track --userspace --vuid=http,999 --vgid=mysql,9
111 While you can also track UIDs and GIDs with the
112 opt:lttng-enable-event(1):--filter option of the `enable-event` command,
113 this dedicated process attribute tracking approach reduces tracing
114 overhead and prevents the creation of <<def-sub-buffer,sub-buffers>> for
115 the users and groups which LTTng doesn't track.
117 In the command manual pages, the term ``whitelist'' is renamed to
118 ``inclusion set'' to clarify the concept.
120 * The <<lttng-relayd,relay daemon>> can now maintain many files
121 virtually opened without using as many file descriptors (FD). It does
122 so by closing and reopening FDs as needed.
124 This feature is meant as a workaround for users who can't bump the
125 system limit because of permission restrictions.
127 The new opt:lttng-relayd(8):--fd-pool-size relay daemon option
128 sets the maximum number of simultaneously opened file descriptors
129 (using the soft `RLIMIT_NOFILE` resource limit of the process by
130 default; see man:getrlimit(2)).
132 * By default, the relay daemon writes its traces under a predefined
134 +$LTTNG_HOME/lttng-traces/__host__/__session__/__domain__+, with:
141 <<tracing-session,Tracing session>> name.
144 <<domain,Tracing domain>> name (`ust` or `kernel`).
147 Change this hierarchy to group traces by tracing session name rather
149 (+$LTTNG_HOME/lttng-traces/__session__/__host__/__domain__+) with the
150 new opt:lttng-relayd(8):--group-output-by-session option of the
153 This feature is especially useful if you're tracing two or more hosts,
154 having different hostnames, which share the same tracing session name as
155 part of their configuration. In this scenario, you can use a descriptive
156 tracing session name (for example, `connection-hang`) across a fleet of
157 machines streaming to a single relay daemon.
159 * The relay daemon has a new opt:lttng-relayd(8):--working-directory
160 option to override its working directory.
162 Linux kernel tracing::
164 * New instrumentation hooks to trace the entry and exit tracepoints of
165 the network reception code paths of the Linux kernel.
167 Use the resulting event records to identify the bounds of a network
168 reception and link the events that occur in the interim (for example,
169 wake-ups) to a specific network reception instance. You can also
170 analyze the latency of the network stack thanks to those event records.
172 * The `thread` field of the `irqaction` structure, which specifies the
173 process to wake up when a threaded interrupt request (IRQ) occurs, is
174 now part of the `lttng_statedump_interrupt` event record.
176 Use this information to discover which processes handle the various
177 IRQs. You can also associate the events occurring in the context of
178 those processes with their respective IRQ.
180 * New `lttng_statedump_cpu_topology` tracepoint to record the active
183 Use this information to discover which CPUs are SMT siblings or part of
184 the same socket. As of LTTng{nbsp}{revision}, only the x86 architecture
185 is supported since all architectures describe their topologies
188 The `architecture` field of the tracepoint is statically defined and
189 exists for all architecture implementations. Analysis tools can
190 therefore anticipate the layout of the event record.
192 Event record example:
196 lttng_statedump_cpu_topology:
202 model_name: Intel(R) Core(TM) i7-7600U CPU @ 2.80GHz
208 * New product UUID metadata environment field, `product_uuid`,
209 which LTTng copies from the
210 https://en.wikipedia.org/wiki/Desktop_Management_Interface[Desktop
211 Management Interface] (DMI).
213 Use this environment field to uniquely identify a machine (virtual or
214 physical) in order to correlate traces from multiple virtual machines.
220 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
221 generation_ is a modern toolkit for tracing Linux systems and
222 applications. So your first question might be:
229 As the history of software engineering progressed and led to what
230 we now take for granted--complex, numerous and
231 interdependent software applications running in parallel on
232 sophisticated operating systems like Linux--the authors of such
233 components, software developers, began feeling a natural
234 urge to have tools that would ensure the robustness and good performance
235 of their masterpieces.
237 One major achievement in this field is, inarguably, the
238 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
239 an essential tool for developers to find and fix bugs. But even the best
240 debugger won't help make your software run faster, and nowadays, faster
241 software means either more work done by the same hardware, or cheaper
242 hardware for the same work.
244 A _profiler_ is often the tool of choice to identify performance
245 bottlenecks. Profiling is suitable to identify _where_ performance is
246 lost in a given software. The profiler outputs a profile, a statistical
247 summary of observed events, which you may use to discover which
248 functions took the most time to execute. However, a profiler won't
249 report _why_ some identified functions are the bottleneck. Bottlenecks
250 might only occur when specific conditions are met, conditions that are
251 sometimes impossible to capture by a statistical profiler, or impossible
252 to reproduce with an application altered by the overhead of an
253 event-based profiler. For a thorough investigation of software
254 performance issues, a history of execution is essential, with the
255 recorded values of variables and context fields you choose, and
256 with as little influence as possible on the instrumented software. This
257 is where tracing comes in handy.
259 _Tracing_ is a technique used to understand what goes on in a running
260 software system. The software used for tracing is called a _tracer_,
261 which is conceptually similar to a tape recorder. When recording,
262 specific instrumentation points placed in the software source code
263 generate events that are saved on a giant tape: a _trace_ file. You
264 can trace user applications and the operating system at the same time,
265 opening the possibility of resolving a wide range of problems that would
266 otherwise be extremely challenging.
268 Tracing is often compared to _logging_. However, tracers and loggers are
269 two different tools, serving two different purposes. Tracers are
270 designed to record much lower-level events that occur much more
271 frequently than log messages, often in the range of thousands per
272 second, with very little execution overhead. Logging is more appropriate
273 for a very high-level analysis of less frequent events: user accesses,
274 exceptional conditions (errors and warnings, for example), database
275 transactions, instant messaging communications, and such. Simply put,
276 logging is one of the many use cases that can be satisfied with tracing.
278 The list of recorded events inside a trace file can be read manually
279 like a log file for the maximum level of detail, but it is generally
280 much more interesting to perform application-specific analyses to
281 produce reduced statistics and graphs that are useful to resolve a
282 given problem. Trace viewers and analyzers are specialized tools
285 In the end, this is what LTTng is: a powerful, open source set of
286 tools to trace the Linux kernel and user applications at the same time.
287 LTTng is composed of several components actively maintained and
288 developed by its link:/community/#where[community].
291 [[lttng-alternatives]]
292 === Alternatives to noch:{LTTng}
294 Excluding proprietary solutions, a few competing software tracers
297 https://github.com/dtrace4linux/linux[dtrace4linux]::
298 A port of Sun Microsystems' DTrace to Linux.
300 The cmd:dtrace tool interprets user scripts and is responsible for
301 loading code into the Linux kernel for further execution and collecting
304 https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF]::
305 A subsystem in the Linux kernel in which a virtual machine can
306 execute programs passed from the user space to the kernel.
308 You can attach such programs to tracepoints and kprobes thanks to a
309 system call, and they can output data to the user space when executed
310 thanks to different mechanisms (pipe, VM register values, and eBPF maps,
313 https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]::
314 The de facto function tracer of the Linux kernel.
316 Its user interface is a set of special files in sysfs.
318 https://perf.wiki.kernel.org/[perf]::
319 A performance analysis tool for Linux which supports hardware
320 performance counters, tracepoints, as well as other counters and
323 The controlling utility of perf is the cmd:perf command line/text UI
326 http://linux.die.net/man/1/strace[strace]::
327 A command-line utility which records system calls made by a
328 user process, as well as signal deliveries and changes of process
331 strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace] to
332 fulfill its function.
334 http://www.sysdig.org/[sysdig]::
335 Like SystemTap, uses scripts to analyze Linux kernel events.
337 You write scripts, or _chisels_ in the jargon of sysdig, in Lua and
338 sysdig executes them while it traces the system or afterwards. The
339 interface of sysdig is the cmd:sysdig command-line tool as well as the
340 text UI-based cmd:csysdig tool.
342 https://sourceware.org/systemtap/[SystemTap]::
343 A Linux kernel and user space tracer which uses custom user scripts
344 to produce plain text traces.
346 SystemTap converts the scripts to the C language, and then compiles them
347 as Linux kernel modules which are loaded to produce trace data. The
348 primary user interface of SystemTap is the cmd:stap command-line tool.
350 The main distinctive features of LTTng is that it produces correlated
351 kernel and user space traces, as well as doing so with the lowest
352 overhead amongst other solutions. It produces trace files in the
353 http://diamon.org/ctf[CTF] format, a file format optimized
354 for the production and analyses of multi-gigabyte data.
356 LTTng is the result of more than 10{nbsp}years of active open source
357 development by a community of passionate developers. LTTng is currently
358 available on major desktop and server Linux distributions.
360 The main interface for tracing control is a single command-line tool
361 named cmd:lttng. The latter can create several tracing sessions, enable
362 and disable events on the fly, filter events efficiently with custom
363 user expressions, start and stop tracing, and much more. LTTng can
364 record the traces on the file system or send them over the network, and
365 keep them totally or partially. You can view the traces once tracing
366 becomes inactive or in real-time.
368 <<installing-lttng,Install LTTng now>> and
369 <<getting-started,start tracing>>!
375 **LTTng** is a set of software <<plumbing,components>> which interact to
376 <<instrumenting,instrument>> the Linux kernel and user applications, and
377 to <<controlling-tracing,control tracing>> (start and stop
378 tracing, enable and disable event rules, and the rest). Those
379 components are bundled into the following packages:
382 Libraries and command-line interface to control tracing.
385 Linux kernel modules to instrument and trace the kernel.
388 Libraries and Java/Python packages to instrument and trace user
391 Most distributions mark the LTTng-modules and LTTng-UST packages as
392 optional when installing LTTng-tools (which is always required). In the
393 following sections, we always provide the steps to install all three,
396 * You only need to install LTTng-modules if you intend to trace the
398 * You only need to install LTTng-UST if you intend to trace user
402 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 3{nbsp}November{nbsp}2023.
404 |Distribution |Available in releases
406 |https://www.ubuntu.com/[Ubuntu]
407 |Ubuntu{nbsp}16.04 _Xenial Xerus_, Ubuntu{nbsp}18.04 _Bionic Beaver_,
408 Ubuntu{nbsp}20.04 _Focal Fossa_, and Ubuntu{nbsp}22.04 _Jammy Jellyfish_:
409 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
411 |https://www.debian.org/[Debian]
412 |<<debian,Debian{nbsp}11 _bullseye_>>.
414 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
415 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
417 |https://alpinelinux.org/[Alpine Linux]
418 |xref:alpine-linux[Alpine Linux{nbsp}3.12, Alpine Linux{nbsp}3.13,
419 Alpine Linux{nbsp}3.14, and Alpine Linux{nbsp}3.15].
421 |https://buildroot.org/[Buildroot]
422 |xref:buildroot[Buildroot{nbsp}2020.08, Buildroot{nbsp}2020.11,
423 Builroot{nbsp}2021.02, Buildroot{nbsp}2021.05, Buildroot{nbsp}2021.08,
424 and Builroot{nbsp}2021.11].
426 |https://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
427 https://www.yoctoproject.org/[Yocto]
428 |xref:oe-yocto[Yocto Project{nbsp}3.2 _Gatesgarth_ and Yocto Project{nbsp}3.3 _Hardknott_].
433 === Ubuntu: noch:{LTTng} Stable {revision} PPA
435 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
436 Stable{nbsp}{revision} PPA] offers the latest stable
437 LTTng{nbsp}{revision} packages for Ubuntu{nbsp}16.04 _Xenial Xerus_,
438 Ubuntu{nbsp}18.04 _Bionic Beaver_, Ubuntu{nbsp}20.04 _Focal Fossa_,
439 and Ubuntu{nbsp}22.04 _Jammy Jellyfish_.
441 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
444 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
450 # apt-add-repository ppa:lttng/stable-2.12
455 . Install the main LTTng{nbsp}{revision} packages:
460 # apt-get install lttng-tools
461 # apt-get install lttng-modules-dkms
462 # apt-get install liblttng-ust-dev
466 . **If you need to instrument and trace
467 <<java-application,Java applications>>**, install the LTTng-UST
473 # apt-get install liblttng-ust-agent-java
477 . **If you need to instrument and trace
478 <<python-application,Python{nbsp}3 applications>>**, install the
479 LTTng-UST Python agent:
484 # apt-get install python3-lttngust
492 To install LTTng{nbsp}{revision} on Debian{nbsp}11 _bullseye_:
494 . Install the main LTTng{nbsp}{revision} packages:
499 # apt-get install lttng-modules-dkms
500 # apt-get install liblttng-ust-dev
501 # apt-get install lttng-tools
505 . **If you need to instrument and trace <<java-application,Java
506 applications>>**, install the LTTng-UST Java agent:
511 # apt-get install liblttng-ust-agent-java
515 . **If you need to instrument and trace <<python-application,Python
516 applications>>**, install the LTTng-UST Python agent:
521 # apt-get install python3-lttngust
528 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
529 Alpine Linux{nbsp}3.12, Alpine Linux{nbsp}3.13, Alpine Linux{nbsp}3.14,
530 or Alpine Linux{nbsp}3.15:
532 . Add the LTTng packages:
537 # apk add lttng-tools
538 # apk add lttng-ust-dev
542 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
548 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
549 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
550 cd lttng-modules-2.12.* &&
552 sudo make modules_install &&
561 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2020.08, Buildroot{nbsp}2020.11,
562 Buildroot{nbsp}2021.02, Buildroot{nbsp}2021.05, Buildroot{nbsp}2021.08, or
563 Buildroot{nbsp}2021.11:
565 . Launch the Buildroot configuration tool:
574 . In **Kernel**, check **Linux kernel**.
575 . In **Toolchain**, check **Enable WCHAR support**.
576 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
577 check **lttng-modules** and **lttng-tools**.
578 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
579 **Other**, check **lttng-libust**.
583 === OpenEmbedded and Yocto
585 LTTng{nbsp}{revision} recipes are available in the
586 https://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
587 layer for Yocto Project{nbsp}3.2 _Gatesgarth_ and Yocto Project{nbsp}3.3 _Hardknott_
588 under the following names:
594 With BitBake, the simplest way to include LTTng recipes in your target
595 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
598 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
603 . Select a machine and an image recipe.
604 . Click **Edit image recipe**.
605 . Under the **All recipes** tab, search for **lttng**.
606 . Check the desired LTTng recipes.
609 [[building-from-source]]
610 === Build from source
612 To build and install LTTng{nbsp}{revision} from source:
614 . Using the package manager of your distribution, or from source,
615 install the following dependencies of LTTng-tools and LTTng-UST:
618 * https://sourceforge.net/projects/libuuid/[libuuid]
619 * http://directory.fsf.org/wiki/Popt[popt]
620 * http://liburcu.org/[Userspace RCU]
621 * http://www.xmlsoft.org/[libxml2]
622 * **Optional**: https://github.com/numactl/numactl[numactl]
625 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
631 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
632 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
633 cd lttng-modules-2.12.* &&
635 sudo make modules_install &&
640 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
646 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.12.tar.bz2 &&
647 tar -xf lttng-ust-latest-2.12.tar.bz2 &&
648 cd lttng-ust-2.12.* &&
656 Add `--disable-numa` to `./configure` if you don't have
657 https://github.com/numactl/numactl[numactl].
661 .Java and Python application tracing
663 If you need to instrument and trace <<java-application,Java
664 applications>>, pass the `--enable-java-agent-jul`,
665 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
666 `configure` script, depending on which Java logging framework you use.
668 If you need to instrument and trace <<python-application,Python
669 applications>>, pass the `--enable-python-agent` option to the
670 `configure` script. You can set the `PYTHON` environment variable to the
671 path to the Python interpreter for which to install the LTTng-UST Python
679 By default, LTTng-UST libraries are installed to
680 dir:{/usr/local/lib}, which is the de facto directory in which to
681 keep self-compiled and third-party libraries.
683 When <<building-tracepoint-providers-and-user-application,linking an
684 instrumented user application with `liblttng-ust`>>:
686 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
688 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
689 man:gcc(1), man:g++(1), or man:clang(1).
693 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
699 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
700 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
701 cd lttng-tools-2.12.* &&
709 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
710 previous steps automatically for a given version of LTTng and confine
711 the installed files in a specific directory. This can be useful to test
712 LTTng without installing it on your system.
718 This is a short guide to get started quickly with LTTng kernel and user
721 Before you follow this guide, make sure to <<installing-lttng,install>>
724 This tutorial walks you through the steps to:
726 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
727 . <<tracing-your-own-user-application,Trace a user application>> written
729 . <<viewing-and-analyzing-your-traces,View and analyze the
733 [[tracing-the-linux-kernel]]
734 === Trace the Linux kernel
736 The following command lines start with the `#` prompt because you need
737 root privileges to trace the Linux kernel. You can also trace the kernel
738 as a regular user if your Unix user is a member of the
739 <<tracing-group,tracing group>>.
741 . Create a <<tracing-session,tracing session>> which writes its traces
742 to dir:{/tmp/my-kernel-trace}:
747 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
751 . List the available kernel tracepoints and system calls:
756 # lttng list --kernel
757 # lttng list --kernel --syscall
761 . Create <<event,event rules>> which match the desired instrumentation
762 point names, for example the `sched_switch` and `sched_process_fork`
763 tracepoints, and the man:open(2) and man:close(2) system calls:
768 # lttng enable-event --kernel sched_switch,sched_process_fork
769 # lttng enable-event --kernel --syscall open,close
773 Create an event rule which matches _all_ the Linux kernel
774 tracepoints with the opt:lttng-enable-event(1):--all option
775 (this will generate a lot of data when tracing):
780 # lttng enable-event --kernel --all
784 . <<basic-tracing-session-control,Start tracing>>:
793 . Do some operation on your system for a few seconds. For example,
794 load a website, or list the files of a directory.
795 . <<creating-destroying-tracing-sessions,Destroy>> the current
805 The man:lttng-destroy(1) command doesn't destroy the trace data; it
806 only destroys the state of the tracing session.
808 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
809 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
810 session>>). You need to stop tracing to make LTTng flush the remaining
811 trace data and make the trace readable.
813 . For the sake of this example, make the recorded trace accessible to
819 # chown -R $(whoami) /tmp/my-kernel-trace
823 See <<viewing-and-analyzing-your-traces,View and analyze the
824 recorded events>> to view the recorded events.
827 [[tracing-your-own-user-application]]
828 === Trace a user application
830 This section steps you through a simple example to trace a
831 _Hello world_ program written in C.
833 To create the traceable user application:
835 . Create the tracepoint provider header file, which defines the
836 tracepoints and the events they can generate:
842 #undef TRACEPOINT_PROVIDER
843 #define TRACEPOINT_PROVIDER hello_world
845 #undef TRACEPOINT_INCLUDE
846 #define TRACEPOINT_INCLUDE "./hello-tp.h"
848 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
851 #include <lttng/tracepoint.h>
861 ctf_string(my_string_field, my_string_arg)
862 ctf_integer(int, my_integer_field, my_integer_arg)
866 #endif /* _HELLO_TP_H */
868 #include <lttng/tracepoint-event.h>
872 . Create the tracepoint provider package source file:
878 #define TRACEPOINT_CREATE_PROBES
879 #define TRACEPOINT_DEFINE
881 #include "hello-tp.h"
885 . Build the tracepoint provider package:
890 $ gcc -c -I. hello-tp.c
894 . Create the _Hello World_ application source file:
901 #include "hello-tp.h"
903 int main(int argc, char *argv[])
907 puts("Hello, World!\nPress Enter to continue...");
910 * The following getchar() call is only placed here for the purpose
911 * of this demonstration, to pause the application in order for
912 * you to have time to list its tracepoints. It's not needed
918 * A tracepoint() call.
920 * Arguments, as defined in hello-tp.h:
922 * 1. Tracepoint provider name (required)
923 * 2. Tracepoint name (required)
924 * 3. my_integer_arg (first user-defined argument)
925 * 4. my_string_arg (second user-defined argument)
927 * Notice the tracepoint provider and tracepoint names are
928 * NOT strings: they are in fact parts of variables that the
929 * macros in hello-tp.h create.
931 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
933 for (x = 0; x < argc; ++x) {
934 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
937 puts("Quitting now!");
938 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
945 . Build the application:
954 . Link the application with the tracepoint provider package,
955 `liblttng-ust`, and `libdl`:
960 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
964 Here's the whole build process:
967 .Build steps of the user space tracing tutorial.
968 image::ust-flow.png[]
970 To trace the user application:
972 . Run the application with a few arguments:
977 $ ./hello world and beyond
986 Press Enter to continue...
990 . Start an LTTng <<lttng-sessiond,session daemon>>:
995 $ lttng-sessiond --daemonize
999 Note that a session daemon might already be running, for example as a
1000 service that the service manager of the distribution started.
1002 . List the available user space tracepoints:
1007 $ lttng list --userspace
1011 You see the `hello_world:my_first_tracepoint` tracepoint listed
1012 under the `./hello` process.
1014 . Create a <<tracing-session,tracing session>>:
1019 $ lttng create my-user-space-session
1023 . Create an <<event,event rule>> which matches the
1024 `hello_world:my_first_tracepoint` event name:
1029 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1033 . <<basic-tracing-session-control,Start tracing>>:
1042 . Go back to the running `hello` application and press Enter. The
1043 program executes all `tracepoint()` instrumentation points and exits.
1044 . <<creating-destroying-tracing-sessions,Destroy>> the current
1054 The man:lttng-destroy(1) command doesn't destroy the trace data; it
1055 only destroys the state of the tracing session.
1057 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1058 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
1059 session>>). You need to stop tracing to make LTTng flush the remaining
1060 trace data and make the trace readable.
1062 By default, LTTng saves the traces in
1063 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1064 where +__name__+ is the tracing session name. The
1065 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1067 See <<viewing-and-analyzing-your-traces,View and analyze the
1068 recorded events>> to view the recorded events.
1071 [[viewing-and-analyzing-your-traces]]
1072 === View and analyze the recorded events
1074 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1075 kernel>> and <<tracing-your-own-user-application,Trace a user
1076 application>> tutorials, you can inspect the recorded events.
1078 There are many tools you can use to read LTTng traces:
1080 https://babeltrace.org/[Babeltrace{nbsp}2]::
1081 A rich, flexible trace manipulation toolkit which includes
1082 a versatile command-line interface
1083 (https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]),
1084 a https://babeltrace.org/docs/v2.0/libbabeltrace2/[C library],
1085 and https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1086 so that you can easily process or convert an LTTng trace with
1089 The Babeltrace{nbsp}2 project ships with a
1090 https://babeltrace.org/docs/v2.0/man7/babeltrace2-plugin-ctf.7/[plugin]
1091 which supports the format of the traces which LTTng produces,
1092 https://diamon.org/ctf/[CTF].
1094 http://tracecompass.org/[Trace Compass]::
1095 A graphical user interface for viewing and analyzing any type of
1096 logs or traces, including those of LTTng.
1098 https://github.com/lttng/lttng-analyses[**LTTng analyses**]::
1099 An experimental project which includes many high-level analyses of
1100 LTTng kernel traces, like scheduling statistics, interrupt
1101 frequency distribution, top CPU usage, and more.
1103 NOTE: This section assumes that LTTng saved the traces it recorded
1104 during the previous tutorials to their default location, in the
1105 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1106 environment variable defaults to `$HOME` if not set.
1109 [[viewing-and-analyzing-your-traces-bt]]
1110 ==== Use the cmd:babeltrace2 command-line tool
1112 The simplest way to list all the recorded events of an LTTng trace is to
1114 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]
1119 $ babeltrace2 ~/lttng-traces/my-user-space-session*
1122 cmd:babeltrace2 finds all traces recursively within the given path and
1123 prints all their events, sorting them chronologically.
1125 Pipe the output of cmd:babeltrace2 into a tool like man:grep(1) for
1130 $ babeltrace2 /tmp/my-kernel-trace | grep _switch
1133 Pipe the output of cmd:babeltrace2 into a tool like man:wc(1) to count
1134 the recorded events:
1138 $ babeltrace2 /tmp/my-kernel-trace | grep _open | wc --lines
1142 [[viewing-and-analyzing-your-traces-bt-python]]
1143 ==== Use the Babeltrace{nbsp}2 Python bindings
1145 The <<viewing-and-analyzing-your-traces-bt,text output of
1146 cmd:babeltrace2>> is useful to isolate events by simple matching using
1147 man:grep(1) and similar utilities. However, more elaborate filters, such
1148 as keeping only event records with a field value falling within a
1149 specific range, are not trivial to write using a shell. Moreover,
1150 reductions and even the most basic computations involving multiple event
1151 records are virtually impossible to implement.
1153 Fortunately, Babeltrace{nbsp}2 ships with
1154 https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1155 which make it easy to read the event records of an LTTng trace
1156 sequentially and compute the desired information.
1158 The following script accepts an LTTng Linux kernel trace path as its
1159 first argument and prints the short names of the top five running
1160 processes on CPU{nbsp}0 during the whole trace:
1170 # Get the trace path from the first command-line argument.
1171 it = bt2.TraceCollectionMessageIterator(sys.argv[1])
1173 # This counter dictionary will hold execution times:
1175 # Task command name -> Total execution time (ns)
1176 exec_times = collections.Counter()
1178 # This holds the last `sched_switch` timestamp.
1182 # We only care about event messages.
1183 if type(msg) is not bt2._EventMessageConst:
1186 # Event of the event message.
1189 # Keep only `sched_switch` events.
1190 if event.cls.name != 'sched_switch':
1193 # Keep only events which occurred on CPU 0.
1194 if event.packet.context_field['cpu_id'] != 0:
1197 # Event timestamp (ns).
1198 cur_ts = msg.default_clock_snapshot.ns_from_origin
1204 # (Short) name of the previous task command.
1205 prev_comm = str(event.payload_field['prev_comm'])
1207 # Initialize an entry in our dictionary if not yet done.
1208 if prev_comm not in exec_times:
1209 exec_times[prev_comm] = 0
1211 # Compute previous command execution time.
1212 diff = cur_ts - last_ts
1214 # Update execution time of this command.
1215 exec_times[prev_comm] += diff
1217 # Update last timestamp.
1221 for name, ns in exec_times.most_common(5):
1222 print('{:20}{} s'.format(name, ns / 1e9))
1225 if __name__ == '__main__':
1233 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1239 swapper/0 48.607245889 s
1240 chromium 7.192738188 s
1241 pavucontrol 0.709894415 s
1242 Compositor 0.660867933 s
1243 Xorg.bin 0.616753786 s
1246 Note that `swapper/0` is the ``idle'' process of CPU{nbsp}0 on Linux;
1247 since we weren't using the CPU that much when tracing, its first
1248 position in the list makes sense.
1252 == [[understanding-lttng]]Core concepts
1254 From a user's perspective, the LTTng system is built on a few concepts,
1255 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1256 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1257 Understanding how those objects relate to eachother is key in mastering
1260 The core concepts are:
1262 * <<tracing-session,Tracing session>>
1263 * <<domain,Tracing domain>>
1264 * <<channel,Channel and ring buffer>>
1265 * <<"event","Instrumentation point, event rule, event, and event record">>
1271 A _tracing session_ is a stateful dialogue between you and
1272 a <<lttng-sessiond,session daemon>>. You can
1273 <<creating-destroying-tracing-sessions,create a new tracing
1274 session>> with the `lttng create` command.
1276 Most of what you do when you control LTTng tracers happens within a
1277 tracing session. In particular, a tracing session:
1280 * Has its own set of trace files.
1281 * Has its own state of activity (started or stopped).
1282 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1284 * Has its own <<channel,channels>> to which are associated their own
1285 <<event,event rules>>.
1286 * Has its own <<pid-tracking,process attribute tracking>> inclusion
1290 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1291 image::concepts.png[]
1293 Those attributes and objects are completely isolated between different
1296 A tracing session is analogous to a cash machine session:
1297 the operations you do on the banking system through the cash machine do
1298 not alter the data of other users of the same system. In the case of
1299 the cash machine, a session lasts as long as your bank card is inside.
1300 In the case of LTTng, a tracing session lasts from the `lttng create`
1301 command to the `lttng destroy` command.
1304 .Each Unix user has its own set of tracing sessions.
1305 image::many-sessions.png[]
1308 [[tracing-session-mode]]
1309 ==== Tracing session mode
1311 LTTng can send the generated trace data to different locations. The
1312 _tracing session mode_ dictates where to send it. The following modes
1313 are available in LTTng{nbsp}{revision}:
1315 [[local-mode]]Local mode::
1316 LTTng writes the traces to the file system of the machine it traces
1319 [[net-streaming-mode]]Network streaming mode::
1320 LTTng sends the traces over the network to a
1321 <<lttng-relayd,relay daemon>> running on a remote system.
1324 LTTng doesn't write the traces by default.
1326 Instead, you can request LTTng to <<taking-a-snapshot,take a snapshot>>,
1327 that is, a copy of the current sub-buffers of the tracing session, and
1328 to write it to the file system of the target or to send it over the
1329 network to a <<lttng-relayd,relay daemon>> running on a remote system.
1331 [[live-mode]]Live mode::
1332 This mode is similar to the network streaming mode, but a live
1333 trace viewer can connect to the distant relay daemon to
1334 <<lttng-live,view event records as LTTng generates them>>.
1340 A _tracing domain_ is a namespace for event sources. A tracing domain
1341 has its own properties and features.
1343 There are currently five available tracing domains:
1347 * `java.util.logging` (JUL)
1351 You must specify a tracing domain when using some commands to avoid
1352 ambiguity. For example, since all the domains support named tracepoints
1353 as event sources (instrumentation points that you manually insert in the
1354 source code), you need to specify a tracing domain when
1355 <<enabling-disabling-events,creating an event rule>> because all the
1356 tracing domains could have tracepoints with the same names.
1358 You can create <<channel,channels>> in the Linux kernel and user space
1359 tracing domains. The other tracing domains have a single default
1364 === Channel and ring buffer
1366 A _channel_ is an object which is responsible for a set of ring buffers.
1367 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1368 tracer emits an event, it can record it to one or more
1369 sub-buffers. The attributes of a channel determine what to do when
1370 there's no space left for a new event record because all sub-buffers
1371 are full, where to send a full sub-buffer, and other behaviours.
1373 A channel is always associated to a <<domain,tracing domain>>. The
1374 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1375 a default channel which you can't configure.
1377 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1378 an event, it records it to the sub-buffers of all
1379 the enabled channels with a satisfied event rule, as long as those
1380 channels are part of active <<tracing-session,tracing sessions>>.
1383 [[channel-buffering-schemes]]
1384 ==== Per-user vs. per-process buffering schemes
1386 A channel has at least one ring buffer _per CPU_. LTTng always
1387 records an event to the ring buffer associated to the CPU on which it
1390 Two _buffering schemes_ are available when you
1391 <<enabling-disabling-channels,create a channel>> in the
1392 user space <<domain,tracing domain>>:
1394 Per-user buffering::
1395 Allocate one set of ring buffers--one per CPU--shared by all the
1396 instrumented processes of each Unix user.
1400 .Per-user buffering scheme.
1401 image::per-user-buffering.png[]
1404 Per-process buffering::
1405 Allocate one set of ring buffers--one per CPU--for each
1406 instrumented process.
1410 .Per-process buffering scheme.
1411 image::per-process-buffering.png[]
1414 The per-process buffering scheme tends to consume more memory than the
1415 per-user option because systems generally have more instrumented
1416 processes than Unix users running instrumented processes. However, the
1417 per-process buffering scheme ensures that one process having a high
1418 event throughput won't fill all the shared sub-buffers of the same
1421 The Linux kernel tracing domain has only one available buffering scheme
1422 which is to allocate a single set of ring buffers for the whole system.
1423 This scheme is similar to the per-user option, but with a single, global
1424 user ``running'' the kernel.
1427 [[channel-overwrite-mode-vs-discard-mode]]
1428 ==== Overwrite vs. discard event record loss modes
1430 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1431 arc in the following animations) of the ring buffer of a specific
1432 channel. When there's no space left in a sub-buffer, the tracer marks it
1433 as consumable (red) and another, empty sub-buffer starts receiving the
1434 following event records. A <<lttng-consumerd,consumer daemon>>
1435 eventually consumes the marked sub-buffer (returns to white).
1438 [role="docsvg-channel-subbuf-anim"]
1443 In an ideal world, sub-buffers are consumed faster than they are filled,
1444 as it is the case in the previous animation. In the real world,
1445 however, all sub-buffers can be full at some point, leaving no space to
1446 record the following events.
1448 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1449 no empty sub-buffer is available, it is acceptable to lose event records
1450 when the alternative would be to cause substantial delays in the
1451 execution of the instrumented application. LTTng privileges performance
1452 over integrity; it aims at perturbing the target system as little as
1453 possible in order to make tracing of subtle race conditions and rare
1454 interrupt cascades possible.
1456 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1457 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1458 example>> to learn how to use the blocking mode.
1460 When it comes to losing event records because no empty sub-buffer is
1461 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1462 reached, the _event record loss mode_ of the channel determines what to
1463 do. The available event record loss modes are:
1466 Drop the newest event records until the tracer releases a sub-buffer.
1468 This is the only available mode when you specify a
1469 <<opt-blocking-timeout,blocking timeout>>.
1472 Clear the sub-buffer containing the oldest event records and start
1473 writing the newest event records there.
1475 This mode is sometimes called _flight recorder mode_ because it's
1477 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1478 always keep a fixed amount of the latest data.
1480 Which mechanism you should choose depends on your context: prioritize
1481 the newest or the oldest event records in the ring buffer?
1483 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1484 as soon as a there's no space left for a new event record, whereas in
1485 discard mode, the tracer only discards the event record that doesn't
1488 In discard mode, LTTng increments a count of lost event records when an
1489 event record is lost and saves this count to the trace. Since
1490 LTTng{nbsp}2.8, in overwrite mode, LTTng writes to a given sub-buffer
1491 its sequence number within its data stream. With a <<local-mode,local>>,
1492 <<net-streaming-mode,network streaming>>, or <<live-mode,live>>
1493 <<tracing-session,tracing session>>, a trace reader can use such
1494 sequence numbers to report lost packets. In overwrite mode, LTTng
1495 doesn't write to the trace the exact number of lost event records in
1496 those lost sub-buffers.
1498 Trace analyses can use saved discarded event record and sub-buffer
1499 (packet) counts of the trace to decide whether or not to perform the
1500 analyses even if trace data is known to be missing.
1502 There are a few ways to decrease your probability of losing event
1504 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1505 how to fine-tune the sub-buffer count and size of a channel to virtually
1506 stop losing event records, though at the cost of greater memory usage.
1509 [[channel-subbuf-size-vs-subbuf-count]]
1510 ==== Sub-buffer count and size
1512 When you <<enabling-disabling-channels,create a channel>>, you can
1513 set its number of sub-buffers and their size.
1515 Note that there is noticeable CPU overhead introduced when
1516 switching sub-buffers (marking a full one as consumable and switching
1517 to an empty one for the following events to be recorded). Knowing this,
1518 the following list presents a few practical situations along with how
1519 to configure the sub-buffer count and size for them:
1521 * **High event throughput**: In general, prefer bigger sub-buffers to
1522 lower the risk of losing event records.
1524 Having bigger sub-buffers also ensures a lower
1525 <<channel-switch-timer,sub-buffer switching frequency>>.
1527 The number of sub-buffers is only meaningful if you create the channel
1528 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1529 other sub-buffers are left unaltered.
1531 * **Low event throughput**: In general, prefer smaller sub-buffers
1532 since the risk of losing event records is low.
1534 Because events occur less frequently, the sub-buffer switching frequency
1535 should remain low and thus the overhead of the tracer shouldn't be a
1538 * **Low memory system**: If your target system has a low memory
1539 limit, prefer fewer first, then smaller sub-buffers.
1541 Even if the system is limited in memory, you want to keep the
1542 sub-buffers as big as possible to avoid a high sub-buffer switching
1545 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1546 which means event data is very compact. For example, the average
1547 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1548 sub-buffer size of 1{nbsp}MiB is considered big.
1550 The previous situations highlight the major trade-off between a few big
1551 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1552 frequency vs. how much data is lost in overwrite mode. Assuming a
1553 constant event throughput and using the overwrite mode, the two
1554 following configurations have the same ring buffer total size:
1557 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1562 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1563 switching frequency, but if a sub-buffer overwrite happens, half of
1564 the event records so far (4{nbsp}MiB) are definitely lost.
1565 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the
1566 overhead of the tracer as the previous configuration, but if a
1567 sub-buffer overwrite happens, only the eighth of event records so far
1568 are definitely lost.
1570 In discard mode, the sub-buffers count parameter is pointless: use two
1571 sub-buffers and set their size according to the requirements of your
1575 [[channel-switch-timer]]
1576 ==== Switch timer period
1578 The _switch timer period_ is an important configurable attribute of
1579 a channel to ensure periodic sub-buffer flushing.
1581 When the _switch timer_ expires, a sub-buffer switch happens. Set
1582 the switch timer period attribute when you
1583 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1584 consumes and commits trace data to trace files or to a distant relay
1585 daemon periodically in case of a low event throughput.
1588 [role="docsvg-channel-switch-timer"]
1593 This attribute is also convenient when you use big sub-buffers to cope
1594 with a sporadic high event throughput, even if the throughput is
1598 [[channel-read-timer]]
1599 ==== Read timer period
1601 By default, the LTTng tracers use a notification mechanism to signal a
1602 full sub-buffer so that a consumer daemon can consume it. When such
1603 notifications must be avoided, for example in real-time applications,
1604 use the _read timer_ of the channel instead. When the read timer fires,
1605 the <<lttng-consumerd,consumer daemon>> checks for full, consumable
1609 [[tracefile-rotation]]
1610 ==== Trace file count and size
1612 By default, trace files can grow as large as needed. Set the maximum
1613 size of each trace file that a channel writes when you
1614 <<enabling-disabling-channels,create a channel>>. When the size of a
1615 trace file reaches the fixed maximum size of the channel, LTTng creates
1616 another file to contain the next event records. LTTng appends a file
1617 count to each trace file name in this case.
1619 If you set the trace file size attribute when you create a channel, the
1620 maximum number of trace files that LTTng creates is _unlimited_ by
1621 default. To limit them, set a maximum number of trace files. When the
1622 number of trace files reaches the fixed maximum count of the channel,
1623 the oldest trace file is overwritten. This mechanism is called _trace
1628 Even if you don't limit the trace file count, you can't assume that
1629 LTTng doesn't manage any trace file.
1631 In other words, there is no safe way to know if LTTng still holds a
1632 given trace file open with the trace file rotation feature.
1634 The only way to obtain an unmanaged, self-contained LTTng trace before
1635 you <<creating-destroying-tracing-sessions,destroy>> the tracing session
1636 is with the <<session-rotation,tracing session rotation>> feature
1637 (available since LTTng{nbsp}2.11).
1642 === Instrumentation point, event rule, event, and event record
1644 An _event rule_ is a set of conditions which must be **all** satisfied
1645 for LTTng to record an occuring event.
1647 You set the conditions when you <<enabling-disabling-events,create
1650 You always attach an event rule to a <<channel,channel>> when you create
1653 When an event passes the conditions of an event rule, LTTng records it
1654 in one of the sub-buffers of the attached channel.
1656 The available conditions, as of LTTng{nbsp}{revision}, are:
1658 * The event rule _is enabled_.
1659 * The type of the instrumentation point _is{nbsp}T_.
1660 * The name of the instrumentation point (sometimes called _event name_)
1661 _matches{nbsp}N_, but _isn't{nbsp}E_.
1662 * The log level of the instrumentation point _is as severe as{nbsp}L_, or
1663 _is exactly{nbsp}L_.
1664 * The fields of the payload of the event _satisfy_ a filter
1665 expression{nbsp}__F__.
1667 As you can see, all the conditions but the dynamic filter are related to
1668 the status of the event rule or to the instrumentation point, not to the
1669 occurring events. This is why, without a filter, checking if an event
1670 passes an event rule isn't a dynamic task: when you create or modify an
1671 event rule, all the tracers of its tracing domain enable or disable the
1672 instrumentation points themselves once. This is possible because the
1673 attributes of an instrumentation point (type, name, and log level) are
1674 defined statically. In other words, without a dynamic filter, the tracer
1675 _doesn't evaluate_ the arguments of an instrumentation point unless it
1676 matches an enabled event rule.
1678 Note that, for LTTng to record an event, the <<channel,channel>> to
1679 which a matching event rule is attached must also be enabled, and the
1680 <<tracing-session,tracing session>> owning this channel must be active
1684 .Logical path from an instrumentation point to an event record.
1685 image::event-rule.png[]
1687 .Event, event record, or event rule?
1689 With so many similar terms, it's easy to get confused.
1691 An **event** is the consequence of the execution of an _instrumentation
1692 point_, like a tracepoint that you manually place in some source code,
1693 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1694 time. Different actions can be taken upon the occurrence of an event,
1695 like record the payload of the event to a buffer.
1697 An **event record** is the representation of an event in a sub-buffer. A
1698 tracer is responsible for capturing the payload of an event, current
1699 context variables, the ID of the event, and its timestamp. LTTng
1700 can append this sub-buffer to a trace file.
1702 An **event rule** is a set of conditions which must _all_ be satisfied
1703 for LTTng to record an occuring event. Events still occur without
1704 satisfying event rules, but LTTng doesn't record them.
1709 == Components of noch:{LTTng}
1711 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1712 to call LTTng a simple _tool_ since it is composed of multiple
1713 interacting components. This section describes those components,
1714 explains their respective roles, and shows how they connect together to
1715 form the LTTng ecosystem.
1717 The following diagram shows how the most important components of LTTng
1718 interact with user applications, the Linux kernel, and you:
1721 .Control and trace data paths between LTTng components.
1722 image::plumbing.png[]
1724 The LTTng project incorporates:
1726 * **LTTng-tools**: Libraries and command-line interface to
1727 control tracing sessions.
1728 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1729 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1730 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1731 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1732 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1733 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1735 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1736 headers to instrument and trace any native user application.
1737 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1738 *** `liblttng-ust-libc-wrapper`
1739 *** `liblttng-ust-pthread-wrapper`
1740 *** `liblttng-ust-cyg-profile`
1741 *** `liblttng-ust-cyg-profile-fast`
1742 *** `liblttng-ust-dl`
1743 ** User space tracepoint provider source files generator command-line
1744 tool (man:lttng-gen-tp(1)).
1745 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1746 Java applications using `java.util.logging` or
1747 Apache log4j{nbsp}1.2 logging.
1748 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1749 Python applications using the standard `logging` package.
1750 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1752 ** LTTng kernel tracer module.
1753 ** Tracing ring buffer kernel modules.
1754 ** Probe kernel modules.
1755 ** LTTng logger kernel module.
1759 === Tracing control command-line interface
1762 .The tracing control command-line interface.
1763 image::plumbing-lttng-cli.png[]
1765 The _man:lttng(1) command-line tool_ is the standard user interface to
1766 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1767 is part of LTTng-tools.
1769 The cmd:lttng tool is linked with
1770 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1771 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1773 The cmd:lttng tool has a Git-like interface:
1777 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1780 The <<controlling-tracing,Tracing control>> section explores the
1781 available features of LTTng using the cmd:lttng tool.
1784 [[liblttng-ctl-lttng]]
1785 === Tracing control library
1788 .The tracing control library.
1789 image::plumbing-liblttng-ctl.png[]
1791 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1792 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1793 underlying details of the protocol. `liblttng-ctl` is part of LTTng-tools.
1795 The <<lttng-cli,cmd:lttng command-line tool>>
1796 is linked with `liblttng-ctl`.
1798 Use `liblttng-ctl` in C or $$C++$$ source code by including its
1803 #include <lttng/lttng.h>
1806 Some objects are referenced by name (C string), such as tracing
1807 sessions, but most of them require to create a handle first using
1808 `lttng_create_handle()`.
1810 As of LTTng{nbsp}{revision}, the best available developer documentation for
1811 `liblttng-ctl` is its installed header files. Every function and structure is
1812 thoroughly documented.
1816 === User space tracing library
1819 .The user space tracing library.
1820 image::plumbing-liblttng-ust.png[]
1822 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1823 is the LTTng user space tracer. It receives commands from a
1824 <<lttng-sessiond,session daemon>>, for example to
1825 enable and disable specific instrumentation points, and writes event
1826 records to ring buffers shared with a
1827 <<lttng-consumerd,consumer daemon>>.
1828 `liblttng-ust` is part of LTTng-UST.
1830 Public C header files are installed beside `liblttng-ust` to
1831 instrument any <<c-application,C or $$C++$$ application>>.
1833 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1834 packages, use their own library providing tracepoints which is
1835 linked with `liblttng-ust`.
1837 An application or library doesn't have to initialize `liblttng-ust`
1838 manually: its constructor does the necessary tasks to properly register
1839 to a session daemon. The initialization phase also enables the
1840 instrumentation points matching the <<event,event rules>> that you
1844 [[lttng-ust-agents]]
1845 === User space tracing agents
1848 .The user space tracing agents.
1849 image::plumbing-lttng-ust-agents.png[]
1851 The _LTTng-UST Java and Python agents_ are regular Java and Python
1852 packages which add LTTng tracing capabilities to the
1853 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1855 In the case of Java, the
1856 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1857 core logging facilities] and
1858 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1859 Note that Apache Log4{nbsp}2 isn't supported.
1861 In the case of Python, the standard
1862 https://docs.python.org/3/library/logging.html[`logging`] package
1863 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1864 LTTng-UST Python agent package.
1866 The applications using the LTTng-UST agents are in the
1867 `java.util.logging` (JUL),
1868 log4j, and Python <<domain,tracing domains>>.
1870 Both agents use the same mechanism to trace the log statements. When an
1871 agent initializes, it creates a log handler that attaches to the root
1872 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1873 When the application executes a log statement, the root logger passes it
1874 to the log handler of the agent. The log handler of the agent calls a
1875 native function in a tracepoint provider package shared library linked
1876 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1877 other fields, like its logger name and its log level. This native
1878 function contains a user space instrumentation point, hence tracing the
1881 The log level condition of an
1882 <<event,event rule>> is considered when tracing
1883 a Java or a Python application, and it's compatible with the standard
1884 JUL, log4j, and Python log levels.
1888 === LTTng kernel modules
1891 .The LTTng kernel modules.
1892 image::plumbing-lttng-modules.png[]
1894 The _LTTng kernel modules_ are a set of Linux kernel modules
1895 which implement the kernel tracer of the LTTng project. The LTTng
1896 kernel modules are part of LTTng-modules.
1898 The LTTng kernel modules include:
1900 * A set of _probe_ modules.
1902 Each module attaches to a specific subsystem
1903 of the Linux kernel using its tracepoint instrument points. There are
1904 also modules to attach to the entry and return points of the Linux
1905 system call functions.
1907 * _Ring buffer_ modules.
1909 A ring buffer implementation is provided as kernel modules. The LTTng
1910 kernel tracer writes to the ring buffer; a
1911 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1913 * The _LTTng kernel tracer_ module.
1914 * The _LTTng logger_ module.
1916 The LTTng logger module implements the special path:{/proc/lttng-logger}
1917 (and path:{/dev/lttng-logger} since LTTng{nbsp}2.11) files so that any
1918 executable can generate LTTng events by opening and writing to those
1921 See <<proc-lttng-logger-abi,LTTng logger>>.
1923 Generally, you don't have to load the LTTng kernel modules manually
1924 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1925 daemon>> loads the necessary modules when starting. If you have extra
1926 probe modules, you can specify to load them to the session daemon on
1929 The LTTng kernel modules are installed in
1930 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1931 the kernel release (see `uname --kernel-release`).
1938 .The session daemon.
1939 image::plumbing-sessiond.png[]
1941 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1942 managing tracing sessions and for controlling the various components of
1943 LTTng. The session daemon is part of LTTng-tools.
1945 The session daemon sends control requests to and receives control
1948 * The <<lttng-ust,user space tracing library>>.
1950 Any instance of the user space tracing library first registers to
1951 a session daemon. Then, the session daemon can send requests to
1952 this instance, such as:
1955 ** Get the list of tracepoints.
1956 ** Share an <<event,event rule>> so that the user space tracing library
1957 can enable or disable tracepoints. Amongst the possible conditions
1958 of an event rule is a filter expression which `liblttng-ust` evalutes
1959 when an event occurs.
1960 ** Share <<channel,channel>> attributes and ring buffer locations.
1963 The session daemon and the user space tracing library use a Unix
1964 domain socket for their communication.
1966 * The <<lttng-ust-agents,user space tracing agents>>.
1968 Any instance of a user space tracing agent first registers to
1969 a session daemon. Then, the session daemon can send requests to
1970 this instance, such as:
1973 ** Get the list of loggers.
1974 ** Enable or disable a specific logger.
1977 The session daemon and the user space tracing agent use a TCP connection
1978 for their communication.
1980 * The <<lttng-modules,LTTng kernel tracer>>.
1981 * The <<lttng-consumerd,consumer daemon>>.
1983 The session daemon sends requests to the consumer daemon to instruct
1984 it where to send the trace data streams, amongst other information.
1986 * The <<lttng-relayd,relay daemon>>.
1988 The session daemon receives commands from the
1989 <<liblttng-ctl-lttng,tracing control library>>.
1991 The root session daemon loads the appropriate
1992 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1993 a <<lttng-consumerd,consumer daemon>> as soon as you create
1994 an <<event,event rule>>.
1996 The session daemon doesn't send and receive trace data: this is the
1997 role of the <<lttng-consumerd,consumer daemon>> and
1998 <<lttng-relayd,relay daemon>>. It does, however, generate the
1999 http://diamon.org/ctf/[CTF] metadata stream.
2001 Each Unix user can have its own session daemon instance. The
2002 tracing sessions which different session daemons manage are completely
2005 The root user's session daemon is the only one which is
2006 allowed to control the LTTng kernel tracer, and its spawned consumer
2007 daemon is the only one which is allowed to consume trace data from the
2008 LTTng kernel tracer. Note, however, that any Unix user which is a member
2009 of the <<tracing-group,tracing group>> is allowed
2010 to create <<channel,channels>> in the
2011 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2014 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2015 session daemon when using its `create` command if none is currently
2016 running. You can also start the session daemon manually.
2023 .The consumer daemon.
2024 image::plumbing-consumerd.png[]
2026 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
2027 ring buffers with user applications or with the LTTng kernel modules to
2028 collect trace data and send it to some location (on disk or to a
2029 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2030 is part of LTTng-tools.
2032 You don't start a consumer daemon manually: a consumer daemon is always
2033 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2034 <<event,event rule>>, that is, before you start tracing. When you kill
2035 its owner session daemon, the consumer daemon also exits because it is
2036 the child process of the session daemon. Command-line options of
2037 man:lttng-sessiond(8) target the consumer daemon process.
2039 There are up to two running consumer daemons per Unix user, whereas only
2040 one session daemon can run per user. This is because each process can be
2041 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2042 and 64-bit processes, it is more efficient to have separate
2043 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2044 exception: it can have up to _three_ running consumer daemons: 32-bit
2045 and 64-bit instances for its user applications, and one more
2046 reserved for collecting kernel trace data.
2054 image::plumbing-relayd.png[]
2056 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2057 between remote session and consumer daemons, local trace files, and a
2058 remote live trace viewer. The relay daemon is part of LTTng-tools.
2060 The main purpose of the relay daemon is to implement a receiver of
2061 <<sending-trace-data-over-the-network,trace data over the network>>.
2062 This is useful when the target system doesn't have much file system
2063 space to record trace files locally.
2065 The relay daemon is also a server to which a
2066 <<lttng-live,live trace viewer>> can
2067 connect. The live trace viewer sends requests to the relay daemon to
2068 receive trace data as the target system emits events. The
2069 communication protocol is named _LTTng live_; it is used over TCP
2072 Note that you can start the relay daemon on the target system directly.
2073 This is the setup of choice when the use case is to view events as
2074 the target system emits them without the need of a remote system.
2078 == [[using-lttng]]Instrumentation
2080 There are many examples of tracing and monitoring in our everyday life:
2082 * You have access to real-time and historical weather reports and
2083 forecasts thanks to weather stations installed around the country.
2084 * You know your heart is safe thanks to an electrocardiogram.
2085 * You make sure not to drive your car too fast and to have enough fuel
2086 to reach your destination thanks to gauges visible on your dashboard.
2088 All the previous examples have something in common: they rely on
2089 **instruments**. Without the electrodes attached to the surface of your
2090 body skin, cardiac monitoring is futile.
2092 LTTng, as a tracer, is no different from those real life examples. If
2093 you're about to trace a software system or, in other words, record its
2094 history of execution, you better have **instrumentation points** in the
2095 subject you're tracing, that is, the actual software.
2097 Various ways were developed to instrument a piece of software for LTTng
2098 tracing. The most straightforward one is to manually place
2099 instrumentation points, called _tracepoints_, in the source code of the
2100 software. It is also possible to add instrumentation points dynamically
2101 in the Linux kernel <<domain,tracing domain>>.
2103 If you're only interested in tracing the Linux kernel, your
2104 instrumentation needs are probably already covered by the built-in
2105 <<lttng-modules,Linux kernel tracepoints>> of LTTng. You may also wish
2106 to trace a user application which is already instrumented for LTTng
2107 tracing. In such cases, skip this whole section and read the topics of
2108 the <<controlling-tracing,Tracing control>> section.
2110 Many methods are available to instrument a piece of software for LTTng
2113 * <<c-application,User space instrumentation for C and $$C++$$
2115 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2116 * <<java-application,User space Java agent>>.
2117 * <<python-application,User space Python agent>>.
2118 * <<proc-lttng-logger-abi,LTTng logger>>.
2119 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2123 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2125 The procedure to instrument a C or $$C++$$ user application with
2126 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2128 . <<tracepoint-provider,Create the source files of a tracepoint provider
2130 . <<probing-the-application-source-code,Add tracepoints to
2131 the source code of the application>>.
2132 . <<building-tracepoint-providers-and-user-application,Build and link
2133 a tracepoint provider package and the user application>>.
2135 If you need quick, man:printf(3)-like instrumentation, skip
2136 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2139 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2140 instrument a user application with `liblttng-ust`.
2143 [[tracepoint-provider]]
2144 ==== Create the source files of a tracepoint provider package
2146 A _tracepoint provider_ is a set of compiled functions which provide
2147 **tracepoints** to an application, the type of instrumentation point
2148 supported by LTTng-UST. Those functions can emit events with
2149 user-defined fields and serialize those events as event records to one
2150 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2151 macro, which you <<probing-the-application-source-code,insert in the
2152 source code of a user application>>, calls those functions.
2154 A _tracepoint provider package_ is an object file (`.o`) or a shared
2155 library (`.so`) which contains one or more tracepoint providers.
2156 Its source files are:
2158 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2159 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2161 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2162 the LTTng user space tracer, at run time.
2165 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2166 image::ust-app.png[]
2168 NOTE: If you need quick, man:printf(3)-like instrumentation,
2169 skip creating and using a tracepoint provider and use
2170 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2174 ===== Create a tracepoint provider header file template
2176 A _tracepoint provider header file_ contains the tracepoint
2177 definitions of a tracepoint provider.
2179 To create a tracepoint provider header file:
2181 . Start from this template:
2185 .Tracepoint provider header file template (`.h` file extension).
2187 #undef TRACEPOINT_PROVIDER
2188 #define TRACEPOINT_PROVIDER provider_name
2190 #undef TRACEPOINT_INCLUDE
2191 #define TRACEPOINT_INCLUDE "./tp.h"
2193 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2196 #include <lttng/tracepoint.h>
2199 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2200 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2205 #include <lttng/tracepoint-event.h>
2211 * `provider_name` with the name of your tracepoint provider.
2212 * `"tp.h"` with the name of your tracepoint provider header file.
2214 . Below the `#include <lttng/tracepoint.h>` line, put your
2215 <<defining-tracepoints,tracepoint definitions>>.
2217 Your tracepoint provider name must be unique amongst all the possible
2218 tracepoint provider names used on the same target system. We
2219 suggest to include the name of your project or company in the name,
2220 for example, `org_lttng_my_project_tpp`.
2222 TIP: [[lttng-gen-tp]]Use the man:lttng-gen-tp(1) tool to create
2223 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2224 write are the <<defining-tracepoints,tracepoint definitions>>.
2227 [[defining-tracepoints]]
2228 ===== Create a tracepoint definition
2230 A _tracepoint definition_ defines, for a given tracepoint:
2232 * Its **input arguments**. They are the macro parameters that the
2233 `tracepoint()` macro accepts for this particular tracepoint
2234 in the source code of the user application.
2235 * Its **output event fields**. They are the sources of event fields
2236 that form the payload of any event that the execution of the
2237 `tracepoint()` macro emits for this particular tracepoint.
2239 Create a tracepoint definition by using the
2240 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2242 <<tpp-header,tracepoint provider header file template>>.
2244 The syntax of the `TRACEPOINT_EVENT()` macro is:
2247 .`TRACEPOINT_EVENT()` macro syntax.
2250 /* Tracepoint provider name */
2253 /* Tracepoint name */
2256 /* Input arguments */
2261 /* Output event fields */
2270 * `provider_name` with your tracepoint provider name.
2271 * `tracepoint_name` with your tracepoint name.
2272 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2273 * `fields` with the <<tpp-def-output-fields,output event field>>
2276 This tracepoint emits events named `provider_name:tracepoint_name`.
2279 .Event name length limitation
2281 The concatenation of the tracepoint provider name and the
2282 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2283 instrumented application compiles and runs, but LTTng throws multiple
2284 warnings and you could experience serious issues.
2287 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2290 .`TP_ARGS()` macro syntax.
2299 * `type` with the C type of the argument.
2300 * `arg_name` with the argument name.
2302 You can repeat `type` and `arg_name` up to 10{nbsp}times to have more
2305 .`TP_ARGS()` usage with three arguments.
2317 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2318 tracepoint definition with no input arguments.
2320 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2321 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2322 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2323 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2326 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2327 C expression that the tracer evalutes at the `tracepoint()` macro site
2328 in the source code of the application. This expression provides the
2329 source of data of a field. The argument expression can include input
2330 argument names listed in the `TP_ARGS()` macro.
2332 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2333 must be unique within a given tracepoint definition.
2335 Here's a complete tracepoint definition example:
2337 .Tracepoint definition.
2339 The following tracepoint definition defines a tracepoint which takes
2340 three input arguments and has four output event fields.
2344 #include "my-custom-structure.h"
2350 const struct my_custom_structure*, my_custom_structure,
2355 ctf_string(query_field, query)
2356 ctf_float(double, ratio_field, ratio)
2357 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2358 ctf_integer(int, send_size, my_custom_structure->send_size)
2363 Refer to this tracepoint definition with the `tracepoint()` macro in
2364 the source code of your application like this:
2368 tracepoint(my_provider, my_tracepoint,
2369 my_structure, some_ratio, the_query);
2373 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2374 if they satisfy an enabled <<event,event rule>>.
2377 [[using-tracepoint-classes]]
2378 ===== Use a tracepoint class
2380 A _tracepoint class_ is a class of tracepoints which share the same
2381 output event field definitions. A _tracepoint instance_ is one
2382 instance of such a defined tracepoint class, with its own tracepoint
2385 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2386 shorthand which defines both a tracepoint class and a tracepoint
2387 instance at the same time.
2389 When you build a tracepoint provider package, the C or $$C++$$ compiler
2390 creates one serialization function for each **tracepoint class**. A
2391 serialization function is responsible for serializing the event fields
2392 of a tracepoint to a sub-buffer when tracing.
2394 For various performance reasons, when your situation requires multiple
2395 tracepoint definitions with different names, but with the same event
2396 fields, we recommend that you manually create a tracepoint class
2397 and instantiate as many tracepoint instances as needed. One positive
2398 effect of such a design, amongst other advantages, is that all
2399 tracepoint instances of the same tracepoint class reuse the same
2400 serialization function, thus reducing
2401 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2403 .Use a tracepoint class and tracepoint instances.
2405 Consider the following three tracepoint definitions:
2417 ctf_integer(int, userid, userid)
2418 ctf_integer(size_t, len, len)
2430 ctf_integer(int, userid, userid)
2431 ctf_integer(size_t, len, len)
2443 ctf_integer(int, userid, userid)
2444 ctf_integer(size_t, len, len)
2449 In this case, we create three tracepoint classes, with one implicit
2450 tracepoint instance for each of them: `get_account`, `get_settings`, and
2451 `get_transaction`. However, they all share the same event field names
2452 and types. Hence three identical, yet independent serialization
2453 functions are created when you build the tracepoint provider package.
2455 A better design choice is to define a single tracepoint class and three
2456 tracepoint instances:
2460 /* The tracepoint class */
2461 TRACEPOINT_EVENT_CLASS(
2462 /* Tracepoint provider name */
2465 /* Tracepoint class name */
2468 /* Input arguments */
2474 /* Output event fields */
2476 ctf_integer(int, userid, userid)
2477 ctf_integer(size_t, len, len)
2481 /* The tracepoint instances */
2482 TRACEPOINT_EVENT_INSTANCE(
2483 /* Tracepoint provider name */
2486 /* Tracepoint class name */
2489 /* Tracepoint name */
2492 /* Input arguments */
2498 TRACEPOINT_EVENT_INSTANCE(
2507 TRACEPOINT_EVENT_INSTANCE(
2520 [[assigning-log-levels]]
2521 ===== Assign a log level to a tracepoint definition
2523 Assign a _log level_ to a <<defining-tracepoints,tracepoint definition>>
2524 with the `TRACEPOINT_LOGLEVEL()` macro.
2526 Assigning different levels of severity to tracepoint definitions can
2527 be useful: when you <<enabling-disabling-events,create an event rule>>,
2528 you can target tracepoints having a log level as severe as a specific
2531 The concept of LTTng-UST log levels is similar to the levels found
2532 in typical logging frameworks:
2534 * In a logging framework, the log level is given by the function
2535 or method name you use at the log statement site: `debug()`,
2536 `info()`, `warn()`, `error()`, and so on.
2537 * In LTTng-UST, you statically assign the log level to a tracepoint
2538 definition; any `tracepoint()` macro invocation which refers to
2539 this definition has this log level.
2541 You must use `TRACEPOINT_LOGLEVEL()` _after_ the
2542 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2543 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2546 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2549 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2551 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2556 * `provider_name` with the tracepoint provider name.
2557 * `tracepoint_name` with the tracepoint name.
2558 * `log_level` with the log level to assign to the tracepoint
2559 definition named `tracepoint_name` in the `provider_name`
2560 tracepoint provider.
2562 See man:lttng-ust(3) for a list of available log level names.
2564 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2568 /* Tracepoint definition */
2577 ctf_integer(int, userid, userid)
2578 ctf_integer(size_t, len, len)
2582 /* Log level assignment */
2583 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2589 ===== Create a tracepoint provider package source file
2591 A _tracepoint provider package source file_ is a C source file which
2592 includes a <<tpp-header,tracepoint provider header file>> to expand its
2593 macros into event serialization and other functions.
2595 Use the following tracepoint provider package source file template:
2598 .Tracepoint provider package source file template.
2600 #define TRACEPOINT_CREATE_PROBES
2605 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2606 header file>> name. You may also include more than one tracepoint
2607 provider header file here to create a tracepoint provider package
2608 holding more than one tracepoint providers.
2611 [[probing-the-application-source-code]]
2612 ==== Add tracepoints to the source code of an application
2614 Once you <<tpp-header,create a tracepoint provider header file>>, use
2615 the `tracepoint()` macro in the source code of your application to
2616 insert the tracepoints that this header
2617 <<defining-tracepoints,defines>>.
2619 The `tracepoint()` macro takes at least two parameters: the tracepoint
2620 provider name and the tracepoint name. The corresponding tracepoint
2621 definition defines the other parameters.
2623 .`tracepoint()` usage.
2625 The following <<defining-tracepoints,tracepoint definition>> defines a
2626 tracepoint which takes two input arguments and has two output event
2630 .Tracepoint provider header file.
2632 #include "my-custom-structure.h"
2639 const char*, cmd_name
2642 ctf_string(cmd_name, cmd_name)
2643 ctf_integer(int, number_of_args, argc)
2648 Refer to this tracepoint definition with the `tracepoint()` macro in
2649 the source code of your application like this:
2652 .Application source file.
2656 int main(int argc, char* argv[])
2658 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2664 Note how the source code of the application includes
2665 the tracepoint provider header file containing the tracepoint
2666 definitions to use, path:{tp.h}.
2669 .`tracepoint()` usage with a complex tracepoint definition.
2671 Consider this complex tracepoint definition, where multiple event
2672 fields refer to the same input arguments in their argument expression
2676 .Tracepoint provider header file.
2678 /* For `struct stat` */
2679 #include <sys/types.h>
2680 #include <sys/stat.h>
2692 ctf_integer(int, my_constant_field, 23 + 17)
2693 ctf_integer(int, my_int_arg_field, my_int_arg)
2694 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2695 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2696 my_str_arg[2] + my_str_arg[3])
2697 ctf_string(my_str_arg_field, my_str_arg)
2698 ctf_integer_hex(off_t, size_field, st->st_size)
2699 ctf_float(double, size_dbl_field, (double) st->st_size)
2700 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2701 size_t, strlen(my_str_arg) / 2)
2706 Refer to this tracepoint definition with the `tracepoint()` macro in
2707 the source code of your application like this:
2710 .Application source file.
2712 #define TRACEPOINT_DEFINE
2719 stat("/etc/fstab", &s);
2720 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2726 If you look at the event record that LTTng writes when tracing this
2727 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2728 it should look like this:
2730 .Event record fields
2732 |Field name |Field value
2733 |`my_constant_field` |40
2734 |`my_int_arg_field` |23
2735 |`my_int_arg_field2` |529
2737 |`my_str_arg_field` |`Hello, World!`
2738 |`size_field` |0x12d
2739 |`size_dbl_field` |301.0
2740 |`half_my_str_arg_field` |`Hello,`
2744 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2745 compute--they use the call stack, for example. To avoid this computation
2746 when the tracepoint is disabled, use the `tracepoint_enabled()` and
2747 `do_tracepoint()` macros.
2749 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2753 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2755 tracepoint_enabled(provider_name, tracepoint_name)
2756 do_tracepoint(provider_name, tracepoint_name, ...)
2761 * `provider_name` with the tracepoint provider name.
2762 * `tracepoint_name` with the tracepoint name.
2764 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2765 `tracepoint_name` from the provider named `provider_name` is enabled
2768 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2769 if the tracepoint is enabled. Using `tracepoint()` with
2770 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2771 the `tracepoint_enabled()` check, thus a race condition is
2772 possible in this situation:
2775 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2777 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2778 stuff = prepare_stuff();
2781 tracepoint(my_provider, my_tracepoint, stuff);
2784 If the tracepoint is enabled after the condition, then `stuff` isn't
2785 prepared: the emitted event will either contain wrong data, or the whole
2786 application could crash (segmentation fault, for example).
2788 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2789 `STAP_PROBEV()` call. If you need it, you must emit
2793 [[building-tracepoint-providers-and-user-application]]
2794 ==== Build and link a tracepoint provider package and an application
2796 Once you have one or more <<tpp-header,tracepoint provider header
2797 files>> and a <<tpp-source,tracepoint provider package source file>>,
2798 create the tracepoint provider package by compiling its source
2799 file. From here, multiple build and run scenarios are possible. The
2800 following table shows common application and library configurations
2801 along with the required command lines to achieve them.
2803 In the following diagrams, we use the following file names:
2806 Executable application.
2809 Application object file.
2812 Tracepoint provider package object file.
2815 Tracepoint provider package archive file.
2818 Tracepoint provider package shared object file.
2821 User library object file.
2824 User library shared object file.
2826 We use the following symbols in the diagrams of table below:
2829 .Symbols used in the build scenario diagrams.
2830 image::ust-sit-symbols.png[]
2832 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2833 variable in the following instructions.
2835 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2836 .Common tracepoint provider package scenarios.
2838 |Scenario |Instructions
2841 The instrumented application is statically linked with
2842 the tracepoint provider package object.
2844 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2847 include::../common/ust-sit-step-tp-o.txt[]
2849 To build the instrumented application:
2851 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2856 #define TRACEPOINT_DEFINE
2860 . Compile the application source file:
2869 . Build the application:
2874 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2878 To run the instrumented application:
2880 * Start the application:
2890 The instrumented application is statically linked with the
2891 tracepoint provider package archive file.
2893 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2896 To create the tracepoint provider package archive file:
2898 . Compile the <<tpp-source,tracepoint provider package source file>>:
2907 . Create the tracepoint provider package archive file:
2912 $ ar rcs tpp.a tpp.o
2916 To build the instrumented application:
2918 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2923 #define TRACEPOINT_DEFINE
2927 . Compile the application source file:
2936 . Build the application:
2941 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2945 To run the instrumented application:
2947 * Start the application:
2957 The instrumented application is linked with the tracepoint provider
2958 package shared object.
2960 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2963 include::../common/ust-sit-step-tp-so.txt[]
2965 To build the instrumented application:
2967 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2972 #define TRACEPOINT_DEFINE
2976 . Compile the application source file:
2985 . Build the application:
2990 $ gcc -o app app.o -ldl -L. -ltpp
2994 To run the instrumented application:
2996 * Start the application:
3006 The tracepoint provider package shared object is preloaded before the
3007 instrumented application starts.
3009 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3012 include::../common/ust-sit-step-tp-so.txt[]
3014 To build the instrumented application:
3016 . In path:{app.c}, before including path:{tpp.h}, add the
3022 #define TRACEPOINT_DEFINE
3023 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3027 . Compile the application source file:
3036 . Build the application:
3041 $ gcc -o app app.o -ldl
3045 To run the instrumented application with tracing support:
3047 * Preload the tracepoint provider package shared object and
3048 start the application:
3053 $ LD_PRELOAD=./libtpp.so ./app
3057 To run the instrumented application without tracing support:
3059 * Start the application:
3069 The instrumented application dynamically loads the tracepoint provider
3070 package shared object.
3072 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3075 include::../common/ust-sit-step-tp-so.txt[]
3077 To build the instrumented application:
3079 . In path:{app.c}, before including path:{tpp.h}, add the
3085 #define TRACEPOINT_DEFINE
3086 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3090 . Compile the application source file:
3099 . Build the application:
3104 $ gcc -o app app.o -ldl
3108 To run the instrumented application:
3110 * Start the application:
3120 The application is linked with the instrumented user library.
3122 The instrumented user library is statically linked with the tracepoint
3123 provider package object file.
3125 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3128 include::../common/ust-sit-step-tp-o-fpic.txt[]
3130 To build the instrumented user library:
3132 . In path:{emon.c}, before including path:{tpp.h}, add the
3138 #define TRACEPOINT_DEFINE
3142 . Compile the user library source file:
3147 $ gcc -I. -fpic -c emon.c
3151 . Build the user library shared object:
3156 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3160 To build the application:
3162 . Compile the application source file:
3171 . Build the application:
3176 $ gcc -o app app.o -L. -lemon
3180 To run the application:
3182 * Start the application:
3192 The application is linked with the instrumented user library.
3194 The instrumented user library is linked with the tracepoint provider
3195 package shared object.
3197 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3200 include::../common/ust-sit-step-tp-so.txt[]
3202 To build the instrumented user library:
3204 . In path:{emon.c}, before including path:{tpp.h}, add the
3210 #define TRACEPOINT_DEFINE
3214 . Compile the user library source file:
3219 $ gcc -I. -fpic -c emon.c
3223 . Build the user library shared object:
3228 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3232 To build the application:
3234 . Compile the application source file:
3243 . Build the application:
3248 $ gcc -o app app.o -L. -lemon
3252 To run the application:
3254 * Start the application:
3264 The tracepoint provider package shared object is preloaded before the
3267 The application is linked with the instrumented user library.
3269 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3272 include::../common/ust-sit-step-tp-so.txt[]
3274 To build the instrumented user library:
3276 . In path:{emon.c}, before including path:{tpp.h}, add the
3282 #define TRACEPOINT_DEFINE
3283 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3287 . Compile the user library source file:
3292 $ gcc -I. -fpic -c emon.c
3296 . Build the user library shared object:
3301 $ gcc -shared -o libemon.so emon.o -ldl
3305 To build the application:
3307 . Compile the application source file:
3316 . Build the application:
3321 $ gcc -o app app.o -L. -lemon
3325 To run the application with tracing support:
3327 * Preload the tracepoint provider package shared object and
3328 start the application:
3333 $ LD_PRELOAD=./libtpp.so ./app
3337 To run the application without tracing support:
3339 * Start the application:
3349 The application is linked with the instrumented user library.
3351 The instrumented user library dynamically loads the tracepoint provider
3352 package shared object.
3354 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3357 include::../common/ust-sit-step-tp-so.txt[]
3359 To build the instrumented user library:
3361 . In path:{emon.c}, before including path:{tpp.h}, add the
3367 #define TRACEPOINT_DEFINE
3368 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3372 . Compile the user library source file:
3377 $ gcc -I. -fpic -c emon.c
3381 . Build the user library shared object:
3386 $ gcc -shared -o libemon.so emon.o -ldl
3390 To build the application:
3392 . Compile the application source file:
3401 . Build the application:
3406 $ gcc -o app app.o -L. -lemon
3410 To run the application:
3412 * Start the application:
3422 The application dynamically loads the instrumented user library.
3424 The instrumented user library is linked with the tracepoint provider
3425 package shared object.
3427 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3430 include::../common/ust-sit-step-tp-so.txt[]
3432 To build the instrumented user library:
3434 . In path:{emon.c}, before including path:{tpp.h}, add the
3440 #define TRACEPOINT_DEFINE
3444 . Compile the user library source file:
3449 $ gcc -I. -fpic -c emon.c
3453 . Build the user library shared object:
3458 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3462 To build the application:
3464 . Compile the application source file:
3473 . Build the application:
3478 $ gcc -o app app.o -ldl -L. -lemon
3482 To run the application:
3484 * Start the application:
3494 The application dynamically loads the instrumented user library.
3496 The instrumented user library dynamically loads the tracepoint provider
3497 package shared object.
3499 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3502 include::../common/ust-sit-step-tp-so.txt[]
3504 To build the instrumented user library:
3506 . In path:{emon.c}, before including path:{tpp.h}, add the
3512 #define TRACEPOINT_DEFINE
3513 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3517 . Compile the user library source file:
3522 $ gcc -I. -fpic -c emon.c
3526 . Build the user library shared object:
3531 $ gcc -shared -o libemon.so emon.o -ldl
3535 To build the application:
3537 . Compile the application source file:
3546 . Build the application:
3551 $ gcc -o app app.o -ldl -L. -lemon
3555 To run the application:
3557 * Start the application:
3567 The tracepoint provider package shared object is preloaded before the
3570 The application dynamically loads the instrumented user library.
3572 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3575 include::../common/ust-sit-step-tp-so.txt[]
3577 To build the instrumented user library:
3579 . In path:{emon.c}, before including path:{tpp.h}, add the
3585 #define TRACEPOINT_DEFINE
3586 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3590 . Compile the user library source file:
3595 $ gcc -I. -fpic -c emon.c
3599 . Build the user library shared object:
3604 $ gcc -shared -o libemon.so emon.o -ldl
3608 To build the application:
3610 . Compile the application source file:
3619 . Build the application:
3624 $ gcc -o app app.o -L. -lemon
3628 To run the application with tracing support:
3630 * Preload the tracepoint provider package shared object and
3631 start the application:
3636 $ LD_PRELOAD=./libtpp.so ./app
3640 To run the application without tracing support:
3642 * Start the application:
3652 The application is statically linked with the tracepoint provider
3653 package object file.
3655 The application is linked with the instrumented user library.
3657 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3660 include::../common/ust-sit-step-tp-o.txt[]
3662 To build the instrumented user library:
3664 . In path:{emon.c}, before including path:{tpp.h}, add the
3670 #define TRACEPOINT_DEFINE
3674 . Compile the user library source file:
3679 $ gcc -I. -fpic -c emon.c
3683 . Build the user library shared object:
3688 $ gcc -shared -o libemon.so emon.o
3692 To build the application:
3694 . Compile the application source file:
3703 . Build the application:
3708 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3712 To run the instrumented application:
3714 * Start the application:
3724 The application is statically linked with the tracepoint provider
3725 package object file.
3727 The application dynamically loads the instrumented user library.
3729 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3732 include::../common/ust-sit-step-tp-o.txt[]
3734 To build the application:
3736 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3741 #define TRACEPOINT_DEFINE
3745 . Compile the application source file:
3754 . Build the application:
3759 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3764 The `--export-dynamic` option passed to the linker is necessary for the
3765 dynamically loaded library to ``see'' the tracepoint symbols defined in
3768 To build the instrumented user library:
3770 . Compile the user library source file:
3775 $ gcc -I. -fpic -c emon.c
3779 . Build the user library shared object:
3784 $ gcc -shared -o libemon.so emon.o
3788 To run the application:
3790 * Start the application:
3801 [[using-lttng-ust-with-daemons]]
3802 ===== Use noch:{LTTng-UST} with daemons
3804 If your instrumented application calls man:fork(2), man:clone(2),
3805 or BSD's man:rfork(2), without a following man:exec(3)-family
3806 system call, you must preload the path:{liblttng-ust-fork.so} shared
3807 object when you start the application.
3811 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3814 If your tracepoint provider package is
3815 a shared library which you also preload, you must put both
3816 shared objects in env:LD_PRELOAD:
3820 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3826 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3828 If your instrumented application closes one or more file descriptors
3829 which it did not open itself, you must preload the
3830 path:{liblttng-ust-fd.so} shared object when you start the application:
3834 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3837 Typical use cases include closing all the file descriptors after
3838 man:fork(2) or man:rfork(2) and buggy applications doing
3842 [[lttng-ust-pkg-config]]
3843 ===== Use noch:{pkg-config}
3845 On some distributions, LTTng-UST ships with a
3846 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3847 metadata file. If this is your case, then use cmd:pkg-config to
3848 build an application on the command line:
3852 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3856 [[instrumenting-32-bit-app-on-64-bit-system]]
3857 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3859 In order to trace a 32-bit application running on a 64-bit system,
3860 LTTng must use a dedicated 32-bit
3861 <<lttng-consumerd,consumer daemon>>.
3863 The following steps show how to build and install a 32-bit consumer
3864 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3865 build and install the 32-bit LTTng-UST libraries, and how to build and
3866 link an instrumented 32-bit application in that context.
3868 To build a 32-bit instrumented application for a 64-bit target system,
3869 assuming you have a fresh target system with no installed Userspace RCU
3872 . Download, build, and install a 32-bit version of Userspace RCU:
3877 $ cd $(mktemp -d) &&
3878 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3879 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3880 cd userspace-rcu-0.9.* &&
3881 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3883 sudo make install &&
3888 . Using the package manager of your distribution, or from source,
3889 install the following 32-bit versions of the following dependencies of
3890 LTTng-tools and LTTng-UST:
3893 * https://sourceforge.net/projects/libuuid/[libuuid]
3894 * http://directory.fsf.org/wiki/Popt[popt]
3895 * http://www.xmlsoft.org/[libxml2]
3898 . Download, build, and install a 32-bit version of the latest
3899 LTTng-UST{nbsp}{revision}:
3904 $ cd $(mktemp -d) &&
3905 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.12.tar.bz2 &&
3906 tar -xf lttng-ust-latest-2.12.tar.bz2 &&
3907 cd lttng-ust-2.12.* &&
3908 ./configure --libdir=/usr/local/lib32 \
3909 CFLAGS=-m32 CXXFLAGS=-m32 \
3910 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3912 sudo make install &&
3919 Depending on your distribution,
3920 32-bit libraries could be installed at a different location than
3921 `/usr/lib32`. For example, Debian is known to install
3922 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3924 In this case, make sure to set `LDFLAGS` to all the
3925 relevant 32-bit library paths, for example:
3929 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3933 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3934 the 32-bit consumer daemon:
3939 $ cd $(mktemp -d) &&
3940 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
3941 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
3942 cd lttng-tools-2.12.* &&
3943 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3944 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3945 --disable-bin-lttng --disable-bin-lttng-crash \
3946 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3948 cd src/bin/lttng-consumerd &&
3949 sudo make install &&
3954 . From your distribution or from source,
3955 <<installing-lttng,install>> the 64-bit versions of
3956 LTTng-UST and Userspace RCU.
3957 . Download, build, and install the 64-bit version of the
3958 latest LTTng-tools{nbsp}{revision}:
3963 $ cd $(mktemp -d) &&
3964 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
3965 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
3966 cd lttng-tools-2.12.* &&
3967 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3968 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3970 sudo make install &&
3975 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3976 when linking your 32-bit application:
3979 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3980 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3983 For example, let's rebuild the quick start example in
3984 <<tracing-your-own-user-application,Trace a user application>> as an
3985 instrumented 32-bit application:
3990 $ gcc -m32 -c -I. hello-tp.c
3991 $ gcc -m32 -c hello.c
3992 $ gcc -m32 -o hello hello.o hello-tp.o \
3993 -L/usr/lib32 -L/usr/local/lib32 \
3994 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3999 No special action is required to execute the 32-bit application and
4000 to trace it: use the command-line man:lttng(1) tool as usual.
4007 man:tracef(3) is a small LTTng-UST API designed for quick,
4008 man:printf(3)-like instrumentation without the burden of
4009 <<tracepoint-provider,creating>> and
4010 <<building-tracepoint-providers-and-user-application,building>>
4011 a tracepoint provider package.
4013 To use `tracef()` in your application:
4015 . In the C or C++ source files where you need to use `tracef()`,
4016 include `<lttng/tracef.h>`:
4021 #include <lttng/tracef.h>
4025 . In the source code of the application, use `tracef()` like you would
4033 tracef("my message: %d (%s)", my_integer, my_string);
4039 . Link your application with `liblttng-ust`:
4044 $ gcc -o app app.c -llttng-ust
4048 To trace the events that `tracef()` calls emit:
4050 * <<enabling-disabling-events,Create an event rule>> which matches the
4051 `lttng_ust_tracef:*` event name:
4056 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4061 .Limitations of `tracef()`
4063 The `tracef()` utility function was developed to make user space tracing
4064 super simple, albeit with notable disadvantages compared to
4065 <<defining-tracepoints,user-defined tracepoints>>:
4067 * All the emitted events have the same tracepoint provider and
4068 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4069 * There is no static type checking.
4070 * The only event record field you actually get, named `msg`, is a string
4071 potentially containing the values you passed to `tracef()`
4072 using your own format string. This also means that you can't filter
4073 events with a custom expression at run time because there are no
4075 * Since `tracef()` uses the man:vasprintf(3) function of the
4076 C{nbsp}standard library behind the scenes to format the strings at run
4077 time, its expected performance is lower than with user-defined
4078 tracepoints, which don't require a conversion to a string.
4080 Taking this into consideration, `tracef()` is useful for some quick
4081 prototyping and debugging, but you shouldn't consider it for any
4082 permanent and serious applicative instrumentation.
4088 ==== Use `tracelog()`
4090 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4091 the difference that it accepts an additional log level parameter.
4093 The goal of `tracelog()` is to ease the migration from logging to
4096 To use `tracelog()` in your application:
4098 . In the C or C++ source files where you need to use `tracelog()`,
4099 include `<lttng/tracelog.h>`:
4104 #include <lttng/tracelog.h>
4108 . In the source code of the application, use `tracelog()` like you would
4109 use man:printf(3), except for the first parameter which is the log
4117 tracelog(TRACE_WARNING, "my message: %d (%s)",
4118 my_integer, my_string);
4124 See man:lttng-ust(3) for a list of available log level names.
4126 . Link your application with `liblttng-ust`:
4131 $ gcc -o app app.c -llttng-ust
4135 To trace the events that `tracelog()` calls emit with a log level
4136 _as severe as_ a specific log level:
4138 * <<enabling-disabling-events,Create an event rule>> which matches the
4139 `lttng_ust_tracelog:*` event name and a minimum level
4145 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4146 --loglevel=TRACE_WARNING
4150 To trace the events that `tracelog()` calls emit with a
4151 _specific log level_:
4153 * Create an event rule which matches the `lttng_ust_tracelog:*`
4154 event name and a specific log level:
4159 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4160 --loglevel-only=TRACE_INFO
4165 [[prebuilt-ust-helpers]]
4166 === Prebuilt user space tracing helpers
4168 The LTTng-UST package provides a few helpers in the form of preloadable
4169 shared objects which automatically instrument system functions and
4172 The helper shared objects are normally found in dir:{/usr/lib}. If you
4173 built LTTng-UST <<building-from-source,from source>>, they are probably
4174 located in dir:{/usr/local/lib}.
4176 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4179 path:{liblttng-ust-libc-wrapper.so}::
4180 path:{liblttng-ust-pthread-wrapper.so}::
4181 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4182 memory and POSIX threads function tracing>>.
4184 path:{liblttng-ust-cyg-profile.so}::
4185 path:{liblttng-ust-cyg-profile-fast.so}::
4186 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4188 path:{liblttng-ust-dl.so}::
4189 <<liblttng-ust-dl,Dynamic linker tracing>>.
4191 To use a user space tracing helper with any user application:
4193 * Preload the helper shared object when you start the application:
4198 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4202 You can preload more than one helper:
4207 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4213 [[liblttng-ust-libc-pthread-wrapper]]
4214 ==== Instrument C standard library memory and POSIX threads functions
4216 The path:{liblttng-ust-libc-wrapper.so} and
4217 path:{liblttng-ust-pthread-wrapper.so} helpers
4218 add instrumentation to some C standard library and POSIX
4222 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4224 |TP provider name |TP name |Instrumented function
4226 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4227 |`calloc` |man:calloc(3)
4228 |`realloc` |man:realloc(3)
4229 |`free` |man:free(3)
4230 |`memalign` |man:memalign(3)
4231 |`posix_memalign` |man:posix_memalign(3)
4235 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4237 |TP provider name |TP name |Instrumented function
4239 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4240 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4241 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4242 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4245 When you preload the shared object, it replaces the functions listed
4246 in the previous tables by wrappers which contain tracepoints and call
4247 the replaced functions.
4250 [[liblttng-ust-cyg-profile]]
4251 ==== Instrument function entry and exit
4253 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4254 to the entry and exit points of functions.
4256 man:gcc(1) and man:clang(1) have an option named
4257 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4258 which generates instrumentation calls for entry and exit to functions.
4259 The LTTng-UST function tracing helpers,
4260 path:{liblttng-ust-cyg-profile.so} and
4261 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4262 to add tracepoints to the two generated functions (which contain
4263 `cyg_profile` in their names, hence the name of the helper).
4265 To use the LTTng-UST function tracing helper, the source files to
4266 instrument must be built using the `-finstrument-functions` compiler
4269 There are two versions of the LTTng-UST function tracing helper:
4271 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4272 that you should only use when it can be _guaranteed_ that the
4273 complete event stream is recorded without any lost event record.
4274 Any kind of duplicate information is left out.
4276 Assuming no event record is lost, having only the function addresses on
4277 entry is enough to create a call graph, since an event record always
4278 contains the ID of the CPU that generated it.
4280 Use a tool like man:addr2line(1) to convert function addresses back to
4281 source file names and line numbers.
4283 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4284 which also works in use cases where event records might get discarded or
4285 not recorded from application startup.
4286 In these cases, the trace analyzer needs more information to be
4287 able to reconstruct the program flow.
4289 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4290 points of this helper.
4292 All the tracepoints that this helper provides have the
4293 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4295 TIP: It's sometimes a good idea to limit the number of source files that
4296 you compile with the `-finstrument-functions` option to prevent LTTng
4297 from writing an excessive amount of trace data at run time. When using
4299 `-finstrument-functions-exclude-function-list` option to avoid
4300 instrument entries and exits of specific function names.
4305 ==== Instrument the dynamic linker
4307 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4308 man:dlopen(3) and man:dlclose(3) function calls.
4310 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4315 [[java-application]]
4316 === User space Java agent
4318 You can instrument any Java application which uses one of the following
4321 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4322 (JUL) core logging facilities.
4323 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4324 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 isn't supported.
4327 .LTTng-UST Java agent imported by a Java application.
4328 image::java-app.png[]
4330 Note that the methods described below are new in LTTng{nbsp}2.8.
4331 Previous LTTng versions use another technique.
4333 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4334 and https://ci.lttng.org/[continuous integration], thus this version is
4335 directly supported. However, the LTTng-UST Java agent is also tested
4336 with OpenJDK{nbsp}7.
4341 ==== Use the LTTng-UST Java agent for `java.util.logging`
4343 To use the LTTng-UST Java agent in a Java application which uses
4344 `java.util.logging` (JUL):
4346 . In the source code of the Java application, import the LTTng-UST log
4347 handler package for `java.util.logging`:
4352 import org.lttng.ust.agent.jul.LttngLogHandler;
4356 . Create an LTTng-UST JUL log handler:
4361 Handler lttngUstLogHandler = new LttngLogHandler();
4365 . Add this handler to the JUL loggers which should emit LTTng events:
4370 Logger myLogger = Logger.getLogger("some-logger");
4372 myLogger.addHandler(lttngUstLogHandler);
4376 . Use `java.util.logging` log statements and configuration as usual.
4377 The loggers with an attached LTTng-UST log handler can emit
4380 . Before exiting the application, remove the LTTng-UST log handler from
4381 the loggers attached to it and call its `close()` method:
4386 myLogger.removeHandler(lttngUstLogHandler);
4387 lttngUstLogHandler.close();
4391 This isn't strictly necessary, but it is recommended for a clean
4392 disposal of the resources of the handler.
4394 . Include the common and JUL-specific JAR files of the LTTng-UST Java agent,
4395 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4397 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4398 path] when you build the Java application.
4400 The JAR files are typically located in dir:{/usr/share/java}.
4402 IMPORTANT: The LTTng-UST Java agent must be
4403 <<installing-lttng,installed>> for the logging framework your
4406 .Use the LTTng-UST Java agent for `java.util.logging`.
4411 import java.io.IOException;
4412 import java.util.logging.Handler;
4413 import java.util.logging.Logger;
4414 import org.lttng.ust.agent.jul.LttngLogHandler;
4418 private static final int answer = 42;
4420 public static void main(String[] argv) throws Exception
4423 Logger logger = Logger.getLogger("jello");
4425 // Create an LTTng-UST log handler
4426 Handler lttngUstLogHandler = new LttngLogHandler();
4428 // Add the LTTng-UST log handler to our logger
4429 logger.addHandler(lttngUstLogHandler);
4432 logger.info("some info");
4433 logger.warning("some warning");
4435 logger.finer("finer information; the answer is " + answer);
4437 logger.severe("error!");
4439 // Not mandatory, but cleaner
4440 logger.removeHandler(lttngUstLogHandler);
4441 lttngUstLogHandler.close();
4450 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4453 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4454 <<enabling-disabling-events,create an event rule>> matching the
4455 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4460 $ lttng enable-event --jul jello
4464 Run the compiled class:
4468 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4471 <<basic-tracing-session-control,Stop tracing>> and inspect the
4481 In the resulting trace, an <<event,event record>> generated by a Java
4482 application using `java.util.logging` is named `lttng_jul:event` and
4483 has the following fields:
4492 Name of the class in which the log statement was executed.
4495 Name of the method in which the log statement was executed.
4498 Logging time (timestamp in milliseconds).
4501 Log level integer value.
4504 ID of the thread in which the log statement was executed.
4506 Use the opt:lttng-enable-event(1):--loglevel or
4507 opt:lttng-enable-event(1):--loglevel-only option of the
4508 man:lttng-enable-event(1) command to target a range of JUL log levels
4509 or a specific JUL log level.
4514 ==== Use the LTTng-UST Java agent for Apache log4j
4516 To use the LTTng-UST Java agent in a Java application which uses
4517 Apache log4j{nbsp}1.2:
4519 . In the source code of the Java application, import the LTTng-UST log
4520 appender package for Apache log4j:
4525 import org.lttng.ust.agent.log4j.LttngLogAppender;
4529 . Create an LTTng-UST log4j log appender:
4534 Appender lttngUstLogAppender = new LttngLogAppender();
4538 . Add this appender to the log4j loggers which should emit LTTng events:
4543 Logger myLogger = Logger.getLogger("some-logger");
4545 myLogger.addAppender(lttngUstLogAppender);
4549 . Use Apache log4j log statements and configuration as usual. The
4550 loggers with an attached LTTng-UST log appender can emit LTTng events.
4552 . Before exiting the application, remove the LTTng-UST log appender from
4553 the loggers attached to it and call its `close()` method:
4558 myLogger.removeAppender(lttngUstLogAppender);
4559 lttngUstLogAppender.close();
4563 This isn't strictly necessary, but it is recommended for a clean
4564 disposal of the resources of the appender.
4566 . Include the common and log4j-specific JAR
4567 files of the LTTng-UST Java agent, path:{lttng-ust-agent-common.jar} and
4568 path:{lttng-ust-agent-log4j.jar}, in the
4569 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4570 path] when you build the Java application.
4572 The JAR files are typically located in dir:{/usr/share/java}.
4574 IMPORTANT: The LTTng-UST Java agent must be
4575 <<installing-lttng,installed>> for the logging framework your
4578 .Use the LTTng-UST Java agent for Apache log4j.
4583 import org.apache.log4j.Appender;
4584 import org.apache.log4j.Logger;
4585 import org.lttng.ust.agent.log4j.LttngLogAppender;
4589 private static final int answer = 42;
4591 public static void main(String[] argv) throws Exception
4594 Logger logger = Logger.getLogger("jello");
4596 // Create an LTTng-UST log appender
4597 Appender lttngUstLogAppender = new LttngLogAppender();
4599 // Add the LTTng-UST log appender to our logger
4600 logger.addAppender(lttngUstLogAppender);
4603 logger.info("some info");
4604 logger.warn("some warning");
4606 logger.debug("debug information; the answer is " + answer);
4608 logger.fatal("error!");
4610 // Not mandatory, but cleaner
4611 logger.removeAppender(lttngUstLogAppender);
4612 lttngUstLogAppender.close();
4618 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4623 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4626 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4627 <<enabling-disabling-events,create an event rule>> matching the
4628 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4633 $ lttng enable-event --log4j jello
4637 Run the compiled class:
4641 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4644 <<basic-tracing-session-control,Stop tracing>> and inspect the
4654 In the resulting trace, an <<event,event record>> generated by a Java
4655 application using log4j is named `lttng_log4j:event` and
4656 has the following fields:
4665 Name of the class in which the log statement was executed.
4668 Name of the method in which the log statement was executed.
4671 Name of the file in which the executed log statement is located.
4674 Line number at which the log statement was executed.
4680 Log level integer value.
4683 Name of the Java thread in which the log statement was executed.
4685 Use the opt:lttng-enable-event(1):--loglevel or
4686 opt:lttng-enable-event(1):--loglevel-only option of the
4687 man:lttng-enable-event(1) command to target a range of Apache log4j
4688 log levels or a specific log4j log level.
4692 [[java-application-context]]
4693 ==== Provide application-specific context fields in a Java application
4695 A Java application-specific context field is a piece of state provided
4696 by the application which <<adding-context,you can add>>, using the
4697 man:lttng-add-context(1) command, to each <<event,event record>>
4698 produced by the log statements of this application.
4700 For example, a given object might have a current request ID variable.
4701 You can create a context information retriever for this object and
4702 assign a name to this current request ID. You can then, using the
4703 man:lttng-add-context(1) command, add this context field by name to
4704 the JUL or log4j <<channel,channel>>.
4706 To provide application-specific context fields in a Java application:
4708 . In the source code of the Java application, import the LTTng-UST
4709 Java agent context classes and interfaces:
4714 import org.lttng.ust.agent.context.ContextInfoManager;
4715 import org.lttng.ust.agent.context.IContextInfoRetriever;
4719 . Create a context information retriever class, that is, a class which
4720 implements the `IContextInfoRetriever` interface:
4725 class MyContextInfoRetriever implements IContextInfoRetriever
4728 public Object retrieveContextInfo(String key)
4730 if (key.equals("intCtx")) {
4732 } else if (key.equals("strContext")) {
4733 return "context value!";
4742 This `retrieveContextInfo()` method is the only member of the
4743 `IContextInfoRetriever` interface. Its role is to return the current
4744 value of a state by name to create a context field. The names of the
4745 context fields and which state variables they return depends on your
4748 All primitive types and objects are supported as context fields.
4749 When `retrieveContextInfo()` returns an object, the context field
4750 serializer calls its `toString()` method to add a string field to
4751 event records. The method can also return `null`, which means that
4752 no context field is available for the required name.
4754 . Register an instance of your context information retriever class to
4755 the context information manager singleton:
4760 IContextInfoRetriever cir = new MyContextInfoRetriever();
4761 ContextInfoManager cim = ContextInfoManager.getInstance();
4762 cim.registerContextInfoRetriever("retrieverName", cir);
4766 . Before exiting the application, remove your context information
4767 retriever from the context information manager singleton:
4772 ContextInfoManager cim = ContextInfoManager.getInstance();
4773 cim.unregisterContextInfoRetriever("retrieverName");
4777 This isn't strictly necessary, but it is recommended for a clean
4778 disposal of some resources of the manager.
4780 . Build your Java application with LTTng-UST Java agent support as
4781 usual, following the procedure for either the <<jul,JUL>> or
4782 <<log4j,Apache log4j>> framework.
4785 .Provide application-specific context fields in a Java application.
4790 import java.util.logging.Handler;
4791 import java.util.logging.Logger;
4792 import org.lttng.ust.agent.jul.LttngLogHandler;
4793 import org.lttng.ust.agent.context.ContextInfoManager;
4794 import org.lttng.ust.agent.context.IContextInfoRetriever;
4798 // Our context information retriever class
4799 private static class MyContextInfoRetriever
4800 implements IContextInfoRetriever
4803 public Object retrieveContextInfo(String key) {
4804 if (key.equals("intCtx")) {
4806 } else if (key.equals("strContext")) {
4807 return "context value!";
4814 private static final int answer = 42;
4816 public static void main(String args[]) throws Exception
4818 // Get the context information manager instance
4819 ContextInfoManager cim = ContextInfoManager.getInstance();
4821 // Create and register our context information retriever
4822 IContextInfoRetriever cir = new MyContextInfoRetriever();
4823 cim.registerContextInfoRetriever("myRetriever", cir);
4826 Logger logger = Logger.getLogger("jello");
4828 // Create an LTTng-UST log handler
4829 Handler lttngUstLogHandler = new LttngLogHandler();
4831 // Add the LTTng-UST log handler to our logger
4832 logger.addHandler(lttngUstLogHandler);
4835 logger.info("some info");
4836 logger.warning("some warning");
4838 logger.finer("finer information; the answer is " + answer);
4840 logger.severe("error!");
4842 // Not mandatory, but cleaner
4843 logger.removeHandler(lttngUstLogHandler);
4844 lttngUstLogHandler.close();
4845 cim.unregisterContextInfoRetriever("myRetriever");
4854 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4857 <<creating-destroying-tracing-sessions,Create a tracing session>>
4858 and <<enabling-disabling-events,create an event rule>> matching the
4864 $ lttng enable-event --jul jello
4867 <<adding-context,Add the application-specific context fields>> to the
4872 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4873 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4876 <<basic-tracing-session-control,Start tracing>>:
4883 Run the compiled class:
4887 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4890 <<basic-tracing-session-control,Stop tracing>> and inspect the
4902 [[python-application]]
4903 === User space Python agent
4905 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4907 https://docs.python.org/3/library/logging.html[`logging`] package.
4909 Each log statement emits an LTTng event once the
4910 application module imports the
4911 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4914 .A Python application importing the LTTng-UST Python agent.
4915 image::python-app.png[]
4917 To use the LTTng-UST Python agent:
4919 . In the source code of the Python application, import the LTTng-UST
4929 The LTTng-UST Python agent automatically adds its logging handler to the
4930 root logger at import time.
4932 Any log statement that the application executes before this import does
4933 not emit an LTTng event.
4935 IMPORTANT: The LTTng-UST Python agent must be
4936 <<installing-lttng,installed>>.
4938 . Use log statements and logging configuration as usual.
4939 Since the LTTng-UST Python agent adds a handler to the _root_
4940 logger, you can trace any log statement from any logger.
4942 .Use the LTTng-UST Python agent.
4953 logging.basicConfig()
4954 logger = logging.getLogger('my-logger')
4957 logger.debug('debug message')
4958 logger.info('info message')
4959 logger.warn('warn message')
4960 logger.error('error message')
4961 logger.critical('critical message')
4965 if __name__ == '__main__':
4969 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4970 logging handler which prints to the standard error stream, isn't
4971 strictly required for LTTng-UST tracing to work, but in versions of
4972 Python preceding{nbsp}3.2, you could see a warning message which indicates
4973 that no handler exists for the logger `my-logger`.
4975 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4976 <<enabling-disabling-events,create an event rule>> matching the
4977 `my-logger` Python logger, and <<basic-tracing-session-control,start
4983 $ lttng enable-event --python my-logger
4987 Run the Python script:
4994 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5004 In the resulting trace, an <<event,event record>> generated by a Python
5005 application is named `lttng_python:event` and has the following fields:
5008 Logging time (string).
5017 Name of the function in which the log statement was executed.
5020 Line number at which the log statement was executed.
5023 Log level integer value.
5026 ID of the Python thread in which the log statement was executed.
5029 Name of the Python thread in which the log statement was executed.
5031 Use the opt:lttng-enable-event(1):--loglevel or
5032 opt:lttng-enable-event(1):--loglevel-only option of the
5033 man:lttng-enable-event(1) command to target a range of Python log levels
5034 or a specific Python log level.
5036 When an application imports the LTTng-UST Python agent, the agent tries
5037 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5038 <<start-sessiond,start the session daemon>> _before_ you run the Python
5039 application. If a session daemon is found, the agent tries to register
5040 to it during five seconds, after which the application continues
5041 without LTTng tracing support. Override this timeout value with
5042 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5045 If the session daemon stops while a Python application with an imported
5046 LTTng-UST Python agent runs, the agent retries to connect and to
5047 register to a session daemon every three seconds. Override this
5048 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5053 [[proc-lttng-logger-abi]]
5056 The `lttng-tracer` Linux kernel module, part of
5057 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger files
5058 path:{/proc/lttng-logger} and path:{/dev/lttng-logger} (since
5059 LTTng{nbsp}2.11) when it's loaded. Any application can write text data
5060 to any of those files to emit an LTTng event.
5063 .An application writes to the LTTng logger file to emit an LTTng event.
5064 image::lttng-logger.png[]
5066 The LTTng logger is the quickest method--not the most efficient,
5067 however--to add instrumentation to an application. It is designed
5068 mostly to instrument shell scripts:
5072 $ echo "Some message, some $variable" > /dev/lttng-logger
5075 Any event that the LTTng logger emits is named `lttng_logger` and
5076 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5077 other instrumentation points in the kernel tracing domain, **any Unix
5078 user** can <<enabling-disabling-events,create an event rule>> which
5079 matches its event name, not only the root user or users in the
5080 <<tracing-group,tracing group>>.
5082 To use the LTTng logger:
5084 * From any application, write text data to the path:{/dev/lttng-logger}
5087 The `msg` field of `lttng_logger` event records contains the
5090 NOTE: The maximum message length of an LTTng logger event is
5091 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5092 than one event to contain the remaining data.
5094 You shouldn't use the LTTng logger to trace a user application which
5095 can be instrumented in a more efficient way, namely:
5097 * <<c-application,C and $$C++$$ applications>>.
5098 * <<java-application,Java applications>>.
5099 * <<python-application,Python applications>>.
5101 .Use the LTTng logger.
5106 echo 'Hello, World!' > /dev/lttng-logger
5108 df --human-readable --print-type / > /dev/lttng-logger
5111 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5112 <<enabling-disabling-events,create an event rule>> matching the
5113 `lttng_logger` Linux kernel tracepoint, and
5114 <<basic-tracing-session-control,start tracing>>:
5119 $ lttng enable-event --kernel lttng_logger
5123 Run the Bash script:
5130 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5141 [[instrumenting-linux-kernel]]
5142 === LTTng kernel tracepoints
5144 NOTE: This section shows how to _add_ instrumentation points to the
5145 Linux kernel. The subsystems of the kernel are already thoroughly
5146 instrumented at strategic places for LTTng when you
5147 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5151 There are two methods to instrument the Linux kernel:
5153 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5154 tracepoint which uses the `TRACE_EVENT()` API.
5156 Choose this if you want to instrumentation a Linux kernel tree with an
5157 instrumentation point compatible with ftrace, perf, and SystemTap.
5159 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5160 instrument an out-of-tree kernel module.
5162 Choose this if you don't need ftrace, perf, or SystemTap support.
5166 [[linux-add-lttng-layer]]
5167 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5169 This section shows how to add an LTTng layer to existing ftrace
5170 instrumentation using the `TRACE_EVENT()` API.
5172 This section doesn't document the `TRACE_EVENT()` macro. Read the
5173 following articles to learn more about this API:
5175 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5176 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5177 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5179 The following procedure assumes that your ftrace tracepoints are
5180 correctly defined in their own header and that they are created in
5181 one source file using the `CREATE_TRACE_POINTS` definition.
5183 To add an LTTng layer over an existing ftrace tracepoint:
5185 . Make sure the following kernel configuration options are
5191 * `CONFIG_HIGH_RES_TIMERS`
5192 * `CONFIG_TRACEPOINTS`
5195 . Build the Linux source tree with your custom ftrace tracepoints.
5196 . Boot the resulting Linux image on your target system.
5198 Confirm that the tracepoints exist by looking for their names in the
5199 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5200 is your subsystem name.
5202 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5207 $ cd $(mktemp -d) &&
5208 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
5209 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
5210 cd lttng-modules-2.12.*
5214 . In dir:{instrumentation/events/lttng-module}, relative to the root
5215 of the LTTng-modules source tree, create a header file named
5216 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5217 LTTng-modules tracepoint definitions using the LTTng-modules
5220 Start with this template:
5224 .path:{instrumentation/events/lttng-module/my_subsys.h}
5227 #define TRACE_SYSTEM my_subsys
5229 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5230 #define _LTTNG_MY_SUBSYS_H
5232 #include "../../../probes/lttng-tracepoint-event.h"
5233 #include <linux/tracepoint.h>
5235 LTTNG_TRACEPOINT_EVENT(
5237 * Format is identical to the TRACE_EVENT() version for the three
5238 * following macro parameters:
5241 TP_PROTO(int my_int, const char *my_string),
5242 TP_ARGS(my_int, my_string),
5244 /* LTTng-modules specific macros */
5246 ctf_integer(int, my_int_field, my_int)
5247 ctf_string(my_bar_field, my_bar)
5251 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5253 #include "../../../probes/define_trace.h"
5257 The entries in the `TP_FIELDS()` section are the list of fields for the
5258 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5259 the `TRACE_EVENT()` ftrace macro.
5261 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5262 complete description of the available `ctf_*()` macros.
5264 . Create the kernel module C{nbsp}source file of the LTTng-modules
5265 probe, +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5270 .path:{probes/lttng-probe-my-subsys.c}
5272 #include <linux/module.h>
5273 #include "../lttng-tracer.h"
5276 * Build-time verification of mismatch between mainline
5277 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5278 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5280 #include <trace/events/my_subsys.h>
5282 /* Create LTTng tracepoint probes */
5283 #define LTTNG_PACKAGE_BUILD
5284 #define CREATE_TRACE_POINTS
5285 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5287 #include "../instrumentation/events/lttng-module/my_subsys.h"
5289 MODULE_LICENSE("GPL and additional rights");
5290 MODULE_AUTHOR("Your name <your-email>");
5291 MODULE_DESCRIPTION("LTTng my_subsys probes");
5292 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5293 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5294 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5295 LTTNG_MODULES_EXTRAVERSION);
5299 . Edit path:{probes/KBuild} and add your new kernel module object
5300 next to the existing ones:
5304 .path:{probes/KBuild}
5308 obj-m += lttng-probe-module.o
5309 obj-m += lttng-probe-power.o
5311 obj-m += lttng-probe-my-subsys.o
5317 . Build and install the LTTng kernel modules:
5322 $ make KERNELDIR=/path/to/linux
5323 # make modules_install && depmod -a
5327 Replace `/path/to/linux` with the path to the Linux source tree where
5328 you defined and used tracepoints with the `TRACE_EVENT()` ftrace macro.
5330 Note that you can also use the
5331 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5332 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5333 C code that need to be executed before the event fields are recorded.
5335 The best way to learn how to use the previous LTTng-modules macros is to
5336 inspect the existing LTTng-modules tracepoint definitions in the
5337 dir:{instrumentation/events/lttng-module} header files. Compare them
5338 with the Linux kernel mainline versions in the
5339 dir:{include/trace/events} directory of the Linux source tree.
5343 [[lttng-tracepoint-event-code]]
5344 ===== Use custom C code to access the data for tracepoint fields
5346 Although we recommended to always use the
5347 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5348 the arguments and fields of an LTTng-modules tracepoint when possible,
5349 sometimes you need a more complex process to access the data that the
5350 tracer records as event record fields. In other words, you need local
5351 variables and multiple C{nbsp}statements instead of simple
5352 argument-based expressions that you pass to the
5353 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5355 Use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5356 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5357 a block of C{nbsp}code to be executed before LTTng records the fields.
5358 The structure of this macro is:
5361 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5363 LTTNG_TRACEPOINT_EVENT_CODE(
5365 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5366 * version for the following three macro parameters:
5369 TP_PROTO(int my_int, const char *my_string),
5370 TP_ARGS(my_int, my_string),
5372 /* Declarations of custom local variables */
5375 unsigned long b = 0;
5376 const char *name = "(undefined)";
5377 struct my_struct *my_struct;
5381 * Custom code which uses both tracepoint arguments
5382 * (in TP_ARGS()) and local variables (in TP_locvar()).
5384 * Local variables are actually members of a structure pointed
5385 * to by the special variable tp_locvar.
5389 tp_locvar->a = my_int + 17;
5390 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5391 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5392 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5393 put_my_struct(tp_locvar->my_struct);
5402 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5403 * version for this, except that tp_locvar members can be
5404 * used in the argument expression parameters of
5405 * the ctf_*() macros.
5408 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5409 ctf_integer(int, my_struct_a, tp_locvar->a)
5410 ctf_string(my_string_field, my_string)
5411 ctf_string(my_struct_name, tp_locvar->name)
5416 IMPORTANT: The C code defined in `TP_code()` must not have any side
5417 effects when executed. In particular, the code must not allocate
5418 memory or get resources without deallocating this memory or putting
5419 those resources afterwards.
5422 [[instrumenting-linux-kernel-tracing]]
5423 ==== Load and unload a custom probe kernel module
5425 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5426 kernel module>> in the kernel before it can emit LTTng events.
5428 To load the default probe kernel modules and a custom probe kernel
5431 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5432 probe modules to load when starting a root <<lttng-sessiond,session
5436 .Load the `my_subsys`, `usb`, and the default probe modules.
5440 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5445 You only need to pass the subsystem name, not the whole kernel module
5448 To load _only_ a given custom probe kernel module:
5450 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5451 modules to load when starting a root session daemon:
5454 .Load only the `my_subsys` and `usb` probe modules.
5458 # lttng-sessiond --kmod-probes=my_subsys,usb
5463 To confirm that a probe module is loaded:
5470 $ lsmod | grep lttng_probe_usb
5474 To unload the loaded probe modules:
5476 * Kill the session daemon with `SIGTERM`:
5481 # pkill lttng-sessiond
5485 You can also use the man:modprobe(8) `--remove` option if the session
5486 daemon terminates abnormally.
5489 [[controlling-tracing]]
5492 Once an application or a Linux kernel is
5493 <<instrumenting,instrumented>> for LTTng tracing,
5496 This section is divided in topics on how to use the various
5497 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5498 command-line tool>>, to _control_ the LTTng daemons and tracers.
5500 NOTE: In the following subsections, we refer to an man:lttng(1) command
5501 using its man page name. For example, instead of _Run the `create`
5502 command to..._, we use _Run the man:lttng-create(1) command to..._.
5506 === Start a session daemon
5508 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5509 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5512 You will see the following error when you run a command while no session
5516 Error: No session daemon is available
5519 The only command that automatically runs a session daemon is
5520 man:lttng-create(1), which you use to
5521 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5522 this is most of the time the first operation that you do, sometimes it's
5523 not. Some examples are:
5525 * <<list-instrumentation-points,List the available instrumentation points>>.
5526 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5528 [[tracing-group]] Each Unix user must have its own running session
5529 daemon to trace user applications. The session daemon that the root user
5530 starts is the only one allowed to control the LTTng kernel tracer. Users
5531 that are part of the _tracing group_ can control the root session
5532 daemon. The default tracing group name is `tracing`; set it to something
5533 else with the opt:lttng-sessiond(8):--group option when you start the
5534 root session daemon.
5536 To start a user session daemon:
5538 * Run man:lttng-sessiond(8):
5543 $ lttng-sessiond --daemonize
5547 To start the root session daemon:
5549 * Run man:lttng-sessiond(8) as the root user:
5554 # lttng-sessiond --daemonize
5558 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5559 start the session daemon in foreground.
5561 To stop a session daemon, use man:kill(1) on its process ID (standard
5564 Note that some Linux distributions could manage the LTTng session daemon
5565 as a service. In this case, you should use the service manager to
5566 start, restart, and stop session daemons.
5569 [[creating-destroying-tracing-sessions]]
5570 === Create and destroy a tracing session
5572 Almost all the LTTng control operations happen in the scope of
5573 a <<tracing-session,tracing session>>, which is the dialogue between the
5574 <<lttng-sessiond,session daemon>> and you.
5576 To create a tracing session with a generated name:
5578 * Use the man:lttng-create(1) command:
5587 The name of the created tracing session is `auto` followed by the
5590 To create a tracing session with a specific name:
5592 * Use the optional argument of the man:lttng-create(1) command:
5597 $ lttng create my-session
5601 Replace `my-session` with the specific tracing session name.
5603 LTTng appends the creation date to the name of the created tracing
5606 LTTng writes the traces of a tracing session in
5607 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5608 name of the tracing session. Note that the env:LTTNG_HOME environment
5609 variable defaults to `$HOME` if not set.
5611 To output LTTng traces to a non-default location:
5613 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5618 $ lttng create my-session --output=/tmp/some-directory
5622 You may create as many tracing sessions as you wish.
5624 To list all the existing tracing sessions for your Unix user:
5626 * Use the man:lttng-list(1) command:
5635 [[cur-tracing-session]]When you create a tracing session, it is set as
5636 the _current tracing session_. The following man:lttng(1) commands
5637 operate on the current tracing session when you don't specify one:
5639 [role="list-3-cols"]
5640 * man:lttng-add-context(1)
5641 * man:lttng-clear(1)
5642 * man:lttng-destroy(1)
5643 * man:lttng-disable-channel(1)
5644 * man:lttng-disable-event(1)
5645 * man:lttng-disable-rotation(1)
5646 * man:lttng-enable-channel(1)
5647 * man:lttng-enable-event(1)
5648 * man:lttng-enable-rotation(1)
5650 * man:lttng-regenerate(1)
5651 * man:lttng-rotate(1)
5653 * man:lttng-snapshot(1)
5654 * man:lttng-start(1)
5655 * man:lttng-status(1)
5657 * man:lttng-track(1)
5658 * man:lttng-untrack(1)
5661 To change the current tracing session:
5663 * Use the man:lttng-set-session(1) command:
5668 $ lttng set-session new-session
5672 Replace `new-session` by the name of the new current tracing session.
5674 When you're done tracing in a given tracing session, destroy it. This
5675 operation frees the resources taken by the tracing session to destroy;
5676 it doesn't destroy the trace data that LTTng wrote for this tracing
5677 session (see <<clear,Clear a tracing session>> for one way to do this).
5679 To destroy the current tracing session:
5681 * Use the man:lttng-destroy(1) command:
5690 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5691 command implicitly (see <<basic-tracing-session-control,Start and stop a
5692 tracing session>>). You need to stop tracing to make LTTng flush the
5693 remaining trace data and make the trace readable.
5696 [[list-instrumentation-points]]
5697 === List the available instrumentation points
5699 The <<lttng-sessiond,session daemon>> can query the running instrumented
5700 user applications and the Linux kernel to get a list of available
5701 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5702 they are tracepoints and system calls. For the user space tracing
5703 domain, they are tracepoints. For the other tracing domains, they are
5706 To list the available instrumentation points:
5708 * Use the man:lttng-list(1) command with the option of the requested
5709 tracing domain amongst:
5712 opt:lttng-list(1):--kernel::
5713 Linux kernel tracepoints (your Unix user must be a root user, or it
5714 must be a member of the <<tracing-group,tracing group>>).
5716 opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall::
5717 Linux kernel system calls (your Unix user must be a root user, or it
5718 must be a member of the tracing group).
5720 opt:lttng-list(1):--userspace::
5721 User space tracepoints.
5723 opt:lttng-list(1):--jul::
5724 `java.util.logging` loggers.
5726 opt:lttng-list(1):--log4j::
5727 Apache log4j loggers.
5729 opt:lttng-list(1):--python::
5733 .List the available user space tracepoints.
5737 $ lttng list --userspace
5741 .List the available Linux kernel system call tracepoints.
5745 $ lttng list --kernel --syscall
5750 [[enabling-disabling-events]]
5751 === Create and enable an event rule
5753 Once you <<creating-destroying-tracing-sessions,create a tracing
5754 session>>, you can create <<event,event rules>> with the
5755 man:lttng-enable-event(1) command.
5757 You specify each condition with a command-line option. The available
5758 condition arguments are shown in the following table.
5760 [role="growable",cols="asciidoc,asciidoc,default"]
5761 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5763 |Argument |Description |Applicable tracing domains
5769 . +--probe=__ADDR__+
5770 . +--function=__ADDR__+
5771 . +--userspace-probe=__PATH__:__SYMBOL__+
5772 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5775 Instead of using the default _tracepoint_ instrumentation type, use:
5777 . A Linux system call (entry and exit).
5778 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5779 . The entry and return points of a Linux function (symbol or address).
5780 . The entry point of a user application or library function (path to
5781 application/library and symbol).
5782 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5783 Statically Defined Tracing] (USDT) probe (path to application/library,
5784 provider and probe names).
5788 |First positional argument.
5791 Tracepoint or system call name.
5793 With the opt:lttng-enable-event(1):--probe,
5794 opt:lttng-enable-event(1):--function, and
5795 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5796 name given to the event rule. With the JUL, log4j, and Python domains,
5797 this is a logger name.
5799 With a tracepoint, logger, or system call name, use the special
5800 `*` globbing character to match anything (for example, `sched_*`,
5808 . +--loglevel=__LEVEL__+
5809 . +--loglevel-only=__LEVEL__+
5812 . Match only tracepoints or log statements with a logging level at
5813 least as severe as +__LEVEL__+.
5814 . Match only tracepoints or log statements with a logging level
5815 equal to +__LEVEL__+.
5817 See man:lttng-enable-event(1) for the list of available logging level
5820 |User space, JUL, log4j, and Python.
5822 |+--exclude=__EXCLUSIONS__+
5825 When you use a `*` character at the end of the tracepoint or logger
5826 name (first positional argument), exclude the specific names in the
5827 comma-delimited list +__EXCLUSIONS__+.
5830 User space, JUL, log4j, and Python.
5832 |+--filter=__EXPR__+
5835 Match only events which satisfy the expression +__EXPR__+.
5837 See man:lttng-enable-event(1) to learn more about the syntax of a
5844 You attach an event rule to a <<channel,channel>> on creation. If you do
5845 not specify the channel with the opt:lttng-enable-event(1):--channel
5846 option, and if the event rule to create is the first in its
5847 <<domain,tracing domain>> for a given tracing session, then LTTng
5848 creates a _default channel_ for you. This default channel is reused in
5849 subsequent invocations of the man:lttng-enable-event(1) command for the
5850 same tracing domain.
5852 An event rule is always enabled at creation time.
5854 The following examples show how to combine the previous
5855 command-line options to create simple to more complex event rules.
5857 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5861 $ lttng enable-event --kernel sched_switch
5865 .Create an event rule matching four Linux kernel system calls (default channel).
5869 $ lttng enable-event --kernel --syscall open,write,read,close
5873 .Create event rules matching tracepoints with filter expressions (default channel).
5877 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5882 $ lttng enable-event --kernel --all \
5883 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5888 $ lttng enable-event --jul my_logger \
5889 --filter='$app.retriever:cur_msg_id > 3'
5892 IMPORTANT: Make sure to always quote the filter string when you
5893 use man:lttng(1) from a shell.
5895 See also <<pid-tracking,Track process attributes>> which offers another,
5896 more efficient filtering mechanism for process ID, user ID, and group
5900 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5904 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5907 IMPORTANT: Make sure to always quote the wildcard character when you
5908 use man:lttng(1) from a shell.
5911 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5915 $ lttng enable-event --python my-app.'*' \
5916 --exclude='my-app.module,my-app.hello'
5920 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5924 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5928 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5932 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5936 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5940 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5945 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
5949 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5950 server_accept_request
5954 The event rules of a given channel form a whitelist: as soon as an
5955 emitted event passes one of them, LTTng can record the event. For
5956 example, an event named `my_app:my_tracepoint` emitted from a user space
5957 tracepoint with a `TRACE_ERROR` log level passes both of the following
5962 $ lttng enable-event --userspace my_app:my_tracepoint
5963 $ lttng enable-event --userspace my_app:my_tracepoint \
5964 --loglevel=TRACE_INFO
5967 The second event rule is redundant: the first one includes
5971 [[disable-event-rule]]
5972 === Disable an event rule
5974 To disable an event rule that you <<enabling-disabling-events,created>>
5975 previously, use the man:lttng-disable-event(1) command. This command
5976 disables _all_ the event rules (of a given tracing domain and channel)
5977 which match an instrumentation point. The other conditions aren't
5978 supported as of LTTng{nbsp}{revision}.
5980 The LTTng tracer doesn't record an emitted event which passes
5981 a _disabled_ event rule.
5983 .Disable an event rule matching a Python logger (default channel).
5987 $ lttng disable-event --python my-logger
5991 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5995 $ lttng disable-event --jul '*'
5999 .Disable _all_ the event rules of the default channel.
6001 The opt:lttng-disable-event(1):--all-events option isn't, like the
6002 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
6003 equivalent of the event name `*` (wildcard): it disables _all_ the event
6004 rules of a given channel.
6008 $ lttng disable-event --jul --all-events
6012 NOTE: You can't delete an event rule once you create it.
6016 === Get the status of a tracing session
6018 To get the status of the <<cur-tracing-session,current tracing
6019 session>>, that is, its parameters, its channels, event rules, and their
6022 * Use the man:lttng-status(1) command:
6031 To get the status of any tracing session:
6033 * Use the man:lttng-list(1) command with the name of the tracing
6039 $ lttng list my-session
6043 Replace `my-session` with the desired tracing session name.
6046 [[basic-tracing-session-control]]
6047 === Start and stop a tracing session
6049 Once you <<creating-destroying-tracing-sessions,create a tracing
6051 <<enabling-disabling-events,create one or more event rules>>,
6052 you can start and stop the tracers for this tracing session.
6054 To start tracing in the <<cur-tracing-session,current tracing session>>:
6056 * Use the man:lttng-start(1) command:
6065 LTTng is very flexible: you can launch user applications before
6066 or after the you start the tracers. The tracers only record the events
6067 if they pass enabled event rules and if they occur while the tracers are
6070 To stop tracing in the current tracing session:
6072 * Use the man:lttng-stop(1) command:
6081 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6082 records>> or lost sub-buffers since the last time you ran
6083 man:lttng-start(1), warnings are printed when you run the
6084 man:lttng-stop(1) command.
6086 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
6087 trace data and make the trace readable. Note that the
6088 man:lttng-destroy(1) command (see
6089 <<creating-destroying-tracing-sessions,Create and destroy a tracing
6090 session>>) also runs the man:lttng-stop(1) command implicitly.
6094 === Clear a tracing session
6096 You might need to remove all the current tracing data of one or more
6097 <<tracing-session,tracing sessions>> between multiple attempts to
6098 reproduce a problem without interrupting the LTTng tracing activity.
6100 To clear the tracing data of the
6101 <<cur-tracing-session,current tracing session>>:
6103 * Use the man:lttng-clear(1) command:
6112 To clear the tracing data of all the tracing sessions:
6114 * Use the `lttng clear` command with the opt:lttng-clear(1):--all
6125 [[enabling-disabling-channels]]
6126 === Create a channel
6128 Once you create a tracing session, you can create a <<channel,channel>>
6129 with the man:lttng-enable-channel(1) command.
6131 Note that LTTng automatically creates a default channel when, for a
6132 given <<domain,tracing domain>>, no channels exist and you
6133 <<enabling-disabling-events,create>> the first event rule. This default
6134 channel is named `channel0` and its attributes are set to reasonable
6135 values. Therefore, you only need to create a channel when you need
6136 non-default attributes.
6138 You specify each non-default channel attribute with a command-line
6139 option when you use the man:lttng-enable-channel(1) command. The
6140 available command-line options are:
6142 [role="growable",cols="asciidoc,asciidoc"]
6143 .Command-line options for the man:lttng-enable-channel(1) command.
6145 |Option |Description
6151 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
6152 of the default _discard_ mode.
6154 |`--buffers-pid` (user space tracing domain only)
6157 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6158 instead of the default per-user buffering scheme.
6160 |+--subbuf-size=__SIZE__+
6163 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6164 either for each Unix user (default), or for each instrumented process.
6166 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6168 |+--num-subbuf=__COUNT__+
6171 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6172 for each Unix user (default), or for each instrumented process.
6174 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6176 |+--tracefile-size=__SIZE__+
6179 Set the maximum size of each trace file that this channel writes within
6180 a stream to +__SIZE__+ bytes instead of no maximum.
6182 See <<tracefile-rotation,Trace file count and size>>.
6184 |+--tracefile-count=__COUNT__+
6187 Limit the number of trace files that this channel creates to
6188 +__COUNT__+ channels instead of no limit.
6190 See <<tracefile-rotation,Trace file count and size>>.
6192 |+--switch-timer=__PERIODUS__+
6195 Set the <<channel-switch-timer,switch timer period>>
6196 to +__PERIODUS__+{nbsp}µs.
6198 |+--read-timer=__PERIODUS__+
6201 Set the <<channel-read-timer,read timer period>>
6202 to +__PERIODUS__+{nbsp}µs.
6204 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6207 Set the timeout of user space applications which load LTTng-UST
6208 in blocking mode to +__TIMEOUTUS__+:
6211 Never block (non-blocking mode).
6214 Block forever until space is available in a sub-buffer to record
6217 __n__, a positive value::
6218 Wait for at most __n__ µs when trying to write into a sub-buffer.
6220 Note that, for this option to have any effect on an instrumented
6221 user space application, you need to run the application with a set
6222 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6224 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6227 Set the output type of the channel to +__TYPE__+, either `mmap` or
6232 You can only create a channel in the Linux kernel and user space
6233 <<domain,tracing domains>>: other tracing domains have their own channel
6234 created on the fly when <<enabling-disabling-events,creating event
6239 Because of a current LTTng limitation, you must create all channels
6240 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6241 tracing session, that is, before the first time you run
6244 Since LTTng automatically creates a default channel when you use the
6245 man:lttng-enable-event(1) command with a specific tracing domain, you
6246 can't, for example, create a Linux kernel event rule, start tracing,
6247 and then create a user space event rule, because no user space channel
6248 exists yet and it's too late to create one.
6250 For this reason, make sure to configure your channels properly
6251 before starting the tracers for the first time!
6254 The following examples show how to combine the previous
6255 command-line options to create simple to more complex channels.
6257 .Create a Linux kernel channel with default attributes.
6261 $ lttng enable-channel --kernel my-channel
6265 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6269 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6270 --buffers-pid my-channel
6274 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6276 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6277 create the channel, <<enabling-disabling-events,create an event rule>>,
6278 and <<basic-tracing-session-control,start tracing>>:
6283 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6284 $ lttng enable-event --userspace --channel=blocking-channel --all
6288 Run an application instrumented with LTTng-UST and allow it to block:
6292 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6296 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6300 $ lttng enable-channel --kernel --tracefile-count=8 \
6301 --tracefile-size=4194304 my-channel
6305 .Create a user space channel in overwrite (or _flight recorder_) mode.
6309 $ lttng enable-channel --userspace --overwrite my-channel
6313 <<enabling-disabling-events,Create>> the same event rule in
6314 two different channels:
6318 $ lttng enable-event --userspace --channel=my-channel app:tp
6319 $ lttng enable-event --userspace --channel=other-channel app:tp
6322 If both channels are enabled, when a tracepoint named `app:tp` is
6323 reached, LTTng records two events, one for each channel.
6327 === Disable a channel
6329 To disable a specific channel that you <<enabling-disabling-channels,created>>
6330 previously, use the man:lttng-disable-channel(1) command.
6332 .Disable a specific Linux kernel channel.
6336 $ lttng disable-channel --kernel my-channel
6340 The state of a channel precedes the individual states of event rules
6341 attached to it: event rules which belong to a disabled channel, even if
6342 they are enabled, are also considered disabled.
6346 === Add context fields to a channel
6348 Event record fields in trace files provide important information about
6349 events that occured previously, but sometimes some external context may
6350 help you solve a problem faster.
6352 Examples of context fields are:
6354 * The **process ID**, **thread ID**, **process name**, and
6355 **process priority** of the thread in which the event occurs.
6356 * The **hostname** of the system on which the event occurs.
6357 * The Linux kernel and user call stacks (since
6359 * The current values of many possible **performance counters** using
6361 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6363 ** Branch instructions, misses, and loads.
6365 * Any context defined at the application level (supported for the
6366 JUL and log4j <<domain,tracing domains>>).
6368 To get the full list of available context fields, see
6369 `lttng add-context --list`. Some context fields are reserved for a
6370 specific <<domain,tracing domain>> (Linux kernel or user space).
6372 You add context fields to <<channel,channels>>. All the events
6373 that a channel with added context fields records contain those fields.
6375 To add context fields to one or all the channels of a given tracing
6378 * Use the man:lttng-add-context(1) command.
6380 .Add context fields to all the channels of the current tracing session.
6382 The following command line adds the virtual process identifier and
6383 the per-thread CPU cycles count fields to all the user space channels
6385 <<cur-tracing-session,current tracing session>>.
6389 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6393 .Add performance counter context fields by raw ID
6395 See man:lttng-add-context(1) for the exact format of the context field
6396 type, which is partly compatible with the format used in
6401 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6402 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6406 .Add context fields to a specific channel.
6408 The following command line adds the thread identifier and user call
6409 stack context fields to the Linux kernel channel named `my-channel` in
6410 the current tracing session.
6414 $ lttng add-context --kernel --channel=my-channel \
6415 --type=tid --type=callstack-user
6419 .Add an application-specific context field to a specific channel.
6421 The following command line adds the `cur_msg_id` context field of the
6422 `retriever` context retriever for all the instrumented
6423 <<java-application,Java applications>> recording <<event,event records>>
6424 in the channel named `my-channel`:
6428 $ lttng add-context --kernel --channel=my-channel \
6429 --type='$app:retriever:cur_msg_id'
6432 IMPORTANT: Make sure to always quote the `$` character when you
6433 use man:lttng-add-context(1) from a shell.
6436 NOTE: You can't remove context fields from a channel once you add it.
6441 === Track process attributes
6443 It's often useful to only allow processes with specific attributes to
6444 emit events. For example, you may wish to record all the system calls
6445 which a given process makes (à la
6446 http://linux.die.net/man/1/strace[strace]).
6448 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6449 purpose. Both commands operate on _inclusion sets_ of process attribute
6450 values. The available process attribute types are:
6452 Linux kernel <<domain,tracing domain>> only::
6456 * Virtual process ID (VPID).
6458 This is the PID as seen by the application.
6460 * Unix user ID (UID) (since LTTng{nbsp}2.12).
6462 * Virtual Unix user ID (VUID) (since LTTng{nbsp}2.12).
6464 This is the UID as seen by the application.
6466 * Unix group ID (GID) (since LTTng{nbsp}2.12).
6468 * Virtual Unix group ID (VGID) (since LTTng{nbsp}2.12).
6470 This is the GID as seen by the application.
6473 User space tracing domain::
6476 * VUID (since LTTng{nbsp}2.12).
6477 * VGID (since LTTng{nbsp}2.12).
6479 Each tracing domain has one inclusion set per process attribute type:
6480 the Linux kernel tracing domain has six while the user space tracing
6483 For a given event which passes an enabled <<event,event rule>> to be
6484 recorded, _all_ the attributes of its executing process must be part of
6485 the inclusion sets of the tracing domain of the event rule.
6487 Add entries to an inclusion set with the man:lttng-track(1) command and
6488 remove entries with the man:lttng-untrack(1) command. A process
6489 attribute is _tracked_ when it's part of an inclusion set and
6490 _untracked_ otherwise.
6494 The process attribute values are _numeric_.
6496 Should a process with a given tracked process ID, for example, exit, and
6497 then a new process be given this ID, then the latter would also be
6498 allowed to emit events.
6500 With the `lttng track` command, you can add Unix user and group _names_
6501 to the user and group inclusion sets: the <<lttng-sessiond,session
6502 daemon>> finds the corresponding UID, VUID, GID, or VGID once on
6503 _addition_ to the inclusion set. This means that if you rename the user
6504 or group after you run `lttng track`, its user/group ID remains tracked.
6507 .Track and untrack virtual process IDs.
6509 For the sake of the following example, assume the target system has
6510 16{nbsp}possible VPIDs.
6513 <<creating-destroying-tracing-sessions,create a tracing session>>,
6514 the user space VPID inclusion set contains _all_ the possible VPIDs:
6517 .All VPIDs are tracked.
6518 image::track-all.png[]
6520 When the inclusion set is full and you use the man:lttng-track(1)
6521 command to specify some VPIDs to track, LTTng first clears the inclusion
6522 set, and then it adds the specific VPIDs to track. After:
6526 $ lttng track --userspace --vpid=3,4,7,10,13
6529 the VPID inclusion set is:
6532 .VPIDs 3, 4, 7, 10, and 13 are tracked.
6533 image::track-3-4-7-10-13.png[]
6535 Add more VPIDs to the inclusion set afterwards:
6539 $ lttng track --userspace --vpid=1,15,16
6545 .VPIDs 1, 15, and 16 are added to the inclusion set.
6546 image::track-1-3-4-7-10-13-15-16.png[]
6548 The man:lttng-untrack(1) command removes entries from process attribute
6549 inclusion sets. Given the previous example, the following command:
6553 $ lttng untrack --userspace --vpid=3,7,10,13
6556 leads to this VPID inclusion set:
6559 .VPIDs 3, 7, 10, and 13 are removed from the inclusion set.
6560 image::track-1-4-15-16.png[]
6562 LTTng can track all the possible VPIDs again using the
6563 opt:lttng-track(1):--all option:
6567 $ lttng track --userspace --vpid --all
6570 The result is, again:
6573 .All VPIDs are tracked.
6574 image::track-all.png[]
6577 .Track only specific process attributes.
6579 A typical use case with process attribute tracking is to start with an
6580 empty inclusion set, then <<basic-tracing-session-control,start the
6581 tracers>>, and then add entries manually while the tracers are active.
6583 Use the opt:lttng-untrack(1):--all option of the
6584 man:lttng-untrack(1) command to clear the inclusion set after you
6585 <<creating-destroying-tracing-sessions,create a tracing session>>, for
6586 example (with UIDs):
6590 $ lttng untrack --kernel --uid --all
6596 .No UIDs are tracked.
6597 image::untrack-all.png[]
6599 If you trace with this inclusion set configuration, the LTTng kernel
6600 tracer records no events within the <<cur-tracing-session,current
6601 tracing session>> because it doesn't track any UID. Use the
6602 man:lttng-track(1) command as usual to track specific UIDs when you need
6607 $ lttng track --kernel --uid=http,11
6613 .UIDs 6 (`http`) and 11 are tracked.
6614 image::track-6-11.png[]
6619 [[saving-loading-tracing-session]]
6620 === Save and load tracing session configurations
6622 Configuring a <<tracing-session,tracing session>> can be long. Some of
6623 the tasks involved are:
6625 * <<enabling-disabling-channels,Create channels>> with
6626 specific attributes.
6627 * <<adding-context,Add context fields>> to specific channels.
6628 * <<enabling-disabling-events,Create event rules>> with specific log
6629 level and filter conditions.
6631 If you use LTTng to solve real world problems, chances are you have to
6632 record events using the same tracing session setup over and over,
6633 modifying a few variables each time in your instrumented program
6634 or environment. To avoid constant tracing session reconfiguration,
6635 the man:lttng(1) command-line tool can save and load tracing session
6636 configurations to/from XML files.
6638 To save a given tracing session configuration:
6640 * Use the man:lttng-save(1) command:
6645 $ lttng save my-session
6649 Replace `my-session` with the name of the tracing session to save.
6651 LTTng saves tracing session configurations to
6652 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6653 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6654 the opt:lttng-save(1):--output-path option to change this destination
6657 LTTng saves all configuration parameters, for example:
6659 * The tracing session name.
6660 * The trace data output path.
6661 * The channels with their state and all their attributes.
6662 * The context fields you added to channels.
6663 * The event rules with their state, log level and filter conditions.
6665 To load a tracing session:
6667 * Use the man:lttng-load(1) command:
6672 $ lttng load my-session
6676 Replace `my-session` with the name of the tracing session to load.
6678 When LTTng loads a configuration, it restores your saved tracing session
6679 as if you just configured it manually.
6681 See man:lttng-load(1) for the complete list of command-line options. You
6682 can also save and load many sessions at a time, and decide in which
6683 directory to output the XML files.
6686 [[sending-trace-data-over-the-network]]
6687 === Send trace data over the network
6689 LTTng can send the recorded trace data to a remote system over the
6690 network instead of writing it to the local file system.
6692 To send the trace data over the network:
6694 . On the _remote_ system (which can also be the target system),
6695 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6704 . On the _target_ system, create a tracing session configured to
6705 send trace data over the network:
6710 $ lttng create my-session --set-url=net://remote-system
6714 Replace `remote-system` by the host name or IP address of the
6715 remote system. See man:lttng-create(1) for the exact URL format.
6717 . On the target system, use the man:lttng(1) command-line tool as usual.
6718 When tracing is active, the consumer daemon of the target sends
6719 sub-buffers to the relay daemon running on the remote system instead
6720 of flushing them to the local file system. The relay daemon writes the
6721 received packets to the local file system.
6723 The relay daemon writes trace files to
6724 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6725 +__hostname__+ is the host name of the target system and +__session__+
6726 is the tracing session name. Note that the env:LTTNG_HOME environment
6727 variable defaults to `$HOME` if not set. Use the
6728 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6729 trace files to another base directory.
6734 === View events as LTTng emits them (noch:{LTTng} live)
6736 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6737 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6738 display events as LTTng emits them on the target system while tracing is
6741 The relay daemon creates a _tee_: it forwards the trace data to both
6742 the local file system and to connected live viewers:
6745 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6750 . On the _target system_, create a <<tracing-session,tracing session>>
6756 $ lttng create my-session --live
6760 This spawns a local relay daemon.
6762 . Start the live viewer and configure it to connect to the relay
6763 daemon. For example, with
6764 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]:
6769 $ babeltrace2 net://localhost/host/hostname/my-session
6776 * `hostname` with the host name of the target system.
6777 * `my-session` with the name of the tracing session to view.
6780 . Configure the tracing session as usual with the man:lttng(1)
6781 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6783 List the available live tracing sessions with Babeltrace{nbsp}2:
6787 $ babeltrace2 net://localhost
6790 You can start the relay daemon on another system. In this case, you need
6791 to specify the URL of the relay daemon when you create the tracing
6792 session with the opt:lttng-create(1):--set-url option. You also need to
6793 replace `localhost` in the procedure above with the host name of the
6794 system on which the relay daemon is running.
6796 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6797 command-line options.
6801 [[taking-a-snapshot]]
6802 === Take a snapshot of the current sub-buffers of a tracing session
6804 The normal behavior of LTTng is to append full sub-buffers to growing
6805 trace data files. This is ideal to keep a full history of the events
6806 that occurred on the target system, but it can
6807 represent too much data in some situations. For example, you may wish
6808 to trace your application continuously until some critical situation
6809 happens, in which case you only need the latest few recorded
6810 events to perform the desired analysis, not multi-gigabyte trace files.
6812 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6813 current sub-buffers of a given <<tracing-session,tracing session>>.
6814 LTTng can write the snapshot to the local file system or send it over
6818 .A snapshot is a copy of the current sub-buffers, which aren't cleared after the operation.
6819 image::snapshot.png[]
6821 If you wish to create unmanaged, self-contained, non-overlapping
6822 trace chunk archives instead of a simple copy of the current
6823 sub-buffers, see the <<session-rotation,tracing session rotation>>
6824 feature (available since LTTng{nbsp}2.11).
6828 . Create a tracing session in _snapshot mode_:
6833 $ lttng create my-session --snapshot
6837 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6838 <<channel,channels>> created in this mode is automatically set to
6839 _overwrite_ (flight recorder mode).
6841 . Configure the tracing session as usual with the man:lttng(1)
6842 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6844 . **Optional**: When you need to take a snapshot,
6845 <<basic-tracing-session-control,stop tracing>>.
6847 You can take a snapshot when the tracers are active, but if you stop
6848 them first, you're sure that the data in the sub-buffers doesn't
6849 change before you actually take the snapshot.
6856 $ lttng snapshot record --name=my-first-snapshot
6860 LTTng writes the current sub-buffers of all the channels of the
6861 <<cur-tracing-session,current tracing session>> to
6862 trace files on the local file system. Those trace files have
6863 `my-first-snapshot` in their name.
6865 There is no difference between the format of a normal trace file and the
6866 format of a snapshot: viewers of LTTng traces also support LTTng
6869 By default, LTTng writes snapshot files to the path shown by
6870 `lttng snapshot list-output`. You can change this path or decide to send
6871 snapshots over the network using either:
6873 . An output path or URL that you specify when you
6874 <<creating-destroying-tracing-sessions,create the tracing session>>.
6875 . A snapshot output path or URL that you add using
6876 `lttng snapshot add-output`.
6877 . An output path or URL that you provide directly to the
6878 `lttng snapshot record` command.
6880 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6881 you specify a URL, a relay daemon must listen on a remote system (see
6882 <<sending-trace-data-over-the-network,Send trace data over the
6887 [[session-rotation]]
6888 === Archive the current trace chunk (rotate a tracing session)
6890 The <<taking-a-snapshot,snapshot user guide>> shows how to dump the
6891 current sub-buffers of a tracing session to the file system or send them
6892 over the network. When you take a snapshot, LTTng doesn't clear the ring
6893 buffers of the tracing session: if you take another snapshot immediately
6894 after, both snapshots could contain overlapping trace data.
6896 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6897 _tracing session rotation_ is a feature which appends the content of the
6898 ring buffers to what's already on the file system or sent over the
6899 network since the creation of the tracing session or since the last
6900 rotation, and then clears those ring buffers to avoid trace data
6903 What LTTng is about to write when performing a tracing session rotation
6904 is called the _current trace chunk_. When this current trace chunk is
6905 written to the file system or sent over the network, it becomes a _trace
6906 chunk archive_. Therefore, a tracing session rotation _archives_ the
6907 current trace chunk.
6910 .A tracing session rotation operation _archives_ the current trace chunk.
6911 image::rotation.png[]
6913 A trace chunk archive is a self-contained LTTng trace which LTTng
6914 doesn't manage anymore: you can read it, modify it, move it, or remove
6917 There are two methods to perform a tracing session rotation: immediately
6918 or with a rotation schedule.
6920 To perform an immediate tracing session rotation:
6922 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6923 in _normal mode_ or _network streaming mode_
6924 (only those two creation modes support tracing session rotation):
6929 $ lttng create my-session
6933 . <<enabling-disabling-events,Create one or more event rules>>
6934 and <<basic-tracing-session-control,start tracing>>:
6939 $ lttng enable-event --kernel sched_'*'
6944 . When needed, immediately rotate the
6945 <<cur-tracing-session,current tracing session>>:
6954 The cmd:lttng-rotate command prints the path to the created trace
6955 chunk archive. See man:lttng-rotate(1) to learn about the format
6956 of trace chunk archive directory names.
6958 Perform other immediate rotations while the tracing session is
6959 active. It is guaranteed that all the trace chunk archives don't
6960 contain overlapping trace data. You can also perform an immediate
6961 rotation once you have <<basic-tracing-session-control,stopped>> the
6964 . When you're done tracing,
6965 <<creating-destroying-tracing-sessions,destroy the current tracing
6975 The tracing session destruction operation creates one last trace
6976 chunk archive from the current trace chunk.
6978 A tracing session rotation schedule is a planned rotation which LTTng
6979 performs automatically based on one of the following conditions:
6981 * A timer with a configured period times out.
6983 * The total size of the flushed part of the current trace chunk
6984 becomes greater than or equal to a configured value.
6986 To schedule a tracing session rotation, set a _rotation schedule_:
6988 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6989 in _normal mode_ or _network streaming mode_
6990 (only those two creation modes support tracing session rotation):
6995 $ lttng create my-session
6999 . <<enabling-disabling-events,Create one or more event rules>>:
7004 $ lttng enable-event --kernel sched_'*'
7008 . Set a tracing session rotation schedule:
7013 $ lttng enable-rotation --timer=10s
7017 In this example, we set a rotation schedule so that LTTng performs a
7018 tracing session rotation every ten seconds.
7020 See man:lttng-enable-rotation(1) to learn more about other ways to set a
7023 . <<basic-tracing-session-control,Start tracing>>:
7032 LTTng performs tracing session rotations automatically while the tracing
7033 session is active thanks to the rotation schedule.
7035 . When you're done tracing,
7036 <<creating-destroying-tracing-sessions,destroy the current tracing
7046 The tracing session destruction operation creates one last trace chunk
7047 archive from the current trace chunk.
7049 Use man:lttng-disable-rotation(1) to unset a tracing session
7052 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
7053 limitations regarding those two commands.
7058 === Use the machine interface
7060 With any command of the man:lttng(1) command-line tool, set the
7061 opt:lttng(1):--mi option to `xml` (before the command name) to get an
7062 XML machine interface output, for example:
7066 $ lttng --mi=xml enable-event --kernel --syscall open
7069 A schema definition (XSD) is
7070 https://github.com/lttng/lttng-tools/blob/stable-{revision}/src/common/src/common/mi-lttng-4.0.xsd[available]
7071 to ease the integration with external tools as much as possible.
7075 [[metadata-regenerate]]
7076 === Regenerate the metadata of an LTTng trace
7078 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
7079 data stream files and a metadata file. This metadata file contains,
7080 amongst other things, information about the offset of the clock sources
7081 used to timestamp <<event,event records>> when tracing.
7083 If, once a <<tracing-session,tracing session>> is
7084 <<basic-tracing-session-control,started>>, a major
7085 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
7086 happens, the clock offset of the trace also needs to be updated. Use
7087 the `metadata` item of the man:lttng-regenerate(1) command to do so.
7089 The main use case of this command is to allow a system to boot with
7090 an incorrect wall time and trace it with LTTng before its wall time
7091 is corrected. Once the system is known to be in a state where its
7092 wall time is correct, it can run `lttng regenerate metadata`.
7094 To regenerate the metadata of an LTTng trace:
7096 * Use the `metadata` item of the man:lttng-regenerate(1) command:
7101 $ lttng regenerate metadata
7107 `lttng regenerate metadata` has the following limitations:
7109 * Tracing session <<creating-destroying-tracing-sessions,created>>
7111 * User space <<channel,channels>>, if any, are using
7112 <<channel-buffering-schemes,per-user buffering>>.
7117 [[regenerate-statedump]]
7118 === Regenerate the state dump of a tracing session
7120 The LTTng kernel and user space tracers generate state dump
7121 <<event,event records>> when the application starts or when you
7122 <<basic-tracing-session-control,start a tracing session>>. An analysis
7123 can use the state dump event records to set an initial state before it
7124 builds the rest of the state from the following event records.
7125 http://tracecompass.org/[Trace Compass] is a notable example of an
7126 application which uses the state dump of an LTTng trace.
7128 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
7129 state dump event records aren't included in the snapshot because they
7130 were recorded to a sub-buffer that has been consumed or overwritten
7133 Use the `lttng regenerate statedump` command to emit the state
7134 dump event records again.
7136 To regenerate the state dump of the current tracing session, provided
7137 create it in snapshot mode, before you take a snapshot:
7139 . Use the `statedump` item of the man:lttng-regenerate(1) command:
7144 $ lttng regenerate statedump
7148 . <<basic-tracing-session-control,Stop the tracing session>>:
7157 . <<taking-a-snapshot,Take a snapshot>>:
7162 $ lttng snapshot record --name=my-snapshot
7166 Depending on the event throughput, you should run steps 1 and 2
7167 as closely as possible.
7169 NOTE: To record the state dump events, you need to
7170 <<enabling-disabling-events,create event rules>> which enable them.
7171 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
7172 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
7176 [[persistent-memory-file-systems]]
7177 === Record trace data on persistent memory file systems
7179 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
7180 (NVRAM) is random-access memory that retains its information when power
7181 is turned off (non-volatile). Systems with such memory can store data
7182 structures in RAM and retrieve them after a reboot, without flushing
7183 to typical _storage_.
7185 Linux supports NVRAM file systems thanks to either
7186 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
7187 (requires Linux{nbsp}4.1+) or http://pramfs.sourceforge.net/[PRAMFS] (requires Linux{nbsp}<{nbsp}4).
7189 This section doesn't describe how to operate such file systems;
7190 we assume that you have a working persistent memory file system.
7192 When you create a <<tracing-session,tracing session>>, you can specify
7193 the path of the shared memory holding the sub-buffers. If you specify a
7194 location on an NVRAM file system, then you can retrieve the latest
7195 recorded trace data when the system reboots after a crash.
7197 To record trace data on a persistent memory file system and retrieve the
7198 trace data after a system crash:
7200 . Create a tracing session with a sub-buffer shared memory path located
7201 on an NVRAM file system:
7206 $ lttng create my-session --shm-path=/path/to/shm
7210 . Configure the tracing session as usual with the man:lttng(1)
7211 command-line tool, and <<basic-tracing-session-control,start tracing>>.
7213 . After a system crash, use the man:lttng-crash(1) command-line tool to
7214 view the trace data recorded on the NVRAM file system:
7219 $ lttng-crash /path/to/shm
7223 The binary layout of the ring buffer files isn't exactly the same as
7224 the trace files layout. This is why you need to use man:lttng-crash(1)
7225 instead of your preferred trace viewer directly.
7227 To convert the ring buffer files to LTTng trace files:
7229 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
7234 $ lttng-crash --extract=/path/to/trace /path/to/shm
7240 [[notif-trigger-api]]
7241 === Get notified when the buffer usage of a channel is too high or too low
7243 With the $$C/C++$$ notification and trigger API of LTTng, your user
7244 application can get notified when the buffer usage of one or more
7245 <<channel,channels>> becomes too low or too high. Use this API
7246 and enable or disable <<event,event rules>> during tracing to avoid
7247 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7249 .Have a user application get notified when the buffer usage of an LTTng channel is too high.
7251 In this example, we create and build an application which gets notified
7252 when the buffer usage of a specific LTTng channel is higher than
7253 75{nbsp}%. We only print that it is the case in the example, but we
7254 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7255 disable event rules when this happens.
7257 . Create the C{nbsp}source file of application:
7265 #include <lttng/domain.h>
7266 #include <lttng/action/action.h>
7267 #include <lttng/action/notify.h>
7268 #include <lttng/condition/condition.h>
7269 #include <lttng/condition/buffer-usage.h>
7270 #include <lttng/condition/evaluation.h>
7271 #include <lttng/notification/channel.h>
7272 #include <lttng/notification/notification.h>
7273 #include <lttng/trigger/trigger.h>
7274 #include <lttng/endpoint.h>
7276 int main(int argc, char *argv[])
7278 int exit_status = 0;
7279 struct lttng_notification_channel *notification_channel;
7280 struct lttng_condition *condition;
7281 struct lttng_action *action;
7282 struct lttng_trigger *trigger;
7283 const char *tracing_session_name;
7284 const char *channel_name;
7287 tracing_session_name = argv[1];
7288 channel_name = argv[2];
7291 * Create a notification channel. A notification channel
7292 * connects the user application to the LTTng session daemon.
7293 * This notification channel can be used to listen to various
7294 * types of notifications.
7296 notification_channel = lttng_notification_channel_create(
7297 lttng_session_daemon_notification_endpoint);
7300 * Create a "high buffer usage" condition. In this case, the
7301 * condition is reached when the buffer usage is greater than or
7302 * equal to 75 %. We create the condition for a specific tracing
7303 * session name, channel name, and for the user space tracing
7306 * The "low buffer usage" condition type also exists.
7308 condition = lttng_condition_buffer_usage_high_create();
7309 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7310 lttng_condition_buffer_usage_set_session_name(
7311 condition, tracing_session_name);
7312 lttng_condition_buffer_usage_set_channel_name(condition,
7314 lttng_condition_buffer_usage_set_domain_type(condition,
7318 * Create an action (get a notification) to take when the
7319 * condition created above is reached.
7321 action = lttng_action_notify_create();
7324 * Create a trigger. A trigger associates a condition to an
7325 * action: the action is executed when the condition is reached.
7327 trigger = lttng_trigger_create(condition, action);
7329 /* Register the trigger to LTTng. */
7330 lttng_register_trigger(trigger);
7333 * Now that we have registered a trigger, a notification will be
7334 * emitted everytime its condition is met. To receive this
7335 * notification, we must subscribe to notifications that match
7336 * the same condition.
7338 lttng_notification_channel_subscribe(notification_channel,
7342 * Notification loop. Put this in a dedicated thread to avoid
7343 * blocking the main thread.
7346 struct lttng_notification *notification;
7347 enum lttng_notification_channel_status status;
7348 const struct lttng_evaluation *notification_evaluation;
7349 const struct lttng_condition *notification_condition;
7350 double buffer_usage;
7352 /* Receive the next notification. */
7353 status = lttng_notification_channel_get_next_notification(
7354 notification_channel, ¬ification);
7357 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7359 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7361 * The session daemon can drop notifications if a monitoring
7362 * application isn't consuming the notifications fast
7366 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7368 * The notification channel has been closed by the
7369 * session daemon. This is typically caused by a session
7370 * daemon shutting down.
7374 /* Unhandled conditions or errors. */
7380 * A notification provides, amongst other things:
7382 * * The condition that caused this notification to be
7384 * * The condition evaluation, which provides more
7385 * specific information on the evaluation of the
7388 * The condition evaluation provides the buffer usage
7389 * value at the moment the condition was reached.
7391 notification_condition = lttng_notification_get_condition(
7393 notification_evaluation = lttng_notification_get_evaluation(
7396 /* We're subscribed to only one condition. */
7397 assert(lttng_condition_get_type(notification_condition) ==
7398 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7401 * Get the exact sampled buffer usage from the
7402 * condition evaluation.
7404 lttng_evaluation_buffer_usage_get_usage_ratio(
7405 notification_evaluation, &buffer_usage);
7408 * At this point, instead of printing a message, we
7409 * could do something to reduce the buffer usage of the channel,
7410 * like disable specific events.
7412 printf("Buffer usage is %f %% in tracing session \"%s\", "
7413 "user space channel \"%s\".\n", buffer_usage * 100,
7414 tracing_session_name, channel_name);
7415 lttng_notification_destroy(notification);
7419 lttng_action_destroy(action);
7420 lttng_condition_destroy(condition);
7421 lttng_trigger_destroy(trigger);
7422 lttng_notification_channel_destroy(notification_channel);
7428 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7433 $ gcc -o notif-app notif-app.c -llttng-ctl
7437 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7438 <<enabling-disabling-events,create an event rule>> matching all the
7439 user space tracepoints, and
7440 <<basic-tracing-session-control,start tracing>>:
7445 $ lttng create my-session
7446 $ lttng enable-event --userspace --all
7451 If you create the channel manually with the man:lttng-enable-channel(1)
7452 command, control how frequently LTTng samples the current values of the
7453 channel properties to evaluate user conditions with the
7454 opt:lttng-enable-channel(1):--monitor-timer option.
7456 . Run the `notif-app` application. This program accepts the
7457 <<tracing-session,tracing session>> name and the user space channel
7458 name as its two first arguments. The channel which LTTng automatically
7459 creates with the man:lttng-enable-event(1) command above is named
7465 $ ./notif-app my-session channel0
7469 . In another terminal, run an application with a very high event
7470 throughput so that the 75{nbsp}% buffer usage condition is reached.
7472 In the first terminal, the application should print lines like this:
7475 Buffer usage is 81.45197 % in tracing session "my-session", user space
7479 If you don't see anything, try modifying the condition in
7480 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7481 (step{nbsp}2) and running it again (step{nbsp}4).
7488 [[lttng-modules-ref]]
7489 === noch:{LTTng-modules}
7493 [[lttng-tracepoint-enum]]
7494 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7496 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7500 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7505 * `name` with the name of the enumeration (C identifier, unique
7506 amongst all the defined enumerations).
7507 * `entries` with a list of enumeration entries.
7509 The available enumeration entry macros are:
7511 +ctf_enum_value(__name__, __value__)+::
7512 Entry named +__name__+ mapped to the integral value +__value__+.
7514 +ctf_enum_range(__name__, __begin__, __end__)+::
7515 Entry named +__name__+ mapped to the range of integral values between
7516 +__begin__+ (included) and +__end__+ (included).
7518 +ctf_enum_auto(__name__)+::
7519 Entry named +__name__+ mapped to the integral value following the
7522 The last value of a `ctf_enum_value()` entry is its +__value__+
7525 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7527 If `ctf_enum_auto()` is the first entry in the list, its integral
7530 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7531 to use a defined enumeration as a tracepoint field.
7533 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7537 LTTNG_TRACEPOINT_ENUM(
7540 ctf_enum_auto("AUTO: EXPECT 0")
7541 ctf_enum_value("VALUE: 23", 23)
7542 ctf_enum_value("VALUE: 27", 27)
7543 ctf_enum_auto("AUTO: EXPECT 28")
7544 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7545 ctf_enum_auto("AUTO: EXPECT 304")
7553 [[lttng-modules-tp-fields]]
7554 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7556 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7557 tracepoint fields, which must be listed within `TP_FIELDS()` in
7558 `LTTNG_TRACEPOINT_EVENT()`, are:
7560 [role="func-desc growable",cols="asciidoc,asciidoc"]
7561 .Available macros to define LTTng-modules tracepoint fields
7563 |Macro |Description and parameters
7566 +ctf_integer(__t__, __n__, __e__)+
7568 +ctf_integer_nowrite(__t__, __n__, __e__)+
7570 +ctf_user_integer(__t__, __n__, __e__)+
7572 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7574 Standard integer, displayed in base{nbsp}10.
7577 Integer C type (`int`, `long`, `size_t`, ...).
7583 Argument expression.
7586 +ctf_integer_hex(__t__, __n__, __e__)+
7588 +ctf_user_integer_hex(__t__, __n__, __e__)+
7590 Standard integer, displayed in base{nbsp}16.
7599 Argument expression.
7601 |+ctf_integer_oct(__t__, __n__, __e__)+
7603 Standard integer, displayed in base{nbsp}8.
7612 Argument expression.
7615 +ctf_integer_network(__t__, __n__, __e__)+
7617 +ctf_user_integer_network(__t__, __n__, __e__)+
7619 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7628 Argument expression.
7631 +ctf_integer_network_hex(__t__, __n__, __e__)+
7633 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7635 Integer in network byte order, displayed in base{nbsp}16.
7644 Argument expression.
7647 +ctf_enum(__N__, __t__, __n__, __e__)+
7649 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7651 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7653 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7658 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7661 Integer C type (`int`, `long`, `size_t`, ...).
7667 Argument expression.
7670 +ctf_string(__n__, __e__)+
7672 +ctf_string_nowrite(__n__, __e__)+
7674 +ctf_user_string(__n__, __e__)+
7676 +ctf_user_string_nowrite(__n__, __e__)+
7678 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7684 Argument expression.
7687 +ctf_array(__t__, __n__, __e__, __s__)+
7689 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7691 +ctf_user_array(__t__, __n__, __e__, __s__)+
7693 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7695 Statically-sized array of integers.
7698 Array element C type.
7704 Argument expression.
7710 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7712 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7714 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7716 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7718 Statically-sized array of bits.
7720 The type of +__e__+ must be an integer type. +__s__+ is the number
7721 of elements of such type in +__e__+, not the number of bits.
7724 Array element C type.
7730 Argument expression.
7736 +ctf_array_text(__t__, __n__, __e__, __s__)+
7738 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7740 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7742 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7744 Statically-sized array, printed as text.
7746 The string doesn't need to be null-terminated.
7749 Array element C type (always `char`).
7755 Argument expression.
7761 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7763 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7765 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7767 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7769 Dynamically-sized array of integers.
7771 The type of +__E__+ must be unsigned.
7774 Array element C type.
7780 Argument expression.
7783 Length expression C type.
7789 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7791 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7793 Dynamically-sized array of integers, displayed in base{nbsp}16.
7795 The type of +__E__+ must be unsigned.
7798 Array element C type.
7804 Argument expression.
7807 Length expression C type.
7812 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7814 Dynamically-sized array of integers in network byte order (big-endian),
7815 displayed in base{nbsp}10.
7817 The type of +__E__+ must be unsigned.
7820 Array element C type.
7826 Argument expression.
7829 Length expression C type.
7835 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7837 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7839 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7841 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7843 Dynamically-sized array of bits.
7845 The type of +__e__+ must be an integer type. +__s__+ is the number
7846 of elements of such type in +__e__+, not the number of bits.
7848 The type of +__E__+ must be unsigned.
7851 Array element C type.
7857 Argument expression.
7860 Length expression C type.
7866 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7868 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7870 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7872 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7874 Dynamically-sized array, displayed as text.
7876 The string doesn't need to be null-terminated.
7878 The type of +__E__+ must be unsigned.
7880 The behaviour is undefined if +__e__+ is `NULL`.
7883 Sequence element C type (always `char`).
7889 Argument expression.
7892 Length expression C type.
7898 Use the `_user` versions when the argument expression, `e`, is
7899 a user space address. In the cases of `ctf_user_integer*()` and
7900 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7903 The `_nowrite` versions omit themselves from the session trace, but are
7904 otherwise identical. This means the `_nowrite` fields won't be written
7905 in the recorded trace. Their primary purpose is to make some
7906 of the event context available to the
7907 <<enabling-disabling-events,event filters>> without having to
7908 commit the data to sub-buffers.
7914 Terms related to LTTng and to tracing in general:
7917 The http://diamon.org/babeltrace[Babeltrace] project, which includes:
7920 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]
7921 command-line interface.
7922 * The libbabeltrace2 library which offers a
7923 https://babeltrace.org/docs/v2.0/libbabeltrace2/[C API].
7924 * https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings].
7927 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7928 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7930 [[def-channel]]<<channel,channel>>::
7931 An entity which is responsible for a set of
7932 <<def-ring-buffer,ring buffers>>.
7934 <<def-event-rule,Event rules>> are always attached to a specific
7938 A source of time for a <<def-tracer,tracer>>.
7940 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
7941 A process which is responsible for consuming the full
7942 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7943 send them over the network.
7945 [[def-current-trace-chunk]]current trace chunk::
7946 A <<def-trace-chunk,trace chunk>> which includes the current content
7947 of all the <<def-sub-buffer,sub-buffers>> of the
7948 <<def-tracing-session-rotation,tracing session>> and the stream files
7949 produced since the latest event amongst:
7951 * The creation of the <<def-tracing-session,tracing session>>.
7952 * The last tracing session rotation, if any.
7954 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
7955 The <<def-event-record-loss-mode,event record loss mode>> in which
7956 the <<def-tracer,tracer>> _discards_ new event records when there's no
7957 <<def-sub-buffer,sub-buffer>> space left to store them.
7959 [[def-event]]event::
7960 The consequence of the execution of an
7961 <<def-instrumentation-point,instrumentation point>>, like a
7962 <<def-tracepoint,tracepoint>> that you manually place in some source
7963 code, or a Linux kernel kprobe.
7965 An event is said to _occur_ at a specific time. <<def-lttng,LTTng>> can
7966 take various actions upon the occurrence of an event, like record its
7967 payload to a <<def-sub-buffer,sub-buffer>>.
7969 [[def-event-name]]event name::
7970 The name of an <<def-event,event>>, which is also the name of the
7971 <<def-event-record,event record>>.
7973 This is also called the _instrumentation point name_.
7975 [[def-event-record]]event record::
7976 A record, in a <<def-trace,trace>>, of the payload of an
7977 <<def-event,event>> which occured.
7979 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7980 The mechanism by which event records of a given
7981 <<def-channel,channel>> are lost (not recorded) when there is no
7982 <<def-sub-buffer,sub-buffer>> space left to store them.
7984 [[def-event-rule]]<<event,event rule>>::
7985 Set of conditions which must be satisfied for one or more occuring
7986 <<def-event,events>> to be recorded.
7988 [[def-incl-set]]inclusion set::
7989 In the <<pid-tracking,process attribute tracking>> context: a
7990 set of <<def-proc-attr,process attributes>> of a given type.
7992 <<instrumenting,instrumentation>>::
7993 The use of <<def-lttng,LTTng>> probes to make a piece of software
7996 [[def-instrumentation-point]]instrumentation point::
7997 A point in the execution path of a piece of software that, when
7998 reached by this execution, can emit an <<def-event,event>>.
8000 instrumentation point name::
8001 See _<<def-event-name,event name>>_.
8003 `java.util.logging`::
8005 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities]
8006 of the Java platform.
8009 A http://logging.apache.org/log4j/1.2/[logging library] for Java
8010 developed by the Apache Software Foundation.
8013 Level of severity of a log statement or user space
8014 <<def-instrumentation-point,instrumentation point>>.
8016 [[def-lttng]]LTTng::
8017 The _Linux Trace Toolkit: next generation_ project.
8019 <<lttng-cli,cmd:lttng>>::
8020 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
8021 project which you can use to send and receive control messages to and
8022 from a <<def-session-daemon,session daemon>>.
8025 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
8026 which is a set of analyzing programs that you can use to obtain a
8027 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
8029 cmd:lttng-consumerd::
8030 The name of the <<def-consumer-daemon,consumer daemon>> program.
8033 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
8034 which can convert <<def-ring-buffer,ring buffer>> files (usually
8035 <<persistent-memory-file-systems,saved on a persistent memory file
8036 system>>) to <<def-trace,trace>> files.
8038 See man:lttng-crash(1).
8040 LTTng Documentation::
8043 <<lttng-live,LTTng live>>::
8044 A communication protocol between the <<lttng-relayd,relay daemon>> and
8045 live viewers which makes it possible to see <<def-event-record,event
8046 records>> ``live'', as they are received by the
8047 <<def-relay-daemon,relay daemon>>.
8049 <<lttng-modules,LTTng-modules>>::
8050 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
8051 which contains the Linux kernel modules to make the Linux kernel
8052 <<def-instrumentation-point,instrumentation points>> available for
8053 <<def-lttng,LTTng>> tracing.
8056 The name of the <<def-relay-daemon,relay daemon>> program.
8058 cmd:lttng-sessiond::
8059 The name of the <<def-session-daemon,session daemon>> program.
8061 [[def-lttng-tools]]LTTng-tools::
8062 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
8063 contains the various programs and libraries used to
8064 <<controlling-tracing,control tracing>>.
8066 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
8067 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
8068 contains libraries to instrument
8069 <<def-user-application,user applications>>.
8071 <<lttng-ust-agents,LTTng-UST Java agent>>::
8072 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
8073 allow the LTTng instrumentation of `java.util.logging` and Apache
8074 log4j{nbsp}1.2 logging statements.
8076 <<lttng-ust-agents,LTTng-UST Python agent>>::
8077 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
8078 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
8081 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
8082 The <<def-event-record-loss-mode,event record loss mode>> in which new
8083 <<def-event-record,event records>> _overwrite_ older event records
8084 when there's no <<def-sub-buffer,sub-buffer>> space left to store
8087 <<channel-buffering-schemes,per-process buffering>>::
8088 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
8089 process has its own <<def-sub-buffer,sub-buffers>> for a given user
8090 space <<def-channel,channel>>.
8092 <<channel-buffering-schemes,per-user buffering>>::
8093 A <<def-buffering-scheme,buffering scheme>> in which all the processes
8094 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
8095 given user space <<def-channel,channel>>.
8097 [[def-proc-attr]]process attribute::
8098 In the <<pid-tracking,process attribute tracking>> context:
8101 * A virtual process ID.
8103 * A virtual Unix user ID.
8105 * A virtual Unix group ID.
8107 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
8108 A process which is responsible for receiving the <<def-trace,trace>>
8109 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
8111 [[def-ring-buffer]]ring buffer::
8112 A set of <<def-sub-buffer,sub-buffers>>.
8115 See _<<def-tracing-session-rotation,tracing session rotation>>_.
8117 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
8118 A process which receives control commands from you and orchestrates
8119 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
8121 <<taking-a-snapshot,snapshot>>::
8122 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
8123 of a given <<def-tracing-session,tracing session>>, saved as
8124 <<def-trace,trace>> files.
8126 [[def-sub-buffer]]sub-buffer::
8127 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
8128 which contains <<def-event-record,event records>>.
8131 The time information attached to an <<def-event,event>> when it is
8134 [[def-trace]]trace (_noun_)::
8137 * One http://diamon.org/ctf/[CTF] metadata stream file.
8138 * One or more CTF data stream files which are the concatenations of one
8139 or more flushed <<def-sub-buffer,sub-buffers>>.
8141 [[def-trace-verb]]trace (_verb_)::
8142 The action of recording the <<def-event,events>> emitted by an
8143 application or by a system, or to initiate such recording by
8144 controlling a <<def-tracer,tracer>>.
8146 [[def-trace-chunk]]trace chunk::
8147 A self-contained <<def-trace,trace>> which is part of a
8148 <<def-tracing-session,tracing session>>. Each
8149 <<def-tracing-session-rotation, tracing session rotation>> produces a
8150 <<def-trace-chunk-archive,trace chunk archive>>.
8152 [[def-trace-chunk-archive]]trace chunk archive::
8153 The result of a <<def-tracing-session-rotation, tracing session rotation>>.
8155 <<def-lttng,LTTng>> doesn't manage any trace chunk archive, even if its
8156 containing <<def-tracing-session,tracing session>> is still active: you
8157 are free to read it, modify it, move it, or remove it.
8160 The http://tracecompass.org[Trace Compass] project and application.
8162 [[def-tracepoint]]tracepoint::
8163 An instrumentation point using the tracepoint mechanism of the Linux
8164 kernel or of <<def-lttng-ust,LTTng-UST>>.
8166 tracepoint definition::
8167 The definition of a single <<def-tracepoint,tracepoint>>.
8170 The name of a <<def-tracepoint,tracepoint>>.
8172 [[def-tracepoint-provider]]tracepoint provider::
8173 A set of functions providing <<def-tracepoint,tracepoints>> to an
8174 instrumented <<def-user-application,user application>>.
8176 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
8177 provider package>>: many tracepoint providers can exist within a
8178 tracepoint provider package.
8180 [[def-tracepoint-provider-package]]tracepoint provider package::
8181 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
8182 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
8183 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
8186 [[def-tracer]]tracer::
8187 A software which records emitted <<def-event,events>>.
8189 <<domain,tracing domain>>::
8190 A namespace for <<def-event,event>> sources.
8192 <<tracing-group,tracing group>>::
8193 The Unix group in which a Unix user can be to be allowed to
8194 <<def-trace-verb,trace>> the Linux kernel.
8196 [[def-tracing-session]]<<tracing-session,tracing session>>::
8197 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
8199 [[def-tracing-session-rotation]]<<session-rotation,tracing session rotation>>::
8200 The action of archiving the
8201 <<def-current-trace-chunk,current trace chunk>> of a
8202 <<def-tracing-session,tracing session>>.
8204 tracked <<def-proc-attr,process attribute>>::
8205 A process attribute which is part of an <<def-incl-set,inclusion
8208 untracked process attribute::
8209 A process attribute which isn't part of an <<def-incl-set,inclusion
8212 [[def-user-application]]user application::
8213 An application running in user space, as opposed to a Linux kernel
8214 module, for example.