xref: /linux/Documentation/trace/user_events.rst (revision e7d759f31ca295d589f7420719c311870bb3166f)
1=========================================
2user_events: User-based Event Tracing
3=========================================
4
5:Author: Beau Belgrave
6
7Overview
8--------
9User based trace events allow user processes to create events and trace data
10that can be viewed via existing tools, such as ftrace and perf.
11To enable this feature, build your kernel with CONFIG_USER_EVENTS=y.
12
13Programs can view status of the events via
14/sys/kernel/tracing/user_events_status and can both register and write
15data out via /sys/kernel/tracing/user_events_data.
16
17Programs can also use /sys/kernel/tracing/dynamic_events to register and
18delete user based events via the u: prefix. The format of the command to
19dynamic_events is the same as the ioctl with the u: prefix applied. This
20requires CAP_PERFMON due to the event persisting, otherwise -EPERM is returned.
21
22Typically programs will register a set of events that they wish to expose to
23tools that can read trace_events (such as ftrace and perf). The registration
24process tells the kernel which address and bit to reflect if any tool has
25enabled the event and data should be written. The registration will give back
26a write index which describes the data when a write() or writev() is called
27on the /sys/kernel/tracing/user_events_data file.
28
29The structures referenced in this document are contained within the
30/include/uapi/linux/user_events.h file in the source tree.
31
32**NOTE:** *Both user_events_status and user_events_data are under the tracefs
33filesystem and may be mounted at different paths than above.*
34
35Registering
36-----------
37Registering within a user process is done via ioctl() out to the
38/sys/kernel/tracing/user_events_data file. The command to issue is
39DIAG_IOCSREG.
40
41This command takes a packed struct user_reg as an argument::
42
43  struct user_reg {
44        /* Input: Size of the user_reg structure being used */
45        __u32 size;
46
47        /* Input: Bit in enable address to use */
48        __u8 enable_bit;
49
50        /* Input: Enable size in bytes at address */
51        __u8 enable_size;
52
53        /* Input: Flags to use, if any */
54        __u16 flags;
55
56        /* Input: Address to update when enabled */
57        __u64 enable_addr;
58
59        /* Input: Pointer to string with event name, description and flags */
60        __u64 name_args;
61
62        /* Output: Index of the event to use when writing data */
63        __u32 write_index;
64  } __attribute__((__packed__));
65
66The struct user_reg requires all the above inputs to be set appropriately.
67
68+ size: This must be set to sizeof(struct user_reg).
69
70+ enable_bit: The bit to reflect the event status at the address specified by
71  enable_addr.
72
73+ enable_size: The size of the value specified by enable_addr.
74  This must be 4 (32-bit) or 8 (64-bit). 64-bit values are only allowed to be
75  used on 64-bit kernels, however, 32-bit can be used on all kernels.
76
77+ flags: The flags to use, if any.
78  Callers should first attempt to use flags and retry without flags to ensure
79  support for lower versions of the kernel. If a flag is not supported -EINVAL
80  is returned.
81
82+ enable_addr: The address of the value to use to reflect event status. This
83  must be naturally aligned and write accessible within the user program.
84
85+ name_args: The name and arguments to describe the event, see command format
86  for details.
87
88The following flags are currently supported.
89
90+ USER_EVENT_REG_PERSIST: The event will not delete upon the last reference
91  closing. Callers may use this if an event should exist even after the
92  process closes or unregisters the event. Requires CAP_PERFMON otherwise
93  -EPERM is returned.
94
95Upon successful registration the following is set.
96
97+ write_index: The index to use for this file descriptor that represents this
98  event when writing out data. The index is unique to this instance of the file
99  descriptor that was used for the registration. See writing data for details.
100
101User based events show up under tracefs like any other event under the
102subsystem named "user_events". This means tools that wish to attach to the
103events need to use /sys/kernel/tracing/events/user_events/[name]/enable
104or perf record -e user_events:[name] when attaching/recording.
105
106**NOTE:** The event subsystem name by default is "user_events". Callers should
107not assume it will always be "user_events". Operators reserve the right in the
108future to change the subsystem name per-process to accommodate event isolation.
109
110Command Format
111^^^^^^^^^^^^^^
112The command string format is as follows::
113
114  name[:FLAG1[,FLAG2...]] [Field1[;Field2...]]
115
116Supported Flags
117^^^^^^^^^^^^^^^
118None yet
119
120Field Format
121^^^^^^^^^^^^
122::
123
124  type name [size]
125
126Basic types are supported (__data_loc, u32, u64, int, char, char[20], etc).
127User programs are encouraged to use clearly sized types like u32.
128
129**NOTE:** *Long is not supported since size can vary between user and kernel.*
130
131The size is only valid for types that start with a struct prefix.
132This allows user programs to describe custom structs out to tools, if required.
133
134For example, a struct in C that looks like this::
135
136  struct mytype {
137    char data[20];
138  };
139
140Would be represented by the following field::
141
142  struct mytype myname 20
143
144Deleting
145--------
146Deleting an event from within a user process is done via ioctl() out to the
147/sys/kernel/tracing/user_events_data file. The command to issue is
148DIAG_IOCSDEL.
149
150This command only requires a single string specifying the event to delete by
151its name. Delete will only succeed if there are no references left to the
152event (in both user and kernel space). User programs should use a separate file
153to request deletes than the one used for registration due to this.
154
155**NOTE:** By default events will auto-delete when there are no references left
156to the event. If programs do not want auto-delete, they must use the
157USER_EVENT_REG_PERSIST flag when registering the event. Once that flag is used
158the event exists until DIAG_IOCSDEL is invoked. Both register and delete of an
159event that persists requires CAP_PERFMON, otherwise -EPERM is returned.
160
161Unregistering
162-------------
163If after registering an event it is no longer wanted to be updated then it can
164be disabled via ioctl() out to the /sys/kernel/tracing/user_events_data file.
165The command to issue is DIAG_IOCSUNREG. This is different than deleting, where
166deleting actually removes the event from the system. Unregistering simply tells
167the kernel your process is no longer interested in updates to the event.
168
169This command takes a packed struct user_unreg as an argument::
170
171  struct user_unreg {
172        /* Input: Size of the user_unreg structure being used */
173        __u32 size;
174
175        /* Input: Bit to unregister */
176        __u8 disable_bit;
177
178        /* Input: Reserved, set to 0 */
179        __u8 __reserved;
180
181        /* Input: Reserved, set to 0 */
182        __u16 __reserved2;
183
184        /* Input: Address to unregister */
185        __u64 disable_addr;
186  } __attribute__((__packed__));
187
188The struct user_unreg requires all the above inputs to be set appropriately.
189
190+ size: This must be set to sizeof(struct user_unreg).
191
192+ disable_bit: This must be set to the bit to disable (same bit that was
193  previously registered via enable_bit).
194
195+ disable_addr: This must be set to the address to disable (same address that was
196  previously registered via enable_addr).
197
198**NOTE:** Events are automatically unregistered when execve() is invoked. During
199fork() the registered events will be retained and must be unregistered manually
200in each process if wanted.
201
202Status
203------
204When tools attach/record user based events the status of the event is updated
205in realtime. This allows user programs to only incur the cost of the write() or
206writev() calls when something is actively attached to the event.
207
208The kernel will update the specified bit that was registered for the event as
209tools attach/detach from the event. User programs simply check if the bit is set
210to see if something is attached or not.
211
212Administrators can easily check the status of all registered events by reading
213the user_events_status file directly via a terminal. The output is as follows::
214
215  Name [# Comments]
216  ...
217
218  Active: ActiveCount
219  Busy: BusyCount
220
221For example, on a system that has a single event the output looks like this::
222
223  test
224
225  Active: 1
226  Busy: 0
227
228If a user enables the user event via ftrace, the output would change to this::
229
230  test # Used by ftrace
231
232  Active: 1
233  Busy: 1
234
235Writing Data
236------------
237After registering an event the same fd that was used to register can be used
238to write an entry for that event. The write_index returned must be at the start
239of the data, then the remaining data is treated as the payload of the event.
240
241For example, if write_index returned was 1 and I wanted to write out an int
242payload of the event. Then the data would have to be 8 bytes (2 ints) in size,
243with the first 4 bytes being equal to 1 and the last 4 bytes being equal to the
244value I want as the payload.
245
246In memory this would look like this::
247
248  int index;
249  int payload;
250
251User programs might have well known structs that they wish to use to emit out
252as payloads. In those cases writev() can be used, with the first vector being
253the index and the following vector(s) being the actual event payload.
254
255For example, if I have a struct like this::
256
257  struct payload {
258        int src;
259        int dst;
260        int flags;
261  } __attribute__((__packed__));
262
263It's advised for user programs to do the following::
264
265  struct iovec io[2];
266  struct payload e;
267
268  io[0].iov_base = &write_index;
269  io[0].iov_len = sizeof(write_index);
270  io[1].iov_base = &e;
271  io[1].iov_len = sizeof(e);
272
273  writev(fd, (const struct iovec*)io, 2);
274
275**NOTE:** *The write_index is not emitted out into the trace being recorded.*
276
277Example Code
278------------
279See sample code in samples/user_events.
280