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 5{nbsp}August{nbsp}2020.
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 and Ubuntu{nbsp}20.04 _Focal Fossa_:
409 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
411 |https://www.debian.org/[Debian]
412 |<<debian,Debian "bullseye" (testing)>>.
414 |https://www.archlinux.org/[Arch Linux]
415 |<<arch-linux,_Community_ repository and AUR>>.
417 |https://alpinelinux.org/[Alpine Linux]
418 |<<alpine-linux,Alpine Linux{nbsp}3.12>>.
420 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
421 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
426 === Ubuntu: noch:{LTTng} Stable {revision} PPA
428 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
429 Stable{nbsp}{revision} PPA] offers the latest stable
430 LTTng{nbsp}{revision} packages for Ubuntu{nbsp}16.04 _Xenial Xerus_,
431 Ubuntu{nbsp}18.04 _Bionic Beaver_, and Ubuntu{nbsp}20.04 _Focal Fossa_.
433 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
436 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
442 # apt-add-repository ppa:lttng/stable-2.12
447 . Install the main LTTng{nbsp}{revision} packages:
452 # apt-get install lttng-tools
453 # apt-get install lttng-modules-dkms
454 # apt-get install liblttng-ust-dev
458 . **If you need to instrument and trace
459 <<java-application,Java applications>>**, install the LTTng-UST
465 # apt-get install liblttng-ust-agent-java
469 . **If you need to instrument and trace
470 <<python-application,Python{nbsp}3 applications>>**, install the
471 LTTng-UST Python agent:
476 # apt-get install python3-lttngust
484 To install LTTng{nbsp}{revision} on Debian "bullseye" (testing):
486 . Install the main LTTng{nbsp}{revision} packages:
491 # apt-get install lttng-modules-dkms
492 # apt-get install liblttng-ust-dev
493 # apt-get install lttng-tools
497 . **If you need to instrument and trace <<java-application,Java
498 applications>>**, install the LTTng-UST Java agent:
503 # apt-get install liblttng-ust-agent-java
507 . **If you need to instrument and trace <<python-application,Python
508 applications>>**, install the LTTng-UST Python agent:
513 # apt-get install python3-lttngust
521 LTTng-UST{nbsp}{revision} is available in the _community_
522 repository of Arch Linux, while LTTng-tools{nbsp}{revision} and
523 LTTng-modules{nbsp}{revision} are available in the
524 https://aur.archlinux.org/[AUR].
526 To install LTTng{nbsp}{revision} on Arch Linux, using
527 https://github.com/Jguer/yay[yay] for the AUR packages:
529 . Install the main LTTng{nbsp}{revision} packages:
534 # pacman -Sy lttng-ust
535 $ yay -Sy lttng-tools
536 $ yay -Sy lttng-modules
540 . **If you need to instrument and trace <<python-application,Python
541 applications>>**, install the LTTng-UST Python agent:
546 # pacman -Sy python-lttngust
547 # pacman -Sy python2-lttngust
555 To install LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision} on
556 Alpine Linux{nbsp}3.12:
558 . Add the LTTng packages:
563 # apk add lttng-tools
564 # apk add lttng-ust-dev
568 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
574 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
575 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
576 cd lttng-modules-2.12.* &&
578 sudo make modules_install &&
584 [[building-from-source]]
585 === Build from source
587 To build and install LTTng{nbsp}{revision} from source:
589 . Using the package manager of your distribution, or from source,
590 install the following dependencies of LTTng-tools and LTTng-UST:
593 * https://sourceforge.net/projects/libuuid/[libuuid]
594 * http://directory.fsf.org/wiki/Popt[popt]
595 * http://liburcu.org/[Userspace RCU]
596 * http://www.xmlsoft.org/[libxml2]
597 * **Optional**: https://github.com/numactl/numactl[numactl]
600 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
606 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
607 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
608 cd lttng-modules-2.12.* &&
610 sudo make modules_install &&
615 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
621 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.12.tar.bz2 &&
622 tar -xf lttng-ust-latest-2.12.tar.bz2 &&
623 cd lttng-ust-2.12.* &&
631 Add `--disable-numa` to `./configure` if you don't have
632 https://github.com/numactl/numactl[numactl].
636 .Java and Python application tracing
638 If you need to instrument and trace <<java-application,Java
639 applications>>, pass the `--enable-java-agent-jul`,
640 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
641 `configure` script, depending on which Java logging framework you use.
643 If you need to instrument and trace <<python-application,Python
644 applications>>, pass the `--enable-python-agent` option to the
645 `configure` script. You can set the `PYTHON` environment variable to the
646 path to the Python interpreter for which to install the LTTng-UST Python
654 By default, LTTng-UST libraries are installed to
655 dir:{/usr/local/lib}, which is the de facto directory in which to
656 keep self-compiled and third-party libraries.
658 When <<building-tracepoint-providers-and-user-application,linking an
659 instrumented user application with `liblttng-ust`>>:
661 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
663 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
664 man:gcc(1), man:g++(1), or man:clang(1).
668 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
674 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
675 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
676 cd lttng-tools-2.12.* &&
684 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
685 previous steps automatically for a given version of LTTng and confine
686 the installed files in a specific directory. This can be useful to test
687 LTTng without installing it on your system.
693 This is a short guide to get started quickly with LTTng kernel and user
696 Before you follow this guide, make sure to <<installing-lttng,install>>
699 This tutorial walks you through the steps to:
701 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
702 . <<tracing-your-own-user-application,Trace a user application>> written
704 . <<viewing-and-analyzing-your-traces,View and analyze the
708 [[tracing-the-linux-kernel]]
709 === Trace the Linux kernel
711 The following command lines start with the `#` prompt because you need
712 root privileges to trace the Linux kernel. You can also trace the kernel
713 as a regular user if your Unix user is a member of the
714 <<tracing-group,tracing group>>.
716 . Create a <<tracing-session,tracing session>> which writes its traces
717 to dir:{/tmp/my-kernel-trace}:
722 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
726 . List the available kernel tracepoints and system calls:
731 # lttng list --kernel
732 # lttng list --kernel --syscall
736 . Create <<event,event rules>> which match the desired instrumentation
737 point names, for example the `sched_switch` and `sched_process_fork`
738 tracepoints, and the man:open(2) and man:close(2) system calls:
743 # lttng enable-event --kernel sched_switch,sched_process_fork
744 # lttng enable-event --kernel --syscall open,close
748 Create an event rule which matches _all_ the Linux kernel
749 tracepoints with the opt:lttng-enable-event(1):--all option
750 (this will generate a lot of data when tracing):
755 # lttng enable-event --kernel --all
759 . <<basic-tracing-session-control,Start tracing>>:
768 . Do some operation on your system for a few seconds. For example,
769 load a website, or list the files of a directory.
770 . <<creating-destroying-tracing-sessions,Destroy>> the current
780 The man:lttng-destroy(1) command doesn't destroy the trace data; it
781 only destroys the state of the tracing session.
783 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
784 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
785 session>>). You need to stop tracing to make LTTng flush the remaining
786 trace data and make the trace readable.
788 . For the sake of this example, make the recorded trace accessible to
794 # chown -R $(whoami) /tmp/my-kernel-trace
798 See <<viewing-and-analyzing-your-traces,View and analyze the
799 recorded events>> to view the recorded events.
802 [[tracing-your-own-user-application]]
803 === Trace a user application
805 This section steps you through a simple example to trace a
806 _Hello world_ program written in C.
808 To create the traceable user application:
810 . Create the tracepoint provider header file, which defines the
811 tracepoints and the events they can generate:
817 #undef TRACEPOINT_PROVIDER
818 #define TRACEPOINT_PROVIDER hello_world
820 #undef TRACEPOINT_INCLUDE
821 #define TRACEPOINT_INCLUDE "./hello-tp.h"
823 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
826 #include <lttng/tracepoint.h>
836 ctf_string(my_string_field, my_string_arg)
837 ctf_integer(int, my_integer_field, my_integer_arg)
841 #endif /* _HELLO_TP_H */
843 #include <lttng/tracepoint-event.h>
847 . Create the tracepoint provider package source file:
853 #define TRACEPOINT_CREATE_PROBES
854 #define TRACEPOINT_DEFINE
856 #include "hello-tp.h"
860 . Build the tracepoint provider package:
865 $ gcc -c -I. hello-tp.c
869 . Create the _Hello World_ application source file:
876 #include "hello-tp.h"
878 int main(int argc, char *argv[])
882 puts("Hello, World!\nPress Enter to continue...");
885 * The following getchar() call is only placed here for the purpose
886 * of this demonstration, to pause the application in order for
887 * you to have time to list its tracepoints. It's not needed
893 * A tracepoint() call.
895 * Arguments, as defined in hello-tp.h:
897 * 1. Tracepoint provider name (required)
898 * 2. Tracepoint name (required)
899 * 3. my_integer_arg (first user-defined argument)
900 * 4. my_string_arg (second user-defined argument)
902 * Notice the tracepoint provider and tracepoint names are
903 * NOT strings: they are in fact parts of variables that the
904 * macros in hello-tp.h create.
906 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
908 for (x = 0; x < argc; ++x) {
909 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
912 puts("Quitting now!");
913 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
920 . Build the application:
929 . Link the application with the tracepoint provider package,
930 `liblttng-ust`, and `libdl`:
935 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
939 Here's the whole build process:
942 .Build steps of the user space tracing tutorial.
943 image::ust-flow.png[]
945 To trace the user application:
947 . Run the application with a few arguments:
952 $ ./hello world and beyond
961 Press Enter to continue...
965 . Start an LTTng <<lttng-sessiond,session daemon>>:
970 $ lttng-sessiond --daemonize
974 Note that a session daemon might already be running, for example as a
975 service that the service manager of the distribution started.
977 . List the available user space tracepoints:
982 $ lttng list --userspace
986 You see the `hello_world:my_first_tracepoint` tracepoint listed
987 under the `./hello` process.
989 . Create a <<tracing-session,tracing session>>:
994 $ lttng create my-user-space-session
998 . Create an <<event,event rule>> which matches the
999 `hello_world:my_first_tracepoint` event name:
1004 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1008 . <<basic-tracing-session-control,Start tracing>>:
1017 . Go back to the running `hello` application and press Enter. The
1018 program executes all `tracepoint()` instrumentation points and exits.
1019 . <<creating-destroying-tracing-sessions,Destroy>> the current
1029 The man:lttng-destroy(1) command doesn't destroy the trace data; it
1030 only destroys the state of the tracing session.
1032 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1033 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
1034 session>>). You need to stop tracing to make LTTng flush the remaining
1035 trace data and make the trace readable.
1037 By default, LTTng saves the traces in
1038 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1039 where +__name__+ is the tracing session name. The
1040 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1042 See <<viewing-and-analyzing-your-traces,View and analyze the
1043 recorded events>> to view the recorded events.
1046 [[viewing-and-analyzing-your-traces]]
1047 === View and analyze the recorded events
1049 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1050 kernel>> and <<tracing-your-own-user-application,Trace a user
1051 application>> tutorials, you can inspect the recorded events.
1053 There are many tools you can use to read LTTng traces:
1055 https://babeltrace.org/[Babeltrace{nbsp}2]::
1056 A rich, flexible trace manipulation toolkit which includes
1057 a versatile command-line interface
1058 (https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]),
1059 a https://babeltrace.org/docs/v2.0/libbabeltrace2/[C library],
1060 and https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1061 so that you can easily process or convert an LTTng trace with
1064 The Babeltrace{nbsp}2 project ships with a
1065 https://babeltrace.org/docs/v2.0/man7/babeltrace2-plugin-ctf.7/[plugin]
1066 which supports the format of the traces which LTTng produces,
1067 https://diamon.org/ctf/[CTF].
1069 http://tracecompass.org/[Trace Compass]::
1070 A graphical user interface for viewing and analyzing any type of
1071 logs or traces, including those of LTTng.
1073 https://github.com/lttng/lttng-analyses[**LTTng analyses**]::
1074 An experimental project which includes many high-level analyses of
1075 LTTng kernel traces, like scheduling statistics, interrupt
1076 frequency distribution, top CPU usage, and more.
1078 NOTE: This section assumes that LTTng saved the traces it recorded
1079 during the previous tutorials to their default location, in the
1080 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1081 environment variable defaults to `$HOME` if not set.
1084 [[viewing-and-analyzing-your-traces-bt]]
1085 ==== Use the cmd:babeltrace2 command-line tool
1087 The simplest way to list all the recorded events of an LTTng trace is to
1089 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]
1094 $ babeltrace2 ~/lttng-traces/my-user-space-session*
1097 cmd:babeltrace2 finds all traces recursively within the given path and
1098 prints all their events, sorting them chronologically.
1100 Pipe the output of cmd:babeltrace2 into a tool like man:grep(1) for
1105 $ babeltrace2 /tmp/my-kernel-trace | grep _switch
1108 Pipe the output of cmd:babeltrace2 into a tool like man:wc(1) to count
1109 the recorded events:
1113 $ babeltrace2 /tmp/my-kernel-trace | grep _open | wc --lines
1117 [[viewing-and-analyzing-your-traces-bt-python]]
1118 ==== Use the Babeltrace{nbsp}2 Python bindings
1120 The <<viewing-and-analyzing-your-traces-bt,text output of
1121 cmd:babeltrace2>> is useful to isolate events by simple matching using
1122 man:grep(1) and similar utilities. However, more elaborate filters, such
1123 as keeping only event records with a field value falling within a
1124 specific range, are not trivial to write using a shell. Moreover,
1125 reductions and even the most basic computations involving multiple event
1126 records are virtually impossible to implement.
1128 Fortunately, Babeltrace{nbsp}2 ships with
1129 https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings]
1130 which make it easy to read the event records of an LTTng trace
1131 sequentially and compute the desired information.
1133 The following script accepts an LTTng Linux kernel trace path as its
1134 first argument and prints the short names of the top five running
1135 processes on CPU{nbsp}0 during the whole trace:
1145 # Get the trace path from the first command-line argument.
1146 it = bt2.TraceCollectionMessageIterator(sys.argv[1])
1148 # This counter dictionary will hold execution times:
1150 # Task command name -> Total execution time (ns)
1151 exec_times = collections.Counter()
1153 # This holds the last `sched_switch` timestamp.
1157 # We only care about event messages.
1158 if type(msg) is not bt2._EventMessageConst:
1161 # Event of the event message.
1164 # Keep only `sched_switch` events.
1165 if event.cls.name != 'sched_switch':
1168 # Keep only events which occurred on CPU 0.
1169 if event.packet.context_field['cpu_id'] != 0:
1172 # Event timestamp (ns).
1173 cur_ts = msg.default_clock_snapshot.ns_from_origin
1179 # (Short) name of the previous task command.
1180 prev_comm = str(event.payload_field['prev_comm'])
1182 # Initialize an entry in our dictionary if not yet done.
1183 if prev_comm not in exec_times:
1184 exec_times[prev_comm] = 0
1186 # Compute previous command execution time.
1187 diff = cur_ts - last_ts
1189 # Update execution time of this command.
1190 exec_times[prev_comm] += diff
1192 # Update last timestamp.
1196 for name, ns in exec_times.most_common(5):
1197 print('{:20}{} s'.format(name, ns / 1e9))
1200 if __name__ == '__main__':
1208 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1214 swapper/0 48.607245889 s
1215 chromium 7.192738188 s
1216 pavucontrol 0.709894415 s
1217 Compositor 0.660867933 s
1218 Xorg.bin 0.616753786 s
1221 Note that `swapper/0` is the ``idle'' process of CPU{nbsp}0 on Linux;
1222 since we weren't using the CPU that much when tracing, its first
1223 position in the list makes sense.
1227 == [[understanding-lttng]]Core concepts
1229 From a user's perspective, the LTTng system is built on a few concepts,
1230 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1231 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1232 Understanding how those objects relate to eachother is key in mastering
1235 The core concepts are:
1237 * <<tracing-session,Tracing session>>
1238 * <<domain,Tracing domain>>
1239 * <<channel,Channel and ring buffer>>
1240 * <<"event","Instrumentation point, event rule, event, and event record">>
1246 A _tracing session_ is a stateful dialogue between you and
1247 a <<lttng-sessiond,session daemon>>. You can
1248 <<creating-destroying-tracing-sessions,create a new tracing
1249 session>> with the `lttng create` command.
1251 Most of what you do when you control LTTng tracers happens within a
1252 tracing session. In particular, a tracing session:
1255 * Has its own set of trace files.
1256 * Has its own state of activity (started or stopped).
1257 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1259 * Has its own <<channel,channels>> to which are associated their own
1260 <<event,event rules>>.
1261 * Has its own <<pid-tracking,process attribute tracking>> inclusion
1265 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1266 image::concepts.png[]
1268 Those attributes and objects are completely isolated between different
1271 A tracing session is analogous to a cash machine session:
1272 the operations you do on the banking system through the cash machine do
1273 not alter the data of other users of the same system. In the case of
1274 the cash machine, a session lasts as long as your bank card is inside.
1275 In the case of LTTng, a tracing session lasts from the `lttng create`
1276 command to the `lttng destroy` command.
1279 .Each Unix user has its own set of tracing sessions.
1280 image::many-sessions.png[]
1283 [[tracing-session-mode]]
1284 ==== Tracing session mode
1286 LTTng can send the generated trace data to different locations. The
1287 _tracing session mode_ dictates where to send it. The following modes
1288 are available in LTTng{nbsp}{revision}:
1291 LTTng writes the traces to the file system of the machine it traces
1294 Network streaming mode::
1295 LTTng sends the traces over the network to a
1296 <<lttng-relayd,relay daemon>> running on a remote system.
1299 LTTng doesn't write the traces by default.
1301 Instead, you can request LTTng to <<taking-a-snapshot,take a snapshot>>,
1302 that is, a copy of the current sub-buffers of the tracing session, and
1303 to write it to the file system of the target or to send it over the
1304 network to a <<lttng-relayd,relay daemon>> running on a remote system.
1307 This mode is similar to the network streaming mode, but a live
1308 trace viewer can connect to the distant relay daemon to
1309 <<lttng-live,view event records as LTTng generates them>>.
1315 A _tracing domain_ is a namespace for event sources. A tracing domain
1316 has its own properties and features.
1318 There are currently five available tracing domains:
1322 * `java.util.logging` (JUL)
1326 You must specify a tracing domain when using some commands to avoid
1327 ambiguity. For example, since all the domains support named tracepoints
1328 as event sources (instrumentation points that you manually insert in the
1329 source code), you need to specify a tracing domain when
1330 <<enabling-disabling-events,creating an event rule>> because all the
1331 tracing domains could have tracepoints with the same names.
1333 You can create <<channel,channels>> in the Linux kernel and user space
1334 tracing domains. The other tracing domains have a single default
1339 === Channel and ring buffer
1341 A _channel_ is an object which is responsible for a set of ring buffers.
1342 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1343 tracer emits an event, it can record it to one or more
1344 sub-buffers. The attributes of a channel determine what to do when
1345 there's no space left for a new event record because all sub-buffers
1346 are full, where to send a full sub-buffer, and other behaviours.
1348 A channel is always associated to a <<domain,tracing domain>>. The
1349 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1350 a default channel which you can't configure.
1352 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1353 an event, it records it to the sub-buffers of all
1354 the enabled channels with a satisfied event rule, as long as those
1355 channels are part of active <<tracing-session,tracing sessions>>.
1358 [[channel-buffering-schemes]]
1359 ==== Per-user vs. per-process buffering schemes
1361 A channel has at least one ring buffer _per CPU_. LTTng always
1362 records an event to the ring buffer associated to the CPU on which it
1365 Two _buffering schemes_ are available when you
1366 <<enabling-disabling-channels,create a channel>> in the
1367 user space <<domain,tracing domain>>:
1369 Per-user buffering::
1370 Allocate one set of ring buffers--one per CPU--shared by all the
1371 instrumented processes of each Unix user.
1375 .Per-user buffering scheme.
1376 image::per-user-buffering.png[]
1379 Per-process buffering::
1380 Allocate one set of ring buffers--one per CPU--for each
1381 instrumented process.
1385 .Per-process buffering scheme.
1386 image::per-process-buffering.png[]
1389 The per-process buffering scheme tends to consume more memory than the
1390 per-user option because systems generally have more instrumented
1391 processes than Unix users running instrumented processes. However, the
1392 per-process buffering scheme ensures that one process having a high
1393 event throughput won't fill all the shared sub-buffers of the same
1396 The Linux kernel tracing domain has only one available buffering scheme
1397 which is to allocate a single set of ring buffers for the whole system.
1398 This scheme is similar to the per-user option, but with a single, global
1399 user ``running'' the kernel.
1402 [[channel-overwrite-mode-vs-discard-mode]]
1403 ==== Overwrite vs. discard event record loss modes
1405 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1406 arc in the following animations) of the ring buffer of a specific
1407 channel. When there's no space left in a sub-buffer, the tracer marks it
1408 as consumable (red) and another, empty sub-buffer starts receiving the
1409 following event records. A <<lttng-consumerd,consumer daemon>>
1410 eventually consumes the marked sub-buffer (returns to white).
1413 [role="docsvg-channel-subbuf-anim"]
1418 In an ideal world, sub-buffers are consumed faster than they are filled,
1419 as it is the case in the previous animation. In the real world,
1420 however, all sub-buffers can be full at some point, leaving no space to
1421 record the following events.
1423 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1424 no empty sub-buffer is available, it is acceptable to lose event records
1425 when the alternative would be to cause substantial delays in the
1426 execution of the instrumented application. LTTng privileges performance
1427 over integrity; it aims at perturbing the target system as little as
1428 possible in order to make tracing of subtle race conditions and rare
1429 interrupt cascades possible.
1431 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1432 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1433 example>> to learn how to use the blocking mode.
1435 When it comes to losing event records because no empty sub-buffer is
1436 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1437 reached, the _event record loss mode_ of the channel determines what to
1438 do. The available event record loss modes are:
1441 Drop the newest event records until the tracer releases a sub-buffer.
1443 This is the only available mode when you specify a
1444 <<opt-blocking-timeout,blocking timeout>>.
1447 Clear the sub-buffer containing the oldest event records and start
1448 writing the newest event records there.
1450 This mode is sometimes called _flight recorder mode_ because it's
1452 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1453 always keep a fixed amount of the latest data.
1455 Which mechanism you should choose depends on your context: prioritize
1456 the newest or the oldest event records in the ring buffer?
1458 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1459 as soon as a there's no space left for a new event record, whereas in
1460 discard mode, the tracer only discards the event record that doesn't
1463 In discard mode, LTTng increments a count of lost event records when an
1464 event record is lost and saves this count to the trace. In overwrite
1465 mode, since LTTng{nbsp}2.8, LTTng increments a count of lost sub-buffers
1466 when a sub-buffer is lost and saves this count to the trace. In this
1467 mode, LTTng doesn't write to the trace the exact number of lost event
1468 records in those lost sub-buffers. Trace analyses can use saved
1469 discarded event record and sub-buffer counts of the trace to decide
1470 whether or not to perform the analyses even if trace data is known to be
1473 There are a few ways to decrease your probability of losing event
1475 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1476 how to fine-tune the sub-buffer count and size of a channel to virtually
1477 stop losing event records, though at the cost of greater memory usage.
1480 [[channel-subbuf-size-vs-subbuf-count]]
1481 ==== Sub-buffer count and size
1483 When you <<enabling-disabling-channels,create a channel>>, you can
1484 set its number of sub-buffers and their size.
1486 Note that there is noticeable CPU overhead introduced when
1487 switching sub-buffers (marking a full one as consumable and switching
1488 to an empty one for the following events to be recorded). Knowing this,
1489 the following list presents a few practical situations along with how
1490 to configure the sub-buffer count and size for them:
1492 * **High event throughput**: In general, prefer bigger sub-buffers to
1493 lower the risk of losing event records.
1495 Having bigger sub-buffers also ensures a lower
1496 <<channel-switch-timer,sub-buffer switching frequency>>.
1498 The number of sub-buffers is only meaningful if you create the channel
1499 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1500 other sub-buffers are left unaltered.
1502 * **Low event throughput**: In general, prefer smaller sub-buffers
1503 since the risk of losing event records is low.
1505 Because events occur less frequently, the sub-buffer switching frequency
1506 should remain low and thus the overhead of the tracer shouldn't be a
1509 * **Low memory system**: If your target system has a low memory
1510 limit, prefer fewer first, then smaller sub-buffers.
1512 Even if the system is limited in memory, you want to keep the
1513 sub-buffers as big as possible to avoid a high sub-buffer switching
1516 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1517 which means event data is very compact. For example, the average
1518 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1519 sub-buffer size of 1{nbsp}MiB is considered big.
1521 The previous situations highlight the major trade-off between a few big
1522 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1523 frequency vs. how much data is lost in overwrite mode. Assuming a
1524 constant event throughput and using the overwrite mode, the two
1525 following configurations have the same ring buffer total size:
1528 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1533 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1534 switching frequency, but if a sub-buffer overwrite happens, half of
1535 the event records so far (4{nbsp}MiB) are definitely lost.
1536 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the
1537 overhead of the tracer as the previous configuration, but if a
1538 sub-buffer overwrite happens, only the eighth of event records so far
1539 are definitely lost.
1541 In discard mode, the sub-buffers count parameter is pointless: use two
1542 sub-buffers and set their size according to the requirements of your
1546 [[channel-switch-timer]]
1547 ==== Switch timer period
1549 The _switch timer period_ is an important configurable attribute of
1550 a channel to ensure periodic sub-buffer flushing.
1552 When the _switch timer_ expires, a sub-buffer switch happens. Set
1553 the switch timer period attribute when you
1554 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1555 consumes and commits trace data to trace files or to a distant relay
1556 daemon periodically in case of a low event throughput.
1559 [role="docsvg-channel-switch-timer"]
1564 This attribute is also convenient when you use big sub-buffers to cope
1565 with a sporadic high event throughput, even if the throughput is
1569 [[channel-read-timer]]
1570 ==== Read timer period
1572 By default, the LTTng tracers use a notification mechanism to signal a
1573 full sub-buffer so that a consumer daemon can consume it. When such
1574 notifications must be avoided, for example in real-time applications,
1575 use the _read timer_ of the channel instead. When the read timer fires,
1576 the <<lttng-consumerd,consumer daemon>> checks for full, consumable
1580 [[tracefile-rotation]]
1581 ==== Trace file count and size
1583 By default, trace files can grow as large as needed. Set the maximum
1584 size of each trace file that a channel writes when you
1585 <<enabling-disabling-channels,create a channel>>. When the size of a
1586 trace file reaches the fixed maximum size of the channel, LTTng creates
1587 another file to contain the next event records. LTTng appends a file
1588 count to each trace file name in this case.
1590 If you set the trace file size attribute when you create a channel, the
1591 maximum number of trace files that LTTng creates is _unlimited_ by
1592 default. To limit them, set a maximum number of trace files. When the
1593 number of trace files reaches the fixed maximum count of the channel,
1594 the oldest trace file is overwritten. This mechanism is called _trace
1599 Even if you don't limit the trace file count, you can't assume that
1600 LTTng doesn't manage any trace file.
1602 In other words, there is no safe way to know if LTTng still holds a
1603 given trace file open with the trace file rotation feature.
1605 The only way to obtain an unmanaged, self-contained LTTng trace before
1606 you <<creating-destroying-tracing-sessions,destroy>> the tracing session
1607 is with the <<session-rotation,tracing session rotation>> feature
1608 (available since LTTng{nbsp}2.11).
1613 === Instrumentation point, event rule, event, and event record
1615 An _event rule_ is a set of conditions which must be **all** satisfied
1616 for LTTng to record an occuring event.
1618 You set the conditions when you <<enabling-disabling-events,create
1621 You always attach an event rule to a <<channel,channel>> when you create
1624 When an event passes the conditions of an event rule, LTTng records it
1625 in one of the sub-buffers of the attached channel.
1627 The available conditions, as of LTTng{nbsp}{revision}, are:
1629 * The event rule _is enabled_.
1630 * The type of the instrumentation point _is{nbsp}T_.
1631 * The name of the instrumentation point (sometimes called _event name_)
1632 _matches{nbsp}N_, but _isn't{nbsp}E_.
1633 * The log level of the instrumentation point _is as severe as{nbsp}L_, or
1634 _is exactly{nbsp}L_.
1635 * The fields of the payload of the event _satisfy_ a filter
1636 expression{nbsp}__F__.
1638 As you can see, all the conditions but the dynamic filter are related to
1639 the status of the event rule or to the instrumentation point, not to the
1640 occurring events. This is why, without a filter, checking if an event
1641 passes an event rule isn't a dynamic task: when you create or modify an
1642 event rule, all the tracers of its tracing domain enable or disable the
1643 instrumentation points themselves once. This is possible because the
1644 attributes of an instrumentation point (type, name, and log level) are
1645 defined statically. In other words, without a dynamic filter, the tracer
1646 _doesn't evaluate_ the arguments of an instrumentation point unless it
1647 matches an enabled event rule.
1649 Note that, for LTTng to record an event, the <<channel,channel>> to
1650 which a matching event rule is attached must also be enabled, and the
1651 <<tracing-session,tracing session>> owning this channel must be active
1655 .Logical path from an instrumentation point to an event record.
1656 image::event-rule.png[]
1658 .Event, event record, or event rule?
1660 With so many similar terms, it's easy to get confused.
1662 An **event** is the consequence of the execution of an _instrumentation
1663 point_, like a tracepoint that you manually place in some source code,
1664 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1665 time. Different actions can be taken upon the occurrence of an event,
1666 like record the payload of the event to a buffer.
1668 An **event record** is the representation of an event in a sub-buffer. A
1669 tracer is responsible for capturing the payload of an event, current
1670 context variables, the ID of the event, and its timestamp. LTTng
1671 can append this sub-buffer to a trace file.
1673 An **event rule** is a set of conditions which must _all_ be satisfied
1674 for LTTng to record an occuring event. Events still occur without
1675 satisfying event rules, but LTTng doesn't record them.
1680 == Components of noch:{LTTng}
1682 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1683 to call LTTng a simple _tool_ since it is composed of multiple
1684 interacting components. This section describes those components,
1685 explains their respective roles, and shows how they connect together to
1686 form the LTTng ecosystem.
1688 The following diagram shows how the most important components of LTTng
1689 interact with user applications, the Linux kernel, and you:
1692 .Control and trace data paths between LTTng components.
1693 image::plumbing.png[]
1695 The LTTng project incorporates:
1697 * **LTTng-tools**: Libraries and command-line interface to
1698 control tracing sessions.
1699 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1700 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1701 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1702 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1703 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1704 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1706 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1707 headers to instrument and trace any native user application.
1708 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1709 *** `liblttng-ust-libc-wrapper`
1710 *** `liblttng-ust-pthread-wrapper`
1711 *** `liblttng-ust-cyg-profile`
1712 *** `liblttng-ust-cyg-profile-fast`
1713 *** `liblttng-ust-dl`
1714 ** User space tracepoint provider source files generator command-line
1715 tool (man:lttng-gen-tp(1)).
1716 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1717 Java applications using `java.util.logging` or
1718 Apache log4j{nbsp}1.2 logging.
1719 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1720 Python applications using the standard `logging` package.
1721 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1723 ** LTTng kernel tracer module.
1724 ** Tracing ring buffer kernel modules.
1725 ** Probe kernel modules.
1726 ** LTTng logger kernel module.
1730 === Tracing control command-line interface
1733 .The tracing control command-line interface.
1734 image::plumbing-lttng-cli.png[]
1736 The _man:lttng(1) command-line tool_ is the standard user interface to
1737 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1738 is part of LTTng-tools.
1740 The cmd:lttng tool is linked with
1741 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1742 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1744 The cmd:lttng tool has a Git-like interface:
1748 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1751 The <<controlling-tracing,Tracing control>> section explores the
1752 available features of LTTng using the cmd:lttng tool.
1755 [[liblttng-ctl-lttng]]
1756 === Tracing control library
1759 .The tracing control library.
1760 image::plumbing-liblttng-ctl.png[]
1762 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1763 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1764 underlying details of the protocol. `liblttng-ctl` is part of LTTng-tools.
1766 The <<lttng-cli,cmd:lttng command-line tool>>
1767 is linked with `liblttng-ctl`.
1769 Use `liblttng-ctl` in C or $$C++$$ source code by including its
1774 #include <lttng/lttng.h>
1777 Some objects are referenced by name (C string), such as tracing
1778 sessions, but most of them require to create a handle first using
1779 `lttng_create_handle()`.
1781 As of LTTng{nbsp}{revision}, the best available developer documentation for
1782 `liblttng-ctl` is its installed header files. Every function and structure is
1783 thoroughly documented.
1787 === User space tracing library
1790 .The user space tracing library.
1791 image::plumbing-liblttng-ust.png[]
1793 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1794 is the LTTng user space tracer. It receives commands from a
1795 <<lttng-sessiond,session daemon>>, for example to
1796 enable and disable specific instrumentation points, and writes event
1797 records to ring buffers shared with a
1798 <<lttng-consumerd,consumer daemon>>.
1799 `liblttng-ust` is part of LTTng-UST.
1801 Public C header files are installed beside `liblttng-ust` to
1802 instrument any <<c-application,C or $$C++$$ application>>.
1804 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1805 packages, use their own library providing tracepoints which is
1806 linked with `liblttng-ust`.
1808 An application or library doesn't have to initialize `liblttng-ust`
1809 manually: its constructor does the necessary tasks to properly register
1810 to a session daemon. The initialization phase also enables the
1811 instrumentation points matching the <<event,event rules>> that you
1815 [[lttng-ust-agents]]
1816 === User space tracing agents
1819 .The user space tracing agents.
1820 image::plumbing-lttng-ust-agents.png[]
1822 The _LTTng-UST Java and Python agents_ are regular Java and Python
1823 packages which add LTTng tracing capabilities to the
1824 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1826 In the case of Java, the
1827 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1828 core logging facilities] and
1829 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1830 Note that Apache Log4{nbsp}2 isn't supported.
1832 In the case of Python, the standard
1833 https://docs.python.org/3/library/logging.html[`logging`] package
1834 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1835 LTTng-UST Python agent package.
1837 The applications using the LTTng-UST agents are in the
1838 `java.util.logging` (JUL),
1839 log4j, and Python <<domain,tracing domains>>.
1841 Both agents use the same mechanism to trace the log statements. When an
1842 agent initializes, it creates a log handler that attaches to the root
1843 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1844 When the application executes a log statement, the root logger passes it
1845 to the log handler of the agent. The log handler of the agent calls a
1846 native function in a tracepoint provider package shared library linked
1847 with <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1848 other fields, like its logger name and its log level. This native
1849 function contains a user space instrumentation point, hence tracing the
1852 The log level condition of an
1853 <<event,event rule>> is considered when tracing
1854 a Java or a Python application, and it's compatible with the standard
1855 JUL, log4j, and Python log levels.
1859 === LTTng kernel modules
1862 .The LTTng kernel modules.
1863 image::plumbing-lttng-modules.png[]
1865 The _LTTng kernel modules_ are a set of Linux kernel modules
1866 which implement the kernel tracer of the LTTng project. The LTTng
1867 kernel modules are part of LTTng-modules.
1869 The LTTng kernel modules include:
1871 * A set of _probe_ modules.
1873 Each module attaches to a specific subsystem
1874 of the Linux kernel using its tracepoint instrument points. There are
1875 also modules to attach to the entry and return points of the Linux
1876 system call functions.
1878 * _Ring buffer_ modules.
1880 A ring buffer implementation is provided as kernel modules. The LTTng
1881 kernel tracer writes to the ring buffer; a
1882 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1884 * The _LTTng kernel tracer_ module.
1885 * The _LTTng logger_ module.
1887 The LTTng logger module implements the special path:{/proc/lttng-logger}
1888 (and path:{/dev/lttng-logger} since LTTng{nbsp}2.11) files so that any
1889 executable can generate LTTng events by opening and writing to those
1892 See <<proc-lttng-logger-abi,LTTng logger>>.
1894 Generally, you don't have to load the LTTng kernel modules manually
1895 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1896 daemon>> loads the necessary modules when starting. If you have extra
1897 probe modules, you can specify to load them to the session daemon on
1900 The LTTng kernel modules are installed in
1901 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1902 the kernel release (see `uname --kernel-release`).
1909 .The session daemon.
1910 image::plumbing-sessiond.png[]
1912 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1913 managing tracing sessions and for controlling the various components of
1914 LTTng. The session daemon is part of LTTng-tools.
1916 The session daemon sends control requests to and receives control
1919 * The <<lttng-ust,user space tracing library>>.
1921 Any instance of the user space tracing library first registers to
1922 a session daemon. Then, the session daemon can send requests to
1923 this instance, such as:
1926 ** Get the list of tracepoints.
1927 ** Share an <<event,event rule>> so that the user space tracing library
1928 can enable or disable tracepoints. Amongst the possible conditions
1929 of an event rule is a filter expression which `liblttng-ust` evalutes
1930 when an event occurs.
1931 ** Share <<channel,channel>> attributes and ring buffer locations.
1934 The session daemon and the user space tracing library use a Unix
1935 domain socket for their communication.
1937 * The <<lttng-ust-agents,user space tracing agents>>.
1939 Any instance of a user space tracing agent first registers to
1940 a session daemon. Then, the session daemon can send requests to
1941 this instance, such as:
1944 ** Get the list of loggers.
1945 ** Enable or disable a specific logger.
1948 The session daemon and the user space tracing agent use a TCP connection
1949 for their communication.
1951 * The <<lttng-modules,LTTng kernel tracer>>.
1952 * The <<lttng-consumerd,consumer daemon>>.
1954 The session daemon sends requests to the consumer daemon to instruct
1955 it where to send the trace data streams, amongst other information.
1957 * The <<lttng-relayd,relay daemon>>.
1959 The session daemon receives commands from the
1960 <<liblttng-ctl-lttng,tracing control library>>.
1962 The root session daemon loads the appropriate
1963 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1964 a <<lttng-consumerd,consumer daemon>> as soon as you create
1965 an <<event,event rule>>.
1967 The session daemon doesn't send and receive trace data: this is the
1968 role of the <<lttng-consumerd,consumer daemon>> and
1969 <<lttng-relayd,relay daemon>>. It does, however, generate the
1970 http://diamon.org/ctf/[CTF] metadata stream.
1972 Each Unix user can have its own session daemon instance. The
1973 tracing sessions which different session daemons manage are completely
1976 The root user's session daemon is the only one which is
1977 allowed to control the LTTng kernel tracer, and its spawned consumer
1978 daemon is the only one which is allowed to consume trace data from the
1979 LTTng kernel tracer. Note, however, that any Unix user which is a member
1980 of the <<tracing-group,tracing group>> is allowed
1981 to create <<channel,channels>> in the
1982 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
1985 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
1986 session daemon when using its `create` command if none is currently
1987 running. You can also start the session daemon manually.
1994 .The consumer daemon.
1995 image::plumbing-consumerd.png[]
1997 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
1998 ring buffers with user applications or with the LTTng kernel modules to
1999 collect trace data and send it to some location (on disk or to a
2000 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2001 is part of LTTng-tools.
2003 You don't start a consumer daemon manually: a consumer daemon is always
2004 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2005 <<event,event rule>>, that is, before you start tracing. When you kill
2006 its owner session daemon, the consumer daemon also exits because it is
2007 the child process of the session daemon. Command-line options of
2008 man:lttng-sessiond(8) target the consumer daemon process.
2010 There are up to two running consumer daemons per Unix user, whereas only
2011 one session daemon can run per user. This is because each process can be
2012 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2013 and 64-bit processes, it is more efficient to have separate
2014 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2015 exception: it can have up to _three_ running consumer daemons: 32-bit
2016 and 64-bit instances for its user applications, and one more
2017 reserved for collecting kernel trace data.
2025 image::plumbing-relayd.png[]
2027 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2028 between remote session and consumer daemons, local trace files, and a
2029 remote live trace viewer. The relay daemon is part of LTTng-tools.
2031 The main purpose of the relay daemon is to implement a receiver of
2032 <<sending-trace-data-over-the-network,trace data over the network>>.
2033 This is useful when the target system doesn't have much file system
2034 space to record trace files locally.
2036 The relay daemon is also a server to which a
2037 <<lttng-live,live trace viewer>> can
2038 connect. The live trace viewer sends requests to the relay daemon to
2039 receive trace data as the target system emits events. The
2040 communication protocol is named _LTTng live_; it is used over TCP
2043 Note that you can start the relay daemon on the target system directly.
2044 This is the setup of choice when the use case is to view events as
2045 the target system emits them without the need of a remote system.
2049 == [[using-lttng]]Instrumentation
2051 There are many examples of tracing and monitoring in our everyday life:
2053 * You have access to real-time and historical weather reports and
2054 forecasts thanks to weather stations installed around the country.
2055 * You know your heart is safe thanks to an electrocardiogram.
2056 * You make sure not to drive your car too fast and to have enough fuel
2057 to reach your destination thanks to gauges visible on your dashboard.
2059 All the previous examples have something in common: they rely on
2060 **instruments**. Without the electrodes attached to the surface of your
2061 body skin, cardiac monitoring is futile.
2063 LTTng, as a tracer, is no different from those real life examples. If
2064 you're about to trace a software system or, in other words, record its
2065 history of execution, you better have **instrumentation points** in the
2066 subject you're tracing, that is, the actual software.
2068 Various ways were developed to instrument a piece of software for LTTng
2069 tracing. The most straightforward one is to manually place
2070 instrumentation points, called _tracepoints_, in the source code of the
2071 software. It is also possible to add instrumentation points dynamically
2072 in the Linux kernel <<domain,tracing domain>>.
2074 If you're only interested in tracing the Linux kernel, your
2075 instrumentation needs are probably already covered by the built-in
2076 <<lttng-modules,Linux kernel tracepoints>> of LTTng. You may also wish
2077 to trace a user application which is already instrumented for LTTng
2078 tracing. In such cases, skip this whole section and read the topics of
2079 the <<controlling-tracing,Tracing control>> section.
2081 Many methods are available to instrument a piece of software for LTTng
2084 * <<c-application,User space instrumentation for C and $$C++$$
2086 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2087 * <<java-application,User space Java agent>>.
2088 * <<python-application,User space Python agent>>.
2089 * <<proc-lttng-logger-abi,LTTng logger>>.
2090 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2094 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2096 The procedure to instrument a C or $$C++$$ user application with
2097 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2099 . <<tracepoint-provider,Create the source files of a tracepoint provider
2101 . <<probing-the-application-source-code,Add tracepoints to
2102 the source code of the application>>.
2103 . <<building-tracepoint-providers-and-user-application,Build and link
2104 a tracepoint provider package and the user application>>.
2106 If you need quick, man:printf(3)-like instrumentation, skip
2107 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2110 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2111 instrument a user application with `liblttng-ust`.
2114 [[tracepoint-provider]]
2115 ==== Create the source files of a tracepoint provider package
2117 A _tracepoint provider_ is a set of compiled functions which provide
2118 **tracepoints** to an application, the type of instrumentation point
2119 supported by LTTng-UST. Those functions can emit events with
2120 user-defined fields and serialize those events as event records to one
2121 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2122 macro, which you <<probing-the-application-source-code,insert in the
2123 source code of a user application>>, calls those functions.
2125 A _tracepoint provider package_ is an object file (`.o`) or a shared
2126 library (`.so`) which contains one or more tracepoint providers.
2127 Its source files are:
2129 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2130 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2132 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2133 the LTTng user space tracer, at run time.
2136 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2137 image::ust-app.png[]
2139 NOTE: If you need quick, man:printf(3)-like instrumentation,
2140 skip creating and using a tracepoint provider and use
2141 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2145 ===== Create a tracepoint provider header file template
2147 A _tracepoint provider header file_ contains the tracepoint
2148 definitions of a tracepoint provider.
2150 To create a tracepoint provider header file:
2152 . Start from this template:
2156 .Tracepoint provider header file template (`.h` file extension).
2158 #undef TRACEPOINT_PROVIDER
2159 #define TRACEPOINT_PROVIDER provider_name
2161 #undef TRACEPOINT_INCLUDE
2162 #define TRACEPOINT_INCLUDE "./tp.h"
2164 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2167 #include <lttng/tracepoint.h>
2170 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2171 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2176 #include <lttng/tracepoint-event.h>
2182 * `provider_name` with the name of your tracepoint provider.
2183 * `"tp.h"` with the name of your tracepoint provider header file.
2185 . Below the `#include <lttng/tracepoint.h>` line, put your
2186 <<defining-tracepoints,tracepoint definitions>>.
2188 Your tracepoint provider name must be unique amongst all the possible
2189 tracepoint provider names used on the same target system. We
2190 suggest to include the name of your project or company in the name,
2191 for example, `org_lttng_my_project_tpp`.
2193 TIP: [[lttng-gen-tp]]Use the man:lttng-gen-tp(1) tool to create
2194 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2195 write are the <<defining-tracepoints,tracepoint definitions>>.
2198 [[defining-tracepoints]]
2199 ===== Create a tracepoint definition
2201 A _tracepoint definition_ defines, for a given tracepoint:
2203 * Its **input arguments**. They are the macro parameters that the
2204 `tracepoint()` macro accepts for this particular tracepoint
2205 in the source code of the user application.
2206 * Its **output event fields**. They are the sources of event fields
2207 that form the payload of any event that the execution of the
2208 `tracepoint()` macro emits for this particular tracepoint.
2210 Create a tracepoint definition by using the
2211 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2213 <<tpp-header,tracepoint provider header file template>>.
2215 The syntax of the `TRACEPOINT_EVENT()` macro is:
2218 .`TRACEPOINT_EVENT()` macro syntax.
2221 /* Tracepoint provider name */
2224 /* Tracepoint name */
2227 /* Input arguments */
2232 /* Output event fields */
2241 * `provider_name` with your tracepoint provider name.
2242 * `tracepoint_name` with your tracepoint name.
2243 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2244 * `fields` with the <<tpp-def-output-fields,output event field>>
2247 This tracepoint emits events named `provider_name:tracepoint_name`.
2250 .Event name length limitation
2252 The concatenation of the tracepoint provider name and the
2253 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2254 instrumented application compiles and runs, but LTTng throws multiple
2255 warnings and you could experience serious issues.
2258 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2261 .`TP_ARGS()` macro syntax.
2270 * `type` with the C type of the argument.
2271 * `arg_name` with the argument name.
2273 You can repeat `type` and `arg_name` up to 10{nbsp}times to have more
2276 .`TP_ARGS()` usage with three arguments.
2288 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2289 tracepoint definition with no input arguments.
2291 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2292 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2293 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2294 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2297 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2298 C expression that the tracer evalutes at the `tracepoint()` macro site
2299 in the source code of the application. This expression provides the
2300 source of data of a field. The argument expression can include input
2301 argument names listed in the `TP_ARGS()` macro.
2303 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2304 must be unique within a given tracepoint definition.
2306 Here's a complete tracepoint definition example:
2308 .Tracepoint definition.
2310 The following tracepoint definition defines a tracepoint which takes
2311 three input arguments and has four output event fields.
2315 #include "my-custom-structure.h"
2321 const struct my_custom_structure*, my_custom_structure,
2326 ctf_string(query_field, query)
2327 ctf_float(double, ratio_field, ratio)
2328 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2329 ctf_integer(int, send_size, my_custom_structure->send_size)
2334 Refer to this tracepoint definition with the `tracepoint()` macro in
2335 the source code of your application like this:
2339 tracepoint(my_provider, my_tracepoint,
2340 my_structure, some_ratio, the_query);
2344 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2345 if they satisfy an enabled <<event,event rule>>.
2348 [[using-tracepoint-classes]]
2349 ===== Use a tracepoint class
2351 A _tracepoint class_ is a class of tracepoints which share the same
2352 output event field definitions. A _tracepoint instance_ is one
2353 instance of such a defined tracepoint class, with its own tracepoint
2356 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2357 shorthand which defines both a tracepoint class and a tracepoint
2358 instance at the same time.
2360 When you build a tracepoint provider package, the C or $$C++$$ compiler
2361 creates one serialization function for each **tracepoint class**. A
2362 serialization function is responsible for serializing the event fields
2363 of a tracepoint to a sub-buffer when tracing.
2365 For various performance reasons, when your situation requires multiple
2366 tracepoint definitions with different names, but with the same event
2367 fields, we recommend that you manually create a tracepoint class
2368 and instantiate as many tracepoint instances as needed. One positive
2369 effect of such a design, amongst other advantages, is that all
2370 tracepoint instances of the same tracepoint class reuse the same
2371 serialization function, thus reducing
2372 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2374 .Use a tracepoint class and tracepoint instances.
2376 Consider the following three tracepoint definitions:
2388 ctf_integer(int, userid, userid)
2389 ctf_integer(size_t, len, len)
2401 ctf_integer(int, userid, userid)
2402 ctf_integer(size_t, len, len)
2414 ctf_integer(int, userid, userid)
2415 ctf_integer(size_t, len, len)
2420 In this case, we create three tracepoint classes, with one implicit
2421 tracepoint instance for each of them: `get_account`, `get_settings`, and
2422 `get_transaction`. However, they all share the same event field names
2423 and types. Hence three identical, yet independent serialization
2424 functions are created when you build the tracepoint provider package.
2426 A better design choice is to define a single tracepoint class and three
2427 tracepoint instances:
2431 /* The tracepoint class */
2432 TRACEPOINT_EVENT_CLASS(
2433 /* Tracepoint provider name */
2436 /* Tracepoint class name */
2439 /* Input arguments */
2445 /* Output event fields */
2447 ctf_integer(int, userid, userid)
2448 ctf_integer(size_t, len, len)
2452 /* The tracepoint instances */
2453 TRACEPOINT_EVENT_INSTANCE(
2454 /* Tracepoint provider name */
2457 /* Tracepoint class name */
2460 /* Tracepoint name */
2463 /* Input arguments */
2469 TRACEPOINT_EVENT_INSTANCE(
2478 TRACEPOINT_EVENT_INSTANCE(
2491 [[assigning-log-levels]]
2492 ===== Assign a log level to a tracepoint definition
2494 Assign a _log level_ to a <<defining-tracepoints,tracepoint definition>>
2495 with the `TRACEPOINT_LOGLEVEL()` macro.
2497 Assigning different levels of severity to tracepoint definitions can
2498 be useful: when you <<enabling-disabling-events,create an event rule>>,
2499 you can target tracepoints having a log level as severe as a specific
2502 The concept of LTTng-UST log levels is similar to the levels found
2503 in typical logging frameworks:
2505 * In a logging framework, the log level is given by the function
2506 or method name you use at the log statement site: `debug()`,
2507 `info()`, `warn()`, `error()`, and so on.
2508 * In LTTng-UST, you statically assign the log level to a tracepoint
2509 definition; any `tracepoint()` macro invocation which refers to
2510 this definition has this log level.
2512 You must use `TRACEPOINT_LOGLEVEL()` _after_ the
2513 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2514 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2517 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2520 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2522 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2527 * `provider_name` with the tracepoint provider name.
2528 * `tracepoint_name` with the tracepoint name.
2529 * `log_level` with the log level to assign to the tracepoint
2530 definition named `tracepoint_name` in the `provider_name`
2531 tracepoint provider.
2533 See man:lttng-ust(3) for a list of available log level names.
2535 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2539 /* Tracepoint definition */
2548 ctf_integer(int, userid, userid)
2549 ctf_integer(size_t, len, len)
2553 /* Log level assignment */
2554 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2560 ===== Create a tracepoint provider package source file
2562 A _tracepoint provider package source file_ is a C source file which
2563 includes a <<tpp-header,tracepoint provider header file>> to expand its
2564 macros into event serialization and other functions.
2566 Use the following tracepoint provider package source file template:
2569 .Tracepoint provider package source file template.
2571 #define TRACEPOINT_CREATE_PROBES
2576 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2577 header file>> name. You may also include more than one tracepoint
2578 provider header file here to create a tracepoint provider package
2579 holding more than one tracepoint providers.
2582 [[probing-the-application-source-code]]
2583 ==== Add tracepoints to the source code of an application
2585 Once you <<tpp-header,create a tracepoint provider header file>>, use
2586 the `tracepoint()` macro in the source code of your application to
2587 insert the tracepoints that this header
2588 <<defining-tracepoints,defines>>.
2590 The `tracepoint()` macro takes at least two parameters: the tracepoint
2591 provider name and the tracepoint name. The corresponding tracepoint
2592 definition defines the other parameters.
2594 .`tracepoint()` usage.
2596 The following <<defining-tracepoints,tracepoint definition>> defines a
2597 tracepoint which takes two input arguments and has two output event
2601 .Tracepoint provider header file.
2603 #include "my-custom-structure.h"
2610 const char*, cmd_name
2613 ctf_string(cmd_name, cmd_name)
2614 ctf_integer(int, number_of_args, argc)
2619 Refer to this tracepoint definition with the `tracepoint()` macro in
2620 the source code of your application like this:
2623 .Application source file.
2627 int main(int argc, char* argv[])
2629 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2635 Note how the source code of the application includes
2636 the tracepoint provider header file containing the tracepoint
2637 definitions to use, path:{tp.h}.
2640 .`tracepoint()` usage with a complex tracepoint definition.
2642 Consider this complex tracepoint definition, where multiple event
2643 fields refer to the same input arguments in their argument expression
2647 .Tracepoint provider header file.
2649 /* For `struct stat` */
2650 #include <sys/types.h>
2651 #include <sys/stat.h>
2663 ctf_integer(int, my_constant_field, 23 + 17)
2664 ctf_integer(int, my_int_arg_field, my_int_arg)
2665 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2666 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2667 my_str_arg[2] + my_str_arg[3])
2668 ctf_string(my_str_arg_field, my_str_arg)
2669 ctf_integer_hex(off_t, size_field, st->st_size)
2670 ctf_float(double, size_dbl_field, (double) st->st_size)
2671 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2672 size_t, strlen(my_str_arg) / 2)
2677 Refer to this tracepoint definition with the `tracepoint()` macro in
2678 the source code of your application like this:
2681 .Application source file.
2683 #define TRACEPOINT_DEFINE
2690 stat("/etc/fstab", &s);
2691 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2697 If you look at the event record that LTTng writes when tracing this
2698 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2699 it should look like this:
2701 .Event record fields
2703 |Field name |Field value
2704 |`my_constant_field` |40
2705 |`my_int_arg_field` |23
2706 |`my_int_arg_field2` |529
2708 |`my_str_arg_field` |`Hello, World!`
2709 |`size_field` |0x12d
2710 |`size_dbl_field` |301.0
2711 |`half_my_str_arg_field` |`Hello,`
2715 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2716 compute--they use the call stack, for example. To avoid this computation
2717 when the tracepoint is disabled, use the `tracepoint_enabled()` and
2718 `do_tracepoint()` macros.
2720 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2724 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2726 tracepoint_enabled(provider_name, tracepoint_name)
2727 do_tracepoint(provider_name, tracepoint_name, ...)
2732 * `provider_name` with the tracepoint provider name.
2733 * `tracepoint_name` with the tracepoint name.
2735 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2736 `tracepoint_name` from the provider named `provider_name` is enabled
2739 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2740 if the tracepoint is enabled. Using `tracepoint()` with
2741 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2742 the `tracepoint_enabled()` check, thus a race condition is
2743 possible in this situation:
2746 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2748 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2749 stuff = prepare_stuff();
2752 tracepoint(my_provider, my_tracepoint, stuff);
2755 If the tracepoint is enabled after the condition, then `stuff` isn't
2756 prepared: the emitted event will either contain wrong data, or the whole
2757 application could crash (segmentation fault, for example).
2759 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2760 `STAP_PROBEV()` call. If you need it, you must emit
2764 [[building-tracepoint-providers-and-user-application]]
2765 ==== Build and link a tracepoint provider package and an application
2767 Once you have one or more <<tpp-header,tracepoint provider header
2768 files>> and a <<tpp-source,tracepoint provider package source file>>,
2769 create the tracepoint provider package by compiling its source
2770 file. From here, multiple build and run scenarios are possible. The
2771 following table shows common application and library configurations
2772 along with the required command lines to achieve them.
2774 In the following diagrams, we use the following file names:
2777 Executable application.
2780 Application object file.
2783 Tracepoint provider package object file.
2786 Tracepoint provider package archive file.
2789 Tracepoint provider package shared object file.
2792 User library object file.
2795 User library shared object file.
2797 We use the following symbols in the diagrams of table below:
2800 .Symbols used in the build scenario diagrams.
2801 image::ust-sit-symbols.png[]
2803 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2804 variable in the following instructions.
2806 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2807 .Common tracepoint provider package scenarios.
2809 |Scenario |Instructions
2812 The instrumented application is statically linked with
2813 the tracepoint provider package object.
2815 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2818 include::../common/ust-sit-step-tp-o.txt[]
2820 To build the instrumented application:
2822 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2827 #define TRACEPOINT_DEFINE
2831 . Compile the application source file:
2840 . Build the application:
2845 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2849 To run the instrumented application:
2851 * Start the application:
2861 The instrumented application is statically linked with the
2862 tracepoint provider package archive file.
2864 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2867 To create the tracepoint provider package archive file:
2869 . Compile the <<tpp-source,tracepoint provider package source file>>:
2878 . Create the tracepoint provider package archive file:
2883 $ ar rcs tpp.a tpp.o
2887 To build the instrumented application:
2889 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2894 #define TRACEPOINT_DEFINE
2898 . Compile the application source file:
2907 . Build the application:
2912 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2916 To run the instrumented application:
2918 * Start the application:
2928 The instrumented application is linked with the tracepoint provider
2929 package shared object.
2931 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2934 include::../common/ust-sit-step-tp-so.txt[]
2936 To build the instrumented application:
2938 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2943 #define TRACEPOINT_DEFINE
2947 . Compile the application source file:
2956 . Build the application:
2961 $ gcc -o app app.o -ldl -L. -ltpp
2965 To run the instrumented application:
2967 * Start the application:
2977 The tracepoint provider package shared object is preloaded before the
2978 instrumented application starts.
2980 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
2983 include::../common/ust-sit-step-tp-so.txt[]
2985 To build the instrumented application:
2987 . In path:{app.c}, before including path:{tpp.h}, add the
2993 #define TRACEPOINT_DEFINE
2994 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
2998 . Compile the application source file:
3007 . Build the application:
3012 $ gcc -o app app.o -ldl
3016 To run the instrumented application with tracing support:
3018 * Preload the tracepoint provider package shared object and
3019 start the application:
3024 $ LD_PRELOAD=./libtpp.so ./app
3028 To run the instrumented application without tracing support:
3030 * Start the application:
3040 The instrumented application dynamically loads the tracepoint provider
3041 package shared object.
3043 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3046 include::../common/ust-sit-step-tp-so.txt[]
3048 To build the instrumented application:
3050 . In path:{app.c}, before including path:{tpp.h}, add the
3056 #define TRACEPOINT_DEFINE
3057 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3061 . Compile the application source file:
3070 . Build the application:
3075 $ gcc -o app app.o -ldl
3079 To run the instrumented application:
3081 * Start the application:
3091 The application is linked with the instrumented user library.
3093 The instrumented user library is statically linked with the tracepoint
3094 provider package object file.
3096 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3099 include::../common/ust-sit-step-tp-o-fpic.txt[]
3101 To build the instrumented user library:
3103 . In path:{emon.c}, before including path:{tpp.h}, add the
3109 #define TRACEPOINT_DEFINE
3113 . Compile the user library source file:
3118 $ gcc -I. -fpic -c emon.c
3122 . Build the user library shared object:
3127 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3131 To build the application:
3133 . Compile the application source file:
3142 . Build the application:
3147 $ gcc -o app app.o -L. -lemon
3151 To run the application:
3153 * Start the application:
3163 The application is linked with the instrumented user library.
3165 The instrumented user library is linked with the tracepoint provider
3166 package shared object.
3168 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3171 include::../common/ust-sit-step-tp-so.txt[]
3173 To build the instrumented user library:
3175 . In path:{emon.c}, before including path:{tpp.h}, add the
3181 #define TRACEPOINT_DEFINE
3185 . Compile the user library source file:
3190 $ gcc -I. -fpic -c emon.c
3194 . Build the user library shared object:
3199 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3203 To build the application:
3205 . Compile the application source file:
3214 . Build the application:
3219 $ gcc -o app app.o -L. -lemon
3223 To run the application:
3225 * Start the application:
3235 The tracepoint provider package shared object is preloaded before the
3238 The application is linked with the instrumented user library.
3240 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3243 include::../common/ust-sit-step-tp-so.txt[]
3245 To build the instrumented user library:
3247 . In path:{emon.c}, before including path:{tpp.h}, add the
3253 #define TRACEPOINT_DEFINE
3254 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3258 . Compile the user library source file:
3263 $ gcc -I. -fpic -c emon.c
3267 . Build the user library shared object:
3272 $ gcc -shared -o libemon.so emon.o -ldl
3276 To build the application:
3278 . Compile the application source file:
3287 . Build the application:
3292 $ gcc -o app app.o -L. -lemon
3296 To run the application with tracing support:
3298 * Preload the tracepoint provider package shared object and
3299 start the application:
3304 $ LD_PRELOAD=./libtpp.so ./app
3308 To run the application without tracing support:
3310 * Start the application:
3320 The application is linked with the instrumented user library.
3322 The instrumented user library dynamically loads the tracepoint provider
3323 package shared object.
3325 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3328 include::../common/ust-sit-step-tp-so.txt[]
3330 To build the instrumented user library:
3332 . In path:{emon.c}, before including path:{tpp.h}, add the
3338 #define TRACEPOINT_DEFINE
3339 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3343 . Compile the user library source file:
3348 $ gcc -I. -fpic -c emon.c
3352 . Build the user library shared object:
3357 $ gcc -shared -o libemon.so emon.o -ldl
3361 To build the application:
3363 . Compile the application source file:
3372 . Build the application:
3377 $ gcc -o app app.o -L. -lemon
3381 To run the application:
3383 * Start the application:
3393 The application dynamically loads the instrumented user library.
3395 The instrumented user library is linked with the tracepoint provider
3396 package shared object.
3398 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3401 include::../common/ust-sit-step-tp-so.txt[]
3403 To build the instrumented user library:
3405 . In path:{emon.c}, before including path:{tpp.h}, add the
3411 #define TRACEPOINT_DEFINE
3415 . Compile the user library source file:
3420 $ gcc -I. -fpic -c emon.c
3424 . Build the user library shared object:
3429 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3433 To build the application:
3435 . Compile the application source file:
3444 . Build the application:
3449 $ gcc -o app app.o -ldl -L. -lemon
3453 To run the application:
3455 * Start the application:
3465 The application dynamically loads the instrumented user library.
3467 The instrumented user library dynamically loads the tracepoint provider
3468 package shared object.
3470 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3473 include::../common/ust-sit-step-tp-so.txt[]
3475 To build the instrumented user library:
3477 . In path:{emon.c}, before including path:{tpp.h}, add the
3483 #define TRACEPOINT_DEFINE
3484 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3488 . Compile the user library source file:
3493 $ gcc -I. -fpic -c emon.c
3497 . Build the user library shared object:
3502 $ gcc -shared -o libemon.so emon.o -ldl
3506 To build the application:
3508 . Compile the application source file:
3517 . Build the application:
3522 $ gcc -o app app.o -ldl -L. -lemon
3526 To run the application:
3528 * Start the application:
3538 The tracepoint provider package shared object is preloaded before the
3541 The application dynamically loads the instrumented user library.
3543 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3546 include::../common/ust-sit-step-tp-so.txt[]
3548 To build the instrumented user library:
3550 . In path:{emon.c}, before including path:{tpp.h}, add the
3556 #define TRACEPOINT_DEFINE
3557 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3561 . Compile the user library source file:
3566 $ gcc -I. -fpic -c emon.c
3570 . Build the user library shared object:
3575 $ gcc -shared -o libemon.so emon.o -ldl
3579 To build the application:
3581 . Compile the application source file:
3590 . Build the application:
3595 $ gcc -o app app.o -L. -lemon
3599 To run the application with tracing support:
3601 * Preload the tracepoint provider package shared object and
3602 start the application:
3607 $ LD_PRELOAD=./libtpp.so ./app
3611 To run the application without tracing support:
3613 * Start the application:
3623 The application is statically linked with the tracepoint provider
3624 package object file.
3626 The application is linked with the instrumented user library.
3628 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3631 include::../common/ust-sit-step-tp-o.txt[]
3633 To build the instrumented user library:
3635 . In path:{emon.c}, before including path:{tpp.h}, add the
3641 #define TRACEPOINT_DEFINE
3645 . Compile the user library source file:
3650 $ gcc -I. -fpic -c emon.c
3654 . Build the user library shared object:
3659 $ gcc -shared -o libemon.so emon.o
3663 To build the application:
3665 . Compile the application source file:
3674 . Build the application:
3679 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3683 To run the instrumented application:
3685 * Start the application:
3695 The application is statically linked with the tracepoint provider
3696 package object file.
3698 The application dynamically loads the instrumented user library.
3700 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3703 include::../common/ust-sit-step-tp-o.txt[]
3705 To build the application:
3707 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3712 #define TRACEPOINT_DEFINE
3716 . Compile the application source file:
3725 . Build the application:
3730 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3735 The `--export-dynamic` option passed to the linker is necessary for the
3736 dynamically loaded library to ``see'' the tracepoint symbols defined in
3739 To build the instrumented user library:
3741 . Compile the user library source file:
3746 $ gcc -I. -fpic -c emon.c
3750 . Build the user library shared object:
3755 $ gcc -shared -o libemon.so emon.o
3759 To run the application:
3761 * Start the application:
3772 [[using-lttng-ust-with-daemons]]
3773 ===== Use noch:{LTTng-UST} with daemons
3775 If your instrumented application calls man:fork(2), man:clone(2),
3776 or BSD's man:rfork(2), without a following man:exec(3)-family
3777 system call, you must preload the path:{liblttng-ust-fork.so} shared
3778 object when you start the application.
3782 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3785 If your tracepoint provider package is
3786 a shared library which you also preload, you must put both
3787 shared objects in env:LD_PRELOAD:
3791 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3797 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3799 If your instrumented application closes one or more file descriptors
3800 which it did not open itself, you must preload the
3801 path:{liblttng-ust-fd.so} shared object when you start the application:
3805 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3808 Typical use cases include closing all the file descriptors after
3809 man:fork(2) or man:rfork(2) and buggy applications doing
3813 [[lttng-ust-pkg-config]]
3814 ===== Use noch:{pkg-config}
3816 On some distributions, LTTng-UST ships with a
3817 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3818 metadata file. If this is your case, then use cmd:pkg-config to
3819 build an application on the command line:
3823 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3827 [[instrumenting-32-bit-app-on-64-bit-system]]
3828 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3830 In order to trace a 32-bit application running on a 64-bit system,
3831 LTTng must use a dedicated 32-bit
3832 <<lttng-consumerd,consumer daemon>>.
3834 The following steps show how to build and install a 32-bit consumer
3835 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3836 build and install the 32-bit LTTng-UST libraries, and how to build and
3837 link an instrumented 32-bit application in that context.
3839 To build a 32-bit instrumented application for a 64-bit target system,
3840 assuming you have a fresh target system with no installed Userspace RCU
3843 . Download, build, and install a 32-bit version of Userspace RCU:
3848 $ cd $(mktemp -d) &&
3849 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3850 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3851 cd userspace-rcu-0.9.* &&
3852 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3854 sudo make install &&
3859 . Using the package manager of your distribution, or from source,
3860 install the following 32-bit versions of the following dependencies of
3861 LTTng-tools and LTTng-UST:
3864 * https://sourceforge.net/projects/libuuid/[libuuid]
3865 * http://directory.fsf.org/wiki/Popt[popt]
3866 * http://www.xmlsoft.org/[libxml2]
3869 . Download, build, and install a 32-bit version of the latest
3870 LTTng-UST{nbsp}{revision}:
3875 $ cd $(mktemp -d) &&
3876 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.12.tar.bz2 &&
3877 tar -xf lttng-ust-latest-2.12.tar.bz2 &&
3878 cd lttng-ust-2.12.* &&
3879 ./configure --libdir=/usr/local/lib32 \
3880 CFLAGS=-m32 CXXFLAGS=-m32 \
3881 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3883 sudo make install &&
3890 Depending on your distribution,
3891 32-bit libraries could be installed at a different location than
3892 `/usr/lib32`. For example, Debian is known to install
3893 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3895 In this case, make sure to set `LDFLAGS` to all the
3896 relevant 32-bit library paths, for example:
3900 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3904 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3905 the 32-bit consumer daemon:
3910 $ cd $(mktemp -d) &&
3911 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
3912 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
3913 cd lttng-tools-2.12.* &&
3914 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3915 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3916 --disable-bin-lttng --disable-bin-lttng-crash \
3917 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3919 cd src/bin/lttng-consumerd &&
3920 sudo make install &&
3925 . From your distribution or from source,
3926 <<installing-lttng,install>> the 64-bit versions of
3927 LTTng-UST and Userspace RCU.
3928 . Download, build, and install the 64-bit version of the
3929 latest LTTng-tools{nbsp}{revision}:
3934 $ cd $(mktemp -d) &&
3935 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.12.tar.bz2 &&
3936 tar -xf lttng-tools-latest-2.12.tar.bz2 &&
3937 cd lttng-tools-2.12.* &&
3938 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3939 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3941 sudo make install &&
3946 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3947 when linking your 32-bit application:
3950 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3951 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3954 For example, let's rebuild the quick start example in
3955 <<tracing-your-own-user-application,Trace a user application>> as an
3956 instrumented 32-bit application:
3961 $ gcc -m32 -c -I. hello-tp.c
3962 $ gcc -m32 -c hello.c
3963 $ gcc -m32 -o hello hello.o hello-tp.o \
3964 -L/usr/lib32 -L/usr/local/lib32 \
3965 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3970 No special action is required to execute the 32-bit application and
3971 to trace it: use the command-line man:lttng(1) tool as usual.
3978 man:tracef(3) is a small LTTng-UST API designed for quick,
3979 man:printf(3)-like instrumentation without the burden of
3980 <<tracepoint-provider,creating>> and
3981 <<building-tracepoint-providers-and-user-application,building>>
3982 a tracepoint provider package.
3984 To use `tracef()` in your application:
3986 . In the C or C++ source files where you need to use `tracef()`,
3987 include `<lttng/tracef.h>`:
3992 #include <lttng/tracef.h>
3996 . In the source code of the application, use `tracef()` like you would
4004 tracef("my message: %d (%s)", my_integer, my_string);
4010 . Link your application with `liblttng-ust`:
4015 $ gcc -o app app.c -llttng-ust
4019 To trace the events that `tracef()` calls emit:
4021 * <<enabling-disabling-events,Create an event rule>> which matches the
4022 `lttng_ust_tracef:*` event name:
4027 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4032 .Limitations of `tracef()`
4034 The `tracef()` utility function was developed to make user space tracing
4035 super simple, albeit with notable disadvantages compared to
4036 <<defining-tracepoints,user-defined tracepoints>>:
4038 * All the emitted events have the same tracepoint provider and
4039 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4040 * There is no static type checking.
4041 * The only event record field you actually get, named `msg`, is a string
4042 potentially containing the values you passed to `tracef()`
4043 using your own format string. This also means that you can't filter
4044 events with a custom expression at run time because there are no
4046 * Since `tracef()` uses the man:vasprintf(3) function of the
4047 C{nbsp}standard library behind the scenes to format the strings at run
4048 time, its expected performance is lower than with user-defined
4049 tracepoints, which don't require a conversion to a string.
4051 Taking this into consideration, `tracef()` is useful for some quick
4052 prototyping and debugging, but you shouldn't consider it for any
4053 permanent and serious applicative instrumentation.
4059 ==== Use `tracelog()`
4061 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4062 the difference that it accepts an additional log level parameter.
4064 The goal of `tracelog()` is to ease the migration from logging to
4067 To use `tracelog()` in your application:
4069 . In the C or C++ source files where you need to use `tracelog()`,
4070 include `<lttng/tracelog.h>`:
4075 #include <lttng/tracelog.h>
4079 . In the source code of the application, use `tracelog()` like you would
4080 use man:printf(3), except for the first parameter which is the log
4088 tracelog(TRACE_WARNING, "my message: %d (%s)",
4089 my_integer, my_string);
4095 See man:lttng-ust(3) for a list of available log level names.
4097 . Link your application with `liblttng-ust`:
4102 $ gcc -o app app.c -llttng-ust
4106 To trace the events that `tracelog()` calls emit with a log level
4107 _as severe as_ a specific log level:
4109 * <<enabling-disabling-events,Create an event rule>> which matches the
4110 `lttng_ust_tracelog:*` event name and a minimum level
4116 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4117 --loglevel=TRACE_WARNING
4121 To trace the events that `tracelog()` calls emit with a
4122 _specific log level_:
4124 * Create an event rule which matches the `lttng_ust_tracelog:*`
4125 event name and a specific log level:
4130 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4131 --loglevel-only=TRACE_INFO
4136 [[prebuilt-ust-helpers]]
4137 === Prebuilt user space tracing helpers
4139 The LTTng-UST package provides a few helpers in the form of preloadable
4140 shared objects which automatically instrument system functions and
4143 The helper shared objects are normally found in dir:{/usr/lib}. If you
4144 built LTTng-UST <<building-from-source,from source>>, they are probably
4145 located in dir:{/usr/local/lib}.
4147 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4150 path:{liblttng-ust-libc-wrapper.so}::
4151 path:{liblttng-ust-pthread-wrapper.so}::
4152 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4153 memory and POSIX threads function tracing>>.
4155 path:{liblttng-ust-cyg-profile.so}::
4156 path:{liblttng-ust-cyg-profile-fast.so}::
4157 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4159 path:{liblttng-ust-dl.so}::
4160 <<liblttng-ust-dl,Dynamic linker tracing>>.
4162 To use a user space tracing helper with any user application:
4164 * Preload the helper shared object when you start the application:
4169 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4173 You can preload more than one helper:
4178 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4184 [[liblttng-ust-libc-pthread-wrapper]]
4185 ==== Instrument C standard library memory and POSIX threads functions
4187 The path:{liblttng-ust-libc-wrapper.so} and
4188 path:{liblttng-ust-pthread-wrapper.so} helpers
4189 add instrumentation to some C standard library and POSIX
4193 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4195 |TP provider name |TP name |Instrumented function
4197 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4198 |`calloc` |man:calloc(3)
4199 |`realloc` |man:realloc(3)
4200 |`free` |man:free(3)
4201 |`memalign` |man:memalign(3)
4202 |`posix_memalign` |man:posix_memalign(3)
4206 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4208 |TP provider name |TP name |Instrumented function
4210 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4211 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4212 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4213 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4216 When you preload the shared object, it replaces the functions listed
4217 in the previous tables by wrappers which contain tracepoints and call
4218 the replaced functions.
4221 [[liblttng-ust-cyg-profile]]
4222 ==== Instrument function entry and exit
4224 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4225 to the entry and exit points of functions.
4227 man:gcc(1) and man:clang(1) have an option named
4228 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4229 which generates instrumentation calls for entry and exit to functions.
4230 The LTTng-UST function tracing helpers,
4231 path:{liblttng-ust-cyg-profile.so} and
4232 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4233 to add tracepoints to the two generated functions (which contain
4234 `cyg_profile` in their names, hence the name of the helper).
4236 To use the LTTng-UST function tracing helper, the source files to
4237 instrument must be built using the `-finstrument-functions` compiler
4240 There are two versions of the LTTng-UST function tracing helper:
4242 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4243 that you should only use when it can be _guaranteed_ that the
4244 complete event stream is recorded without any lost event record.
4245 Any kind of duplicate information is left out.
4247 Assuming no event record is lost, having only the function addresses on
4248 entry is enough to create a call graph, since an event record always
4249 contains the ID of the CPU that generated it.
4251 Use a tool like man:addr2line(1) to convert function addresses back to
4252 source file names and line numbers.
4254 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4255 which also works in use cases where event records might get discarded or
4256 not recorded from application startup.
4257 In these cases, the trace analyzer needs more information to be
4258 able to reconstruct the program flow.
4260 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4261 points of this helper.
4263 All the tracepoints that this helper provides have the
4264 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4266 TIP: It's sometimes a good idea to limit the number of source files that
4267 you compile with the `-finstrument-functions` option to prevent LTTng
4268 from writing an excessive amount of trace data at run time. When using
4270 `-finstrument-functions-exclude-function-list` option to avoid
4271 instrument entries and exits of specific function names.
4276 ==== Instrument the dynamic linker
4278 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4279 man:dlopen(3) and man:dlclose(3) function calls.
4281 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4286 [[java-application]]
4287 === User space Java agent
4289 You can instrument any Java application which uses one of the following
4292 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4293 (JUL) core logging facilities.
4294 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4295 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 isn't supported.
4298 .LTTng-UST Java agent imported by a Java application.
4299 image::java-app.png[]
4301 Note that the methods described below are new in LTTng{nbsp}2.8.
4302 Previous LTTng versions use another technique.
4304 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4305 and https://ci.lttng.org/[continuous integration], thus this version is
4306 directly supported. However, the LTTng-UST Java agent is also tested
4307 with OpenJDK{nbsp}7.
4312 ==== Use the LTTng-UST Java agent for `java.util.logging`
4314 To use the LTTng-UST Java agent in a Java application which uses
4315 `java.util.logging` (JUL):
4317 . In the source code of the Java application, import the LTTng-UST log
4318 handler package for `java.util.logging`:
4323 import org.lttng.ust.agent.jul.LttngLogHandler;
4327 . Create an LTTng-UST JUL log handler:
4332 Handler lttngUstLogHandler = new LttngLogHandler();
4336 . Add this handler to the JUL loggers which should emit LTTng events:
4341 Logger myLogger = Logger.getLogger("some-logger");
4343 myLogger.addHandler(lttngUstLogHandler);
4347 . Use `java.util.logging` log statements and configuration as usual.
4348 The loggers with an attached LTTng-UST log handler can emit
4351 . Before exiting the application, remove the LTTng-UST log handler from
4352 the loggers attached to it and call its `close()` method:
4357 myLogger.removeHandler(lttngUstLogHandler);
4358 lttngUstLogHandler.close();
4362 This isn't strictly necessary, but it is recommended for a clean
4363 disposal of the resources of the handler.
4365 . Include the common and JUL-specific JAR files of the LTTng-UST Java agent,
4366 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4368 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4369 path] when you build the Java application.
4371 The JAR files are typically located in dir:{/usr/share/java}.
4373 IMPORTANT: The LTTng-UST Java agent must be
4374 <<installing-lttng,installed>> for the logging framework your
4377 .Use the LTTng-UST Java agent for `java.util.logging`.
4382 import java.io.IOException;
4383 import java.util.logging.Handler;
4384 import java.util.logging.Logger;
4385 import org.lttng.ust.agent.jul.LttngLogHandler;
4389 private static final int answer = 42;
4391 public static void main(String[] argv) throws Exception
4394 Logger logger = Logger.getLogger("jello");
4396 // Create an LTTng-UST log handler
4397 Handler lttngUstLogHandler = new LttngLogHandler();
4399 // Add the LTTng-UST log handler to our logger
4400 logger.addHandler(lttngUstLogHandler);
4403 logger.info("some info");
4404 logger.warning("some warning");
4406 logger.finer("finer information; the answer is " + answer);
4408 logger.severe("error!");
4410 // Not mandatory, but cleaner
4411 logger.removeHandler(lttngUstLogHandler);
4412 lttngUstLogHandler.close();
4421 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4424 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4425 <<enabling-disabling-events,create an event rule>> matching the
4426 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4431 $ lttng enable-event --jul jello
4435 Run the compiled class:
4439 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4442 <<basic-tracing-session-control,Stop tracing>> and inspect the
4452 In the resulting trace, an <<event,event record>> generated by a Java
4453 application using `java.util.logging` is named `lttng_jul:event` and
4454 has the following fields:
4463 Name of the class in which the log statement was executed.
4466 Name of the method in which the log statement was executed.
4469 Logging time (timestamp in milliseconds).
4472 Log level integer value.
4475 ID of the thread in which the log statement was executed.
4477 Use the opt:lttng-enable-event(1):--loglevel or
4478 opt:lttng-enable-event(1):--loglevel-only option of the
4479 man:lttng-enable-event(1) command to target a range of JUL log levels
4480 or a specific JUL log level.
4485 ==== Use the LTTng-UST Java agent for Apache log4j
4487 To use the LTTng-UST Java agent in a Java application which uses
4488 Apache log4j{nbsp}1.2:
4490 . In the source code of the Java application, import the LTTng-UST log
4491 appender package for Apache log4j:
4496 import org.lttng.ust.agent.log4j.LttngLogAppender;
4500 . Create an LTTng-UST log4j log appender:
4505 Appender lttngUstLogAppender = new LttngLogAppender();
4509 . Add this appender to the log4j loggers which should emit LTTng events:
4514 Logger myLogger = Logger.getLogger("some-logger");
4516 myLogger.addAppender(lttngUstLogAppender);
4520 . Use Apache log4j log statements and configuration as usual. The
4521 loggers with an attached LTTng-UST log appender can emit LTTng events.
4523 . Before exiting the application, remove the LTTng-UST log appender from
4524 the loggers attached to it and call its `close()` method:
4529 myLogger.removeAppender(lttngUstLogAppender);
4530 lttngUstLogAppender.close();
4534 This isn't strictly necessary, but it is recommended for a clean
4535 disposal of the resources of the appender.
4537 . Include the common and log4j-specific JAR
4538 files of the LTTng-UST Java agent, path:{lttng-ust-agent-common.jar} and
4539 path:{lttng-ust-agent-log4j.jar}, in the
4540 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4541 path] when you build the Java application.
4543 The JAR files are typically located in dir:{/usr/share/java}.
4545 IMPORTANT: The LTTng-UST Java agent must be
4546 <<installing-lttng,installed>> for the logging framework your
4549 .Use the LTTng-UST Java agent for Apache log4j.
4554 import org.apache.log4j.Appender;
4555 import org.apache.log4j.Logger;
4556 import org.lttng.ust.agent.log4j.LttngLogAppender;
4560 private static final int answer = 42;
4562 public static void main(String[] argv) throws Exception
4565 Logger logger = Logger.getLogger("jello");
4567 // Create an LTTng-UST log appender
4568 Appender lttngUstLogAppender = new LttngLogAppender();
4570 // Add the LTTng-UST log appender to our logger
4571 logger.addAppender(lttngUstLogAppender);
4574 logger.info("some info");
4575 logger.warn("some warning");
4577 logger.debug("debug information; the answer is " + answer);
4579 logger.fatal("error!");
4581 // Not mandatory, but cleaner
4582 logger.removeAppender(lttngUstLogAppender);
4583 lttngUstLogAppender.close();
4589 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4594 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4597 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4598 <<enabling-disabling-events,create an event rule>> matching the
4599 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4604 $ lttng enable-event --log4j jello
4608 Run the compiled class:
4612 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4615 <<basic-tracing-session-control,Stop tracing>> and inspect the
4625 In the resulting trace, an <<event,event record>> generated by a Java
4626 application using log4j is named `lttng_log4j:event` and
4627 has the following fields:
4636 Name of the class in which the log statement was executed.
4639 Name of the method in which the log statement was executed.
4642 Name of the file in which the executed log statement is located.
4645 Line number at which the log statement was executed.
4651 Log level integer value.
4654 Name of the Java thread in which the log statement was executed.
4656 Use the opt:lttng-enable-event(1):--loglevel or
4657 opt:lttng-enable-event(1):--loglevel-only option of the
4658 man:lttng-enable-event(1) command to target a range of Apache log4j
4659 log levels or a specific log4j log level.
4663 [[java-application-context]]
4664 ==== Provide application-specific context fields in a Java application
4666 A Java application-specific context field is a piece of state provided
4667 by the application which <<adding-context,you can add>>, using the
4668 man:lttng-add-context(1) command, to each <<event,event record>>
4669 produced by the log statements of this application.
4671 For example, a given object might have a current request ID variable.
4672 You can create a context information retriever for this object and
4673 assign a name to this current request ID. You can then, using the
4674 man:lttng-add-context(1) command, add this context field by name to
4675 the JUL or log4j <<channel,channel>>.
4677 To provide application-specific context fields in a Java application:
4679 . In the source code of the Java application, import the LTTng-UST
4680 Java agent context classes and interfaces:
4685 import org.lttng.ust.agent.context.ContextInfoManager;
4686 import org.lttng.ust.agent.context.IContextInfoRetriever;
4690 . Create a context information retriever class, that is, a class which
4691 implements the `IContextInfoRetriever` interface:
4696 class MyContextInfoRetriever implements IContextInfoRetriever
4699 public Object retrieveContextInfo(String key)
4701 if (key.equals("intCtx")) {
4703 } else if (key.equals("strContext")) {
4704 return "context value!";
4713 This `retrieveContextInfo()` method is the only member of the
4714 `IContextInfoRetriever` interface. Its role is to return the current
4715 value of a state by name to create a context field. The names of the
4716 context fields and which state variables they return depends on your
4719 All primitive types and objects are supported as context fields.
4720 When `retrieveContextInfo()` returns an object, the context field
4721 serializer calls its `toString()` method to add a string field to
4722 event records. The method can also return `null`, which means that
4723 no context field is available for the required name.
4725 . Register an instance of your context information retriever class to
4726 the context information manager singleton:
4731 IContextInfoRetriever cir = new MyContextInfoRetriever();
4732 ContextInfoManager cim = ContextInfoManager.getInstance();
4733 cim.registerContextInfoRetriever("retrieverName", cir);
4737 . Before exiting the application, remove your context information
4738 retriever from the context information manager singleton:
4743 ContextInfoManager cim = ContextInfoManager.getInstance();
4744 cim.unregisterContextInfoRetriever("retrieverName");
4748 This isn't strictly necessary, but it is recommended for a clean
4749 disposal of some resources of the manager.
4751 . Build your Java application with LTTng-UST Java agent support as
4752 usual, following the procedure for either the <<jul,JUL>> or
4753 <<log4j,Apache log4j>> framework.
4756 .Provide application-specific context fields in a Java application.
4761 import java.util.logging.Handler;
4762 import java.util.logging.Logger;
4763 import org.lttng.ust.agent.jul.LttngLogHandler;
4764 import org.lttng.ust.agent.context.ContextInfoManager;
4765 import org.lttng.ust.agent.context.IContextInfoRetriever;
4769 // Our context information retriever class
4770 private static class MyContextInfoRetriever
4771 implements IContextInfoRetriever
4774 public Object retrieveContextInfo(String key) {
4775 if (key.equals("intCtx")) {
4777 } else if (key.equals("strContext")) {
4778 return "context value!";
4785 private static final int answer = 42;
4787 public static void main(String args[]) throws Exception
4789 // Get the context information manager instance
4790 ContextInfoManager cim = ContextInfoManager.getInstance();
4792 // Create and register our context information retriever
4793 IContextInfoRetriever cir = new MyContextInfoRetriever();
4794 cim.registerContextInfoRetriever("myRetriever", cir);
4797 Logger logger = Logger.getLogger("jello");
4799 // Create an LTTng-UST log handler
4800 Handler lttngUstLogHandler = new LttngLogHandler();
4802 // Add the LTTng-UST log handler to our logger
4803 logger.addHandler(lttngUstLogHandler);
4806 logger.info("some info");
4807 logger.warning("some warning");
4809 logger.finer("finer information; the answer is " + answer);
4811 logger.severe("error!");
4813 // Not mandatory, but cleaner
4814 logger.removeHandler(lttngUstLogHandler);
4815 lttngUstLogHandler.close();
4816 cim.unregisterContextInfoRetriever("myRetriever");
4825 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4828 <<creating-destroying-tracing-sessions,Create a tracing session>>
4829 and <<enabling-disabling-events,create an event rule>> matching the
4835 $ lttng enable-event --jul jello
4838 <<adding-context,Add the application-specific context fields>> to the
4843 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4844 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4847 <<basic-tracing-session-control,Start tracing>>:
4854 Run the compiled class:
4858 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4861 <<basic-tracing-session-control,Stop tracing>> and inspect the
4873 [[python-application]]
4874 === User space Python agent
4876 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4878 https://docs.python.org/3/library/logging.html[`logging`] package.
4880 Each log statement emits an LTTng event once the
4881 application module imports the
4882 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4885 .A Python application importing the LTTng-UST Python agent.
4886 image::python-app.png[]
4888 To use the LTTng-UST Python agent:
4890 . In the source code of the Python application, import the LTTng-UST
4900 The LTTng-UST Python agent automatically adds its logging handler to the
4901 root logger at import time.
4903 Any log statement that the application executes before this import does
4904 not emit an LTTng event.
4906 IMPORTANT: The LTTng-UST Python agent must be
4907 <<installing-lttng,installed>>.
4909 . Use log statements and logging configuration as usual.
4910 Since the LTTng-UST Python agent adds a handler to the _root_
4911 logger, you can trace any log statement from any logger.
4913 .Use the LTTng-UST Python agent.
4924 logging.basicConfig()
4925 logger = logging.getLogger('my-logger')
4928 logger.debug('debug message')
4929 logger.info('info message')
4930 logger.warn('warn message')
4931 logger.error('error message')
4932 logger.critical('critical message')
4936 if __name__ == '__main__':
4940 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4941 logging handler which prints to the standard error stream, isn't
4942 strictly required for LTTng-UST tracing to work, but in versions of
4943 Python preceding{nbsp}3.2, you could see a warning message which indicates
4944 that no handler exists for the logger `my-logger`.
4946 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4947 <<enabling-disabling-events,create an event rule>> matching the
4948 `my-logger` Python logger, and <<basic-tracing-session-control,start
4954 $ lttng enable-event --python my-logger
4958 Run the Python script:
4965 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4975 In the resulting trace, an <<event,event record>> generated by a Python
4976 application is named `lttng_python:event` and has the following fields:
4979 Logging time (string).
4988 Name of the function in which the log statement was executed.
4991 Line number at which the log statement was executed.
4994 Log level integer value.
4997 ID of the Python thread in which the log statement was executed.
5000 Name of the Python thread in which the log statement was executed.
5002 Use the opt:lttng-enable-event(1):--loglevel or
5003 opt:lttng-enable-event(1):--loglevel-only option of the
5004 man:lttng-enable-event(1) command to target a range of Python log levels
5005 or a specific Python log level.
5007 When an application imports the LTTng-UST Python agent, the agent tries
5008 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5009 <<start-sessiond,start the session daemon>> _before_ you run the Python
5010 application. If a session daemon is found, the agent tries to register
5011 to it during five seconds, after which the application continues
5012 without LTTng tracing support. Override this timeout value with
5013 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5016 If the session daemon stops while a Python application with an imported
5017 LTTng-UST Python agent runs, the agent retries to connect and to
5018 register to a session daemon every three seconds. Override this
5019 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5024 [[proc-lttng-logger-abi]]
5027 The `lttng-tracer` Linux kernel module, part of
5028 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger files
5029 path:{/proc/lttng-logger} and path:{/dev/lttng-logger} (since
5030 LTTng{nbsp}2.11) when it's loaded. Any application can write text data
5031 to any of those files to emit an LTTng event.
5034 .An application writes to the LTTng logger file to emit an LTTng event.
5035 image::lttng-logger.png[]
5037 The LTTng logger is the quickest method--not the most efficient,
5038 however--to add instrumentation to an application. It is designed
5039 mostly to instrument shell scripts:
5043 $ echo "Some message, some $variable" > /dev/lttng-logger
5046 Any event that the LTTng logger emits is named `lttng_logger` and
5047 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5048 other instrumentation points in the kernel tracing domain, **any Unix
5049 user** can <<enabling-disabling-events,create an event rule>> which
5050 matches its event name, not only the root user or users in the
5051 <<tracing-group,tracing group>>.
5053 To use the LTTng logger:
5055 * From any application, write text data to the path:{/dev/lttng-logger}
5058 The `msg` field of `lttng_logger` event records contains the
5061 NOTE: The maximum message length of an LTTng logger event is
5062 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5063 than one event to contain the remaining data.
5065 You shouldn't use the LTTng logger to trace a user application which
5066 can be instrumented in a more efficient way, namely:
5068 * <<c-application,C and $$C++$$ applications>>.
5069 * <<java-application,Java applications>>.
5070 * <<python-application,Python applications>>.
5072 .Use the LTTng logger.
5077 echo 'Hello, World!' > /dev/lttng-logger
5079 df --human-readable --print-type / > /dev/lttng-logger
5082 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5083 <<enabling-disabling-events,create an event rule>> matching the
5084 `lttng_logger` Linux kernel tracepoint, and
5085 <<basic-tracing-session-control,start tracing>>:
5090 $ lttng enable-event --kernel lttng_logger
5094 Run the Bash script:
5101 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5112 [[instrumenting-linux-kernel]]
5113 === LTTng kernel tracepoints
5115 NOTE: This section shows how to _add_ instrumentation points to the
5116 Linux kernel. The subsystems of the kernel are already thoroughly
5117 instrumented at strategic places for LTTng when you
5118 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5122 There are two methods to instrument the Linux kernel:
5124 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5125 tracepoint which uses the `TRACE_EVENT()` API.
5127 Choose this if you want to instrumentation a Linux kernel tree with an
5128 instrumentation point compatible with ftrace, perf, and SystemTap.
5130 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5131 instrument an out-of-tree kernel module.
5133 Choose this if you don't need ftrace, perf, or SystemTap support.
5137 [[linux-add-lttng-layer]]
5138 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5140 This section shows how to add an LTTng layer to existing ftrace
5141 instrumentation using the `TRACE_EVENT()` API.
5143 This section doesn't document the `TRACE_EVENT()` macro. Read the
5144 following articles to learn more about this API:
5146 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5147 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5148 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5150 The following procedure assumes that your ftrace tracepoints are
5151 correctly defined in their own header and that they are created in
5152 one source file using the `CREATE_TRACE_POINTS` definition.
5154 To add an LTTng layer over an existing ftrace tracepoint:
5156 . Make sure the following kernel configuration options are
5162 * `CONFIG_HIGH_RES_TIMERS`
5163 * `CONFIG_TRACEPOINTS`
5166 . Build the Linux source tree with your custom ftrace tracepoints.
5167 . Boot the resulting Linux image on your target system.
5169 Confirm that the tracepoints exist by looking for their names in the
5170 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5171 is your subsystem name.
5173 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5178 $ cd $(mktemp -d) &&
5179 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.12.tar.bz2 &&
5180 tar -xf lttng-modules-latest-2.12.tar.bz2 &&
5181 cd lttng-modules-2.12.*
5185 . In dir:{instrumentation/events/lttng-module}, relative to the root
5186 of the LTTng-modules source tree, create a header file named
5187 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5188 LTTng-modules tracepoint definitions using the LTTng-modules
5191 Start with this template:
5195 .path:{instrumentation/events/lttng-module/my_subsys.h}
5198 #define TRACE_SYSTEM my_subsys
5200 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5201 #define _LTTNG_MY_SUBSYS_H
5203 #include "../../../probes/lttng-tracepoint-event.h"
5204 #include <linux/tracepoint.h>
5206 LTTNG_TRACEPOINT_EVENT(
5208 * Format is identical to the TRACE_EVENT() version for the three
5209 * following macro parameters:
5212 TP_PROTO(int my_int, const char *my_string),
5213 TP_ARGS(my_int, my_string),
5215 /* LTTng-modules specific macros */
5217 ctf_integer(int, my_int_field, my_int)
5218 ctf_string(my_bar_field, my_bar)
5222 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5224 #include "../../../probes/define_trace.h"
5228 The entries in the `TP_FIELDS()` section are the list of fields for the
5229 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5230 the `TRACE_EVENT()` ftrace macro.
5232 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5233 complete description of the available `ctf_*()` macros.
5235 . Create the kernel module C{nbsp}source file of the LTTng-modules
5236 probe, +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5241 .path:{probes/lttng-probe-my-subsys.c}
5243 #include <linux/module.h>
5244 #include "../lttng-tracer.h"
5247 * Build-time verification of mismatch between mainline
5248 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5249 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5251 #include <trace/events/my_subsys.h>
5253 /* Create LTTng tracepoint probes */
5254 #define LTTNG_PACKAGE_BUILD
5255 #define CREATE_TRACE_POINTS
5256 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5258 #include "../instrumentation/events/lttng-module/my_subsys.h"
5260 MODULE_LICENSE("GPL and additional rights");
5261 MODULE_AUTHOR("Your name <your-email>");
5262 MODULE_DESCRIPTION("LTTng my_subsys probes");
5263 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5264 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5265 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5266 LTTNG_MODULES_EXTRAVERSION);
5270 . Edit path:{probes/KBuild} and add your new kernel module object
5271 next to the existing ones:
5275 .path:{probes/KBuild}
5279 obj-m += lttng-probe-module.o
5280 obj-m += lttng-probe-power.o
5282 obj-m += lttng-probe-my-subsys.o
5288 . Build and install the LTTng kernel modules:
5293 $ make KERNELDIR=/path/to/linux
5294 # make modules_install && depmod -a
5298 Replace `/path/to/linux` with the path to the Linux source tree where
5299 you defined and used tracepoints with the `TRACE_EVENT()` ftrace macro.
5301 Note that you can also use the
5302 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5303 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5304 C code that need to be executed before the event fields are recorded.
5306 The best way to learn how to use the previous LTTng-modules macros is to
5307 inspect the existing LTTng-modules tracepoint definitions in the
5308 dir:{instrumentation/events/lttng-module} header files. Compare them
5309 with the Linux kernel mainline versions in the
5310 dir:{include/trace/events} directory of the Linux source tree.
5314 [[lttng-tracepoint-event-code]]
5315 ===== Use custom C code to access the data for tracepoint fields
5317 Although we recommended to always use the
5318 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5319 the arguments and fields of an LTTng-modules tracepoint when possible,
5320 sometimes you need a more complex process to access the data that the
5321 tracer records as event record fields. In other words, you need local
5322 variables and multiple C{nbsp}statements instead of simple
5323 argument-based expressions that you pass to the
5324 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5326 Use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5327 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5328 a block of C{nbsp}code to be executed before LTTng records the fields.
5329 The structure of this macro is:
5332 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5334 LTTNG_TRACEPOINT_EVENT_CODE(
5336 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5337 * version for the following three macro parameters:
5340 TP_PROTO(int my_int, const char *my_string),
5341 TP_ARGS(my_int, my_string),
5343 /* Declarations of custom local variables */
5346 unsigned long b = 0;
5347 const char *name = "(undefined)";
5348 struct my_struct *my_struct;
5352 * Custom code which uses both tracepoint arguments
5353 * (in TP_ARGS()) and local variables (in TP_locvar()).
5355 * Local variables are actually members of a structure pointed
5356 * to by the special variable tp_locvar.
5360 tp_locvar->a = my_int + 17;
5361 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5362 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5363 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5364 put_my_struct(tp_locvar->my_struct);
5373 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5374 * version for this, except that tp_locvar members can be
5375 * used in the argument expression parameters of
5376 * the ctf_*() macros.
5379 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5380 ctf_integer(int, my_struct_a, tp_locvar->a)
5381 ctf_string(my_string_field, my_string)
5382 ctf_string(my_struct_name, tp_locvar->name)
5387 IMPORTANT: The C code defined in `TP_code()` must not have any side
5388 effects when executed. In particular, the code must not allocate
5389 memory or get resources without deallocating this memory or putting
5390 those resources afterwards.
5393 [[instrumenting-linux-kernel-tracing]]
5394 ==== Load and unload a custom probe kernel module
5396 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5397 kernel module>> in the kernel before it can emit LTTng events.
5399 To load the default probe kernel modules and a custom probe kernel
5402 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5403 probe modules to load when starting a root <<lttng-sessiond,session
5407 .Load the `my_subsys`, `usb`, and the default probe modules.
5411 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5416 You only need to pass the subsystem name, not the whole kernel module
5419 To load _only_ a given custom probe kernel module:
5421 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5422 modules to load when starting a root session daemon:
5425 .Load only the `my_subsys` and `usb` probe modules.
5429 # lttng-sessiond --kmod-probes=my_subsys,usb
5434 To confirm that a probe module is loaded:
5441 $ lsmod | grep lttng_probe_usb
5445 To unload the loaded probe modules:
5447 * Kill the session daemon with `SIGTERM`:
5452 # pkill lttng-sessiond
5456 You can also use the man:modprobe(8) `--remove` option if the session
5457 daemon terminates abnormally.
5460 [[controlling-tracing]]
5463 Once an application or a Linux kernel is
5464 <<instrumenting,instrumented>> for LTTng tracing,
5467 This section is divided in topics on how to use the various
5468 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5469 command-line tool>>, to _control_ the LTTng daemons and tracers.
5471 NOTE: In the following subsections, we refer to an man:lttng(1) command
5472 using its man page name. For example, instead of _Run the `create`
5473 command to..._, we use _Run the man:lttng-create(1) command to..._.
5477 === Start a session daemon
5479 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5480 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5483 You will see the following error when you run a command while no session
5487 Error: No session daemon is available
5490 The only command that automatically runs a session daemon is
5491 man:lttng-create(1), which you use to
5492 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5493 this is most of the time the first operation that you do, sometimes it's
5494 not. Some examples are:
5496 * <<list-instrumentation-points,List the available instrumentation points>>.
5497 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5499 [[tracing-group]] Each Unix user must have its own running session
5500 daemon to trace user applications. The session daemon that the root user
5501 starts is the only one allowed to control the LTTng kernel tracer. Users
5502 that are part of the _tracing group_ can control the root session
5503 daemon. The default tracing group name is `tracing`; set it to something
5504 else with the opt:lttng-sessiond(8):--group option when you start the
5505 root session daemon.
5507 To start a user session daemon:
5509 * Run man:lttng-sessiond(8):
5514 $ lttng-sessiond --daemonize
5518 To start the root session daemon:
5520 * Run man:lttng-sessiond(8) as the root user:
5525 # lttng-sessiond --daemonize
5529 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5530 start the session daemon in foreground.
5532 To stop a session daemon, use man:kill(1) on its process ID (standard
5535 Note that some Linux distributions could manage the LTTng session daemon
5536 as a service. In this case, you should use the service manager to
5537 start, restart, and stop session daemons.
5540 [[creating-destroying-tracing-sessions]]
5541 === Create and destroy a tracing session
5543 Almost all the LTTng control operations happen in the scope of
5544 a <<tracing-session,tracing session>>, which is the dialogue between the
5545 <<lttng-sessiond,session daemon>> and you.
5547 To create a tracing session with a generated name:
5549 * Use the man:lttng-create(1) command:
5558 The name of the created tracing session is `auto` followed by the
5561 To create a tracing session with a specific name:
5563 * Use the optional argument of the man:lttng-create(1) command:
5568 $ lttng create my-session
5572 Replace `my-session` with the specific tracing session name.
5574 LTTng appends the creation date to the name of the created tracing
5577 LTTng writes the traces of a tracing session in
5578 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5579 name of the tracing session. Note that the env:LTTNG_HOME environment
5580 variable defaults to `$HOME` if not set.
5582 To output LTTng traces to a non-default location:
5584 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5589 $ lttng create my-session --output=/tmp/some-directory
5593 You may create as many tracing sessions as you wish.
5595 To list all the existing tracing sessions for your Unix user:
5597 * Use the man:lttng-list(1) command:
5606 [[cur-tracing-session]]When you create a tracing session, it is set as
5607 the _current tracing session_. The following man:lttng(1) commands
5608 operate on the current tracing session when you don't specify one:
5610 [role="list-3-cols"]
5611 * man:lttng-add-context(1)
5612 * man:lttng-clear(1)
5613 * man:lttng-destroy(1)
5614 * man:lttng-disable-channel(1)
5615 * man:lttng-disable-event(1)
5616 * man:lttng-disable-rotation(1)
5617 * man:lttng-enable-channel(1)
5618 * man:lttng-enable-event(1)
5619 * man:lttng-enable-rotation(1)
5621 * man:lttng-regenerate(1)
5622 * man:lttng-rotate(1)
5624 * man:lttng-snapshot(1)
5625 * man:lttng-start(1)
5626 * man:lttng-status(1)
5628 * man:lttng-track(1)
5629 * man:lttng-untrack(1)
5632 To change the current tracing session:
5634 * Use the man:lttng-set-session(1) command:
5639 $ lttng set-session new-session
5643 Replace `new-session` by the name of the new current tracing session.
5645 When you're done tracing in a given tracing session, destroy it. This
5646 operation frees the resources taken by the tracing session to destroy;
5647 it doesn't destroy the trace data that LTTng wrote for this tracing
5648 session (see <<clear,Clear a tracing session>> for one way to do this).
5650 To destroy the current tracing session:
5652 * Use the man:lttng-destroy(1) command:
5661 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5662 command implicitly (see <<basic-tracing-session-control,Start and stop a
5663 tracing session>>). You need to stop tracing to make LTTng flush the
5664 remaining trace data and make the trace readable.
5667 [[list-instrumentation-points]]
5668 === List the available instrumentation points
5670 The <<lttng-sessiond,session daemon>> can query the running instrumented
5671 user applications and the Linux kernel to get a list of available
5672 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5673 they are tracepoints and system calls. For the user space tracing
5674 domain, they are tracepoints. For the other tracing domains, they are
5677 To list the available instrumentation points:
5679 * Use the man:lttng-list(1) command with the option of the requested
5680 tracing domain amongst:
5683 opt:lttng-list(1):--kernel::
5684 Linux kernel tracepoints (your Unix user must be a root user, or it
5685 must be a member of the <<tracing-group,tracing group>>).
5687 opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall::
5688 Linux kernel system calls (your Unix user must be a root user, or it
5689 must be a member of the tracing group).
5691 opt:lttng-list(1):--userspace::
5692 User space tracepoints.
5694 opt:lttng-list(1):--jul::
5695 `java.util.logging` loggers.
5697 opt:lttng-list(1):--log4j::
5698 Apache log4j loggers.
5700 opt:lttng-list(1):--python::
5704 .List the available user space tracepoints.
5708 $ lttng list --userspace
5712 .List the available Linux kernel system call tracepoints.
5716 $ lttng list --kernel --syscall
5721 [[enabling-disabling-events]]
5722 === Create and enable an event rule
5724 Once you <<creating-destroying-tracing-sessions,create a tracing
5725 session>>, you can create <<event,event rules>> with the
5726 man:lttng-enable-event(1) command.
5728 You specify each condition with a command-line option. The available
5729 condition arguments are shown in the following table.
5731 [role="growable",cols="asciidoc,asciidoc,default"]
5732 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5734 |Argument |Description |Applicable tracing domains
5740 . +--probe=__ADDR__+
5741 . +--function=__ADDR__+
5742 . +--userspace-probe=__PATH__:__SYMBOL__+
5743 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5746 Instead of using the default _tracepoint_ instrumentation type, use:
5748 . A Linux system call (entry and exit).
5749 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5750 . The entry and return points of a Linux function (symbol or address).
5751 . The entry point of a user application or library function (path to
5752 application/library and symbol).
5753 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5754 Statically Defined Tracing] (USDT) probe (path to application/library,
5755 provider and probe names).
5759 |First positional argument.
5762 Tracepoint or system call name.
5764 With the opt:lttng-enable-event(1):--probe,
5765 opt:lttng-enable-event(1):--function, and
5766 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5767 name given to the event rule. With the JUL, log4j, and Python domains,
5768 this is a logger name.
5770 With a tracepoint, logger, or system call name, use the special
5771 `*` globbing character to match anything (for example, `sched_*`,
5779 . +--loglevel=__LEVEL__+
5780 . +--loglevel-only=__LEVEL__+
5783 . Match only tracepoints or log statements with a logging level at
5784 least as severe as +__LEVEL__+.
5785 . Match only tracepoints or log statements with a logging level
5786 equal to +__LEVEL__+.
5788 See man:lttng-enable-event(1) for the list of available logging level
5791 |User space, JUL, log4j, and Python.
5793 |+--exclude=__EXCLUSIONS__+
5796 When you use a `*` character at the end of the tracepoint or logger
5797 name (first positional argument), exclude the specific names in the
5798 comma-delimited list +__EXCLUSIONS__+.
5801 User space, JUL, log4j, and Python.
5803 |+--filter=__EXPR__+
5806 Match only events which satisfy the expression +__EXPR__+.
5808 See man:lttng-enable-event(1) to learn more about the syntax of a
5815 You attach an event rule to a <<channel,channel>> on creation. If you do
5816 not specify the channel with the opt:lttng-enable-event(1):--channel
5817 option, and if the event rule to create is the first in its
5818 <<domain,tracing domain>> for a given tracing session, then LTTng
5819 creates a _default channel_ for you. This default channel is reused in
5820 subsequent invocations of the man:lttng-enable-event(1) command for the
5821 same tracing domain.
5823 An event rule is always enabled at creation time.
5825 The following examples show how to combine the previous
5826 command-line options to create simple to more complex event rules.
5828 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5832 $ lttng enable-event --kernel sched_switch
5836 .Create an event rule matching four Linux kernel system calls (default channel).
5840 $ lttng enable-event --kernel --syscall open,write,read,close
5844 .Create event rules matching tracepoints with filter expressions (default channel).
5848 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5853 $ lttng enable-event --kernel --all \
5854 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5859 $ lttng enable-event --jul my_logger \
5860 --filter='$app.retriever:cur_msg_id > 3'
5863 IMPORTANT: Make sure to always quote the filter string when you
5864 use man:lttng(1) from a shell.
5866 See also <<pid-tracking,Track process attributes>> which offers another,
5867 more efficient filtering mechanism for process ID, user ID, and group
5871 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5875 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5878 IMPORTANT: Make sure to always quote the wildcard character when you
5879 use man:lttng(1) from a shell.
5882 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5886 $ lttng enable-event --python my-app.'*' \
5887 --exclude='my-app.module,my-app.hello'
5891 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5895 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5899 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5903 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5907 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5911 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5916 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
5920 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5921 server_accept_request
5925 The event rules of a given channel form a whitelist: as soon as an
5926 emitted event passes one of them, LTTng can record the event. For
5927 example, an event named `my_app:my_tracepoint` emitted from a user space
5928 tracepoint with a `TRACE_ERROR` log level passes both of the following
5933 $ lttng enable-event --userspace my_app:my_tracepoint
5934 $ lttng enable-event --userspace my_app:my_tracepoint \
5935 --loglevel=TRACE_INFO
5938 The second event rule is redundant: the first one includes
5942 [[disable-event-rule]]
5943 === Disable an event rule
5945 To disable an event rule that you <<enabling-disabling-events,created>>
5946 previously, use the man:lttng-disable-event(1) command. This command
5947 disables _all_ the event rules (of a given tracing domain and channel)
5948 which match an instrumentation point. The other conditions aren't
5949 supported as of LTTng{nbsp}{revision}.
5951 The LTTng tracer doesn't record an emitted event which passes
5952 a _disabled_ event rule.
5954 .Disable an event rule matching a Python logger (default channel).
5958 $ lttng disable-event --python my-logger
5962 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5966 $ lttng disable-event --jul '*'
5970 .Disable _all_ the event rules of the default channel.
5972 The opt:lttng-disable-event(1):--all-events option isn't, like the
5973 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5974 equivalent of the event name `*` (wildcard): it disables _all_ the event
5975 rules of a given channel.
5979 $ lttng disable-event --jul --all-events
5983 NOTE: You can't delete an event rule once you create it.
5987 === Get the status of a tracing session
5989 To get the status of the <<cur-tracing-session,current tracing
5990 session>>, that is, its parameters, its channels, event rules, and their
5993 * Use the man:lttng-status(1) command:
6002 To get the status of any tracing session:
6004 * Use the man:lttng-list(1) command with the name of the tracing
6010 $ lttng list my-session
6014 Replace `my-session` with the desired tracing session name.
6017 [[basic-tracing-session-control]]
6018 === Start and stop a tracing session
6020 Once you <<creating-destroying-tracing-sessions,create a tracing
6022 <<enabling-disabling-events,create one or more event rules>>,
6023 you can start and stop the tracers for this tracing session.
6025 To start tracing in the <<cur-tracing-session,current tracing session>>:
6027 * Use the man:lttng-start(1) command:
6036 LTTng is very flexible: you can launch user applications before
6037 or after the you start the tracers. The tracers only record the events
6038 if they pass enabled event rules and if they occur while the tracers are
6041 To stop tracing in the current tracing session:
6043 * Use the man:lttng-stop(1) command:
6052 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6053 records>> or lost sub-buffers since the last time you ran
6054 man:lttng-start(1), warnings are printed when you run the
6055 man:lttng-stop(1) command.
6057 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
6058 trace data and make the trace readable. Note that the
6059 man:lttng-destroy(1) command (see
6060 <<creating-destroying-tracing-sessions,Create and destroy a tracing
6061 session>>) also runs the man:lttng-stop(1) command implicitly.
6065 === Clear a tracing session
6067 You might need to remove all the current tracing data of one or more
6068 <<tracing-session,tracing sessions>> between multiple attempts to
6069 reproduce a problem without interrupting the LTTng tracing activity.
6071 To clear the tracing data of the
6072 <<cur-tracing-session,current tracing session>>:
6074 * Use the man:lttng-clear(1) command:
6083 To clear the tracing data of all the tracing sessions:
6085 * Use the `lttng clear` command with the opt:lttng-clear(1):--all
6096 [[enabling-disabling-channels]]
6097 === Create a channel
6099 Once you create a tracing session, you can create a <<channel,channel>>
6100 with the man:lttng-enable-channel(1) command.
6102 Note that LTTng automatically creates a default channel when, for a
6103 given <<domain,tracing domain>>, no channels exist and you
6104 <<enabling-disabling-events,create>> the first event rule. This default
6105 channel is named `channel0` and its attributes are set to reasonable
6106 values. Therefore, you only need to create a channel when you need
6107 non-default attributes.
6109 You specify each non-default channel attribute with a command-line
6110 option when you use the man:lttng-enable-channel(1) command. The
6111 available command-line options are:
6113 [role="growable",cols="asciidoc,asciidoc"]
6114 .Command-line options for the man:lttng-enable-channel(1) command.
6116 |Option |Description
6122 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
6123 of the default _discard_ mode.
6125 |`--buffers-pid` (user space tracing domain only)
6128 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6129 instead of the default per-user buffering scheme.
6131 |+--subbuf-size=__SIZE__+
6134 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6135 either for each Unix user (default), or for each instrumented process.
6137 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6139 |+--num-subbuf=__COUNT__+
6142 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6143 for each Unix user (default), or for each instrumented process.
6145 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6147 |+--tracefile-size=__SIZE__+
6150 Set the maximum size of each trace file that this channel writes within
6151 a stream to +__SIZE__+ bytes instead of no maximum.
6153 See <<tracefile-rotation,Trace file count and size>>.
6155 |+--tracefile-count=__COUNT__+
6158 Limit the number of trace files that this channel creates to
6159 +__COUNT__+ channels instead of no limit.
6161 See <<tracefile-rotation,Trace file count and size>>.
6163 |+--switch-timer=__PERIODUS__+
6166 Set the <<channel-switch-timer,switch timer period>>
6167 to +__PERIODUS__+{nbsp}µs.
6169 |+--read-timer=__PERIODUS__+
6172 Set the <<channel-read-timer,read timer period>>
6173 to +__PERIODUS__+{nbsp}µs.
6175 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6178 Set the timeout of user space applications which load LTTng-UST
6179 in blocking mode to +__TIMEOUTUS__+:
6182 Never block (non-blocking mode).
6185 Block forever until space is available in a sub-buffer to record
6188 __n__, a positive value::
6189 Wait for at most __n__ µs when trying to write into a sub-buffer.
6191 Note that, for this option to have any effect on an instrumented
6192 user space application, you need to run the application with a set
6193 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6195 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6198 Set the output type of the channel to +__TYPE__+, either `mmap` or
6203 You can only create a channel in the Linux kernel and user space
6204 <<domain,tracing domains>>: other tracing domains have their own channel
6205 created on the fly when <<enabling-disabling-events,creating event
6210 Because of a current LTTng limitation, you must create all channels
6211 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6212 tracing session, that is, before the first time you run
6215 Since LTTng automatically creates a default channel when you use the
6216 man:lttng-enable-event(1) command with a specific tracing domain, you
6217 can't, for example, create a Linux kernel event rule, start tracing,
6218 and then create a user space event rule, because no user space channel
6219 exists yet and it's too late to create one.
6221 For this reason, make sure to configure your channels properly
6222 before starting the tracers for the first time!
6225 The following examples show how to combine the previous
6226 command-line options to create simple to more complex channels.
6228 .Create a Linux kernel channel with default attributes.
6232 $ lttng enable-channel --kernel my-channel
6236 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6240 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6241 --buffers-pid my-channel
6245 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6247 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6248 create the channel, <<enabling-disabling-events,create an event rule>>,
6249 and <<basic-tracing-session-control,start tracing>>:
6254 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6255 $ lttng enable-event --userspace --channel=blocking-channel --all
6259 Run an application instrumented with LTTng-UST and allow it to block:
6263 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6267 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6271 $ lttng enable-channel --kernel --tracefile-count=8 \
6272 --tracefile-size=4194304 my-channel
6276 .Create a user space channel in overwrite (or _flight recorder_) mode.
6280 $ lttng enable-channel --userspace --overwrite my-channel
6284 <<enabling-disabling-events,Create>> the same event rule in
6285 two different channels:
6289 $ lttng enable-event --userspace --channel=my-channel app:tp
6290 $ lttng enable-event --userspace --channel=other-channel app:tp
6293 If both channels are enabled, when a tracepoint named `app:tp` is
6294 reached, LTTng records two events, one for each channel.
6298 === Disable a channel
6300 To disable a specific channel that you <<enabling-disabling-channels,created>>
6301 previously, use the man:lttng-disable-channel(1) command.
6303 .Disable a specific Linux kernel channel.
6307 $ lttng disable-channel --kernel my-channel
6311 The state of a channel precedes the individual states of event rules
6312 attached to it: event rules which belong to a disabled channel, even if
6313 they are enabled, are also considered disabled.
6317 === Add context fields to a channel
6319 Event record fields in trace files provide important information about
6320 events that occured previously, but sometimes some external context may
6321 help you solve a problem faster.
6323 Examples of context fields are:
6325 * The **process ID**, **thread ID**, **process name**, and
6326 **process priority** of the thread in which the event occurs.
6327 * The **hostname** of the system on which the event occurs.
6328 * The Linux kernel and user call stacks (since
6330 * The current values of many possible **performance counters** using
6332 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6334 ** Branch instructions, misses, and loads.
6336 * Any context defined at the application level (supported for the
6337 JUL and log4j <<domain,tracing domains>>).
6339 To get the full list of available context fields, see
6340 `lttng add-context --list`. Some context fields are reserved for a
6341 specific <<domain,tracing domain>> (Linux kernel or user space).
6343 You add context fields to <<channel,channels>>. All the events
6344 that a channel with added context fields records contain those fields.
6346 To add context fields to one or all the channels of a given tracing
6349 * Use the man:lttng-add-context(1) command.
6351 .Add context fields to all the channels of the current tracing session.
6353 The following command line adds the virtual process identifier and
6354 the per-thread CPU cycles count fields to all the user space channels
6356 <<cur-tracing-session,current tracing session>>.
6360 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6364 .Add performance counter context fields by raw ID
6366 See man:lttng-add-context(1) for the exact format of the context field
6367 type, which is partly compatible with the format used in
6372 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6373 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6377 .Add context fields to a specific channel.
6379 The following command line adds the thread identifier and user call
6380 stack context fields to the Linux kernel channel named `my-channel` in
6381 the current tracing session.
6385 $ lttng add-context --kernel --channel=my-channel \
6386 --type=tid --type=callstack-user
6390 .Add an application-specific context field to a specific channel.
6392 The following command line adds the `cur_msg_id` context field of the
6393 `retriever` context retriever for all the instrumented
6394 <<java-application,Java applications>> recording <<event,event records>>
6395 in the channel named `my-channel`:
6399 $ lttng add-context --kernel --channel=my-channel \
6400 --type='$app:retriever:cur_msg_id'
6403 IMPORTANT: Make sure to always quote the `$` character when you
6404 use man:lttng-add-context(1) from a shell.
6407 NOTE: You can't remove context fields from a channel once you add it.
6412 === Track process attributes
6414 It's often useful to only allow processes with specific attributes to
6415 emit events. For example, you may wish to record all the system calls
6416 which a given process makes (à la
6417 http://linux.die.net/man/1/strace[strace]).
6419 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6420 purpose. Both commands operate on _inclusion sets_ of process attribute
6421 values. The available process attribute types are:
6423 Linux kernel <<domain,tracing domain>> only::
6427 * Virtual process ID (VPID).
6429 This is the PID as seen by the application.
6431 * Unix user ID (UID) (since LTTng{nbsp}2.12).
6433 * Virtual Unix user ID (VUID) (since LTTng{nbsp}2.12).
6435 This is the UID as seen by the application.
6437 * Unix group ID (GID) (since LTTng{nbsp}2.12).
6439 * Virtual Unix group ID (VGID) (since LTTng{nbsp}2.12).
6441 This is the GID as seen by the application.
6444 User space tracing domain::
6447 * VUID (since LTTng{nbsp}2.12).
6448 * VGID (since LTTng{nbsp}2.12).
6450 Each tracing domain has one inclusion set per process attribute type:
6451 the Linux kernel tracing domain has six while the user space tracing
6454 For a given event which passes an enabled <<event,event rule>> to be
6455 recorded, _all_ the attributes of its executing process must be part of
6456 the inclusion sets of the tracing domain of the event rule.
6458 Add entries to an inclusion set with the man:lttng-track(1) command and
6459 remove entries with the man:lttng-untrack(1) command. A process
6460 attribute is _tracked_ when it's part of an inclusion set and
6461 _untracked_ otherwise.
6465 The process attribute values are _numeric_.
6467 Should a process with a given tracked process ID, for example, exit, and
6468 then a new process be given this ID, then the latter would also be
6469 allowed to emit events.
6471 With the `lttng track` command, you can add Unix user and group _names_
6472 to the user and group inclusion sets: the <<lttng-sessiond,session
6473 daemon>> finds the corresponding UID, VUID, GID, or VGID once on
6474 _addition_ to the inclusion set. This means that if you rename the user
6475 or group after you run `lttng track`, its user/group ID remains tracked.
6478 .Track and untrack virtual process IDs.
6480 For the sake of the following example, assume the target system has
6481 16{nbsp}possible VPIDs.
6484 <<creating-destroying-tracing-sessions,create a tracing session>>,
6485 the user space VPID inclusion set contains _all_ the possible VPIDs:
6488 .All VPIDs are tracked.
6489 image::track-all.png[]
6491 When the inclusion set is full and you use the man:lttng-track(1)
6492 command to specify some VPIDs to track, LTTng first clears the inclusion
6493 set, and then it adds the specific VPIDs to track. After:
6497 $ lttng track --userspace --vpid=3,4,7,10,13
6500 the VPID inclusion set is:
6503 .VPIDs 3, 4, 7, 10, and 13 are tracked.
6504 image::track-3-4-7-10-13.png[]
6506 Add more VPIDs to the inclusion set afterwards:
6510 $ lttng track --userspace --vpid=1,15,16
6516 .VPIDs 1, 15, and 16 are added to the inclusion set.
6517 image::track-1-3-4-7-10-13-15-16.png[]
6519 The man:lttng-untrack(1) command removes entries from process attribute
6520 inclusion sets. Given the previous example, the following command:
6524 $ lttng untrack --userspace --vpid=3,7,10,13
6527 leads to this VPID inclusion set:
6530 .VPIDs 3, 7, 10, and 13 are removed from the inclusion set.
6531 image::track-1-4-15-16.png[]
6533 LTTng can track all the possible VPIDs again using the
6534 opt:lttng-track(1):--all option:
6538 $ lttng track --userspace --vpid --all
6541 The result is, again:
6544 .All VPIDs are tracked.
6545 image::track-all.png[]
6548 .Track only specific process attributes.
6550 A typical use case with process attribute tracking is to start with an
6551 empty inclusion set, then <<basic-tracing-session-control,start the
6552 tracers>>, and then add entries manually while the tracers are active.
6554 Use the opt:lttng-untrack(1):--all option of the
6555 man:lttng-untrack(1) command to clear the inclusion set after you
6556 <<creating-destroying-tracing-sessions,create a tracing session>>, for
6557 example (with UIDs):
6561 $ lttng untrack --kernel --uid --all
6567 .No UIDs are tracked.
6568 image::untrack-all.png[]
6570 If you trace with this inclusion set configuration, the LTTng kernel
6571 tracer records no events within the <<cur-tracing-session,current
6572 tracing session>> because it doesn't track any UID. Use the
6573 man:lttng-track(1) command as usual to track specific UIDs when you need
6578 $ lttng track --kernel --uid=http,11
6584 .UIDs 6 (`http`) and 11 are tracked.
6585 image::track-6-11.png[]
6590 [[saving-loading-tracing-session]]
6591 === Save and load tracing session configurations
6593 Configuring a <<tracing-session,tracing session>> can be long. Some of
6594 the tasks involved are:
6596 * <<enabling-disabling-channels,Create channels>> with
6597 specific attributes.
6598 * <<adding-context,Add context fields>> to specific channels.
6599 * <<enabling-disabling-events,Create event rules>> with specific log
6600 level and filter conditions.
6602 If you use LTTng to solve real world problems, chances are you have to
6603 record events using the same tracing session setup over and over,
6604 modifying a few variables each time in your instrumented program
6605 or environment. To avoid constant tracing session reconfiguration,
6606 the man:lttng(1) command-line tool can save and load tracing session
6607 configurations to/from XML files.
6609 To save a given tracing session configuration:
6611 * Use the man:lttng-save(1) command:
6616 $ lttng save my-session
6620 Replace `my-session` with the name of the tracing session to save.
6622 LTTng saves tracing session configurations to
6623 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6624 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6625 the opt:lttng-save(1):--output-path option to change this destination
6628 LTTng saves all configuration parameters, for example:
6630 * The tracing session name.
6631 * The trace data output path.
6632 * The channels with their state and all their attributes.
6633 * The context fields you added to channels.
6634 * The event rules with their state, log level and filter conditions.
6636 To load a tracing session:
6638 * Use the man:lttng-load(1) command:
6643 $ lttng load my-session
6647 Replace `my-session` with the name of the tracing session to load.
6649 When LTTng loads a configuration, it restores your saved tracing session
6650 as if you just configured it manually.
6652 See man:lttng-load(1) for the complete list of command-line options. You
6653 can also save and load many sessions at a time, and decide in which
6654 directory to output the XML files.
6657 [[sending-trace-data-over-the-network]]
6658 === Send trace data over the network
6660 LTTng can send the recorded trace data to a remote system over the
6661 network instead of writing it to the local file system.
6663 To send the trace data over the network:
6665 . On the _remote_ system (which can also be the target system),
6666 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6675 . On the _target_ system, create a tracing session configured to
6676 send trace data over the network:
6681 $ lttng create my-session --set-url=net://remote-system
6685 Replace `remote-system` by the host name or IP address of the
6686 remote system. See man:lttng-create(1) for the exact URL format.
6688 . On the target system, use the man:lttng(1) command-line tool as usual.
6689 When tracing is active, the consumer daemon of the target sends
6690 sub-buffers to the relay daemon running on the remote system instead
6691 of flushing them to the local file system. The relay daemon writes the
6692 received packets to the local file system.
6694 The relay daemon writes trace files to
6695 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6696 +__hostname__+ is the host name of the target system and +__session__+
6697 is the tracing session name. Note that the env:LTTNG_HOME environment
6698 variable defaults to `$HOME` if not set. Use the
6699 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6700 trace files to another base directory.
6705 === View events as LTTng emits them (noch:{LTTng} live)
6707 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6708 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6709 display events as LTTng emits them on the target system while tracing is
6712 The relay daemon creates a _tee_: it forwards the trace data to both
6713 the local file system and to connected live viewers:
6716 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6721 . On the _target system_, create a <<tracing-session,tracing session>>
6727 $ lttng create my-session --live
6731 This spawns a local relay daemon.
6733 . Start the live viewer and configure it to connect to the relay
6734 daemon. For example, with
6735 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]:
6740 $ babeltrace2 net://localhost/host/hostname/my-session
6747 * `hostname` with the host name of the target system.
6748 * `my-session` with the name of the tracing session to view.
6751 . Configure the tracing session as usual with the man:lttng(1)
6752 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6754 List the available live tracing sessions with Babeltrace{nbsp}2:
6758 $ babeltrace2 net://localhost
6761 You can start the relay daemon on another system. In this case, you need
6762 to specify the URL of the relay daemon when you create the tracing
6763 session with the opt:lttng-create(1):--set-url option. You also need to
6764 replace `localhost` in the procedure above with the host name of the
6765 system on which the relay daemon is running.
6767 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6768 command-line options.
6772 [[taking-a-snapshot]]
6773 === Take a snapshot of the current sub-buffers of a tracing session
6775 The normal behavior of LTTng is to append full sub-buffers to growing
6776 trace data files. This is ideal to keep a full history of the events
6777 that occurred on the target system, but it can
6778 represent too much data in some situations. For example, you may wish
6779 to trace your application continuously until some critical situation
6780 happens, in which case you only need the latest few recorded
6781 events to perform the desired analysis, not multi-gigabyte trace files.
6783 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6784 current sub-buffers of a given <<tracing-session,tracing session>>.
6785 LTTng can write the snapshot to the local file system or send it over
6789 .A snapshot is a copy of the current sub-buffers, which aren't cleared after the operation.
6790 image::snapshot.png[]
6792 If you wish to create unmanaged, self-contained, non-overlapping
6793 trace chunk archives instead of a simple copy of the current
6794 sub-buffers, see the <<session-rotation,tracing session rotation>>
6795 feature (available since LTTng{nbsp}2.11).
6799 . Create a tracing session in _snapshot mode_:
6804 $ lttng create my-session --snapshot
6808 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6809 <<channel,channels>> created in this mode is automatically set to
6810 _overwrite_ (flight recorder mode).
6812 . Configure the tracing session as usual with the man:lttng(1)
6813 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6815 . **Optional**: When you need to take a snapshot,
6816 <<basic-tracing-session-control,stop tracing>>.
6818 You can take a snapshot when the tracers are active, but if you stop
6819 them first, you're sure that the data in the sub-buffers doesn't
6820 change before you actually take the snapshot.
6827 $ lttng snapshot record --name=my-first-snapshot
6831 LTTng writes the current sub-buffers of all the channels of the
6832 <<cur-tracing-session,current tracing session>> to
6833 trace files on the local file system. Those trace files have
6834 `my-first-snapshot` in their name.
6836 There is no difference between the format of a normal trace file and the
6837 format of a snapshot: viewers of LTTng traces also support LTTng
6840 By default, LTTng writes snapshot files to the path shown by
6841 `lttng snapshot list-output`. You can change this path or decide to send
6842 snapshots over the network using either:
6844 . An output path or URL that you specify when you
6845 <<creating-destroying-tracing-sessions,create the tracing session>>.
6846 . A snapshot output path or URL that you add using
6847 `lttng snapshot add-output`.
6848 . An output path or URL that you provide directly to the
6849 `lttng snapshot record` command.
6851 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6852 you specify a URL, a relay daemon must listen on a remote system (see
6853 <<sending-trace-data-over-the-network,Send trace data over the
6858 [[session-rotation]]
6859 === Archive the current trace chunk (rotate a tracing session)
6861 The <<taking-a-snapshot,snapshot user guide>> shows how to dump the
6862 current sub-buffers of a tracing session to the file system or send them
6863 over the network. When you take a snapshot, LTTng doesn't clear the ring
6864 buffers of the tracing session: if you take another snapshot immediately
6865 after, both snapshots could contain overlapping trace data.
6867 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6868 _tracing session rotation_ is a feature which appends the content of the
6869 ring buffers to what's already on the file system or sent over the
6870 network since the creation of the tracing session or since the last
6871 rotation, and then clears those ring buffers to avoid trace data
6874 What LTTng is about to write when performing a tracing session rotation
6875 is called the _current trace chunk_. When this current trace chunk is
6876 written to the file system or sent over the network, it becomes a _trace
6877 chunk archive_. Therefore, a tracing session rotation _archives_ the
6878 current trace chunk.
6881 .A tracing session rotation operation _archives_ the current trace chunk.
6882 image::rotation.png[]
6884 A trace chunk archive is a self-contained LTTng trace which LTTng
6885 doesn't manage anymore: you can read it, modify it, move it, or remove
6888 There are two methods to perform a tracing session rotation: immediately
6889 or with a rotation schedule.
6891 To perform an immediate tracing session rotation:
6893 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6894 in _normal mode_ or _network streaming mode_
6895 (only those two creation modes support tracing session rotation):
6900 $ lttng create my-session
6904 . <<enabling-disabling-events,Create one or more event rules>>
6905 and <<basic-tracing-session-control,start tracing>>:
6910 $ lttng enable-event --kernel sched_'*'
6915 . When needed, immediately rotate the
6916 <<cur-tracing-session,current tracing session>>:
6925 The cmd:lttng-rotate command prints the path to the created trace
6926 chunk archive. See man:lttng-rotate(1) to learn about the format
6927 of trace chunk archive directory names.
6929 Perform other immediate rotations while the tracing session is
6930 active. It is guaranteed that all the trace chunk archives don't
6931 contain overlapping trace data. You can also perform an immediate
6932 rotation once you have <<basic-tracing-session-control,stopped>> the
6935 . When you're done tracing,
6936 <<creating-destroying-tracing-sessions,destroy the current tracing
6946 The tracing session destruction operation creates one last trace
6947 chunk archive from the current trace chunk.
6949 A tracing session rotation schedule is a planned rotation which LTTng
6950 performs automatically based on one of the following conditions:
6952 * A timer with a configured period times out.
6954 * The total size of the flushed part of the current trace chunk
6955 becomes greater than or equal to a configured value.
6957 To schedule a tracing session rotation, set a _rotation schedule_:
6959 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6960 in _normal mode_ or _network streaming mode_
6961 (only those two creation modes support tracing session rotation):
6966 $ lttng create my-session
6970 . <<enabling-disabling-events,Create one or more event rules>>:
6975 $ lttng enable-event --kernel sched_'*'
6979 . Set a tracing session rotation schedule:
6984 $ lttng enable-rotation --timer=10s
6988 In this example, we set a rotation schedule so that LTTng performs a
6989 tracing session rotation every ten seconds.
6991 See man:lttng-enable-rotation(1) to learn more about other ways to set a
6994 . <<basic-tracing-session-control,Start tracing>>:
7003 LTTng performs tracing session rotations automatically while the tracing
7004 session is active thanks to the rotation schedule.
7006 . When you're done tracing,
7007 <<creating-destroying-tracing-sessions,destroy the current tracing
7017 The tracing session destruction operation creates one last trace chunk
7018 archive from the current trace chunk.
7020 Use man:lttng-disable-rotation(1) to unset a tracing session
7023 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
7024 limitations regarding those two commands.
7029 === Use the machine interface
7031 With any command of the man:lttng(1) command-line tool, set the
7032 opt:lttng(1):--mi option to `xml` (before the command name) to get an
7033 XML machine interface output, for example:
7037 $ lttng --mi=xml enable-event --kernel --syscall open
7040 A schema definition (XSD) is
7041 https://github.com/lttng/lttng-tools/blob/stable-{revision}/src/common/src/common/mi-lttng-4.0.xsd[available]
7042 to ease the integration with external tools as much as possible.
7046 [[metadata-regenerate]]
7047 === Regenerate the metadata of an LTTng trace
7049 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
7050 data stream files and a metadata file. This metadata file contains,
7051 amongst other things, information about the offset of the clock sources
7052 used to timestamp <<event,event records>> when tracing.
7054 If, once a <<tracing-session,tracing session>> is
7055 <<basic-tracing-session-control,started>>, a major
7056 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
7057 happens, the clock offset of the trace also needs to be updated. Use
7058 the `metadata` item of the man:lttng-regenerate(1) command to do so.
7060 The main use case of this command is to allow a system to boot with
7061 an incorrect wall time and trace it with LTTng before its wall time
7062 is corrected. Once the system is known to be in a state where its
7063 wall time is correct, it can run `lttng regenerate metadata`.
7065 To regenerate the metadata of an LTTng trace:
7067 * Use the `metadata` item of the man:lttng-regenerate(1) command:
7072 $ lttng regenerate metadata
7078 `lttng regenerate metadata` has the following limitations:
7080 * Tracing session <<creating-destroying-tracing-sessions,created>>
7082 * User space <<channel,channels>>, if any, are using
7083 <<channel-buffering-schemes,per-user buffering>>.
7088 [[regenerate-statedump]]
7089 === Regenerate the state dump of a tracing session
7091 The LTTng kernel and user space tracers generate state dump
7092 <<event,event records>> when the application starts or when you
7093 <<basic-tracing-session-control,start a tracing session>>. An analysis
7094 can use the state dump event records to set an initial state before it
7095 builds the rest of the state from the following event records.
7096 http://tracecompass.org/[Trace Compass] is a notable example of an
7097 application which uses the state dump of an LTTng trace.
7099 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
7100 state dump event records aren't included in the snapshot because they
7101 were recorded to a sub-buffer that has been consumed or overwritten
7104 Use the `lttng regenerate statedump` command to emit the state
7105 dump event records again.
7107 To regenerate the state dump of the current tracing session, provided
7108 create it in snapshot mode, before you take a snapshot:
7110 . Use the `statedump` item of the man:lttng-regenerate(1) command:
7115 $ lttng regenerate statedump
7119 . <<basic-tracing-session-control,Stop the tracing session>>:
7128 . <<taking-a-snapshot,Take a snapshot>>:
7133 $ lttng snapshot record --name=my-snapshot
7137 Depending on the event throughput, you should run steps 1 and 2
7138 as closely as possible.
7140 NOTE: To record the state dump events, you need to
7141 <<enabling-disabling-events,create event rules>> which enable them.
7142 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
7143 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
7147 [[persistent-memory-file-systems]]
7148 === Record trace data on persistent memory file systems
7150 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
7151 (NVRAM) is random-access memory that retains its information when power
7152 is turned off (non-volatile). Systems with such memory can store data
7153 structures in RAM and retrieve them after a reboot, without flushing
7154 to typical _storage_.
7156 Linux supports NVRAM file systems thanks to either
7157 http://pramfs.sourceforge.net/[PRAMFS] or
7158 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
7159 (requires Linux{nbsp}4.1+).
7161 This section doesn't describe how to operate such file systems;
7162 we assume that you have a working persistent memory file system.
7164 When you create a <<tracing-session,tracing session>>, you can specify
7165 the path of the shared memory holding the sub-buffers. If you specify a
7166 location on an NVRAM file system, then you can retrieve the latest
7167 recorded trace data when the system reboots after a crash.
7169 To record trace data on a persistent memory file system and retrieve the
7170 trace data after a system crash:
7172 . Create a tracing session with a sub-buffer shared memory path located
7173 on an NVRAM file system:
7178 $ lttng create my-session --shm-path=/path/to/shm
7182 . Configure the tracing session as usual with the man:lttng(1)
7183 command-line tool, and <<basic-tracing-session-control,start tracing>>.
7185 . After a system crash, use the man:lttng-crash(1) command-line tool to
7186 view the trace data recorded on the NVRAM file system:
7191 $ lttng-crash /path/to/shm
7195 The binary layout of the ring buffer files isn't exactly the same as
7196 the trace files layout. This is why you need to use man:lttng-crash(1)
7197 instead of your preferred trace viewer directly.
7199 To convert the ring buffer files to LTTng trace files:
7201 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
7206 $ lttng-crash --extract=/path/to/trace /path/to/shm
7212 [[notif-trigger-api]]
7213 === Get notified when the buffer usage of a channel is too high or too low
7215 With the $$C/C++$$ notification and trigger API of LTTng, your user
7216 application can get notified when the buffer usage of one or more
7217 <<channel,channels>> becomes too low or too high. Use this API
7218 and enable or disable <<event,event rules>> during tracing to avoid
7219 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7221 .Have a user application get notified when the buffer usage of an LTTng channel is too high.
7223 In this example, we create and build an application which gets notified
7224 when the buffer usage of a specific LTTng channel is higher than
7225 75{nbsp}%. We only print that it is the case in the example, but we
7226 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7227 disable event rules when this happens.
7229 . Create the C{nbsp}source file of application:
7237 #include <lttng/domain.h>
7238 #include <lttng/action/action.h>
7239 #include <lttng/action/notify.h>
7240 #include <lttng/condition/condition.h>
7241 #include <lttng/condition/buffer-usage.h>
7242 #include <lttng/condition/evaluation.h>
7243 #include <lttng/notification/channel.h>
7244 #include <lttng/notification/notification.h>
7245 #include <lttng/trigger/trigger.h>
7246 #include <lttng/endpoint.h>
7248 int main(int argc, char *argv[])
7250 int exit_status = 0;
7251 struct lttng_notification_channel *notification_channel;
7252 struct lttng_condition *condition;
7253 struct lttng_action *action;
7254 struct lttng_trigger *trigger;
7255 const char *tracing_session_name;
7256 const char *channel_name;
7259 tracing_session_name = argv[1];
7260 channel_name = argv[2];
7263 * Create a notification channel. A notification channel
7264 * connects the user application to the LTTng session daemon.
7265 * This notification channel can be used to listen to various
7266 * types of notifications.
7268 notification_channel = lttng_notification_channel_create(
7269 lttng_session_daemon_notification_endpoint);
7272 * Create a "high buffer usage" condition. In this case, the
7273 * condition is reached when the buffer usage is greater than or
7274 * equal to 75 %. We create the condition for a specific tracing
7275 * session name, channel name, and for the user space tracing
7278 * The "low buffer usage" condition type also exists.
7280 condition = lttng_condition_buffer_usage_high_create();
7281 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7282 lttng_condition_buffer_usage_set_session_name(
7283 condition, tracing_session_name);
7284 lttng_condition_buffer_usage_set_channel_name(condition,
7286 lttng_condition_buffer_usage_set_domain_type(condition,
7290 * Create an action (get a notification) to take when the
7291 * condition created above is reached.
7293 action = lttng_action_notify_create();
7296 * Create a trigger. A trigger associates a condition to an
7297 * action: the action is executed when the condition is reached.
7299 trigger = lttng_trigger_create(condition, action);
7301 /* Register the trigger to LTTng. */
7302 lttng_register_trigger(trigger);
7305 * Now that we have registered a trigger, a notification will be
7306 * emitted everytime its condition is met. To receive this
7307 * notification, we must subscribe to notifications that match
7308 * the same condition.
7310 lttng_notification_channel_subscribe(notification_channel,
7314 * Notification loop. Put this in a dedicated thread to avoid
7315 * blocking the main thread.
7318 struct lttng_notification *notification;
7319 enum lttng_notification_channel_status status;
7320 const struct lttng_evaluation *notification_evaluation;
7321 const struct lttng_condition *notification_condition;
7322 double buffer_usage;
7324 /* Receive the next notification. */
7325 status = lttng_notification_channel_get_next_notification(
7326 notification_channel, ¬ification);
7329 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7331 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7333 * The session daemon can drop notifications if a monitoring
7334 * application isn't consuming the notifications fast
7338 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7340 * The notification channel has been closed by the
7341 * session daemon. This is typically caused by a session
7342 * daemon shutting down.
7346 /* Unhandled conditions or errors. */
7352 * A notification provides, amongst other things:
7354 * * The condition that caused this notification to be
7356 * * The condition evaluation, which provides more
7357 * specific information on the evaluation of the
7360 * The condition evaluation provides the buffer usage
7361 * value at the moment the condition was reached.
7363 notification_condition = lttng_notification_get_condition(
7365 notification_evaluation = lttng_notification_get_evaluation(
7368 /* We're subscribed to only one condition. */
7369 assert(lttng_condition_get_type(notification_condition) ==
7370 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7373 * Get the exact sampled buffer usage from the
7374 * condition evaluation.
7376 lttng_evaluation_buffer_usage_get_usage_ratio(
7377 notification_evaluation, &buffer_usage);
7380 * At this point, instead of printing a message, we
7381 * could do something to reduce the buffer usage of the channel,
7382 * like disable specific events.
7384 printf("Buffer usage is %f %% in tracing session \"%s\", "
7385 "user space channel \"%s\".\n", buffer_usage * 100,
7386 tracing_session_name, channel_name);
7387 lttng_notification_destroy(notification);
7391 lttng_action_destroy(action);
7392 lttng_condition_destroy(condition);
7393 lttng_trigger_destroy(trigger);
7394 lttng_notification_channel_destroy(notification_channel);
7400 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7405 $ gcc -o notif-app notif-app.c -llttng-ctl
7409 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7410 <<enabling-disabling-events,create an event rule>> matching all the
7411 user space tracepoints, and
7412 <<basic-tracing-session-control,start tracing>>:
7417 $ lttng create my-session
7418 $ lttng enable-event --userspace --all
7423 If you create the channel manually with the man:lttng-enable-channel(1)
7424 command, control how frequently LTTng samples the current values of the
7425 channel properties to evaluate user conditions with the
7426 opt:lttng-enable-channel(1):--monitor-timer option.
7428 . Run the `notif-app` application. This program accepts the
7429 <<tracing-session,tracing session>> name and the user space channel
7430 name as its two first arguments. The channel which LTTng automatically
7431 creates with the man:lttng-enable-event(1) command above is named
7437 $ ./notif-app my-session channel0
7441 . In another terminal, run an application with a very high event
7442 throughput so that the 75{nbsp}% buffer usage condition is reached.
7444 In the first terminal, the application should print lines like this:
7447 Buffer usage is 81.45197 % in tracing session "my-session", user space
7451 If you don't see anything, try modifying the condition in
7452 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7453 (step{nbsp}2) and running it again (step{nbsp}4).
7460 [[lttng-modules-ref]]
7461 === noch:{LTTng-modules}
7465 [[lttng-tracepoint-enum]]
7466 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7468 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7472 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7477 * `name` with the name of the enumeration (C identifier, unique
7478 amongst all the defined enumerations).
7479 * `entries` with a list of enumeration entries.
7481 The available enumeration entry macros are:
7483 +ctf_enum_value(__name__, __value__)+::
7484 Entry named +__name__+ mapped to the integral value +__value__+.
7486 +ctf_enum_range(__name__, __begin__, __end__)+::
7487 Entry named +__name__+ mapped to the range of integral values between
7488 +__begin__+ (included) and +__end__+ (included).
7490 +ctf_enum_auto(__name__)+::
7491 Entry named +__name__+ mapped to the integral value following the
7494 The last value of a `ctf_enum_value()` entry is its +__value__+
7497 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7499 If `ctf_enum_auto()` is the first entry in the list, its integral
7502 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7503 to use a defined enumeration as a tracepoint field.
7505 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7509 LTTNG_TRACEPOINT_ENUM(
7512 ctf_enum_auto("AUTO: EXPECT 0")
7513 ctf_enum_value("VALUE: 23", 23)
7514 ctf_enum_value("VALUE: 27", 27)
7515 ctf_enum_auto("AUTO: EXPECT 28")
7516 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7517 ctf_enum_auto("AUTO: EXPECT 304")
7525 [[lttng-modules-tp-fields]]
7526 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7528 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7529 tracepoint fields, which must be listed within `TP_FIELDS()` in
7530 `LTTNG_TRACEPOINT_EVENT()`, are:
7532 [role="func-desc growable",cols="asciidoc,asciidoc"]
7533 .Available macros to define LTTng-modules tracepoint fields
7535 |Macro |Description and parameters
7538 +ctf_integer(__t__, __n__, __e__)+
7540 +ctf_integer_nowrite(__t__, __n__, __e__)+
7542 +ctf_user_integer(__t__, __n__, __e__)+
7544 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7546 Standard integer, displayed in base{nbsp}10.
7549 Integer C type (`int`, `long`, `size_t`, ...).
7555 Argument expression.
7558 +ctf_integer_hex(__t__, __n__, __e__)+
7560 +ctf_user_integer_hex(__t__, __n__, __e__)+
7562 Standard integer, displayed in base{nbsp}16.
7571 Argument expression.
7573 |+ctf_integer_oct(__t__, __n__, __e__)+
7575 Standard integer, displayed in base{nbsp}8.
7584 Argument expression.
7587 +ctf_integer_network(__t__, __n__, __e__)+
7589 +ctf_user_integer_network(__t__, __n__, __e__)+
7591 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7600 Argument expression.
7603 +ctf_integer_network_hex(__t__, __n__, __e__)+
7605 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7607 Integer in network byte order, displayed in base{nbsp}16.
7616 Argument expression.
7619 +ctf_enum(__N__, __t__, __n__, __e__)+
7621 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7623 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7625 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7630 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7633 Integer C type (`int`, `long`, `size_t`, ...).
7639 Argument expression.
7642 +ctf_string(__n__, __e__)+
7644 +ctf_string_nowrite(__n__, __e__)+
7646 +ctf_user_string(__n__, __e__)+
7648 +ctf_user_string_nowrite(__n__, __e__)+
7650 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7656 Argument expression.
7659 +ctf_array(__t__, __n__, __e__, __s__)+
7661 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7663 +ctf_user_array(__t__, __n__, __e__, __s__)+
7665 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7667 Statically-sized array of integers.
7670 Array element C type.
7676 Argument expression.
7682 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7684 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7686 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7688 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7690 Statically-sized array of bits.
7692 The type of +__e__+ must be an integer type. +__s__+ is the number
7693 of elements of such type in +__e__+, not the number of bits.
7696 Array element C type.
7702 Argument expression.
7708 +ctf_array_text(__t__, __n__, __e__, __s__)+
7710 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7712 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7714 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7716 Statically-sized array, printed as text.
7718 The string doesn't need to be null-terminated.
7721 Array element C type (always `char`).
7727 Argument expression.
7733 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7735 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7737 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7739 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7741 Dynamically-sized array of integers.
7743 The type of +__E__+ must be unsigned.
7746 Array element C type.
7752 Argument expression.
7755 Length expression C type.
7761 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7763 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7765 Dynamically-sized array of integers, displayed in base{nbsp}16.
7767 The type of +__E__+ must be unsigned.
7770 Array element C type.
7776 Argument expression.
7779 Length expression C type.
7784 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7786 Dynamically-sized array of integers in network byte order (big-endian),
7787 displayed in base{nbsp}10.
7789 The type of +__E__+ must be unsigned.
7792 Array element C type.
7798 Argument expression.
7801 Length expression C type.
7807 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7809 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7811 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7813 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7815 Dynamically-sized array of bits.
7817 The type of +__e__+ must be an integer type. +__s__+ is the number
7818 of elements of such type in +__e__+, not the number of bits.
7820 The type of +__E__+ must be unsigned.
7823 Array element C type.
7829 Argument expression.
7832 Length expression C type.
7838 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7840 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7842 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7844 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7846 Dynamically-sized array, displayed as text.
7848 The string doesn't need to be null-terminated.
7850 The type of +__E__+ must be unsigned.
7852 The behaviour is undefined if +__e__+ is `NULL`.
7855 Sequence element C type (always `char`).
7861 Argument expression.
7864 Length expression C type.
7870 Use the `_user` versions when the argument expression, `e`, is
7871 a user space address. In the cases of `ctf_user_integer*()` and
7872 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7875 The `_nowrite` versions omit themselves from the session trace, but are
7876 otherwise identical. This means the `_nowrite` fields won't be written
7877 in the recorded trace. Their primary purpose is to make some
7878 of the event context available to the
7879 <<enabling-disabling-events,event filters>> without having to
7880 commit the data to sub-buffers.
7886 Terms related to LTTng and to tracing in general:
7889 The http://diamon.org/babeltrace[Babeltrace] project, which includes:
7892 https://babeltrace.org/docs/v2.0/man1/babeltrace2.1/[cmd:babeltrace2]
7893 command-line interface.
7894 * The libbabeltrace2 library which offers a
7895 https://babeltrace.org/docs/v2.0/libbabeltrace2/[C API].
7896 * https://babeltrace.org/docs/v2.0/python/bt2/[Python{nbsp}3 bindings].
7899 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7900 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7902 [[def-channel]]<<channel,channel>>::
7903 An entity which is responsible for a set of
7904 <<def-ring-buffer,ring buffers>>.
7906 <<def-event-rule,Event rules>> are always attached to a specific
7910 A source of time for a <<def-tracer,tracer>>.
7912 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
7913 A process which is responsible for consuming the full
7914 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7915 send them over the network.
7917 [[def-current-trace-chunk]]current trace chunk::
7918 A <<def-trace-chunk,trace chunk>> which includes the current content
7919 of all the <<def-sub-buffer,sub-buffers>> of the
7920 <<def-tracing-session-rotation,tracing session>> and the stream files
7921 produced since the latest event amongst:
7923 * The creation of the <<def-tracing-session,tracing session>>.
7924 * The last tracing session rotation, if any.
7926 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
7927 The <<def-event-record-loss-mode,event record loss mode>> in which
7928 the <<def-tracer,tracer>> _discards_ new event records when there's no
7929 <<def-sub-buffer,sub-buffer>> space left to store them.
7931 [[def-event]]event::
7932 The consequence of the execution of an
7933 <<def-instrumentation-point,instrumentation point>>, like a
7934 <<def-tracepoint,tracepoint>> that you manually place in some source
7935 code, or a Linux kernel kprobe.
7937 An event is said to _occur_ at a specific time. <<def-lttng,LTTng>> can
7938 take various actions upon the occurrence of an event, like record its
7939 payload to a <<def-sub-buffer,sub-buffer>>.
7941 [[def-event-name]]event name::
7942 The name of an <<def-event,event>>, which is also the name of the
7943 <<def-event-record,event record>>.
7945 This is also called the _instrumentation point name_.
7947 [[def-event-record]]event record::
7948 A record, in a <<def-trace,trace>>, of the payload of an
7949 <<def-event,event>> which occured.
7951 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7952 The mechanism by which event records of a given
7953 <<def-channel,channel>> are lost (not recorded) when there is no
7954 <<def-sub-buffer,sub-buffer>> space left to store them.
7956 [[def-event-rule]]<<event,event rule>>::
7957 Set of conditions which must be satisfied for one or more occuring
7958 <<def-event,events>> to be recorded.
7960 [[def-incl-set]]inclusion set::
7961 In the <<pid-tracking,process attribute tracking>> context: a
7962 set of <<def-proc-attr,process attributes>> of a given type.
7964 <<instrumenting,instrumentation>>::
7965 The use of <<def-lttng,LTTng>> probes to make a piece of software
7968 [[def-instrumentation-point]]instrumentation point::
7969 A point in the execution path of a piece of software that, when
7970 reached by this execution, can emit an <<def-event,event>>.
7972 instrumentation point name::
7973 See _<<def-event-name,event name>>_.
7975 `java.util.logging`::
7977 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities]
7978 of the Java platform.
7981 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7982 developed by the Apache Software Foundation.
7985 Level of severity of a log statement or user space
7986 <<def-instrumentation-point,instrumentation point>>.
7988 [[def-lttng]]LTTng::
7989 The _Linux Trace Toolkit: next generation_ project.
7991 <<lttng-cli,cmd:lttng>>::
7992 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
7993 project which you can use to send and receive control messages to and
7994 from a <<def-session-daemon,session daemon>>.
7997 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7998 which is a set of analyzing programs that you can use to obtain a
7999 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
8001 cmd:lttng-consumerd::
8002 The name of the <<def-consumer-daemon,consumer daemon>> program.
8005 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
8006 which can convert <<def-ring-buffer,ring buffer>> files (usually
8007 <<persistent-memory-file-systems,saved on a persistent memory file
8008 system>>) to <<def-trace,trace>> files.
8010 See man:lttng-crash(1).
8012 LTTng Documentation::
8015 <<lttng-live,LTTng live>>::
8016 A communication protocol between the <<lttng-relayd,relay daemon>> and
8017 live viewers which makes it possible to see <<def-event-record,event
8018 records>> ``live'', as they are received by the
8019 <<def-relay-daemon,relay daemon>>.
8021 <<lttng-modules,LTTng-modules>>::
8022 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
8023 which contains the Linux kernel modules to make the Linux kernel
8024 <<def-instrumentation-point,instrumentation points>> available for
8025 <<def-lttng,LTTng>> tracing.
8028 The name of the <<def-relay-daemon,relay daemon>> program.
8030 cmd:lttng-sessiond::
8031 The name of the <<def-session-daemon,session daemon>> program.
8033 [[def-lttng-tools]]LTTng-tools::
8034 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
8035 contains the various programs and libraries used to
8036 <<controlling-tracing,control tracing>>.
8038 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
8039 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
8040 contains libraries to instrument
8041 <<def-user-application,user applications>>.
8043 <<lttng-ust-agents,LTTng-UST Java agent>>::
8044 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
8045 allow the LTTng instrumentation of `java.util.logging` and Apache
8046 log4j{nbsp}1.2 logging statements.
8048 <<lttng-ust-agents,LTTng-UST Python agent>>::
8049 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
8050 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
8053 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
8054 The <<def-event-record-loss-mode,event record loss mode>> in which new
8055 <<def-event-record,event records>> _overwrite_ older event records
8056 when there's no <<def-sub-buffer,sub-buffer>> space left to store
8059 <<channel-buffering-schemes,per-process buffering>>::
8060 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
8061 process has its own <<def-sub-buffer,sub-buffers>> for a given user
8062 space <<def-channel,channel>>.
8064 <<channel-buffering-schemes,per-user buffering>>::
8065 A <<def-buffering-scheme,buffering scheme>> in which all the processes
8066 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
8067 given user space <<def-channel,channel>>.
8069 [[def-proc-attr]]process attribute::
8070 In the <<pid-tracking,process attribute tracking>> context:
8073 * A virtual process ID.
8075 * A virtual Unix user ID.
8077 * A virtual Unix group ID.
8079 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
8080 A process which is responsible for receiving the <<def-trace,trace>>
8081 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
8083 [[def-ring-buffer]]ring buffer::
8084 A set of <<def-sub-buffer,sub-buffers>>.
8087 See _<<def-tracing-session-rotation,tracing session rotation>>_.
8089 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
8090 A process which receives control commands from you and orchestrates
8091 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
8093 <<taking-a-snapshot,snapshot>>::
8094 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
8095 of a given <<def-tracing-session,tracing session>>, saved as
8096 <<def-trace,trace>> files.
8098 [[def-sub-buffer]]sub-buffer::
8099 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
8100 which contains <<def-event-record,event records>>.
8103 The time information attached to an <<def-event,event>> when it is
8106 [[def-trace]]trace (_noun_)::
8109 * One http://diamon.org/ctf/[CTF] metadata stream file.
8110 * One or more CTF data stream files which are the concatenations of one
8111 or more flushed <<def-sub-buffer,sub-buffers>>.
8113 [[def-trace-verb]]trace (_verb_)::
8114 The action of recording the <<def-event,events>> emitted by an
8115 application or by a system, or to initiate such recording by
8116 controlling a <<def-tracer,tracer>>.
8118 [[def-trace-chunk]]trace chunk::
8119 A self-contained <<def-trace,trace>> which is part of a
8120 <<def-tracing-session,tracing session>>. Each
8121 <<def-tracing-session-rotation, tracing session rotation>> produces a
8122 <<def-trace-chunk-archive,trace chunk archive>>.
8124 [[def-trace-chunk-archive]]trace chunk archive::
8125 The result of a <<def-tracing-session-rotation, tracing session rotation>>.
8127 <<def-lttng,LTTng>> doesn't manage any trace chunk archive, even if its
8128 containing <<def-tracing-session,tracing session>> is still active: you
8129 are free to read it, modify it, move it, or remove it.
8132 The http://tracecompass.org[Trace Compass] project and application.
8134 [[def-tracepoint]]tracepoint::
8135 An instrumentation point using the tracepoint mechanism of the Linux
8136 kernel or of <<def-lttng-ust,LTTng-UST>>.
8138 tracepoint definition::
8139 The definition of a single <<def-tracepoint,tracepoint>>.
8142 The name of a <<def-tracepoint,tracepoint>>.
8144 [[def-tracepoint-provider]]tracepoint provider::
8145 A set of functions providing <<def-tracepoint,tracepoints>> to an
8146 instrumented <<def-user-application,user application>>.
8148 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
8149 provider package>>: many tracepoint providers can exist within a
8150 tracepoint provider package.
8152 [[def-tracepoint-provider-package]]tracepoint provider package::
8153 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
8154 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
8155 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
8158 [[def-tracer]]tracer::
8159 A software which records emitted <<def-event,events>>.
8161 <<domain,tracing domain>>::
8162 A namespace for <<def-event,event>> sources.
8164 <<tracing-group,tracing group>>::
8165 The Unix group in which a Unix user can be to be allowed to
8166 <<def-trace-verb,trace>> the Linux kernel.
8168 [[def-tracing-session]]<<tracing-session,tracing session>>::
8169 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
8171 [[def-tracing-session-rotation]]<<session-rotation,tracing session rotation>>::
8172 The action of archiving the
8173 <<def-current-trace-chunk,current trace chunk>> of a
8174 <<def-tracing-session,tracing session>>.
8176 tracked <<def-proc-attr,process attribute>>::
8177 A process attribute which is part of an <<def-incl-set,inclusion
8180 untracked process attribute::
8181 A process attribute which isn't part of an <<def-incl-set,inclusion
8184 [[def-user-application]]user application::
8185 An application running in user space, as opposed to a Linux kernel
8186 module, for example.