1 The LTTng Documentation
2 =======================
3 Philippe Proulx <pproulx@efficios.com>
4 v2.11, 25 February 2021
7 include::../common/copyright.txt[]
10 include::../common/welcome.txt[]
13 include::../common/audience.txt[]
17 === What's in this documentation?
19 The LTTng Documentation is divided into the following sections:
21 * **<<nuts-and-bolts,Nuts and bolts>>** explains the
22 rudiments of software tracing and the rationale behind the
25 You can skip this section if you’re familiar with software tracing and
26 with the LTTng project.
28 * **<<installing-lttng,Installation>>** describes the steps to
29 install the LTTng packages on common Linux distributions and from
32 You can skip this section if you already properly installed LTTng on
35 * **<<getting-started,Quick start>>** is a concise guide to
36 getting started quickly with LTTng kernel and user space tracing.
38 We recommend this section if you're new to LTTng or to software tracing
41 You can skip this section if you're not new to LTTng.
43 * **<<core-concepts,Core concepts>>** explains the concepts at
46 It's a good idea to become familiar with the core concepts
47 before attempting to use the toolkit.
49 * **<<plumbing,Components of LTTng>>** describes the various components
50 of the LTTng machinery, like the daemons, the libraries, and the
51 command-line interface.
52 * **<<instrumenting,Instrumentation>>** shows different ways to
53 instrument user applications and the Linux kernel.
55 Instrumenting source code is essential to provide a meaningful
58 You can skip this section if you do not have a programming background.
60 * **<<controlling-tracing,Tracing control>>** is divided into topics
61 which demonstrate how to use the vast array of features that
62 LTTng{nbsp}{revision} offers.
63 * **<<reference,Reference>>** contains reference tables.
64 * **<<glossary,Glossary>>** is a specialized dictionary of terms related
65 to LTTng or to the field of software tracing.
68 include::../common/convention.txt[]
71 include::../common/acknowledgements.txt[]
75 == What's new in LTTng{nbsp}{revision}?
77 LTTng{nbsp}{revision} bears the name _Lafontaine_. This modern
78 https://en.wikipedia.org/wiki/saison[saison] from the
79 https://oshlag.com/[Oshlag] microbrewery is a refreshing--zesty--rice
80 beer with hints of fruit and spices. Some even say it makes for a great
81 https://en.wikipedia.org/wiki/Somaek[Somaek] when mixed with
82 Chamisul Soju, not that we've tried!
84 New features and changes in LTTng{nbsp}{revision}:
86 * Just like you can typically perform
87 https://en.wikipedia.org/wiki/Log_rotation[log rotation], you can
88 now <<session-rotation,_rotate_ a tracing session>>, that
89 is, according to man:lttng-rotate(1), archive the current trace
90 chunk (all the tracing session's trace data since the last rotation
91 or since its inception) so that LTTng does not manage it anymore.
93 Once LTTng archives a trace chunk, you are free to read it, modify it,
94 move it, or remove it.
96 You can rotate a tracing session immediately or set a rotation schedule
97 to automate rotations.
99 * When you <<enabling-disabling-events,create an event rule>>, the
100 filter expression syntax now supports the following new operators:
104 ** `<<` (bitwise left shift)
105 ** `>>` (bitwise right shift)
111 The syntax also supports array indexing with the usual square brackets:
114 regs[3][1] & 0xff7 == 0x240
117 There are peculiarities for both the new operators and the array
118 indexing brackets, like a custom precedence table and implicit casting.
119 See man:lttng-enable-event(1) to get all the details about the filter
122 * You can now dynamically instrument any application's or library's
123 function entry by symbol name thanks to the new
124 opt:lttng-enable-event(1):--userspace-probe option of
125 the `lttng enable-event` command:
129 $ lttng enable-event --kernel \
130 --userspace-probe=/usr/lib/libc.so.6:malloc libc_malloc
133 The option also supports tracing existing
134 https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
135 Statically Defined Tracing] (USDT) probe (DTrace-style marker). For
136 example, given the following probe:
140 DTRACE_PROBE2("server", "accept-request", request_id, ip_addr);
143 You can trace this probe with:
146 $ lttng enable-event --kernel \
147 --userspace-probe=sdt:/path/to/server:server:accept-request \
148 server_accept_request
151 This feature makes use of Linux's
152 https://www.kernel.org/doc/Documentation/trace/uprobetracer.txt[uprobe]
153 mechanism, therefore you must use the `--userspace-probe`
154 instrumentation option with the opt:lttng-enable-event(1):--kernel
157 NOTE: As of LTTng{nbsp}{revision}, LTTng does not record function
158 parameters with the opt:lttng-enable-event(1):--userspace-probe option.
160 * Two new <<adding-context,context>> fields are available for Linux
161 kernel <<channel,channels>>:
164 ** `callstack-kernel`
168 Thanks to those, you can record the Linux kernel and user call stacks
169 when a kernel event occurs. For example:
173 $ lttng enable-event --kernel --syscall open
174 $ lttng add-context --kernel --type=callstack-kernel --type=callstack-user
177 When an man:open(2) system call occurs, LTTng attaches the kernel and
178 user call stacks to the recorded event.
180 NOTE: LTTng cannot always sample the user space call stack reliably.
181 For instance, LTTng cannot sample the call stack of user applications
182 and libraries compiled with the
183 https://gcc.gnu.org/onlinedocs/gcc/Optimize-Options.html[`-fomit-frame-pointer`]
184 option. In such a case, the tracing is not affected, but the sampled
185 user space call stack may only contain the user call stack's topmost
188 * User applications and libraries instrumented with
189 <<lttng-ust,LTTng-UST>> can now safely unload (man:dlclose(3)) a
191 <<building-tracepoint-providers-and-user-application,tracepoint
194 * The <<lttng-relayd,relay daemon>> is more efficient and presents fewer
195 connectivity issues, especially when a large number of targets send
196 trace data to a given relay daemon.
198 * LTTng-UST uses https://github.com/numactl/numactl[libnuma]
199 when available to allocate <<def-sub-buffer,sub-buffers>>, making them
201 https://en.wikipedia.org/wiki/Non-uniform_memory_access[NUMA] node.
203 This change makes the tracer more efficient on NUMA systems.
205 * The <<proc-lttng-logger-abi,LTTng logger>> kernel module now also
206 creates a ``misc'' device named `lttng-logger`, which udev will
207 make accessible as the path:{/dev/lttng-logger} special file.
209 The `lttng-logger` device shares the `/proc/lttng-logger` file's ABI,
210 but it works from within containers when the path is made accessible to
217 What is LTTng? As its name suggests, the _Linux Trace Toolkit: next
218 generation_ is a modern toolkit for tracing Linux systems and
219 applications. So your first question might be:
226 As the history of software engineering progressed and led to what
227 we now take for granted--complex, numerous and
228 interdependent software applications running in parallel on
229 sophisticated operating systems like Linux--the authors of such
230 components, software developers, began feeling a natural
231 urge to have tools that would ensure the robustness and good performance
232 of their masterpieces.
234 One major achievement in this field is, inarguably, the
235 https://www.gnu.org/software/gdb/[GNU debugger (GDB)],
236 an essential tool for developers to find and fix bugs. But even the best
237 debugger won't help make your software run faster, and nowadays, faster
238 software means either more work done by the same hardware, or cheaper
239 hardware for the same work.
241 A _profiler_ is often the tool of choice to identify performance
242 bottlenecks. Profiling is suitable to identify _where_ performance is
243 lost in a given software. The profiler outputs a profile, a statistical
244 summary of observed events, which you may use to discover which
245 functions took the most time to execute. However, a profiler won't
246 report _why_ some identified functions are the bottleneck. Bottlenecks
247 might only occur when specific conditions are met, conditions that are
248 sometimes impossible to capture by a statistical profiler, or impossible
249 to reproduce with an application altered by the overhead of an
250 event-based profiler. For a thorough investigation of software
251 performance issues, a history of execution is essential, with the
252 recorded values of variables and context fields you choose, and
253 with as little influence as possible on the instrumented software. This
254 is where tracing comes in handy.
256 _Tracing_ is a technique used to understand what goes on in a running
257 software system. The software used for tracing is called a _tracer_,
258 which is conceptually similar to a tape recorder. When recording,
259 specific instrumentation points placed in the software source code
260 generate events that are saved on a giant tape: a _trace_ file. You
261 can trace user applications and the operating system at the same time,
262 opening the possibility of resolving a wide range of problems that would
263 otherwise be extremely challenging.
265 Tracing is often compared to _logging_. However, tracers and loggers are
266 two different tools, serving two different purposes. Tracers are
267 designed to record much lower-level events that occur much more
268 frequently than log messages, often in the range of thousands per
269 second, with very little execution overhead. Logging is more appropriate
270 for a very high-level analysis of less frequent events: user accesses,
271 exceptional conditions (errors and warnings, for example), database
272 transactions, instant messaging communications, and such. Simply put,
273 logging is one of the many use cases that can be satisfied with tracing.
275 The list of recorded events inside a trace file can be read manually
276 like a log file for the maximum level of detail, but it is generally
277 much more interesting to perform application-specific analyses to
278 produce reduced statistics and graphs that are useful to resolve a
279 given problem. Trace viewers and analyzers are specialized tools
282 In the end, this is what LTTng is: a powerful, open source set of
283 tools to trace the Linux kernel and user applications at the same time.
284 LTTng is composed of several components actively maintained and
285 developed by its link:/community/#where[community].
288 [[lttng-alternatives]]
289 === Alternatives to noch:{LTTng}
291 Excluding proprietary solutions, a few competing software tracers
294 * https://github.com/dtrace4linux/linux[dtrace4linux] is a port of
295 Sun Microsystems's DTrace to Linux. The cmd:dtrace tool interprets
296 user scripts and is responsible for loading code into the
297 Linux kernel for further execution and collecting the outputted data.
298 * https://en.wikipedia.org/wiki/Berkeley_Packet_Filter[eBPF] is a
299 subsystem in the Linux kernel in which a virtual machine can execute
300 programs passed from the user space to the kernel. You can attach
301 such programs to tracepoints and kprobes thanks to a system call, and
302 they can output data to the user space when executed thanks to
303 different mechanisms (pipe, VM register values, and eBPF maps, to name
305 * https://www.kernel.org/doc/Documentation/trace/ftrace.txt[ftrace]
306 is the de facto function tracer of the Linux kernel. Its user
307 interface is a set of special files in sysfs.
308 * https://perf.wiki.kernel.org/[perf] is
309 a performance analysis tool for Linux which supports hardware
310 performance counters, tracepoints, as well as other counters and
311 types of probes. perf's controlling utility is the cmd:perf command
313 * http://linux.die.net/man/1/strace[strace]
314 is a command-line utility which records system calls made by a
315 user process, as well as signal deliveries and changes of process
316 state. strace makes use of https://en.wikipedia.org/wiki/Ptrace[ptrace]
317 to fulfill its function.
318 * http://www.sysdig.org/[sysdig], like SystemTap, uses scripts to
319 analyze Linux kernel events. You write scripts, or _chisels_ in
320 sysdig's jargon, in Lua and sysdig executes them while it traces the
321 system or afterwards. sysdig's interface is the cmd:sysdig
322 command-line tool as well as the curses-based cmd:csysdig tool.
323 * https://sourceware.org/systemtap/[SystemTap] is a Linux kernel and
324 user space tracer which uses custom user scripts to produce plain text
325 traces. SystemTap converts the scripts to the C language, and then
326 compiles them as Linux kernel modules which are loaded to produce
327 trace data. SystemTap's primary user interface is the cmd:stap
330 The main distinctive features of LTTng is that it produces correlated
331 kernel and user space traces, as well as doing so with the lowest
332 overhead amongst other solutions. It produces trace files in the
333 http://diamon.org/ctf[CTF] format, a file format optimized
334 for the production and analyses of multi-gigabyte data.
336 LTTng is the result of more than 10{nbsp}years of active open source
337 development by a community of passionate developers.
338 LTTng{nbsp}{revision} is currently available on major desktop and server
341 The main interface for tracing control is a single command-line tool
342 named cmd:lttng. The latter can create several tracing sessions, enable
343 and disable events on the fly, filter events efficiently with custom
344 user expressions, start and stop tracing, and much more. LTTng can
345 record the traces on the file system or send them over the network, and
346 keep them totally or partially. You can view the traces once tracing
347 becomes inactive or in real-time.
349 <<installing-lttng,Install LTTng now>> and
350 <<getting-started,start tracing>>!
356 **LTTng** is a set of software <<plumbing,components>> which interact to
357 <<instrumenting,instrument>> the Linux kernel and user applications, and
358 to <<controlling-tracing,control tracing>> (start and stop
359 tracing, enable and disable event rules, and the rest). Those
360 components are bundled into the following packages:
363 Libraries and command-line interface to control tracing.
366 Linux kernel modules to instrument and trace the kernel.
369 Libraries and Java/Python packages to instrument and trace user
372 Most distributions mark the LTTng-modules and LTTng-UST packages as
373 optional when installing LTTng-tools (which is always required). In the
374 following sections, we always provide the steps to install all three,
377 * You only need to install LTTng-modules if you intend to trace the
379 * You only need to install LTTng-UST if you intend to trace user
383 .Availability of LTTng{nbsp}{revision} for major Linux distributions as of 5{nbsp}August{nbsp}2020.
385 |Distribution |Available in releases
387 |https://www.ubuntu.com/[Ubuntu]
388 |<<ubuntu,Ubuntu{nbsp}20.04 _Focal Fossa_>>.
390 Ubuntu{nbsp}16.04 _Xenial Xerus_ and Ubuntu{nbsp}18.04 _Bionic Beaver_:
391 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
393 |https://getfedora.org/[Fedora]
394 |<<fedora,Fedora{nbsp}32>>.
396 |https://www.redhat.com/[RHEL] and https://www.suse.com/[SLES]
397 |See http://packages.efficios.com/[EfficiOS Enterprise Packages].
399 |https://buildroot.org/[Buildroot]
400 |xref:buildroot[Buildroot{nbsp}2019.11, Buildroot{nbsp}2020.02, and
401 Buildroot{nbsp}2020.05].
403 |https://www.openembedded.org/wiki/Main_Page[OpenEmbedded] and
404 https://www.yoctoproject.org/[Yocto]
405 |<<oe-yocto,Yocto Project{nbsp}3.1 _Dunfell_>>
406 (`openembedded-core` layer).
411 === [[ubuntu-official-repositories]]Ubuntu
413 LTTng{nbsp}{revision} is available on Ubuntu{nbsp}20.04 _Focal Fossa_.
414 For previous supported releases of Ubuntu,
415 <<ubuntu-ppa,use the LTTng Stable{nbsp}{revision} PPA>>.
417 To install LTTng{nbsp}{revision} on Ubuntu{nbsp}20.04 _Focal Fossa_:
419 . Install the main LTTng{nbsp}{revision} packages:
424 # apt-get install lttng-tools
425 # apt-get install lttng-modules-dkms
426 # apt-get install liblttng-ust-dev
430 . **If you need to instrument and trace
431 <<java-application,Java applications>>**, install the LTTng-UST
437 # apt-get install liblttng-ust-agent-java
441 . **If you need to instrument and trace
442 <<python-application,Python{nbsp}3 applications>>**, install the
443 LTTng-UST Python agent:
448 # apt-get install python3-lttngust
453 === Ubuntu: noch:{LTTng} Stable {revision} PPA
455 The https://launchpad.net/~lttng/+archive/ubuntu/stable-{revision}[LTTng
456 Stable{nbsp}{revision} PPA] offers the latest stable
457 LTTng{nbsp}{revision} packages for Ubuntu{nbsp}16.04 _Xenial Xerus_ and
458 Ubuntu{nbsp}18.04 _Bionic Beaver_.
460 To install LTTng{nbsp}{revision} from the LTTng Stable{nbsp}{revision}
463 . Add the LTTng Stable{nbsp}{revision} PPA repository and update the
469 # apt-add-repository ppa:lttng/stable-2.11
474 . Install the main LTTng{nbsp}{revision} packages:
479 # apt-get install lttng-tools
480 # apt-get install lttng-modules-dkms
481 # apt-get install liblttng-ust-dev
485 . **If you need to instrument and trace
486 <<java-application,Java applications>>**, install the LTTng-UST
492 # apt-get install liblttng-ust-agent-java
496 . **If you need to instrument and trace
497 <<python-application,Python{nbsp}3 applications>>**, install the
498 LTTng-UST Python agent:
503 # apt-get install python3-lttngust
511 To install LTTng{nbsp}{revision} on Fedora{nbsp}32:
513 . Install the LTTng-tools{nbsp}{revision} and LTTng-UST{nbsp}{revision}
519 # yum install lttng-tools
520 # yum install lttng-ust
524 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
530 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
531 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
532 cd lttng-modules-2.11.* &&
534 sudo make modules_install &&
540 .Java and Python application instrumentation and tracing
542 If you need to instrument and trace <<java-application,Java
543 applications>> on Fedora, you need to build and install
544 LTTng-UST{nbsp}{revision} <<building-from-source,from source>> and pass
545 the `--enable-java-agent-jul`, `--enable-java-agent-log4j`, or
546 `--enable-java-agent-all` options to the `configure` script, depending
547 on which Java logging framework you use.
549 If you need to instrument and trace <<python-application,Python
550 applications>> on Fedora, you need to build and install
551 LTTng-UST{nbsp}{revision} from source and pass the
552 `--enable-python-agent` option to the `configure` script.
559 To install LTTng{nbsp}{revision} on Buildroot{nbsp}2019.11,
560 Buildroot{nbsp}2020.02, or Buildroot{nbsp}2020.05:
562 . Launch the Buildroot configuration tool:
571 . In **Kernel**, check **Linux kernel**.
572 . In **Toolchain**, check **Enable WCHAR support**.
573 . In **Target packages**{nbsp}→ **Debugging, profiling and benchmark**,
574 check **lttng-modules** and **lttng-tools**.
575 . In **Target packages**{nbsp}→ **Libraries**{nbsp}→
576 **Other**, check **lttng-libust**.
580 === OpenEmbedded and Yocto
582 LTTng{nbsp}{revision} recipes are available in the
583 https://layers.openembedded.org/layerindex/branch/master/layer/openembedded-core/[`openembedded-core`]
584 layer for Yocto Project{nbsp}3.1 _Dunfell_ under the following names:
590 With BitBake, the simplest way to include LTTng recipes in your target
591 image is to add them to `IMAGE_INSTALL_append` in path:{conf/local.conf}:
594 IMAGE_INSTALL_append = " lttng-tools lttng-modules lttng-ust"
599 . Select a machine and an image recipe.
600 . Click **Edit image recipe**.
601 . Under the **All recipes** tab, search for **lttng**.
602 . Check the desired LTTng recipes.
605 [[building-from-source]]
606 === Build from source
608 To build and install LTTng{nbsp}{revision} from source:
610 . Using your distribution's package manager, or from source, install
611 the following dependencies of LTTng-tools and LTTng-UST:
614 * https://sourceforge.net/projects/libuuid/[libuuid]
615 * http://directory.fsf.org/wiki/Popt[popt]
616 * http://liburcu.org/[Userspace RCU]
617 * http://www.xmlsoft.org/[libxml2]
618 * **Optional**: https://github.com/numactl/numactl[numactl]
621 . Download, build, and install the latest LTTng-modules{nbsp}{revision}:
627 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
628 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
629 cd lttng-modules-2.11.* &&
631 sudo make modules_install &&
636 . Download, build, and install the latest LTTng-UST{nbsp}{revision}:
642 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
643 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
644 cd lttng-ust-2.11.* &&
652 Add `--disable-numa` to `./configure` if you don't have
653 https://github.com/numactl/numactl[numactl].
657 .Java and Python application tracing
659 If you need to instrument and trace <<java-application,Java
660 applications>>, pass the `--enable-java-agent-jul`,
661 `--enable-java-agent-log4j`, or `--enable-java-agent-all` options to the
662 `configure` script, depending on which Java logging framework you use.
664 If you need to instrument and trace <<python-application,Python
665 applications>>, pass the `--enable-python-agent` option to the
666 `configure` script. You can set the `PYTHON` environment variable to the
667 path to the Python interpreter for which to install the LTTng-UST Python
675 By default, LTTng-UST libraries are installed to
676 dir:{/usr/local/lib}, which is the de facto directory in which to
677 keep self-compiled and third-party libraries.
679 When <<building-tracepoint-providers-and-user-application,linking an
680 instrumented user application with `liblttng-ust`>>:
682 * Append `/usr/local/lib` to the env:LD_LIBRARY_PATH environment
684 * Pass the `-L/usr/local/lib` and `-Wl,-rpath,/usr/local/lib` options to
685 man:gcc(1), man:g++(1), or man:clang(1).
689 . Download, build, and install the latest LTTng-tools{nbsp}{revision}:
695 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
696 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
697 cd lttng-tools-2.11.* &&
705 TIP: The https://github.com/eepp/vlttng[vlttng tool] can do all the
706 previous steps automatically for a given version of LTTng and confine
707 the installed files in a specific directory. This can be useful to test
708 LTTng without installing it on your system.
714 This is a short guide to get started quickly with LTTng kernel and user
717 Before you follow this guide, make sure to <<installing-lttng,install>>
720 This tutorial walks you through the steps to:
722 . <<tracing-the-linux-kernel,Trace the Linux kernel>>.
723 . <<tracing-your-own-user-application,Trace a user application>> written
725 . <<viewing-and-analyzing-your-traces,View and analyze the
729 [[tracing-the-linux-kernel]]
730 === Trace the Linux kernel
732 The following command lines start with the `#` prompt because you need
733 root privileges to trace the Linux kernel. You can also trace the kernel
734 as a regular user if your Unix user is a member of the
735 <<tracing-group,tracing group>>.
737 . Create a <<tracing-session,tracing session>> which writes its traces
738 to dir:{/tmp/my-kernel-trace}:
743 # lttng create my-kernel-session --output=/tmp/my-kernel-trace
747 . List the available kernel tracepoints and system calls:
752 # lttng list --kernel
753 # lttng list --kernel --syscall
757 . Create <<event,event rules>> which match the desired instrumentation
758 point names, for example the `sched_switch` and `sched_process_fork`
759 tracepoints, and the man:open(2) and man:close(2) system calls:
764 # lttng enable-event --kernel sched_switch,sched_process_fork
765 # lttng enable-event --kernel --syscall open,close
769 You can also create an event rule which matches _all_ the Linux kernel
770 tracepoints (this will generate a lot of data when tracing):
775 # lttng enable-event --kernel --all
779 . <<basic-tracing-session-control,Start tracing>>:
788 . Do some operation on your system for a few seconds. For example,
789 load a website, or list the files of a directory.
790 . <<creating-destroying-tracing-sessions,Destroy>> the current
800 The man:lttng-destroy(1) command does not destroy the trace data; it
801 only destroys the state of the tracing session.
803 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
804 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
805 session>>). You need to stop tracing to make LTTng flush the remaining
806 trace data and make the trace readable.
808 . For the sake of this example, make the recorded trace accessible to
814 # chown -R $(whoami) /tmp/my-kernel-trace
818 See <<viewing-and-analyzing-your-traces,View and analyze the
819 recorded events>> to view the recorded events.
822 [[tracing-your-own-user-application]]
823 === Trace a user application
825 This section steps you through a simple example to trace a
826 _Hello world_ program written in C.
828 To create the traceable user application:
830 . Create the tracepoint provider header file, which defines the
831 tracepoints and the events they can generate:
837 #undef TRACEPOINT_PROVIDER
838 #define TRACEPOINT_PROVIDER hello_world
840 #undef TRACEPOINT_INCLUDE
841 #define TRACEPOINT_INCLUDE "./hello-tp.h"
843 #if !defined(_HELLO_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
846 #include <lttng/tracepoint.h>
856 ctf_string(my_string_field, my_string_arg)
857 ctf_integer(int, my_integer_field, my_integer_arg)
861 #endif /* _HELLO_TP_H */
863 #include <lttng/tracepoint-event.h>
867 . Create the tracepoint provider package source file:
873 #define TRACEPOINT_CREATE_PROBES
874 #define TRACEPOINT_DEFINE
876 #include "hello-tp.h"
880 . Build the tracepoint provider package:
885 $ gcc -c -I. hello-tp.c
889 . Create the _Hello World_ application source file:
896 #include "hello-tp.h"
898 int main(int argc, char *argv[])
902 puts("Hello, World!\nPress Enter to continue...");
905 * The following getchar() call is only placed here for the purpose
906 * of this demonstration, to pause the application in order for
907 * you to have time to list its tracepoints. It is not
913 * A tracepoint() call.
915 * Arguments, as defined in hello-tp.h:
917 * 1. Tracepoint provider name (required)
918 * 2. Tracepoint name (required)
919 * 3. my_integer_arg (first user-defined argument)
920 * 4. my_string_arg (second user-defined argument)
922 * Notice the tracepoint provider and tracepoint names are
923 * NOT strings: they are in fact parts of variables that the
924 * macros in hello-tp.h create.
926 tracepoint(hello_world, my_first_tracepoint, 23, "hi there!");
928 for (x = 0; x < argc; ++x) {
929 tracepoint(hello_world, my_first_tracepoint, x, argv[x]);
932 puts("Quitting now!");
933 tracepoint(hello_world, my_first_tracepoint, x * x, "x^2");
940 . Build the application:
949 . Link the application with the tracepoint provider package,
950 `liblttng-ust`, and `libdl`:
955 $ gcc -o hello hello.o hello-tp.o -llttng-ust -ldl
959 Here's the whole build process:
962 .User space tracing tutorial's build steps.
963 image::ust-flow.png[]
965 To trace the user application:
967 . Run the application with a few arguments:
972 $ ./hello world and beyond
981 Press Enter to continue...
985 . Start an LTTng <<lttng-sessiond,session daemon>>:
990 $ lttng-sessiond --daemonize
994 Note that a session daemon might already be running, for example as
995 a service that the distribution's service manager started.
997 . List the available user space tracepoints:
1002 $ lttng list --userspace
1006 You see the `hello_world:my_first_tracepoint` tracepoint listed
1007 under the `./hello` process.
1009 . Create a <<tracing-session,tracing session>>:
1014 $ lttng create my-user-space-session
1018 . Create an <<event,event rule>> which matches the
1019 `hello_world:my_first_tracepoint` event name:
1024 $ lttng enable-event --userspace hello_world:my_first_tracepoint
1028 . <<basic-tracing-session-control,Start tracing>>:
1037 . Go back to the running `hello` application and press Enter. The
1038 program executes all `tracepoint()` instrumentation points and exits.
1039 . <<creating-destroying-tracing-sessions,Destroy>> the current
1049 The man:lttng-destroy(1) command does not destroy the trace data; it
1050 only destroys the state of the tracing session.
1052 The man:lttng-destroy(1) command also runs the man:lttng-stop(1) command
1053 implicitly (see <<basic-tracing-session-control,Start and stop a tracing
1054 session>>). You need to stop tracing to make LTTng flush the remaining
1055 trace data and make the trace readable.
1057 By default, LTTng saves the traces in
1058 +$LTTNG_HOME/lttng-traces/__name__-__date__-__time__+,
1059 where +__name__+ is the tracing session name. The
1060 env:LTTNG_HOME environment variable defaults to `$HOME` if not set.
1062 See <<viewing-and-analyzing-your-traces,View and analyze the
1063 recorded events>> to view the recorded events.
1066 [[viewing-and-analyzing-your-traces]]
1067 === View and analyze the recorded events
1069 Once you have completed the <<tracing-the-linux-kernel,Trace the Linux
1070 kernel>> and <<tracing-your-own-user-application,Trace a user
1071 application>> tutorials, you can inspect the recorded events.
1073 There are many tools you can use to read LTTng traces:
1075 * **cmd:babeltrace** is a command-line utility which converts trace
1076 formats; it supports the format that LTTng produces, CTF, as well as a
1077 basic text output which can be ++grep++ed. The cmd:babeltrace command
1078 is part of the http://diamon.org/babeltrace[Babeltrace] project.
1079 * Babeltrace also includes
1080 **https://www.python.org/[Python{nbsp}3] bindings** so
1081 that you can easily open and read an LTTng trace with your own script,
1082 benefiting from the power of Python.
1083 * http://tracecompass.org/[**Trace Compass**]
1084 is a graphical user interface for viewing and analyzing any type of
1085 logs or traces, including LTTng's.
1086 * https://github.com/lttng/lttng-analyses[**LTTng analyses**] is a
1087 project which includes many high-level analyses of LTTng kernel
1088 traces, like scheduling statistics, interrupt frequency distribution,
1089 top CPU usage, and more.
1091 NOTE: This section assumes that LTTng saved the traces it recorded
1092 during the previous tutorials to their default location, in the
1093 dir:{$LTTNG_HOME/lttng-traces} directory. The env:LTTNG_HOME
1094 environment variable defaults to `$HOME` if not set.
1097 [[viewing-and-analyzing-your-traces-bt]]
1098 ==== Use the cmd:babeltrace command-line tool
1100 The simplest way to list all the recorded events of a trace is to pass
1101 its path to cmd:babeltrace with no options:
1105 $ babeltrace ~/lttng-traces/my-user-space-session*
1108 cmd:babeltrace finds all traces recursively within the given path and
1109 prints all their events, merging them in chronological order.
1111 You can pipe the output of cmd:babeltrace into a tool like man:grep(1) for
1116 $ babeltrace /tmp/my-kernel-trace | grep _switch
1119 You can pipe the output of cmd:babeltrace into a tool like man:wc(1) to
1120 count the recorded events:
1124 $ babeltrace /tmp/my-kernel-trace | grep _open | wc --lines
1128 [[viewing-and-analyzing-your-traces-bt-python]]
1129 ==== Use the Babeltrace{nbsp}1 Python bindings
1131 The <<viewing-and-analyzing-your-traces-bt,text output of cmd:babeltrace>>
1132 is useful to isolate events by simple matching using man:grep(1) and
1133 similar utilities. However, more elaborate filters, such as keeping only
1134 event records with a field value falling within a specific range, are
1135 not trivial to write using a shell. Moreover, reductions and even the
1136 most basic computations involving multiple event records are virtually
1137 impossible to implement.
1139 Fortunately, Babeltrace{nbsp}1 ships with Python{nbsp}3 bindings which
1140 makes it easy to read the event records of an LTTng trace sequentially
1141 and compute the desired information.
1143 The following script accepts an LTTng Linux kernel trace path as its
1144 first argument and prints the short names of the top five running
1145 processes on CPU{nbsp}0 during the whole trace:
1150 from collections import Counter
1156 if len(sys.argv) != 2:
1157 msg = 'Usage: python3 {} TRACEPATH'.format(sys.argv[0])
1158 print(msg, file=sys.stderr)
1161 # A trace collection contains one or more traces
1162 col = babeltrace.TraceCollection()
1164 # Add the trace provided by the user (LTTng traces always have
1166 if col.add_trace(sys.argv[1], 'ctf') is None:
1167 raise RuntimeError('Cannot add trace')
1169 # This counter dict contains execution times:
1171 # task command name -> total execution time (ns)
1172 exec_times = Counter()
1174 # This contains the last `sched_switch` timestamp
1178 for event in col.events:
1179 # Keep only `sched_switch` events
1180 if event.name != 'sched_switch':
1183 # Keep only events which happened on CPU 0
1184 if event['cpu_id'] != 0:
1188 cur_ts = event.timestamp
1194 # Previous task command (short) name
1195 prev_comm = event['prev_comm']
1197 # Initialize entry in our dict if not yet done
1198 if prev_comm not in exec_times:
1199 exec_times[prev_comm] = 0
1201 # Compute previous command execution time
1202 diff = cur_ts - last_ts
1204 # Update execution time of this command
1205 exec_times[prev_comm] += diff
1207 # Update last timestamp
1211 for name, ns in exec_times.most_common(5):
1213 print('{:20}{} s'.format(name, s))
1218 if __name__ == '__main__':
1219 sys.exit(0 if top5proc() else 1)
1226 $ python3 top5proc.py /tmp/my-kernel-trace/kernel
1232 swapper/0 48.607245889 s
1233 chromium 7.192738188 s
1234 pavucontrol 0.709894415 s
1235 Compositor 0.660867933 s
1236 Xorg.bin 0.616753786 s
1239 Note that `swapper/0` is the "idle" process of CPU{nbsp}0 on Linux;
1240 since we weren't using the CPU that much when tracing, its first
1241 position in the list makes sense.
1245 == [[understanding-lttng]]Core concepts
1247 From a user's perspective, the LTTng system is built on a few concepts,
1248 or objects, on which the <<lttng-cli,cmd:lttng command-line tool>>
1249 operates by sending commands to the <<lttng-sessiond,session daemon>>.
1250 Understanding how those objects relate to eachother is key in mastering
1253 The core concepts are:
1255 * <<tracing-session,Tracing session>>
1256 * <<domain,Tracing domain>>
1257 * <<channel,Channel and ring buffer>>
1258 * <<"event","Instrumentation point, event rule, event, and event record">>
1264 A _tracing session_ is a stateful dialogue between you and
1265 a <<lttng-sessiond,session daemon>>. You can
1266 <<creating-destroying-tracing-sessions,create a new tracing
1267 session>> with the `lttng create` command.
1269 Anything that you do when you control LTTng tracers happens within a
1270 tracing session. In particular, a tracing session:
1273 * Has its own set of trace files.
1274 * Has its own state of activity (started or stopped).
1275 * Has its own <<tracing-session-mode,mode>> (local, network streaming,
1277 * Has its own <<channel,channels>> to which are associated their own
1278 <<event,event rules>>.
1281 .A _tracing session_ contains <<channel,channels>> that are members of <<domain,tracing domains>> and contain <<event,event rules>>.
1282 image::concepts.png[]
1284 Those attributes and objects are completely isolated between different
1287 A tracing session is analogous to a cash machine session:
1288 the operations you do on the banking system through the cash machine do
1289 not alter the data of other users of the same system. In the case of
1290 the cash machine, a session lasts as long as your bank card is inside.
1291 In the case of LTTng, a tracing session lasts from the `lttng create`
1292 command to the `lttng destroy` command.
1295 .Each Unix user has its own set of tracing sessions.
1296 image::many-sessions.png[]
1299 [[tracing-session-mode]]
1300 ==== Tracing session mode
1302 LTTng can send the generated trace data to different locations. The
1303 _tracing session mode_ dictates where to send it. The following modes
1304 are available in LTTng{nbsp}{revision}:
1306 [[local-mode]]Local mode::
1307 LTTng writes the traces to the file system of the machine it traces
1310 [[net-streaming-mode]]Network streaming mode::
1311 LTTng sends the traces over the network to a
1312 <<lttng-relayd,relay daemon>> running on a remote system.
1315 LTTng does not write the traces by default. Instead, you can request
1316 LTTng to <<taking-a-snapshot,take a snapshot>>, that is, a copy of the
1317 tracing session's current sub-buffers, and to write it to the
1318 target's file system or to send it over the network to a
1319 <<lttng-relayd,relay daemon>> running on a remote system.
1321 [[live-mode]]Live mode::
1322 This mode is similar to the network streaming mode, but a live
1323 trace viewer can connect to the distant relay daemon to
1324 <<lttng-live,view event records as LTTng generates them>>.
1330 A _tracing domain_ is a namespace for event sources. A tracing domain
1331 has its own properties and features.
1333 There are currently five available tracing domains:
1337 * `java.util.logging` (JUL)
1341 You must specify a tracing domain when using some commands to avoid
1342 ambiguity. For example, since all the domains support named tracepoints
1343 as event sources (instrumentation points that you manually insert in the
1344 source code), you need to specify a tracing domain when
1345 <<enabling-disabling-events,creating an event rule>> because all the
1346 tracing domains could have tracepoints with the same names.
1348 You can create <<channel,channels>> in the Linux kernel and user space
1349 tracing domains. The other tracing domains have a single default
1354 === Channel and ring buffer
1356 A _channel_ is an object which is responsible for a set of ring buffers.
1357 Each ring buffer is divided into multiple sub-buffers. When an LTTng
1358 tracer emits an event, it can record it to one or more
1359 sub-buffers. The attributes of a channel determine what to do when
1360 there's no space left for a new event record because all sub-buffers
1361 are full, where to send a full sub-buffer, and other behaviours.
1363 A channel is always associated to a <<domain,tracing domain>>. The
1364 `java.util.logging` (JUL), log4j, and Python tracing domains each have
1365 a default channel which you cannot configure.
1367 A channel also owns <<event,event rules>>. When an LTTng tracer emits
1368 an event, it records it to the sub-buffers of all
1369 the enabled channels with a satisfied event rule, as long as those
1370 channels are part of active <<tracing-session,tracing sessions>>.
1373 [[channel-buffering-schemes]]
1374 ==== Per-user vs. per-process buffering schemes
1376 A channel has at least one ring buffer _per CPU_. LTTng always
1377 records an event to the ring buffer associated to the CPU on which it
1380 Two _buffering schemes_ are available when you
1381 <<enabling-disabling-channels,create a channel>> in the
1382 user space <<domain,tracing domain>>:
1384 Per-user buffering::
1385 Allocate one set of ring buffers--one per CPU--shared by all the
1386 instrumented processes of each Unix user.
1390 .Per-user buffering scheme.
1391 image::per-user-buffering.png[]
1394 Per-process buffering::
1395 Allocate one set of ring buffers--one per CPU--for each
1396 instrumented process.
1400 .Per-process buffering scheme.
1401 image::per-process-buffering.png[]
1404 The per-process buffering scheme tends to consume more memory than the
1405 per-user option because systems generally have more instrumented
1406 processes than Unix users running instrumented processes. However, the
1407 per-process buffering scheme ensures that one process having a high
1408 event throughput won't fill all the shared sub-buffers of the same
1411 The Linux kernel tracing domain has only one available buffering scheme
1412 which is to allocate a single set of ring buffers for the whole system.
1413 This scheme is similar to the per-user option, but with a single, global
1414 user "running" the kernel.
1417 [[channel-overwrite-mode-vs-discard-mode]]
1418 ==== Overwrite vs. discard event record loss modes
1420 When an event occurs, LTTng records it to a specific sub-buffer (yellow
1421 arc in the following animations) of a specific channel's ring buffer.
1422 When there's no space left in a sub-buffer, the tracer marks it as
1423 consumable (red) and another, empty sub-buffer starts receiving the
1424 following event records. A <<lttng-consumerd,consumer daemon>>
1425 eventually consumes the marked sub-buffer (returns to white).
1428 [role="docsvg-channel-subbuf-anim"]
1433 In an ideal world, sub-buffers are consumed faster than they are filled,
1434 as it is the case in the previous animation. In the real world,
1435 however, all sub-buffers can be full at some point, leaving no space to
1436 record the following events.
1438 By default, LTTng-modules and LTTng-UST are _non-blocking_ tracers: when
1439 no empty sub-buffer is available, it is acceptable to lose event records
1440 when the alternative would be to cause substantial delays in the
1441 instrumented application's execution. LTTng privileges performance over
1442 integrity; it aims at perturbing the target system as little as possible
1443 in order to make tracing of subtle race conditions and rare interrupt
1446 Since LTTng{nbsp}2.10, the LTTng user space tracer, LTTng-UST, supports
1447 a _blocking mode_. See the <<blocking-timeout-example,blocking timeout
1448 example>> to learn how to use the blocking mode.
1450 When it comes to losing event records because no empty sub-buffer is
1451 available, or because the <<opt-blocking-timeout,blocking timeout>> is
1452 reached, the channel's _event record loss mode_ determines what to do.
1453 The available event record loss modes are:
1456 Drop the newest event records until the tracer releases a sub-buffer.
1458 This is the only available mode when you specify a
1459 <<opt-blocking-timeout,blocking timeout>>.
1462 Clear the sub-buffer containing the oldest event records and start
1463 writing the newest event records there.
1465 This mode is sometimes called _flight recorder mode_ because it's
1467 https://en.wikipedia.org/wiki/Flight_recorder[flight recorder]:
1468 always keep a fixed amount of the latest data.
1470 Which mechanism you should choose depends on your context: prioritize
1471 the newest or the oldest event records in the ring buffer?
1473 Beware that, in overwrite mode, the tracer abandons a _whole sub-buffer_
1474 as soon as a there's no space left for a new event record, whereas in
1475 discard mode, the tracer only discards the event record that doesn't
1478 In discard mode, LTTng increments a count of lost event records when an
1479 event record is lost and saves this count to the trace. Since
1480 LTTng{nbsp}2.8, in overwrite mode, LTTng writes to a given sub-buffer
1481 its sequence number within its data stream. With a <<local-mode,local>>,
1482 <<net-streaming-mode,network streaming>>, or <<live-mode,live>>
1483 <<tracing-session,tracing session>>, a trace reader can use such
1484 sequence numbers to report lost packets. In overwrite mode, LTTng
1485 doesn't write to the trace the exact number of lost event records in
1486 those lost sub-buffers.
1488 Trace analyses can use saved discarded event record and sub-buffer
1489 (packet) counts of the trace to decide whether or not to perform the
1490 analyses even if trace data is known to be missing.
1492 There are a few ways to decrease your probability of losing event
1494 <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>> shows
1495 how you can fine-tune the sub-buffer count and size of a channel to
1496 virtually stop losing event records, though at the cost of greater
1500 [[channel-subbuf-size-vs-subbuf-count]]
1501 ==== Sub-buffer count and size
1503 When you <<enabling-disabling-channels,create a channel>>, you can
1504 set its number of sub-buffers and their size.
1506 Note that there is noticeable CPU overhead introduced when
1507 switching sub-buffers (marking a full one as consumable and switching
1508 to an empty one for the following events to be recorded). Knowing this,
1509 the following list presents a few practical situations along with how
1510 to configure the sub-buffer count and size for them:
1512 * **High event throughput**: In general, prefer bigger sub-buffers to
1513 lower the risk of losing event records.
1515 Having bigger sub-buffers also ensures a lower
1516 <<channel-switch-timer,sub-buffer switching frequency>>.
1518 The number of sub-buffers is only meaningful if you create the channel
1519 in overwrite mode: in this case, if a sub-buffer overwrite happens, the
1520 other sub-buffers are left unaltered.
1522 * **Low event throughput**: In general, prefer smaller sub-buffers
1523 since the risk of losing event records is low.
1525 Because events occur less frequently, the sub-buffer switching frequency
1526 should remain low and thus the tracer's overhead should not be a
1529 * **Low memory system**: If your target system has a low memory
1530 limit, prefer fewer first, then smaller sub-buffers.
1532 Even if the system is limited in memory, you want to keep the
1533 sub-buffers as big as possible to avoid a high sub-buffer switching
1536 Note that LTTng uses http://diamon.org/ctf/[CTF] as its trace format,
1537 which means event data is very compact. For example, the average
1538 LTTng kernel event record weights about 32{nbsp}bytes. Thus, a
1539 sub-buffer size of 1{nbsp}MiB is considered big.
1541 The previous situations highlight the major trade-off between a few big
1542 sub-buffers and more, smaller sub-buffers: sub-buffer switching
1543 frequency vs. how much data is lost in overwrite mode. Assuming a
1544 constant event throughput and using the overwrite mode, the two
1545 following configurations have the same ring buffer total size:
1548 [role="docsvg-channel-subbuf-size-vs-count-anim"]
1553 * **Two sub-buffers of 4{nbsp}MiB each**: Expect a very low sub-buffer
1554 switching frequency, but if a sub-buffer overwrite happens, half of
1555 the event records so far (4{nbsp}MiB) are definitely lost.
1556 * **Eight sub-buffers of 1{nbsp}MiB each**: Expect four times the tracer's
1557 overhead as the previous configuration, but if a sub-buffer
1558 overwrite happens, only the eighth of event records so far are
1561 In discard mode, the sub-buffers count parameter is pointless: use two
1562 sub-buffers and set their size according to the requirements of your
1566 [[channel-switch-timer]]
1567 ==== Switch timer period
1569 The _switch timer period_ is an important configurable attribute of
1570 a channel to ensure periodic sub-buffer flushing.
1572 When the _switch timer_ expires, a sub-buffer switch happens. You can
1573 set the switch timer period attribute when you
1574 <<enabling-disabling-channels,create a channel>> to ensure that LTTng
1575 consumes and commits trace data to trace files or to a distant relay
1576 daemon periodically in case of a low event throughput.
1579 [role="docsvg-channel-switch-timer"]
1584 This attribute is also convenient when you use big sub-buffers to cope
1585 with a sporadic high event throughput, even if the throughput is
1589 [[channel-read-timer]]
1590 ==== Read timer period
1592 By default, the LTTng tracers use a notification mechanism to signal a
1593 full sub-buffer so that a consumer daemon can consume it. When such
1594 notifications must be avoided, for example in real-time applications,
1595 you can use the channel's _read timer_ instead. When the read timer
1596 fires, the <<lttng-consumerd,consumer daemon>> checks for full,
1597 consumable sub-buffers.
1600 [[tracefile-rotation]]
1601 ==== Trace file count and size
1603 By default, trace files can grow as large as needed. You can set the
1604 maximum size of each trace file that a channel writes when you
1605 <<enabling-disabling-channels,create a channel>>. When the size of
1606 a trace file reaches the channel's fixed maximum size, LTTng creates
1607 another file to contain the next event records. LTTng appends a file
1608 count to each trace file name in this case.
1610 If you set the trace file size attribute when you create a channel, the
1611 maximum number of trace files that LTTng creates is _unlimited_ by
1612 default. To limit them, you can also set a maximum number of trace
1613 files. When the number of trace files reaches the channel's fixed
1614 maximum count, the oldest trace file is overwritten. This mechanism is
1615 called _trace file rotation_.
1619 Even if you don't limit the trace file count, you cannot assume that
1620 LTTng doesn't manage any trace file.
1622 In other words, there is no safe way to know if LTTng still holds a
1623 given trace file open with the trace file rotation feature.
1625 The only way to obtain an unmanaged, self-contained LTTng trace before
1626 you <<creating-destroying-tracing-sessions,destroy>> the tracing session
1627 is with the <<session-rotation,tracing session rotation>> feature
1628 (available since LTTng{nbsp}2.11).
1633 === Instrumentation point, event rule, event, and event record
1635 An _event rule_ is a set of conditions which must be **all** satisfied
1636 for LTTng to record an occuring event.
1638 You set the conditions when you <<enabling-disabling-events,create
1641 You always attach an event rule to a <<channel,channel>> when you create
1644 When an event passes the conditions of an event rule, LTTng records it
1645 in one of the attached channel's sub-buffers.
1647 The available conditions, as of LTTng{nbsp}{revision}, are:
1649 * The event rule _is enabled_.
1650 * The instrumentation point's type _is{nbsp}T_.
1651 * The instrumentation point's name (sometimes called _event name_)
1652 _matches{nbsp}N_, but _is not{nbsp}E_.
1653 * The instrumentation point's log level _is as severe as{nbsp}L_, or
1654 _is exactly{nbsp}L_.
1655 * The fields of the event's payload _satisfy_ a filter
1656 expression{nbsp}__F__.
1658 As you can see, all the conditions but the dynamic filter are related to
1659 the event rule's status or to the instrumentation point, not to the
1660 occurring events. This is why, without a filter, checking if an event
1661 passes an event rule is not a dynamic task: when you create or modify an
1662 event rule, all the tracers of its tracing domain enable or disable the
1663 instrumentation points themselves once. This is possible because the
1664 attributes of an instrumentation point (type, name, and log level) are
1665 defined statically. In other words, without a dynamic filter, the tracer
1666 _does not evaluate_ the arguments of an instrumentation point unless it
1667 matches an enabled event rule.
1669 Note that, for LTTng to record an event, the <<channel,channel>> to
1670 which a matching event rule is attached must also be enabled, and the
1671 <<tracing-session,tracing session>> owning this channel must be active
1675 .Logical path from an instrumentation point to an event record.
1676 image::event-rule.png[]
1678 .Event, event record, or event rule?
1680 With so many similar terms, it's easy to get confused.
1682 An **event** is the consequence of the execution of an _instrumentation
1683 point_, like a tracepoint that you manually place in some source code,
1684 or a Linux kernel kprobe. An event is said to _occur_ at a specific
1685 time. Different actions can be taken upon the occurrence of an event,
1686 like record the event's payload to a buffer.
1688 An **event record** is the representation of an event in a sub-buffer. A
1689 tracer is responsible for capturing the payload of an event, current
1690 context variables, the event's ID, and the event's timestamp. LTTng
1691 can append this sub-buffer to a trace file.
1693 An **event rule** is a set of conditions which must _all_ be satisfied
1694 for LTTng to record an occuring event. Events still occur without
1695 satisfying event rules, but LTTng does not record them.
1700 == Components of noch:{LTTng}
1702 The second _T_ in _LTTng_ stands for _toolkit_: it would be wrong
1703 to call LTTng a simple _tool_ since it is composed of multiple
1704 interacting components. This section describes those components,
1705 explains their respective roles, and shows how they connect together to
1706 form the LTTng ecosystem.
1708 The following diagram shows how the most important components of LTTng
1709 interact with user applications, the Linux kernel, and you:
1712 .Control and trace data paths between LTTng components.
1713 image::plumbing.png[]
1715 The LTTng project incorporates:
1717 * **LTTng-tools**: Libraries and command-line interface to
1718 control tracing sessions.
1719 ** <<lttng-sessiond,Session daemon>> (man:lttng-sessiond(8)).
1720 ** <<lttng-consumerd,Consumer daemon>> (cmd:lttng-consumerd).
1721 ** <<lttng-relayd,Relay daemon>> (man:lttng-relayd(8)).
1722 ** <<liblttng-ctl-lttng,Tracing control library>> (`liblttng-ctl`).
1723 ** <<lttng-cli,Tracing control command-line tool>> (man:lttng(1)).
1724 * **LTTng-UST**: Libraries and Java/Python packages to trace user
1726 ** <<lttng-ust,User space tracing library>> (`liblttng-ust`) and its
1727 headers to instrument and trace any native user application.
1728 ** <<prebuilt-ust-helpers,Preloadable user space tracing helpers>>:
1729 *** `liblttng-ust-libc-wrapper`
1730 *** `liblttng-ust-pthread-wrapper`
1731 *** `liblttng-ust-cyg-profile`
1732 *** `liblttng-ust-cyg-profile-fast`
1733 *** `liblttng-ust-dl`
1734 ** User space tracepoint provider source files generator command-line
1735 tool (man:lttng-gen-tp(1)).
1736 ** <<lttng-ust-agents,LTTng-UST Java agent>> to instrument and trace
1737 Java applications using `java.util.logging` or
1738 Apache log4j{nbsp}1.2 logging.
1739 ** <<lttng-ust-agents,LTTng-UST Python agent>> to instrument
1740 Python applications using the standard `logging` package.
1741 * **LTTng-modules**: <<lttng-modules,Linux kernel modules>> to trace
1743 ** LTTng kernel tracer module.
1744 ** Tracing ring buffer kernel modules.
1745 ** Probe kernel modules.
1746 ** LTTng logger kernel module.
1750 === Tracing control command-line interface
1753 .The tracing control command-line interface.
1754 image::plumbing-lttng-cli.png[]
1756 The _man:lttng(1) command-line tool_ is the standard user interface to
1757 control LTTng <<tracing-session,tracing sessions>>. The cmd:lttng tool
1758 is part of LTTng-tools.
1760 The cmd:lttng tool is linked with
1761 <<liblttng-ctl-lttng,`liblttng-ctl`>> to communicate with
1762 one or more <<lttng-sessiond,session daemons>> behind the scenes.
1764 The cmd:lttng tool has a Git-like interface:
1768 $ lttng <GENERAL OPTIONS> <COMMAND> <COMMAND OPTIONS>
1771 The <<controlling-tracing,Tracing control>> section explores the
1772 available features of LTTng using the cmd:lttng tool.
1775 [[liblttng-ctl-lttng]]
1776 === Tracing control library
1779 .The tracing control library.
1780 image::plumbing-liblttng-ctl.png[]
1782 The _LTTng control library_, `liblttng-ctl`, is used to communicate
1783 with a <<lttng-sessiond,session daemon>> using a C API that hides the
1784 underlying protocol's details. `liblttng-ctl` is part of LTTng-tools.
1786 The <<lttng-cli,cmd:lttng command-line tool>>
1787 is linked with `liblttng-ctl`.
1789 You can use `liblttng-ctl` in C or $$C++$$ source code by including its
1794 #include <lttng/lttng.h>
1797 Some objects are referenced by name (C string), such as tracing
1798 sessions, but most of them require to create a handle first using
1799 `lttng_create_handle()`.
1801 As of LTTng{nbsp}{revision}, the best available developer documentation for
1802 `liblttng-ctl` is its installed header files. Every function and structure is
1803 thoroughly documented.
1807 === User space tracing library
1810 .The user space tracing library.
1811 image::plumbing-liblttng-ust.png[]
1813 The _user space tracing library_, `liblttng-ust` (see man:lttng-ust(3)),
1814 is the LTTng user space tracer. It receives commands from a
1815 <<lttng-sessiond,session daemon>>, for example to
1816 enable and disable specific instrumentation points, and writes event
1817 records to ring buffers shared with a
1818 <<lttng-consumerd,consumer daemon>>.
1819 `liblttng-ust` is part of LTTng-UST.
1821 Public C header files are installed beside `liblttng-ust` to
1822 instrument any <<c-application,C or $$C++$$ application>>.
1824 <<lttng-ust-agents,LTTng-UST agents>>, which are regular Java and Python
1825 packages, use their own library providing tracepoints which is
1826 linked with `liblttng-ust`.
1828 An application or library does not have to initialize `liblttng-ust`
1829 manually: its constructor does the necessary tasks to properly register
1830 to a session daemon. The initialization phase also enables the
1831 instrumentation points matching the <<event,event rules>> that you
1835 [[lttng-ust-agents]]
1836 === User space tracing agents
1839 .The user space tracing agents.
1840 image::plumbing-lttng-ust-agents.png[]
1842 The _LTTng-UST Java and Python agents_ are regular Java and Python
1843 packages which add LTTng tracing capabilities to the
1844 native logging frameworks. The LTTng-UST agents are part of LTTng-UST.
1846 In the case of Java, the
1847 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[`java.util.logging`
1848 core logging facilities] and
1849 https://logging.apache.org/log4j/1.2/[Apache log4j{nbsp}1.2] are supported.
1850 Note that Apache Log4{nbsp}2 is not supported.
1852 In the case of Python, the standard
1853 https://docs.python.org/3/library/logging.html[`logging`] package
1854 is supported. Both Python{nbsp}2 and Python{nbsp}3 modules can import the
1855 LTTng-UST Python agent package.
1857 The applications using the LTTng-UST agents are in the
1858 `java.util.logging` (JUL),
1859 log4j, and Python <<domain,tracing domains>>.
1861 Both agents use the same mechanism to trace the log statements. When an
1862 agent initializes, it creates a log handler that attaches to the root
1863 logger. The agent also registers to a <<lttng-sessiond,session daemon>>.
1864 When the application executes a log statement, the root logger passes it
1865 to the agent's log handler. The agent's log handler calls a native
1866 function in a tracepoint provider package shared library linked with
1867 <<lttng-ust,`liblttng-ust`>>, passing the formatted log message and
1868 other fields, like its logger name and its log level. This native
1869 function contains a user space instrumentation point, hence tracing the
1872 The log level condition of an
1873 <<event,event rule>> is considered when tracing
1874 a Java or a Python application, and it's compatible with the standard
1875 JUL, log4j, and Python log levels.
1879 === LTTng kernel modules
1882 .The LTTng kernel modules.
1883 image::plumbing-lttng-modules.png[]
1885 The _LTTng kernel modules_ are a set of Linux kernel modules
1886 which implement the kernel tracer of the LTTng project. The LTTng
1887 kernel modules are part of LTTng-modules.
1889 The LTTng kernel modules include:
1891 * A set of _probe_ modules.
1893 Each module attaches to a specific subsystem
1894 of the Linux kernel using its tracepoint instrument points. There are
1895 also modules to attach to the entry and return points of the Linux
1896 system call functions.
1898 * _Ring buffer_ modules.
1900 A ring buffer implementation is provided as kernel modules. The LTTng
1901 kernel tracer writes to the ring buffer; a
1902 <<lttng-consumerd,consumer daemon>> reads from the ring buffer.
1904 * The _LTTng kernel tracer_ module.
1905 * The _LTTng logger_ module.
1907 The LTTng logger module implements the special path:{/proc/lttng-logger}
1908 (and path:{/dev/lttng-logger} since LTTng{nbsp}2.11) files so that any
1909 executable can generate LTTng events by opening and writing to those
1912 See <<proc-lttng-logger-abi,LTTng logger>>.
1914 Generally, you do not have to load the LTTng kernel modules manually
1915 (using man:modprobe(8), for example): a root <<lttng-sessiond,session
1916 daemon>> loads the necessary modules when starting. If you have extra
1917 probe modules, you can specify to load them to the session daemon on
1920 The LTTng kernel modules are installed in
1921 +/usr/lib/modules/__release__/extra+ by default, where +__release__+ is
1922 the kernel release (see `uname --kernel-release`).
1929 .The session daemon.
1930 image::plumbing-sessiond.png[]
1932 The _session daemon_, man:lttng-sessiond(8), is a daemon responsible for
1933 managing tracing sessions and for controlling the various components of
1934 LTTng. The session daemon is part of LTTng-tools.
1936 The session daemon sends control requests to and receives control
1939 * The <<lttng-ust,user space tracing library>>.
1941 Any instance of the user space tracing library first registers to
1942 a session daemon. Then, the session daemon can send requests to
1943 this instance, such as:
1946 ** Get the list of tracepoints.
1947 ** Share an <<event,event rule>> so that the user space tracing library
1948 can enable or disable tracepoints. Amongst the possible conditions
1949 of an event rule is a filter expression which `liblttng-ust` evalutes
1950 when an event occurs.
1951 ** Share <<channel,channel>> attributes and ring buffer locations.
1954 The session daemon and the user space tracing library use a Unix
1955 domain socket for their communication.
1957 * The <<lttng-ust-agents,user space tracing agents>>.
1959 Any instance of a user space tracing agent first registers to
1960 a session daemon. Then, the session daemon can send requests to
1961 this instance, such as:
1964 ** Get the list of loggers.
1965 ** Enable or disable a specific logger.
1968 The session daemon and the user space tracing agent use a TCP connection
1969 for their communication.
1971 * The <<lttng-modules,LTTng kernel tracer>>.
1972 * The <<lttng-consumerd,consumer daemon>>.
1974 The session daemon sends requests to the consumer daemon to instruct
1975 it where to send the trace data streams, amongst other information.
1977 * The <<lttng-relayd,relay daemon>>.
1979 The session daemon receives commands from the
1980 <<liblttng-ctl-lttng,tracing control library>>.
1982 The root session daemon loads the appropriate
1983 <<lttng-modules,LTTng kernel modules>> on startup. It also spawns
1984 a <<lttng-consumerd,consumer daemon>> as soon as you create
1985 an <<event,event rule>>.
1987 The session daemon does not send and receive trace data: this is the
1988 role of the <<lttng-consumerd,consumer daemon>> and
1989 <<lttng-relayd,relay daemon>>. It does, however, generate the
1990 http://diamon.org/ctf/[CTF] metadata stream.
1992 Each Unix user can have its own session daemon instance. The
1993 tracing sessions which different session daemons manage are completely
1996 The root user's session daemon is the only one which is
1997 allowed to control the LTTng kernel tracer, and its spawned consumer
1998 daemon is the only one which is allowed to consume trace data from the
1999 LTTng kernel tracer. Note, however, that any Unix user which is a member
2000 of the <<tracing-group,tracing group>> is allowed
2001 to create <<channel,channels>> in the
2002 Linux kernel <<domain,tracing domain>>, and thus to trace the Linux
2005 The <<lttng-cli,cmd:lttng command-line tool>> automatically starts a
2006 session daemon when using its `create` command if none is currently
2007 running. You can also start the session daemon manually.
2014 .The consumer daemon.
2015 image::plumbing-consumerd.png[]
2017 The _consumer daemon_, cmd:lttng-consumerd, is a daemon which shares
2018 ring buffers with user applications or with the LTTng kernel modules to
2019 collect trace data and send it to some location (on disk or to a
2020 <<lttng-relayd,relay daemon>> over the network). The consumer daemon
2021 is part of LTTng-tools.
2023 You do not start a consumer daemon manually: a consumer daemon is always
2024 spawned by a <<lttng-sessiond,session daemon>> as soon as you create an
2025 <<event,event rule>>, that is, before you start tracing. When you kill
2026 its owner session daemon, the consumer daemon also exits because it is
2027 the session daemon's child process. Command-line options of
2028 man:lttng-sessiond(8) target the consumer daemon process.
2030 There are up to two running consumer daemons per Unix user, whereas only
2031 one session daemon can run per user. This is because each process can be
2032 either 32-bit or 64-bit: if the target system runs a mixture of 32-bit
2033 and 64-bit processes, it is more efficient to have separate
2034 corresponding 32-bit and 64-bit consumer daemons. The root user is an
2035 exception: it can have up to _three_ running consumer daemons: 32-bit
2036 and 64-bit instances for its user applications, and one more
2037 reserved for collecting kernel trace data.
2045 image::plumbing-relayd.png[]
2047 The _relay daemon_, man:lttng-relayd(8), is a daemon acting as a bridge
2048 between remote session and consumer daemons, local trace files, and a
2049 remote live trace viewer. The relay daemon is part of LTTng-tools.
2051 The main purpose of the relay daemon is to implement a receiver of
2052 <<sending-trace-data-over-the-network,trace data over the network>>.
2053 This is useful when the target system does not have much file system
2054 space to record trace files locally.
2056 The relay daemon is also a server to which a
2057 <<lttng-live,live trace viewer>> can
2058 connect. The live trace viewer sends requests to the relay daemon to
2059 receive trace data as the target system emits events. The
2060 communication protocol is named _LTTng live_; it is used over TCP
2063 Note that you can start the relay daemon on the target system directly.
2064 This is the setup of choice when the use case is to view events as
2065 the target system emits them without the need of a remote system.
2069 == [[using-lttng]]Instrumentation
2071 There are many examples of tracing and monitoring in our everyday life:
2073 * You have access to real-time and historical weather reports and
2074 forecasts thanks to weather stations installed around the country.
2075 * You know your heart is safe thanks to an electrocardiogram.
2076 * You make sure not to drive your car too fast and to have enough fuel
2077 to reach your destination thanks to gauges visible on your dashboard.
2079 All the previous examples have something in common: they rely on
2080 **instruments**. Without the electrodes attached to the surface of your
2081 body's skin, cardiac monitoring is futile.
2083 LTTng, as a tracer, is no different from those real life examples. If
2084 you're about to trace a software system or, in other words, record its
2085 history of execution, you better have **instrumentation points** in the
2086 subject you're tracing, that is, the actual software.
2088 Various ways were developed to instrument a piece of software for LTTng
2089 tracing. The most straightforward one is to manually place
2090 instrumentation points, called _tracepoints_, in the software's source
2091 code. It is also possible to add instrumentation points dynamically in
2092 the Linux kernel <<domain,tracing domain>>.
2094 If you're only interested in tracing the Linux kernel, your
2095 instrumentation needs are probably already covered by LTTng's built-in
2096 <<lttng-modules,Linux kernel tracepoints>>. You may also wish to trace a
2097 user application which is already instrumented for LTTng tracing.
2098 In such cases, you can skip this whole section and read the topics of
2099 the <<controlling-tracing,Tracing control>> section.
2101 Many methods are available to instrument a piece of software for LTTng
2104 * <<c-application,User space instrumentation for C and $$C++$$
2106 * <<prebuilt-ust-helpers,Prebuilt user space tracing helpers>>.
2107 * <<java-application,User space Java agent>>.
2108 * <<python-application,User space Python agent>>.
2109 * <<proc-lttng-logger-abi,LTTng logger>>.
2110 * <<instrumenting-linux-kernel,LTTng kernel tracepoints>>.
2114 === [[cxx-application]]User space instrumentation for C and $$C++$$ applications
2116 The procedure to instrument a C or $$C++$$ user application with
2117 the <<lttng-ust,LTTng user space tracing library>>, `liblttng-ust`, is:
2119 . <<tracepoint-provider,Create the source files of a tracepoint provider
2121 . <<probing-the-application-source-code,Add tracepoints to
2122 the application's source code>>.
2123 . <<building-tracepoint-providers-and-user-application,Build and link
2124 a tracepoint provider package and the user application>>.
2126 If you need quick, man:printf(3)-like instrumentation, you can skip
2127 those steps and use <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>>
2130 IMPORTANT: You need to <<installing-lttng,install>> LTTng-UST to
2131 instrument a user application with `liblttng-ust`.
2134 [[tracepoint-provider]]
2135 ==== Create the source files of a tracepoint provider package
2137 A _tracepoint provider_ is a set of compiled functions which provide
2138 **tracepoints** to an application, the type of instrumentation point
2139 supported by LTTng-UST. Those functions can emit events with
2140 user-defined fields and serialize those events as event records to one
2141 or more LTTng-UST <<channel,channel>> sub-buffers. The `tracepoint()`
2142 macro, which you <<probing-the-application-source-code,insert in a user
2143 application's source code>>, calls those functions.
2145 A _tracepoint provider package_ is an object file (`.o`) or a shared
2146 library (`.so`) which contains one or more tracepoint providers.
2147 Its source files are:
2149 * One or more <<tpp-header,tracepoint provider header>> (`.h`).
2150 * A <<tpp-source,tracepoint provider package source>> (`.c`).
2152 A tracepoint provider package is dynamically linked with `liblttng-ust`,
2153 the LTTng user space tracer, at run time.
2156 .User application linked with `liblttng-ust` and containing a tracepoint provider.
2157 image::ust-app.png[]
2159 NOTE: If you need quick, man:printf(3)-like instrumentation, you can
2160 skip creating and using a tracepoint provider and use
2161 <<tracef,`tracef()`>> or <<tracelog,`tracelog()`>> instead.
2165 ===== Create a tracepoint provider header file template
2167 A _tracepoint provider header file_ contains the tracepoint
2168 definitions of a tracepoint provider.
2170 To create a tracepoint provider header file:
2172 . Start from this template:
2176 .Tracepoint provider header file template (`.h` file extension).
2178 #undef TRACEPOINT_PROVIDER
2179 #define TRACEPOINT_PROVIDER provider_name
2181 #undef TRACEPOINT_INCLUDE
2182 #define TRACEPOINT_INCLUDE "./tp.h"
2184 #if !defined(_TP_H) || defined(TRACEPOINT_HEADER_MULTI_READ)
2187 #include <lttng/tracepoint.h>
2190 * Use TRACEPOINT_EVENT(), TRACEPOINT_EVENT_CLASS(),
2191 * TRACEPOINT_EVENT_INSTANCE(), and TRACEPOINT_LOGLEVEL() here.
2196 #include <lttng/tracepoint-event.h>
2202 * `provider_name` with the name of your tracepoint provider.
2203 * `"tp.h"` with the name of your tracepoint provider header file.
2205 . Below the `#include <lttng/tracepoint.h>` line, put your
2206 <<defining-tracepoints,tracepoint definitions>>.
2208 Your tracepoint provider name must be unique amongst all the possible
2209 tracepoint provider names used on the same target system. We
2210 suggest to include the name of your project or company in the name,
2211 for example, `org_lttng_my_project_tpp`.
2213 TIP: [[lttng-gen-tp]]You can use the man:lttng-gen-tp(1) tool to create
2214 this boilerplate for you. When using cmd:lttng-gen-tp, all you need to
2215 write are the <<defining-tracepoints,tracepoint definitions>>.
2218 [[defining-tracepoints]]
2219 ===== Create a tracepoint definition
2221 A _tracepoint definition_ defines, for a given tracepoint:
2223 * Its **input arguments**. They are the macro parameters that the
2224 `tracepoint()` macro accepts for this particular tracepoint
2225 in the user application's source code.
2226 * Its **output event fields**. They are the sources of event fields
2227 that form the payload of any event that the execution of the
2228 `tracepoint()` macro emits for this particular tracepoint.
2230 You can create a tracepoint definition by using the
2231 `TRACEPOINT_EVENT()` macro below the `#include <lttng/tracepoint.h>`
2233 <<tpp-header,tracepoint provider header file template>>.
2235 The syntax of the `TRACEPOINT_EVENT()` macro is:
2238 .`TRACEPOINT_EVENT()` macro syntax.
2241 /* Tracepoint provider name */
2244 /* Tracepoint name */
2247 /* Input arguments */
2252 /* Output event fields */
2261 * `provider_name` with your tracepoint provider name.
2262 * `tracepoint_name` with your tracepoint name.
2263 * `arguments` with the <<tpp-def-input-args,input arguments>>.
2264 * `fields` with the <<tpp-def-output-fields,output event field>>
2267 This tracepoint emits events named `provider_name:tracepoint_name`.
2270 .Event name's length limitation
2272 The concatenation of the tracepoint provider name and the
2273 tracepoint name must not exceed **254{nbsp}characters**. If it does, the
2274 instrumented application compiles and runs, but LTTng throws multiple
2275 warnings and you could experience serious issues.
2278 [[tpp-def-input-args]]The syntax of the `TP_ARGS()` macro is:
2281 .`TP_ARGS()` macro syntax.
2290 * `type` with the C type of the argument.
2291 * `arg_name` with the argument name.
2293 You can repeat `type` and `arg_name` up to 10{nbsp}times to have
2294 more than one argument.
2296 .`TP_ARGS()` usage with three arguments.
2308 The `TP_ARGS()` and `TP_ARGS(void)` forms are valid to create a
2309 tracepoint definition with no input arguments.
2311 [[tpp-def-output-fields]]The `TP_FIELDS()` macro contains a list of
2312 `ctf_*()` macros. Each `ctf_*()` macro defines one event field. See
2313 man:lttng-ust(3) for a complete description of the available `ctf_*()`
2314 macros. A `ctf_*()` macro specifies the type, size, and byte order of
2317 Each `ctf_*()` macro takes an _argument expression_ parameter. This is a
2318 C expression that the tracer evalutes at the `tracepoint()` macro site
2319 in the application's source code. This expression provides a field's
2320 source of data. The argument expression can include input argument names
2321 listed in the `TP_ARGS()` macro.
2323 Each `ctf_*()` macro also takes a _field name_ parameter. Field names
2324 must be unique within a given tracepoint definition.
2326 Here's a complete tracepoint definition example:
2328 .Tracepoint definition.
2330 The following tracepoint definition defines a tracepoint which takes
2331 three input arguments and has four output event fields.
2335 #include "my-custom-structure.h"
2341 const struct my_custom_structure*, my_custom_structure,
2346 ctf_string(query_field, query)
2347 ctf_float(double, ratio_field, ratio)
2348 ctf_integer(int, recv_size, my_custom_structure->recv_size)
2349 ctf_integer(int, send_size, my_custom_structure->send_size)
2354 You can refer to this tracepoint definition with the `tracepoint()`
2355 macro in your application's source code like this:
2359 tracepoint(my_provider, my_tracepoint,
2360 my_structure, some_ratio, the_query);
2364 NOTE: The LTTng tracer only evaluates tracepoint arguments at run time
2365 if they satisfy an enabled <<event,event rule>>.
2368 [[using-tracepoint-classes]]
2369 ===== Use a tracepoint class
2371 A _tracepoint class_ is a class of tracepoints which share the same
2372 output event field definitions. A _tracepoint instance_ is one
2373 instance of such a defined tracepoint class, with its own tracepoint
2376 The <<defining-tracepoints,`TRACEPOINT_EVENT()` macro>> is actually a
2377 shorthand which defines both a tracepoint class and a tracepoint
2378 instance at the same time.
2380 When you build a tracepoint provider package, the C or $$C++$$ compiler
2381 creates one serialization function for each **tracepoint class**. A
2382 serialization function is responsible for serializing the event fields
2383 of a tracepoint to a sub-buffer when tracing.
2385 For various performance reasons, when your situation requires multiple
2386 tracepoint definitions with different names, but with the same event
2387 fields, we recommend that you manually create a tracepoint class
2388 and instantiate as many tracepoint instances as needed. One positive
2389 effect of such a design, amongst other advantages, is that all
2390 tracepoint instances of the same tracepoint class reuse the same
2391 serialization function, thus reducing
2392 https://en.wikipedia.org/wiki/Cache_pollution[cache pollution].
2394 .Use a tracepoint class and tracepoint instances.
2396 Consider the following three tracepoint definitions:
2408 ctf_integer(int, userid, userid)
2409 ctf_integer(size_t, len, len)
2421 ctf_integer(int, userid, userid)
2422 ctf_integer(size_t, len, len)
2434 ctf_integer(int, userid, userid)
2435 ctf_integer(size_t, len, len)
2440 In this case, we create three tracepoint classes, with one implicit
2441 tracepoint instance for each of them: `get_account`, `get_settings`, and
2442 `get_transaction`. However, they all share the same event field names
2443 and types. Hence three identical, yet independent serialization
2444 functions are created when you build the tracepoint provider package.
2446 A better design choice is to define a single tracepoint class and three
2447 tracepoint instances:
2451 /* The tracepoint class */
2452 TRACEPOINT_EVENT_CLASS(
2453 /* Tracepoint provider name */
2456 /* Tracepoint class name */
2459 /* Input arguments */
2465 /* Output event fields */
2467 ctf_integer(int, userid, userid)
2468 ctf_integer(size_t, len, len)
2472 /* The tracepoint instances */
2473 TRACEPOINT_EVENT_INSTANCE(
2474 /* Tracepoint provider name */
2477 /* Tracepoint class name */
2480 /* Tracepoint name */
2483 /* Input arguments */
2489 TRACEPOINT_EVENT_INSTANCE(
2498 TRACEPOINT_EVENT_INSTANCE(
2511 [[assigning-log-levels]]
2512 ===== Assign a log level to a tracepoint definition
2514 You can assign an optional _log level_ to a
2515 <<defining-tracepoints,tracepoint definition>>.
2517 Assigning different levels of severity to tracepoint definitions can
2518 be useful: when you <<enabling-disabling-events,create an event rule>>,
2519 you can target tracepoints having a log level as severe as a specific
2522 The concept of LTTng-UST log levels is similar to the levels found
2523 in typical logging frameworks:
2525 * In a logging framework, the log level is given by the function
2526 or method name you use at the log statement site: `debug()`,
2527 `info()`, `warn()`, `error()`, and so on.
2528 * In LTTng-UST, you statically assign the log level to a tracepoint
2529 definition; any `tracepoint()` macro invocation which refers to
2530 this definition has this log level.
2532 You can assign a log level to a tracepoint definition with the
2533 `TRACEPOINT_LOGLEVEL()` macro. You must use this macro _after_ the
2534 <<defining-tracepoints,`TRACEPOINT_EVENT()`>> or
2535 <<using-tracepoint-classes,`TRACEPOINT_INSTANCE()`>> macro for a given
2538 The syntax of the `TRACEPOINT_LOGLEVEL()` macro is:
2541 .`TRACEPOINT_LOGLEVEL()` macro syntax.
2543 TRACEPOINT_LOGLEVEL(provider_name, tracepoint_name, log_level)
2548 * `provider_name` with the tracepoint provider name.
2549 * `tracepoint_name` with the tracepoint name.
2550 * `log_level` with the log level to assign to the tracepoint
2551 definition named `tracepoint_name` in the `provider_name`
2552 tracepoint provider.
2554 See man:lttng-ust(3) for a list of available log level names.
2556 .Assign the `TRACE_DEBUG_UNIT` log level to a tracepoint definition.
2560 /* Tracepoint definition */
2569 ctf_integer(int, userid, userid)
2570 ctf_integer(size_t, len, len)
2574 /* Log level assignment */
2575 TRACEPOINT_LOGLEVEL(my_app, get_transaction, TRACE_DEBUG_UNIT)
2581 ===== Create a tracepoint provider package source file
2583 A _tracepoint provider package source file_ is a C source file which
2584 includes a <<tpp-header,tracepoint provider header file>> to expand its
2585 macros into event serialization and other functions.
2587 You can always use the following tracepoint provider package source
2591 .Tracepoint provider package source file template.
2593 #define TRACEPOINT_CREATE_PROBES
2598 Replace `tp.h` with the name of your <<tpp-header,tracepoint provider
2599 header file>> name. You may also include more than one tracepoint
2600 provider header file here to create a tracepoint provider package
2601 holding more than one tracepoint providers.
2604 [[probing-the-application-source-code]]
2605 ==== Add tracepoints to an application's source code
2607 Once you <<tpp-header,create a tracepoint provider header file>>, you
2608 can use the `tracepoint()` macro in your application's
2609 source code to insert the tracepoints that this header
2610 <<defining-tracepoints,defines>>.
2612 The `tracepoint()` macro takes at least two parameters: the tracepoint
2613 provider name and the tracepoint name. The corresponding tracepoint
2614 definition defines the other parameters.
2616 .`tracepoint()` usage.
2618 The following <<defining-tracepoints,tracepoint definition>> defines a
2619 tracepoint which takes two input arguments and has two output event
2623 .Tracepoint provider header file.
2625 #include "my-custom-structure.h"
2632 const char*, cmd_name
2635 ctf_string(cmd_name, cmd_name)
2636 ctf_integer(int, number_of_args, argc)
2641 You can refer to this tracepoint definition with the `tracepoint()`
2642 macro in your application's source code like this:
2645 .Application's source file.
2649 int main(int argc, char* argv[])
2651 tracepoint(my_provider, my_tracepoint, argc, argv[0]);
2657 Note how the application's source code includes
2658 the tracepoint provider header file containing the tracepoint
2659 definitions to use, path:{tp.h}.
2662 .`tracepoint()` usage with a complex tracepoint definition.
2664 Consider this complex tracepoint definition, where multiple event
2665 fields refer to the same input arguments in their argument expression
2669 .Tracepoint provider header file.
2671 /* For `struct stat` */
2672 #include <sys/types.h>
2673 #include <sys/stat.h>
2685 ctf_integer(int, my_constant_field, 23 + 17)
2686 ctf_integer(int, my_int_arg_field, my_int_arg)
2687 ctf_integer(int, my_int_arg_field2, my_int_arg * my_int_arg)
2688 ctf_integer(int, sum4_field, my_str_arg[0] + my_str_arg[1] +
2689 my_str_arg[2] + my_str_arg[3])
2690 ctf_string(my_str_arg_field, my_str_arg)
2691 ctf_integer_hex(off_t, size_field, st->st_size)
2692 ctf_float(double, size_dbl_field, (double) st->st_size)
2693 ctf_sequence_text(char, half_my_str_arg_field, my_str_arg,
2694 size_t, strlen(my_str_arg) / 2)
2699 You can refer to this tracepoint definition with the `tracepoint()`
2700 macro in your application's source code like this:
2703 .Application's source file.
2705 #define TRACEPOINT_DEFINE
2712 stat("/etc/fstab", &s);
2713 tracepoint(my_provider, my_tracepoint, 23, "Hello, World!", &s);
2719 If you look at the event record that LTTng writes when tracing this
2720 program, assuming the file size of path:{/etc/fstab} is 301{nbsp}bytes,
2721 it should look like this:
2723 .Event record fields
2725 |Field's name |Field's value
2726 |`my_constant_field` |40
2727 |`my_int_arg_field` |23
2728 |`my_int_arg_field2` |529
2730 |`my_str_arg_field` |`Hello, World!`
2731 |`size_field` |0x12d
2732 |`size_dbl_field` |301.0
2733 |`half_my_str_arg_field` |`Hello,`
2737 Sometimes, the arguments you pass to `tracepoint()` are expensive to
2738 compute--they use the call stack, for example. To avoid this
2739 computation when the tracepoint is disabled, you can use the
2740 `tracepoint_enabled()` and `do_tracepoint()` macros.
2742 The syntax of the `tracepoint_enabled()` and `do_tracepoint()` macros
2746 .`tracepoint_enabled()` and `do_tracepoint()` macros syntax.
2748 tracepoint_enabled(provider_name, tracepoint_name)
2749 do_tracepoint(provider_name, tracepoint_name, ...)
2754 * `provider_name` with the tracepoint provider name.
2755 * `tracepoint_name` with the tracepoint name.
2757 `tracepoint_enabled()` returns a non-zero value if the tracepoint named
2758 `tracepoint_name` from the provider named `provider_name` is enabled
2761 `do_tracepoint()` is like `tracepoint()`, except that it doesn't check
2762 if the tracepoint is enabled. Using `tracepoint()` with
2763 `tracepoint_enabled()` is dangerous since `tracepoint()` also contains
2764 the `tracepoint_enabled()` check, thus a race condition is
2765 possible in this situation:
2768 .Possible race condition when using `tracepoint_enabled()` with `tracepoint()`.
2770 if (tracepoint_enabled(my_provider, my_tracepoint)) {
2771 stuff = prepare_stuff();
2774 tracepoint(my_provider, my_tracepoint, stuff);
2777 If the tracepoint is enabled after the condition, then `stuff` is not
2778 prepared: the emitted event will either contain wrong data, or the whole
2779 application could crash (segmentation fault, for example).
2781 NOTE: Neither `tracepoint_enabled()` nor `do_tracepoint()` have an
2782 `STAP_PROBEV()` call. If you need it, you must emit
2786 [[building-tracepoint-providers-and-user-application]]
2787 ==== Build and link a tracepoint provider package and an application
2789 Once you have one or more <<tpp-header,tracepoint provider header
2790 files>> and a <<tpp-source,tracepoint provider package source file>>,
2791 you can create the tracepoint provider package by compiling its source
2792 file. From here, multiple build and run scenarios are possible. The
2793 following table shows common application and library configurations
2794 along with the required command lines to achieve them.
2796 In the following diagrams, we use the following file names:
2799 Executable application.
2802 Application's object file.
2805 Tracepoint provider package object file.
2808 Tracepoint provider package archive file.
2811 Tracepoint provider package shared object file.
2814 User library object file.
2817 User library shared object file.
2819 We use the following symbols in the diagrams of table below:
2822 .Symbols used in the build scenario diagrams.
2823 image::ust-sit-symbols.png[]
2825 We assume that path:{.} is part of the env:LD_LIBRARY_PATH environment
2826 variable in the following instructions.
2828 [role="growable ust-scenarios",cols="asciidoc,asciidoc"]
2829 .Common tracepoint provider package scenarios.
2831 |Scenario |Instructions
2834 The instrumented application is statically linked with
2835 the tracepoint provider package object.
2837 image::ust-sit+app-linked-with-tp-o+app-instrumented.png[]
2840 include::../common/ust-sit-step-tp-o.txt[]
2842 To build the instrumented application:
2844 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2849 #define TRACEPOINT_DEFINE
2853 . Compile the application source file:
2862 . Build the application:
2867 $ gcc -o app app.o tpp.o -llttng-ust -ldl
2871 To run the instrumented application:
2873 * Start the application:
2883 The instrumented application is statically linked with the
2884 tracepoint provider package archive file.
2886 image::ust-sit+app-linked-with-tp-a+app-instrumented.png[]
2889 To create the tracepoint provider package archive file:
2891 . Compile the <<tpp-source,tracepoint provider package source file>>:
2900 . Create the tracepoint provider package archive file:
2905 $ ar rcs tpp.a tpp.o
2909 To build the instrumented application:
2911 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2916 #define TRACEPOINT_DEFINE
2920 . Compile the application source file:
2929 . Build the application:
2934 $ gcc -o app app.o tpp.a -llttng-ust -ldl
2938 To run the instrumented application:
2940 * Start the application:
2950 The instrumented application is linked with the tracepoint provider
2951 package shared object.
2953 image::ust-sit+app-linked-with-tp-so+app-instrumented.png[]
2956 include::../common/ust-sit-step-tp-so.txt[]
2958 To build the instrumented application:
2960 . In path:{app.c}, before including path:{tpp.h}, add the following line:
2965 #define TRACEPOINT_DEFINE
2969 . Compile the application source file:
2978 . Build the application:
2983 $ gcc -o app app.o -ldl -L. -ltpp
2987 To run the instrumented application:
2989 * Start the application:
2999 The tracepoint provider package shared object is preloaded before the
3000 instrumented application starts.
3002 image::ust-sit+tp-so-preloaded+app-instrumented.png[]
3005 include::../common/ust-sit-step-tp-so.txt[]
3007 To build the instrumented application:
3009 . In path:{app.c}, before including path:{tpp.h}, add the
3015 #define TRACEPOINT_DEFINE
3016 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3020 . Compile the application source file:
3029 . Build the application:
3034 $ gcc -o app app.o -ldl
3038 To run the instrumented application with tracing support:
3040 * Preload the tracepoint provider package shared object and
3041 start the application:
3046 $ LD_PRELOAD=./libtpp.so ./app
3050 To run the instrumented application without tracing support:
3052 * Start the application:
3062 The instrumented application dynamically loads the tracepoint provider
3063 package shared object.
3065 image::ust-sit+app-dlopens-tp-so+app-instrumented.png[]
3068 include::../common/ust-sit-step-tp-so.txt[]
3070 To build the instrumented application:
3072 . In path:{app.c}, before including path:{tpp.h}, add the
3078 #define TRACEPOINT_DEFINE
3079 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3083 . Compile the application source file:
3092 . Build the application:
3097 $ gcc -o app app.o -ldl
3101 To run the instrumented application:
3103 * Start the application:
3113 The application is linked with the instrumented user library.
3115 The instrumented user library is statically linked with the tracepoint
3116 provider package object file.
3118 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-o+lib-instrumented.png[]
3121 include::../common/ust-sit-step-tp-o-fpic.txt[]
3123 To build the instrumented user library:
3125 . In path:{emon.c}, before including path:{tpp.h}, add the
3131 #define TRACEPOINT_DEFINE
3135 . Compile the user library source file:
3140 $ gcc -I. -fpic -c emon.c
3144 . Build the user library shared object:
3149 $ gcc -shared -o libemon.so emon.o tpp.o -llttng-ust -ldl
3153 To build the application:
3155 . Compile the application source file:
3164 . Build the application:
3169 $ gcc -o app app.o -L. -lemon
3173 To run the application:
3175 * Start the application:
3185 The application is linked with the instrumented user library.
3187 The instrumented user library is linked with the tracepoint provider
3188 package shared object.
3190 image::ust-sit+app-linked-with-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3193 include::../common/ust-sit-step-tp-so.txt[]
3195 To build the instrumented user library:
3197 . In path:{emon.c}, before including path:{tpp.h}, add the
3203 #define TRACEPOINT_DEFINE
3207 . Compile the user library source file:
3212 $ gcc -I. -fpic -c emon.c
3216 . Build the user library shared object:
3221 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3225 To build the application:
3227 . Compile the application source file:
3236 . Build the application:
3241 $ gcc -o app app.o -L. -lemon
3245 To run the application:
3247 * Start the application:
3257 The tracepoint provider package shared object is preloaded before the
3260 The application is linked with the instrumented user library.
3262 image::ust-sit+tp-so-preloaded+app-linked-with-lib+lib-instrumented.png[]
3265 include::../common/ust-sit-step-tp-so.txt[]
3267 To build the instrumented user library:
3269 . In path:{emon.c}, before including path:{tpp.h}, add the
3275 #define TRACEPOINT_DEFINE
3276 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3280 . Compile the user library source file:
3285 $ gcc -I. -fpic -c emon.c
3289 . Build the user library shared object:
3294 $ gcc -shared -o libemon.so emon.o -ldl
3298 To build the application:
3300 . Compile the application source file:
3309 . Build the application:
3314 $ gcc -o app app.o -L. -lemon
3318 To run the application with tracing support:
3320 * Preload the tracepoint provider package shared object and
3321 start the application:
3326 $ LD_PRELOAD=./libtpp.so ./app
3330 To run the application without tracing support:
3332 * Start the application:
3342 The application is linked with the instrumented user library.
3344 The instrumented user library dynamically loads the tracepoint provider
3345 package shared object.
3347 image::ust-sit+app-linked-with-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3350 include::../common/ust-sit-step-tp-so.txt[]
3352 To build the instrumented user library:
3354 . In path:{emon.c}, before including path:{tpp.h}, add the
3360 #define TRACEPOINT_DEFINE
3361 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3365 . Compile the user library source file:
3370 $ gcc -I. -fpic -c emon.c
3374 . Build the user library shared object:
3379 $ gcc -shared -o libemon.so emon.o -ldl
3383 To build the application:
3385 . Compile the application source file:
3394 . Build the application:
3399 $ gcc -o app app.o -L. -lemon
3403 To run the application:
3405 * Start the application:
3415 The application dynamically loads the instrumented user library.
3417 The instrumented user library is linked with the tracepoint provider
3418 package shared object.
3420 image::ust-sit+app-dlopens-lib+lib-linked-with-tp-so+lib-instrumented.png[]
3423 include::../common/ust-sit-step-tp-so.txt[]
3425 To build the instrumented user library:
3427 . In path:{emon.c}, before including path:{tpp.h}, add the
3433 #define TRACEPOINT_DEFINE
3437 . Compile the user library source file:
3442 $ gcc -I. -fpic -c emon.c
3446 . Build the user library shared object:
3451 $ gcc -shared -o libemon.so emon.o -ldl -L. -ltpp
3455 To build the application:
3457 . Compile the application source file:
3466 . Build the application:
3471 $ gcc -o app app.o -ldl -L. -lemon
3475 To run the application:
3477 * Start the application:
3487 The application dynamically loads the instrumented user library.
3489 The instrumented user library dynamically loads the tracepoint provider
3490 package shared object.
3492 image::ust-sit+app-dlopens-lib+lib-dlopens-tp-so+lib-instrumented.png[]
3495 include::../common/ust-sit-step-tp-so.txt[]
3497 To build the instrumented user library:
3499 . In path:{emon.c}, before including path:{tpp.h}, add the
3505 #define TRACEPOINT_DEFINE
3506 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3510 . Compile the user library source file:
3515 $ gcc -I. -fpic -c emon.c
3519 . Build the user library shared object:
3524 $ gcc -shared -o libemon.so emon.o -ldl
3528 To build the application:
3530 . Compile the application source file:
3539 . Build the application:
3544 $ gcc -o app app.o -ldl -L. -lemon
3548 To run the application:
3550 * Start the application:
3560 The tracepoint provider package shared object is preloaded before the
3563 The application dynamically loads the instrumented user library.
3565 image::ust-sit+tp-so-preloaded+app-dlopens-lib+lib-instrumented.png[]
3568 include::../common/ust-sit-step-tp-so.txt[]
3570 To build the instrumented user library:
3572 . In path:{emon.c}, before including path:{tpp.h}, add the
3578 #define TRACEPOINT_DEFINE
3579 #define TRACEPOINT_PROBE_DYNAMIC_LINKAGE
3583 . Compile the user library source file:
3588 $ gcc -I. -fpic -c emon.c
3592 . Build the user library shared object:
3597 $ gcc -shared -o libemon.so emon.o -ldl
3601 To build the application:
3603 . Compile the application source file:
3612 . Build the application:
3617 $ gcc -o app app.o -L. -lemon
3621 To run the application with tracing support:
3623 * Preload the tracepoint provider package shared object and
3624 start the application:
3629 $ LD_PRELOAD=./libtpp.so ./app
3633 To run the application without tracing support:
3635 * Start the application:
3645 The application is statically linked with the tracepoint provider
3646 package object file.
3648 The application is linked with the instrumented user library.
3650 image::ust-sit+app-linked-with-tp-o+app-linked-with-lib+lib-instrumented.png[]
3653 include::../common/ust-sit-step-tp-o.txt[]
3655 To build the instrumented user library:
3657 . In path:{emon.c}, before including path:{tpp.h}, add the
3663 #define TRACEPOINT_DEFINE
3667 . Compile the user library source file:
3672 $ gcc -I. -fpic -c emon.c
3676 . Build the user library shared object:
3681 $ gcc -shared -o libemon.so emon.o
3685 To build the application:
3687 . Compile the application source file:
3696 . Build the application:
3701 $ gcc -o app app.o tpp.o -llttng-ust -ldl -L. -lemon
3705 To run the instrumented application:
3707 * Start the application:
3717 The application is statically linked with the tracepoint provider
3718 package object file.
3720 The application dynamically loads the instrumented user library.
3722 image::ust-sit+app-linked-with-tp-o+app-dlopens-lib+lib-instrumented.png[]
3725 include::../common/ust-sit-step-tp-o.txt[]
3727 To build the application:
3729 . In path:{app.c}, before including path:{tpp.h}, add the following line:
3734 #define TRACEPOINT_DEFINE
3738 . Compile the application source file:
3747 . Build the application:
3752 $ gcc -Wl,--export-dynamic -o app app.o tpp.o \
3757 The `--export-dynamic` option passed to the linker is necessary for the
3758 dynamically loaded library to ``see'' the tracepoint symbols defined in
3761 To build the instrumented user library:
3763 . Compile the user library source file:
3768 $ gcc -I. -fpic -c emon.c
3772 . Build the user library shared object:
3777 $ gcc -shared -o libemon.so emon.o
3781 To run the application:
3783 * Start the application:
3794 [[using-lttng-ust-with-daemons]]
3795 ===== Use noch:{LTTng-UST} with daemons
3797 If your instrumented application calls man:fork(2), man:clone(2),
3798 or BSD's man:rfork(2), without a following man:exec(3)-family
3799 system call, you must preload the path:{liblttng-ust-fork.so} shared
3800 object when you start the application.
3804 $ LD_PRELOAD=liblttng-ust-fork.so ./my-app
3807 If your tracepoint provider package is
3808 a shared library which you also preload, you must put both
3809 shared objects in env:LD_PRELOAD:
3813 $ LD_PRELOAD=liblttng-ust-fork.so:/path/to/tp.so ./my-app
3819 ===== Use noch:{LTTng-UST} with applications which close file descriptors that don't belong to them
3821 If your instrumented application closes one or more file descriptors
3822 which it did not open itself, you must preload the
3823 path:{liblttng-ust-fd.so} shared object when you start the application:
3827 $ LD_PRELOAD=liblttng-ust-fd.so ./my-app
3830 Typical use cases include closing all the file descriptors after
3831 man:fork(2) or man:rfork(2) and buggy applications doing
3835 [[lttng-ust-pkg-config]]
3836 ===== Use noch:{pkg-config}
3838 On some distributions, LTTng-UST ships with a
3839 https://www.freedesktop.org/wiki/Software/pkg-config/[pkg-config]
3840 metadata file. If this is your case, then you can use cmd:pkg-config to
3841 build an application on the command line:
3845 $ gcc -o my-app my-app.o tp.o $(pkg-config --cflags --libs lttng-ust)
3849 [[instrumenting-32-bit-app-on-64-bit-system]]
3850 ===== [[advanced-instrumenting-techniques]]Build a 32-bit instrumented application for a 64-bit target system
3852 In order to trace a 32-bit application running on a 64-bit system,
3853 LTTng must use a dedicated 32-bit
3854 <<lttng-consumerd,consumer daemon>>.
3856 The following steps show how to build and install a 32-bit consumer
3857 daemon, which is _not_ part of the default 64-bit LTTng build, how to
3858 build and install the 32-bit LTTng-UST libraries, and how to build and
3859 link an instrumented 32-bit application in that context.
3861 To build a 32-bit instrumented application for a 64-bit target system,
3862 assuming you have a fresh target system with no installed Userspace RCU
3865 . Download, build, and install a 32-bit version of Userspace RCU:
3870 $ cd $(mktemp -d) &&
3871 wget http://lttng.org/files/urcu/userspace-rcu-latest-0.9.tar.bz2 &&
3872 tar -xf userspace-rcu-latest-0.9.tar.bz2 &&
3873 cd userspace-rcu-0.9.* &&
3874 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 &&
3876 sudo make install &&
3881 . Using your distribution's package manager, or from source, install
3882 the following 32-bit versions of the following dependencies of
3883 LTTng-tools and LTTng-UST:
3886 * https://sourceforge.net/projects/libuuid/[libuuid]
3887 * http://directory.fsf.org/wiki/Popt[popt]
3888 * http://www.xmlsoft.org/[libxml2]
3891 . Download, build, and install a 32-bit version of the latest
3892 LTTng-UST{nbsp}{revision}:
3897 $ cd $(mktemp -d) &&
3898 wget http://lttng.org/files/lttng-ust/lttng-ust-latest-2.11.tar.bz2 &&
3899 tar -xf lttng-ust-latest-2.11.tar.bz2 &&
3900 cd lttng-ust-2.11.* &&
3901 ./configure --libdir=/usr/local/lib32 \
3902 CFLAGS=-m32 CXXFLAGS=-m32 \
3903 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' &&
3905 sudo make install &&
3912 Depending on your distribution,
3913 32-bit libraries could be installed at a different location than
3914 `/usr/lib32`. For example, Debian is known to install
3915 some 32-bit libraries in `/usr/lib/i386-linux-gnu`.
3917 In this case, make sure to set `LDFLAGS` to all the
3918 relevant 32-bit library paths, for example:
3922 $ LDFLAGS='-L/usr/lib/i386-linux-gnu -L/usr/lib32'
3926 . Download the latest LTTng-tools{nbsp}{revision}, build, and install
3927 the 32-bit consumer daemon:
3932 $ cd $(mktemp -d) &&
3933 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3934 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3935 cd lttng-tools-2.11.* &&
3936 ./configure --libdir=/usr/local/lib32 CFLAGS=-m32 CXXFLAGS=-m32 \
3937 LDFLAGS='-L/usr/local/lib32 -L/usr/lib32' \
3938 --disable-bin-lttng --disable-bin-lttng-crash \
3939 --disable-bin-lttng-relayd --disable-bin-lttng-sessiond &&
3941 cd src/bin/lttng-consumerd &&
3942 sudo make install &&
3947 . From your distribution or from source,
3948 <<installing-lttng,install>> the 64-bit versions of
3949 LTTng-UST and Userspace RCU.
3950 . Download, build, and install the 64-bit version of the
3951 latest LTTng-tools{nbsp}{revision}:
3956 $ cd $(mktemp -d) &&
3957 wget http://lttng.org/files/lttng-tools/lttng-tools-latest-2.11.tar.bz2 &&
3958 tar -xf lttng-tools-latest-2.11.tar.bz2 &&
3959 cd lttng-tools-2.11.* &&
3960 ./configure --with-consumerd32-libdir=/usr/local/lib32 \
3961 --with-consumerd32-bin=/usr/local/lib32/lttng/libexec/lttng-consumerd &&
3963 sudo make install &&
3968 . Pass the following options to man:gcc(1), man:g++(1), or man:clang(1)
3969 when linking your 32-bit application:
3972 -m32 -L/usr/lib32 -L/usr/local/lib32 \
3973 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32
3976 For example, let's rebuild the quick start example in
3977 <<tracing-your-own-user-application,Trace a user application>> as an
3978 instrumented 32-bit application:
3983 $ gcc -m32 -c -I. hello-tp.c
3984 $ gcc -m32 -c hello.c
3985 $ gcc -m32 -o hello hello.o hello-tp.o \
3986 -L/usr/lib32 -L/usr/local/lib32 \
3987 -Wl,-rpath,/usr/lib32,-rpath,/usr/local/lib32 \
3992 No special action is required to execute the 32-bit application and
3993 to trace it: use the command-line man:lttng(1) tool as usual.
4000 man:tracef(3) is a small LTTng-UST API designed for quick,
4001 man:printf(3)-like instrumentation without the burden of
4002 <<tracepoint-provider,creating>> and
4003 <<building-tracepoint-providers-and-user-application,building>>
4004 a tracepoint provider package.
4006 To use `tracef()` in your application:
4008 . In the C or C++ source files where you need to use `tracef()`,
4009 include `<lttng/tracef.h>`:
4014 #include <lttng/tracef.h>
4018 . In the application's source code, use `tracef()` like you would use
4026 tracef("my message: %d (%s)", my_integer, my_string);
4032 . Link your application with `liblttng-ust`:
4037 $ gcc -o app app.c -llttng-ust
4041 To trace the events that `tracef()` calls emit:
4043 * <<enabling-disabling-events,Create an event rule>> which matches the
4044 `lttng_ust_tracef:*` event name:
4049 $ lttng enable-event --userspace 'lttng_ust_tracef:*'
4054 .Limitations of `tracef()`
4056 The `tracef()` utility function was developed to make user space tracing
4057 super simple, albeit with notable disadvantages compared to
4058 <<defining-tracepoints,user-defined tracepoints>>:
4060 * All the emitted events have the same tracepoint provider and
4061 tracepoint names, respectively `lttng_ust_tracef` and `event`.
4062 * There is no static type checking.
4063 * The only event record field you actually get, named `msg`, is a string
4064 potentially containing the values you passed to `tracef()`
4065 using your own format string. This also means that you cannot filter
4066 events with a custom expression at run time because there are no
4068 * Since `tracef()` uses the C standard library's man:vasprintf(3)
4069 function behind the scenes to format the strings at run time, its
4070 expected performance is lower than with user-defined tracepoints,
4071 which do not require a conversion to a string.
4073 Taking this into consideration, `tracef()` is useful for some quick
4074 prototyping and debugging, but you should not consider it for any
4075 permanent and serious applicative instrumentation.
4081 ==== Use `tracelog()`
4083 The man:tracelog(3) API is very similar to <<tracef,`tracef()`>>, with
4084 the difference that it accepts an additional log level parameter.
4086 The goal of `tracelog()` is to ease the migration from logging to
4089 To use `tracelog()` in your application:
4091 . In the C or C++ source files where you need to use `tracelog()`,
4092 include `<lttng/tracelog.h>`:
4097 #include <lttng/tracelog.h>
4101 . In the application's source code, use `tracelog()` like you would use
4102 man:printf(3), except for the first parameter which is the log
4110 tracelog(TRACE_WARNING, "my message: %d (%s)",
4111 my_integer, my_string);
4117 See man:lttng-ust(3) for a list of available log level names.
4119 . Link your application with `liblttng-ust`:
4124 $ gcc -o app app.c -llttng-ust
4128 To trace the events that `tracelog()` calls emit with a log level
4129 _as severe as_ a specific log level:
4131 * <<enabling-disabling-events,Create an event rule>> which matches the
4132 `lttng_ust_tracelog:*` event name and a minimum level
4138 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4139 --loglevel=TRACE_WARNING
4143 To trace the events that `tracelog()` calls emit with a
4144 _specific log level_:
4146 * Create an event rule which matches the `lttng_ust_tracelog:*`
4147 event name and a specific log level:
4152 $ lttng enable-event --userspace 'lttng_ust_tracelog:*'
4153 --loglevel-only=TRACE_INFO
4158 [[prebuilt-ust-helpers]]
4159 === Prebuilt user space tracing helpers
4161 The LTTng-UST package provides a few helpers in the form of preloadable
4162 shared objects which automatically instrument system functions and
4165 The helper shared objects are normally found in dir:{/usr/lib}. If you
4166 built LTTng-UST <<building-from-source,from source>>, they are probably
4167 located in dir:{/usr/local/lib}.
4169 The installed user space tracing helpers in LTTng-UST{nbsp}{revision}
4172 path:{liblttng-ust-libc-wrapper.so}::
4173 path:{liblttng-ust-pthread-wrapper.so}::
4174 <<liblttng-ust-libc-pthread-wrapper,C{nbsp}standard library
4175 memory and POSIX threads function tracing>>.
4177 path:{liblttng-ust-cyg-profile.so}::
4178 path:{liblttng-ust-cyg-profile-fast.so}::
4179 <<liblttng-ust-cyg-profile,Function entry and exit tracing>>.
4181 path:{liblttng-ust-dl.so}::
4182 <<liblttng-ust-dl,Dynamic linker tracing>>.
4184 To use a user space tracing helper with any user application:
4186 * Preload the helper shared object when you start the application:
4191 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so my-app
4195 You can preload more than one helper:
4200 $ LD_PRELOAD=liblttng-ust-libc-wrapper.so:liblttng-ust-dl.so my-app
4206 [[liblttng-ust-libc-pthread-wrapper]]
4207 ==== Instrument C standard library memory and POSIX threads functions
4209 The path:{liblttng-ust-libc-wrapper.so} and
4210 path:{liblttng-ust-pthread-wrapper.so} helpers
4211 add instrumentation to some C standard library and POSIX
4215 .Functions instrumented by preloading path:{liblttng-ust-libc-wrapper.so}.
4217 |TP provider name |TP name |Instrumented function
4219 .6+|`lttng_ust_libc` |`malloc` |man:malloc(3)
4220 |`calloc` |man:calloc(3)
4221 |`realloc` |man:realloc(3)
4222 |`free` |man:free(3)
4223 |`memalign` |man:memalign(3)
4224 |`posix_memalign` |man:posix_memalign(3)
4228 .Functions instrumented by preloading path:{liblttng-ust-pthread-wrapper.so}.
4230 |TP provider name |TP name |Instrumented function
4232 .4+|`lttng_ust_pthread` |`pthread_mutex_lock_req` |man:pthread_mutex_lock(3p) (request time)
4233 |`pthread_mutex_lock_acq` |man:pthread_mutex_lock(3p) (acquire time)
4234 |`pthread_mutex_trylock` |man:pthread_mutex_trylock(3p)
4235 |`pthread_mutex_unlock` |man:pthread_mutex_unlock(3p)
4238 When you preload the shared object, it replaces the functions listed
4239 in the previous tables by wrappers which contain tracepoints and call
4240 the replaced functions.
4243 [[liblttng-ust-cyg-profile]]
4244 ==== Instrument function entry and exit
4246 The path:{liblttng-ust-cyg-profile*.so} helpers can add instrumentation
4247 to the entry and exit points of functions.
4249 man:gcc(1) and man:clang(1) have an option named
4250 https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html[`-finstrument-functions`]
4251 which generates instrumentation calls for entry and exit to functions.
4252 The LTTng-UST function tracing helpers,
4253 path:{liblttng-ust-cyg-profile.so} and
4254 path:{liblttng-ust-cyg-profile-fast.so}, take advantage of this feature
4255 to add tracepoints to the two generated functions (which contain
4256 `cyg_profile` in their names, hence the helper's name).
4258 To use the LTTng-UST function tracing helper, the source files to
4259 instrument must be built using the `-finstrument-functions` compiler
4262 There are two versions of the LTTng-UST function tracing helper:
4264 * **path:{liblttng-ust-cyg-profile-fast.so}** is a lightweight variant
4265 that you should only use when it can be _guaranteed_ that the
4266 complete event stream is recorded without any lost event record.
4267 Any kind of duplicate information is left out.
4269 Assuming no event record is lost, having only the function addresses on
4270 entry is enough to create a call graph, since an event record always
4271 contains the ID of the CPU that generated it.
4273 You can use a tool like man:addr2line(1) to convert function addresses
4274 back to source file names and line numbers.
4276 * **path:{liblttng-ust-cyg-profile.so}** is a more robust variant
4277 which also works in use cases where event records might get discarded or
4278 not recorded from application startup.
4279 In these cases, the trace analyzer needs more information to be
4280 able to reconstruct the program flow.
4282 See man:lttng-ust-cyg-profile(3) to learn more about the instrumentation
4283 points of this helper.
4285 All the tracepoints that this helper provides have the
4286 log level `TRACE_DEBUG_FUNCTION` (see man:lttng-ust(3)).
4288 TIP: It's sometimes a good idea to limit the number of source files that
4289 you compile with the `-finstrument-functions` option to prevent LTTng
4290 from writing an excessive amount of trace data at run time. When using
4291 man:gcc(1), you can use the
4292 `-finstrument-functions-exclude-function-list` option to avoid
4293 instrument entries and exits of specific function names.
4298 ==== Instrument the dynamic linker
4300 The path:{liblttng-ust-dl.so} helper adds instrumentation to the
4301 man:dlopen(3) and man:dlclose(3) function calls.
4303 See man:lttng-ust-dl(3) to learn more about the instrumentation points
4308 [[java-application]]
4309 === User space Java agent
4311 You can instrument any Java application which uses one of the following
4314 * The https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[**`java.util.logging`**]
4315 (JUL) core logging facilities.
4316 * http://logging.apache.org/log4j/1.2/[**Apache log4j{nbsp}1.2**], since
4317 LTTng{nbsp}2.6. Note that Apache Log4j{nbsp}2 is not supported.
4320 .LTTng-UST Java agent imported by a Java application.
4321 image::java-app.png[]
4323 Note that the methods described below are new in LTTng{nbsp}2.8.
4324 Previous LTTng versions use another technique.
4326 NOTE: We use http://openjdk.java.net/[OpenJDK]{nbsp}8 for development
4327 and https://ci.lttng.org/[continuous integration], thus this version is
4328 directly supported. However, the LTTng-UST Java agent is also tested
4329 with OpenJDK{nbsp}7.
4334 ==== Use the LTTng-UST Java agent for `java.util.logging`
4336 To use the LTTng-UST Java agent in a Java application which uses
4337 `java.util.logging` (JUL):
4339 . In the Java application's source code, import the LTTng-UST
4340 log handler package for `java.util.logging`:
4345 import org.lttng.ust.agent.jul.LttngLogHandler;
4349 . Create an LTTng-UST JUL log handler:
4354 Handler lttngUstLogHandler = new LttngLogHandler();
4358 . Add this handler to the JUL loggers which should emit LTTng events:
4363 Logger myLogger = Logger.getLogger("some-logger");
4365 myLogger.addHandler(lttngUstLogHandler);
4369 . Use `java.util.logging` log statements and configuration as usual.
4370 The loggers with an attached LTTng-UST log handler can emit
4373 . Before exiting the application, remove the LTTng-UST log handler from
4374 the loggers attached to it and call its `close()` method:
4379 myLogger.removeHandler(lttngUstLogHandler);
4380 lttngUstLogHandler.close();
4384 This is not strictly necessary, but it is recommended for a clean
4385 disposal of the handler's resources.
4387 . Include the LTTng-UST Java agent's common and JUL-specific JAR files,
4388 path:{lttng-ust-agent-common.jar} and path:{lttng-ust-agent-jul.jar},
4390 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4391 path] when you build the Java application.
4393 The JAR files are typically located in dir:{/usr/share/java}.
4395 IMPORTANT: The LTTng-UST Java agent must be
4396 <<installing-lttng,installed>> for the logging framework your
4399 .Use the LTTng-UST Java agent for `java.util.logging`.
4404 import java.io.IOException;
4405 import java.util.logging.Handler;
4406 import java.util.logging.Logger;
4407 import org.lttng.ust.agent.jul.LttngLogHandler;
4411 private static final int answer = 42;
4413 public static void main(String[] argv) throws Exception
4416 Logger logger = Logger.getLogger("jello");
4418 // Create an LTTng-UST log handler
4419 Handler lttngUstLogHandler = new LttngLogHandler();
4421 // Add the LTTng-UST log handler to our logger
4422 logger.addHandler(lttngUstLogHandler);
4425 logger.info("some info");
4426 logger.warning("some warning");
4428 logger.finer("finer information; the answer is " + answer);
4430 logger.severe("error!");
4432 // Not mandatory, but cleaner
4433 logger.removeHandler(lttngUstLogHandler);
4434 lttngUstLogHandler.close();
4443 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4446 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4447 <<enabling-disabling-events,create an event rule>> matching the
4448 `jello` JUL logger, and <<basic-tracing-session-control,start tracing>>:
4453 $ lttng enable-event --jul jello
4457 Run the compiled class:
4461 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4464 <<basic-tracing-session-control,Stop tracing>> and inspect the
4474 In the resulting trace, an <<event,event record>> generated by a Java
4475 application using `java.util.logging` is named `lttng_jul:event` and
4476 has the following fields:
4479 Log record's message.
4485 Name of the class in which the log statement was executed.
4488 Name of the method in which the log statement was executed.
4491 Logging time (timestamp in milliseconds).
4494 Log level integer value.
4497 ID of the thread in which the log statement was executed.
4499 You can use the opt:lttng-enable-event(1):--loglevel or
4500 opt:lttng-enable-event(1):--loglevel-only option of the
4501 man:lttng-enable-event(1) command to target a range of JUL log levels
4502 or a specific JUL log level.
4507 ==== Use the LTTng-UST Java agent for Apache log4j
4509 To use the LTTng-UST Java agent in a Java application which uses
4510 Apache log4j{nbsp}1.2:
4512 . In the Java application's source code, import the LTTng-UST
4513 log appender package for Apache log4j:
4518 import org.lttng.ust.agent.log4j.LttngLogAppender;
4522 . Create an LTTng-UST log4j log appender:
4527 Appender lttngUstLogAppender = new LttngLogAppender();
4531 . Add this appender to the log4j loggers which should emit LTTng events:
4536 Logger myLogger = Logger.getLogger("some-logger");
4538 myLogger.addAppender(lttngUstLogAppender);
4542 . Use Apache log4j log statements and configuration as usual. The
4543 loggers with an attached LTTng-UST log appender can emit LTTng events.
4545 . Before exiting the application, remove the LTTng-UST log appender from
4546 the loggers attached to it and call its `close()` method:
4551 myLogger.removeAppender(lttngUstLogAppender);
4552 lttngUstLogAppender.close();
4556 This is not strictly necessary, but it is recommended for a clean
4557 disposal of the appender's resources.
4559 . Include the LTTng-UST Java agent's common and log4j-specific JAR
4560 files, path:{lttng-ust-agent-common.jar} and
4561 path:{lttng-ust-agent-log4j.jar}, in the
4562 https://docs.oracle.com/javase/tutorial/essential/environment/paths.html[class
4563 path] when you build the Java application.
4565 The JAR files are typically located in dir:{/usr/share/java}.
4567 IMPORTANT: The LTTng-UST Java agent must be
4568 <<installing-lttng,installed>> for the logging framework your
4571 .Use the LTTng-UST Java agent for Apache log4j.
4576 import org.apache.log4j.Appender;
4577 import org.apache.log4j.Logger;
4578 import org.lttng.ust.agent.log4j.LttngLogAppender;
4582 private static final int answer = 42;
4584 public static void main(String[] argv) throws Exception
4587 Logger logger = Logger.getLogger("jello");
4589 // Create an LTTng-UST log appender
4590 Appender lttngUstLogAppender = new LttngLogAppender();
4592 // Add the LTTng-UST log appender to our logger
4593 logger.addAppender(lttngUstLogAppender);
4596 logger.info("some info");
4597 logger.warn("some warning");
4599 logger.debug("debug information; the answer is " + answer);
4601 logger.fatal("error!");
4603 // Not mandatory, but cleaner
4604 logger.removeAppender(lttngUstLogAppender);
4605 lttngUstLogAppender.close();
4611 Build this example (`$LOG4JPATH` is the path to the Apache log4j JAR
4616 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH Test.java
4619 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4620 <<enabling-disabling-events,create an event rule>> matching the
4621 `jello` log4j logger, and <<basic-tracing-session-control,start tracing>>:
4626 $ lttng enable-event --log4j jello
4630 Run the compiled class:
4634 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-log4j.jar:$LOG4JPATH:. Test
4637 <<basic-tracing-session-control,Stop tracing>> and inspect the
4647 In the resulting trace, an <<event,event record>> generated by a Java
4648 application using log4j is named `lttng_log4j:event` and
4649 has the following fields:
4652 Log record's message.
4658 Name of the class in which the log statement was executed.
4661 Name of the method in which the log statement was executed.
4664 Name of the file in which the executed log statement is located.
4667 Line number at which the log statement was executed.
4673 Log level integer value.
4676 Name of the Java thread in which the log statement was executed.
4678 You can use the opt:lttng-enable-event(1):--loglevel or
4679 opt:lttng-enable-event(1):--loglevel-only option of the
4680 man:lttng-enable-event(1) command to target a range of Apache log4j log levels
4681 or a specific log4j log level.
4685 [[java-application-context]]
4686 ==== Provide application-specific context fields in a Java application
4688 A Java application-specific context field is a piece of state provided
4689 by the application which <<adding-context,you can add>>, using the
4690 man:lttng-add-context(1) command, to each <<event,event record>>
4691 produced by the log statements of this application.
4693 For example, a given object might have a current request ID variable.
4694 You can create a context information retriever for this object and
4695 assign a name to this current request ID. You can then, using the
4696 man:lttng-add-context(1) command, add this context field by name to
4697 the JUL or log4j <<channel,channel>>.
4699 To provide application-specific context fields in a Java application:
4701 . In the Java application's source code, import the LTTng-UST
4702 Java agent context classes and interfaces:
4707 import org.lttng.ust.agent.context.ContextInfoManager;
4708 import org.lttng.ust.agent.context.IContextInfoRetriever;
4712 . Create a context information retriever class, that is, a class which
4713 implements the `IContextInfoRetriever` interface:
4718 class MyContextInfoRetriever implements IContextInfoRetriever
4721 public Object retrieveContextInfo(String key)
4723 if (key.equals("intCtx")) {
4725 } else if (key.equals("strContext")) {
4726 return "context value!";
4735 This `retrieveContextInfo()` method is the only member of the
4736 `IContextInfoRetriever` interface. Its role is to return the current
4737 value of a state by name to create a context field. The names of the
4738 context fields and which state variables they return depends on your
4741 All primitive types and objects are supported as context fields.
4742 When `retrieveContextInfo()` returns an object, the context field
4743 serializer calls its `toString()` method to add a string field to
4744 event records. The method can also return `null`, which means that
4745 no context field is available for the required name.
4747 . Register an instance of your context information retriever class to
4748 the context information manager singleton:
4753 IContextInfoRetriever cir = new MyContextInfoRetriever();
4754 ContextInfoManager cim = ContextInfoManager.getInstance();
4755 cim.registerContextInfoRetriever("retrieverName", cir);
4759 . Before exiting the application, remove your context information
4760 retriever from the context information manager singleton:
4765 ContextInfoManager cim = ContextInfoManager.getInstance();
4766 cim.unregisterContextInfoRetriever("retrieverName");
4770 This is not strictly necessary, but it is recommended for a clean
4771 disposal of some manager's resources.
4773 . Build your Java application with LTTng-UST Java agent support as
4774 usual, following the procedure for either the <<jul,JUL>> or
4775 <<log4j,Apache log4j>> framework.
4778 .Provide application-specific context fields in a Java application.
4783 import java.util.logging.Handler;
4784 import java.util.logging.Logger;
4785 import org.lttng.ust.agent.jul.LttngLogHandler;
4786 import org.lttng.ust.agent.context.ContextInfoManager;
4787 import org.lttng.ust.agent.context.IContextInfoRetriever;
4791 // Our context information retriever class
4792 private static class MyContextInfoRetriever
4793 implements IContextInfoRetriever
4796 public Object retrieveContextInfo(String key) {
4797 if (key.equals("intCtx")) {
4799 } else if (key.equals("strContext")) {
4800 return "context value!";
4807 private static final int answer = 42;
4809 public static void main(String args[]) throws Exception
4811 // Get the context information manager instance
4812 ContextInfoManager cim = ContextInfoManager.getInstance();
4814 // Create and register our context information retriever
4815 IContextInfoRetriever cir = new MyContextInfoRetriever();
4816 cim.registerContextInfoRetriever("myRetriever", cir);
4819 Logger logger = Logger.getLogger("jello");
4821 // Create an LTTng-UST log handler
4822 Handler lttngUstLogHandler = new LttngLogHandler();
4824 // Add the LTTng-UST log handler to our logger
4825 logger.addHandler(lttngUstLogHandler);
4828 logger.info("some info");
4829 logger.warning("some warning");
4831 logger.finer("finer information; the answer is " + answer);
4833 logger.severe("error!");
4835 // Not mandatory, but cleaner
4836 logger.removeHandler(lttngUstLogHandler);
4837 lttngUstLogHandler.close();
4838 cim.unregisterContextInfoRetriever("myRetriever");
4847 $ javac -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar Test.java
4850 <<creating-destroying-tracing-sessions,Create a tracing session>>
4851 and <<enabling-disabling-events,create an event rule>> matching the
4857 $ lttng enable-event --jul jello
4860 <<adding-context,Add the application-specific context fields>> to the
4865 $ lttng add-context --jul --type='$app.myRetriever:intCtx'
4866 $ lttng add-context --jul --type='$app.myRetriever:strContext'
4869 <<basic-tracing-session-control,Start tracing>>:
4876 Run the compiled class:
4880 $ java -cp /usr/share/java/jarpath/lttng-ust-agent-common.jar:/usr/share/java/jarpath/lttng-ust-agent-jul.jar:. Test
4883 <<basic-tracing-session-control,Stop tracing>> and inspect the
4895 [[python-application]]
4896 === User space Python agent
4898 You can instrument a Python{nbsp}2 or Python{nbsp}3 application which
4900 https://docs.python.org/3/library/logging.html[`logging`] package.
4902 Each log statement emits an LTTng event once the
4903 application module imports the
4904 <<lttng-ust-agents,LTTng-UST Python agent>> package.
4907 .A Python application importing the LTTng-UST Python agent.
4908 image::python-app.png[]
4910 To use the LTTng-UST Python agent:
4912 . In the Python application's source code, import the LTTng-UST Python
4922 The LTTng-UST Python agent automatically adds its logging handler to the
4923 root logger at import time.
4925 Any log statement that the application executes before this import does
4926 not emit an LTTng event.
4928 IMPORTANT: The LTTng-UST Python agent must be
4929 <<installing-lttng,installed>>.
4931 . Use log statements and logging configuration as usual.
4932 Since the LTTng-UST Python agent adds a handler to the _root_
4933 logger, you can trace any log statement from any logger.
4935 .Use the LTTng-UST Python agent.
4946 logging.basicConfig()
4947 logger = logging.getLogger('my-logger')
4950 logger.debug('debug message')
4951 logger.info('info message')
4952 logger.warn('warn message')
4953 logger.error('error message')
4954 logger.critical('critical message')
4958 if __name__ == '__main__':
4962 NOTE: `logging.basicConfig()`, which adds to the root logger a basic
4963 logging handler which prints to the standard error stream, is not
4964 strictly required for LTTng-UST tracing to work, but in versions of
4965 Python preceding{nbsp}3.2, you could see a warning message which indicates
4966 that no handler exists for the logger `my-logger`.
4968 <<creating-destroying-tracing-sessions,Create a tracing session>>,
4969 <<enabling-disabling-events,create an event rule>> matching the
4970 `my-logger` Python logger, and <<basic-tracing-session-control,start
4976 $ lttng enable-event --python my-logger
4980 Run the Python script:
4987 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
4997 In the resulting trace, an <<event,event record>> generated by a Python
4998 application is named `lttng_python:event` and has the following fields:
5001 Logging time (string).
5004 Log record's message.
5010 Name of the function in which the log statement was executed.
5013 Line number at which the log statement was executed.
5016 Log level integer value.
5019 ID of the Python thread in which the log statement was executed.
5022 Name of the Python thread in which the log statement was executed.
5024 You can use the opt:lttng-enable-event(1):--loglevel or
5025 opt:lttng-enable-event(1):--loglevel-only option of the
5026 man:lttng-enable-event(1) command to target a range of Python log levels
5027 or a specific Python log level.
5029 When an application imports the LTTng-UST Python agent, the agent tries
5030 to register to a <<lttng-sessiond,session daemon>>. Note that you must
5031 <<start-sessiond,start the session daemon>> _before_ you run the Python
5032 application. If a session daemon is found, the agent tries to register
5033 to it during five seconds, after which the application continues
5034 without LTTng tracing support. You can override this timeout value with
5035 the env:LTTNG_UST_PYTHON_REGISTER_TIMEOUT environment variable
5038 If the session daemon stops while a Python application with an imported
5039 LTTng-UST Python agent runs, the agent retries to connect and to
5040 register to a session daemon every three seconds. You can override this
5041 delay with the env:LTTNG_UST_PYTHON_REGISTER_RETRY_DELAY environment
5046 [[proc-lttng-logger-abi]]
5049 The `lttng-tracer` Linux kernel module, part of
5050 <<lttng-modules,LTTng-modules>>, creates the special LTTng logger files
5051 path:{/proc/lttng-logger} and path:{/dev/lttng-logger} (since
5052 LTTng{nbsp}2.11) when it's loaded. Any application can write text data
5053 to any of those files to emit an LTTng event.
5056 .An application writes to the LTTng logger file to emit an LTTng event.
5057 image::lttng-logger.png[]
5059 The LTTng logger is the quickest method--not the most efficient,
5060 however--to add instrumentation to an application. It is designed
5061 mostly to instrument shell scripts:
5065 $ echo "Some message, some $variable" > /dev/lttng-logger
5068 Any event that the LTTng logger emits is named `lttng_logger` and
5069 belongs to the Linux kernel <<domain,tracing domain>>. However, unlike
5070 other instrumentation points in the kernel tracing domain, **any Unix
5071 user** can <<enabling-disabling-events,create an event rule>> which
5072 matches its event name, not only the root user or users in the
5073 <<tracing-group,tracing group>>.
5075 To use the LTTng logger:
5077 * From any application, write text data to the path:{/dev/lttng-logger}
5080 The `msg` field of `lttng_logger` event records contains the
5083 NOTE: The maximum message length of an LTTng logger event is
5084 1024{nbsp}bytes. Writing more than this makes the LTTng logger emit more
5085 than one event to contain the remaining data.
5087 You should not use the LTTng logger to trace a user application which
5088 can be instrumented in a more efficient way, namely:
5090 * <<c-application,C and $$C++$$ applications>>.
5091 * <<java-application,Java applications>>.
5092 * <<python-application,Python applications>>.
5094 .Use the LTTng logger.
5099 echo 'Hello, World!' > /dev/lttng-logger
5101 df --human-readable --print-type / > /dev/lttng-logger
5104 <<creating-destroying-tracing-sessions,Create a tracing session>>,
5105 <<enabling-disabling-events,create an event rule>> matching the
5106 `lttng_logger` Linux kernel tracepoint, and
5107 <<basic-tracing-session-control,start tracing>>:
5112 $ lttng enable-event --kernel lttng_logger
5116 Run the Bash script:
5123 <<basic-tracing-session-control,Stop tracing>> and inspect the recorded
5134 [[instrumenting-linux-kernel]]
5135 === LTTng kernel tracepoints
5137 NOTE: This section shows how to _add_ instrumentation points to the
5138 Linux kernel. The kernel's subsystems are already thoroughly
5139 instrumented at strategic places for LTTng when you
5140 <<installing-lttng,install>> the <<lttng-modules,LTTng-modules>>
5144 There are two methods to instrument the Linux kernel:
5146 . <<linux-add-lttng-layer,Add an LTTng layer>> over an existing ftrace
5147 tracepoint which uses the `TRACE_EVENT()` API.
5149 Choose this if you want to instrumentation a Linux kernel tree with an
5150 instrumentation point compatible with ftrace, perf, and SystemTap.
5152 . Use an <<linux-lttng-tracepoint-event,LTTng-only approach>> to
5153 instrument an out-of-tree kernel module.
5155 Choose this if you don't need ftrace, perf, or SystemTap support.
5159 [[linux-add-lttng-layer]]
5160 ==== [[instrumenting-linux-kernel-itself]][[mainline-trace-event]][[lttng-adaptation-layer]]Add an LTTng layer to an existing ftrace tracepoint
5162 This section shows how to add an LTTng layer to existing ftrace
5163 instrumentation using the `TRACE_EVENT()` API.
5165 This section does not document the `TRACE_EVENT()` macro. You can
5166 read the following articles to learn more about this API:
5168 * http://lwn.net/Articles/379903/[Using the TRACE_EVENT() macro (Part{nbsp}1)]
5169 * http://lwn.net/Articles/381064/[Using the TRACE_EVENT() macro (Part{nbsp}2)]
5170 * http://lwn.net/Articles/383362/[Using the TRACE_EVENT() macro (Part{nbsp}3)]
5172 The following procedure assumes that your ftrace tracepoints are
5173 correctly defined in their own header and that they are created in
5174 one source file using the `CREATE_TRACE_POINTS` definition.
5176 To add an LTTng layer over an existing ftrace tracepoint:
5178 . Make sure the following kernel configuration options are
5184 * `CONFIG_HIGH_RES_TIMERS`
5185 * `CONFIG_TRACEPOINTS`
5188 . Build the Linux source tree with your custom ftrace tracepoints.
5189 . Boot the resulting Linux image on your target system.
5191 Confirm that the tracepoints exist by looking for their names in the
5192 dir:{/sys/kernel/debug/tracing/events/subsys} directory, where `subsys`
5193 is your subsystem's name.
5195 . Get a copy of the latest LTTng-modules{nbsp}{revision}:
5200 $ cd $(mktemp -d) &&
5201 wget http://lttng.org/files/lttng-modules/lttng-modules-latest-2.11.tar.bz2 &&
5202 tar -xf lttng-modules-latest-2.11.tar.bz2 &&
5203 cd lttng-modules-2.11.*
5207 . In dir:{instrumentation/events/lttng-module}, relative to the root
5208 of the LTTng-modules source tree, create a header file named
5209 +__subsys__.h+ for your custom subsystem +__subsys__+ and write your
5210 LTTng-modules tracepoint definitions using the LTTng-modules
5213 Start with this template:
5217 .path:{instrumentation/events/lttng-module/my_subsys.h}
5220 #define TRACE_SYSTEM my_subsys
5222 #if !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ)
5223 #define _LTTNG_MY_SUBSYS_H
5225 #include "../../../probes/lttng-tracepoint-event.h"
5226 #include <linux/tracepoint.h>
5228 LTTNG_TRACEPOINT_EVENT(
5230 * Format is identical to TRACE_EVENT()'s version for the three
5231 * following macro parameters:
5234 TP_PROTO(int my_int, const char *my_string),
5235 TP_ARGS(my_int, my_string),
5237 /* LTTng-modules specific macros */
5239 ctf_integer(int, my_int_field, my_int)
5240 ctf_string(my_bar_field, my_bar)
5244 #endif /* !defined(_LTTNG_MY_SUBSYS_H) || defined(TRACE_HEADER_MULTI_READ) */
5246 #include "../../../probes/define_trace.h"
5250 The entries in the `TP_FIELDS()` section are the list of fields for the
5251 LTTng tracepoint. This is similar to the `TP_STRUCT__entry()` part of
5252 ftrace's `TRACE_EVENT()` macro.
5254 See <<lttng-modules-tp-fields,Tracepoint fields macros>> for a
5255 complete description of the available `ctf_*()` macros.
5257 . Create the LTTng-modules probe's kernel module C source file,
5258 +probes/lttng-probe-__subsys__.c+, where +__subsys__+ is your
5263 .path:{probes/lttng-probe-my-subsys.c}
5265 #include <linux/module.h>
5266 #include "../lttng-tracer.h"
5269 * Build-time verification of mismatch between mainline
5270 * TRACE_EVENT() arguments and the LTTng-modules adaptation
5271 * layer LTTNG_TRACEPOINT_EVENT() arguments.
5273 #include <trace/events/my_subsys.h>
5275 /* Create LTTng tracepoint probes */
5276 #define LTTNG_PACKAGE_BUILD
5277 #define CREATE_TRACE_POINTS
5278 #define TRACE_INCLUDE_PATH ../instrumentation/events/lttng-module
5280 #include "../instrumentation/events/lttng-module/my_subsys.h"
5282 MODULE_LICENSE("GPL and additional rights");
5283 MODULE_AUTHOR("Your name <your-email>");
5284 MODULE_DESCRIPTION("LTTng my_subsys probes");
5285 MODULE_VERSION(__stringify(LTTNG_MODULES_MAJOR_VERSION) "."
5286 __stringify(LTTNG_MODULES_MINOR_VERSION) "."
5287 __stringify(LTTNG_MODULES_PATCHLEVEL_VERSION)
5288 LTTNG_MODULES_EXTRAVERSION);
5292 . Edit path:{probes/KBuild} and add your new kernel module object
5293 next to the existing ones:
5297 .path:{probes/KBuild}
5301 obj-m += lttng-probe-module.o
5302 obj-m += lttng-probe-power.o
5304 obj-m += lttng-probe-my-subsys.o
5310 . Build and install the LTTng kernel modules:
5315 $ make KERNELDIR=/path/to/linux
5316 # make modules_install && depmod -a
5320 Replace `/path/to/linux` with the path to the Linux source tree where
5321 you defined and used tracepoints with ftrace's `TRACE_EVENT()` macro.
5323 Note that you can also use the
5324 <<lttng-tracepoint-event-code,`LTTNG_TRACEPOINT_EVENT_CODE()` macro>>
5325 instead of `LTTNG_TRACEPOINT_EVENT()` to use custom local variables and
5326 C code that need to be executed before the event fields are recorded.
5328 The best way to learn how to use the previous LTTng-modules macros is to
5329 inspect the existing LTTng-modules tracepoint definitions in the
5330 dir:{instrumentation/events/lttng-module} header files. Compare them
5331 with the Linux kernel mainline versions in the
5332 dir:{include/trace/events} directory of the Linux source tree.
5336 [[lttng-tracepoint-event-code]]
5337 ===== Use custom C code to access the data for tracepoint fields
5339 Although we recommended to always use the
5340 <<lttng-adaptation-layer,`LTTNG_TRACEPOINT_EVENT()`>> macro to describe
5341 the arguments and fields of an LTTng-modules tracepoint when possible,
5342 sometimes you need a more complex process to access the data that the
5343 tracer records as event record fields. In other words, you need local
5344 variables and multiple C{nbsp}statements instead of simple
5345 argument-based expressions that you pass to the
5346 <<lttng-modules-tp-fields,`ctf_*()` macros of `TP_FIELDS()`>>.
5348 You can use the `LTTNG_TRACEPOINT_EVENT_CODE()` macro instead of
5349 `LTTNG_TRACEPOINT_EVENT()` to declare custom local variables and define
5350 a block of C{nbsp}code to be executed before LTTng records the fields.
5351 The structure of this macro is:
5354 .`LTTNG_TRACEPOINT_EVENT_CODE()` macro syntax.
5356 LTTNG_TRACEPOINT_EVENT_CODE(
5358 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5359 * version for the following three macro parameters:
5362 TP_PROTO(int my_int, const char *my_string),
5363 TP_ARGS(my_int, my_string),
5365 /* Declarations of custom local variables */
5368 unsigned long b = 0;
5369 const char *name = "(undefined)";
5370 struct my_struct *my_struct;
5374 * Custom code which uses both tracepoint arguments
5375 * (in TP_ARGS()) and local variables (in TP_locvar()).
5377 * Local variables are actually members of a structure pointed
5378 * to by the special variable tp_locvar.
5382 tp_locvar->a = my_int + 17;
5383 tp_locvar->my_struct = get_my_struct_at(tp_locvar->a);
5384 tp_locvar->b = my_struct_compute_b(tp_locvar->my_struct);
5385 tp_locvar->name = my_struct_get_name(tp_locvar->my_struct);
5386 put_my_struct(tp_locvar->my_struct);
5395 * Format identical to the LTTNG_TRACEPOINT_EVENT()
5396 * version for this, except that tp_locvar members can be
5397 * used in the argument expression parameters of
5398 * the ctf_*() macros.
5401 ctf_integer(unsigned long, my_struct_b, tp_locvar->b)
5402 ctf_integer(int, my_struct_a, tp_locvar->a)
5403 ctf_string(my_string_field, my_string)
5404 ctf_string(my_struct_name, tp_locvar->name)
5409 IMPORTANT: The C code defined in `TP_code()` must not have any side
5410 effects when executed. In particular, the code must not allocate
5411 memory or get resources without deallocating this memory or putting
5412 those resources afterwards.
5415 [[instrumenting-linux-kernel-tracing]]
5416 ==== Load and unload a custom probe kernel module
5418 You must load a <<lttng-adaptation-layer,created LTTng-modules probe
5419 kernel module>> in the kernel before it can emit LTTng events.
5421 To load the default probe kernel modules and a custom probe kernel
5424 * Use the opt:lttng-sessiond(8):--extra-kmod-probes option to give extra
5425 probe modules to load when starting a root <<lttng-sessiond,session
5429 .Load the `my_subsys`, `usb`, and the default probe modules.
5433 # lttng-sessiond --extra-kmod-probes=my_subsys,usb
5438 You only need to pass the subsystem name, not the whole kernel module
5441 To load _only_ a given custom probe kernel module:
5443 * Use the opt:lttng-sessiond(8):--kmod-probes option to give the probe
5444 modules to load when starting a root session daemon:
5447 .Load only the `my_subsys` and `usb` probe modules.
5451 # lttng-sessiond --kmod-probes=my_subsys,usb
5456 To confirm that a probe module is loaded:
5463 $ lsmod | grep lttng_probe_usb
5467 To unload the loaded probe modules:
5469 * Kill the session daemon with `SIGTERM`:
5474 # pkill lttng-sessiond
5478 You can also use man:modprobe(8)'s `--remove` option if the session
5479 daemon terminates abnormally.
5482 [[controlling-tracing]]
5485 Once an application or a Linux kernel is
5486 <<instrumenting,instrumented>> for LTTng tracing,
5489 This section is divided in topics on how to use the various
5490 <<plumbing,components of LTTng>>, in particular the <<lttng-cli,cmd:lttng
5491 command-line tool>>, to _control_ the LTTng daemons and tracers.
5493 NOTE: In the following subsections, we refer to an man:lttng(1) command
5494 using its man page name. For example, instead of _Run the `create`
5495 command to..._, we use _Run the man:lttng-create(1) command to..._.
5499 === Start a session daemon
5501 In some situations, you need to run a <<lttng-sessiond,session daemon>>
5502 (man:lttng-sessiond(8)) _before_ you can use the man:lttng(1)
5505 You will see the following error when you run a command while no session
5509 Error: No session daemon is available
5512 The only command that automatically runs a session daemon is
5513 man:lttng-create(1), which you use to
5514 <<creating-destroying-tracing-sessions,create a tracing session>>. While
5515 this is most of the time the first operation that you do, sometimes it's
5516 not. Some examples are:
5518 * <<list-instrumentation-points,List the available instrumentation points>>.
5519 * <<saving-loading-tracing-session,Load a tracing session configuration>>.
5521 [[tracing-group]] Each Unix user must have its own running session
5522 daemon to trace user applications. The session daemon that the root user
5523 starts is the only one allowed to control the LTTng kernel tracer. Users
5524 that are part of the _tracing group_ can control the root session
5525 daemon. The default tracing group name is `tracing`; you can set it to
5526 something else with the opt:lttng-sessiond(8):--group option when you
5527 start the root session daemon.
5529 To start a user session daemon:
5531 * Run man:lttng-sessiond(8):
5536 $ lttng-sessiond --daemonize
5540 To start the root session daemon:
5542 * Run man:lttng-sessiond(8) as the root user:
5547 # lttng-sessiond --daemonize
5551 In both cases, remove the opt:lttng-sessiond(8):--daemonize option to
5552 start the session daemon in foreground.
5554 To stop a session daemon, use man:kill(1) on its process ID (standard
5557 Note that some Linux distributions could manage the LTTng session daemon
5558 as a service. In this case, you should use the service manager to
5559 start, restart, and stop session daemons.
5562 [[creating-destroying-tracing-sessions]]
5563 === Create and destroy a tracing session
5565 Almost all the LTTng control operations happen in the scope of
5566 a <<tracing-session,tracing session>>, which is the dialogue between the
5567 <<lttng-sessiond,session daemon>> and you.
5569 To create a tracing session with a generated name:
5571 * Use the man:lttng-create(1) command:
5580 The created tracing session's name is `auto` followed by the
5583 To create a tracing session with a specific name:
5585 * Use the optional argument of the man:lttng-create(1) command:
5590 $ lttng create my-session
5594 Replace `my-session` with the specific tracing session name.
5596 LTTng appends the creation date to the created tracing session's name.
5598 LTTng writes the traces of a tracing session in
5599 +$LTTNG_HOME/lttng-trace/__name__+ by default, where +__name__+ is the
5600 name of the tracing session. Note that the env:LTTNG_HOME environment
5601 variable defaults to `$HOME` if not set.
5603 To output LTTng traces to a non-default location:
5605 * Use the opt:lttng-create(1):--output option of the man:lttng-create(1) command:
5610 $ lttng create my-session --output=/tmp/some-directory
5614 You may create as many tracing sessions as you wish.
5616 To list all the existing tracing sessions for your Unix user:
5618 * Use the man:lttng-list(1) command:
5627 When you create a tracing session, it is set as the _current tracing
5628 session_. The following man:lttng(1) commands operate on the current
5629 tracing session when you don't specify one:
5631 [role="list-3-cols"]
5632 * man:lttng-add-context(1)
5633 * man:lttng-destroy(1)
5634 * man:lttng-disable-channel(1)
5635 * man:lttng-disable-event(1)
5636 * man:lttng-disable-rotation(1)
5637 * man:lttng-enable-channel(1)
5638 * man:lttng-enable-event(1)
5639 * man:lttng-enable-rotation(1)
5641 * man:lttng-regenerate(1)
5642 * man:lttng-rotate(1)
5644 * man:lttng-snapshot(1)
5645 * man:lttng-start(1)
5646 * man:lttng-status(1)
5648 * man:lttng-track(1)
5649 * man:lttng-untrack(1)
5652 To change the current tracing session:
5654 * Use the man:lttng-set-session(1) command:
5659 $ lttng set-session new-session
5663 Replace `new-session` by the name of the new current tracing session.
5665 When you are done tracing in a given tracing session, you can destroy
5666 it. This operation frees the resources taken by the tracing session
5667 to destroy; it does not destroy the trace data that LTTng wrote for
5668 this tracing session.
5670 To destroy the current tracing session:
5672 * Use the man:lttng-destroy(1) command:
5681 The man:lttng-destroy(1) command also runs the man:lttng-stop(1)
5682 command implicitly (see <<basic-tracing-session-control,Start and stop a
5683 tracing session>>). You need to stop tracing to make LTTng flush the
5684 remaining trace data and make the trace readable.
5687 [[list-instrumentation-points]]
5688 === List the available instrumentation points
5690 The <<lttng-sessiond,session daemon>> can query the running instrumented
5691 user applications and the Linux kernel to get a list of available
5692 instrumentation points. For the Linux kernel <<domain,tracing domain>>,
5693 they are tracepoints and system calls. For the user space tracing
5694 domain, they are tracepoints. For the other tracing domains, they are
5697 To list the available instrumentation points:
5699 * Use the man:lttng-list(1) command with the requested tracing domain's
5703 * opt:lttng-list(1):--kernel: Linux kernel tracepoints (your Unix user
5704 must be a root user, or it must be a member of the
5705 <<tracing-group,tracing group>>).
5706 * opt:lttng-list(1):--kernel with opt:lttng-list(1):--syscall: Linux
5707 kernel system calls (your Unix user must be a root user, or it must be
5708 a member of the tracing group).
5709 * opt:lttng-list(1):--userspace: user space tracepoints.
5710 * opt:lttng-list(1):--jul: `java.util.logging` loggers.
5711 * opt:lttng-list(1):--log4j: Apache log4j loggers.
5712 * opt:lttng-list(1):--python: Python loggers.
5715 .List the available user space tracepoints.
5719 $ lttng list --userspace
5723 .List the available Linux kernel system call tracepoints.
5727 $ lttng list --kernel --syscall
5732 [[enabling-disabling-events]]
5733 === Create and enable an event rule
5735 Once you <<creating-destroying-tracing-sessions,create a tracing
5736 session>>, you can create <<event,event rules>> with the
5737 man:lttng-enable-event(1) command.
5739 You specify each condition with a command-line option. The available
5740 condition arguments are shown in the following table.
5742 [role="growable",cols="asciidoc,asciidoc,default"]
5743 .Condition command-line arguments for the man:lttng-enable-event(1) command.
5745 |Argument |Description |Applicable tracing domains
5751 . +--probe=__ADDR__+
5752 . +--function=__ADDR__+
5753 . +--userspace-probe=__PATH__:__SYMBOL__+
5754 . +--userspace-probe=sdt:__PATH__:__PROVIDER__:__NAME__+
5757 Instead of using the default _tracepoint_ instrumentation type, use:
5759 . A Linux system call (entry and exit).
5760 . A Linux https://lwn.net/Articles/132196/[kprobe] (symbol or address).
5761 . The entry and return points of a Linux function (symbol or address).
5762 . The entry point of a user application or library function (path to
5763 application/library and symbol).
5764 . A https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[SystemTap
5765 Statically Defined Tracing] (USDT) probe (path to application/library,
5766 provider and probe names).
5770 |First positional argument.
5773 Tracepoint or system call name.
5775 With the opt:lttng-enable-event(1):--probe,
5776 opt:lttng-enable-event(1):--function, and
5777 opt:lttng-enable-event(1):--userspace-probe options, this is a custom
5778 name given to the event rule. With the JUL, log4j, and Python domains,
5779 this is a logger name.
5781 With a tracepoint, logger, or system call name, you can use the special
5782 `*` globbing character to match anything (for example, `sched_*`,
5790 . +--loglevel=__LEVEL__+
5791 . +--loglevel-only=__LEVEL__+
5794 . Match only tracepoints or log statements with a logging level at
5795 least as severe as +__LEVEL__+.
5796 . Match only tracepoints or log statements with a logging level
5797 equal to +__LEVEL__+.
5799 See man:lttng-enable-event(1) for the list of available logging level
5802 |User space, JUL, log4j, and Python.
5804 |+--exclude=__EXCLUSIONS__+
5807 When you use a `*` character at the end of the tracepoint or logger
5808 name (first positional argument), exclude the specific names in the
5809 comma-delimited list +__EXCLUSIONS__+.
5812 User space, JUL, log4j, and Python.
5814 |+--filter=__EXPR__+
5817 Match only events which satisfy the expression +__EXPR__+.
5819 See man:lttng-enable-event(1) to learn more about the syntax of a
5826 You attach an event rule to a <<channel,channel>> on creation. If you do
5827 not specify the channel with the opt:lttng-enable-event(1):--channel
5828 option, and if the event rule to create is the first in its
5829 <<domain,tracing domain>> for a given tracing session, then LTTng
5830 creates a _default channel_ for you. This default channel is reused in
5831 subsequent invocations of the man:lttng-enable-event(1) command for the
5832 same tracing domain.
5834 An event rule is always enabled at creation time.
5836 The following examples show how you can combine the previous
5837 command-line options to create simple to more complex event rules.
5839 .Create an event rule targetting a Linux kernel tracepoint (default channel).
5843 $ lttng enable-event --kernel sched_switch
5847 .Create an event rule matching four Linux kernel system calls (default channel).
5851 $ lttng enable-event --kernel --syscall open,write,read,close
5855 .Create event rules matching tracepoints with filter expressions (default channel).
5859 $ lttng enable-event --kernel sched_switch --filter='prev_comm == "bash"'
5864 $ lttng enable-event --kernel --all \
5865 --filter='$ctx.tid == 1988 || $ctx.tid == 1534'
5870 $ lttng enable-event --jul my_logger \
5871 --filter='$app.retriever:cur_msg_id > 3'
5874 IMPORTANT: Make sure to always quote the filter string when you
5875 use man:lttng(1) from a shell.
5878 .Create an event rule matching any user space tracepoint of a given tracepoint provider with a log level range (default channel).
5882 $ lttng enable-event --userspace my_app:'*' --loglevel=TRACE_INFO
5885 IMPORTANT: Make sure to always quote the wildcard character when you
5886 use man:lttng(1) from a shell.
5889 .Create an event rule matching multiple Python loggers with a wildcard and with exclusions (default channel).
5893 $ lttng enable-event --python my-app.'*' \
5894 --exclude='my-app.module,my-app.hello'
5898 .Create an event rule matching any Apache log4j logger with a specific log level (default channel).
5902 $ lttng enable-event --log4j --all --loglevel-only=LOG4J_WARN
5906 .Create an event rule attached to a specific channel matching a specific user space tracepoint provider and tracepoint.
5910 $ lttng enable-event --userspace my_app:my_tracepoint --channel=my-channel
5914 .Create an event rule matching the `malloc` function entry in path:{/usr/lib/libc.so.6}:
5918 $ lttng enable-event --kernel --userspace-probe=/usr/lib/libc.so.6:malloc \
5923 .Create an event rule matching the `server`/`accept_request` https://www.sourceware.org/systemtap/wiki/AddingUserSpaceProbingToApps[USDT probe] in path:{/usr/bin/serv}:
5927 $ lttng enable-event --kernel --userspace-probe=sdt:serv:server:accept_request \
5928 server_accept_request
5932 The event rules of a given channel form a whitelist: as soon as an
5933 emitted event passes one of them, LTTng can record the event. For
5934 example, an event named `my_app:my_tracepoint` emitted from a user space
5935 tracepoint with a `TRACE_ERROR` log level passes both of the following
5940 $ lttng enable-event --userspace my_app:my_tracepoint
5941 $ lttng enable-event --userspace my_app:my_tracepoint \
5942 --loglevel=TRACE_INFO
5945 The second event rule is redundant: the first one includes
5949 [[disable-event-rule]]
5950 === Disable an event rule
5952 To disable an event rule that you <<enabling-disabling-events,created>>
5953 previously, use the man:lttng-disable-event(1) command. This command
5954 disables _all_ the event rules (of a given tracing domain and channel)
5955 which match an instrumentation point. The other conditions are not
5956 supported as of LTTng{nbsp}{revision}.
5958 The LTTng tracer does not record an emitted event which passes
5959 a _disabled_ event rule.
5961 .Disable an event rule matching a Python logger (default channel).
5965 $ lttng disable-event --python my-logger
5969 .Disable an event rule matching all `java.util.logging` loggers (default channel).
5973 $ lttng disable-event --jul '*'
5977 .Disable _all_ the event rules of the default channel.
5979 The opt:lttng-disable-event(1):--all-events option is not, like the
5980 opt:lttng-enable-event(1):--all option of man:lttng-enable-event(1), the
5981 equivalent of the event name `*` (wildcard): it disables _all_ the event
5982 rules of a given channel.
5986 $ lttng disable-event --jul --all-events
5990 NOTE: You cannot delete an event rule once you create it.
5994 === Get the status of a tracing session
5996 To get the status of the current tracing session, that is, its
5997 parameters, its channels, event rules, and their attributes:
5999 * Use the man:lttng-status(1) command:
6009 To get the status of any tracing session:
6011 * Use the man:lttng-list(1) command with the tracing session's name:
6016 $ lttng list my-session
6020 Replace `my-session` with the desired tracing session's name.
6023 [[basic-tracing-session-control]]
6024 === Start and stop a tracing session
6026 Once you <<creating-destroying-tracing-sessions,create a tracing
6028 <<enabling-disabling-events,create one or more event rules>>,
6029 you can start and stop the tracers for this tracing session.
6031 To start tracing in the current tracing session:
6033 * Use the man:lttng-start(1) command:
6042 LTTng is very flexible: you can launch user applications before
6043 or after the you start the tracers. The tracers only record the events
6044 if they pass enabled event rules and if they occur while the tracers are
6047 To stop tracing in the current tracing session:
6049 * Use the man:lttng-stop(1) command:
6058 If there were <<channel-overwrite-mode-vs-discard-mode,lost event
6059 records>> or lost sub-buffers since the last time you ran
6060 man:lttng-start(1), warnings are printed when you run the
6061 man:lttng-stop(1) command.
6063 IMPORTANT: You need to stop tracing to make LTTng flush the remaining
6064 trace data and make the trace readable. Note that the
6065 man:lttng-destroy(1) command (see
6066 <<creating-destroying-tracing-sessions,Create and destroy a tracing
6067 session>>) also runs the man:lttng-stop(1) command implicitly.
6070 [[enabling-disabling-channels]]
6071 === Create a channel
6073 Once you create a tracing session, you can create a <<channel,channel>>
6074 with the man:lttng-enable-channel(1) command.
6076 Note that LTTng automatically creates a default channel when, for a
6077 given <<domain,tracing domain>>, no channels exist and you
6078 <<enabling-disabling-events,create>> the first event rule. This default
6079 channel is named `channel0` and its attributes are set to reasonable
6080 values. Therefore, you only need to create a channel when you need
6081 non-default attributes.
6083 You specify each non-default channel attribute with a command-line
6084 option when you use the man:lttng-enable-channel(1) command. The
6085 available command-line options are:
6087 [role="growable",cols="asciidoc,asciidoc"]
6088 .Command-line options for the man:lttng-enable-channel(1) command.
6090 |Option |Description
6096 <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> instead
6097 of the default _discard_ mode.
6099 |`--buffers-pid` (user space tracing domain only)
6102 Use the per-process <<channel-buffering-schemes,buffering scheme>>
6103 instead of the default per-user buffering scheme.
6105 |+--subbuf-size=__SIZE__+
6108 Allocate sub-buffers of +__SIZE__+ bytes (power of two), for each CPU,
6109 either for each Unix user (default), or for each instrumented process.
6111 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6113 |+--num-subbuf=__COUNT__+
6116 Allocate +__COUNT__+ sub-buffers (power of two), for each CPU, either
6117 for each Unix user (default), or for each instrumented process.
6119 See <<channel-subbuf-size-vs-subbuf-count,Sub-buffer count and size>>.
6121 |+--tracefile-size=__SIZE__+
6124 Set the maximum size of each trace file that this channel writes within
6125 a stream to +__SIZE__+ bytes instead of no maximum.
6127 See <<tracefile-rotation,Trace file count and size>>.
6129 |+--tracefile-count=__COUNT__+
6132 Limit the number of trace files that this channel creates to
6133 +__COUNT__+ channels instead of no limit.
6135 See <<tracefile-rotation,Trace file count and size>>.
6137 |+--switch-timer=__PERIODUS__+
6140 Set the <<channel-switch-timer,switch timer period>>
6141 to +__PERIODUS__+{nbsp}µs.
6143 |+--read-timer=__PERIODUS__+
6146 Set the <<channel-read-timer,read timer period>>
6147 to +__PERIODUS__+{nbsp}µs.
6149 |[[opt-blocking-timeout]]+--blocking-timeout=__TIMEOUTUS__+
6152 Set the timeout of user space applications which load LTTng-UST
6153 in blocking mode to +__TIMEOUTUS__+:
6156 Never block (non-blocking mode).
6159 Block forever until space is available in a sub-buffer to record
6162 __n__, a positive value::
6163 Wait for at most __n__ µs when trying to write into a sub-buffer.
6165 Note that, for this option to have any effect on an instrumented
6166 user space application, you need to run the application with a set
6167 env:LTTNG_UST_ALLOW_BLOCKING environment variable.
6169 |+--output=__TYPE__+ (Linux kernel tracing domain only)
6172 Set the channel's output type to +__TYPE__+, either `mmap` or `splice`.
6176 You can only create a channel in the Linux kernel and user space
6177 <<domain,tracing domains>>: other tracing domains have their own channel
6178 created on the fly when <<enabling-disabling-events,creating event
6183 Because of a current LTTng limitation, you must create all channels
6184 _before_ you <<basic-tracing-session-control,start tracing>> in a given
6185 tracing session, that is, before the first time you run
6188 Since LTTng automatically creates a default channel when you use the
6189 man:lttng-enable-event(1) command with a specific tracing domain, you
6190 cannot, for example, create a Linux kernel event rule, start tracing,
6191 and then create a user space event rule, because no user space channel
6192 exists yet and it's too late to create one.
6194 For this reason, make sure to configure your channels properly
6195 before starting the tracers for the first time!
6198 The following examples show how you can combine the previous
6199 command-line options to create simple to more complex channels.
6201 .Create a Linux kernel channel with default attributes.
6205 $ lttng enable-channel --kernel my-channel
6209 .Create a user space channel with four sub-buffers or 1{nbsp}MiB each, per CPU, per instrumented process.
6213 $ lttng enable-channel --userspace --num-subbuf=4 --subbuf-size=1M \
6214 --buffers-pid my-channel
6218 .[[blocking-timeout-example]]Create a default user space channel with an infinite blocking timeout.
6220 <<creating-destroying-tracing-sessions,Create a tracing-session>>,
6221 create the channel, <<enabling-disabling-events,create an event rule>>,
6222 and <<basic-tracing-session-control,start tracing>>:
6227 $ lttng enable-channel --userspace --blocking-timeout=inf blocking-channel
6228 $ lttng enable-event --userspace --channel=blocking-channel --all
6232 Run an application instrumented with LTTng-UST and allow it to block:
6236 $ LTTNG_UST_ALLOW_BLOCKING=1 my-app
6240 .Create a Linux kernel channel which rotates eight trace files of 4{nbsp}MiB each for each stream
6244 $ lttng enable-channel --kernel --tracefile-count=8 \
6245 --tracefile-size=4194304 my-channel
6249 .Create a user space channel in overwrite (or _flight recorder_) mode.
6253 $ lttng enable-channel --userspace --overwrite my-channel
6257 You can <<enabling-disabling-events,create>> the same event rule in
6258 two different channels:
6262 $ lttng enable-event --userspace --channel=my-channel app:tp
6263 $ lttng enable-event --userspace --channel=other-channel app:tp
6266 If both channels are enabled, when a tracepoint named `app:tp` is
6267 reached, LTTng records two events, one for each channel.
6271 === Disable a channel
6273 To disable a specific channel that you <<enabling-disabling-channels,created>>
6274 previously, use the man:lttng-disable-channel(1) command.
6276 .Disable a specific Linux kernel channel.
6280 $ lttng disable-channel --kernel my-channel
6284 The state of a channel precedes the individual states of event rules
6285 attached to it: event rules which belong to a disabled channel, even if
6286 they are enabled, are also considered disabled.
6290 === Add context fields to a channel
6292 Event record fields in trace files provide important information about
6293 events that occured previously, but sometimes some external context may
6294 help you solve a problem faster. Examples of context fields are:
6296 * The **process ID**, **thread ID**, **process name**, and
6297 **process priority** of the thread in which the event occurs.
6298 * The **hostname** of the system on which the event occurs.
6299 * The Linux kernel and user call stacks (since
6301 * The current values of many possible **performance counters** using
6303 ** CPU cycles, stalled cycles, idle cycles, and the other cycle types.
6305 ** Branch instructions, misses, and loads.
6307 * Any context defined at the application level (supported for the
6308 JUL and log4j <<domain,tracing domains>>).
6310 To get the full list of available context fields, see
6311 `lttng add-context --list`. Some context fields are reserved for a
6312 specific <<domain,tracing domain>> (Linux kernel or user space).
6314 You add context fields to <<channel,channels>>. All the events
6315 that a channel with added context fields records contain those fields.
6317 To add context fields to one or all the channels of a given tracing
6320 * Use the man:lttng-add-context(1) command.
6322 .Add context fields to all the channels of the current tracing session.
6324 The following command line adds the virtual process identifier and
6325 the per-thread CPU cycles count fields to all the user space channels
6326 of the current tracing session.
6330 $ lttng add-context --userspace --type=vpid --type=perf:thread:cpu-cycles
6334 .Add performance counter context fields by raw ID
6336 See man:lttng-add-context(1) for the exact format of the context field
6337 type, which is partly compatible with the format used in
6342 $ lttng add-context --userspace --type=perf:thread:raw:r0110:test
6343 $ lttng add-context --kernel --type=perf:cpu:raw:r0013c:x86unhalted
6347 .Add context fields to a specific channel.
6349 The following command line adds the thread identifier and user call
6350 stack context fields to the Linux kernel channel named `my-channel` in
6351 the current tracing session.
6355 $ lttng add-context --kernel --channel=my-channel \
6356 --type=tid --type=callstack-user
6360 .Add an application-specific context field to a specific channel.
6362 The following command line adds the `cur_msg_id` context field of the
6363 `retriever` context retriever for all the instrumented
6364 <<java-application,Java applications>> recording <<event,event records>>
6365 in the channel named `my-channel`:
6369 $ lttng add-context --kernel --channel=my-channel \
6370 --type='$app:retriever:cur_msg_id'
6373 IMPORTANT: Make sure to always quote the `$` character when you
6374 use man:lttng-add-context(1) from a shell.
6377 NOTE: You cannot remove context fields from a channel once you add it.
6382 === Track process IDs
6384 It's often useful to allow only specific process IDs (PIDs) to emit
6385 events. For example, you may wish to record all the system calls made by
6386 a given process (Ă la http://linux.die.net/man/1/strace[strace]).
6388 The man:lttng-track(1) and man:lttng-untrack(1) commands serve this
6389 purpose. Both commands operate on a whitelist of process IDs. You _add_
6390 entries to this whitelist with the man:lttng-track(1) command and remove
6391 entries with the man:lttng-untrack(1) command. Any process which has one
6392 of the PIDs in the whitelist is allowed to emit LTTng events which pass
6393 an enabled <<event,event rule>>.
6395 NOTE: The PID tracker tracks the _numeric process IDs_. Should a
6396 process with a given tracked ID exit and another process be given this
6397 ID, then the latter would also be allowed to emit events.
6399 .Track and untrack process IDs.
6401 For the sake of the following example, assume the target system has
6402 16{nbsp}possible PIDs.
6405 <<creating-destroying-tracing-sessions,create a tracing session>>,
6406 the whitelist contains all the possible PIDs:
6409 .All PIDs are tracked.
6410 image::track-all.png[]
6412 When the whitelist is full and you use the man:lttng-track(1) command to
6413 specify some PIDs to track, LTTng first clears the whitelist, then it
6414 tracks the specific PIDs. After:
6418 $ lttng track --pid=3,4,7,10,13
6424 .PIDs 3, 4, 7, 10, and 13 are tracked.
6425 image::track-3-4-7-10-13.png[]
6427 You can add more PIDs to the whitelist afterwards:
6431 $ lttng track --pid=1,15,16
6437 .PIDs 1, 15, and 16 are added to the whitelist.
6438 image::track-1-3-4-7-10-13-15-16.png[]
6440 The man:lttng-untrack(1) command removes entries from the PID tracker's
6441 whitelist. Given the previous example, the following command:
6445 $ lttng untrack --pid=3,7,10,13
6448 leads to this whitelist:
6451 .PIDs 3, 7, 10, and 13 are removed from the whitelist.
6452 image::track-1-4-15-16.png[]
6454 LTTng can track all possible PIDs again using the
6455 opt:lttng-track(1):--all option:
6459 $ lttng track --pid --all
6462 The result is, again:
6465 .All PIDs are tracked.
6466 image::track-all.png[]
6469 .Track only specific PIDs
6471 A very typical use case with PID tracking is to start with an empty
6472 whitelist, then <<basic-tracing-session-control,start the tracers>>, and
6473 then add PIDs manually while tracers are active. You can accomplish this
6474 by using the opt:lttng-untrack(1):--all option of the
6475 man:lttng-untrack(1) command to clear the whitelist after you
6476 <<creating-destroying-tracing-sessions,create a tracing session>>:
6480 $ lttng untrack --pid --all
6486 .No PIDs are tracked.
6487 image::untrack-all.png[]
6489 If you trace with this whitelist configuration, the tracer records no
6490 events for this <<domain,tracing domain>> because no processes are
6491 tracked. You can use the man:lttng-track(1) command as usual to track
6492 specific PIDs, for example:
6496 $ lttng track --pid=6,11
6502 .PIDs 6 and 11 are tracked.
6503 image::track-6-11.png[]
6508 [[saving-loading-tracing-session]]
6509 === Save and load tracing session configurations
6511 Configuring a <<tracing-session,tracing session>> can be long. Some of
6512 the tasks involved are:
6514 * <<enabling-disabling-channels,Create channels>> with
6515 specific attributes.
6516 * <<adding-context,Add context fields>> to specific channels.
6517 * <<enabling-disabling-events,Create event rules>> with specific log
6518 level and filter conditions.
6520 If you use LTTng to solve real world problems, chances are you have to
6521 record events using the same tracing session setup over and over,
6522 modifying a few variables each time in your instrumented program
6523 or environment. To avoid constant tracing session reconfiguration,
6524 the man:lttng(1) command-line tool can save and load tracing session
6525 configurations to/from XML files.
6527 To save a given tracing session configuration:
6529 * Use the man:lttng-save(1) command:
6534 $ lttng save my-session
6538 Replace `my-session` with the name of the tracing session to save.
6540 LTTng saves tracing session configurations to
6541 dir:{$LTTNG_HOME/.lttng/sessions} by default. Note that the
6542 env:LTTNG_HOME environment variable defaults to `$HOME` if not set. Use
6543 the opt:lttng-save(1):--output-path option to change this destination
6546 LTTng saves all configuration parameters, for example:
6548 * The tracing session name.
6549 * The trace data output path.
6550 * The channels with their state and all their attributes.
6551 * The context fields you added to channels.
6552 * The event rules with their state, log level and filter conditions.
6554 To load a tracing session:
6556 * Use the man:lttng-load(1) command:
6561 $ lttng load my-session
6565 Replace `my-session` with the name of the tracing session to load.
6567 When LTTng loads a configuration, it restores your saved tracing session
6568 as if you just configured it manually.
6570 See man:lttng-load(1) for the complete list of command-line options. You
6571 can also save and load many sessions at a time, and decide in which
6572 directory to output the XML files.
6575 [[sending-trace-data-over-the-network]]
6576 === Send trace data over the network
6578 LTTng can send the recorded trace data to a remote system over the
6579 network instead of writing it to the local file system.
6581 To send the trace data over the network:
6583 . On the _remote_ system (which can also be the target system),
6584 start an LTTng <<lttng-relayd,relay daemon>> (man:lttng-relayd(8)):
6593 . On the _target_ system, create a tracing session configured to
6594 send trace data over the network:
6599 $ lttng create my-session --set-url=net://remote-system
6603 Replace `remote-system` by the host name or IP address of the
6604 remote system. See man:lttng-create(1) for the exact URL format.
6606 . On the target system, use the man:lttng(1) command-line tool as usual.
6607 When tracing is active, the target's consumer daemon sends sub-buffers
6608 to the relay daemon running on the remote system instead of flushing
6609 them to the local file system. The relay daemon writes the received
6610 packets to the local file system.
6612 The relay daemon writes trace files to
6613 +$LTTNG_HOME/lttng-traces/__hostname__/__session__+ by default, where
6614 +__hostname__+ is the host name of the target system and +__session__+
6615 is the tracing session name. Note that the env:LTTNG_HOME environment
6616 variable defaults to `$HOME` if not set. Use the
6617 opt:lttng-relayd(8):--output option of man:lttng-relayd(8) to write
6618 trace files to another base directory.
6623 === View events as LTTng emits them (noch:{LTTng} live)
6625 LTTng live is a network protocol implemented by the <<lttng-relayd,relay
6626 daemon>> (man:lttng-relayd(8)) to allow compatible trace viewers to
6627 display events as LTTng emits them on the target system while tracing is
6630 The relay daemon creates a _tee_: it forwards the trace data to both
6631 the local file system and to connected live viewers:
6634 .The relay daemon creates a _tee_, forwarding the trace data to both trace files and a connected live viewer.
6639 . On the _target system_, create a <<tracing-session,tracing session>>
6645 $ lttng create my-session --live
6649 This spawns a local relay daemon.
6651 . Start the live viewer and configure it to connect to the relay
6652 daemon. For example, with http://diamon.org/babeltrace[Babeltrace]:
6657 $ babeltrace --input-format=lttng-live \
6658 net://localhost/host/hostname/my-session
6665 * `hostname` with the host name of the target system.
6666 * `my-session` with the name of the tracing session to view.
6669 . Configure the tracing session as usual with the man:lttng(1)
6670 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6672 You can list the available live tracing sessions with Babeltrace:
6676 $ babeltrace --input-format=lttng-live net://localhost
6679 You can start the relay daemon on another system. In this case, you need
6680 to specify the relay daemon's URL when you create the tracing session
6681 with the opt:lttng-create(1):--set-url option. You also need to replace
6682 `localhost` in the procedure above with the host name of the system on
6683 which the relay daemon is running.
6685 See man:lttng-create(1) and man:lttng-relayd(8) for the complete list of
6686 command-line options.
6690 [[taking-a-snapshot]]
6691 === Take a snapshot of the current sub-buffers of a tracing session
6693 The normal behavior of LTTng is to append full sub-buffers to growing
6694 trace data files. This is ideal to keep a full history of the events
6695 that occurred on the target system, but it can
6696 represent too much data in some situations. For example, you may wish
6697 to trace your application continuously until some critical situation
6698 happens, in which case you only need the latest few recorded
6699 events to perform the desired analysis, not multi-gigabyte trace files.
6701 With the man:lttng-snapshot(1) command, you can take a snapshot of the
6702 current sub-buffers of a given <<tracing-session,tracing session>>.
6703 LTTng can write the snapshot to the local file system or send it over
6707 .A snapshot is a copy of the current sub-buffers, which are not cleared after the operation.
6708 image::snapshot.png[]
6710 If you wish to create unmanaged, self-contained, non-overlapping
6711 trace chunk archives instead of a simple copy of the current
6712 sub-buffers, see the <<session-rotation,tracing session rotation>>
6713 feature (available since LTTng{nbsp}2.11).
6717 . Create a tracing session in _snapshot mode_:
6722 $ lttng create my-session --snapshot
6726 The <<channel-overwrite-mode-vs-discard-mode,event record loss mode>> of
6727 <<channel,channels>> created in this mode is automatically set to
6728 _overwrite_ (flight recorder mode).
6730 . Configure the tracing session as usual with the man:lttng(1)
6731 command-line tool, and <<basic-tracing-session-control,start tracing>>.
6733 . **Optional**: When you need to take a snapshot,
6734 <<basic-tracing-session-control,stop tracing>>.
6736 You can take a snapshot when the tracers are active, but if you stop
6737 them first, you are sure that the data in the sub-buffers does not
6738 change before you actually take the snapshot.
6745 $ lttng snapshot record --name=my-first-snapshot
6749 LTTng writes the current sub-buffers of all the current tracing
6750 session's channels to trace files on the local file system. Those trace
6751 files have `my-first-snapshot` in their name.
6753 There is no difference between the format of a normal trace file and the
6754 format of a snapshot: viewers of LTTng traces also support LTTng
6757 By default, LTTng writes snapshot files to the path shown by
6758 `lttng snapshot list-output`. You can change this path or decide to send
6759 snapshots over the network using either:
6761 . An output path or URL that you specify when you
6762 <<creating-destroying-tracing-sessions,create the tracing session>>.
6763 . A snapshot output path or URL that you add using
6764 `lttng snapshot add-output`.
6765 . An output path or URL that you provide directly to the
6766 `lttng snapshot record` command.
6768 Method{nbsp}3 overrides method{nbsp}2, which overrides method 1. When
6769 you specify a URL, a relay daemon must listen on a remote system (see
6770 <<sending-trace-data-over-the-network,Send trace data over the
6775 [[session-rotation]]
6776 === Archive the current trace chunk (rotate a tracing session)
6778 The <<taking-a-snapshot,snapshot user guide>> shows how you can dump
6779 a tracing session's current sub-buffers to the file system or send them
6780 over the network. When you take a snapshot, LTTng does not clear the
6781 tracing session's ring buffers: if you take another snapshot immediately
6782 after, both snapshots could contain overlapping trace data.
6784 Inspired by https://en.wikipedia.org/wiki/Log_rotation[log rotation],
6785 _tracing session rotation_ is a feature which appends the content of the
6786 ring buffers to what's already on the file system or sent over the
6787 network since the tracing session's creation or since the last
6788 rotation, and then clears those ring buffers to avoid trace data
6791 What LTTng is about to write when performing a tracing session rotation
6792 is called the _current trace chunk_. When this current trace chunk is
6793 written to the file system or sent over the network, it becomes a _trace
6794 chunk archive_. Therefore, a tracing session rotation _archives_ the
6795 current trace chunk.
6798 .A tracing session rotation operation _archives_ the current trace chunk.
6799 image::rotation.png[]
6801 A trace chunk archive is a self-contained LTTng trace which LTTng
6802 doesn't manage anymore: you can read it, modify it, move it, or remove
6805 There are two methods to perform a tracing session rotation: immediately
6806 or with a rotation schedule.
6808 To perform an immediate tracing session rotation:
6810 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6811 in _normal mode_ or _network streaming mode_
6812 (only those two creation modes support tracing session rotation):
6817 $ lttng create my-session
6821 . <<enabling-disabling-events,Create one or more event rules>>
6822 and <<basic-tracing-session-control,start tracing>>:
6827 $ lttng enable-event --kernel sched_'*'
6832 . When needed, immediately rotate the current tracing session:
6841 The cmd:lttng-rotate command prints the path to the created trace
6842 chunk archive. See man:lttng-rotate(1) to learn about the format
6843 of trace chunk archive directory names.
6845 You can perform other immediate rotations while the tracing session is
6846 active. It is guaranteed that all the trace chunk archives do not
6847 contain overlapping trace data. You can also perform an immediate
6848 rotation once you have <<basic-tracing-session-control,stopped>> the
6851 . When you are done tracing,
6852 <<creating-destroying-tracing-sessions,destroy the current tracing
6862 The tracing session destruction operation creates one last trace
6863 chunk archive from the current trace chunk.
6865 A tracing session rotation schedule is a planned rotation which LTTng
6866 performs automatically based on one of the following conditions:
6868 * A timer with a configured period times out.
6870 * The total size of the flushed part of the current trace chunk
6871 becomes greater than or equal to a configured value.
6873 To schedule a tracing session rotation, set a _rotation schedule_:
6875 . <<creating-destroying-tracing-sessions,Create a tracing session>>
6876 in _normal mode_ or _network streaming mode_
6877 (only those two creation modes support tracing session rotation):
6882 $ lttng create my-session
6886 . <<enabling-disabling-events,Create one or more event rules>>:
6891 $ lttng enable-event --kernel sched_'*'
6895 . Set a tracing session rotation schedule:
6900 $ lttng enable-rotation --timer=10s
6904 In this example, we set a rotation schedule so that LTTng performs a
6905 tracing session rotation every ten seconds.
6907 See man:lttng-enable-rotation(1) to learn more about other ways to set a
6910 . <<basic-tracing-session-control,Start tracing>>:
6919 LTTng performs tracing session rotations automatically while the tracing
6920 session is active thanks to the rotation schedule.
6922 . When you are done tracing,
6923 <<creating-destroying-tracing-sessions,destroy the current tracing
6933 The tracing session destruction operation creates one last trace chunk
6934 archive from the current trace chunk.
6936 You can use man:lttng-disable-rotation(1) to unset a tracing session
6939 NOTE: man:lttng-rotate(1) and man:lttng-enable-rotation(1) list
6940 limitations regarding those two commands.
6945 === Use the machine interface
6947 With any command of the man:lttng(1) command-line tool, you can set the
6948 opt:lttng(1):--mi option to `xml` (before the command name) to get an
6949 XML machine interface output, for example:
6953 $ lttng --mi=xml enable-event --kernel --syscall open
6956 A schema definition (XSD) is
6957 https://github.com/lttng/lttng-tools/blob/stable-2.11/src/common/mi-lttng-3.0.xsd[available]
6958 to ease the integration with external tools as much as possible.
6962 [[metadata-regenerate]]
6963 === Regenerate the metadata of an LTTng trace
6965 An LTTng trace, which is a http://diamon.org/ctf[CTF] trace, has both
6966 data stream files and a metadata file. This metadata file contains,
6967 amongst other things, information about the offset of the clock sources
6968 used to timestamp <<event,event records>> when tracing.
6970 If, once a <<tracing-session,tracing session>> is
6971 <<basic-tracing-session-control,started>>, a major
6972 https://en.wikipedia.org/wiki/Network_Time_Protocol[NTP] correction
6973 happens, the trace's clock offset also needs to be updated. You
6974 can use the `metadata` item of the man:lttng-regenerate(1) command
6977 The main use case of this command is to allow a system to boot with
6978 an incorrect wall time and trace it with LTTng before its wall time
6979 is corrected. Once the system is known to be in a state where its
6980 wall time is correct, it can run `lttng regenerate metadata`.
6982 To regenerate the metadata of an LTTng trace:
6984 * Use the `metadata` item of the man:lttng-regenerate(1) command:
6989 $ lttng regenerate metadata
6995 `lttng regenerate metadata` has the following limitations:
6997 * Tracing session <<creating-destroying-tracing-sessions,created>>
6999 * User space <<channel,channels>>, if any, are using
7000 <<channel-buffering-schemes,per-user buffering>>.
7005 [[regenerate-statedump]]
7006 === Regenerate the state dump of a tracing session
7008 The LTTng kernel and user space tracers generate state dump
7009 <<event,event records>> when the application starts or when you
7010 <<basic-tracing-session-control,start a tracing session>>. An analysis
7011 can use the state dump event records to set an initial state before it
7012 builds the rest of the state from the following event records.
7013 http://tracecompass.org/[Trace Compass] is a notable example of an
7014 application which uses the state dump of an LTTng trace.
7016 When you <<taking-a-snapshot,take a snapshot>>, it's possible that the
7017 state dump event records are not included in the snapshot because they
7018 were recorded to a sub-buffer that has been consumed or overwritten
7021 You can use the `lttng regenerate statedump` command to emit the state
7022 dump event records again.
7024 To regenerate the state dump of the current tracing session, provided
7025 create it in snapshot mode, before you take a snapshot:
7027 . Use the `statedump` item of the man:lttng-regenerate(1) command:
7032 $ lttng regenerate statedump
7036 . <<basic-tracing-session-control,Stop the tracing session>>:
7045 . <<taking-a-snapshot,Take a snapshot>>:
7050 $ lttng snapshot record --name=my-snapshot
7054 Depending on the event throughput, you should run steps 1 and 2
7055 as closely as possible.
7057 NOTE: To record the state dump events, you need to
7058 <<enabling-disabling-events,create event rules>> which enable them.
7059 LTTng-UST state dump tracepoints start with `lttng_ust_statedump:`.
7060 LTTng-modules state dump tracepoints start with `lttng_statedump_`.
7064 [[persistent-memory-file-systems]]
7065 === Record trace data on persistent memory file systems
7067 https://en.wikipedia.org/wiki/Non-volatile_random-access_memory[Non-volatile random-access memory]
7068 (NVRAM) is random-access memory that retains its information when power
7069 is turned off (non-volatile). Systems with such memory can store data
7070 structures in RAM and retrieve them after a reboot, without flushing
7071 to typical _storage_.
7073 Linux supports NVRAM file systems thanks to either
7074 http://pramfs.sourceforge.net/[PRAMFS] or
7075 https://www.kernel.org/doc/Documentation/filesystems/dax.txt[DAX]{nbsp}+{nbsp}http://lkml.iu.edu/hypermail/linux/kernel/1504.1/03463.html[pmem]
7076 (requires Linux{nbsp}4.1+).
7078 This section does not describe how to operate such file systems;
7079 we assume that you have a working persistent memory file system.
7081 When you create a <<tracing-session,tracing session>>, you can specify
7082 the path of the shared memory holding the sub-buffers. If you specify a
7083 location on an NVRAM file system, then you can retrieve the latest
7084 recorded trace data when the system reboots after a crash.
7086 To record trace data on a persistent memory file system and retrieve the
7087 trace data after a system crash:
7089 . Create a tracing session with a sub-buffer shared memory path located
7090 on an NVRAM file system:
7095 $ lttng create my-session --shm-path=/path/to/shm
7099 . Configure the tracing session as usual with the man:lttng(1)
7100 command-line tool, and <<basic-tracing-session-control,start tracing>>.
7102 . After a system crash, use the man:lttng-crash(1) command-line tool to
7103 view the trace data recorded on the NVRAM file system:
7108 $ lttng-crash /path/to/shm
7112 The binary layout of the ring buffer files is not exactly the same as
7113 the trace files layout. This is why you need to use man:lttng-crash(1)
7114 instead of your preferred trace viewer directly.
7116 To convert the ring buffer files to LTTng trace files:
7118 * Use the opt:lttng-crash(1):--extract option of man:lttng-crash(1):
7123 $ lttng-crash --extract=/path/to/trace /path/to/shm
7129 [[notif-trigger-api]]
7130 === Get notified when a channel's buffer usage is too high or too low
7132 With LTTng's $$C/C++$$ notification and trigger API, your user
7133 application can get notified when the buffer usage of one or more
7134 <<channel,channels>> becomes too low or too high. You can use this API
7135 and enable or disable <<event,event rules>> during tracing to avoid
7136 <<channel-overwrite-mode-vs-discard-mode,discarded event records>>.
7138 .Have a user application get notified when an LTTng channel's buffer usage is too high.
7140 In this example, we create and build an application which gets notified
7141 when the buffer usage of a specific LTTng channel is higher than
7142 75{nbsp}%. We only print that it is the case in the example, but we
7143 could as well use the API of <<liblttng-ctl-lttng,`liblttng-ctl`>> to
7144 disable event rules when this happens.
7146 . Create the application's C source file:
7154 #include <lttng/domain.h>
7155 #include <lttng/action/action.h>
7156 #include <lttng/action/notify.h>
7157 #include <lttng/condition/condition.h>
7158 #include <lttng/condition/buffer-usage.h>
7159 #include <lttng/condition/evaluation.h>
7160 #include <lttng/notification/channel.h>
7161 #include <lttng/notification/notification.h>
7162 #include <lttng/trigger/trigger.h>
7163 #include <lttng/endpoint.h>
7165 int main(int argc, char *argv[])
7167 int exit_status = 0;
7168 struct lttng_notification_channel *notification_channel;
7169 struct lttng_condition *condition;
7170 struct lttng_action *action;
7171 struct lttng_trigger *trigger;
7172 const char *tracing_session_name;
7173 const char *channel_name;
7176 tracing_session_name = argv[1];
7177 channel_name = argv[2];
7180 * Create a notification channel. A notification channel
7181 * connects the user application to the LTTng session daemon.
7182 * This notification channel can be used to listen to various
7183 * types of notifications.
7185 notification_channel = lttng_notification_channel_create(
7186 lttng_session_daemon_notification_endpoint);
7189 * Create a "high buffer usage" condition. In this case, the
7190 * condition is reached when the buffer usage is greater than or
7191 * equal to 75 %. We create the condition for a specific tracing
7192 * session name, channel name, and for the user space tracing
7195 * The "low buffer usage" condition type also exists.
7197 condition = lttng_condition_buffer_usage_high_create();
7198 lttng_condition_buffer_usage_set_threshold_ratio(condition, .75);
7199 lttng_condition_buffer_usage_set_session_name(
7200 condition, tracing_session_name);
7201 lttng_condition_buffer_usage_set_channel_name(condition,
7203 lttng_condition_buffer_usage_set_domain_type(condition,
7207 * Create an action (get a notification) to take when the
7208 * condition created above is reached.
7210 action = lttng_action_notify_create();
7213 * Create a trigger. A trigger associates a condition to an
7214 * action: the action is executed when the condition is reached.
7216 trigger = lttng_trigger_create(condition, action);
7218 /* Register the trigger to LTTng. */
7219 lttng_register_trigger(trigger);
7222 * Now that we have registered a trigger, a notification will be
7223 * emitted everytime its condition is met. To receive this
7224 * notification, we must subscribe to notifications that match
7225 * the same condition.
7227 lttng_notification_channel_subscribe(notification_channel,
7231 * Notification loop. You can put this in a dedicated thread to
7232 * avoid blocking the main thread.
7235 struct lttng_notification *notification;
7236 enum lttng_notification_channel_status status;
7237 const struct lttng_evaluation *notification_evaluation;
7238 const struct lttng_condition *notification_condition;
7239 double buffer_usage;
7241 /* Receive the next notification. */
7242 status = lttng_notification_channel_get_next_notification(
7243 notification_channel, ¬ification);
7246 case LTTNG_NOTIFICATION_CHANNEL_STATUS_OK:
7248 case LTTNG_NOTIFICATION_CHANNEL_STATUS_NOTIFICATIONS_DROPPED:
7250 * The session daemon can drop notifications if
7251 * a monitoring application is not consuming the
7252 * notifications fast enough.
7255 case LTTNG_NOTIFICATION_CHANNEL_STATUS_CLOSED:
7257 * The notification channel has been closed by the
7258 * session daemon. This is typically caused by a session
7259 * daemon shutting down.
7263 /* Unhandled conditions or errors. */
7269 * A notification provides, amongst other things:
7271 * * The condition that caused this notification to be
7273 * * The condition evaluation, which provides more
7274 * specific information on the evaluation of the
7277 * The condition evaluation provides the buffer usage
7278 * value at the moment the condition was reached.
7280 notification_condition = lttng_notification_get_condition(
7282 notification_evaluation = lttng_notification_get_evaluation(
7285 /* We're subscribed to only one condition. */
7286 assert(lttng_condition_get_type(notification_condition) ==
7287 LTTNG_CONDITION_TYPE_BUFFER_USAGE_HIGH);
7290 * Get the exact sampled buffer usage from the
7291 * condition evaluation.
7293 lttng_evaluation_buffer_usage_get_usage_ratio(
7294 notification_evaluation, &buffer_usage);
7297 * At this point, instead of printing a message, we
7298 * could do something to reduce the channel's buffer
7299 * usage, like disable specific events.
7301 printf("Buffer usage is %f %% in tracing session \"%s\", "
7302 "user space channel \"%s\".\n", buffer_usage * 100,
7303 tracing_session_name, channel_name);
7304 lttng_notification_destroy(notification);
7308 lttng_action_destroy(action);
7309 lttng_condition_destroy(condition);
7310 lttng_trigger_destroy(trigger);
7311 lttng_notification_channel_destroy(notification_channel);
7317 . Build the `notif-app` application, linking it to `liblttng-ctl`:
7322 $ gcc -o notif-app notif-app.c -llttng-ctl
7326 . <<creating-destroying-tracing-sessions,Create a tracing session>>,
7327 <<enabling-disabling-events,create an event rule>> matching all the
7328 user space tracepoints, and
7329 <<basic-tracing-session-control,start tracing>>:
7334 $ lttng create my-session
7335 $ lttng enable-event --userspace --all
7340 If you create the channel manually with the man:lttng-enable-channel(1)
7341 command, you can control how frequently are the current values of the
7342 channel's properties sampled to evaluate user conditions with the
7343 opt:lttng-enable-channel(1):--monitor-timer option.
7345 . Run the `notif-app` application. This program accepts the
7346 <<tracing-session,tracing session>> name and the user space channel
7347 name as its two first arguments. The channel which LTTng automatically
7348 creates with the man:lttng-enable-event(1) command above is named
7354 $ ./notif-app my-session channel0
7358 . In another terminal, run an application with a very high event
7359 throughput so that the 75{nbsp}% buffer usage condition is reached.
7361 In the first terminal, the application should print lines like this:
7364 Buffer usage is 81.45197 % in tracing session "my-session", user space
7368 If you don't see anything, try modifying the condition in
7369 path:{notif-app.c} to a lower value (0.1, for example), rebuilding it
7370 (step{nbsp}2) and running it again (step{nbsp}4).
7377 [[lttng-modules-ref]]
7378 === noch:{LTTng-modules}
7382 [[lttng-tracepoint-enum]]
7383 ==== `LTTNG_TRACEPOINT_ENUM()` usage
7385 Use the `LTTNG_TRACEPOINT_ENUM()` macro to define an enumeration:
7389 LTTNG_TRACEPOINT_ENUM(name, TP_ENUM_VALUES(entries))
7394 * `name` with the name of the enumeration (C identifier, unique
7395 amongst all the defined enumerations).
7396 * `entries` with a list of enumeration entries.
7398 The available enumeration entry macros are:
7400 +ctf_enum_value(__name__, __value__)+::
7401 Entry named +__name__+ mapped to the integral value +__value__+.
7403 +ctf_enum_range(__name__, __begin__, __end__)+::
7404 Entry named +__name__+ mapped to the range of integral values between
7405 +__begin__+ (included) and +__end__+ (included).
7407 +ctf_enum_auto(__name__)+::
7408 Entry named +__name__+ mapped to the integral value following the
7409 last mapping's value.
7411 The last value of a `ctf_enum_value()` entry is its +__value__+
7414 The last value of a `ctf_enum_range()` entry is its +__end__+ parameter.
7416 If `ctf_enum_auto()` is the first entry in the list, its integral
7419 Use the `ctf_enum()` <<lttng-modules-tp-fields,field definition macro>>
7420 to use a defined enumeration as a tracepoint field.
7422 .Define an enumeration with `LTTNG_TRACEPOINT_ENUM()`.
7426 LTTNG_TRACEPOINT_ENUM(
7429 ctf_enum_auto("AUTO: EXPECT 0")
7430 ctf_enum_value("VALUE: 23", 23)
7431 ctf_enum_value("VALUE: 27", 27)
7432 ctf_enum_auto("AUTO: EXPECT 28")
7433 ctf_enum_range("RANGE: 101 TO 303", 101, 303)
7434 ctf_enum_auto("AUTO: EXPECT 304")
7442 [[lttng-modules-tp-fields]]
7443 ==== Tracepoint fields macros (for `TP_FIELDS()`)
7445 [[tp-fast-assign]][[tp-struct-entry]]The available macros to define
7446 tracepoint fields, which must be listed within `TP_FIELDS()` in
7447 `LTTNG_TRACEPOINT_EVENT()`, are:
7449 [role="func-desc growable",cols="asciidoc,asciidoc"]
7450 .Available macros to define LTTng-modules tracepoint fields
7452 |Macro |Description and parameters
7455 +ctf_integer(__t__, __n__, __e__)+
7457 +ctf_integer_nowrite(__t__, __n__, __e__)+
7459 +ctf_user_integer(__t__, __n__, __e__)+
7461 +ctf_user_integer_nowrite(__t__, __n__, __e__)+
7463 Standard integer, displayed in base{nbsp}10.
7466 Integer C type (`int`, `long`, `size_t`, ...).
7472 Argument expression.
7475 +ctf_integer_hex(__t__, __n__, __e__)+
7477 +ctf_user_integer_hex(__t__, __n__, __e__)+
7479 Standard integer, displayed in base{nbsp}16.
7488 Argument expression.
7490 |+ctf_integer_oct(__t__, __n__, __e__)+
7492 Standard integer, displayed in base{nbsp}8.
7501 Argument expression.
7504 +ctf_integer_network(__t__, __n__, __e__)+
7506 +ctf_user_integer_network(__t__, __n__, __e__)+
7508 Integer in network byte order (big-endian), displayed in base{nbsp}10.
7517 Argument expression.
7520 +ctf_integer_network_hex(__t__, __n__, __e__)+
7522 +ctf_user_integer_network_hex(__t__, __n__, __e__)+
7524 Integer in network byte order, displayed in base{nbsp}16.
7533 Argument expression.
7536 +ctf_enum(__N__, __t__, __n__, __e__)+
7538 +ctf_enum_nowrite(__N__, __t__, __n__, __e__)+
7540 +ctf_user_enum(__N__, __t__, __n__, __e__)+
7542 +ctf_user_enum_nowrite(__N__, __t__, __n__, __e__)+
7547 Name of a <<lttng-tracepoint-enum,previously defined enumeration>>.
7550 Integer C type (`int`, `long`, `size_t`, ...).
7556 Argument expression.
7559 +ctf_string(__n__, __e__)+
7561 +ctf_string_nowrite(__n__, __e__)+
7563 +ctf_user_string(__n__, __e__)+
7565 +ctf_user_string_nowrite(__n__, __e__)+
7567 Null-terminated string; undefined behavior if +__e__+ is `NULL`.
7573 Argument expression.
7576 +ctf_array(__t__, __n__, __e__, __s__)+
7578 +ctf_array_nowrite(__t__, __n__, __e__, __s__)+
7580 +ctf_user_array(__t__, __n__, __e__, __s__)+
7582 +ctf_user_array_nowrite(__t__, __n__, __e__, __s__)+
7584 Statically-sized array of integers.
7587 Array element C type.
7593 Argument expression.
7599 +ctf_array_bitfield(__t__, __n__, __e__, __s__)+
7601 +ctf_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7603 +ctf_user_array_bitfield(__t__, __n__, __e__, __s__)+
7605 +ctf_user_array_bitfield_nowrite(__t__, __n__, __e__, __s__)+
7607 Statically-sized array of bits.
7609 The type of +__e__+ must be an integer type. +__s__+ is the number
7610 of elements of such type in +__e__+, not the number of bits.
7613 Array element C type.
7619 Argument expression.
7625 +ctf_array_text(__t__, __n__, __e__, __s__)+
7627 +ctf_array_text_nowrite(__t__, __n__, __e__, __s__)+
7629 +ctf_user_array_text(__t__, __n__, __e__, __s__)+
7631 +ctf_user_array_text_nowrite(__t__, __n__, __e__, __s__)+
7633 Statically-sized array, printed as text.
7635 The string does not need to be null-terminated.
7638 Array element C type (always `char`).
7644 Argument expression.
7650 +ctf_sequence(__t__, __n__, __e__, __T__, __E__)+
7652 +ctf_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7654 +ctf_user_sequence(__t__, __n__, __e__, __T__, __E__)+
7656 +ctf_user_sequence_nowrite(__t__, __n__, __e__, __T__, __E__)+
7658 Dynamically-sized array of integers.
7660 The type of +__E__+ must be unsigned.
7663 Array element C type.
7669 Argument expression.
7672 Length expression C type.
7678 +ctf_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7680 +ctf_user_sequence_hex(__t__, __n__, __e__, __T__, __E__)+
7682 Dynamically-sized array of integers, displayed in base{nbsp}16.
7684 The type of +__E__+ must be unsigned.
7687 Array element C type.
7693 Argument expression.
7696 Length expression C type.
7701 |+ctf_sequence_network(__t__, __n__, __e__, __T__, __E__)+
7703 Dynamically-sized array of integers in network byte order (big-endian),
7704 displayed in base{nbsp}10.
7706 The type of +__E__+ must be unsigned.
7709 Array element C type.
7715 Argument expression.
7718 Length expression C type.
7724 +ctf_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7726 +ctf_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7728 +ctf_user_sequence_bitfield(__t__, __n__, __e__, __T__, __E__)+
7730 +ctf_user_sequence_bitfield_nowrite(__t__, __n__, __e__, __T__, __E__)+
7732 Dynamically-sized array of bits.
7734 The type of +__e__+ must be an integer type. +__s__+ is the number
7735 of elements of such type in +__e__+, not the number of bits.
7737 The type of +__E__+ must be unsigned.
7740 Array element C type.
7746 Argument expression.
7749 Length expression C type.
7755 +ctf_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7757 +ctf_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7759 +ctf_user_sequence_text(__t__, __n__, __e__, __T__, __E__)+
7761 +ctf_user_sequence_text_nowrite(__t__, __n__, __e__, __T__, __E__)+
7763 Dynamically-sized array, displayed as text.
7765 The string does not need to be null-terminated.
7767 The type of +__E__+ must be unsigned.
7769 The behaviour is undefined if +__e__+ is `NULL`.
7772 Sequence element C type (always `char`).
7778 Argument expression.
7781 Length expression C type.
7787 Use the `_user` versions when the argument expression, `e`, is
7788 a user space address. In the cases of `ctf_user_integer*()` and
7789 `ctf_user_float*()`, `&e` must be a user space address, thus `e` must
7792 The `_nowrite` versions omit themselves from the session trace, but are
7793 otherwise identical. This means the `_nowrite` fields won't be written
7794 in the recorded trace. Their primary purpose is to make some
7795 of the event context available to the
7796 <<enabling-disabling-events,event filters>> without having to
7797 commit the data to sub-buffers.
7803 Terms related to LTTng and to tracing in general:
7806 The http://diamon.org/babeltrace[Babeltrace] project, which includes:
7808 * The cmd:babeltrace (Babeltrace{nbsp}1) or cmd:babeltrace2
7809 (Babeltrace{nbsp}2) command.
7810 * Libraries with a C{nbsp}API.
7811 * Python{nbsp}3 bindings.
7812 * Plugins (Babeltrace{nbsp}2).
7814 [[def-buffering-scheme]]<<channel-buffering-schemes,buffering scheme>>::
7815 A layout of <<def-sub-buffer,sub-buffers>> applied to a given channel.
7817 [[def-channel]]<<channel,channel>>::
7818 An entity which is responsible for a set of
7819 <<def-ring-buffer,ring buffers>>.
7821 <<def-event-rule,Event rules>> are always attached to a specific
7825 A source of time for a <<def-tracer,tracer>>.
7827 [[def-consumer-daemon]]<<lttng-consumerd,consumer daemon>>::
7828 A process which is responsible for consuming the full
7829 <<def-sub-buffer,sub-buffers>> and write them to a file system or
7830 send them over the network.
7832 [[def-current-trace-chunk]]current trace chunk::
7833 A <<def-trace-chunk,trace chunk>> which includes the current content
7834 of all the <<def-tracing-session-rotation,tracing session>>'s
7835 <<def-sub-buffer,sub-buffers>> and the stream files produced since the
7836 latest event amongst:
7838 * The creation of the <<def-tracing-session,tracing session>>.
7839 * The last tracing session rotation, if any.
7841 <<channel-overwrite-mode-vs-discard-mode,discard mode>>::
7842 The <<def-event-record-loss-mode,event record loss mode>> in which
7843 the <<def-tracer,tracer>> _discards_ new event records when there's no
7844 <<def-sub-buffer,sub-buffer>> space left to store them.
7846 [[def-event]]event::
7847 The consequence of the execution of an
7848 <<def-instrumentation-point,instrumentation point>>, like a
7849 <<def-tracepoint,tracepoint>> that you manually place in some source
7850 code, or a Linux kernel kprobe.
7852 An event is said to _occur_ at a specific time. <<def-lttng,LTTng>> can
7853 take various actions upon the occurrence of an event, like record the
7854 event's payload to a <<def-sub-buffer,sub-buffer>>.
7856 [[def-event-name]]event name::
7857 The name of an <<def-event,event>>, which is also the name of the
7858 <<def-event-record,event record>>.
7860 This is also called the _instrumentation point name_.
7862 [[def-event-record]]event record::
7863 A record, in a <<def-trace,trace>>, of the payload of an
7864 <<def-event,event>> which occured.
7866 [[def-event-record-loss-mode]]<<channel-overwrite-mode-vs-discard-mode,event record loss mode>>::
7867 The mechanism by which event records of a given
7868 <<def-channel,channel>> are lost (not recorded) when there is no
7869 <<def-sub-buffer,sub-buffer>> space left to store them.
7871 [[def-event-rule]]<<event,event rule>>::
7872 Set of conditions which must be satisfied for one or more occuring
7873 <<def-event,events>> to be recorded.
7875 `java.util.logging`::
7877 https://docs.oracle.com/javase/7/docs/api/java/util/logging/package-summary.html[core logging facilities].
7879 <<instrumenting,instrumentation>>::
7880 The use of <<def-lttng,LTTng>> probes to make a piece of software
7883 [[def-instrumentation-point]]instrumentation point::
7884 A point in the execution path of a piece of software that, when
7885 reached by this execution, can emit an <<def-event,event>>.
7887 instrumentation point name::
7888 See _<<def-event-name,event name>>_.
7891 A http://logging.apache.org/log4j/1.2/[logging library] for Java
7892 developed by the Apache Software Foundation.
7895 Level of severity of a log statement or user space
7896 <<def-instrumentation-point,instrumentation point>>.
7898 [[def-lttng]]LTTng::
7899 The _Linux Trace Toolkit: next generation_ project.
7901 <<lttng-cli,cmd:lttng>>::
7902 A command-line tool provided by the <<def-lttng-tools,LTTng-tools>>
7903 project which you can use to send and receive control messages to and
7904 from a <<def-session-daemon,session daemon>>.
7907 The https://github.com/lttng/lttng-analyses[LTTng analyses] project,
7908 which is a set of analyzing programs that you can use to obtain a
7909 higher level view of an <<def-lttng,LTTng>> <<def-trace,trace>>.
7911 cmd:lttng-consumerd::
7912 The name of the <<def-consumer-daemon,consumer daemon>> program.
7915 A utility provided by the <<def-lttng-tools,LTTng-tools>> project
7916 which can convert <<def-ring-buffer,ring buffer>> files (usually
7917 <<persistent-memory-file-systems,saved on a persistent memory file
7918 system>>) to <<def-trace,trace>> files.
7920 See man:lttng-crash(1).
7922 LTTng Documentation::
7925 <<lttng-live,LTTng live>>::
7926 A communication protocol between the <<lttng-relayd,relay daemon>> and
7927 live viewers which makes it possible to see <<def-event-record,event
7928 records>> "live", as they are received by the
7929 <<def-relay-daemon,relay daemon>>.
7931 <<lttng-modules,LTTng-modules>>::
7932 The https://github.com/lttng/lttng-modules[LTTng-modules] project,
7933 which contains the Linux kernel modules to make the Linux kernel
7934 <<def-instrumentation-point,instrumentation points>> available for
7935 <<def-lttng,LTTng>> tracing.
7938 The name of the <<def-relay-daemon,relay daemon>> program.
7940 cmd:lttng-sessiond::
7941 The name of the <<def-session-daemon,session daemon>> program.
7943 [[def-lttng-tools]]LTTng-tools::
7944 The https://github.com/lttng/lttng-tools[LTTng-tools] project, which
7945 contains the various programs and libraries used to
7946 <<controlling-tracing,control tracing>>.
7948 [[def-lttng-ust]]<<lttng-ust,LTTng-UST>>::
7949 The https://github.com/lttng/lttng-ust[LTTng-UST] project, which
7950 contains libraries to instrument
7951 <<def-user-application,user applications>>.
7953 <<lttng-ust-agents,LTTng-UST Java agent>>::
7954 A Java package provided by the <<def-lttng-ust,LTTng-UST>> project to
7955 allow the LTTng instrumentation of `java.util.logging` and Apache
7956 log4j{nbsp}1.2 logging statements.
7958 <<lttng-ust-agents,LTTng-UST Python agent>>::
7959 A Python package provided by the <<def-lttng-ust,LTTng-UST>> project
7960 to allow the <<def-lttng,LTTng>> instrumentation of Python logging
7963 <<channel-overwrite-mode-vs-discard-mode,overwrite mode>>::
7964 The <<def-event-record-loss-mode,event record loss mode>> in which new
7965 <<def-event-record,event records>> _overwrite_ older event records
7966 when there's no <<def-sub-buffer,sub-buffer>> space left to store
7969 <<channel-buffering-schemes,per-process buffering>>::
7970 A <<def-buffering-scheme,buffering scheme>> in which each instrumented
7971 process has its own <<def-sub-buffer,sub-buffers>> for a given user
7972 space <<def-channel,channel>>.
7974 <<channel-buffering-schemes,per-user buffering>>::
7975 A <<def-buffering-scheme,buffering scheme>> in which all the processes
7976 of a Unix user share the same <<def-sub-buffer,sub-buffers>> for a
7977 given user space <<def-channel,channel>>.
7979 [[def-relay-daemon]]<<lttng-relayd,relay daemon>>::
7980 A process which is responsible for receiving the <<def-trace,trace>>
7981 data which a distant <<def-consumer-daemon,consumer daemon>> sends.
7983 [[def-ring-buffer]]ring buffer::
7984 A set of <<def-sub-buffer,sub-buffers>>.
7987 See _<<def-tracing-session-rotation,tracing session rotation>>_.
7989 [[def-session-daemon]]<<lttng-sessiond,session daemon>>::
7990 A process which receives control commands from you and orchestrates
7991 the <<def-tracer,tracers>> and various <<def-lttng,LTTng>> daemons.
7993 <<taking-a-snapshot,snapshot>>::
7994 A copy of the current data of all the <<def-sub-buffer,sub-buffers>>
7995 of a given <<def-tracing-session,tracing session>>, saved as
7996 <<def-trace,trace>> files.
7998 [[def-sub-buffer]]sub-buffer::
7999 One part of an <<def-lttng,LTTng>> <<def-ring-buffer,ring buffer>>
8000 which contains <<def-event-record,event records>>.
8003 The time information attached to an <<def-event,event>> when it is
8006 [[def-trace]]trace (_noun_)::
8009 * One http://diamon.org/ctf/[CTF] metadata stream file.
8010 * One or more CTF data stream files which are the concatenations of one
8011 or more flushed <<def-sub-buffer,sub-buffers>>.
8013 [[def-trace-verb]]trace (_verb_)::
8014 The action of recording the <<def-event,events>> emitted by an
8015 application or by a system, or to initiate such recording by
8016 controlling a <<def-tracer,tracer>>.
8018 [[def-trace-chunk]]trace chunk::
8019 A self-contained <<def-trace,trace>> which is part of a
8020 <<def-tracing-session,tracing session>>. Each
8021 <<def-tracing-session-rotation, tracing session rotation>> produces a
8022 <<def-trace-chunk-archive,trace chunk archive>>.
8024 [[def-trace-chunk-archive]]trace chunk archive::
8025 The result of a <<def-tracing-session-rotation, tracing session rotation>>.
8027 <<def-lttng,LTTng>> does not manage any trace chunk archive, even if its
8028 containing <<def-tracing-session,tracing session>> is still active: you
8029 are free to read it, modify it, move it, or remove it.
8032 The http://tracecompass.org[Trace Compass] project and application.
8034 [[def-tracepoint]]tracepoint::
8035 An instrumentation point using the tracepoint mechanism of the Linux
8036 kernel or of <<def-lttng-ust,LTTng-UST>>.
8038 tracepoint definition::
8039 The definition of a single <<def-tracepoint,tracepoint>>.
8042 The name of a <<def-tracepoint,tracepoint>>.
8044 [[def-tracepoint-provider]]tracepoint provider::
8045 A set of functions providing <<def-tracepoint,tracepoints>> to an
8046 instrumented <<def-user-application,user application>>.
8048 Not to be confused with a <<def-tracepoint-provider-package,tracepoint
8049 provider package>>: many tracepoint providers can exist within a
8050 tracepoint provider package.
8052 [[def-tracepoint-provider-package]]tracepoint provider package::
8053 One or more <<def-tracepoint-provider,tracepoint providers>> compiled
8054 as an https://en.wikipedia.org/wiki/Object_file[object file] or as a
8055 link:https://en.wikipedia.org/wiki/Library_(computing)#Shared_libraries[shared
8058 [[def-tracer]]tracer::
8059 A software which records emitted <<def-event,events>>.
8061 <<domain,tracing domain>>::
8062 A namespace for <<def-event,event>> sources.
8064 <<tracing-group,tracing group>>::
8065 The Unix group in which a Unix user can be to be allowed to
8066 <<def-trace-verb,trace>> the Linux kernel.
8068 [[def-tracing-session]]<<tracing-session,tracing session>>::
8069 A stateful dialogue between you and a <<lttng-sessiond,session daemon>>.
8071 [[def-tracing-session-rotation]]<<session-rotation,tracing session rotation>>::
8072 The action of archiving the
8073 <<def-current-trace-chunk,current trace chunk>> of a
8074 <<def-tracing-session,tracing session>>.
8076 [[def-user-application]]user application::
8077 An application running in user space, as opposed to a Linux kernel
8078 module, for example.