1                     Using the Linux Kernel Markers
   3                            Mathieu Desnoyers
   6This document introduces Linux Kernel Markers and their use. It provides
   7examples of how to insert markers in the kernel and connect probe functions to
   8them and provides some examples of probe functions.
  11* Purpose of markers
  13A marker placed in code provides a hook to call a function (probe) that you can
  14provide at runtime. A marker can be "on" (a probe is connected to it) or "off"
  15(no probe is attached). When a marker is "off" it has no effect, except for
  16adding a tiny time penalty (checking a condition for a branch) and space
  17penalty (adding a few bytes for the function call at the end of the
  18instrumented function and adds a data structure in a separate section).  When a
  19marker is "on", the function you provide is called each time the marker is
  20executed, in the execution context of the caller. When the function provided
  21ends its execution, it returns to the caller (continuing from the marker site).
  23You can put markers at important locations in the code. Markers are
  24lightweight hooks that can pass an arbitrary number of parameters,
  25described in a printk-like format string, to the attached probe function.
  27They can be used for tracing and performance accounting.
  30* Usage
  32In order to use the macro trace_mark, you should include linux/marker.h.
  34#include <linux/marker.h>
  38trace_mark(subsystem_event, "myint %d mystring %s", someint, somestring);
  39Where :
  40- subsystem_event is an identifier unique to your event
  41    - subsystem is the name of your subsystem.
  42    - event is the name of the event to mark.
  43- "myint %d mystring %s" is the formatted string for the serializer. "myint" and
  44  "mystring" are repectively the field names associated with the first and
  45  second parameter.
  46- someint is an integer.
  47- somestring is a char pointer.
  49Connecting a function (probe) to a marker is done by providing a probe (function
  50to call) for the specific marker through marker_probe_register() and can be
  51activated by calling marker_arm(). Marker deactivation can be done by calling
  52marker_disarm() as many times as marker_arm() has been called. Removing a probe
  53is done through marker_probe_unregister(); it will disarm the probe.
  55marker_synchronize_unregister() must be called between probe unregistration and
  56the first occurrence of
  57- the end of module exit function,
  58  to make sure there is no caller left using the probe;
  59- the free of any resource used by the probes,
  60  to make sure the probes wont be accessing invalid data.
  61This, and the fact that preemption is disabled around the probe call, make sure
  62that probe removal and module unload are safe. See the "Probe example" section
  63below for a sample probe module.
  65The marker mechanism supports inserting multiple instances of the same marker.
  66Markers can be put in inline functions, inlined static functions, and
  67unrolled loops as well as regular functions.
  69The naming scheme "subsystem_event" is suggested here as a convention intended
  70to limit collisions. Marker names are global to the kernel: they are considered
  71as being the same whether they are in the core kernel image or in modules.
  72Conflicting format strings for markers with the same name will cause the markers
  73to be detected to have a different format string not to be armed and will output
  74a printk warning which identifies the inconsistency:
  76"Format mismatch for probe probe_name (format), marker (format)"
  78Another way to use markers is to simply define the marker without generating any
  79function call to actually call into the marker. This is useful in combination
  80with tracepoint probes in a scheme like this :
  82void probe_tracepoint_name(unsigned int arg1, struct task_struct *tsk);
  84DEFINE_MARKER_TP(marker_eventname, tracepoint_name, probe_tracepoint_name,
  85        "arg1 %u pid %d");
  87notrace void probe_tracepoint_name(unsigned int arg1, struct task_struct *tsk)
  89        struct marker *marker = &GET_MARKER(kernel_irq_entry);
  90        /* write data to trace buffers ... */
  93* Probe / marker example
  95See the example provided in samples/markers/src
  97Compile them with your kernel.
  99Run, as root :
 100modprobe marker-example (insmod order is not important)
 101modprobe probe-example
 102cat /proc/marker-example (returns an expected error)
 103rmmod marker-example probe-example
 105 kindly hosted by Redpill Linpro AS, provider of Linux consulting and operations services since 1995.