1 ======================== 1 ========================
2 ftrace - Function Tracer 2 ftrace - Function Tracer
3 ======================== 3 ========================
4 4
5 Copyright 2008 Red Hat Inc. 5 Copyright 2008 Red Hat Inc.
6 6
7 :Author: Steven Rostedt <srostedt@redhat.com> 7 :Author: Steven Rostedt <srostedt@redhat.com>
8 :License: The GNU Free Documentation License, 8 :License: The GNU Free Documentation License, Version 1.2
9 (dual licensed under the GPL v2) 9 (dual licensed under the GPL v2)
10 :Original Reviewers: Elias Oltmanns, Randy Du 10 :Original Reviewers: Elias Oltmanns, Randy Dunlap, Andrew Morton,
11 John Kacur, and David Te 11 John Kacur, and David Teigland.
12 12
13 - Written for: 2.6.28-rc2 13 - Written for: 2.6.28-rc2
14 - Updated for: 3.10 14 - Updated for: 3.10
15 - Updated for: 4.13 - Copyright 2017 VMware In 15 - Updated for: 4.13 - Copyright 2017 VMware Inc. Steven Rostedt
16 - Converted to rst format - Changbin Du <changb 16 - Converted to rst format - Changbin Du <changbin.du@intel.com>
17 17
18 Introduction 18 Introduction
19 ------------ 19 ------------
20 20
21 Ftrace is an internal tracer designed to help 21 Ftrace is an internal tracer designed to help out developers and
22 designers of systems to find what is going on 22 designers of systems to find what is going on inside the kernel.
23 It can be used for debugging or analyzing late 23 It can be used for debugging or analyzing latencies and
24 performance issues that take place outside of 24 performance issues that take place outside of user-space.
25 25
26 Although ftrace is typically considered the fu 26 Although ftrace is typically considered the function tracer, it
27 is really a framework of several assorted trac 27 is really a framework of several assorted tracing utilities.
28 There's latency tracing to examine what occurs 28 There's latency tracing to examine what occurs between interrupts
29 disabled and enabled, as well as for preemptio 29 disabled and enabled, as well as for preemption and from a time
30 a task is woken to the task is actually schedu 30 a task is woken to the task is actually scheduled in.
31 31
32 One of the most common uses of ftrace is the e 32 One of the most common uses of ftrace is the event tracing.
33 Throughout the kernel is hundreds of static ev 33 Throughout the kernel is hundreds of static event points that
34 can be enabled via the tracefs file system to 34 can be enabled via the tracefs file system to see what is
35 going on in certain parts of the kernel. 35 going on in certain parts of the kernel.
36 36
37 See events.rst for more information. 37 See events.rst for more information.
38 38
39 39
40 Implementation Details 40 Implementation Details
41 ---------------------- 41 ----------------------
42 42
43 See Documentation/trace/ftrace-design.rst for !! 43 See :doc:`ftrace-design` for details for arch porters and such.
44 44
45 45
46 The File System 46 The File System
47 --------------- 47 ---------------
48 48
49 Ftrace uses the tracefs file system to hold th 49 Ftrace uses the tracefs file system to hold the control files as
50 well as the files to display output. 50 well as the files to display output.
51 51
52 When tracefs is configured into the kernel (wh 52 When tracefs is configured into the kernel (which selecting any ftrace
53 option will do) the directory /sys/kernel/trac 53 option will do) the directory /sys/kernel/tracing will be created. To mount
54 this directory, you can add to your /etc/fstab 54 this directory, you can add to your /etc/fstab file::
55 55
56 tracefs /sys/kernel/tracing trace 56 tracefs /sys/kernel/tracing tracefs defaults 0 0
57 57
58 Or you can mount it at run time with:: 58 Or you can mount it at run time with::
59 59
60 mount -t tracefs nodev /sys/kernel/tracing 60 mount -t tracefs nodev /sys/kernel/tracing
61 61
62 For quicker access to that directory you may w 62 For quicker access to that directory you may want to make a soft link to
63 it:: 63 it::
64 64
65 ln -s /sys/kernel/tracing /tracing 65 ln -s /sys/kernel/tracing /tracing
66 66
67 .. attention:: 67 .. attention::
68 68
69 Before 4.1, all ftrace tracing control files 69 Before 4.1, all ftrace tracing control files were within the debugfs
70 file system, which is typically located at / 70 file system, which is typically located at /sys/kernel/debug/tracing.
71 For backward compatibility, when mounting th 71 For backward compatibility, when mounting the debugfs file system,
72 the tracefs file system will be automaticall 72 the tracefs file system will be automatically mounted at:
73 73
74 /sys/kernel/debug/tracing 74 /sys/kernel/debug/tracing
75 75
76 All files located in the tracefs file system 76 All files located in the tracefs file system will be located in that
77 debugfs file system directory as well. 77 debugfs file system directory as well.
78 78
79 .. attention:: 79 .. attention::
80 80
81 Any selected ftrace option will also create 81 Any selected ftrace option will also create the tracefs file system.
82 The rest of the document will assume that yo 82 The rest of the document will assume that you are in the ftrace directory
83 (cd /sys/kernel/tracing) and will only conce 83 (cd /sys/kernel/tracing) and will only concentrate on the files within that
84 directory and not distract from the content 84 directory and not distract from the content with the extended
85 "/sys/kernel/tracing" path name. 85 "/sys/kernel/tracing" path name.
86 86
87 That's it! (assuming that you have ftrace conf 87 That's it! (assuming that you have ftrace configured into your kernel)
88 88
89 After mounting tracefs you will have access to 89 After mounting tracefs you will have access to the control and output files
90 of ftrace. Here is a list of some of the key f 90 of ftrace. Here is a list of some of the key files:
91 91
92 92
93 Note: all time values are in microseconds. 93 Note: all time values are in microseconds.
94 94
95 current_tracer: 95 current_tracer:
96 96
97 This is used to set or display the cur 97 This is used to set or display the current tracer
98 that is configured. Changing the curre 98 that is configured. Changing the current tracer clears
99 the ring buffer content as well as the 99 the ring buffer content as well as the "snapshot" buffer.
100 100
101 available_tracers: 101 available_tracers:
102 102
103 This holds the different types of trac 103 This holds the different types of tracers that
104 have been compiled into the kernel. Th 104 have been compiled into the kernel. The
105 tracers listed here can be configured 105 tracers listed here can be configured by
106 echoing their name into current_tracer 106 echoing their name into current_tracer.
107 107
108 tracing_on: 108 tracing_on:
109 109
110 This sets or displays whether writing 110 This sets or displays whether writing to the trace
111 ring buffer is enabled. Echo 0 into th 111 ring buffer is enabled. Echo 0 into this file to disable
112 the tracer or 1 to enable it. Note, th 112 the tracer or 1 to enable it. Note, this only disables
113 writing to the ring buffer, the tracin 113 writing to the ring buffer, the tracing overhead may
114 still be occurring. 114 still be occurring.
115 115
116 The kernel function tracing_off() can 116 The kernel function tracing_off() can be used within the
117 kernel to disable writing to the ring 117 kernel to disable writing to the ring buffer, which will
118 set this file to "0". User space can r 118 set this file to "0". User space can re-enable tracing by
119 echoing "1" into the file. 119 echoing "1" into the file.
120 120
121 Note, the function and event trigger " 121 Note, the function and event trigger "traceoff" will also
122 set this file to zero and stop tracing 122 set this file to zero and stop tracing. Which can also
123 be re-enabled by user space using this 123 be re-enabled by user space using this file.
124 124
125 trace: 125 trace:
126 126
127 This file holds the output of the trac 127 This file holds the output of the trace in a human
128 readable format (described below). Ope 128 readable format (described below). Opening this file for
129 writing with the O_TRUNC flag clears t 129 writing with the O_TRUNC flag clears the ring buffer content.
130 Note, this file is not a consumer. If 130 Note, this file is not a consumer. If tracing is off
131 (no tracer running, or tracing_on is z 131 (no tracer running, or tracing_on is zero), it will produce
132 the same output each time it is read. 132 the same output each time it is read. When tracing is on,
133 it may produce inconsistent results as 133 it may produce inconsistent results as it tries to read
134 the entire buffer without consuming it 134 the entire buffer without consuming it.
135 135
136 trace_pipe: 136 trace_pipe:
137 137
138 The output is the same as the "trace" 138 The output is the same as the "trace" file but this
139 file is meant to be streamed with live 139 file is meant to be streamed with live tracing.
140 Reads from this file will block until 140 Reads from this file will block until new data is
141 retrieved. Unlike the "trace" file, t 141 retrieved. Unlike the "trace" file, this file is a
142 consumer. This means reading from this 142 consumer. This means reading from this file causes
143 sequential reads to display more curre 143 sequential reads to display more current data. Once
144 data is read from this file, it is con 144 data is read from this file, it is consumed, and
145 will not be read again with a sequenti 145 will not be read again with a sequential read. The
146 "trace" file is static, and if the tra 146 "trace" file is static, and if the tracer is not
147 adding more data, it will display the 147 adding more data, it will display the same
148 information every time it is read. 148 information every time it is read.
149 149
150 trace_options: 150 trace_options:
151 151
152 This file lets the user control the am 152 This file lets the user control the amount of data
153 that is displayed in one of the above 153 that is displayed in one of the above output
154 files. Options also exist to modify ho 154 files. Options also exist to modify how a tracer
155 or events work (stack traces, timestam 155 or events work (stack traces, timestamps, etc).
156 156
157 options: 157 options:
158 158
159 This is a directory that has a file fo 159 This is a directory that has a file for every available
160 trace option (also in trace_options). 160 trace option (also in trace_options). Options may also be set
161 or cleared by writing a "1" or "0" res 161 or cleared by writing a "1" or "0" respectively into the
162 corresponding file with the option nam 162 corresponding file with the option name.
163 163
164 tracing_max_latency: 164 tracing_max_latency:
165 165
166 Some of the tracers record the max lat 166 Some of the tracers record the max latency.
167 For example, the maximum time that int 167 For example, the maximum time that interrupts are disabled.
168 The maximum time is saved in this file 168 The maximum time is saved in this file. The max trace will also be
169 stored, and displayed by "trace". A ne 169 stored, and displayed by "trace". A new max trace will only be
170 recorded if the latency is greater tha 170 recorded if the latency is greater than the value in this file
171 (in microseconds). 171 (in microseconds).
172 172
173 By echoing in a time into this file, n 173 By echoing in a time into this file, no latency will be recorded
174 unless it is greater than the time in 174 unless it is greater than the time in this file.
175 175
176 tracing_thresh: 176 tracing_thresh:
177 177
178 Some latency tracers will record a tra 178 Some latency tracers will record a trace whenever the
179 latency is greater than the number in 179 latency is greater than the number in this file.
180 Only active when the file contains a n 180 Only active when the file contains a number greater than 0.
181 (in microseconds) 181 (in microseconds)
182 182
183 buffer_percent: <<
184 <<
185 This is the watermark for how much the <<
186 before a waiter is woken up. That is, <<
187 blocking read syscall on one of the pe <<
188 will block until the given amount of d <<
189 is in the ring buffer before it wakes <<
190 controls how the splice system calls a <<
191 <<
192 0 - means to wake up as soon as th <<
193 50 - means to wake up when roughly <<
194 are full. <<
195 100 - means to block until the ring <<
196 about to start overwriting the <<
197 <<
198 buffer_size_kb: 183 buffer_size_kb:
199 184
200 This sets or displays the number of ki 185 This sets or displays the number of kilobytes each CPU
201 buffer holds. By default, the trace bu 186 buffer holds. By default, the trace buffers are the same size
202 for each CPU. The displayed number is 187 for each CPU. The displayed number is the size of the
203 CPU buffer and not total size of all b 188 CPU buffer and not total size of all buffers. The
204 trace buffers are allocated in pages ( 189 trace buffers are allocated in pages (blocks of memory
205 that the kernel uses for allocation, u 190 that the kernel uses for allocation, usually 4 KB in size).
206 A few extra pages may be allocated to 191 A few extra pages may be allocated to accommodate buffer management
207 meta-data. If the last page allocated 192 meta-data. If the last page allocated has room for more bytes
208 than requested, the rest of the page w 193 than requested, the rest of the page will be used,
209 making the actual allocation bigger th 194 making the actual allocation bigger than requested or shown.
210 ( Note, the size may not be a multiple 195 ( Note, the size may not be a multiple of the page size
211 due to buffer management meta-data. ) 196 due to buffer management meta-data. )
212 197
213 Buffer sizes for individual CPUs may v 198 Buffer sizes for individual CPUs may vary
214 (see "per_cpu/cpu0/buffer_size_kb" bel 199 (see "per_cpu/cpu0/buffer_size_kb" below), and if they do
215 this file will show "X". 200 this file will show "X".
216 201
217 buffer_total_size_kb: 202 buffer_total_size_kb:
218 203
219 This displays the total combined size 204 This displays the total combined size of all the trace buffers.
220 205
221 buffer_subbuf_size_kb: <<
222 <<
223 This sets or displays the sub buffer s <<
224 into several same size "sub buffers". <<
225 the size of the sub buffer. Normally, <<
226 architecture's page (4K on x86). The s <<
227 at the start which also limits the siz <<
228 the sub buffer is a page size, no even <<
229 size minus the sub buffer meta data. <<
230 <<
231 Note, the buffer_subbuf_size_kb is a w <<
232 minimum size of the subbuffer. The ker <<
233 implementation details, or simply fail <<
234 not handle the request. <<
235 <<
236 Changing the sub buffer size allows fo <<
237 page size. <<
238 <<
239 Note: When changing the sub-buffer siz <<
240 data in the ring buffer and the snapsh <<
241 <<
242 free_buffer: 206 free_buffer:
243 207
244 If a process is performing tracing, an 208 If a process is performing tracing, and the ring buffer should be
245 shrunk "freed" when the process is fin 209 shrunk "freed" when the process is finished, even if it were to be
246 killed by a signal, this file can be u 210 killed by a signal, this file can be used for that purpose. On close
247 of this file, the ring buffer will be 211 of this file, the ring buffer will be resized to its minimum size.
248 Having a process that is tracing also 212 Having a process that is tracing also open this file, when the process
249 exits its file descriptor for this fil 213 exits its file descriptor for this file will be closed, and in doing so,
250 the ring buffer will be "freed". 214 the ring buffer will be "freed".
251 215
252 It may also stop tracing if disable_on 216 It may also stop tracing if disable_on_free option is set.
253 217
254 tracing_cpumask: 218 tracing_cpumask:
255 219
256 This is a mask that lets the user only 220 This is a mask that lets the user only trace on specified CPUs.
257 The format is a hex string representin 221 The format is a hex string representing the CPUs.
258 222
259 set_ftrace_filter: 223 set_ftrace_filter:
260 224
261 When dynamic ftrace is configured in ( 225 When dynamic ftrace is configured in (see the
262 section below "dynamic ftrace"), the c 226 section below "dynamic ftrace"), the code is dynamically
263 modified (code text rewrite) to disabl 227 modified (code text rewrite) to disable calling of the
264 function profiler (mcount). This lets 228 function profiler (mcount). This lets tracing be configured
265 in with practically no overhead in per 229 in with practically no overhead in performance. This also
266 has a side effect of enabling or disab 230 has a side effect of enabling or disabling specific functions
267 to be traced. Echoing names of functio 231 to be traced. Echoing names of functions into this file
268 will limit the trace to only those fun 232 will limit the trace to only those functions.
269 This influences the tracers "function" 233 This influences the tracers "function" and "function_graph"
270 and thus also function profiling (see 234 and thus also function profiling (see "function_profile_enabled").
271 235
272 The functions listed in "available_fil 236 The functions listed in "available_filter_functions" are what
273 can be written into this file. 237 can be written into this file.
274 238
275 This interface also allows for command 239 This interface also allows for commands to be used. See the
276 "Filter commands" section for more det 240 "Filter commands" section for more details.
277 241
278 As a speed up, since processing string 242 As a speed up, since processing strings can be quite expensive
279 and requires a check of all functions 243 and requires a check of all functions registered to tracing, instead
280 an index can be written into this file 244 an index can be written into this file. A number (starting with "1")
281 written will instead select the same c 245 written will instead select the same corresponding at the line position
282 of the "available_filter_functions" fi 246 of the "available_filter_functions" file.
283 247
284 set_ftrace_notrace: 248 set_ftrace_notrace:
285 249
286 This has an effect opposite to that of 250 This has an effect opposite to that of
287 set_ftrace_filter. Any function that i 251 set_ftrace_filter. Any function that is added here will not
288 be traced. If a function exists in bot 252 be traced. If a function exists in both set_ftrace_filter
289 and set_ftrace_notrace, the function w 253 and set_ftrace_notrace, the function will _not_ be traced.
290 254
291 set_ftrace_pid: 255 set_ftrace_pid:
292 256
293 Have the function tracer only trace th 257 Have the function tracer only trace the threads whose PID are
294 listed in this file. 258 listed in this file.
295 259
296 If the "function-fork" option is set, 260 If the "function-fork" option is set, then when a task whose
297 PID is listed in this file forks, the 261 PID is listed in this file forks, the child's PID will
298 automatically be added to this file, a 262 automatically be added to this file, and the child will be
299 traced by the function tracer as well. 263 traced by the function tracer as well. This option will also
300 cause PIDs of tasks that exit to be re 264 cause PIDs of tasks that exit to be removed from the file.
301 265
302 set_ftrace_notrace_pid: 266 set_ftrace_notrace_pid:
303 267
304 Have the function tracer ignore thread 268 Have the function tracer ignore threads whose PID are listed in
305 this file. 269 this file.
306 270
307 If the "function-fork" option is set, 271 If the "function-fork" option is set, then when a task whose
308 PID is listed in this file forks, the 272 PID is listed in this file forks, the child's PID will
309 automatically be added to this file, a 273 automatically be added to this file, and the child will not be
310 traced by the function tracer as well. 274 traced by the function tracer as well. This option will also
311 cause PIDs of tasks that exit to be re 275 cause PIDs of tasks that exit to be removed from the file.
312 276
313 If a PID is in both this file and "set 277 If a PID is in both this file and "set_ftrace_pid", then this
314 file takes precedence, and the thread 278 file takes precedence, and the thread will not be traced.
315 279
316 set_event_pid: 280 set_event_pid:
317 281
318 Have the events only trace a task with 282 Have the events only trace a task with a PID listed in this file.
319 Note, sched_switch and sched_wake_up w 283 Note, sched_switch and sched_wake_up will also trace events
320 listed in this file. 284 listed in this file.
321 285
322 To have the PIDs of children of tasks 286 To have the PIDs of children of tasks with their PID in this file
323 added on fork, enable the "event-fork" 287 added on fork, enable the "event-fork" option. That option will also
324 cause the PIDs of tasks to be removed 288 cause the PIDs of tasks to be removed from this file when the task
325 exits. 289 exits.
326 290
327 set_event_notrace_pid: 291 set_event_notrace_pid:
328 292
329 Have the events not trace a task with 293 Have the events not trace a task with a PID listed in this file.
330 Note, sched_switch and sched_wakeup wi 294 Note, sched_switch and sched_wakeup will trace threads not listed
331 in this file, even if a thread's PID i 295 in this file, even if a thread's PID is in the file if the
332 sched_switch or sched_wakeup events al 296 sched_switch or sched_wakeup events also trace a thread that should
333 be traced. 297 be traced.
334 298
335 To have the PIDs of children of tasks 299 To have the PIDs of children of tasks with their PID in this file
336 added on fork, enable the "event-fork" 300 added on fork, enable the "event-fork" option. That option will also
337 cause the PIDs of tasks to be removed 301 cause the PIDs of tasks to be removed from this file when the task
338 exits. 302 exits.
339 303
340 set_graph_function: 304 set_graph_function:
341 305
342 Functions listed in this file will cau 306 Functions listed in this file will cause the function graph
343 tracer to only trace these functions a 307 tracer to only trace these functions and the functions that
344 they call. (See the section "dynamic f 308 they call. (See the section "dynamic ftrace" for more details).
345 Note, set_ftrace_filter and set_ftrace 309 Note, set_ftrace_filter and set_ftrace_notrace still affects
346 what functions are being traced. 310 what functions are being traced.
347 311
348 set_graph_notrace: 312 set_graph_notrace:
349 313
350 Similar to set_graph_function, but wil 314 Similar to set_graph_function, but will disable function graph
351 tracing when the function is hit until 315 tracing when the function is hit until it exits the function.
352 This makes it possible to ignore traci 316 This makes it possible to ignore tracing functions that are called
353 by a specific function. 317 by a specific function.
354 318
355 available_filter_functions: 319 available_filter_functions:
356 320
357 This lists the functions that ftrace h 321 This lists the functions that ftrace has processed and can trace.
358 These are the function names that you 322 These are the function names that you can pass to
359 "set_ftrace_filter", "set_ftrace_notra 323 "set_ftrace_filter", "set_ftrace_notrace",
360 "set_graph_function", or "set_graph_no 324 "set_graph_function", or "set_graph_notrace".
361 (See the section "dynamic ftrace" belo 325 (See the section "dynamic ftrace" below for more details.)
362 326
363 available_filter_functions_addrs: <<
364 <<
365 Similar to available_filter_functions, <<
366 for each function. The displayed addre <<
367 and can differ from /proc/kallsyms add <<
368 <<
369 dyn_ftrace_total_info: 327 dyn_ftrace_total_info:
370 328
371 This file is for debugging purposes. T 329 This file is for debugging purposes. The number of functions that
372 have been converted to nops and are av 330 have been converted to nops and are available to be traced.
373 331
374 enabled_functions: 332 enabled_functions:
375 333
376 This file is more for debugging ftrace 334 This file is more for debugging ftrace, but can also be useful
377 in seeing if any function has a callba 335 in seeing if any function has a callback attached to it.
378 Not only does the trace infrastructure 336 Not only does the trace infrastructure use ftrace function
379 trace utility, but other subsystems mi 337 trace utility, but other subsystems might too. This file
380 displays all functions that have a cal 338 displays all functions that have a callback attached to them
381 as well as the number of callbacks tha 339 as well as the number of callbacks that have been attached.
382 Note, a callback may also call multipl 340 Note, a callback may also call multiple functions which will
383 not be listed in this count. 341 not be listed in this count.
384 342
385 If the callback registered to be trace 343 If the callback registered to be traced by a function with
386 the "save regs" attribute (thus even m 344 the "save regs" attribute (thus even more overhead), a 'R'
387 will be displayed on the same line as 345 will be displayed on the same line as the function that
388 is returning registers. 346 is returning registers.
389 347
390 If the callback registered to be trace 348 If the callback registered to be traced by a function with
391 the "ip modify" attribute (thus the re 349 the "ip modify" attribute (thus the regs->ip can be changed),
392 an 'I' will be displayed on the same l 350 an 'I' will be displayed on the same line as the function that
393 can be overridden. 351 can be overridden.
394 352
395 If a non ftrace trampoline is attached <<
396 Note, normal ftrace trampolines can al <<
397 "direct" trampoline can be attached to <<
398 <<
399 Some architectures can not call direct <<
400 the ftrace ops function located above <<
401 such cases an 'O' will be displayed. <<
402 <<
403 If a function had either the "ip modif <<
404 it in the past, a 'M' will be shown. T <<
405 used to know if a function was every m <<
406 and can be used for debugging. <<
407 <<
408 If the architecture supports it, it wi 353 If the architecture supports it, it will also show what callback
409 is being directly called by the functi 354 is being directly called by the function. If the count is greater
410 than 1 it most likely will be ftrace_o 355 than 1 it most likely will be ftrace_ops_list_func().
411 356
412 If the callback of a function jumps to !! 357 If the callback of the function jumps to a trampoline that is
413 specific to the callback and which is !! 358 specific to a the callback and not the standard trampoline,
414 its address will be printed as well as 359 its address will be printed as well as the function that the
415 trampoline calls. 360 trampoline calls.
416 361
417 touched_functions: <<
418 <<
419 This file contains all the functions t <<
420 to it via the ftrace infrastructure. I <<
421 enabled_functions but shows all functi <<
422 traced. <<
423 <<
424 To see any function that has every bee <<
425 direct trampoline, one can perform the <<
426 <<
427 grep ' M ' /sys/kernel/tracing/touched <<
428 <<
429 function_profile_enabled: 362 function_profile_enabled:
430 363
431 When set it will enable all functions 364 When set it will enable all functions with either the function
432 tracer, or if configured, the function 365 tracer, or if configured, the function graph tracer. It will
433 keep a histogram of the number of func 366 keep a histogram of the number of functions that were called
434 and if the function graph tracer was c 367 and if the function graph tracer was configured, it will also keep
435 track of the time spent in those funct 368 track of the time spent in those functions. The histogram
436 content can be displayed in the files: 369 content can be displayed in the files:
437 370
438 trace_stat/function<cpu> ( function0, 371 trace_stat/function<cpu> ( function0, function1, etc).
439 372
440 trace_stat: 373 trace_stat:
441 374
442 A directory that holds different traci 375 A directory that holds different tracing stats.
443 376
444 kprobe_events: 377 kprobe_events:
445 378
446 Enable dynamic trace points. See kprob 379 Enable dynamic trace points. See kprobetrace.rst.
447 380
448 kprobe_profile: 381 kprobe_profile:
449 382
450 Dynamic trace points stats. See kprobe 383 Dynamic trace points stats. See kprobetrace.rst.
451 384
452 max_graph_depth: 385 max_graph_depth:
453 386
454 Used with the function graph tracer. T 387 Used with the function graph tracer. This is the max depth
455 it will trace into a function. Setting 388 it will trace into a function. Setting this to a value of
456 one will show only the first kernel fu 389 one will show only the first kernel function that is called
457 from user space. 390 from user space.
458 391
459 printk_formats: 392 printk_formats:
460 393
461 This is for tools that read the raw fo 394 This is for tools that read the raw format files. If an event in
462 the ring buffer references a string, o 395 the ring buffer references a string, only a pointer to the string
463 is recorded into the buffer and not th 396 is recorded into the buffer and not the string itself. This prevents
464 tools from knowing what that string wa 397 tools from knowing what that string was. This file displays the string
465 and address for the string allowing to 398 and address for the string allowing tools to map the pointers to what
466 the strings were. 399 the strings were.
467 400
468 saved_cmdlines: 401 saved_cmdlines:
469 402
470 Only the pid of the task is recorded i 403 Only the pid of the task is recorded in a trace event unless
471 the event specifically saves the task 404 the event specifically saves the task comm as well. Ftrace
472 makes a cache of pid mappings to comms 405 makes a cache of pid mappings to comms to try to display
473 comms for events. If a pid for a comm 406 comms for events. If a pid for a comm is not listed, then
474 "<...>" is displayed in the output. 407 "<...>" is displayed in the output.
475 408
476 If the option "record-cmd" is set to " 409 If the option "record-cmd" is set to "0", then comms of tasks
477 will not be saved during recording. By 410 will not be saved during recording. By default, it is enabled.
478 411
479 saved_cmdlines_size: 412 saved_cmdlines_size:
480 413
481 By default, 128 comms are saved (see " 414 By default, 128 comms are saved (see "saved_cmdlines" above). To
482 increase or decrease the amount of com 415 increase or decrease the amount of comms that are cached, echo
483 the number of comms to cache into this 416 the number of comms to cache into this file.
484 417
485 saved_tgids: 418 saved_tgids:
486 419
487 If the option "record-tgid" is set, on 420 If the option "record-tgid" is set, on each scheduling context switch
488 the Task Group ID of a task is saved i 421 the Task Group ID of a task is saved in a table mapping the PID of
489 the thread to its TGID. By default, th 422 the thread to its TGID. By default, the "record-tgid" option is
490 disabled. 423 disabled.
491 424
492 snapshot: 425 snapshot:
493 426
494 This displays the "snapshot" buffer an 427 This displays the "snapshot" buffer and also lets the user
495 take a snapshot of the current running 428 take a snapshot of the current running trace.
496 See the "Snapshot" section below for m 429 See the "Snapshot" section below for more details.
497 430
498 stack_max_size: 431 stack_max_size:
499 432
500 When the stack tracer is activated, th 433 When the stack tracer is activated, this will display the
501 maximum stack size it has encountered. 434 maximum stack size it has encountered.
502 See the "Stack Trace" section below. 435 See the "Stack Trace" section below.
503 436
504 stack_trace: 437 stack_trace:
505 438
506 This displays the stack back trace of 439 This displays the stack back trace of the largest stack
507 that was encountered when the stack tr 440 that was encountered when the stack tracer is activated.
508 See the "Stack Trace" section below. 441 See the "Stack Trace" section below.
509 442
510 stack_trace_filter: 443 stack_trace_filter:
511 444
512 This is similar to "set_ftrace_filter" 445 This is similar to "set_ftrace_filter" but it limits what
513 functions the stack tracer will check. 446 functions the stack tracer will check.
514 447
515 trace_clock: 448 trace_clock:
516 449
517 Whenever an event is recorded into the 450 Whenever an event is recorded into the ring buffer, a
518 "timestamp" is added. This stamp comes 451 "timestamp" is added. This stamp comes from a specified
519 clock. By default, ftrace uses the "lo 452 clock. By default, ftrace uses the "local" clock. This
520 clock is very fast and strictly per cp 453 clock is very fast and strictly per cpu, but on some
521 systems it may not be monotonic with r 454 systems it may not be monotonic with respect to other
522 CPUs. In other words, the local clocks 455 CPUs. In other words, the local clocks may not be in sync
523 with local clocks on other CPUs. 456 with local clocks on other CPUs.
524 457
525 Usual clocks for tracing:: 458 Usual clocks for tracing::
526 459
527 # cat trace_clock 460 # cat trace_clock
528 [local] global counter x86-tsc 461 [local] global counter x86-tsc
529 462
530 The clock with the square brackets aro 463 The clock with the square brackets around it is the one in effect.
531 464
532 local: 465 local:
533 Default clock, but may not be 466 Default clock, but may not be in sync across CPUs
534 467
535 global: 468 global:
536 This clock is in sync with all 469 This clock is in sync with all CPUs but may
537 be a bit slower than the local 470 be a bit slower than the local clock.
538 471
539 counter: 472 counter:
540 This is not a clock at all, bu 473 This is not a clock at all, but literally an atomic
541 counter. It counts up one by o 474 counter. It counts up one by one, but is in sync
542 with all CPUs. This is useful 475 with all CPUs. This is useful when you need to
543 know exactly the order events 476 know exactly the order events occurred with respect to
544 each other on different CPUs. 477 each other on different CPUs.
545 478
546 uptime: 479 uptime:
547 This uses the jiffies counter 480 This uses the jiffies counter and the time stamp
548 is relative to the time since 481 is relative to the time since boot up.
549 482
550 perf: 483 perf:
551 This makes ftrace use the same 484 This makes ftrace use the same clock that perf uses.
552 Eventually perf will be able t 485 Eventually perf will be able to read ftrace buffers
553 and this will help out in inte 486 and this will help out in interleaving the data.
554 487
555 x86-tsc: 488 x86-tsc:
556 Architectures may define their 489 Architectures may define their own clocks. For
557 example, x86 uses its own TSC 490 example, x86 uses its own TSC cycle clock here.
558 491
559 ppc-tb: 492 ppc-tb:
560 This uses the powerpc timebase 493 This uses the powerpc timebase register value.
561 This is in sync across CPUs an 494 This is in sync across CPUs and can also be used
562 to correlate events across hyp 495 to correlate events across hypervisor/guest if
563 tb_offset is known. 496 tb_offset is known.
564 497
565 mono: 498 mono:
566 This uses the fast monotonic c 499 This uses the fast monotonic clock (CLOCK_MONOTONIC)
567 which is monotonic and is subj 500 which is monotonic and is subject to NTP rate adjustments.
568 501
569 mono_raw: 502 mono_raw:
570 This is the raw monotonic cloc 503 This is the raw monotonic clock (CLOCK_MONOTONIC_RAW)
571 which is monotonic but is not 504 which is monotonic but is not subject to any rate adjustments
572 and ticks at the same rate as 505 and ticks at the same rate as the hardware clocksource.
573 506
574 boot: 507 boot:
575 This is the boot clock (CLOCK_ 508 This is the boot clock (CLOCK_BOOTTIME) and is based on the
576 fast monotonic clock, but also 509 fast monotonic clock, but also accounts for time spent in
577 suspend. Since the clock acces 510 suspend. Since the clock access is designed for use in
578 tracing in the suspend path, s 511 tracing in the suspend path, some side effects are possible
579 if clock is accessed after the 512 if clock is accessed after the suspend time is accounted before
580 the fast mono clock is updated 513 the fast mono clock is updated. In this case, the clock update
581 appears to happen slightly soo 514 appears to happen slightly sooner than it normally would have.
582 Also on 32-bit systems, it's p 515 Also on 32-bit systems, it's possible that the 64-bit boot offset
583 sees a partial update. These e 516 sees a partial update. These effects are rare and post
584 processing should be able to h 517 processing should be able to handle them. See comments in the
585 ktime_get_boot_fast_ns() funct 518 ktime_get_boot_fast_ns() function for more information.
586 519
587 tai: <<
588 This is the tai clock (CLOCK_T <<
589 clock time. However, this cloc <<
590 discontinuities and backwards <<
591 seconds. Since the clock acces <<
592 side effects are possible. The <<
593 readouts in case the internal <<
594 by setting the system time or <<
595 These effects are rare and pos <<
596 handle them. See comments in t <<
597 function for more information. <<
598 <<
599 To set a clock, simply echo the clock 520 To set a clock, simply echo the clock name into this file::
600 521
601 # echo global > trace_clock 522 # echo global > trace_clock
602 523
603 Setting a clock clears the ring buffer 524 Setting a clock clears the ring buffer content as well as the
604 "snapshot" buffer. 525 "snapshot" buffer.
605 526
606 trace_marker: 527 trace_marker:
607 528
608 This is a very useful file for synchro 529 This is a very useful file for synchronizing user space
609 with events happening in the kernel. W 530 with events happening in the kernel. Writing strings into
610 this file will be written into the ftr 531 this file will be written into the ftrace buffer.
611 532
612 It is useful in applications to open t 533 It is useful in applications to open this file at the start
613 of the application and just reference 534 of the application and just reference the file descriptor
614 for the file:: 535 for the file::
615 536
616 void trace_write(const char *f 537 void trace_write(const char *fmt, ...)
617 { 538 {
618 va_list ap; 539 va_list ap;
619 char buf[256]; 540 char buf[256];
620 int n; 541 int n;
621 542
622 if (trace_fd < 0) 543 if (trace_fd < 0)
623 return; 544 return;
624 545
625 va_start(ap, fmt); 546 va_start(ap, fmt);
626 n = vsnprintf(buf, 256 547 n = vsnprintf(buf, 256, fmt, ap);
627 va_end(ap); 548 va_end(ap);
628 549
629 write(trace_fd, buf, n 550 write(trace_fd, buf, n);
630 } 551 }
631 552
632 start:: 553 start::
633 554
634 trace_fd = open("trace_marker" !! 555 trace_fd = open("trace_marker", WR_ONLY);
635 556
636 Note: Writing into the trace_marker fi 557 Note: Writing into the trace_marker file can also initiate triggers
637 that are written into /sys/kerne 558 that are written into /sys/kernel/tracing/events/ftrace/print/trigger
638 See "Event triggers" in Document 559 See "Event triggers" in Documentation/trace/events.rst and an
639 example in Documentation/trace/h 560 example in Documentation/trace/histogram.rst (Section 3.)
640 561
641 trace_marker_raw: 562 trace_marker_raw:
642 563
643 This is similar to trace_marker above, 564 This is similar to trace_marker above, but is meant for binary data
644 to be written to it, where a tool can 565 to be written to it, where a tool can be used to parse the data
645 from trace_pipe_raw. 566 from trace_pipe_raw.
646 567
647 uprobe_events: 568 uprobe_events:
648 569
649 Add dynamic tracepoints in programs. 570 Add dynamic tracepoints in programs.
650 See uprobetracer.rst 571 See uprobetracer.rst
651 572
652 uprobe_profile: 573 uprobe_profile:
653 574
654 Uprobe statistics. See uprobetrace.txt 575 Uprobe statistics. See uprobetrace.txt
655 576
656 instances: 577 instances:
657 578
658 This is a way to make multiple trace b 579 This is a way to make multiple trace buffers where different
659 events can be recorded in different bu 580 events can be recorded in different buffers.
660 See "Instances" section below. 581 See "Instances" section below.
661 582
662 events: 583 events:
663 584
664 This is the trace event directory. It 585 This is the trace event directory. It holds event tracepoints
665 (also known as static tracepoints) tha 586 (also known as static tracepoints) that have been compiled
666 into the kernel. It shows what event t 587 into the kernel. It shows what event tracepoints exist
667 and how they are grouped by system. Th 588 and how they are grouped by system. There are "enable"
668 files at various levels that can enabl 589 files at various levels that can enable the tracepoints
669 when a "1" is written to them. 590 when a "1" is written to them.
670 591
671 See events.rst for more information. 592 See events.rst for more information.
672 593
673 set_event: 594 set_event:
674 595
675 By echoing in the event into this file 596 By echoing in the event into this file, will enable that event.
676 597
677 See events.rst for more information. 598 See events.rst for more information.
678 599
679 available_events: 600 available_events:
680 601
681 A list of events that can be enabled i 602 A list of events that can be enabled in tracing.
682 603
683 See events.rst for more information. 604 See events.rst for more information.
684 605
685 timestamp_mode: 606 timestamp_mode:
686 607
687 Certain tracers may change the timesta 608 Certain tracers may change the timestamp mode used when
688 logging trace events into the event bu 609 logging trace events into the event buffer. Events with
689 different modes can coexist within a b 610 different modes can coexist within a buffer but the mode in
690 effect when an event is logged determi 611 effect when an event is logged determines which timestamp mode
691 is used for that event. The default t 612 is used for that event. The default timestamp mode is
692 'delta'. 613 'delta'.
693 614
694 Usual timestamp modes for tracing: 615 Usual timestamp modes for tracing:
695 616
696 # cat timestamp_mode 617 # cat timestamp_mode
697 [delta] absolute 618 [delta] absolute
698 619
699 The timestamp mode with the square b 620 The timestamp mode with the square brackets around it is the
700 one in effect. 621 one in effect.
701 622
702 delta: Default timestamp mode - time 623 delta: Default timestamp mode - timestamp is a delta against
703 a per-buffer timestamp. 624 a per-buffer timestamp.
704 625
705 absolute: The timestamp is a full ti 626 absolute: The timestamp is a full timestamp, not a delta
706 against some other value. As 627 against some other value. As such it takes up more
707 space and is less efficient. 628 space and is less efficient.
708 629
709 hwlat_detector: 630 hwlat_detector:
710 631
711 Directory for the Hardware Latency Det 632 Directory for the Hardware Latency Detector.
712 See "Hardware Latency Detector" sectio 633 See "Hardware Latency Detector" section below.
713 634
714 per_cpu: 635 per_cpu:
715 636
716 This is a directory that contains the 637 This is a directory that contains the trace per_cpu information.
717 638
718 per_cpu/cpu0/buffer_size_kb: 639 per_cpu/cpu0/buffer_size_kb:
719 640
720 The ftrace buffer is defined per_cpu. 641 The ftrace buffer is defined per_cpu. That is, there's a separate
721 buffer for each CPU to allow writes to 642 buffer for each CPU to allow writes to be done atomically,
722 and free from cache bouncing. These bu 643 and free from cache bouncing. These buffers may have different
723 size buffers. This file is similar to 644 size buffers. This file is similar to the buffer_size_kb
724 file, but it only displays or sets the 645 file, but it only displays or sets the buffer size for the
725 specific CPU. (here cpu0). 646 specific CPU. (here cpu0).
726 647
727 per_cpu/cpu0/trace: 648 per_cpu/cpu0/trace:
728 649
729 This is similar to the "trace" file, b 650 This is similar to the "trace" file, but it will only display
730 the data specific for the CPU. If writ 651 the data specific for the CPU. If written to, it only clears
731 the specific CPU buffer. 652 the specific CPU buffer.
732 653
733 per_cpu/cpu0/trace_pipe 654 per_cpu/cpu0/trace_pipe
734 655
735 This is similar to the "trace_pipe" fi 656 This is similar to the "trace_pipe" file, and is a consuming
736 read, but it will only display (and co 657 read, but it will only display (and consume) the data specific
737 for the CPU. 658 for the CPU.
738 659
739 per_cpu/cpu0/trace_pipe_raw 660 per_cpu/cpu0/trace_pipe_raw
740 661
741 For tools that can parse the ftrace ri 662 For tools that can parse the ftrace ring buffer binary format,
742 the trace_pipe_raw file can be used to 663 the trace_pipe_raw file can be used to extract the data
743 from the ring buffer directly. With th 664 from the ring buffer directly. With the use of the splice()
744 system call, the buffer data can be qu 665 system call, the buffer data can be quickly transferred to
745 a file or to the network where a serve 666 a file or to the network where a server is collecting the
746 data. 667 data.
747 668
748 Like trace_pipe, this is a consuming r 669 Like trace_pipe, this is a consuming reader, where multiple
749 reads will always produce different da 670 reads will always produce different data.
750 671
751 per_cpu/cpu0/snapshot: 672 per_cpu/cpu0/snapshot:
752 673
753 This is similar to the main "snapshot" 674 This is similar to the main "snapshot" file, but will only
754 snapshot the current CPU (if supported 675 snapshot the current CPU (if supported). It only displays
755 the content of the snapshot for a give 676 the content of the snapshot for a given CPU, and if
756 written to, only clears this CPU buffe 677 written to, only clears this CPU buffer.
757 678
758 per_cpu/cpu0/snapshot_raw: 679 per_cpu/cpu0/snapshot_raw:
759 680
760 Similar to the trace_pipe_raw, but wil 681 Similar to the trace_pipe_raw, but will read the binary format
761 from the snapshot buffer for the given 682 from the snapshot buffer for the given CPU.
762 683
763 per_cpu/cpu0/stats: 684 per_cpu/cpu0/stats:
764 685
765 This displays certain stats about the 686 This displays certain stats about the ring buffer:
766 687
767 entries: 688 entries:
768 The number of events that are 689 The number of events that are still in the buffer.
769 690
770 overrun: 691 overrun:
771 The number of lost events due 692 The number of lost events due to overwriting when
772 the buffer was full. 693 the buffer was full.
773 694
774 commit overrun: 695 commit overrun:
775 Should always be zero. 696 Should always be zero.
776 This gets set if so many event 697 This gets set if so many events happened within a nested
777 event (ring buffer is re-entra 698 event (ring buffer is re-entrant), that it fills the
778 buffer and starts dropping eve 699 buffer and starts dropping events.
779 700
780 bytes: 701 bytes:
781 Bytes actually read (not overw 702 Bytes actually read (not overwritten).
782 703
783 oldest event ts: 704 oldest event ts:
784 The oldest timestamp in the bu 705 The oldest timestamp in the buffer
785 706
786 now ts: 707 now ts:
787 The current timestamp 708 The current timestamp
788 709
789 dropped events: 710 dropped events:
790 Events lost due to overwrite o 711 Events lost due to overwrite option being off.
791 712
792 read events: 713 read events:
793 The number of events read. 714 The number of events read.
794 715
795 The Tracers 716 The Tracers
796 ----------- 717 -----------
797 718
798 Here is the list of current tracers that may b 719 Here is the list of current tracers that may be configured.
799 720
800 "function" 721 "function"
801 722
802 Function call tracer to trace all kern 723 Function call tracer to trace all kernel functions.
803 724
804 "function_graph" 725 "function_graph"
805 726
806 Similar to the function tracer except 727 Similar to the function tracer except that the
807 function tracer probes the functions o 728 function tracer probes the functions on their entry
808 whereas the function graph tracer trac 729 whereas the function graph tracer traces on both entry
809 and exit of the functions. It then pro 730 and exit of the functions. It then provides the ability
810 to draw a graph of function calls simi 731 to draw a graph of function calls similar to C code
811 source. 732 source.
812 733
813 "blk" 734 "blk"
814 735
815 The block tracer. The tracer used by t 736 The block tracer. The tracer used by the blktrace user
816 application. 737 application.
817 738
818 "hwlat" 739 "hwlat"
819 740
820 The Hardware Latency tracer is used to 741 The Hardware Latency tracer is used to detect if the hardware
821 produces any latency. See "Hardware La 742 produces any latency. See "Hardware Latency Detector" section
822 below. 743 below.
823 744
824 "irqsoff" 745 "irqsoff"
825 746
826 Traces the areas that disable interrup 747 Traces the areas that disable interrupts and saves
827 the trace with the longest max latency 748 the trace with the longest max latency.
828 See tracing_max_latency. When a new ma 749 See tracing_max_latency. When a new max is recorded,
829 it replaces the old trace. It is best 750 it replaces the old trace. It is best to view this
830 trace with the latency-format option e 751 trace with the latency-format option enabled, which
831 happens automatically when the tracer 752 happens automatically when the tracer is selected.
832 753
833 "preemptoff" 754 "preemptoff"
834 755
835 Similar to irqsoff but traces and reco 756 Similar to irqsoff but traces and records the amount of
836 time for which preemption is disabled. 757 time for which preemption is disabled.
837 758
838 "preemptirqsoff" 759 "preemptirqsoff"
839 760
840 Similar to irqsoff and preemptoff, but 761 Similar to irqsoff and preemptoff, but traces and
841 records the largest time for which irq 762 records the largest time for which irqs and/or preemption
842 is disabled. 763 is disabled.
843 764
844 "wakeup" 765 "wakeup"
845 766
846 Traces and records the max latency tha 767 Traces and records the max latency that it takes for
847 the highest priority task to get sched 768 the highest priority task to get scheduled after
848 it has been woken up. 769 it has been woken up.
849 Traces all tasks as an average develop 770 Traces all tasks as an average developer would expect.
850 771
851 "wakeup_rt" 772 "wakeup_rt"
852 773
853 Traces and records the max latency tha 774 Traces and records the max latency that it takes for just
854 RT tasks (as the current "wakeup" does 775 RT tasks (as the current "wakeup" does). This is useful
855 for those interested in wake up timing 776 for those interested in wake up timings of RT tasks.
856 777
857 "wakeup_dl" 778 "wakeup_dl"
858 779
859 Traces and records the max latency tha 780 Traces and records the max latency that it takes for
860 a SCHED_DEADLINE task to be woken (as 781 a SCHED_DEADLINE task to be woken (as the "wakeup" and
861 "wakeup_rt" does). 782 "wakeup_rt" does).
862 783
863 "mmiotrace" 784 "mmiotrace"
864 785
865 A special tracer that is used to trace 786 A special tracer that is used to trace binary module.
866 It will trace all the calls that a mod 787 It will trace all the calls that a module makes to the
867 hardware. Everything it writes and rea 788 hardware. Everything it writes and reads from the I/O
868 as well. 789 as well.
869 790
870 "branch" 791 "branch"
871 792
872 This tracer can be configured when tra 793 This tracer can be configured when tracing likely/unlikely
873 calls within the kernel. It will trace 794 calls within the kernel. It will trace when a likely and
874 unlikely branch is hit and if it was c 795 unlikely branch is hit and if it was correct in its prediction
875 of being correct. 796 of being correct.
876 797
877 "nop" 798 "nop"
878 799
879 This is the "trace nothing" tracer. To 800 This is the "trace nothing" tracer. To remove all
880 tracers from tracing simply echo "nop" 801 tracers from tracing simply echo "nop" into
881 current_tracer. 802 current_tracer.
882 803
883 Error conditions 804 Error conditions
884 ---------------- 805 ----------------
885 806
886 For most ftrace commands, failure modes are 807 For most ftrace commands, failure modes are obvious and communicated
887 using standard return codes. 808 using standard return codes.
888 809
889 For other more involved commands, extended e 810 For other more involved commands, extended error information may be
890 available via the tracing/error_log file. F 811 available via the tracing/error_log file. For the commands that
891 support it, reading the tracing/error_log fi 812 support it, reading the tracing/error_log file after an error will
892 display more detailed information about what 813 display more detailed information about what went wrong, if
893 information is available. The tracing/error 814 information is available. The tracing/error_log file is a circular
894 error log displaying a small number (current 815 error log displaying a small number (currently, 8) of ftrace errors
895 for the last (8) failed commands. 816 for the last (8) failed commands.
896 817
897 The extended error information and usage tak 818 The extended error information and usage takes the form shown in
898 this example:: 819 this example::
899 820
900 # echo xxx > /sys/kernel/tracing/events/sc !! 821 # echo xxx > /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger
901 echo: write error: Invalid argument 822 echo: write error: Invalid argument
902 823
903 # cat /sys/kernel/tracing/error_log !! 824 # cat /sys/kernel/debug/tracing/error_log
904 [ 5348.887237] location: error: Couldn't y 825 [ 5348.887237] location: error: Couldn't yyy: zzz
905 Command: xxx 826 Command: xxx
906 ^ 827 ^
907 [ 7517.023364] location: error: Bad rrr: s 828 [ 7517.023364] location: error: Bad rrr: sss
908 Command: ppp qqq 829 Command: ppp qqq
909 ^ 830 ^
910 831
911 To clear the error log, echo the empty strin 832 To clear the error log, echo the empty string into it::
912 833
913 # echo > /sys/kernel/tracing/error_log !! 834 # echo > /sys/kernel/debug/tracing/error_log
914 835
915 Examples of using the tracer 836 Examples of using the tracer
916 ---------------------------- 837 ----------------------------
917 838
918 Here are typical examples of using the tracers 839 Here are typical examples of using the tracers when controlling
919 them only with the tracefs interface (without 840 them only with the tracefs interface (without using any
920 user-land utilities). 841 user-land utilities).
921 842
922 Output format: 843 Output format:
923 -------------- 844 --------------
924 845
925 Here is an example of the output format of the 846 Here is an example of the output format of the file "trace"::
926 847
927 # tracer: function 848 # tracer: function
928 # 849 #
929 # entries-in-buffer/entries-written: 140080/ 850 # entries-in-buffer/entries-written: 140080/250280 #P:4
930 # 851 #
931 # _-----=> irqs 852 # _-----=> irqs-off
932 # / _----=> need 853 # / _----=> need-resched
933 # | / _---=> hard 854 # | / _---=> hardirq/softirq
934 # || / _--=> pree 855 # || / _--=> preempt-depth
935 # ||| / delay 856 # ||| / delay
936 # TASK-PID CPU# |||| TIMESTA 857 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
937 # | | | |||| | 858 # | | | |||| | |
938 bash-1977 [000] .... 17284.9936 859 bash-1977 [000] .... 17284.993652: sys_close <-system_call_fastpath
939 bash-1977 [000] .... 17284.9936 860 bash-1977 [000] .... 17284.993653: __close_fd <-sys_close
940 bash-1977 [000] .... 17284.9936 861 bash-1977 [000] .... 17284.993653: _raw_spin_lock <-__close_fd
941 sshd-1974 [003] .... 17284.9936 862 sshd-1974 [003] .... 17284.993653: __srcu_read_unlock <-fsnotify
942 bash-1977 [000] .... 17284.9936 863 bash-1977 [000] .... 17284.993654: add_preempt_count <-_raw_spin_lock
943 bash-1977 [000] ...1 17284.9936 864 bash-1977 [000] ...1 17284.993655: _raw_spin_unlock <-__close_fd
944 bash-1977 [000] ...1 17284.9936 865 bash-1977 [000] ...1 17284.993656: sub_preempt_count <-_raw_spin_unlock
945 bash-1977 [000] .... 17284.9936 866 bash-1977 [000] .... 17284.993657: filp_close <-__close_fd
946 bash-1977 [000] .... 17284.9936 867 bash-1977 [000] .... 17284.993657: dnotify_flush <-filp_close
947 sshd-1974 [003] .... 17284.9936 868 sshd-1974 [003] .... 17284.993658: sys_select <-system_call_fastpath
948 .... 869 ....
949 870
950 A header is printed with the tracer name that 871 A header is printed with the tracer name that is represented by
951 the trace. In this case the tracer is "functio 872 the trace. In this case the tracer is "function". Then it shows the
952 number of events in the buffer as well as the 873 number of events in the buffer as well as the total number of entries
953 that were written. The difference is the numbe 874 that were written. The difference is the number of entries that were
954 lost due to the buffer filling up (250280 - 14 875 lost due to the buffer filling up (250280 - 140080 = 110200 events
955 lost). 876 lost).
956 877
957 The header explains the content of the events. 878 The header explains the content of the events. Task name "bash", the task
958 PID "1977", the CPU that it was running on "00 879 PID "1977", the CPU that it was running on "000", the latency format
959 (explained below), the timestamp in <secs>.<us 880 (explained below), the timestamp in <secs>.<usecs> format, the
960 function name that was traced "sys_close" and 881 function name that was traced "sys_close" and the parent function that
961 called this function "system_call_fastpath". T 882 called this function "system_call_fastpath". The timestamp is the time
962 at which the function was entered. 883 at which the function was entered.
963 884
964 Latency trace format 885 Latency trace format
965 -------------------- 886 --------------------
966 887
967 When the latency-format option is enabled or w 888 When the latency-format option is enabled or when one of the latency
968 tracers is set, the trace file gives somewhat 889 tracers is set, the trace file gives somewhat more information to see
969 why a latency happened. Here is a typical trac 890 why a latency happened. Here is a typical trace::
970 891
971 # tracer: irqsoff 892 # tracer: irqsoff
972 # 893 #
973 # irqsoff latency trace v1.1.5 on 3.8.0-test 894 # irqsoff latency trace v1.1.5 on 3.8.0-test+
974 # ------------------------------------------ 895 # --------------------------------------------------------------------
975 # latency: 259 us, #4/4, CPU#2 | (M:preempt 896 # latency: 259 us, #4/4, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
976 # ----------------- 897 # -----------------
977 # | task: ps-6143 (uid:0 nice:0 policy:0 898 # | task: ps-6143 (uid:0 nice:0 policy:0 rt_prio:0)
978 # ----------------- 899 # -----------------
979 # => started at: __lock_task_sighand 900 # => started at: __lock_task_sighand
980 # => ended at: _raw_spin_unlock_irqrestor 901 # => ended at: _raw_spin_unlock_irqrestore
981 # 902 #
982 # 903 #
983 # _------=> CPU# 904 # _------=> CPU#
984 # / _-----=> irqs-off 905 # / _-----=> irqs-off
985 # | / _----=> need-resched 906 # | / _----=> need-resched
986 # || / _---=> hardirq/softirq 907 # || / _---=> hardirq/softirq
987 # ||| / _--=> preempt-depth 908 # ||| / _--=> preempt-depth
988 # |||| / delay 909 # |||| / delay
989 # cmd pid ||||| time | caller 910 # cmd pid ||||| time | caller
990 # \ / ||||| \ | / 911 # \ / ||||| \ | /
991 ps-6143 2d... 0us!: trace_hardir 912 ps-6143 2d... 0us!: trace_hardirqs_off <-__lock_task_sighand
992 ps-6143 2d..1 259us+: trace_hardir 913 ps-6143 2d..1 259us+: trace_hardirqs_on <-_raw_spin_unlock_irqrestore
993 ps-6143 2d..1 263us+: time_hardirq 914 ps-6143 2d..1 263us+: time_hardirqs_on <-_raw_spin_unlock_irqrestore
994 ps-6143 2d..1 306us : <stack trace 915 ps-6143 2d..1 306us : <stack trace>
995 => trace_hardirqs_on_caller 916 => trace_hardirqs_on_caller
996 => trace_hardirqs_on 917 => trace_hardirqs_on
997 => _raw_spin_unlock_irqrestore 918 => _raw_spin_unlock_irqrestore
998 => do_task_stat 919 => do_task_stat
999 => proc_tgid_stat 920 => proc_tgid_stat
1000 => proc_single_show 921 => proc_single_show
1001 => seq_read 922 => seq_read
1002 => vfs_read 923 => vfs_read
1003 => sys_read 924 => sys_read
1004 => system_call_fastpath 925 => system_call_fastpath
1005 926
1006 927
1007 This shows that the current tracer is "irqsof 928 This shows that the current tracer is "irqsoff" tracing the time
1008 for which interrupts were disabled. It gives 929 for which interrupts were disabled. It gives the trace version (which
1009 never changes) and the version of the kernel 930 never changes) and the version of the kernel upon which this was executed on
1010 (3.8). Then it displays the max latency in mi 931 (3.8). Then it displays the max latency in microseconds (259 us). The number
1011 of trace entries displayed and the total numb 932 of trace entries displayed and the total number (both are four: #4/4).
1012 VP, KP, SP, and HP are always zero and are re 933 VP, KP, SP, and HP are always zero and are reserved for later use.
1013 #P is the number of online CPUs (#P:4). 934 #P is the number of online CPUs (#P:4).
1014 935
1015 The task is the process that was running when 936 The task is the process that was running when the latency
1016 occurred. (ps pid: 6143). 937 occurred. (ps pid: 6143).
1017 938
1018 The start and stop (the functions in which th 939 The start and stop (the functions in which the interrupts were
1019 disabled and enabled respectively) that cause 940 disabled and enabled respectively) that caused the latencies:
1020 941
1021 - __lock_task_sighand is where the interrup 942 - __lock_task_sighand is where the interrupts were disabled.
1022 - _raw_spin_unlock_irqrestore is where they 943 - _raw_spin_unlock_irqrestore is where they were enabled again.
1023 944
1024 The next lines after the header are the trace 945 The next lines after the header are the trace itself. The header
1025 explains which is which. 946 explains which is which.
1026 947
1027 cmd: The name of the process in the trace. 948 cmd: The name of the process in the trace.
1028 949
1029 pid: The PID of that process. 950 pid: The PID of that process.
1030 951
1031 CPU#: The CPU which the process was running 952 CPU#: The CPU which the process was running on.
1032 953
1033 irqs-off: 'd' interrupts are disabled. '.' 954 irqs-off: 'd' interrupts are disabled. '.' otherwise.
1034 .. caution:: If the architecture does 955 .. caution:: If the architecture does not support a way to
1035 read the irq flags variable, 956 read the irq flags variable, an 'X' will always
1036 be printed here. 957 be printed here.
1037 958
1038 need-resched: 959 need-resched:
1039 - 'N' both TIF_NEED_RESCHED and PREEM 960 - 'N' both TIF_NEED_RESCHED and PREEMPT_NEED_RESCHED is set,
1040 - 'n' only TIF_NEED_RESCHED is set, 961 - 'n' only TIF_NEED_RESCHED is set,
1041 - 'p' only PREEMPT_NEED_RESCHED is se 962 - 'p' only PREEMPT_NEED_RESCHED is set,
1042 - '.' otherwise. 963 - '.' otherwise.
1043 964
1044 hardirq/softirq: 965 hardirq/softirq:
1045 - 'Z' - NMI occurred inside a hardirq 966 - 'Z' - NMI occurred inside a hardirq
1046 - 'z' - NMI is running 967 - 'z' - NMI is running
1047 - 'H' - hard irq occurred inside a so 968 - 'H' - hard irq occurred inside a softirq.
1048 - 'h' - hard irq is running 969 - 'h' - hard irq is running
1049 - 's' - soft irq is running 970 - 's' - soft irq is running
1050 - '.' - normal context. 971 - '.' - normal context.
1051 972
1052 preempt-depth: The level of preempt_disable 973 preempt-depth: The level of preempt_disabled
1053 974
1054 The above is mostly meaningful for kernel dev 975 The above is mostly meaningful for kernel developers.
1055 976
1056 time: 977 time:
1057 When the latency-format option is ena 978 When the latency-format option is enabled, the trace file
1058 output includes a timestamp relative 979 output includes a timestamp relative to the start of the
1059 trace. This differs from the output w 980 trace. This differs from the output when latency-format
1060 is disabled, which includes an absolu 981 is disabled, which includes an absolute timestamp.
1061 982
1062 delay: 983 delay:
1063 This is just to help catch your eye a 984 This is just to help catch your eye a bit better. And
1064 needs to be fixed to be only relative 985 needs to be fixed to be only relative to the same CPU.
1065 The marks are determined by the diffe 986 The marks are determined by the difference between this
1066 current trace and the next trace. 987 current trace and the next trace.
1067 988
1068 - '$' - greater than 1 second 989 - '$' - greater than 1 second
1069 - '@' - greater than 100 millisecon 990 - '@' - greater than 100 millisecond
1070 - '*' - greater than 10 millisecond 991 - '*' - greater than 10 millisecond
1071 - '#' - greater than 1000 microseco 992 - '#' - greater than 1000 microsecond
1072 - '!' - greater than 100 microsecon 993 - '!' - greater than 100 microsecond
1073 - '+' - greater than 10 microsecond 994 - '+' - greater than 10 microsecond
1074 - ' ' - less than or equal to 10 mi 995 - ' ' - less than or equal to 10 microsecond.
1075 996
1076 The rest is the same as the 'trace' file. 997 The rest is the same as the 'trace' file.
1077 998
1078 Note, the latency tracers will usually end 999 Note, the latency tracers will usually end with a back trace
1079 to easily find where the latency occurred. 1000 to easily find where the latency occurred.
1080 1001
1081 trace_options 1002 trace_options
1082 ------------- 1003 -------------
1083 1004
1084 The trace_options file (or the options direct 1005 The trace_options file (or the options directory) is used to control
1085 what gets printed in the trace output, or man 1006 what gets printed in the trace output, or manipulate the tracers.
1086 To see what is available, simply cat the file 1007 To see what is available, simply cat the file::
1087 1008
1088 cat trace_options 1009 cat trace_options
1089 print-parent 1010 print-parent
1090 nosym-offset 1011 nosym-offset
1091 nosym-addr 1012 nosym-addr
1092 noverbose 1013 noverbose
1093 noraw 1014 noraw
1094 nohex 1015 nohex
1095 nobin 1016 nobin
1096 noblock 1017 noblock
1097 nofields <<
1098 trace_printk 1018 trace_printk
1099 annotate 1019 annotate
1100 nouserstacktrace 1020 nouserstacktrace
1101 nosym-userobj 1021 nosym-userobj
1102 noprintk-msg-only 1022 noprintk-msg-only
1103 context-info 1023 context-info
1104 nolatency-format 1024 nolatency-format
1105 record-cmd 1025 record-cmd
1106 norecord-tgid 1026 norecord-tgid
1107 overwrite 1027 overwrite
1108 nodisable_on_free 1028 nodisable_on_free
1109 irq-info 1029 irq-info
1110 markers 1030 markers
1111 noevent-fork 1031 noevent-fork
1112 function-trace 1032 function-trace
1113 nofunction-fork 1033 nofunction-fork
1114 nodisplay-graph 1034 nodisplay-graph
1115 nostacktrace 1035 nostacktrace
1116 nobranch 1036 nobranch
1117 1037
1118 To disable one of the options, echo in the op 1038 To disable one of the options, echo in the option prepended with
1119 "no":: 1039 "no"::
1120 1040
1121 echo noprint-parent > trace_options 1041 echo noprint-parent > trace_options
1122 1042
1123 To enable an option, leave off the "no":: 1043 To enable an option, leave off the "no"::
1124 1044
1125 echo sym-offset > trace_options 1045 echo sym-offset > trace_options
1126 1046
1127 Here are the available options: 1047 Here are the available options:
1128 1048
1129 print-parent 1049 print-parent
1130 On function traces, display the calli 1050 On function traces, display the calling (parent)
1131 function as well as the function bein 1051 function as well as the function being traced.
1132 :: 1052 ::
1133 1053
1134 print-parent: 1054 print-parent:
1135 bash-4000 [01] 1477.606694: simp 1055 bash-4000 [01] 1477.606694: simple_strtoul <-kstrtoul
1136 1056
1137 noprint-parent: 1057 noprint-parent:
1138 bash-4000 [01] 1477.606694: simp 1058 bash-4000 [01] 1477.606694: simple_strtoul
1139 1059
1140 1060
1141 sym-offset 1061 sym-offset
1142 Display not only the function name, b 1062 Display not only the function name, but also the
1143 offset in the function. For example, 1063 offset in the function. For example, instead of
1144 seeing just "ktime_get", you will see 1064 seeing just "ktime_get", you will see
1145 "ktime_get+0xb/0x20". 1065 "ktime_get+0xb/0x20".
1146 :: 1066 ::
1147 1067
1148 sym-offset: 1068 sym-offset:
1149 bash-4000 [01] 1477.606694: simp 1069 bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0
1150 1070
1151 sym-addr 1071 sym-addr
1152 This will also display the function a 1072 This will also display the function address as well
1153 as the function name. 1073 as the function name.
1154 :: 1074 ::
1155 1075
1156 sym-addr: 1076 sym-addr:
1157 bash-4000 [01] 1477.606694: simp 1077 bash-4000 [01] 1477.606694: simple_strtoul <c0339346>
1158 1078
1159 verbose 1079 verbose
1160 This deals with the trace file when t 1080 This deals with the trace file when the
1161 latency-format option is enabled. 1081 latency-format option is enabled.
1162 :: 1082 ::
1163 1083
1164 bash 4000 1 0 00000000 00010a95 1084 bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \
1165 (+0.000ms): simple_strtoul (kstrt 1085 (+0.000ms): simple_strtoul (kstrtoul)
1166 1086
1167 raw 1087 raw
1168 This will display raw numbers. This o 1088 This will display raw numbers. This option is best for
1169 use with user applications that can t 1089 use with user applications that can translate the raw
1170 numbers better than having it done in 1090 numbers better than having it done in the kernel.
1171 1091
1172 hex 1092 hex
1173 Similar to raw, but the numbers will 1093 Similar to raw, but the numbers will be in a hexadecimal format.
1174 1094
1175 bin 1095 bin
1176 This will print out the formats in ra 1096 This will print out the formats in raw binary.
1177 1097
1178 block 1098 block
1179 When set, reading trace_pipe will not 1099 When set, reading trace_pipe will not block when polled.
1180 1100
1181 fields <<
1182 Print the fields as described by thei <<
1183 option than using hex, bin or raw, as <<
1184 of the content of the event. <<
1185 <<
1186 trace_printk 1101 trace_printk
1187 Can disable trace_printk() from writi 1102 Can disable trace_printk() from writing into the buffer.
1188 1103
1189 trace_printk_dest <<
1190 Set to have trace_printk() and simila <<
1191 write into this instance. Note, only <<
1192 this set. By setting this flag, it cl <<
1193 of the instance that had it set previ <<
1194 level trace has this set, and will ge <<
1195 instance has it set then clears it. <<
1196 <<
1197 This flag cannot be cleared by the to <<
1198 default instance. The only way the to <<
1199 cleared, is by it being set in anothe <<
1200 <<
1201 annotate 1104 annotate
1202 It is sometimes confusing when the CP 1105 It is sometimes confusing when the CPU buffers are full
1203 and one CPU buffer had a lot of event 1106 and one CPU buffer had a lot of events recently, thus
1204 a shorter time frame, were another CP 1107 a shorter time frame, were another CPU may have only had
1205 a few events, which lets it have olde 1108 a few events, which lets it have older events. When
1206 the trace is reported, it shows the o 1109 the trace is reported, it shows the oldest events first,
1207 and it may look like only one CPU ran 1110 and it may look like only one CPU ran (the one with the
1208 oldest events). When the annotate opt 1111 oldest events). When the annotate option is set, it will
1209 display when a new CPU buffer started 1112 display when a new CPU buffer started::
1210 1113
1211 <idle>-0 [001] 1114 <idle>-0 [001] dNs4 21169.031481: wake_up_idle_cpu <-add_timer_on
1212 <idle>-0 [001] 1115 <idle>-0 [001] dNs4 21169.031482: _raw_spin_unlock_irqrestore <-add_timer_on
1213 <idle>-0 [001] 1116 <idle>-0 [001] .Ns4 21169.031484: sub_preempt_count <-_raw_spin_unlock_irqrestore
1214 ##### CPU 2 buffer started ## 1117 ##### CPU 2 buffer started ####
1215 <idle>-0 [002] 1118 <idle>-0 [002] .N.1 21169.031484: rcu_idle_exit <-cpu_idle
1216 <idle>-0 [001] 1119 <idle>-0 [001] .Ns3 21169.031484: _raw_spin_unlock <-clocksource_watchdog
1217 <idle>-0 [001] 1120 <idle>-0 [001] .Ns3 21169.031485: sub_preempt_count <-_raw_spin_unlock
1218 1121
1219 userstacktrace 1122 userstacktrace
1220 This option changes the trace. It rec 1123 This option changes the trace. It records a
1221 stacktrace of the current user space 1124 stacktrace of the current user space thread after
1222 each trace event. 1125 each trace event.
1223 1126
1224 sym-userobj 1127 sym-userobj
1225 when user stacktrace are enabled, loo 1128 when user stacktrace are enabled, look up which
1226 object the address belongs to, and pr 1129 object the address belongs to, and print a
1227 relative address. This is especially 1130 relative address. This is especially useful when
1228 ASLR is on, otherwise you don't get a 1131 ASLR is on, otherwise you don't get a chance to
1229 resolve the address to object/file/li 1132 resolve the address to object/file/line after
1230 the app is no longer running 1133 the app is no longer running
1231 1134
1232 The lookup is performed when you read 1135 The lookup is performed when you read
1233 trace,trace_pipe. Example:: 1136 trace,trace_pipe. Example::
1234 1137
1235 a.out-1623 [000] 40874.465 1138 a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0
1236 x494] <- /root/a.out[+0x4a8 1139 x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6]
1237 1140
1238 1141
1239 printk-msg-only 1142 printk-msg-only
1240 When set, trace_printk()s will only s 1143 When set, trace_printk()s will only show the format
1241 and not their parameters (if trace_bp 1144 and not their parameters (if trace_bprintk() or
1242 trace_bputs() was used to save the tr 1145 trace_bputs() was used to save the trace_printk()).
1243 1146
1244 context-info 1147 context-info
1245 Show only the event data. Hides the c 1148 Show only the event data. Hides the comm, PID,
1246 timestamp, CPU, and other useful data 1149 timestamp, CPU, and other useful data.
1247 1150
1248 latency-format 1151 latency-format
1249 This option changes the trace output. 1152 This option changes the trace output. When it is enabled,
1250 the trace displays additional informa 1153 the trace displays additional information about the
1251 latency, as described in "Latency tra 1154 latency, as described in "Latency trace format".
1252 1155
1253 pause-on-trace 1156 pause-on-trace
1254 When set, opening the trace file for 1157 When set, opening the trace file for read, will pause
1255 writing to the ring buffer (as if tra 1158 writing to the ring buffer (as if tracing_on was set to zero).
1256 This simulates the original behavior 1159 This simulates the original behavior of the trace file.
1257 When the file is closed, tracing will 1160 When the file is closed, tracing will be enabled again.
1258 1161
1259 hash-ptr 1162 hash-ptr
1260 When set, "%p" in the event printk fo 1163 When set, "%p" in the event printk format displays the
1261 hashed pointer value instead of real 1164 hashed pointer value instead of real address.
1262 This will be useful if you want to fi 1165 This will be useful if you want to find out which hashed
1263 value is corresponding to the real va 1166 value is corresponding to the real value in trace log.
1264 1167
1265 record-cmd 1168 record-cmd
1266 When any event or tracer is enabled, 1169 When any event or tracer is enabled, a hook is enabled
1267 in the sched_switch trace point to fi 1170 in the sched_switch trace point to fill comm cache
1268 with mapped pids and comms. But this 1171 with mapped pids and comms. But this may cause some
1269 overhead, and if you only care about 1172 overhead, and if you only care about pids, and not the
1270 name of the task, disabling this opti 1173 name of the task, disabling this option can lower the
1271 impact of tracing. See "saved_cmdline 1174 impact of tracing. See "saved_cmdlines".
1272 1175
1273 record-tgid 1176 record-tgid
1274 When any event or tracer is enabled, 1177 When any event or tracer is enabled, a hook is enabled
1275 in the sched_switch trace point to fi 1178 in the sched_switch trace point to fill the cache of
1276 mapped Thread Group IDs (TGID) mappin 1179 mapped Thread Group IDs (TGID) mapping to pids. See
1277 "saved_tgids". 1180 "saved_tgids".
1278 1181
1279 overwrite 1182 overwrite
1280 This controls what happens when the t 1183 This controls what happens when the trace buffer is
1281 full. If "1" (default), the oldest ev 1184 full. If "1" (default), the oldest events are
1282 discarded and overwritten. If "0", th 1185 discarded and overwritten. If "0", then the newest
1283 events are discarded. 1186 events are discarded.
1284 (see per_cpu/cpu0/stats for overrun a 1187 (see per_cpu/cpu0/stats for overrun and dropped)
1285 1188
1286 disable_on_free 1189 disable_on_free
1287 When the free_buffer is closed, traci 1190 When the free_buffer is closed, tracing will
1288 stop (tracing_on set to 0). 1191 stop (tracing_on set to 0).
1289 1192
1290 irq-info 1193 irq-info
1291 Shows the interrupt, preempt count, n 1194 Shows the interrupt, preempt count, need resched data.
1292 When disabled, the trace looks like:: 1195 When disabled, the trace looks like::
1293 1196
1294 # tracer: function 1197 # tracer: function
1295 # 1198 #
1296 # entries-in-buffer/entries-w 1199 # entries-in-buffer/entries-written: 144405/9452052 #P:4
1297 # 1200 #
1298 # TASK-PID CPU# 1201 # TASK-PID CPU# TIMESTAMP FUNCTION
1299 # | | | 1202 # | | | | |
1300 <idle>-0 [002] 1203 <idle>-0 [002] 23636.756054: ttwu_do_activate.constprop.89 <-try_to_wake_up
1301 <idle>-0 [002] 1204 <idle>-0 [002] 23636.756054: activate_task <-ttwu_do_activate.constprop.89
1302 <idle>-0 [002] 1205 <idle>-0 [002] 23636.756055: enqueue_task <-activate_task
1303 1206
1304 1207
1305 markers 1208 markers
1306 When set, the trace_marker is writabl 1209 When set, the trace_marker is writable (only by root).
1307 When disabled, the trace_marker will 1210 When disabled, the trace_marker will error with EINVAL
1308 on write. 1211 on write.
1309 1212
1310 event-fork 1213 event-fork
1311 When set, tasks with PIDs listed in s 1214 When set, tasks with PIDs listed in set_event_pid will have
1312 the PIDs of their children added to s 1215 the PIDs of their children added to set_event_pid when those
1313 tasks fork. Also, when tasks with PID 1216 tasks fork. Also, when tasks with PIDs in set_event_pid exit,
1314 their PIDs will be removed from the f 1217 their PIDs will be removed from the file.
1315 1218
1316 This affects PIDs listed in set_event 1219 This affects PIDs listed in set_event_notrace_pid as well.
1317 1220
1318 function-trace 1221 function-trace
1319 The latency tracers will enable funct 1222 The latency tracers will enable function tracing
1320 if this option is enabled (default it 1223 if this option is enabled (default it is). When
1321 it is disabled, the latency tracers d 1224 it is disabled, the latency tracers do not trace
1322 functions. This keeps the overhead of 1225 functions. This keeps the overhead of the tracer down
1323 when performing latency tests. 1226 when performing latency tests.
1324 1227
1325 function-fork 1228 function-fork
1326 When set, tasks with PIDs listed in s 1229 When set, tasks with PIDs listed in set_ftrace_pid will
1327 have the PIDs of their children added 1230 have the PIDs of their children added to set_ftrace_pid
1328 when those tasks fork. Also, when tas 1231 when those tasks fork. Also, when tasks with PIDs in
1329 set_ftrace_pid exit, their PIDs will 1232 set_ftrace_pid exit, their PIDs will be removed from the
1330 file. 1233 file.
1331 1234
1332 This affects PIDs in set_ftrace_notra 1235 This affects PIDs in set_ftrace_notrace_pid as well.
1333 1236
1334 display-graph 1237 display-graph
1335 When set, the latency tracers (irqsof 1238 When set, the latency tracers (irqsoff, wakeup, etc) will
1336 use function graph tracing instead of 1239 use function graph tracing instead of function tracing.
1337 1240
1338 stacktrace 1241 stacktrace
1339 When set, a stack trace is recorded a 1242 When set, a stack trace is recorded after any trace event
1340 is recorded. 1243 is recorded.
1341 1244
1342 branch 1245 branch
1343 Enable branch tracing with the tracer 1246 Enable branch tracing with the tracer. This enables branch
1344 tracer along with the currently set t 1247 tracer along with the currently set tracer. Enabling this
1345 with the "nop" tracer is the same as 1248 with the "nop" tracer is the same as just enabling the
1346 "branch" tracer. 1249 "branch" tracer.
1347 1250
1348 .. tip:: Some tracers have their own options. 1251 .. tip:: Some tracers have their own options. They only appear in this
1349 file when the tracer is active. They a 1252 file when the tracer is active. They always appear in the
1350 options directory. 1253 options directory.
1351 1254
1352 1255
1353 Here are the per tracer options: 1256 Here are the per tracer options:
1354 1257
1355 Options for function tracer: 1258 Options for function tracer:
1356 1259
1357 func_stack_trace 1260 func_stack_trace
1358 When set, a stack trace is recorded a 1261 When set, a stack trace is recorded after every
1359 function that is recorded. NOTE! Limi 1262 function that is recorded. NOTE! Limit the functions
1360 that are recorded before enabling thi 1263 that are recorded before enabling this, with
1361 "set_ftrace_filter" otherwise the sys 1264 "set_ftrace_filter" otherwise the system performance
1362 will be critically degraded. Remember 1265 will be critically degraded. Remember to disable
1363 this option before clearing the funct 1266 this option before clearing the function filter.
1364 1267
1365 Options for function_graph tracer: 1268 Options for function_graph tracer:
1366 1269
1367 Since the function_graph tracer has a slight 1270 Since the function_graph tracer has a slightly different output
1368 it has its own options to control what is di 1271 it has its own options to control what is displayed.
1369 1272
1370 funcgraph-overrun 1273 funcgraph-overrun
1371 When set, the "overrun" of the graph 1274 When set, the "overrun" of the graph stack is
1372 displayed after each function traced. 1275 displayed after each function traced. The
1373 overrun, is when the stack depth of t 1276 overrun, is when the stack depth of the calls
1374 is greater than what is reserved for 1277 is greater than what is reserved for each task.
1375 Each task has a fixed array of functi 1278 Each task has a fixed array of functions to
1376 trace in the call graph. If the depth 1279 trace in the call graph. If the depth of the
1377 calls exceeds that, the function is n 1280 calls exceeds that, the function is not traced.
1378 The overrun is the number of function 1281 The overrun is the number of functions missed
1379 due to exceeding this array. 1282 due to exceeding this array.
1380 1283
1381 funcgraph-cpu 1284 funcgraph-cpu
1382 When set, the CPU number of the CPU w 1285 When set, the CPU number of the CPU where the trace
1383 occurred is displayed. 1286 occurred is displayed.
1384 1287
1385 funcgraph-overhead 1288 funcgraph-overhead
1386 When set, if the function takes longe 1289 When set, if the function takes longer than
1387 A certain amount, then a delay marker 1290 A certain amount, then a delay marker is
1388 displayed. See "delay" above, under t 1291 displayed. See "delay" above, under the
1389 header description. 1292 header description.
1390 1293
1391 funcgraph-proc 1294 funcgraph-proc
1392 Unlike other tracers, the process' co 1295 Unlike other tracers, the process' command line
1393 is not displayed by default, but inst 1296 is not displayed by default, but instead only
1394 when a task is traced in and out duri 1297 when a task is traced in and out during a context
1395 switch. Enabling this options has the 1298 switch. Enabling this options has the command
1396 of each process displayed at every li 1299 of each process displayed at every line.
1397 1300
1398 funcgraph-duration 1301 funcgraph-duration
1399 At the end of each function (the retu 1302 At the end of each function (the return)
1400 the duration of the amount of time in 1303 the duration of the amount of time in the
1401 function is displayed in microseconds 1304 function is displayed in microseconds.
1402 1305
1403 funcgraph-abstime 1306 funcgraph-abstime
1404 When set, the timestamp is displayed 1307 When set, the timestamp is displayed at each line.
1405 1308
1406 funcgraph-irqs 1309 funcgraph-irqs
1407 When disabled, functions that happen 1310 When disabled, functions that happen inside an
1408 interrupt will not be traced. 1311 interrupt will not be traced.
1409 1312
1410 funcgraph-tail 1313 funcgraph-tail
1411 When set, the return event will inclu 1314 When set, the return event will include the function
1412 that it represents. By default this i 1315 that it represents. By default this is off, and
1413 only a closing curly bracket "}" is d 1316 only a closing curly bracket "}" is displayed for
1414 the return of a function. 1317 the return of a function.
1415 1318
1416 funcgraph-retval <<
1417 When set, the return value of each tr <<
1418 will be printed after an equal sign " <<
1419 this is off. <<
1420 <<
1421 funcgraph-retval-hex <<
1422 When set, the return value will alway <<
1423 in hexadecimal format. If the option <<
1424 the return value is an error code, it <<
1425 in signed decimal format; otherwise i <<
1426 printed in hexadecimal format. By def <<
1427 is off. <<
1428 <<
1429 sleep-time 1319 sleep-time
1430 When running function graph tracer, t 1320 When running function graph tracer, to include
1431 the time a task schedules out in its 1321 the time a task schedules out in its function.
1432 When enabled, it will account time th 1322 When enabled, it will account time the task has been
1433 scheduled out as part of the function 1323 scheduled out as part of the function call.
1434 1324
1435 graph-time 1325 graph-time
1436 When running function profiler with f 1326 When running function profiler with function graph tracer,
1437 to include the time to call nested fu 1327 to include the time to call nested functions. When this is
1438 not set, the time reported for the fu 1328 not set, the time reported for the function will only
1439 include the time the function itself 1329 include the time the function itself executed for, not the
1440 time for functions that it called. 1330 time for functions that it called.
1441 1331
1442 Options for blk tracer: 1332 Options for blk tracer:
1443 1333
1444 blk_classic 1334 blk_classic
1445 Shows a more minimalistic output. 1335 Shows a more minimalistic output.
1446 1336
1447 1337
1448 irqsoff 1338 irqsoff
1449 ------- 1339 -------
1450 1340
1451 When interrupts are disabled, the CPU can not 1341 When interrupts are disabled, the CPU can not react to any other
1452 external event (besides NMIs and SMIs). This 1342 external event (besides NMIs and SMIs). This prevents the timer
1453 interrupt from triggering or the mouse interr 1343 interrupt from triggering or the mouse interrupt from letting
1454 the kernel know of a new mouse event. The res 1344 the kernel know of a new mouse event. The result is a latency
1455 with the reaction time. 1345 with the reaction time.
1456 1346
1457 The irqsoff tracer tracks the time for which 1347 The irqsoff tracer tracks the time for which interrupts are
1458 disabled. When a new maximum latency is hit, 1348 disabled. When a new maximum latency is hit, the tracer saves
1459 the trace leading up to that latency point so 1349 the trace leading up to that latency point so that every time a
1460 new maximum is reached, the old saved trace i 1350 new maximum is reached, the old saved trace is discarded and the
1461 new trace is saved. 1351 new trace is saved.
1462 1352
1463 To reset the maximum, echo 0 into tracing_max 1353 To reset the maximum, echo 0 into tracing_max_latency. Here is
1464 an example:: 1354 an example::
1465 1355
1466 # echo 0 > options/function-trace 1356 # echo 0 > options/function-trace
1467 # echo irqsoff > current_tracer 1357 # echo irqsoff > current_tracer
1468 # echo 1 > tracing_on 1358 # echo 1 > tracing_on
1469 # echo 0 > tracing_max_latency 1359 # echo 0 > tracing_max_latency
1470 # ls -ltr 1360 # ls -ltr
1471 [...] 1361 [...]
1472 # echo 0 > tracing_on 1362 # echo 0 > tracing_on
1473 # cat trace 1363 # cat trace
1474 # tracer: irqsoff 1364 # tracer: irqsoff
1475 # 1365 #
1476 # irqsoff latency trace v1.1.5 on 3.8.0-tes 1366 # irqsoff latency trace v1.1.5 on 3.8.0-test+
1477 # ----------------------------------------- 1367 # --------------------------------------------------------------------
1478 # latency: 16 us, #4/4, CPU#0 | (M:preempt 1368 # latency: 16 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1479 # ----------------- 1369 # -----------------
1480 # | task: swapper/0-0 (uid:0 nice:0 poli 1370 # | task: swapper/0-0 (uid:0 nice:0 policy:0 rt_prio:0)
1481 # ----------------- 1371 # -----------------
1482 # => started at: run_timer_softirq 1372 # => started at: run_timer_softirq
1483 # => ended at: run_timer_softirq 1373 # => ended at: run_timer_softirq
1484 # 1374 #
1485 # 1375 #
1486 # _------=> CPU# 1376 # _------=> CPU#
1487 # / _-----=> irqs-off 1377 # / _-----=> irqs-off
1488 # | / _----=> need-resched 1378 # | / _----=> need-resched
1489 # || / _---=> hardirq/softir 1379 # || / _---=> hardirq/softirq
1490 # ||| / _--=> preempt-depth 1380 # ||| / _--=> preempt-depth
1491 # |||| / delay 1381 # |||| / delay
1492 # cmd pid ||||| time | caller 1382 # cmd pid ||||| time | caller
1493 # \ / ||||| \ | / 1383 # \ / ||||| \ | /
1494 <idle>-0 0d.s2 0us+: _raw_spin_l 1384 <idle>-0 0d.s2 0us+: _raw_spin_lock_irq <-run_timer_softirq
1495 <idle>-0 0dNs3 17us : _raw_spin_u 1385 <idle>-0 0dNs3 17us : _raw_spin_unlock_irq <-run_timer_softirq
1496 <idle>-0 0dNs3 17us+: trace_hardi 1386 <idle>-0 0dNs3 17us+: trace_hardirqs_on <-run_timer_softirq
1497 <idle>-0 0dNs3 25us : <stack trac 1387 <idle>-0 0dNs3 25us : <stack trace>
1498 => _raw_spin_unlock_irq 1388 => _raw_spin_unlock_irq
1499 => run_timer_softirq 1389 => run_timer_softirq
1500 => __do_softirq 1390 => __do_softirq
1501 => call_softirq 1391 => call_softirq
1502 => do_softirq 1392 => do_softirq
1503 => irq_exit 1393 => irq_exit
1504 => smp_apic_timer_interrupt 1394 => smp_apic_timer_interrupt
1505 => apic_timer_interrupt 1395 => apic_timer_interrupt
1506 => rcu_idle_exit 1396 => rcu_idle_exit
1507 => cpu_idle 1397 => cpu_idle
1508 => rest_init 1398 => rest_init
1509 => start_kernel 1399 => start_kernel
1510 => x86_64_start_reservations 1400 => x86_64_start_reservations
1511 => x86_64_start_kernel 1401 => x86_64_start_kernel
1512 1402
1513 Here we see that we had a latency of 16 micro 1403 Here we see that we had a latency of 16 microseconds (which is
1514 very good). The _raw_spin_lock_irq in run_tim 1404 very good). The _raw_spin_lock_irq in run_timer_softirq disabled
1515 interrupts. The difference between the 16 and 1405 interrupts. The difference between the 16 and the displayed
1516 timestamp 25us occurred because the clock was 1406 timestamp 25us occurred because the clock was incremented
1517 between the time of recording the max latency 1407 between the time of recording the max latency and the time of
1518 recording the function that had that latency. 1408 recording the function that had that latency.
1519 1409
1520 Note the above example had function-trace not 1410 Note the above example had function-trace not set. If we set
1521 function-trace, we get a much larger output:: 1411 function-trace, we get a much larger output::
1522 1412
1523 with echo 1 > options/function-trace 1413 with echo 1 > options/function-trace
1524 1414
1525 # tracer: irqsoff 1415 # tracer: irqsoff
1526 # 1416 #
1527 # irqsoff latency trace v1.1.5 on 3.8.0-tes 1417 # irqsoff latency trace v1.1.5 on 3.8.0-test+
1528 # ----------------------------------------- 1418 # --------------------------------------------------------------------
1529 # latency: 71 us, #168/168, CPU#3 | (M:pree 1419 # latency: 71 us, #168/168, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1530 # ----------------- 1420 # -----------------
1531 # | task: bash-2042 (uid:0 nice:0 policy 1421 # | task: bash-2042 (uid:0 nice:0 policy:0 rt_prio:0)
1532 # ----------------- 1422 # -----------------
1533 # => started at: ata_scsi_queuecmd 1423 # => started at: ata_scsi_queuecmd
1534 # => ended at: ata_scsi_queuecmd 1424 # => ended at: ata_scsi_queuecmd
1535 # 1425 #
1536 # 1426 #
1537 # _------=> CPU# 1427 # _------=> CPU#
1538 # / _-----=> irqs-off 1428 # / _-----=> irqs-off
1539 # | / _----=> need-resched 1429 # | / _----=> need-resched
1540 # || / _---=> hardirq/softir 1430 # || / _---=> hardirq/softirq
1541 # ||| / _--=> preempt-depth 1431 # ||| / _--=> preempt-depth
1542 # |||| / delay 1432 # |||| / delay
1543 # cmd pid ||||| time | caller 1433 # cmd pid ||||| time | caller
1544 # \ / ||||| \ | / 1434 # \ / ||||| \ | /
1545 bash-2042 3d... 0us : _raw_spin_l 1435 bash-2042 3d... 0us : _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1546 bash-2042 3d... 0us : add_preempt 1436 bash-2042 3d... 0us : add_preempt_count <-_raw_spin_lock_irqsave
1547 bash-2042 3d..1 1us : ata_scsi_fi 1437 bash-2042 3d..1 1us : ata_scsi_find_dev <-ata_scsi_queuecmd
1548 bash-2042 3d..1 1us : __ata_scsi_ 1438 bash-2042 3d..1 1us : __ata_scsi_find_dev <-ata_scsi_find_dev
1549 bash-2042 3d..1 2us : ata_find_de 1439 bash-2042 3d..1 2us : ata_find_dev.part.14 <-__ata_scsi_find_dev
1550 bash-2042 3d..1 2us : ata_qc_new_ 1440 bash-2042 3d..1 2us : ata_qc_new_init <-__ata_scsi_queuecmd
1551 bash-2042 3d..1 3us : ata_sg_init 1441 bash-2042 3d..1 3us : ata_sg_init <-__ata_scsi_queuecmd
1552 bash-2042 3d..1 4us : ata_scsi_rw 1442 bash-2042 3d..1 4us : ata_scsi_rw_xlat <-__ata_scsi_queuecmd
1553 bash-2042 3d..1 4us : ata_build_r 1443 bash-2042 3d..1 4us : ata_build_rw_tf <-ata_scsi_rw_xlat
1554 [...] 1444 [...]
1555 bash-2042 3d..1 67us : delay_tsc < 1445 bash-2042 3d..1 67us : delay_tsc <-__delay
1556 bash-2042 3d..1 67us : add_preempt 1446 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1557 bash-2042 3d..2 67us : sub_preempt 1447 bash-2042 3d..2 67us : sub_preempt_count <-delay_tsc
1558 bash-2042 3d..1 67us : add_preempt 1448 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1559 bash-2042 3d..2 68us : sub_preempt 1449 bash-2042 3d..2 68us : sub_preempt_count <-delay_tsc
1560 bash-2042 3d..1 68us+: ata_bmdma_s 1450 bash-2042 3d..1 68us+: ata_bmdma_start <-ata_bmdma_qc_issue
1561 bash-2042 3d..1 71us : _raw_spin_u 1451 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1562 bash-2042 3d..1 71us : _raw_spin_u 1452 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1563 bash-2042 3d..1 72us+: trace_hardi 1453 bash-2042 3d..1 72us+: trace_hardirqs_on <-ata_scsi_queuecmd
1564 bash-2042 3d..1 120us : <stack trac 1454 bash-2042 3d..1 120us : <stack trace>
1565 => _raw_spin_unlock_irqrestore 1455 => _raw_spin_unlock_irqrestore
1566 => ata_scsi_queuecmd 1456 => ata_scsi_queuecmd
1567 => scsi_dispatch_cmd 1457 => scsi_dispatch_cmd
1568 => scsi_request_fn 1458 => scsi_request_fn
1569 => __blk_run_queue_uncond 1459 => __blk_run_queue_uncond
1570 => __blk_run_queue 1460 => __blk_run_queue
1571 => blk_queue_bio 1461 => blk_queue_bio
1572 => submit_bio_noacct 1462 => submit_bio_noacct
1573 => submit_bio 1463 => submit_bio
1574 => submit_bh 1464 => submit_bh
1575 => __ext3_get_inode_loc 1465 => __ext3_get_inode_loc
1576 => ext3_iget 1466 => ext3_iget
1577 => ext3_lookup 1467 => ext3_lookup
1578 => lookup_real 1468 => lookup_real
1579 => __lookup_hash 1469 => __lookup_hash
1580 => walk_component 1470 => walk_component
1581 => lookup_last 1471 => lookup_last
1582 => path_lookupat 1472 => path_lookupat
1583 => filename_lookup 1473 => filename_lookup
1584 => user_path_at_empty 1474 => user_path_at_empty
1585 => user_path_at 1475 => user_path_at
1586 => vfs_fstatat 1476 => vfs_fstatat
1587 => vfs_stat 1477 => vfs_stat
1588 => sys_newstat 1478 => sys_newstat
1589 => system_call_fastpath 1479 => system_call_fastpath
1590 1480
1591 1481
1592 Here we traced a 71 microsecond latency. But 1482 Here we traced a 71 microsecond latency. But we also see all the
1593 functions that were called during that time. 1483 functions that were called during that time. Note that by
1594 enabling function tracing, we incur an added 1484 enabling function tracing, we incur an added overhead. This
1595 overhead may extend the latency times. But ne 1485 overhead may extend the latency times. But nevertheless, this
1596 trace has provided some very helpful debuggin 1486 trace has provided some very helpful debugging information.
1597 1487
1598 If we prefer function graph output instead of 1488 If we prefer function graph output instead of function, we can set
1599 display-graph option:: 1489 display-graph option::
1600 1490
1601 with echo 1 > options/display-graph 1491 with echo 1 > options/display-graph
1602 1492
1603 # tracer: irqsoff 1493 # tracer: irqsoff
1604 # 1494 #
1605 # irqsoff latency trace v1.1.5 on 4.20.0-rc 1495 # irqsoff latency trace v1.1.5 on 4.20.0-rc6+
1606 # ----------------------------------------- 1496 # --------------------------------------------------------------------
1607 # latency: 3751 us, #274/274, CPU#0 | (M:de 1497 # latency: 3751 us, #274/274, CPU#0 | (M:desktop VP:0, KP:0, SP:0 HP:0 #P:4)
1608 # ----------------- 1498 # -----------------
1609 # | task: bash-1507 (uid:0 nice:0 policy 1499 # | task: bash-1507 (uid:0 nice:0 policy:0 rt_prio:0)
1610 # ----------------- 1500 # -----------------
1611 # => started at: free_debug_processing 1501 # => started at: free_debug_processing
1612 # => ended at: return_to_handler 1502 # => ended at: return_to_handler
1613 # 1503 #
1614 # 1504 #
1615 # _-- 1505 # _-----=> irqs-off
1616 # / _- 1506 # / _----=> need-resched
1617 # | / _ 1507 # | / _---=> hardirq/softirq
1618 # || / 1508 # || / _--=> preempt-depth
1619 # ||| / 1509 # ||| /
1620 # REL TIME CPU TASK/PID |||| 1510 # REL TIME CPU TASK/PID |||| DURATION FUNCTION CALLS
1621 # | | | | |||| 1511 # | | | | |||| | | | | | |
1622 0 us | 0) bash-1507 | d... 1512 0 us | 0) bash-1507 | d... | 0.000 us | _raw_spin_lock_irqsave();
1623 0 us | 0) bash-1507 | d..1 1513 0 us | 0) bash-1507 | d..1 | 0.378 us | do_raw_spin_trylock();
1624 1 us | 0) bash-1507 | d..2 1514 1 us | 0) bash-1507 | d..2 | | set_track() {
1625 2 us | 0) bash-1507 | d..2 1515 2 us | 0) bash-1507 | d..2 | | save_stack_trace() {
1626 2 us | 0) bash-1507 | d..2 1516 2 us | 0) bash-1507 | d..2 | | __save_stack_trace() {
1627 3 us | 0) bash-1507 | d..2 1517 3 us | 0) bash-1507 | d..2 | | __unwind_start() {
1628 3 us | 0) bash-1507 | d..2 1518 3 us | 0) bash-1507 | d..2 | | get_stack_info() {
1629 3 us | 0) bash-1507 | d..2 1519 3 us | 0) bash-1507 | d..2 | 0.351 us | in_task_stack();
1630 4 us | 0) bash-1507 | d..2 1520 4 us | 0) bash-1507 | d..2 | 1.107 us | }
1631 [...] 1521 [...]
1632 3750 us | 0) bash-1507 | d..1 1522 3750 us | 0) bash-1507 | d..1 | 0.516 us | do_raw_spin_unlock();
1633 3750 us | 0) bash-1507 | d..1 1523 3750 us | 0) bash-1507 | d..1 | 0.000 us | _raw_spin_unlock_irqrestore();
1634 3764 us | 0) bash-1507 | d..1 1524 3764 us | 0) bash-1507 | d..1 | 0.000 us | tracer_hardirqs_on();
1635 bash-1507 0d..1 3792us : <stack trac 1525 bash-1507 0d..1 3792us : <stack trace>
1636 => free_debug_processing 1526 => free_debug_processing
1637 => __slab_free 1527 => __slab_free
1638 => kmem_cache_free 1528 => kmem_cache_free
1639 => vm_area_free 1529 => vm_area_free
1640 => remove_vma 1530 => remove_vma
1641 => exit_mmap 1531 => exit_mmap
1642 => mmput 1532 => mmput
1643 => begin_new_exec 1533 => begin_new_exec
1644 => load_elf_binary 1534 => load_elf_binary
1645 => search_binary_handler 1535 => search_binary_handler
1646 => __do_execve_file.isra.32 1536 => __do_execve_file.isra.32
1647 => __x64_sys_execve 1537 => __x64_sys_execve
1648 => do_syscall_64 1538 => do_syscall_64
1649 => entry_SYSCALL_64_after_hwframe 1539 => entry_SYSCALL_64_after_hwframe
1650 1540
1651 preemptoff 1541 preemptoff
1652 ---------- 1542 ----------
1653 1543
1654 When preemption is disabled, we may be able t 1544 When preemption is disabled, we may be able to receive
1655 interrupts but the task cannot be preempted a 1545 interrupts but the task cannot be preempted and a higher
1656 priority task must wait for preemption to be 1546 priority task must wait for preemption to be enabled again
1657 before it can preempt a lower priority task. 1547 before it can preempt a lower priority task.
1658 1548
1659 The preemptoff tracer traces the places that 1549 The preemptoff tracer traces the places that disable preemption.
1660 Like the irqsoff tracer, it records the maxim 1550 Like the irqsoff tracer, it records the maximum latency for
1661 which preemption was disabled. The control of 1551 which preemption was disabled. The control of preemptoff tracer
1662 is much like the irqsoff tracer. 1552 is much like the irqsoff tracer.
1663 :: 1553 ::
1664 1554
1665 # echo 0 > options/function-trace 1555 # echo 0 > options/function-trace
1666 # echo preemptoff > current_tracer 1556 # echo preemptoff > current_tracer
1667 # echo 1 > tracing_on 1557 # echo 1 > tracing_on
1668 # echo 0 > tracing_max_latency 1558 # echo 0 > tracing_max_latency
1669 # ls -ltr 1559 # ls -ltr
1670 [...] 1560 [...]
1671 # echo 0 > tracing_on 1561 # echo 0 > tracing_on
1672 # cat trace 1562 # cat trace
1673 # tracer: preemptoff 1563 # tracer: preemptoff
1674 # 1564 #
1675 # preemptoff latency trace v1.1.5 on 3.8.0- 1565 # preemptoff latency trace v1.1.5 on 3.8.0-test+
1676 # ----------------------------------------- 1566 # --------------------------------------------------------------------
1677 # latency: 46 us, #4/4, CPU#1 | (M:preempt 1567 # latency: 46 us, #4/4, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1678 # ----------------- 1568 # -----------------
1679 # | task: sshd-1991 (uid:0 nice:0 policy 1569 # | task: sshd-1991 (uid:0 nice:0 policy:0 rt_prio:0)
1680 # ----------------- 1570 # -----------------
1681 # => started at: do_IRQ 1571 # => started at: do_IRQ
1682 # => ended at: do_IRQ 1572 # => ended at: do_IRQ
1683 # 1573 #
1684 # 1574 #
1685 # _------=> CPU# 1575 # _------=> CPU#
1686 # / _-----=> irqs-off 1576 # / _-----=> irqs-off
1687 # | / _----=> need-resched 1577 # | / _----=> need-resched
1688 # || / _---=> hardirq/softir 1578 # || / _---=> hardirq/softirq
1689 # ||| / _--=> preempt-depth 1579 # ||| / _--=> preempt-depth
1690 # |||| / delay 1580 # |||| / delay
1691 # cmd pid ||||| time | caller 1581 # cmd pid ||||| time | caller
1692 # \ / ||||| \ | / 1582 # \ / ||||| \ | /
1693 sshd-1991 1d.h. 0us+: irq_enter < 1583 sshd-1991 1d.h. 0us+: irq_enter <-do_IRQ
1694 sshd-1991 1d..1 46us : irq_exit <- 1584 sshd-1991 1d..1 46us : irq_exit <-do_IRQ
1695 sshd-1991 1d..1 47us+: trace_preem 1585 sshd-1991 1d..1 47us+: trace_preempt_on <-do_IRQ
1696 sshd-1991 1d..1 52us : <stack trac 1586 sshd-1991 1d..1 52us : <stack trace>
1697 => sub_preempt_count 1587 => sub_preempt_count
1698 => irq_exit 1588 => irq_exit
1699 => do_IRQ 1589 => do_IRQ
1700 => ret_from_intr 1590 => ret_from_intr
1701 1591
1702 1592
1703 This has some more changes. Preemption was di 1593 This has some more changes. Preemption was disabled when an
1704 interrupt came in (notice the 'h'), and was e 1594 interrupt came in (notice the 'h'), and was enabled on exit.
1705 But we also see that interrupts have been dis 1595 But we also see that interrupts have been disabled when entering
1706 the preempt off section and leaving it (the ' 1596 the preempt off section and leaving it (the 'd'). We do not know if
1707 interrupts were enabled in the mean time or s 1597 interrupts were enabled in the mean time or shortly after this
1708 was over. 1598 was over.
1709 :: 1599 ::
1710 1600
1711 # tracer: preemptoff 1601 # tracer: preemptoff
1712 # 1602 #
1713 # preemptoff latency trace v1.1.5 on 3.8.0- 1603 # preemptoff latency trace v1.1.5 on 3.8.0-test+
1714 # ----------------------------------------- 1604 # --------------------------------------------------------------------
1715 # latency: 83 us, #241/241, CPU#1 | (M:pree 1605 # latency: 83 us, #241/241, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1716 # ----------------- 1606 # -----------------
1717 # | task: bash-1994 (uid:0 nice:0 policy 1607 # | task: bash-1994 (uid:0 nice:0 policy:0 rt_prio:0)
1718 # ----------------- 1608 # -----------------
1719 # => started at: wake_up_new_task 1609 # => started at: wake_up_new_task
1720 # => ended at: task_rq_unlock 1610 # => ended at: task_rq_unlock
1721 # 1611 #
1722 # 1612 #
1723 # _------=> CPU# 1613 # _------=> CPU#
1724 # / _-----=> irqs-off 1614 # / _-----=> irqs-off
1725 # | / _----=> need-resched 1615 # | / _----=> need-resched
1726 # || / _---=> hardirq/softir 1616 # || / _---=> hardirq/softirq
1727 # ||| / _--=> preempt-depth 1617 # ||| / _--=> preempt-depth
1728 # |||| / delay 1618 # |||| / delay
1729 # cmd pid ||||| time | caller 1619 # cmd pid ||||| time | caller
1730 # \ / ||||| \ | / 1620 # \ / ||||| \ | /
1731 bash-1994 1d..1 0us : _raw_spin_l 1621 bash-1994 1d..1 0us : _raw_spin_lock_irqsave <-wake_up_new_task
1732 bash-1994 1d..1 0us : select_task 1622 bash-1994 1d..1 0us : select_task_rq_fair <-select_task_rq
1733 bash-1994 1d..1 1us : __rcu_read_ 1623 bash-1994 1d..1 1us : __rcu_read_lock <-select_task_rq_fair
1734 bash-1994 1d..1 1us : source_load 1624 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
1735 bash-1994 1d..1 1us : source_load 1625 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
1736 [...] 1626 [...]
1737 bash-1994 1d..1 12us : irq_enter < 1627 bash-1994 1d..1 12us : irq_enter <-smp_apic_timer_interrupt
1738 bash-1994 1d..1 12us : rcu_irq_ent 1628 bash-1994 1d..1 12us : rcu_irq_enter <-irq_enter
1739 bash-1994 1d..1 13us : add_preempt 1629 bash-1994 1d..1 13us : add_preempt_count <-irq_enter
1740 bash-1994 1d.h1 13us : exit_idle < 1630 bash-1994 1d.h1 13us : exit_idle <-smp_apic_timer_interrupt
1741 bash-1994 1d.h1 13us : hrtimer_int 1631 bash-1994 1d.h1 13us : hrtimer_interrupt <-smp_apic_timer_interrupt
1742 bash-1994 1d.h1 13us : _raw_spin_l 1632 bash-1994 1d.h1 13us : _raw_spin_lock <-hrtimer_interrupt
1743 bash-1994 1d.h1 14us : add_preempt 1633 bash-1994 1d.h1 14us : add_preempt_count <-_raw_spin_lock
1744 bash-1994 1d.h2 14us : ktime_get_u 1634 bash-1994 1d.h2 14us : ktime_get_update_offsets <-hrtimer_interrupt
1745 [...] 1635 [...]
1746 bash-1994 1d.h1 35us : lapic_next_ 1636 bash-1994 1d.h1 35us : lapic_next_event <-clockevents_program_event
1747 bash-1994 1d.h1 35us : irq_exit <- 1637 bash-1994 1d.h1 35us : irq_exit <-smp_apic_timer_interrupt
1748 bash-1994 1d.h1 36us : sub_preempt 1638 bash-1994 1d.h1 36us : sub_preempt_count <-irq_exit
1749 bash-1994 1d..2 36us : do_softirq 1639 bash-1994 1d..2 36us : do_softirq <-irq_exit
1750 bash-1994 1d..2 36us : __do_softir 1640 bash-1994 1d..2 36us : __do_softirq <-call_softirq
1751 bash-1994 1d..2 36us : __local_bh_ 1641 bash-1994 1d..2 36us : __local_bh_disable <-__do_softirq
1752 bash-1994 1d.s2 37us : add_preempt 1642 bash-1994 1d.s2 37us : add_preempt_count <-_raw_spin_lock_irq
1753 bash-1994 1d.s3 38us : _raw_spin_u 1643 bash-1994 1d.s3 38us : _raw_spin_unlock <-run_timer_softirq
1754 bash-1994 1d.s3 39us : sub_preempt 1644 bash-1994 1d.s3 39us : sub_preempt_count <-_raw_spin_unlock
1755 bash-1994 1d.s2 39us : call_timer_ 1645 bash-1994 1d.s2 39us : call_timer_fn <-run_timer_softirq
1756 [...] 1646 [...]
1757 bash-1994 1dNs2 81us : cpu_needs_a 1647 bash-1994 1dNs2 81us : cpu_needs_another_gp <-rcu_process_callbacks
1758 bash-1994 1dNs2 82us : __local_bh_ 1648 bash-1994 1dNs2 82us : __local_bh_enable <-__do_softirq
1759 bash-1994 1dNs2 82us : sub_preempt 1649 bash-1994 1dNs2 82us : sub_preempt_count <-__local_bh_enable
1760 bash-1994 1dN.2 82us : idle_cpu <- 1650 bash-1994 1dN.2 82us : idle_cpu <-irq_exit
1761 bash-1994 1dN.2 83us : rcu_irq_exi 1651 bash-1994 1dN.2 83us : rcu_irq_exit <-irq_exit
1762 bash-1994 1dN.2 83us : sub_preempt 1652 bash-1994 1dN.2 83us : sub_preempt_count <-irq_exit
1763 bash-1994 1.N.1 84us : _raw_spin_u 1653 bash-1994 1.N.1 84us : _raw_spin_unlock_irqrestore <-task_rq_unlock
1764 bash-1994 1.N.1 84us+: trace_preem 1654 bash-1994 1.N.1 84us+: trace_preempt_on <-task_rq_unlock
1765 bash-1994 1.N.1 104us : <stack trac 1655 bash-1994 1.N.1 104us : <stack trace>
1766 => sub_preempt_count 1656 => sub_preempt_count
1767 => _raw_spin_unlock_irqrestore 1657 => _raw_spin_unlock_irqrestore
1768 => task_rq_unlock 1658 => task_rq_unlock
1769 => wake_up_new_task 1659 => wake_up_new_task
1770 => do_fork 1660 => do_fork
1771 => sys_clone 1661 => sys_clone
1772 => stub_clone 1662 => stub_clone
1773 1663
1774 1664
1775 The above is an example of the preemptoff tra 1665 The above is an example of the preemptoff trace with
1776 function-trace set. Here we see that interrup 1666 function-trace set. Here we see that interrupts were not disabled
1777 the entire time. The irq_enter code lets us k 1667 the entire time. The irq_enter code lets us know that we entered
1778 an interrupt 'h'. Before that, the functions 1668 an interrupt 'h'. Before that, the functions being traced still
1779 show that it is not in an interrupt, but we c 1669 show that it is not in an interrupt, but we can see from the
1780 functions themselves that this is not the cas 1670 functions themselves that this is not the case.
1781 1671
1782 preemptirqsoff 1672 preemptirqsoff
1783 -------------- 1673 --------------
1784 1674
1785 Knowing the locations that have interrupts di 1675 Knowing the locations that have interrupts disabled or
1786 preemption disabled for the longest times is 1676 preemption disabled for the longest times is helpful. But
1787 sometimes we would like to know when either p 1677 sometimes we would like to know when either preemption and/or
1788 interrupts are disabled. 1678 interrupts are disabled.
1789 1679
1790 Consider the following code:: 1680 Consider the following code::
1791 1681
1792 local_irq_disable(); 1682 local_irq_disable();
1793 call_function_with_irqs_off(); 1683 call_function_with_irqs_off();
1794 preempt_disable(); 1684 preempt_disable();
1795 call_function_with_irqs_and_preemption_of 1685 call_function_with_irqs_and_preemption_off();
1796 local_irq_enable(); 1686 local_irq_enable();
1797 call_function_with_preemption_off(); 1687 call_function_with_preemption_off();
1798 preempt_enable(); 1688 preempt_enable();
1799 1689
1800 The irqsoff tracer will record the total leng 1690 The irqsoff tracer will record the total length of
1801 call_function_with_irqs_off() and 1691 call_function_with_irqs_off() and
1802 call_function_with_irqs_and_preemption_off(). 1692 call_function_with_irqs_and_preemption_off().
1803 1693
1804 The preemptoff tracer will record the total l 1694 The preemptoff tracer will record the total length of
1805 call_function_with_irqs_and_preemption_off() 1695 call_function_with_irqs_and_preemption_off() and
1806 call_function_with_preemption_off(). 1696 call_function_with_preemption_off().
1807 1697
1808 But neither will trace the time that interrup 1698 But neither will trace the time that interrupts and/or
1809 preemption is disabled. This total time is th 1699 preemption is disabled. This total time is the time that we can
1810 not schedule. To record this time, use the pr 1700 not schedule. To record this time, use the preemptirqsoff
1811 tracer. 1701 tracer.
1812 1702
1813 Again, using this trace is much like the irqs 1703 Again, using this trace is much like the irqsoff and preemptoff
1814 tracers. 1704 tracers.
1815 :: 1705 ::
1816 1706
1817 # echo 0 > options/function-trace 1707 # echo 0 > options/function-trace
1818 # echo preemptirqsoff > current_tracer 1708 # echo preemptirqsoff > current_tracer
1819 # echo 1 > tracing_on 1709 # echo 1 > tracing_on
1820 # echo 0 > tracing_max_latency 1710 # echo 0 > tracing_max_latency
1821 # ls -ltr 1711 # ls -ltr
1822 [...] 1712 [...]
1823 # echo 0 > tracing_on 1713 # echo 0 > tracing_on
1824 # cat trace 1714 # cat trace
1825 # tracer: preemptirqsoff 1715 # tracer: preemptirqsoff
1826 # 1716 #
1827 # preemptirqsoff latency trace v1.1.5 on 3. 1717 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1828 # ----------------------------------------- 1718 # --------------------------------------------------------------------
1829 # latency: 100 us, #4/4, CPU#3 | (M:preempt 1719 # latency: 100 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1830 # ----------------- 1720 # -----------------
1831 # | task: ls-2230 (uid:0 nice:0 policy:0 1721 # | task: ls-2230 (uid:0 nice:0 policy:0 rt_prio:0)
1832 # ----------------- 1722 # -----------------
1833 # => started at: ata_scsi_queuecmd 1723 # => started at: ata_scsi_queuecmd
1834 # => ended at: ata_scsi_queuecmd 1724 # => ended at: ata_scsi_queuecmd
1835 # 1725 #
1836 # 1726 #
1837 # _------=> CPU# 1727 # _------=> CPU#
1838 # / _-----=> irqs-off 1728 # / _-----=> irqs-off
1839 # | / _----=> need-resched 1729 # | / _----=> need-resched
1840 # || / _---=> hardirq/softir 1730 # || / _---=> hardirq/softirq
1841 # ||| / _--=> preempt-depth 1731 # ||| / _--=> preempt-depth
1842 # |||| / delay 1732 # |||| / delay
1843 # cmd pid ||||| time | caller 1733 # cmd pid ||||| time | caller
1844 # \ / ||||| \ | / 1734 # \ / ||||| \ | /
1845 ls-2230 3d... 0us+: _raw_spin_l 1735 ls-2230 3d... 0us+: _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1846 ls-2230 3...1 100us : _raw_spin_u 1736 ls-2230 3...1 100us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1847 ls-2230 3...1 101us+: trace_preem 1737 ls-2230 3...1 101us+: trace_preempt_on <-ata_scsi_queuecmd
1848 ls-2230 3...1 111us : <stack trac 1738 ls-2230 3...1 111us : <stack trace>
1849 => sub_preempt_count 1739 => sub_preempt_count
1850 => _raw_spin_unlock_irqrestore 1740 => _raw_spin_unlock_irqrestore
1851 => ata_scsi_queuecmd 1741 => ata_scsi_queuecmd
1852 => scsi_dispatch_cmd 1742 => scsi_dispatch_cmd
1853 => scsi_request_fn 1743 => scsi_request_fn
1854 => __blk_run_queue_uncond 1744 => __blk_run_queue_uncond
1855 => __blk_run_queue 1745 => __blk_run_queue
1856 => blk_queue_bio 1746 => blk_queue_bio
1857 => submit_bio_noacct 1747 => submit_bio_noacct
1858 => submit_bio 1748 => submit_bio
1859 => submit_bh 1749 => submit_bh
1860 => ext3_bread 1750 => ext3_bread
1861 => ext3_dir_bread 1751 => ext3_dir_bread
1862 => htree_dirblock_to_tree 1752 => htree_dirblock_to_tree
1863 => ext3_htree_fill_tree 1753 => ext3_htree_fill_tree
1864 => ext3_readdir 1754 => ext3_readdir
1865 => vfs_readdir 1755 => vfs_readdir
1866 => sys_getdents 1756 => sys_getdents
1867 => system_call_fastpath 1757 => system_call_fastpath
1868 1758
1869 1759
1870 The trace_hardirqs_off_thunk is called from a 1760 The trace_hardirqs_off_thunk is called from assembly on x86 when
1871 interrupts are disabled in the assembly code. 1761 interrupts are disabled in the assembly code. Without the
1872 function tracing, we do not know if interrupt 1762 function tracing, we do not know if interrupts were enabled
1873 within the preemption points. We do see that 1763 within the preemption points. We do see that it started with
1874 preemption enabled. 1764 preemption enabled.
1875 1765
1876 Here is a trace with function-trace set:: 1766 Here is a trace with function-trace set::
1877 1767
1878 # tracer: preemptirqsoff 1768 # tracer: preemptirqsoff
1879 # 1769 #
1880 # preemptirqsoff latency trace v1.1.5 on 3. 1770 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1881 # ----------------------------------------- 1771 # --------------------------------------------------------------------
1882 # latency: 161 us, #339/339, CPU#3 | (M:pre 1772 # latency: 161 us, #339/339, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1883 # ----------------- 1773 # -----------------
1884 # | task: ls-2269 (uid:0 nice:0 policy:0 1774 # | task: ls-2269 (uid:0 nice:0 policy:0 rt_prio:0)
1885 # ----------------- 1775 # -----------------
1886 # => started at: schedule 1776 # => started at: schedule
1887 # => ended at: mutex_unlock 1777 # => ended at: mutex_unlock
1888 # 1778 #
1889 # 1779 #
1890 # _------=> CPU# 1780 # _------=> CPU#
1891 # / _-----=> irqs-off 1781 # / _-----=> irqs-off
1892 # | / _----=> need-resched 1782 # | / _----=> need-resched
1893 # || / _---=> hardirq/softir 1783 # || / _---=> hardirq/softirq
1894 # ||| / _--=> preempt-depth 1784 # ||| / _--=> preempt-depth
1895 # |||| / delay 1785 # |||| / delay
1896 # cmd pid ||||| time | caller 1786 # cmd pid ||||| time | caller
1897 # \ / ||||| \ | / 1787 # \ / ||||| \ | /
1898 kworker/-59 3...1 0us : __schedule 1788 kworker/-59 3...1 0us : __schedule <-schedule
1899 kworker/-59 3d..1 0us : rcu_preempt 1789 kworker/-59 3d..1 0us : rcu_preempt_qs <-rcu_note_context_switch
1900 kworker/-59 3d..1 1us : add_preempt 1790 kworker/-59 3d..1 1us : add_preempt_count <-_raw_spin_lock_irq
1901 kworker/-59 3d..2 1us : deactivate_ 1791 kworker/-59 3d..2 1us : deactivate_task <-__schedule
1902 kworker/-59 3d..2 1us : dequeue_tas 1792 kworker/-59 3d..2 1us : dequeue_task <-deactivate_task
1903 kworker/-59 3d..2 2us : update_rq_c 1793 kworker/-59 3d..2 2us : update_rq_clock <-dequeue_task
1904 kworker/-59 3d..2 2us : dequeue_tas 1794 kworker/-59 3d..2 2us : dequeue_task_fair <-dequeue_task
1905 kworker/-59 3d..2 2us : update_curr 1795 kworker/-59 3d..2 2us : update_curr <-dequeue_task_fair
1906 kworker/-59 3d..2 2us : update_min_ 1796 kworker/-59 3d..2 2us : update_min_vruntime <-update_curr
1907 kworker/-59 3d..2 3us : cpuacct_cha 1797 kworker/-59 3d..2 3us : cpuacct_charge <-update_curr
1908 kworker/-59 3d..2 3us : __rcu_read_ 1798 kworker/-59 3d..2 3us : __rcu_read_lock <-cpuacct_charge
1909 kworker/-59 3d..2 3us : __rcu_read_ 1799 kworker/-59 3d..2 3us : __rcu_read_unlock <-cpuacct_charge
1910 kworker/-59 3d..2 3us : update_cfs_ 1800 kworker/-59 3d..2 3us : update_cfs_rq_blocked_load <-dequeue_task_fair
1911 kworker/-59 3d..2 4us : clear_buddi 1801 kworker/-59 3d..2 4us : clear_buddies <-dequeue_task_fair
1912 kworker/-59 3d..2 4us : account_ent 1802 kworker/-59 3d..2 4us : account_entity_dequeue <-dequeue_task_fair
1913 kworker/-59 3d..2 4us : update_min_ 1803 kworker/-59 3d..2 4us : update_min_vruntime <-dequeue_task_fair
1914 kworker/-59 3d..2 4us : update_cfs_ 1804 kworker/-59 3d..2 4us : update_cfs_shares <-dequeue_task_fair
1915 kworker/-59 3d..2 5us : hrtick_upda 1805 kworker/-59 3d..2 5us : hrtick_update <-dequeue_task_fair
1916 kworker/-59 3d..2 5us : wq_worker_s 1806 kworker/-59 3d..2 5us : wq_worker_sleeping <-__schedule
1917 kworker/-59 3d..2 5us : kthread_dat 1807 kworker/-59 3d..2 5us : kthread_data <-wq_worker_sleeping
1918 kworker/-59 3d..2 5us : put_prev_ta 1808 kworker/-59 3d..2 5us : put_prev_task_fair <-__schedule
1919 kworker/-59 3d..2 6us : pick_next_t 1809 kworker/-59 3d..2 6us : pick_next_task_fair <-pick_next_task
1920 kworker/-59 3d..2 6us : clear_buddi 1810 kworker/-59 3d..2 6us : clear_buddies <-pick_next_task_fair
1921 kworker/-59 3d..2 6us : set_next_en 1811 kworker/-59 3d..2 6us : set_next_entity <-pick_next_task_fair
1922 kworker/-59 3d..2 6us : update_stat 1812 kworker/-59 3d..2 6us : update_stats_wait_end <-set_next_entity
1923 ls-2269 3d..2 7us : finish_task 1813 ls-2269 3d..2 7us : finish_task_switch <-__schedule
1924 ls-2269 3d..2 7us : _raw_spin_u 1814 ls-2269 3d..2 7us : _raw_spin_unlock_irq <-finish_task_switch
1925 ls-2269 3d..2 8us : do_IRQ <-re 1815 ls-2269 3d..2 8us : do_IRQ <-ret_from_intr
1926 ls-2269 3d..2 8us : irq_enter < 1816 ls-2269 3d..2 8us : irq_enter <-do_IRQ
1927 ls-2269 3d..2 8us : rcu_irq_ent 1817 ls-2269 3d..2 8us : rcu_irq_enter <-irq_enter
1928 ls-2269 3d..2 9us : add_preempt 1818 ls-2269 3d..2 9us : add_preempt_count <-irq_enter
1929 ls-2269 3d.h2 9us : exit_idle < 1819 ls-2269 3d.h2 9us : exit_idle <-do_IRQ
1930 [...] 1820 [...]
1931 ls-2269 3d.h3 20us : sub_preempt 1821 ls-2269 3d.h3 20us : sub_preempt_count <-_raw_spin_unlock
1932 ls-2269 3d.h2 20us : irq_exit <- 1822 ls-2269 3d.h2 20us : irq_exit <-do_IRQ
1933 ls-2269 3d.h2 21us : sub_preempt 1823 ls-2269 3d.h2 21us : sub_preempt_count <-irq_exit
1934 ls-2269 3d..3 21us : do_softirq 1824 ls-2269 3d..3 21us : do_softirq <-irq_exit
1935 ls-2269 3d..3 21us : __do_softir 1825 ls-2269 3d..3 21us : __do_softirq <-call_softirq
1936 ls-2269 3d..3 21us+: __local_bh_ 1826 ls-2269 3d..3 21us+: __local_bh_disable <-__do_softirq
1937 ls-2269 3d.s4 29us : sub_preempt 1827 ls-2269 3d.s4 29us : sub_preempt_count <-_local_bh_enable_ip
1938 ls-2269 3d.s5 29us : sub_preempt 1828 ls-2269 3d.s5 29us : sub_preempt_count <-_local_bh_enable_ip
1939 ls-2269 3d.s5 31us : do_IRQ <-re 1829 ls-2269 3d.s5 31us : do_IRQ <-ret_from_intr
1940 ls-2269 3d.s5 31us : irq_enter < 1830 ls-2269 3d.s5 31us : irq_enter <-do_IRQ
1941 ls-2269 3d.s5 31us : rcu_irq_ent 1831 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1942 [...] 1832 [...]
1943 ls-2269 3d.s5 31us : rcu_irq_ent 1833 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1944 ls-2269 3d.s5 32us : add_preempt 1834 ls-2269 3d.s5 32us : add_preempt_count <-irq_enter
1945 ls-2269 3d.H5 32us : exit_idle < 1835 ls-2269 3d.H5 32us : exit_idle <-do_IRQ
1946 ls-2269 3d.H5 32us : handle_irq 1836 ls-2269 3d.H5 32us : handle_irq <-do_IRQ
1947 ls-2269 3d.H5 32us : irq_to_desc 1837 ls-2269 3d.H5 32us : irq_to_desc <-handle_irq
1948 ls-2269 3d.H5 33us : handle_fast 1838 ls-2269 3d.H5 33us : handle_fasteoi_irq <-handle_irq
1949 [...] 1839 [...]
1950 ls-2269 3d.s5 158us : _raw_spin_u 1840 ls-2269 3d.s5 158us : _raw_spin_unlock_irqrestore <-rtl8139_poll
1951 ls-2269 3d.s3 158us : net_rps_act 1841 ls-2269 3d.s3 158us : net_rps_action_and_irq_enable.isra.65 <-net_rx_action
1952 ls-2269 3d.s3 159us : __local_bh_ 1842 ls-2269 3d.s3 159us : __local_bh_enable <-__do_softirq
1953 ls-2269 3d.s3 159us : sub_preempt 1843 ls-2269 3d.s3 159us : sub_preempt_count <-__local_bh_enable
1954 ls-2269 3d..3 159us : idle_cpu <- 1844 ls-2269 3d..3 159us : idle_cpu <-irq_exit
1955 ls-2269 3d..3 159us : rcu_irq_exi 1845 ls-2269 3d..3 159us : rcu_irq_exit <-irq_exit
1956 ls-2269 3d..3 160us : sub_preempt 1846 ls-2269 3d..3 160us : sub_preempt_count <-irq_exit
1957 ls-2269 3d... 161us : __mutex_unl 1847 ls-2269 3d... 161us : __mutex_unlock_slowpath <-mutex_unlock
1958 ls-2269 3d... 162us+: trace_hardi 1848 ls-2269 3d... 162us+: trace_hardirqs_on <-mutex_unlock
1959 ls-2269 3d... 186us : <stack trac 1849 ls-2269 3d... 186us : <stack trace>
1960 => __mutex_unlock_slowpath 1850 => __mutex_unlock_slowpath
1961 => mutex_unlock 1851 => mutex_unlock
1962 => process_output 1852 => process_output
1963 => n_tty_write 1853 => n_tty_write
1964 => tty_write 1854 => tty_write
1965 => vfs_write 1855 => vfs_write
1966 => sys_write 1856 => sys_write
1967 => system_call_fastpath 1857 => system_call_fastpath
1968 1858
1969 This is an interesting trace. It started with 1859 This is an interesting trace. It started with kworker running and
1970 scheduling out and ls taking over. But as soo 1860 scheduling out and ls taking over. But as soon as ls released the
1971 rq lock and enabled interrupts (but not preem 1861 rq lock and enabled interrupts (but not preemption) an interrupt
1972 triggered. When the interrupt finished, it st 1862 triggered. When the interrupt finished, it started running softirqs.
1973 But while the softirq was running, another in 1863 But while the softirq was running, another interrupt triggered.
1974 When an interrupt is running inside a softirq 1864 When an interrupt is running inside a softirq, the annotation is 'H'.
1975 1865
1976 1866
1977 wakeup 1867 wakeup
1978 ------ 1868 ------
1979 1869
1980 One common case that people are interested in 1870 One common case that people are interested in tracing is the
1981 time it takes for a task that is woken to act 1871 time it takes for a task that is woken to actually wake up.
1982 Now for non Real-Time tasks, this can be arbi 1872 Now for non Real-Time tasks, this can be arbitrary. But tracing
1983 it nonetheless can be interesting. !! 1873 it none the less can be interesting.
1984 1874
1985 Without function tracing:: 1875 Without function tracing::
1986 1876
1987 # echo 0 > options/function-trace 1877 # echo 0 > options/function-trace
1988 # echo wakeup > current_tracer 1878 # echo wakeup > current_tracer
1989 # echo 1 > tracing_on 1879 # echo 1 > tracing_on
1990 # echo 0 > tracing_max_latency 1880 # echo 0 > tracing_max_latency
1991 # chrt -f 5 sleep 1 1881 # chrt -f 5 sleep 1
1992 # echo 0 > tracing_on 1882 # echo 0 > tracing_on
1993 # cat trace 1883 # cat trace
1994 # tracer: wakeup 1884 # tracer: wakeup
1995 # 1885 #
1996 # wakeup latency trace v1.1.5 on 3.8.0-test 1886 # wakeup latency trace v1.1.5 on 3.8.0-test+
1997 # ----------------------------------------- 1887 # --------------------------------------------------------------------
1998 # latency: 15 us, #4/4, CPU#3 | (M:preempt 1888 # latency: 15 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1999 # ----------------- 1889 # -----------------
2000 # | task: kworker/3:1H-312 (uid:0 nice:- 1890 # | task: kworker/3:1H-312 (uid:0 nice:-20 policy:0 rt_prio:0)
2001 # ----------------- 1891 # -----------------
2002 # 1892 #
2003 # _------=> CPU# 1893 # _------=> CPU#
2004 # / _-----=> irqs-off 1894 # / _-----=> irqs-off
2005 # | / _----=> need-resched 1895 # | / _----=> need-resched
2006 # || / _---=> hardirq/softir 1896 # || / _---=> hardirq/softirq
2007 # ||| / _--=> preempt-depth 1897 # ||| / _--=> preempt-depth
2008 # |||| / delay 1898 # |||| / delay
2009 # cmd pid ||||| time | caller 1899 # cmd pid ||||| time | caller
2010 # \ / ||||| \ | / 1900 # \ / ||||| \ | /
2011 <idle>-0 3dNs7 0us : 0:120: 1901 <idle>-0 3dNs7 0us : 0:120:R + [003] 312:100:R kworker/3:1H
2012 <idle>-0 3dNs7 1us+: ttwu_do_act 1902 <idle>-0 3dNs7 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
2013 <idle>-0 3d..3 15us : __schedule 1903 <idle>-0 3d..3 15us : __schedule <-schedule
2014 <idle>-0 3d..3 15us : 0:120: 1904 <idle>-0 3d..3 15us : 0:120:R ==> [003] 312:100:R kworker/3:1H
2015 1905
2016 The tracer only traces the highest priority t 1906 The tracer only traces the highest priority task in the system
2017 to avoid tracing the normal circumstances. He 1907 to avoid tracing the normal circumstances. Here we see that
2018 the kworker with a nice priority of -20 (not 1908 the kworker with a nice priority of -20 (not very nice), took
2019 just 15 microseconds from the time it woke up 1909 just 15 microseconds from the time it woke up, to the time it
2020 ran. 1910 ran.
2021 1911
2022 Non Real-Time tasks are not that interesting. 1912 Non Real-Time tasks are not that interesting. A more interesting
2023 trace is to concentrate only on Real-Time tas 1913 trace is to concentrate only on Real-Time tasks.
2024 1914
2025 wakeup_rt 1915 wakeup_rt
2026 --------- 1916 ---------
2027 1917
2028 In a Real-Time environment it is very importa 1918 In a Real-Time environment it is very important to know the
2029 wakeup time it takes for the highest priority 1919 wakeup time it takes for the highest priority task that is woken
2030 up to the time that it executes. This is also 1920 up to the time that it executes. This is also known as "schedule
2031 latency". I stress the point that this is abo 1921 latency". I stress the point that this is about RT tasks. It is
2032 also important to know the scheduling latency 1922 also important to know the scheduling latency of non-RT tasks,
2033 but the average schedule latency is better fo 1923 but the average schedule latency is better for non-RT tasks.
2034 Tools like LatencyTop are more appropriate fo 1924 Tools like LatencyTop are more appropriate for such
2035 measurements. 1925 measurements.
2036 1926
2037 Real-Time environments are interested in the 1927 Real-Time environments are interested in the worst case latency.
2038 That is the longest latency it takes for some 1928 That is the longest latency it takes for something to happen,
2039 and not the average. We can have a very fast 1929 and not the average. We can have a very fast scheduler that may
2040 only have a large latency once in a while, bu 1930 only have a large latency once in a while, but that would not
2041 work well with Real-Time tasks. The wakeup_r 1931 work well with Real-Time tasks. The wakeup_rt tracer was designed
2042 to record the worst case wakeups of RT tasks. 1932 to record the worst case wakeups of RT tasks. Non-RT tasks are
2043 not recorded because the tracer only records 1933 not recorded because the tracer only records one worst case and
2044 tracing non-RT tasks that are unpredictable w 1934 tracing non-RT tasks that are unpredictable will overwrite the
2045 worst case latency of RT tasks (just run the 1935 worst case latency of RT tasks (just run the normal wakeup
2046 tracer for a while to see that effect). 1936 tracer for a while to see that effect).
2047 1937
2048 Since this tracer only deals with RT tasks, w 1938 Since this tracer only deals with RT tasks, we will run this
2049 slightly differently than we did with the pre 1939 slightly differently than we did with the previous tracers.
2050 Instead of performing an 'ls', we will run 's 1940 Instead of performing an 'ls', we will run 'sleep 1' under
2051 'chrt' which changes the priority of the task 1941 'chrt' which changes the priority of the task.
2052 :: 1942 ::
2053 1943
2054 # echo 0 > options/function-trace 1944 # echo 0 > options/function-trace
2055 # echo wakeup_rt > current_tracer 1945 # echo wakeup_rt > current_tracer
2056 # echo 1 > tracing_on 1946 # echo 1 > tracing_on
2057 # echo 0 > tracing_max_latency 1947 # echo 0 > tracing_max_latency
2058 # chrt -f 5 sleep 1 1948 # chrt -f 5 sleep 1
2059 # echo 0 > tracing_on 1949 # echo 0 > tracing_on
2060 # cat trace 1950 # cat trace
2061 # tracer: wakeup 1951 # tracer: wakeup
2062 # 1952 #
2063 # tracer: wakeup_rt 1953 # tracer: wakeup_rt
2064 # 1954 #
2065 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 1955 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2066 # ----------------------------------------- 1956 # --------------------------------------------------------------------
2067 # latency: 5 us, #4/4, CPU#3 | (M:preempt V 1957 # latency: 5 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2068 # ----------------- 1958 # -----------------
2069 # | task: sleep-2389 (uid:0 nice:0 polic 1959 # | task: sleep-2389 (uid:0 nice:0 policy:1 rt_prio:5)
2070 # ----------------- 1960 # -----------------
2071 # 1961 #
2072 # _------=> CPU# 1962 # _------=> CPU#
2073 # / _-----=> irqs-off 1963 # / _-----=> irqs-off
2074 # | / _----=> need-resched 1964 # | / _----=> need-resched
2075 # || / _---=> hardirq/softir 1965 # || / _---=> hardirq/softirq
2076 # ||| / _--=> preempt-depth 1966 # ||| / _--=> preempt-depth
2077 # |||| / delay 1967 # |||| / delay
2078 # cmd pid ||||| time | caller 1968 # cmd pid ||||| time | caller
2079 # \ / ||||| \ | / 1969 # \ / ||||| \ | /
2080 <idle>-0 3d.h4 0us : 0:120: 1970 <idle>-0 3d.h4 0us : 0:120:R + [003] 2389: 94:R sleep
2081 <idle>-0 3d.h4 1us+: ttwu_do_act 1971 <idle>-0 3d.h4 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
2082 <idle>-0 3d..3 5us : __schedule 1972 <idle>-0 3d..3 5us : __schedule <-schedule
2083 <idle>-0 3d..3 5us : 0:120: 1973 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
2084 1974
2085 1975
2086 Running this on an idle system, we see that i 1976 Running this on an idle system, we see that it only took 5 microseconds
2087 to perform the task switch. Note, since the 1977 to perform the task switch. Note, since the trace point in the schedule
2088 is before the actual "switch", we stop the tr 1978 is before the actual "switch", we stop the tracing when the recorded task
2089 is about to schedule in. This may change if w 1979 is about to schedule in. This may change if we add a new marker at the
2090 end of the scheduler. 1980 end of the scheduler.
2091 1981
2092 Notice that the recorded task is 'sleep' with 1982 Notice that the recorded task is 'sleep' with the PID of 2389
2093 and it has an rt_prio of 5. This priority is 1983 and it has an rt_prio of 5. This priority is user-space priority
2094 and not the internal kernel priority. The pol 1984 and not the internal kernel priority. The policy is 1 for
2095 SCHED_FIFO and 2 for SCHED_RR. 1985 SCHED_FIFO and 2 for SCHED_RR.
2096 1986
2097 Note, that the trace data shows the internal 1987 Note, that the trace data shows the internal priority (99 - rtprio).
2098 :: 1988 ::
2099 1989
2100 <idle>-0 3d..3 5us : 0:120:R 1990 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
2101 1991
2102 The 0:120:R means idle was running with a nic 1992 The 0:120:R means idle was running with a nice priority of 0 (120 - 120)
2103 and in the running state 'R'. The sleep task 1993 and in the running state 'R'. The sleep task was scheduled in with
2104 2389: 94:R. That is the priority is the kerne 1994 2389: 94:R. That is the priority is the kernel rtprio (99 - 5 = 94)
2105 and it too is in the running state. 1995 and it too is in the running state.
2106 1996
2107 Doing the same with chrt -r 5 and function-tr 1997 Doing the same with chrt -r 5 and function-trace set.
2108 :: 1998 ::
2109 1999
2110 echo 1 > options/function-trace 2000 echo 1 > options/function-trace
2111 2001
2112 # tracer: wakeup_rt 2002 # tracer: wakeup_rt
2113 # 2003 #
2114 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 2004 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2115 # ----------------------------------------- 2005 # --------------------------------------------------------------------
2116 # latency: 29 us, #85/85, CPU#3 | (M:preemp 2006 # latency: 29 us, #85/85, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2117 # ----------------- 2007 # -----------------
2118 # | task: sleep-2448 (uid:0 nice:0 polic 2008 # | task: sleep-2448 (uid:0 nice:0 policy:1 rt_prio:5)
2119 # ----------------- 2009 # -----------------
2120 # 2010 #
2121 # _------=> CPU# 2011 # _------=> CPU#
2122 # / _-----=> irqs-off 2012 # / _-----=> irqs-off
2123 # | / _----=> need-resched 2013 # | / _----=> need-resched
2124 # || / _---=> hardirq/softir 2014 # || / _---=> hardirq/softirq
2125 # ||| / _--=> preempt-depth 2015 # ||| / _--=> preempt-depth
2126 # |||| / delay 2016 # |||| / delay
2127 # cmd pid ||||| time | caller 2017 # cmd pid ||||| time | caller
2128 # \ / ||||| \ | / 2018 # \ / ||||| \ | /
2129 <idle>-0 3d.h4 1us+: 0:120: 2019 <idle>-0 3d.h4 1us+: 0:120:R + [003] 2448: 94:R sleep
2130 <idle>-0 3d.h4 2us : ttwu_do_act 2020 <idle>-0 3d.h4 2us : ttwu_do_activate.constprop.87 <-try_to_wake_up
2131 <idle>-0 3d.h3 3us : check_preem 2021 <idle>-0 3d.h3 3us : check_preempt_curr <-ttwu_do_wakeup
2132 <idle>-0 3d.h3 3us : resched_cur 2022 <idle>-0 3d.h3 3us : resched_curr <-check_preempt_curr
2133 <idle>-0 3dNh3 4us : task_woken_ 2023 <idle>-0 3dNh3 4us : task_woken_rt <-ttwu_do_wakeup
2134 <idle>-0 3dNh3 4us : _raw_spin_u 2024 <idle>-0 3dNh3 4us : _raw_spin_unlock <-try_to_wake_up
2135 <idle>-0 3dNh3 4us : sub_preempt 2025 <idle>-0 3dNh3 4us : sub_preempt_count <-_raw_spin_unlock
2136 <idle>-0 3dNh2 5us : ttwu_stat < 2026 <idle>-0 3dNh2 5us : ttwu_stat <-try_to_wake_up
2137 <idle>-0 3dNh2 5us : _raw_spin_u 2027 <idle>-0 3dNh2 5us : _raw_spin_unlock_irqrestore <-try_to_wake_up
2138 <idle>-0 3dNh2 6us : sub_preempt 2028 <idle>-0 3dNh2 6us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2139 <idle>-0 3dNh1 6us : _raw_spin_l 2029 <idle>-0 3dNh1 6us : _raw_spin_lock <-__run_hrtimer
2140 <idle>-0 3dNh1 6us : add_preempt 2030 <idle>-0 3dNh1 6us : add_preempt_count <-_raw_spin_lock
2141 <idle>-0 3dNh2 7us : _raw_spin_u 2031 <idle>-0 3dNh2 7us : _raw_spin_unlock <-hrtimer_interrupt
2142 <idle>-0 3dNh2 7us : sub_preempt 2032 <idle>-0 3dNh2 7us : sub_preempt_count <-_raw_spin_unlock
2143 <idle>-0 3dNh1 7us : tick_progra 2033 <idle>-0 3dNh1 7us : tick_program_event <-hrtimer_interrupt
2144 <idle>-0 3dNh1 7us : clockevents 2034 <idle>-0 3dNh1 7us : clockevents_program_event <-tick_program_event
2145 <idle>-0 3dNh1 8us : ktime_get < 2035 <idle>-0 3dNh1 8us : ktime_get <-clockevents_program_event
2146 <idle>-0 3dNh1 8us : lapic_next_ 2036 <idle>-0 3dNh1 8us : lapic_next_event <-clockevents_program_event
2147 <idle>-0 3dNh1 8us : irq_exit <- 2037 <idle>-0 3dNh1 8us : irq_exit <-smp_apic_timer_interrupt
2148 <idle>-0 3dNh1 9us : sub_preempt 2038 <idle>-0 3dNh1 9us : sub_preempt_count <-irq_exit
2149 <idle>-0 3dN.2 9us : idle_cpu <- 2039 <idle>-0 3dN.2 9us : idle_cpu <-irq_exit
2150 <idle>-0 3dN.2 9us : rcu_irq_exi 2040 <idle>-0 3dN.2 9us : rcu_irq_exit <-irq_exit
2151 <idle>-0 3dN.2 10us : rcu_eqs_ent 2041 <idle>-0 3dN.2 10us : rcu_eqs_enter_common.isra.45 <-rcu_irq_exit
2152 <idle>-0 3dN.2 10us : sub_preempt 2042 <idle>-0 3dN.2 10us : sub_preempt_count <-irq_exit
2153 <idle>-0 3.N.1 11us : rcu_idle_ex 2043 <idle>-0 3.N.1 11us : rcu_idle_exit <-cpu_idle
2154 <idle>-0 3dN.1 11us : rcu_eqs_exi 2044 <idle>-0 3dN.1 11us : rcu_eqs_exit_common.isra.43 <-rcu_idle_exit
2155 <idle>-0 3.N.1 11us : tick_nohz_i 2045 <idle>-0 3.N.1 11us : tick_nohz_idle_exit <-cpu_idle
2156 <idle>-0 3dN.1 12us : menu_hrtime 2046 <idle>-0 3dN.1 12us : menu_hrtimer_cancel <-tick_nohz_idle_exit
2157 <idle>-0 3dN.1 12us : ktime_get < 2047 <idle>-0 3dN.1 12us : ktime_get <-tick_nohz_idle_exit
2158 <idle>-0 3dN.1 12us : tick_do_upd 2048 <idle>-0 3dN.1 12us : tick_do_update_jiffies64 <-tick_nohz_idle_exit
2159 <idle>-0 3dN.1 13us : cpu_load_up 2049 <idle>-0 3dN.1 13us : cpu_load_update_nohz <-tick_nohz_idle_exit
2160 <idle>-0 3dN.1 13us : _raw_spin_l 2050 <idle>-0 3dN.1 13us : _raw_spin_lock <-cpu_load_update_nohz
2161 <idle>-0 3dN.1 13us : add_preempt 2051 <idle>-0 3dN.1 13us : add_preempt_count <-_raw_spin_lock
2162 <idle>-0 3dN.2 13us : __cpu_load_ 2052 <idle>-0 3dN.2 13us : __cpu_load_update <-cpu_load_update_nohz
2163 <idle>-0 3dN.2 14us : sched_avg_u 2053 <idle>-0 3dN.2 14us : sched_avg_update <-__cpu_load_update
2164 <idle>-0 3dN.2 14us : _raw_spin_u 2054 <idle>-0 3dN.2 14us : _raw_spin_unlock <-cpu_load_update_nohz
2165 <idle>-0 3dN.2 14us : sub_preempt 2055 <idle>-0 3dN.2 14us : sub_preempt_count <-_raw_spin_unlock
2166 <idle>-0 3dN.1 15us : calc_load_n 2056 <idle>-0 3dN.1 15us : calc_load_nohz_stop <-tick_nohz_idle_exit
2167 <idle>-0 3dN.1 15us : touch_softl 2057 <idle>-0 3dN.1 15us : touch_softlockup_watchdog <-tick_nohz_idle_exit
2168 <idle>-0 3dN.1 15us : hrtimer_can 2058 <idle>-0 3dN.1 15us : hrtimer_cancel <-tick_nohz_idle_exit
2169 <idle>-0 3dN.1 15us : hrtimer_try 2059 <idle>-0 3dN.1 15us : hrtimer_try_to_cancel <-hrtimer_cancel
2170 <idle>-0 3dN.1 16us : lock_hrtime 2060 <idle>-0 3dN.1 16us : lock_hrtimer_base.isra.18 <-hrtimer_try_to_cancel
2171 <idle>-0 3dN.1 16us : _raw_spin_l 2061 <idle>-0 3dN.1 16us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
2172 <idle>-0 3dN.1 16us : add_preempt 2062 <idle>-0 3dN.1 16us : add_preempt_count <-_raw_spin_lock_irqsave
2173 <idle>-0 3dN.2 17us : __remove_hr 2063 <idle>-0 3dN.2 17us : __remove_hrtimer <-remove_hrtimer.part.16
2174 <idle>-0 3dN.2 17us : hrtimer_for 2064 <idle>-0 3dN.2 17us : hrtimer_force_reprogram <-__remove_hrtimer
2175 <idle>-0 3dN.2 17us : tick_progra 2065 <idle>-0 3dN.2 17us : tick_program_event <-hrtimer_force_reprogram
2176 <idle>-0 3dN.2 18us : clockevents 2066 <idle>-0 3dN.2 18us : clockevents_program_event <-tick_program_event
2177 <idle>-0 3dN.2 18us : ktime_get < 2067 <idle>-0 3dN.2 18us : ktime_get <-clockevents_program_event
2178 <idle>-0 3dN.2 18us : lapic_next_ 2068 <idle>-0 3dN.2 18us : lapic_next_event <-clockevents_program_event
2179 <idle>-0 3dN.2 19us : _raw_spin_u 2069 <idle>-0 3dN.2 19us : _raw_spin_unlock_irqrestore <-hrtimer_try_to_cancel
2180 <idle>-0 3dN.2 19us : sub_preempt 2070 <idle>-0 3dN.2 19us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2181 <idle>-0 3dN.1 19us : hrtimer_for 2071 <idle>-0 3dN.1 19us : hrtimer_forward <-tick_nohz_idle_exit
2182 <idle>-0 3dN.1 20us : ktime_add_s 2072 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
2183 <idle>-0 3dN.1 20us : ktime_add_s 2073 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
2184 <idle>-0 3dN.1 20us : hrtimer_sta 2074 <idle>-0 3dN.1 20us : hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
2185 <idle>-0 3dN.1 20us : __hrtimer_s 2075 <idle>-0 3dN.1 20us : __hrtimer_start_range_ns <-hrtimer_start_range_ns
2186 <idle>-0 3dN.1 21us : lock_hrtime 2076 <idle>-0 3dN.1 21us : lock_hrtimer_base.isra.18 <-__hrtimer_start_range_ns
2187 <idle>-0 3dN.1 21us : _raw_spin_l 2077 <idle>-0 3dN.1 21us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
2188 <idle>-0 3dN.1 21us : add_preempt 2078 <idle>-0 3dN.1 21us : add_preempt_count <-_raw_spin_lock_irqsave
2189 <idle>-0 3dN.2 22us : ktime_add_s 2079 <idle>-0 3dN.2 22us : ktime_add_safe <-__hrtimer_start_range_ns
2190 <idle>-0 3dN.2 22us : enqueue_hrt 2080 <idle>-0 3dN.2 22us : enqueue_hrtimer <-__hrtimer_start_range_ns
2191 <idle>-0 3dN.2 22us : tick_progra 2081 <idle>-0 3dN.2 22us : tick_program_event <-__hrtimer_start_range_ns
2192 <idle>-0 3dN.2 23us : clockevents 2082 <idle>-0 3dN.2 23us : clockevents_program_event <-tick_program_event
2193 <idle>-0 3dN.2 23us : ktime_get < 2083 <idle>-0 3dN.2 23us : ktime_get <-clockevents_program_event
2194 <idle>-0 3dN.2 23us : lapic_next_ 2084 <idle>-0 3dN.2 23us : lapic_next_event <-clockevents_program_event
2195 <idle>-0 3dN.2 24us : _raw_spin_u 2085 <idle>-0 3dN.2 24us : _raw_spin_unlock_irqrestore <-__hrtimer_start_range_ns
2196 <idle>-0 3dN.2 24us : sub_preempt 2086 <idle>-0 3dN.2 24us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2197 <idle>-0 3dN.1 24us : account_idl 2087 <idle>-0 3dN.1 24us : account_idle_ticks <-tick_nohz_idle_exit
2198 <idle>-0 3dN.1 24us : account_idl 2088 <idle>-0 3dN.1 24us : account_idle_time <-account_idle_ticks
2199 <idle>-0 3.N.1 25us : sub_preempt 2089 <idle>-0 3.N.1 25us : sub_preempt_count <-cpu_idle
2200 <idle>-0 3.N.. 25us : schedule <- 2090 <idle>-0 3.N.. 25us : schedule <-cpu_idle
2201 <idle>-0 3.N.. 25us : __schedule 2091 <idle>-0 3.N.. 25us : __schedule <-preempt_schedule
2202 <idle>-0 3.N.. 26us : add_preempt 2092 <idle>-0 3.N.. 26us : add_preempt_count <-__schedule
2203 <idle>-0 3.N.1 26us : rcu_note_co 2093 <idle>-0 3.N.1 26us : rcu_note_context_switch <-__schedule
2204 <idle>-0 3.N.1 26us : rcu_sched_q 2094 <idle>-0 3.N.1 26us : rcu_sched_qs <-rcu_note_context_switch
2205 <idle>-0 3dN.1 27us : rcu_preempt 2095 <idle>-0 3dN.1 27us : rcu_preempt_qs <-rcu_note_context_switch
2206 <idle>-0 3.N.1 27us : _raw_spin_l 2096 <idle>-0 3.N.1 27us : _raw_spin_lock_irq <-__schedule
2207 <idle>-0 3dN.1 27us : add_preempt 2097 <idle>-0 3dN.1 27us : add_preempt_count <-_raw_spin_lock_irq
2208 <idle>-0 3dN.2 28us : put_prev_ta 2098 <idle>-0 3dN.2 28us : put_prev_task_idle <-__schedule
2209 <idle>-0 3dN.2 28us : pick_next_t 2099 <idle>-0 3dN.2 28us : pick_next_task_stop <-pick_next_task
2210 <idle>-0 3dN.2 28us : pick_next_t 2100 <idle>-0 3dN.2 28us : pick_next_task_rt <-pick_next_task
2211 <idle>-0 3dN.2 29us : dequeue_pus 2101 <idle>-0 3dN.2 29us : dequeue_pushable_task <-pick_next_task_rt
2212 <idle>-0 3d..3 29us : __schedule 2102 <idle>-0 3d..3 29us : __schedule <-preempt_schedule
2213 <idle>-0 3d..3 30us : 0:120: 2103 <idle>-0 3d..3 30us : 0:120:R ==> [003] 2448: 94:R sleep
2214 2104
2215 This isn't that big of a trace, even with fun 2105 This isn't that big of a trace, even with function tracing enabled,
2216 so I included the entire trace. 2106 so I included the entire trace.
2217 2107
2218 The interrupt went off while when the system 2108 The interrupt went off while when the system was idle. Somewhere
2219 before task_woken_rt() was called, the NEED_R 2109 before task_woken_rt() was called, the NEED_RESCHED flag was set,
2220 this is indicated by the first occurrence of 2110 this is indicated by the first occurrence of the 'N' flag.
2221 2111
2222 Latency tracing and events 2112 Latency tracing and events
2223 -------------------------- 2113 --------------------------
2224 As function tracing can induce a much larger 2114 As function tracing can induce a much larger latency, but without
2225 seeing what happens within the latency it is 2115 seeing what happens within the latency it is hard to know what
2226 caused it. There is a middle ground, and that 2116 caused it. There is a middle ground, and that is with enabling
2227 events. 2117 events.
2228 :: 2118 ::
2229 2119
2230 # echo 0 > options/function-trace 2120 # echo 0 > options/function-trace
2231 # echo wakeup_rt > current_tracer 2121 # echo wakeup_rt > current_tracer
2232 # echo 1 > events/enable 2122 # echo 1 > events/enable
2233 # echo 1 > tracing_on 2123 # echo 1 > tracing_on
2234 # echo 0 > tracing_max_latency 2124 # echo 0 > tracing_max_latency
2235 # chrt -f 5 sleep 1 2125 # chrt -f 5 sleep 1
2236 # echo 0 > tracing_on 2126 # echo 0 > tracing_on
2237 # cat trace 2127 # cat trace
2238 # tracer: wakeup_rt 2128 # tracer: wakeup_rt
2239 # 2129 #
2240 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 2130 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2241 # ----------------------------------------- 2131 # --------------------------------------------------------------------
2242 # latency: 6 us, #12/12, CPU#2 | (M:preempt 2132 # latency: 6 us, #12/12, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2243 # ----------------- 2133 # -----------------
2244 # | task: sleep-5882 (uid:0 nice:0 polic 2134 # | task: sleep-5882 (uid:0 nice:0 policy:1 rt_prio:5)
2245 # ----------------- 2135 # -----------------
2246 # 2136 #
2247 # _------=> CPU# 2137 # _------=> CPU#
2248 # / _-----=> irqs-off 2138 # / _-----=> irqs-off
2249 # | / _----=> need-resched 2139 # | / _----=> need-resched
2250 # || / _---=> hardirq/softir 2140 # || / _---=> hardirq/softirq
2251 # ||| / _--=> preempt-depth 2141 # ||| / _--=> preempt-depth
2252 # |||| / delay 2142 # |||| / delay
2253 # cmd pid ||||| time | caller 2143 # cmd pid ||||| time | caller
2254 # \ / ||||| \ | / 2144 # \ / ||||| \ | /
2255 <idle>-0 2d.h4 0us : 0:120: 2145 <idle>-0 2d.h4 0us : 0:120:R + [002] 5882: 94:R sleep
2256 <idle>-0 2d.h4 0us : ttwu_do_act 2146 <idle>-0 2d.h4 0us : ttwu_do_activate.constprop.87 <-try_to_wake_up
2257 <idle>-0 2d.h4 1us : sched_wakeu 2147 <idle>-0 2d.h4 1us : sched_wakeup: comm=sleep pid=5882 prio=94 success=1 target_cpu=002
2258 <idle>-0 2dNh2 1us : hrtimer_exp 2148 <idle>-0 2dNh2 1us : hrtimer_expire_exit: hrtimer=ffff88007796feb8
2259 <idle>-0 2.N.2 2us : power_end: 2149 <idle>-0 2.N.2 2us : power_end: cpu_id=2
2260 <idle>-0 2.N.2 3us : cpu_idle: s 2150 <idle>-0 2.N.2 3us : cpu_idle: state=4294967295 cpu_id=2
2261 <idle>-0 2dN.3 4us : hrtimer_can 2151 <idle>-0 2dN.3 4us : hrtimer_cancel: hrtimer=ffff88007d50d5e0
2262 <idle>-0 2dN.3 4us : hrtimer_sta 2152 <idle>-0 2dN.3 4us : hrtimer_start: hrtimer=ffff88007d50d5e0 function=tick_sched_timer expires=34311211000000 softexpires=34311211000000
2263 <idle>-0 2.N.2 5us : rcu_utiliza 2153 <idle>-0 2.N.2 5us : rcu_utilization: Start context switch
2264 <idle>-0 2.N.2 5us : rcu_utiliza 2154 <idle>-0 2.N.2 5us : rcu_utilization: End context switch
2265 <idle>-0 2d..3 6us : __schedule 2155 <idle>-0 2d..3 6us : __schedule <-schedule
2266 <idle>-0 2d..3 6us : 0:120: 2156 <idle>-0 2d..3 6us : 0:120:R ==> [002] 5882: 94:R sleep
2267 2157
2268 2158
2269 Hardware Latency Detector 2159 Hardware Latency Detector
2270 ------------------------- 2160 -------------------------
2271 2161
2272 The hardware latency detector is executed by 2162 The hardware latency detector is executed by enabling the "hwlat" tracer.
2273 2163
2274 NOTE, this tracer will affect the performance 2164 NOTE, this tracer will affect the performance of the system as it will
2275 periodically make a CPU constantly busy with 2165 periodically make a CPU constantly busy with interrupts disabled.
2276 :: 2166 ::
2277 2167
2278 # echo hwlat > current_tracer 2168 # echo hwlat > current_tracer
2279 # sleep 100 2169 # sleep 100
2280 # cat trace 2170 # cat trace
2281 # tracer: hwlat 2171 # tracer: hwlat
2282 # 2172 #
2283 # entries-in-buffer/entries-written: 13/13 2173 # entries-in-buffer/entries-written: 13/13 #P:8
2284 # 2174 #
2285 # _-----=> irq 2175 # _-----=> irqs-off
2286 # / _----=> nee 2176 # / _----=> need-resched
2287 # | / _---=> har 2177 # | / _---=> hardirq/softirq
2288 # || / _--=> pre 2178 # || / _--=> preempt-depth
2289 # ||| / dela 2179 # ||| / delay
2290 # TASK-PID CPU# |||| TIMEST 2180 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
2291 # | | | |||| | 2181 # | | | |||| | |
2292 <...>-1729 [001] d... 678.473 2182 <...>-1729 [001] d... 678.473449: #1 inner/outer(us): 11/12 ts:1581527483.343962693 count:6
2293 <...>-1729 [004] d... 689.556 2183 <...>-1729 [004] d... 689.556542: #2 inner/outer(us): 16/9 ts:1581527494.889008092 count:1
2294 <...>-1729 [005] d... 714.756 2184 <...>-1729 [005] d... 714.756290: #3 inner/outer(us): 16/16 ts:1581527519.678961629 count:5
2295 <...>-1729 [001] d... 718.788 2185 <...>-1729 [001] d... 718.788247: #4 inner/outer(us): 9/17 ts:1581527523.889012713 count:1
2296 <...>-1729 [002] d... 719.796 2186 <...>-1729 [002] d... 719.796341: #5 inner/outer(us): 13/9 ts:1581527524.912872606 count:1
2297 <...>-1729 [006] d... 844.787 2187 <...>-1729 [006] d... 844.787091: #6 inner/outer(us): 9/12 ts:1581527649.889048502 count:2
2298 <...>-1729 [003] d... 849.827 2188 <...>-1729 [003] d... 849.827033: #7 inner/outer(us): 18/9 ts:1581527654.889013793 count:1
2299 <...>-1729 [007] d... 853.859 2189 <...>-1729 [007] d... 853.859002: #8 inner/outer(us): 9/12 ts:1581527658.889065736 count:1
2300 <...>-1729 [001] d... 855.874 2190 <...>-1729 [001] d... 855.874978: #9 inner/outer(us): 9/11 ts:1581527660.861991877 count:1
2301 <...>-1729 [001] d... 863.938 2191 <...>-1729 [001] d... 863.938932: #10 inner/outer(us): 9/11 ts:1581527668.970010500 count:1 nmi-total:7 nmi-count:1
2302 <...>-1729 [007] d... 878.050 2192 <...>-1729 [007] d... 878.050780: #11 inner/outer(us): 9/12 ts:1581527683.385002600 count:1 nmi-total:5 nmi-count:1
2303 <...>-1729 [007] d... 886.114 2193 <...>-1729 [007] d... 886.114702: #12 inner/outer(us): 9/12 ts:1581527691.385001600 count:1
2304 2194
2305 2195
2306 The above output is somewhat the same in the 2196 The above output is somewhat the same in the header. All events will have
2307 interrupts disabled 'd'. Under the FUNCTION t 2197 interrupts disabled 'd'. Under the FUNCTION title there is:
2308 2198
2309 #1 2199 #1
2310 This is the count of events recorded 2200 This is the count of events recorded that were greater than the
2311 tracing_threshold (See below). 2201 tracing_threshold (See below).
2312 2202
2313 inner/outer(us): 11/11 2203 inner/outer(us): 11/11
2314 2204
2315 This shows two numbers as "inner latenc 2205 This shows two numbers as "inner latency" and "outer latency". The test
2316 runs in a loop checking a timestamp twi 2206 runs in a loop checking a timestamp twice. The latency detected within
2317 the two timestamps is the "inner latenc 2207 the two timestamps is the "inner latency" and the latency detected
2318 after the previous timestamp and the ne 2208 after the previous timestamp and the next timestamp in the loop is
2319 the "outer latency". 2209 the "outer latency".
2320 2210
2321 ts:1581527483.343962693 2211 ts:1581527483.343962693
2322 2212
2323 The absolute timestamp that the first l 2213 The absolute timestamp that the first latency was recorded in the window.
2324 2214
2325 count:6 2215 count:6
2326 2216
2327 The number of times a latency was detec 2217 The number of times a latency was detected during the window.
2328 2218
2329 nmi-total:7 nmi-count:1 2219 nmi-total:7 nmi-count:1
2330 2220
2331 On architectures that support it, if an 2221 On architectures that support it, if an NMI comes in during the
2332 test, the time spent in NMI is reported 2222 test, the time spent in NMI is reported in "nmi-total" (in
2333 microseconds). 2223 microseconds).
2334 2224
2335 All architectures that have NMIs will s 2225 All architectures that have NMIs will show the "nmi-count" if an
2336 NMI comes in during the test. 2226 NMI comes in during the test.
2337 2227
2338 hwlat files: 2228 hwlat files:
2339 2229
2340 tracing_threshold 2230 tracing_threshold
2341 This gets automatically set to "10" t 2231 This gets automatically set to "10" to represent 10
2342 microseconds. This is the threshold o 2232 microseconds. This is the threshold of latency that
2343 needs to be detected before the trace 2233 needs to be detected before the trace will be recorded.
2344 2234
2345 Note, when hwlat tracer is finished ( 2235 Note, when hwlat tracer is finished (another tracer is
2346 written into "current_tracer"), the o 2236 written into "current_tracer"), the original value for
2347 tracing_threshold is placed back into 2237 tracing_threshold is placed back into this file.
2348 2238
2349 hwlat_detector/width 2239 hwlat_detector/width
2350 The length of time the test runs with 2240 The length of time the test runs with interrupts disabled.
2351 2241
2352 hwlat_detector/window 2242 hwlat_detector/window
2353 The length of time of the window whic 2243 The length of time of the window which the test
2354 runs. That is, the test will run for 2244 runs. That is, the test will run for "width"
2355 microseconds per "window" microsecond 2245 microseconds per "window" microseconds
2356 2246
2357 tracing_cpumask 2247 tracing_cpumask
2358 When the test is started. A kernel th 2248 When the test is started. A kernel thread is created that
2359 runs the test. This thread will alter 2249 runs the test. This thread will alternate between CPUs
2360 listed in the tracing_cpumask between 2250 listed in the tracing_cpumask between each period
2361 (one "window"). To limit the test to 2251 (one "window"). To limit the test to specific CPUs
2362 set the mask in this file to only the 2252 set the mask in this file to only the CPUs that the test
2363 should run on. 2253 should run on.
2364 2254
2365 function 2255 function
2366 -------- 2256 --------
2367 2257
2368 This tracer is the function tracer. Enabling 2258 This tracer is the function tracer. Enabling the function tracer
2369 can be done from the debug file system. Make 2259 can be done from the debug file system. Make sure the
2370 ftrace_enabled is set; otherwise this tracer 2260 ftrace_enabled is set; otherwise this tracer is a nop.
2371 See the "ftrace_enabled" section below. 2261 See the "ftrace_enabled" section below.
2372 :: 2262 ::
2373 2263
2374 # sysctl kernel.ftrace_enabled=1 2264 # sysctl kernel.ftrace_enabled=1
2375 # echo function > current_tracer 2265 # echo function > current_tracer
2376 # echo 1 > tracing_on 2266 # echo 1 > tracing_on
2377 # usleep 1 2267 # usleep 1
2378 # echo 0 > tracing_on 2268 # echo 0 > tracing_on
2379 # cat trace 2269 # cat trace
2380 # tracer: function 2270 # tracer: function
2381 # 2271 #
2382 # entries-in-buffer/entries-written: 24799/ 2272 # entries-in-buffer/entries-written: 24799/24799 #P:4
2383 # 2273 #
2384 # _-----=> irq 2274 # _-----=> irqs-off
2385 # / _----=> nee 2275 # / _----=> need-resched
2386 # | / _---=> har 2276 # | / _---=> hardirq/softirq
2387 # || / _--=> pre 2277 # || / _--=> preempt-depth
2388 # ||| / dela 2278 # ||| / delay
2389 # TASK-PID CPU# |||| TIMEST 2279 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
2390 # | | | |||| | 2280 # | | | |||| | |
2391 bash-1994 [002] .... 3082.063 2281 bash-1994 [002] .... 3082.063030: mutex_unlock <-rb_simple_write
2392 bash-1994 [002] .... 3082.063 2282 bash-1994 [002] .... 3082.063031: __mutex_unlock_slowpath <-mutex_unlock
2393 bash-1994 [002] .... 3082.063 2283 bash-1994 [002] .... 3082.063031: __fsnotify_parent <-fsnotify_modify
2394 bash-1994 [002] .... 3082.063 2284 bash-1994 [002] .... 3082.063032: fsnotify <-fsnotify_modify
2395 bash-1994 [002] .... 3082.063 2285 bash-1994 [002] .... 3082.063032: __srcu_read_lock <-fsnotify
2396 bash-1994 [002] .... 3082.063 2286 bash-1994 [002] .... 3082.063032: add_preempt_count <-__srcu_read_lock
2397 bash-1994 [002] ...1 3082.063 2287 bash-1994 [002] ...1 3082.063032: sub_preempt_count <-__srcu_read_lock
2398 bash-1994 [002] .... 3082.063 2288 bash-1994 [002] .... 3082.063033: __srcu_read_unlock <-fsnotify
2399 [...] 2289 [...]
2400 2290
2401 2291
2402 Note: function tracer uses ring buffers to st 2292 Note: function tracer uses ring buffers to store the above
2403 entries. The newest data may overwrite the ol 2293 entries. The newest data may overwrite the oldest data.
2404 Sometimes using echo to stop the trace is not 2294 Sometimes using echo to stop the trace is not sufficient because
2405 the tracing could have overwritten the data t 2295 the tracing could have overwritten the data that you wanted to
2406 record. For this reason, it is sometimes bett 2296 record. For this reason, it is sometimes better to disable
2407 tracing directly from a program. This allows 2297 tracing directly from a program. This allows you to stop the
2408 tracing at the point that you hit the part th 2298 tracing at the point that you hit the part that you are
2409 interested in. To disable the tracing directl 2299 interested in. To disable the tracing directly from a C program,
2410 something like following code snippet can be 2300 something like following code snippet can be used::
2411 2301
2412 int trace_fd; 2302 int trace_fd;
2413 [...] 2303 [...]
2414 int main(int argc, char *argv[]) { 2304 int main(int argc, char *argv[]) {
2415 [...] 2305 [...]
2416 trace_fd = open(tracing_file( 2306 trace_fd = open(tracing_file("tracing_on"), O_WRONLY);
2417 [...] 2307 [...]
2418 if (condition_hit()) { 2308 if (condition_hit()) {
2419 write(trace_fd, "0", 2309 write(trace_fd, "0", 1);
2420 } 2310 }
2421 [...] 2311 [...]
2422 } 2312 }
2423 2313
2424 2314
2425 Single thread tracing 2315 Single thread tracing
2426 --------------------- 2316 ---------------------
2427 2317
2428 By writing into set_ftrace_pid you can trace 2318 By writing into set_ftrace_pid you can trace a
2429 single thread. For example:: 2319 single thread. For example::
2430 2320
2431 # cat set_ftrace_pid 2321 # cat set_ftrace_pid
2432 no pid 2322 no pid
2433 # echo 3111 > set_ftrace_pid 2323 # echo 3111 > set_ftrace_pid
2434 # cat set_ftrace_pid 2324 # cat set_ftrace_pid
2435 3111 2325 3111
2436 # echo function > current_tracer 2326 # echo function > current_tracer
2437 # cat trace | head 2327 # cat trace | head
2438 # tracer: function 2328 # tracer: function
2439 # 2329 #
2440 # TASK-PID CPU# TIMESTAMP 2330 # TASK-PID CPU# TIMESTAMP FUNCTION
2441 # | | | | 2331 # | | | | |
2442 yum-updatesd-3111 [003] 1637.254676: 2332 yum-updatesd-3111 [003] 1637.254676: finish_task_switch <-thread_return
2443 yum-updatesd-3111 [003] 1637.254681: 2333 yum-updatesd-3111 [003] 1637.254681: hrtimer_cancel <-schedule_hrtimeout_range
2444 yum-updatesd-3111 [003] 1637.254682: 2334 yum-updatesd-3111 [003] 1637.254682: hrtimer_try_to_cancel <-hrtimer_cancel
2445 yum-updatesd-3111 [003] 1637.254683: 2335 yum-updatesd-3111 [003] 1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel
2446 yum-updatesd-3111 [003] 1637.254685: 2336 yum-updatesd-3111 [003] 1637.254685: fget_light <-do_sys_poll
2447 yum-updatesd-3111 [003] 1637.254686: 2337 yum-updatesd-3111 [003] 1637.254686: pipe_poll <-do_sys_poll
2448 # echo > set_ftrace_pid 2338 # echo > set_ftrace_pid
2449 # cat trace |head 2339 # cat trace |head
2450 # tracer: function 2340 # tracer: function
2451 # 2341 #
2452 # TASK-PID CPU# TIMESTAMP 2342 # TASK-PID CPU# TIMESTAMP FUNCTION
2453 # | | | | 2343 # | | | | |
2454 ##### CPU 3 buffer started #### 2344 ##### CPU 3 buffer started ####
2455 yum-updatesd-3111 [003] 1701.957688: 2345 yum-updatesd-3111 [003] 1701.957688: free_poll_entry <-poll_freewait
2456 yum-updatesd-3111 [003] 1701.957689: 2346 yum-updatesd-3111 [003] 1701.957689: remove_wait_queue <-free_poll_entry
2457 yum-updatesd-3111 [003] 1701.957691: 2347 yum-updatesd-3111 [003] 1701.957691: fput <-free_poll_entry
2458 yum-updatesd-3111 [003] 1701.957692: 2348 yum-updatesd-3111 [003] 1701.957692: audit_syscall_exit <-sysret_audit
2459 yum-updatesd-3111 [003] 1701.957693: 2349 yum-updatesd-3111 [003] 1701.957693: path_put <-audit_syscall_exit
2460 2350
2461 If you want to trace a function when executin 2351 If you want to trace a function when executing, you could use
2462 something like this simple program. 2352 something like this simple program.
2463 :: 2353 ::
2464 2354
2465 #include <stdio.h> 2355 #include <stdio.h>
2466 #include <stdlib.h> 2356 #include <stdlib.h>
2467 #include <sys/types.h> 2357 #include <sys/types.h>
2468 #include <sys/stat.h> 2358 #include <sys/stat.h>
2469 #include <fcntl.h> 2359 #include <fcntl.h>
2470 #include <unistd.h> 2360 #include <unistd.h>
2471 #include <string.h> 2361 #include <string.h>
2472 2362
2473 #define _STR(x) #x 2363 #define _STR(x) #x
2474 #define STR(x) _STR(x) 2364 #define STR(x) _STR(x)
2475 #define MAX_PATH 256 2365 #define MAX_PATH 256
2476 2366
2477 const char *find_tracefs(void) 2367 const char *find_tracefs(void)
2478 { 2368 {
2479 static char tracefs[MAX_PATH+1 2369 static char tracefs[MAX_PATH+1];
2480 static int tracefs_found; 2370 static int tracefs_found;
2481 char type[100]; 2371 char type[100];
2482 FILE *fp; 2372 FILE *fp;
2483 2373
2484 if (tracefs_found) 2374 if (tracefs_found)
2485 return tracefs; 2375 return tracefs;
2486 2376
2487 if ((fp = fopen("/proc/mounts" 2377 if ((fp = fopen("/proc/mounts","r")) == NULL) {
2488 perror("/proc/mounts") 2378 perror("/proc/mounts");
2489 return NULL; 2379 return NULL;
2490 } 2380 }
2491 2381
2492 while (fscanf(fp, "%*s %" 2382 while (fscanf(fp, "%*s %"
2493 STR(MAX_PATH) 2383 STR(MAX_PATH)
2494 "s %99s %*s %*d 2384 "s %99s %*s %*d %*d\n",
2495 tracefs, type) = 2385 tracefs, type) == 2) {
2496 if (strcmp(type, "trac 2386 if (strcmp(type, "tracefs") == 0)
2497 break; 2387 break;
2498 } 2388 }
2499 fclose(fp); 2389 fclose(fp);
2500 2390
2501 if (strcmp(type, "tracefs") != 2391 if (strcmp(type, "tracefs") != 0) {
2502 fprintf(stderr, "trace 2392 fprintf(stderr, "tracefs not mounted");
2503 return NULL; 2393 return NULL;
2504 } 2394 }
2505 2395
2506 strcat(tracefs, "/tracing/"); 2396 strcat(tracefs, "/tracing/");
2507 tracefs_found = 1; 2397 tracefs_found = 1;
2508 2398
2509 return tracefs; 2399 return tracefs;
2510 } 2400 }
2511 2401
2512 const char *tracing_file(const char * 2402 const char *tracing_file(const char *file_name)
2513 { 2403 {
2514 static char trace_file[MAX_PAT 2404 static char trace_file[MAX_PATH+1];
2515 snprintf(trace_file, MAX_PATH, 2405 snprintf(trace_file, MAX_PATH, "%s/%s", find_tracefs(), file_name);
2516 return trace_file; 2406 return trace_file;
2517 } 2407 }
2518 2408
2519 int main (int argc, char **argv) 2409 int main (int argc, char **argv)
2520 { 2410 {
2521 if (argc < 1) 2411 if (argc < 1)
2522 exit(-1); 2412 exit(-1);
2523 2413
2524 if (fork() > 0) { 2414 if (fork() > 0) {
2525 int fd, ffd; 2415 int fd, ffd;
2526 char line[64]; 2416 char line[64];
2527 int s; 2417 int s;
2528 2418
2529 ffd = open(tracing_fi 2419 ffd = open(tracing_file("current_tracer"), O_WRONLY);
2530 if (ffd < 0) 2420 if (ffd < 0)
2531 exit(-1); 2421 exit(-1);
2532 write(ffd, "nop", 3); 2422 write(ffd, "nop", 3);
2533 2423
2534 fd = open(tracing_fil 2424 fd = open(tracing_file("set_ftrace_pid"), O_WRONLY);
2535 s = sprintf(line, "%d 2425 s = sprintf(line, "%d\n", getpid());
2536 write(fd, line, s); 2426 write(fd, line, s);
2537 2427
2538 write(ffd, "function" 2428 write(ffd, "function", 8);
2539 2429
2540 close(fd); 2430 close(fd);
2541 close(ffd); 2431 close(ffd);
2542 2432
2543 execvp(argv[1], argv+ 2433 execvp(argv[1], argv+1);
2544 } 2434 }
2545 2435
2546 return 0; 2436 return 0;
2547 } 2437 }
2548 2438
2549 Or this simple script! 2439 Or this simple script!
2550 :: 2440 ::
2551 2441
2552 #!/bin/bash 2442 #!/bin/bash
2553 2443
2554 tracefs=`sed -ne 's/^tracefs \(.*\) tracefs 2444 tracefs=`sed -ne 's/^tracefs \(.*\) tracefs.*/\1/p' /proc/mounts`
2555 echo 0 > $tracefs/tracing_on !! 2445 echo nop > $tracefs/tracing/current_tracer
2556 echo $$ > $tracefs/set_ftrace_pid !! 2446 echo 0 > $tracefs/tracing/tracing_on
2557 echo function > $tracefs/current_tracer !! 2447 echo $$ > $tracefs/tracing/set_ftrace_pid
2558 echo 1 > $tracefs/tracing_on !! 2448 echo function > $tracefs/tracing/current_tracer
>> 2449 echo 1 > $tracefs/tracing/tracing_on
2559 exec "$@" 2450 exec "$@"
2560 2451
2561 2452
2562 function graph tracer 2453 function graph tracer
2563 --------------------------- 2454 ---------------------------
2564 2455
2565 This tracer is similar to the function tracer 2456 This tracer is similar to the function tracer except that it
2566 probes a function on its entry and its exit. 2457 probes a function on its entry and its exit. This is done by
2567 using a dynamically allocated stack of return 2458 using a dynamically allocated stack of return addresses in each
2568 task_struct. On function entry the tracer ove 2459 task_struct. On function entry the tracer overwrites the return
2569 address of each function traced to set a cust 2460 address of each function traced to set a custom probe. Thus the
2570 original return address is stored on the stac 2461 original return address is stored on the stack of return address
2571 in the task_struct. 2462 in the task_struct.
2572 2463
2573 Probing on both ends of a function leads to s 2464 Probing on both ends of a function leads to special features
2574 such as: 2465 such as:
2575 2466
2576 - measure of a function's time execution 2467 - measure of a function's time execution
2577 - having a reliable call stack to draw functi 2468 - having a reliable call stack to draw function calls graph
2578 2469
2579 This tracer is useful in several situations: 2470 This tracer is useful in several situations:
2580 2471
2581 - you want to find the reason of a strange ke 2472 - you want to find the reason of a strange kernel behavior and
2582 need to see what happens in detail on any a 2473 need to see what happens in detail on any areas (or specific
2583 ones). 2474 ones).
2584 2475
2585 - you are experiencing weird latencies but it 2476 - you are experiencing weird latencies but it's difficult to
2586 find its origin. 2477 find its origin.
2587 2478
2588 - you want to find quickly which path is take 2479 - you want to find quickly which path is taken by a specific
2589 function 2480 function
2590 2481
2591 - you just want to peek inside a working kern 2482 - you just want to peek inside a working kernel and want to see
2592 what happens there. 2483 what happens there.
2593 2484
2594 :: 2485 ::
2595 2486
2596 # tracer: function_graph 2487 # tracer: function_graph
2597 # 2488 #
2598 # CPU DURATION FUNCTION C 2489 # CPU DURATION FUNCTION CALLS
2599 # | | | | | | 2490 # | | | | | | |
2600 2491
2601 0) | sys_open() { 2492 0) | sys_open() {
2602 0) | do_sys_open() { 2493 0) | do_sys_open() {
2603 0) | getname() { 2494 0) | getname() {
2604 0) | kmem_cache_alloc 2495 0) | kmem_cache_alloc() {
2605 0) 1.382 us | __might_sleep( 2496 0) 1.382 us | __might_sleep();
2606 0) 2.478 us | } 2497 0) 2.478 us | }
2607 0) | strncpy_from_use 2498 0) | strncpy_from_user() {
2608 0) | might_fault() 2499 0) | might_fault() {
2609 0) 1.389 us | __might_slee 2500 0) 1.389 us | __might_sleep();
2610 0) 2.553 us | } 2501 0) 2.553 us | }
2611 0) 3.807 us | } 2502 0) 3.807 us | }
2612 0) 7.876 us | } 2503 0) 7.876 us | }
2613 0) | alloc_fd() { 2504 0) | alloc_fd() {
2614 0) 0.668 us | _spin_lock(); 2505 0) 0.668 us | _spin_lock();
2615 0) 0.570 us | expand_files(); 2506 0) 0.570 us | expand_files();
2616 0) 0.586 us | _spin_unlock(); 2507 0) 0.586 us | _spin_unlock();
2617 2508
2618 2509
2619 There are several columns that can be dynamic 2510 There are several columns that can be dynamically
2620 enabled/disabled. You can use every combinati 2511 enabled/disabled. You can use every combination of options you
2621 want, depending on your needs. 2512 want, depending on your needs.
2622 2513
2623 - The cpu number on which the function execut 2514 - The cpu number on which the function executed is default
2624 enabled. It is sometimes better to only tr 2515 enabled. It is sometimes better to only trace one cpu (see
2625 tracing_cpumask file) or you might sometime !! 2516 tracing_cpu_mask file) or you might sometimes see unordered
2626 function calls while cpu tracing switch. 2517 function calls while cpu tracing switch.
2627 2518
2628 - hide: echo nofuncgraph-cpu > trace_ 2519 - hide: echo nofuncgraph-cpu > trace_options
2629 - show: echo funcgraph-cpu > trace_op 2520 - show: echo funcgraph-cpu > trace_options
2630 2521
2631 - The duration (function's time of execution) 2522 - The duration (function's time of execution) is displayed on
2632 the closing bracket line of a function or o 2523 the closing bracket line of a function or on the same line
2633 than the current function in case of a leaf 2524 than the current function in case of a leaf one. It is default
2634 enabled. 2525 enabled.
2635 2526
2636 - hide: echo nofuncgraph-duration > t 2527 - hide: echo nofuncgraph-duration > trace_options
2637 - show: echo funcgraph-duration > tra 2528 - show: echo funcgraph-duration > trace_options
2638 2529
2639 - The overhead field precedes the duration fi 2530 - The overhead field precedes the duration field in case of
2640 reached duration thresholds. 2531 reached duration thresholds.
2641 2532
2642 - hide: echo nofuncgraph-overhead > t 2533 - hide: echo nofuncgraph-overhead > trace_options
2643 - show: echo funcgraph-overhead > tra 2534 - show: echo funcgraph-overhead > trace_options
2644 - depends on: funcgraph-duration 2535 - depends on: funcgraph-duration
2645 2536
2646 ie:: 2537 ie::
2647 2538
2648 3) # 1837.709 us | } /* __switch 2539 3) # 1837.709 us | } /* __switch_to */
2649 3) | finish_task_s 2540 3) | finish_task_switch() {
2650 3) 0.313 us | _raw_spin_u 2541 3) 0.313 us | _raw_spin_unlock_irq();
2651 3) 3.177 us | } 2542 3) 3.177 us | }
2652 3) # 1889.063 us | } /* __schedule 2543 3) # 1889.063 us | } /* __schedule */
2653 3) ! 140.417 us | } /* __schedule * 2544 3) ! 140.417 us | } /* __schedule */
2654 3) # 2034.948 us | } /* schedule */ 2545 3) # 2034.948 us | } /* schedule */
2655 3) * 33998.59 us | } /* schedule_preempt 2546 3) * 33998.59 us | } /* schedule_preempt_disabled */
2656 2547
2657 [...] 2548 [...]
2658 2549
2659 1) 0.260 us | msecs_to_ 2550 1) 0.260 us | msecs_to_jiffies();
2660 1) 0.313 us | __rcu_rea 2551 1) 0.313 us | __rcu_read_unlock();
2661 1) + 61.770 us | } 2552 1) + 61.770 us | }
2662 1) + 64.479 us | } 2553 1) + 64.479 us | }
2663 1) 0.313 us | rcu_bh_qs(); 2554 1) 0.313 us | rcu_bh_qs();
2664 1) 0.313 us | __local_bh_en 2555 1) 0.313 us | __local_bh_enable();
2665 1) ! 217.240 us | } 2556 1) ! 217.240 us | }
2666 1) 0.365 us | idle_cpu(); 2557 1) 0.365 us | idle_cpu();
2667 1) | rcu_irq_exit() 2558 1) | rcu_irq_exit() {
2668 1) 0.417 us | rcu_eqs_enter 2559 1) 0.417 us | rcu_eqs_enter_common.isra.47();
2669 1) 3.125 us | } 2560 1) 3.125 us | }
2670 1) ! 227.812 us | } 2561 1) ! 227.812 us | }
2671 1) ! 457.395 us | } 2562 1) ! 457.395 us | }
2672 1) @ 119760.2 us | } 2563 1) @ 119760.2 us | }
2673 2564
2674 [...] 2565 [...]
2675 2566
2676 2) | handle_IPI() { 2567 2) | handle_IPI() {
2677 1) 6.979 us | } 2568 1) 6.979 us | }
2678 2) 0.417 us | scheduler_ipi(); 2569 2) 0.417 us | scheduler_ipi();
2679 1) 9.791 us | } 2570 1) 9.791 us | }
2680 1) + 12.917 us | } 2571 1) + 12.917 us | }
2681 2) 3.490 us | } 2572 2) 3.490 us | }
2682 1) + 15.729 us | } 2573 1) + 15.729 us | }
2683 1) + 18.542 us | } 2574 1) + 18.542 us | }
2684 2) $ 3594274 us | } 2575 2) $ 3594274 us | }
2685 2576
2686 Flags:: 2577 Flags::
2687 2578
2688 + means that the function exceeded 10 usecs 2579 + means that the function exceeded 10 usecs.
2689 ! means that the function exceeded 100 usec 2580 ! means that the function exceeded 100 usecs.
2690 # means that the function exceeded 1000 use 2581 # means that the function exceeded 1000 usecs.
2691 * means that the function exceeded 10 msecs 2582 * means that the function exceeded 10 msecs.
2692 @ means that the function exceeded 100 msec 2583 @ means that the function exceeded 100 msecs.
2693 $ means that the function exceeded 1 sec. 2584 $ means that the function exceeded 1 sec.
2694 2585
2695 2586
2696 - The task/pid field displays the thread cmdl 2587 - The task/pid field displays the thread cmdline and pid which
2697 executed the function. It is default disabl 2588 executed the function. It is default disabled.
2698 2589
2699 - hide: echo nofuncgraph-proc > trace 2590 - hide: echo nofuncgraph-proc > trace_options
2700 - show: echo funcgraph-proc > trace_o 2591 - show: echo funcgraph-proc > trace_options
2701 2592
2702 ie:: 2593 ie::
2703 2594
2704 # tracer: function_graph 2595 # tracer: function_graph
2705 # 2596 #
2706 # CPU TASK/PID DURATION 2597 # CPU TASK/PID DURATION FUNCTION CALLS
2707 # | | | | | 2598 # | | | | | | | | |
2708 0) sh-4802 | | 2599 0) sh-4802 | | d_free() {
2709 0) sh-4802 | | 2600 0) sh-4802 | | call_rcu() {
2710 0) sh-4802 | | 2601 0) sh-4802 | | __call_rcu() {
2711 0) sh-4802 | 0.616 us | 2602 0) sh-4802 | 0.616 us | rcu_process_gp_end();
2712 0) sh-4802 | 0.586 us | 2603 0) sh-4802 | 0.586 us | check_for_new_grace_period();
2713 0) sh-4802 | 2.899 us | 2604 0) sh-4802 | 2.899 us | }
2714 0) sh-4802 | 4.040 us | 2605 0) sh-4802 | 4.040 us | }
2715 0) sh-4802 | 5.151 us | 2606 0) sh-4802 | 5.151 us | }
2716 0) sh-4802 | + 49.370 us | 2607 0) sh-4802 | + 49.370 us | }
2717 2608
2718 2609
2719 - The absolute time field is an absolute time 2610 - The absolute time field is an absolute timestamp given by the
2720 system clock since it started. A snapshot o 2611 system clock since it started. A snapshot of this time is
2721 given on each entry/exit of functions 2612 given on each entry/exit of functions
2722 2613
2723 - hide: echo nofuncgraph-abstime > tr 2614 - hide: echo nofuncgraph-abstime > trace_options
2724 - show: echo funcgraph-abstime > trac 2615 - show: echo funcgraph-abstime > trace_options
2725 2616
2726 ie:: 2617 ie::
2727 2618
2728 # 2619 #
2729 # TIME CPU DURATION 2620 # TIME CPU DURATION FUNCTION CALLS
2730 # | | | | 2621 # | | | | | | | |
2731 360.774522 | 1) 0.541 us | 2622 360.774522 | 1) 0.541 us | }
2732 360.774522 | 1) 4.663 us | 2623 360.774522 | 1) 4.663 us | }
2733 360.774523 | 1) 0.541 us | 2624 360.774523 | 1) 0.541 us | __wake_up_bit();
2734 360.774524 | 1) 6.796 us | 2625 360.774524 | 1) 6.796 us | }
2735 360.774524 | 1) 7.952 us | 2626 360.774524 | 1) 7.952 us | }
2736 360.774525 | 1) 9.063 us | 2627 360.774525 | 1) 9.063 us | }
2737 360.774525 | 1) 0.615 us | 2628 360.774525 | 1) 0.615 us | journal_mark_dirty();
2738 360.774527 | 1) 0.578 us | 2629 360.774527 | 1) 0.578 us | __brelse();
2739 360.774528 | 1) | 2630 360.774528 | 1) | reiserfs_prepare_for_journal() {
2740 360.774528 | 1) | 2631 360.774528 | 1) | unlock_buffer() {
2741 360.774529 | 1) | 2632 360.774529 | 1) | wake_up_bit() {
2742 360.774529 | 1) | 2633 360.774529 | 1) | bit_waitqueue() {
2743 360.774530 | 1) 0.594 us | 2634 360.774530 | 1) 0.594 us | __phys_addr();
2744 2635
2745 2636
2746 The function name is always displayed after t 2637 The function name is always displayed after the closing bracket
2747 for a function if the start of that function 2638 for a function if the start of that function is not in the
2748 trace buffer. 2639 trace buffer.
2749 2640
2750 Display of the function name after the closin 2641 Display of the function name after the closing bracket may be
2751 enabled for functions whose start is in the t 2642 enabled for functions whose start is in the trace buffer,
2752 allowing easier searching with grep for funct 2643 allowing easier searching with grep for function durations.
2753 It is default disabled. 2644 It is default disabled.
2754 2645
2755 - hide: echo nofuncgraph-tail > trace 2646 - hide: echo nofuncgraph-tail > trace_options
2756 - show: echo funcgraph-tail > trace_o 2647 - show: echo funcgraph-tail > trace_options
2757 2648
2758 Example with nofuncgraph-tail (default):: 2649 Example with nofuncgraph-tail (default)::
2759 2650
2760 0) | putname() { 2651 0) | putname() {
2761 0) | kmem_cache_free 2652 0) | kmem_cache_free() {
2762 0) 0.518 us | __phys_addr() 2653 0) 0.518 us | __phys_addr();
2763 0) 1.757 us | } 2654 0) 1.757 us | }
2764 0) 2.861 us | } 2655 0) 2.861 us | }
2765 2656
2766 Example with funcgraph-tail:: 2657 Example with funcgraph-tail::
2767 2658
2768 0) | putname() { 2659 0) | putname() {
2769 0) | kmem_cache_free 2660 0) | kmem_cache_free() {
2770 0) 0.518 us | __phys_addr() 2661 0) 0.518 us | __phys_addr();
2771 0) 1.757 us | } /* kmem_cache 2662 0) 1.757 us | } /* kmem_cache_free() */
2772 0) 2.861 us | } /* putname() */ 2663 0) 2.861 us | } /* putname() */
2773 2664
2774 The return value of each traced function can <<
2775 an equal sign "=". When encountering system c <<
2776 can be very helpful to quickly locate the fun <<
2777 returns an error code. <<
2778 <<
2779 - hide: echo nofuncgraph-retval > tra <<
2780 - show: echo funcgraph-retval > trace <<
2781 <<
2782 Example with funcgraph-retval:: <<
2783 <<
2784 1) | cgroup_migrate() { <<
2785 1) 0.651 us | cgroup_migrate_ad <<
2786 1) | cgroup_migrate_ex <<
2787 1) | cpu_cgroup_can_ <<
2788 1) | cgroup_taskse <<
2789 1) 0.732 us | cgroup_task <<
2790 1) 1.232 us | } /* cgroup_t <<
2791 1) 0.380 us | sched_rt_can_ <<
2792 1) 2.335 us | } /* cpu_cgroup <<
2793 1) 4.369 us | } /* cgroup_migra <<
2794 1) 7.143 us | } /* cgroup_migrate <<
2795 <<
2796 The above example shows that the function cpu <<
2797 returned the error code -22 firstly, then we <<
2798 of this function to get the root cause. <<
2799 <<
2800 When the option funcgraph-retval-hex is not s <<
2801 be displayed in a smart way. Specifically, if <<
2802 it will be printed in signed decimal format, <<
2803 printed in hexadecimal format. <<
2804 <<
2805 - smart: echo nofuncgraph-retval-hex <<
2806 - hexadecimal: echo funcgraph-retval- <<
2807 <<
2808 Example with funcgraph-retval-hex:: <<
2809 <<
2810 1) | cgroup_migrate() <<
2811 1) 0.651 us | cgroup_migrate_ <<
2812 1) | cgroup_migrate_ <<
2813 1) | cpu_cgroup_ca <<
2814 1) | cgroup_task <<
2815 1) 0.732 us | cgroup_ta <<
2816 1) 1.232 us | } /* cgroup <<
2817 1) 0.380 us | sched_rt_ca <<
2818 1) 2.335 us | } /* cpu_cgro <<
2819 1) 4.369 us | } /* cgroup_mig <<
2820 1) 7.143 us | } /* cgroup_migra <<
2821 <<
2822 At present, there are some limitations when u <<
2823 option, and these limitations will be elimina <<
2824 <<
2825 - Even if the function return type is void, a <<
2826 be printed, and you can just ignore it. <<
2827 <<
2828 - Even if return values are stored in multipl <<
2829 value contained in the first register will <<
2830 To illustrate, in the x86 architecture, eax <<
2831 a 64-bit return value, with the lower 32 bi <<
2832 upper 32 bits saved in edx. However, only t <<
2833 will be recorded and printed. <<
2834 <<
2835 - In certain procedure call standards, such a <<
2836 type is smaller than a GPR, it is the respo <<
2837 to perform the narrowing, and the upper bit <<
2838 Therefore, it is advisable to check the cod <<
2839 when using a u8 in a 64-bit GPR, bits [63:8 <<
2840 especially when larger types are truncated, <<
2841 Here are some specific cases to illustrate <<
2842 <<
2843 **Case One**: <<
2844 <<
2845 The function narrow_to_u8 is defined as fol <<
2846 <<
2847 u8 narrow_to_u8(u64 val) <<
2848 { <<
2849 // implicitly truncated <<
2850 return val; <<
2851 } <<
2852 <<
2853 It may be compiled to:: <<
2854 <<
2855 narrow_to_u8: <<
2856 < ... ftrace instrumentation <<
2857 RET <<
2858 <<
2859 If you pass 0x123456789abcdef to this funct <<
2860 it may be recorded as 0x123456789abcdef ins <<
2861 <<
2862 **Case Two**: <<
2863 <<
2864 The function error_if_not_4g_aligned is def <<
2865 <<
2866 int error_if_not_4g_aligned(u64 val) <<
2867 { <<
2868 if (val & GENMASK(31, 0)) <<
2869 return -EINVAL; <<
2870 <<
2871 return 0; <<
2872 } <<
2873 <<
2874 It could be compiled to:: <<
2875 <<
2876 error_if_not_4g_aligned: <<
2877 CBNZ w0, .Lnot_aligned <<
2878 RET // bi <<
2879 // [6 <<
2880 .Lnot_aligned: <<
2881 MOV x0, #-EINVAL <<
2882 RET <<
2883 <<
2884 When passing 0x2_0000_0000 to it, the retur <<
2885 0x2_0000_0000 instead of 0. <<
2886 <<
2887 You can put some comments on specific functio 2665 You can put some comments on specific functions by using
2888 trace_printk() For example, if you want to pu 2666 trace_printk() For example, if you want to put a comment inside
2889 the __might_sleep() function, you just have t 2667 the __might_sleep() function, you just have to include
2890 <linux/ftrace.h> and call trace_printk() insi 2668 <linux/ftrace.h> and call trace_printk() inside __might_sleep()::
2891 2669
2892 trace_printk("I'm a comment!\n") 2670 trace_printk("I'm a comment!\n")
2893 2671
2894 will produce:: 2672 will produce::
2895 2673
2896 1) | __might_sle 2674 1) | __might_sleep() {
2897 1) | /* I'm a 2675 1) | /* I'm a comment! */
2898 1) 1.449 us | } 2676 1) 1.449 us | }
2899 2677
2900 2678
2901 You might find other useful features for this 2679 You might find other useful features for this tracer in the
2902 following "dynamic ftrace" section such as tr 2680 following "dynamic ftrace" section such as tracing only specific
2903 functions or tasks. 2681 functions or tasks.
2904 2682
2905 dynamic ftrace 2683 dynamic ftrace
2906 -------------- 2684 --------------
2907 2685
2908 If CONFIG_DYNAMIC_FTRACE is set, the system w 2686 If CONFIG_DYNAMIC_FTRACE is set, the system will run with
2909 virtually no overhead when function tracing i 2687 virtually no overhead when function tracing is disabled. The way
2910 this works is the mcount function call (place 2688 this works is the mcount function call (placed at the start of
2911 every kernel function, produced by the -pg sw 2689 every kernel function, produced by the -pg switch in gcc),
2912 starts of pointing to a simple return. (Enabl 2690 starts of pointing to a simple return. (Enabling FTRACE will
2913 include the -pg switch in the compiling of th 2691 include the -pg switch in the compiling of the kernel.)
2914 2692
2915 At compile time every C file object is run th 2693 At compile time every C file object is run through the
2916 recordmcount program (located in the scripts 2694 recordmcount program (located in the scripts directory). This
2917 program will parse the ELF headers in the C o 2695 program will parse the ELF headers in the C object to find all
2918 the locations in the .text section that call 2696 the locations in the .text section that call mcount. Starting
2919 with gcc version 4.6, the -mfentry has been a 2697 with gcc version 4.6, the -mfentry has been added for x86, which
2920 calls "__fentry__" instead of "mcount". Which 2698 calls "__fentry__" instead of "mcount". Which is called before
2921 the creation of the stack frame. 2699 the creation of the stack frame.
2922 2700
2923 Note, not all sections are traced. They may b 2701 Note, not all sections are traced. They may be prevented by either
2924 a notrace, or blocked another way and all inl 2702 a notrace, or blocked another way and all inline functions are not
2925 traced. Check the "available_filter_functions 2703 traced. Check the "available_filter_functions" file to see what functions
2926 can be traced. 2704 can be traced.
2927 2705
2928 A section called "__mcount_loc" is created th 2706 A section called "__mcount_loc" is created that holds
2929 references to all the mcount/fentry call site 2707 references to all the mcount/fentry call sites in the .text section.
2930 The recordmcount program re-links this sectio 2708 The recordmcount program re-links this section back into the
2931 original object. The final linking stage of t 2709 original object. The final linking stage of the kernel will add all these
2932 references into a single table. 2710 references into a single table.
2933 2711
2934 On boot up, before SMP is initialized, the dy 2712 On boot up, before SMP is initialized, the dynamic ftrace code
2935 scans this table and updates all the location 2713 scans this table and updates all the locations into nops. It
2936 also records the locations, which are added t 2714 also records the locations, which are added to the
2937 available_filter_functions list. Modules are 2715 available_filter_functions list. Modules are processed as they
2938 are loaded and before they are executed. Whe 2716 are loaded and before they are executed. When a module is
2939 unloaded, it also removes its functions from 2717 unloaded, it also removes its functions from the ftrace function
2940 list. This is automatic in the module unload 2718 list. This is automatic in the module unload code, and the
2941 module author does not need to worry about it 2719 module author does not need to worry about it.
2942 2720
2943 When tracing is enabled, the process of modif 2721 When tracing is enabled, the process of modifying the function
2944 tracepoints is dependent on architecture. The 2722 tracepoints is dependent on architecture. The old method is to use
2945 kstop_machine to prevent races with the CPUs 2723 kstop_machine to prevent races with the CPUs executing code being
2946 modified (which can cause the CPU to do undes 2724 modified (which can cause the CPU to do undesirable things, especially
2947 if the modified code crosses cache (or page) 2725 if the modified code crosses cache (or page) boundaries), and the nops are
2948 patched back to calls. But this time, they do 2726 patched back to calls. But this time, they do not call mcount
2949 (which is just a function stub). They now cal 2727 (which is just a function stub). They now call into the ftrace
2950 infrastructure. 2728 infrastructure.
2951 2729
2952 The new method of modifying the function trac 2730 The new method of modifying the function tracepoints is to place
2953 a breakpoint at the location to be modified, 2731 a breakpoint at the location to be modified, sync all CPUs, modify
2954 the rest of the instruction not covered by th 2732 the rest of the instruction not covered by the breakpoint. Sync
2955 all CPUs again, and then remove the breakpoin 2733 all CPUs again, and then remove the breakpoint with the finished
2956 version to the ftrace call site. 2734 version to the ftrace call site.
2957 2735
2958 Some archs do not even need to monkey around 2736 Some archs do not even need to monkey around with the synchronization,
2959 and can just slap the new code on top of the 2737 and can just slap the new code on top of the old without any
2960 problems with other CPUs executing it at the 2738 problems with other CPUs executing it at the same time.
2961 2739
2962 One special side-effect to the recording of t 2740 One special side-effect to the recording of the functions being
2963 traced is that we can now selectively choose 2741 traced is that we can now selectively choose which functions we
2964 wish to trace and which ones we want the mcou 2742 wish to trace and which ones we want the mcount calls to remain
2965 as nops. 2743 as nops.
2966 2744
2967 Two files are used, one for enabling and one 2745 Two files are used, one for enabling and one for disabling the
2968 tracing of specified functions. They are: 2746 tracing of specified functions. They are:
2969 2747
2970 set_ftrace_filter 2748 set_ftrace_filter
2971 2749
2972 and 2750 and
2973 2751
2974 set_ftrace_notrace 2752 set_ftrace_notrace
2975 2753
2976 A list of available functions that you can ad 2754 A list of available functions that you can add to these files is
2977 listed in: 2755 listed in:
2978 2756
2979 available_filter_functions 2757 available_filter_functions
2980 2758
2981 :: 2759 ::
2982 2760
2983 # cat available_filter_functions 2761 # cat available_filter_functions
2984 put_prev_task_idle 2762 put_prev_task_idle
2985 kmem_cache_create 2763 kmem_cache_create
2986 pick_next_task_rt 2764 pick_next_task_rt
2987 cpus_read_lock !! 2765 get_online_cpus
2988 pick_next_task_fair 2766 pick_next_task_fair
2989 mutex_lock 2767 mutex_lock
2990 [...] 2768 [...]
2991 2769
2992 If I am only interested in sys_nanosleep and 2770 If I am only interested in sys_nanosleep and hrtimer_interrupt::
2993 2771
2994 # echo sys_nanosleep hrtimer_interrupt > se 2772 # echo sys_nanosleep hrtimer_interrupt > set_ftrace_filter
2995 # echo function > current_tracer 2773 # echo function > current_tracer
2996 # echo 1 > tracing_on 2774 # echo 1 > tracing_on
2997 # usleep 1 2775 # usleep 1
2998 # echo 0 > tracing_on 2776 # echo 0 > tracing_on
2999 # cat trace 2777 # cat trace
3000 # tracer: function 2778 # tracer: function
3001 # 2779 #
3002 # entries-in-buffer/entries-written: 5/5 2780 # entries-in-buffer/entries-written: 5/5 #P:4
3003 # 2781 #
3004 # _-----=> irq 2782 # _-----=> irqs-off
3005 # / _----=> nee 2783 # / _----=> need-resched
3006 # | / _---=> har 2784 # | / _---=> hardirq/softirq
3007 # || / _--=> pre 2785 # || / _--=> preempt-depth
3008 # ||| / dela 2786 # ||| / delay
3009 # TASK-PID CPU# |||| TIMEST 2787 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3010 # | | | |||| | 2788 # | | | |||| | |
3011 usleep-2665 [001] .... 4186.475 2789 usleep-2665 [001] .... 4186.475355: sys_nanosleep <-system_call_fastpath
3012 <idle>-0 [001] d.h1 4186.475 2790 <idle>-0 [001] d.h1 4186.475409: hrtimer_interrupt <-smp_apic_timer_interrupt
3013 usleep-2665 [001] d.h1 4186.475 2791 usleep-2665 [001] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
3014 <idle>-0 [003] d.h1 4186.475 2792 <idle>-0 [003] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
3015 <idle>-0 [002] d.h1 4186.475 2793 <idle>-0 [002] d.h1 4186.475427: hrtimer_interrupt <-smp_apic_timer_interrupt
3016 2794
3017 To see which functions are being traced, you 2795 To see which functions are being traced, you can cat the file:
3018 :: 2796 ::
3019 2797
3020 # cat set_ftrace_filter 2798 # cat set_ftrace_filter
3021 hrtimer_interrupt 2799 hrtimer_interrupt
3022 sys_nanosleep 2800 sys_nanosleep
3023 2801
3024 2802
3025 Perhaps this is not enough. The filters also 2803 Perhaps this is not enough. The filters also allow glob(7) matching.
3026 2804
3027 ``<match>*`` 2805 ``<match>*``
3028 will match functions that begin with 2806 will match functions that begin with <match>
3029 ``*<match>`` 2807 ``*<match>``
3030 will match functions that end with <m 2808 will match functions that end with <match>
3031 ``*<match>*`` 2809 ``*<match>*``
3032 will match functions that have <match 2810 will match functions that have <match> in it
3033 ``<match1>*<match2>`` 2811 ``<match1>*<match2>``
3034 will match functions that begin with 2812 will match functions that begin with <match1> and end with <match2>
3035 2813
3036 .. note:: 2814 .. note::
3037 It is better to use quotes to enclose t 2815 It is better to use quotes to enclose the wild cards,
3038 otherwise the shell may expand the para 2816 otherwise the shell may expand the parameters into names
3039 of files in the local directory. 2817 of files in the local directory.
3040 2818
3041 :: 2819 ::
3042 2820
3043 # echo 'hrtimer_*' > set_ftrace_filter 2821 # echo 'hrtimer_*' > set_ftrace_filter
3044 2822
3045 Produces:: 2823 Produces::
3046 2824
3047 # tracer: function 2825 # tracer: function
3048 # 2826 #
3049 # entries-in-buffer/entries-written: 897/89 2827 # entries-in-buffer/entries-written: 897/897 #P:4
3050 # 2828 #
3051 # _-----=> irq 2829 # _-----=> irqs-off
3052 # / _----=> nee 2830 # / _----=> need-resched
3053 # | / _---=> har 2831 # | / _---=> hardirq/softirq
3054 # || / _--=> pre 2832 # || / _--=> preempt-depth
3055 # ||| / dela 2833 # ||| / delay
3056 # TASK-PID CPU# |||| TIMEST 2834 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3057 # | | | |||| | 2835 # | | | |||| | |
3058 <idle>-0 [003] dN.1 4228.547 2836 <idle>-0 [003] dN.1 4228.547803: hrtimer_cancel <-tick_nohz_idle_exit
3059 <idle>-0 [003] dN.1 4228.547 2837 <idle>-0 [003] dN.1 4228.547804: hrtimer_try_to_cancel <-hrtimer_cancel
3060 <idle>-0 [003] dN.2 4228.547 2838 <idle>-0 [003] dN.2 4228.547805: hrtimer_force_reprogram <-__remove_hrtimer
3061 <idle>-0 [003] dN.1 4228.547 2839 <idle>-0 [003] dN.1 4228.547805: hrtimer_forward <-tick_nohz_idle_exit
3062 <idle>-0 [003] dN.1 4228.547 2840 <idle>-0 [003] dN.1 4228.547805: hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
3063 <idle>-0 [003] d..1 4228.547 2841 <idle>-0 [003] d..1 4228.547858: hrtimer_get_next_event <-get_next_timer_interrupt
3064 <idle>-0 [003] d..1 4228.547 2842 <idle>-0 [003] d..1 4228.547859: hrtimer_start <-__tick_nohz_idle_enter
3065 <idle>-0 [003] d..2 4228.547 2843 <idle>-0 [003] d..2 4228.547860: hrtimer_force_reprogram <-__rem
3066 2844
3067 Notice that we lost the sys_nanosleep. 2845 Notice that we lost the sys_nanosleep.
3068 :: 2846 ::
3069 2847
3070 # cat set_ftrace_filter 2848 # cat set_ftrace_filter
3071 hrtimer_run_queues 2849 hrtimer_run_queues
3072 hrtimer_run_pending 2850 hrtimer_run_pending
3073 hrtimer_init 2851 hrtimer_init
3074 hrtimer_cancel 2852 hrtimer_cancel
3075 hrtimer_try_to_cancel 2853 hrtimer_try_to_cancel
3076 hrtimer_forward 2854 hrtimer_forward
3077 hrtimer_start 2855 hrtimer_start
3078 hrtimer_reprogram 2856 hrtimer_reprogram
3079 hrtimer_force_reprogram 2857 hrtimer_force_reprogram
3080 hrtimer_get_next_event 2858 hrtimer_get_next_event
3081 hrtimer_interrupt 2859 hrtimer_interrupt
3082 hrtimer_nanosleep 2860 hrtimer_nanosleep
3083 hrtimer_wakeup 2861 hrtimer_wakeup
3084 hrtimer_get_remaining 2862 hrtimer_get_remaining
3085 hrtimer_get_res 2863 hrtimer_get_res
3086 hrtimer_init_sleeper 2864 hrtimer_init_sleeper
3087 2865
3088 2866
3089 This is because the '>' and '>>' act just lik 2867 This is because the '>' and '>>' act just like they do in bash.
3090 To rewrite the filters, use '>' 2868 To rewrite the filters, use '>'
3091 To append to the filters, use '>>' 2869 To append to the filters, use '>>'
3092 2870
3093 To clear out a filter so that all functions w 2871 To clear out a filter so that all functions will be recorded
3094 again:: 2872 again::
3095 2873
3096 # echo > set_ftrace_filter 2874 # echo > set_ftrace_filter
3097 # cat set_ftrace_filter 2875 # cat set_ftrace_filter
3098 # 2876 #
3099 2877
3100 Again, now we want to append. 2878 Again, now we want to append.
3101 2879
3102 :: 2880 ::
3103 2881
3104 # echo sys_nanosleep > set_ftrace_filter 2882 # echo sys_nanosleep > set_ftrace_filter
3105 # cat set_ftrace_filter 2883 # cat set_ftrace_filter
3106 sys_nanosleep 2884 sys_nanosleep
3107 # echo 'hrtimer_*' >> set_ftrace_filter 2885 # echo 'hrtimer_*' >> set_ftrace_filter
3108 # cat set_ftrace_filter 2886 # cat set_ftrace_filter
3109 hrtimer_run_queues 2887 hrtimer_run_queues
3110 hrtimer_run_pending 2888 hrtimer_run_pending
3111 hrtimer_init 2889 hrtimer_init
3112 hrtimer_cancel 2890 hrtimer_cancel
3113 hrtimer_try_to_cancel 2891 hrtimer_try_to_cancel
3114 hrtimer_forward 2892 hrtimer_forward
3115 hrtimer_start 2893 hrtimer_start
3116 hrtimer_reprogram 2894 hrtimer_reprogram
3117 hrtimer_force_reprogram 2895 hrtimer_force_reprogram
3118 hrtimer_get_next_event 2896 hrtimer_get_next_event
3119 hrtimer_interrupt 2897 hrtimer_interrupt
3120 sys_nanosleep 2898 sys_nanosleep
3121 hrtimer_nanosleep 2899 hrtimer_nanosleep
3122 hrtimer_wakeup 2900 hrtimer_wakeup
3123 hrtimer_get_remaining 2901 hrtimer_get_remaining
3124 hrtimer_get_res 2902 hrtimer_get_res
3125 hrtimer_init_sleeper 2903 hrtimer_init_sleeper
3126 2904
3127 2905
3128 The set_ftrace_notrace prevents those functio 2906 The set_ftrace_notrace prevents those functions from being
3129 traced. 2907 traced.
3130 :: 2908 ::
3131 2909
3132 # echo '*preempt*' '*lock*' > set_ftrace_no 2910 # echo '*preempt*' '*lock*' > set_ftrace_notrace
3133 2911
3134 Produces:: 2912 Produces::
3135 2913
3136 # tracer: function 2914 # tracer: function
3137 # 2915 #
3138 # entries-in-buffer/entries-written: 39608/ 2916 # entries-in-buffer/entries-written: 39608/39608 #P:4
3139 # 2917 #
3140 # _-----=> irq 2918 # _-----=> irqs-off
3141 # / _----=> nee 2919 # / _----=> need-resched
3142 # | / _---=> har 2920 # | / _---=> hardirq/softirq
3143 # || / _--=> pre 2921 # || / _--=> preempt-depth
3144 # ||| / dela 2922 # ||| / delay
3145 # TASK-PID CPU# |||| TIMEST 2923 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3146 # | | | |||| | 2924 # | | | |||| | |
3147 bash-1994 [000] .... 4342.324 2925 bash-1994 [000] .... 4342.324896: file_ra_state_init <-do_dentry_open
3148 bash-1994 [000] .... 4342.324 2926 bash-1994 [000] .... 4342.324897: open_check_o_direct <-do_last
3149 bash-1994 [000] .... 4342.324 2927 bash-1994 [000] .... 4342.324897: ima_file_check <-do_last
3150 bash-1994 [000] .... 4342.324 2928 bash-1994 [000] .... 4342.324898: process_measurement <-ima_file_check
3151 bash-1994 [000] .... 4342.324 2929 bash-1994 [000] .... 4342.324898: ima_get_action <-process_measurement
3152 bash-1994 [000] .... 4342.324 2930 bash-1994 [000] .... 4342.324898: ima_match_policy <-ima_get_action
3153 bash-1994 [000] .... 4342.324 2931 bash-1994 [000] .... 4342.324899: do_truncate <-do_last
3154 bash-1994 [000] .... 4342.324 !! 2932 bash-1994 [000] .... 4342.324899: should_remove_suid <-do_truncate
3155 bash-1994 [000] .... 4342.324 2933 bash-1994 [000] .... 4342.324899: notify_change <-do_truncate
3156 bash-1994 [000] .... 4342.324 2934 bash-1994 [000] .... 4342.324900: current_fs_time <-notify_change
3157 bash-1994 [000] .... 4342.324 2935 bash-1994 [000] .... 4342.324900: current_kernel_time <-current_fs_time
3158 bash-1994 [000] .... 4342.324 2936 bash-1994 [000] .... 4342.324900: timespec_trunc <-current_fs_time
3159 2937
3160 We can see that there's no more lock or preem 2938 We can see that there's no more lock or preempt tracing.
3161 2939
3162 Selecting function filters via index 2940 Selecting function filters via index
3163 ------------------------------------ 2941 ------------------------------------
3164 2942
3165 Because processing of strings is expensive (t 2943 Because processing of strings is expensive (the address of the function
3166 needs to be looked up before comparing to the 2944 needs to be looked up before comparing to the string being passed in),
3167 an index can be used as well to enable functi 2945 an index can be used as well to enable functions. This is useful in the
3168 case of setting thousands of specific functio 2946 case of setting thousands of specific functions at a time. By passing
3169 in a list of numbers, no string processing wi 2947 in a list of numbers, no string processing will occur. Instead, the function
3170 at the specific location in the internal arra 2948 at the specific location in the internal array (which corresponds to the
3171 functions in the "available_filter_functions" 2949 functions in the "available_filter_functions" file), is selected.
3172 2950
3173 :: 2951 ::
3174 2952
3175 # echo 1 > set_ftrace_filter 2953 # echo 1 > set_ftrace_filter
3176 2954
3177 Will select the first function listed in "ava 2955 Will select the first function listed in "available_filter_functions"
3178 2956
3179 :: 2957 ::
3180 2958
3181 # head -1 available_filter_functions 2959 # head -1 available_filter_functions
3182 trace_initcall_finish_cb 2960 trace_initcall_finish_cb
3183 2961
3184 # cat set_ftrace_filter 2962 # cat set_ftrace_filter
3185 trace_initcall_finish_cb 2963 trace_initcall_finish_cb
3186 2964
3187 # head -50 available_filter_functions | tai 2965 # head -50 available_filter_functions | tail -1
3188 x86_pmu_commit_txn 2966 x86_pmu_commit_txn
3189 2967
3190 # echo 1 50 > set_ftrace_filter 2968 # echo 1 50 > set_ftrace_filter
3191 # cat set_ftrace_filter 2969 # cat set_ftrace_filter
3192 trace_initcall_finish_cb 2970 trace_initcall_finish_cb
3193 x86_pmu_commit_txn 2971 x86_pmu_commit_txn
3194 2972
3195 Dynamic ftrace with the function graph tracer 2973 Dynamic ftrace with the function graph tracer
3196 --------------------------------------------- 2974 ---------------------------------------------
3197 2975
3198 Although what has been explained above concer 2976 Although what has been explained above concerns both the
3199 function tracer and the function-graph-tracer 2977 function tracer and the function-graph-tracer, there are some
3200 special features only available in the functi 2978 special features only available in the function-graph tracer.
3201 2979
3202 If you want to trace only one function and al 2980 If you want to trace only one function and all of its children,
3203 you just have to echo its name into set_graph 2981 you just have to echo its name into set_graph_function::
3204 2982
3205 echo __do_fault > set_graph_function 2983 echo __do_fault > set_graph_function
3206 2984
3207 will produce the following "expanded" trace o 2985 will produce the following "expanded" trace of the __do_fault()
3208 function:: 2986 function::
3209 2987
3210 0) | __do_fault() { 2988 0) | __do_fault() {
3211 0) | filemap_fault() { 2989 0) | filemap_fault() {
3212 0) | find_lock_page() { 2990 0) | find_lock_page() {
3213 0) 0.804 us | find_get_page(); 2991 0) 0.804 us | find_get_page();
3214 0) | __might_sleep() 2992 0) | __might_sleep() {
3215 0) 1.329 us | } 2993 0) 1.329 us | }
3216 0) 3.904 us | } 2994 0) 3.904 us | }
3217 0) 4.979 us | } 2995 0) 4.979 us | }
3218 0) 0.653 us | _spin_lock(); 2996 0) 0.653 us | _spin_lock();
3219 0) 0.578 us | page_add_file_rmap() 2997 0) 0.578 us | page_add_file_rmap();
3220 0) 0.525 us | native_set_pte_at(); 2998 0) 0.525 us | native_set_pte_at();
3221 0) 0.585 us | _spin_unlock(); 2999 0) 0.585 us | _spin_unlock();
3222 0) | unlock_page() { 3000 0) | unlock_page() {
3223 0) 0.541 us | page_waitqueue(); 3001 0) 0.541 us | page_waitqueue();
3224 0) 0.639 us | __wake_up_bit(); 3002 0) 0.639 us | __wake_up_bit();
3225 0) 2.786 us | } 3003 0) 2.786 us | }
3226 0) + 14.237 us | } 3004 0) + 14.237 us | }
3227 0) | __do_fault() { 3005 0) | __do_fault() {
3228 0) | filemap_fault() { 3006 0) | filemap_fault() {
3229 0) | find_lock_page() { 3007 0) | find_lock_page() {
3230 0) 0.698 us | find_get_page(); 3008 0) 0.698 us | find_get_page();
3231 0) | __might_sleep() 3009 0) | __might_sleep() {
3232 0) 1.412 us | } 3010 0) 1.412 us | }
3233 0) 3.950 us | } 3011 0) 3.950 us | }
3234 0) 5.098 us | } 3012 0) 5.098 us | }
3235 0) 0.631 us | _spin_lock(); 3013 0) 0.631 us | _spin_lock();
3236 0) 0.571 us | page_add_file_rmap() 3014 0) 0.571 us | page_add_file_rmap();
3237 0) 0.526 us | native_set_pte_at(); 3015 0) 0.526 us | native_set_pte_at();
3238 0) 0.586 us | _spin_unlock(); 3016 0) 0.586 us | _spin_unlock();
3239 0) | unlock_page() { 3017 0) | unlock_page() {
3240 0) 0.533 us | page_waitqueue(); 3018 0) 0.533 us | page_waitqueue();
3241 0) 0.638 us | __wake_up_bit(); 3019 0) 0.638 us | __wake_up_bit();
3242 0) 2.793 us | } 3020 0) 2.793 us | }
3243 0) + 14.012 us | } 3021 0) + 14.012 us | }
3244 3022
3245 You can also expand several functions at once 3023 You can also expand several functions at once::
3246 3024
3247 echo sys_open > set_graph_function 3025 echo sys_open > set_graph_function
3248 echo sys_close >> set_graph_function 3026 echo sys_close >> set_graph_function
3249 3027
3250 Now if you want to go back to trace all funct 3028 Now if you want to go back to trace all functions you can clear
3251 this special filter via:: 3029 this special filter via::
3252 3030
3253 echo > set_graph_function 3031 echo > set_graph_function
3254 3032
3255 3033
3256 ftrace_enabled 3034 ftrace_enabled
3257 -------------- 3035 --------------
3258 3036
3259 Note, the proc sysctl ftrace_enable is a big 3037 Note, the proc sysctl ftrace_enable is a big on/off switch for the
3260 function tracer. By default it is enabled (wh 3038 function tracer. By default it is enabled (when function tracing is
3261 enabled in the kernel). If it is disabled, al 3039 enabled in the kernel). If it is disabled, all function tracing is
3262 disabled. This includes not only the function 3040 disabled. This includes not only the function tracers for ftrace, but
3263 also for any other uses (perf, kprobes, stack 3041 also for any other uses (perf, kprobes, stack tracing, profiling, etc). It
3264 cannot be disabled if there is a callback wit 3042 cannot be disabled if there is a callback with FTRACE_OPS_FL_PERMANENT set
3265 registered. 3043 registered.
3266 3044
3267 Please disable this with care. 3045 Please disable this with care.
3268 3046
3269 This can be disable (and enabled) with:: 3047 This can be disable (and enabled) with::
3270 3048
3271 sysctl kernel.ftrace_enabled=0 3049 sysctl kernel.ftrace_enabled=0
3272 sysctl kernel.ftrace_enabled=1 3050 sysctl kernel.ftrace_enabled=1
3273 3051
3274 or 3052 or
3275 3053
3276 echo 0 > /proc/sys/kernel/ftrace_enabled 3054 echo 0 > /proc/sys/kernel/ftrace_enabled
3277 echo 1 > /proc/sys/kernel/ftrace_enabled 3055 echo 1 > /proc/sys/kernel/ftrace_enabled
3278 3056
3279 3057
3280 Filter commands 3058 Filter commands
3281 --------------- 3059 ---------------
3282 3060
3283 A few commands are supported by the set_ftrac 3061 A few commands are supported by the set_ftrace_filter interface.
3284 Trace commands have the following format:: 3062 Trace commands have the following format::
3285 3063
3286 <function>:<command>:<parameter> 3064 <function>:<command>:<parameter>
3287 3065
3288 The following commands are supported: 3066 The following commands are supported:
3289 3067
3290 - mod: 3068 - mod:
3291 This command enables function filtering per 3069 This command enables function filtering per module. The
3292 parameter defines the module. For example, 3070 parameter defines the module. For example, if only the write*
3293 functions in the ext3 module are desired, r 3071 functions in the ext3 module are desired, run:
3294 3072
3295 echo 'write*:mod:ext3' > set_ftrace_filter 3073 echo 'write*:mod:ext3' > set_ftrace_filter
3296 3074
3297 This command interacts with the filter in t 3075 This command interacts with the filter in the same way as
3298 filtering based on function names. Thus, ad 3076 filtering based on function names. Thus, adding more functions
3299 in a different module is accomplished by ap 3077 in a different module is accomplished by appending (>>) to the
3300 filter file. Remove specific module functio 3078 filter file. Remove specific module functions by prepending
3301 '!':: 3079 '!'::
3302 3080
3303 echo '!writeback*:mod:ext3' >> set_ftrace_ 3081 echo '!writeback*:mod:ext3' >> set_ftrace_filter
3304 3082
3305 Mod command supports module globbing. Disab 3083 Mod command supports module globbing. Disable tracing for all
3306 functions except a specific module:: 3084 functions except a specific module::
3307 3085
3308 echo '!*:mod:!ext3' >> set_ftrace_filter 3086 echo '!*:mod:!ext3' >> set_ftrace_filter
3309 3087
3310 Disable tracing for all modules, but still 3088 Disable tracing for all modules, but still trace kernel::
3311 3089
3312 echo '!*:mod:*' >> set_ftrace_filter 3090 echo '!*:mod:*' >> set_ftrace_filter
3313 3091
3314 Enable filter only for kernel:: 3092 Enable filter only for kernel::
3315 3093
3316 echo '*write*:mod:!*' >> set_ftrace_filter 3094 echo '*write*:mod:!*' >> set_ftrace_filter
3317 3095
3318 Enable filter for module globbing:: 3096 Enable filter for module globbing::
3319 3097
3320 echo '*write*:mod:*snd*' >> set_ftrace_fil 3098 echo '*write*:mod:*snd*' >> set_ftrace_filter
3321 3099
3322 - traceon/traceoff: 3100 - traceon/traceoff:
3323 These commands turn tracing on and off when 3101 These commands turn tracing on and off when the specified
3324 functions are hit. The parameter determines 3102 functions are hit. The parameter determines how many times the
3325 tracing system is turned on and off. If uns 3103 tracing system is turned on and off. If unspecified, there is
3326 no limit. For example, to disable tracing w 3104 no limit. For example, to disable tracing when a schedule bug
3327 is hit the first 5 times, run:: 3105 is hit the first 5 times, run::
3328 3106
3329 echo '__schedule_bug:traceoff:5' > set_ftr 3107 echo '__schedule_bug:traceoff:5' > set_ftrace_filter
3330 3108
3331 To always disable tracing when __schedule_b 3109 To always disable tracing when __schedule_bug is hit::
3332 3110
3333 echo '__schedule_bug:traceoff' > set_ftrac 3111 echo '__schedule_bug:traceoff' > set_ftrace_filter
3334 3112
3335 These commands are cumulative whether or no 3113 These commands are cumulative whether or not they are appended
3336 to set_ftrace_filter. To remove a command, 3114 to set_ftrace_filter. To remove a command, prepend it by '!'
3337 and drop the parameter:: 3115 and drop the parameter::
3338 3116
3339 echo '!__schedule_bug:traceoff:0' > set_ft 3117 echo '!__schedule_bug:traceoff:0' > set_ftrace_filter
3340 3118
3341 The above removes the traceoff command for 3119 The above removes the traceoff command for __schedule_bug
3342 that have a counter. To remove commands wit 3120 that have a counter. To remove commands without counters::
3343 3121
3344 echo '!__schedule_bug:traceoff' > set_ftra 3122 echo '!__schedule_bug:traceoff' > set_ftrace_filter
3345 3123
3346 - snapshot: 3124 - snapshot:
3347 Will cause a snapshot to be triggered when 3125 Will cause a snapshot to be triggered when the function is hit.
3348 :: 3126 ::
3349 3127
3350 echo 'native_flush_tlb_others:snapshot' > 3128 echo 'native_flush_tlb_others:snapshot' > set_ftrace_filter
3351 3129
3352 To only snapshot once: 3130 To only snapshot once:
3353 :: 3131 ::
3354 3132
3355 echo 'native_flush_tlb_others:snapshot:1' 3133 echo 'native_flush_tlb_others:snapshot:1' > set_ftrace_filter
3356 3134
3357 To remove the above commands:: 3135 To remove the above commands::
3358 3136
3359 echo '!native_flush_tlb_others:snapshot' > 3137 echo '!native_flush_tlb_others:snapshot' > set_ftrace_filter
3360 echo '!native_flush_tlb_others:snapshot:0' 3138 echo '!native_flush_tlb_others:snapshot:0' > set_ftrace_filter
3361 3139
3362 - enable_event/disable_event: 3140 - enable_event/disable_event:
3363 These commands can enable or disable a trac 3141 These commands can enable or disable a trace event. Note, because
3364 function tracing callbacks are very sensiti 3142 function tracing callbacks are very sensitive, when these commands
3365 are registered, the trace point is activate 3143 are registered, the trace point is activated, but disabled in
3366 a "soft" mode. That is, the tracepoint will 3144 a "soft" mode. That is, the tracepoint will be called, but
3367 just will not be traced. The event tracepoi 3145 just will not be traced. The event tracepoint stays in this mode
3368 as long as there's a command that triggers 3146 as long as there's a command that triggers it.
3369 :: 3147 ::
3370 3148
3371 echo 'try_to_wake_up:enable_event:sched:sc 3149 echo 'try_to_wake_up:enable_event:sched:sched_switch:2' > \
3372 set_ftrace_filter 3150 set_ftrace_filter
3373 3151
3374 The format is:: 3152 The format is::
3375 3153
3376 <function>:enable_event:<system>:<event>[ 3154 <function>:enable_event:<system>:<event>[:count]
3377 <function>:disable_event:<system>:<event> 3155 <function>:disable_event:<system>:<event>[:count]
3378 3156
3379 To remove the events commands:: 3157 To remove the events commands::
3380 3158
3381 echo '!try_to_wake_up:enable_event:sched:s 3159 echo '!try_to_wake_up:enable_event:sched:sched_switch:0' > \
3382 set_ftrace_filter 3160 set_ftrace_filter
3383 echo '!schedule:disable_event:sched:sched_ 3161 echo '!schedule:disable_event:sched:sched_switch' > \
3384 set_ftrace_filter 3162 set_ftrace_filter
3385 3163
3386 - dump: 3164 - dump:
3387 When the function is hit, it will dump the 3165 When the function is hit, it will dump the contents of the ftrace
3388 ring buffer to the console. This is useful 3166 ring buffer to the console. This is useful if you need to debug
3389 something, and want to dump the trace when 3167 something, and want to dump the trace when a certain function
3390 is hit. Perhaps it's a function that is cal 3168 is hit. Perhaps it's a function that is called before a triple
3391 fault happens and does not allow you to get 3169 fault happens and does not allow you to get a regular dump.
3392 3170
3393 - cpudump: 3171 - cpudump:
3394 When the function is hit, it will dump the 3172 When the function is hit, it will dump the contents of the ftrace
3395 ring buffer for the current CPU to the cons 3173 ring buffer for the current CPU to the console. Unlike the "dump"
3396 command, it only prints out the contents of 3174 command, it only prints out the contents of the ring buffer for the
3397 CPU that executed the function that trigger 3175 CPU that executed the function that triggered the dump.
3398 3176
3399 - stacktrace: 3177 - stacktrace:
3400 When the function is hit, a stack trace is 3178 When the function is hit, a stack trace is recorded.
3401 3179
3402 trace_pipe 3180 trace_pipe
3403 ---------- 3181 ----------
3404 3182
3405 The trace_pipe outputs the same content as th 3183 The trace_pipe outputs the same content as the trace file, but
3406 the effect on the tracing is different. Every 3184 the effect on the tracing is different. Every read from
3407 trace_pipe is consumed. This means that subse 3185 trace_pipe is consumed. This means that subsequent reads will be
3408 different. The trace is live. 3186 different. The trace is live.
3409 :: 3187 ::
3410 3188
3411 # echo function > current_tracer 3189 # echo function > current_tracer
3412 # cat trace_pipe > /tmp/trace.out & 3190 # cat trace_pipe > /tmp/trace.out &
3413 [1] 4153 3191 [1] 4153
3414 # echo 1 > tracing_on 3192 # echo 1 > tracing_on
3415 # usleep 1 3193 # usleep 1
3416 # echo 0 > tracing_on 3194 # echo 0 > tracing_on
3417 # cat trace 3195 # cat trace
3418 # tracer: function 3196 # tracer: function
3419 # 3197 #
3420 # entries-in-buffer/entries-written: 0/0 3198 # entries-in-buffer/entries-written: 0/0 #P:4
3421 # 3199 #
3422 # _-----=> irq 3200 # _-----=> irqs-off
3423 # / _----=> nee 3201 # / _----=> need-resched
3424 # | / _---=> har 3202 # | / _---=> hardirq/softirq
3425 # || / _--=> pre 3203 # || / _--=> preempt-depth
3426 # ||| / dela 3204 # ||| / delay
3427 # TASK-PID CPU# |||| TIMEST 3205 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3428 # | | | |||| | 3206 # | | | |||| | |
3429 3207
3430 # 3208 #
3431 # cat /tmp/trace.out 3209 # cat /tmp/trace.out
3432 bash-1994 [000] .... 5281.5689 3210 bash-1994 [000] .... 5281.568961: mutex_unlock <-rb_simple_write
3433 bash-1994 [000] .... 5281.5689 3211 bash-1994 [000] .... 5281.568963: __mutex_unlock_slowpath <-mutex_unlock
3434 bash-1994 [000] .... 5281.5689 3212 bash-1994 [000] .... 5281.568963: __fsnotify_parent <-fsnotify_modify
3435 bash-1994 [000] .... 5281.5689 3213 bash-1994 [000] .... 5281.568964: fsnotify <-fsnotify_modify
3436 bash-1994 [000] .... 5281.5689 3214 bash-1994 [000] .... 5281.568964: __srcu_read_lock <-fsnotify
3437 bash-1994 [000] .... 5281.5689 3215 bash-1994 [000] .... 5281.568964: add_preempt_count <-__srcu_read_lock
3438 bash-1994 [000] ...1 5281.5689 3216 bash-1994 [000] ...1 5281.568965: sub_preempt_count <-__srcu_read_lock
3439 bash-1994 [000] .... 5281.5689 3217 bash-1994 [000] .... 5281.568965: __srcu_read_unlock <-fsnotify
3440 bash-1994 [000] .... 5281.5689 3218 bash-1994 [000] .... 5281.568967: sys_dup2 <-system_call_fastpath
3441 3219
3442 3220
3443 Note, reading the trace_pipe file will block 3221 Note, reading the trace_pipe file will block until more input is
3444 added. This is contrary to the trace file. If 3222 added. This is contrary to the trace file. If any process opened
3445 the trace file for reading, it will actually 3223 the trace file for reading, it will actually disable tracing and
3446 prevent new entries from being added. The tra 3224 prevent new entries from being added. The trace_pipe file does
3447 not have this limitation. 3225 not have this limitation.
3448 3226
3449 trace entries 3227 trace entries
3450 ------------- 3228 -------------
3451 3229
3452 Having too much or not enough data can be tro 3230 Having too much or not enough data can be troublesome in
3453 diagnosing an issue in the kernel. The file b 3231 diagnosing an issue in the kernel. The file buffer_size_kb is
3454 used to modify the size of the internal trace 3232 used to modify the size of the internal trace buffers. The
3455 number listed is the number of entries that c 3233 number listed is the number of entries that can be recorded per
3456 CPU. To know the full size, multiply the numb 3234 CPU. To know the full size, multiply the number of possible CPUs
3457 with the number of entries. 3235 with the number of entries.
3458 :: 3236 ::
3459 3237
3460 # cat buffer_size_kb 3238 # cat buffer_size_kb
3461 1408 (units kilobytes) 3239 1408 (units kilobytes)
3462 3240
3463 Or simply read buffer_total_size_kb 3241 Or simply read buffer_total_size_kb
3464 :: 3242 ::
3465 3243
3466 # cat buffer_total_size_kb 3244 # cat buffer_total_size_kb
3467 5632 3245 5632
3468 3246
3469 To modify the buffer, simple echo in a number 3247 To modify the buffer, simple echo in a number (in 1024 byte segments).
3470 :: 3248 ::
3471 3249
3472 # echo 10000 > buffer_size_kb 3250 # echo 10000 > buffer_size_kb
3473 # cat buffer_size_kb 3251 # cat buffer_size_kb
3474 10000 (units kilobytes) 3252 10000 (units kilobytes)
3475 3253
3476 It will try to allocate as much as possible. 3254 It will try to allocate as much as possible. If you allocate too
3477 much, it can cause Out-Of-Memory to trigger. 3255 much, it can cause Out-Of-Memory to trigger.
3478 :: 3256 ::
3479 3257
3480 # echo 1000000000000 > buffer_size_kb 3258 # echo 1000000000000 > buffer_size_kb
3481 -bash: echo: write error: Cannot allocate m 3259 -bash: echo: write error: Cannot allocate memory
3482 # cat buffer_size_kb 3260 # cat buffer_size_kb
3483 85 3261 85
3484 3262
3485 The per_cpu buffers can be changed individual 3263 The per_cpu buffers can be changed individually as well:
3486 :: 3264 ::
3487 3265
3488 # echo 10000 > per_cpu/cpu0/buffer_size_kb 3266 # echo 10000 > per_cpu/cpu0/buffer_size_kb
3489 # echo 100 > per_cpu/cpu1/buffer_size_kb 3267 # echo 100 > per_cpu/cpu1/buffer_size_kb
3490 3268
3491 When the per_cpu buffers are not the same, th 3269 When the per_cpu buffers are not the same, the buffer_size_kb
3492 at the top level will just show an X 3270 at the top level will just show an X
3493 :: 3271 ::
3494 3272
3495 # cat buffer_size_kb 3273 # cat buffer_size_kb
3496 X 3274 X
3497 3275
3498 This is where the buffer_total_size_kb is use 3276 This is where the buffer_total_size_kb is useful:
3499 :: 3277 ::
3500 3278
3501 # cat buffer_total_size_kb 3279 # cat buffer_total_size_kb
3502 12916 3280 12916
3503 3281
3504 Writing to the top level buffer_size_kb will 3282 Writing to the top level buffer_size_kb will reset all the buffers
3505 to be the same again. 3283 to be the same again.
3506 3284
3507 Snapshot 3285 Snapshot
3508 -------- 3286 --------
3509 CONFIG_TRACER_SNAPSHOT makes a generic snapsh 3287 CONFIG_TRACER_SNAPSHOT makes a generic snapshot feature
3510 available to all non latency tracers. (Latenc 3288 available to all non latency tracers. (Latency tracers which
3511 record max latency, such as "irqsoff" or "wak 3289 record max latency, such as "irqsoff" or "wakeup", can't use
3512 this feature, since those are already using t 3290 this feature, since those are already using the snapshot
3513 mechanism internally.) 3291 mechanism internally.)
3514 3292
3515 Snapshot preserves a current trace buffer at 3293 Snapshot preserves a current trace buffer at a particular point
3516 in time without stopping tracing. Ftrace swap 3294 in time without stopping tracing. Ftrace swaps the current
3517 buffer with a spare buffer, and tracing conti 3295 buffer with a spare buffer, and tracing continues in the new
3518 current (=previous spare) buffer. 3296 current (=previous spare) buffer.
3519 3297
3520 The following tracefs files in "tracing" are 3298 The following tracefs files in "tracing" are related to this
3521 feature: 3299 feature:
3522 3300
3523 snapshot: 3301 snapshot:
3524 3302
3525 This is used to take a snapshot and t 3303 This is used to take a snapshot and to read the output
3526 of the snapshot. Echo 1 into this fil 3304 of the snapshot. Echo 1 into this file to allocate a
3527 spare buffer and to take a snapshot ( 3305 spare buffer and to take a snapshot (swap), then read
3528 the snapshot from this file in the sa 3306 the snapshot from this file in the same format as
3529 "trace" (described above in the secti 3307 "trace" (described above in the section "The File
3530 System"). Both reads snapshot and tra 3308 System"). Both reads snapshot and tracing are executable
3531 in parallel. When the spare buffer is 3309 in parallel. When the spare buffer is allocated, echoing
3532 0 frees it, and echoing else (positiv 3310 0 frees it, and echoing else (positive) values clear the
3533 snapshot contents. 3311 snapshot contents.
3534 More details are shown in the table b 3312 More details are shown in the table below.
3535 3313
3536 +--------------+------------+-------- 3314 +--------------+------------+------------+------------+
3537 |status\\input | 0 | 1 3315 |status\\input | 0 | 1 | else |
3538 +==============+============+======== 3316 +==============+============+============+============+
3539 |not allocated |(do nothing)| alloc+s 3317 |not allocated |(do nothing)| alloc+swap |(do nothing)|
3540 +--------------+------------+-------- 3318 +--------------+------------+------------+------------+
3541 |allocated | free | swap 3319 |allocated | free | swap | clear |
3542 +--------------+------------+-------- 3320 +--------------+------------+------------+------------+
3543 3321
3544 Here is an example of using the snapshot feat 3322 Here is an example of using the snapshot feature.
3545 :: 3323 ::
3546 3324
3547 # echo 1 > events/sched/enable 3325 # echo 1 > events/sched/enable
3548 # echo 1 > snapshot 3326 # echo 1 > snapshot
3549 # cat snapshot 3327 # cat snapshot
3550 # tracer: nop 3328 # tracer: nop
3551 # 3329 #
3552 # entries-in-buffer/entries-written: 71/71 3330 # entries-in-buffer/entries-written: 71/71 #P:8
3553 # 3331 #
3554 # _-----=> irq 3332 # _-----=> irqs-off
3555 # / _----=> nee 3333 # / _----=> need-resched
3556 # | / _---=> har 3334 # | / _---=> hardirq/softirq
3557 # || / _--=> pre 3335 # || / _--=> preempt-depth
3558 # ||| / dela 3336 # ||| / delay
3559 # TASK-PID CPU# |||| TIMEST 3337 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3560 # | | | |||| | 3338 # | | | |||| | |
3561 <idle>-0 [005] d... 2440.603 3339 <idle>-0 [005] d... 2440.603828: sched_switch: prev_comm=swapper/5 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2242 next_prio=120
3562 sleep-2242 [005] d... 2440.603 3340 sleep-2242 [005] d... 2440.603846: sched_switch: prev_comm=snapshot-test-2 prev_pid=2242 prev_prio=120 prev_state=R ==> next_comm=kworker/5:1 next_pid=60 next_prio=120
3563 [...] 3341 [...]
3564 <idle>-0 [002] d... 2440.70723 3342 <idle>-0 [002] d... 2440.707230: sched_switch: prev_comm=swapper/2 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2229 next_prio=120
3565 3343
3566 # cat trace 3344 # cat trace
3567 # tracer: nop 3345 # tracer: nop
3568 # 3346 #
3569 # entries-in-buffer/entries-written: 77/77 3347 # entries-in-buffer/entries-written: 77/77 #P:8
3570 # 3348 #
3571 # _-----=> irq 3349 # _-----=> irqs-off
3572 # / _----=> nee 3350 # / _----=> need-resched
3573 # | / _---=> har 3351 # | / _---=> hardirq/softirq
3574 # || / _--=> pre 3352 # || / _--=> preempt-depth
3575 # ||| / dela 3353 # ||| / delay
3576 # TASK-PID CPU# |||| TIMEST 3354 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3577 # | | | |||| | 3355 # | | | |||| | |
3578 <idle>-0 [007] d... 2440.707 3356 <idle>-0 [007] d... 2440.707395: sched_switch: prev_comm=swapper/7 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2243 next_prio=120
3579 snapshot-test-2-2229 [002] d... 2440.707 3357 snapshot-test-2-2229 [002] d... 2440.707438: sched_switch: prev_comm=snapshot-test-2 prev_pid=2229 prev_prio=120 prev_state=S ==> next_comm=swapper/2 next_pid=0 next_prio=120
3580 [...] 3358 [...]
3581 3359
3582 3360
3583 If you try to use this snapshot feature when 3361 If you try to use this snapshot feature when current tracer is
3584 one of the latency tracers, you will get the 3362 one of the latency tracers, you will get the following results.
3585 :: 3363 ::
3586 3364
3587 # echo wakeup > current_tracer 3365 # echo wakeup > current_tracer
3588 # echo 1 > snapshot 3366 # echo 1 > snapshot
3589 bash: echo: write error: Device or resource 3367 bash: echo: write error: Device or resource busy
3590 # cat snapshot 3368 # cat snapshot
3591 cat: snapshot: Device or resource busy 3369 cat: snapshot: Device or resource busy
3592 3370
3593 3371
3594 Instances 3372 Instances
3595 --------- 3373 ---------
3596 In the tracefs tracing directory, there is a !! 3374 In the tracefs tracing directory is a directory called "instances".
3597 This directory can have new directories creat 3375 This directory can have new directories created inside of it using
3598 mkdir, and removing directories with rmdir. T 3376 mkdir, and removing directories with rmdir. The directory created
3599 with mkdir in this directory will already con 3377 with mkdir in this directory will already contain files and other
3600 directories after it is created. 3378 directories after it is created.
3601 :: 3379 ::
3602 3380
3603 # mkdir instances/foo 3381 # mkdir instances/foo
3604 # ls instances/foo 3382 # ls instances/foo
3605 buffer_size_kb buffer_total_size_kb event 3383 buffer_size_kb buffer_total_size_kb events free_buffer per_cpu
3606 set_event snapshot trace trace_clock tr 3384 set_event snapshot trace trace_clock trace_marker trace_options
3607 trace_pipe tracing_on 3385 trace_pipe tracing_on
3608 3386
3609 As you can see, the new directory looks simil 3387 As you can see, the new directory looks similar to the tracing directory
3610 itself. In fact, it is very similar, except t 3388 itself. In fact, it is very similar, except that the buffer and
3611 events are agnostic from the main directory, 3389 events are agnostic from the main directory, or from any other
3612 instances that are created. 3390 instances that are created.
3613 3391
3614 The files in the new directory work just like 3392 The files in the new directory work just like the files with the
3615 same name in the tracing directory except the 3393 same name in the tracing directory except the buffer that is used
3616 is a separate and new buffer. The files affec 3394 is a separate and new buffer. The files affect that buffer but do not
3617 affect the main buffer with the exception of 3395 affect the main buffer with the exception of trace_options. Currently,
3618 the trace_options affect all instances and th 3396 the trace_options affect all instances and the top level buffer
3619 the same, but this may change in future relea 3397 the same, but this may change in future releases. That is, options
3620 may become specific to the instance they resi 3398 may become specific to the instance they reside in.
3621 3399
3622 Notice that none of the function tracer files 3400 Notice that none of the function tracer files are there, nor is
3623 current_tracer and available_tracers. This is 3401 current_tracer and available_tracers. This is because the buffers
3624 can currently only have events enabled for th 3402 can currently only have events enabled for them.
3625 :: 3403 ::
3626 3404
3627 # mkdir instances/foo 3405 # mkdir instances/foo
3628 # mkdir instances/bar 3406 # mkdir instances/bar
3629 # mkdir instances/zoot 3407 # mkdir instances/zoot
3630 # echo 100000 > buffer_size_kb 3408 # echo 100000 > buffer_size_kb
3631 # echo 1000 > instances/foo/buffer_size_kb 3409 # echo 1000 > instances/foo/buffer_size_kb
3632 # echo 5000 > instances/bar/per_cpu/cpu1/bu 3410 # echo 5000 > instances/bar/per_cpu/cpu1/buffer_size_kb
3633 # echo function > current_trace 3411 # echo function > current_trace
3634 # echo 1 > instances/foo/events/sched/sched 3412 # echo 1 > instances/foo/events/sched/sched_wakeup/enable
3635 # echo 1 > instances/foo/events/sched/sched 3413 # echo 1 > instances/foo/events/sched/sched_wakeup_new/enable
3636 # echo 1 > instances/foo/events/sched/sched 3414 # echo 1 > instances/foo/events/sched/sched_switch/enable
3637 # echo 1 > instances/bar/events/irq/enable 3415 # echo 1 > instances/bar/events/irq/enable
3638 # echo 1 > instances/zoot/events/syscalls/e 3416 # echo 1 > instances/zoot/events/syscalls/enable
3639 # cat trace_pipe 3417 # cat trace_pipe
3640 CPU:2 [LOST 11745 EVENTS] 3418 CPU:2 [LOST 11745 EVENTS]
3641 bash-2044 [002] .... 10594.481 3419 bash-2044 [002] .... 10594.481032: _raw_spin_lock_irqsave <-get_page_from_freelist
3642 bash-2044 [002] d... 10594.481 3420 bash-2044 [002] d... 10594.481032: add_preempt_count <-_raw_spin_lock_irqsave
3643 bash-2044 [002] d..1 10594.481 3421 bash-2044 [002] d..1 10594.481032: __rmqueue <-get_page_from_freelist
3644 bash-2044 [002] d..1 10594.481 3422 bash-2044 [002] d..1 10594.481033: _raw_spin_unlock <-get_page_from_freelist
3645 bash-2044 [002] d..1 10594.481 3423 bash-2044 [002] d..1 10594.481033: sub_preempt_count <-_raw_spin_unlock
3646 bash-2044 [002] d... 10594.481 3424 bash-2044 [002] d... 10594.481033: get_pageblock_flags_group <-get_pageblock_migratetype
3647 bash-2044 [002] d... 10594.481 3425 bash-2044 [002] d... 10594.481034: __mod_zone_page_state <-get_page_from_freelist
3648 bash-2044 [002] d... 10594.481 3426 bash-2044 [002] d... 10594.481034: zone_statistics <-get_page_from_freelist
3649 bash-2044 [002] d... 10594.481 3427 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
3650 bash-2044 [002] d... 10594.481 3428 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
3651 bash-2044 [002] .... 10594.481 3429 bash-2044 [002] .... 10594.481035: arch_dup_task_struct <-copy_process
3652 [...] 3430 [...]
3653 3431
3654 # cat instances/foo/trace_pipe 3432 # cat instances/foo/trace_pipe
3655 bash-1998 [000] d..4 136.676 3433 bash-1998 [000] d..4 136.676759: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000
3656 bash-1998 [000] dN.4 136.676 3434 bash-1998 [000] dN.4 136.676760: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000
3657 <idle>-0 [003] d.h3 136.676 3435 <idle>-0 [003] d.h3 136.676906: sched_wakeup: comm=rcu_preempt pid=9 prio=120 success=1 target_cpu=003
3658 <idle>-0 [003] d..3 136.676 3436 <idle>-0 [003] d..3 136.676909: sched_switch: prev_comm=swapper/3 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=rcu_preempt next_pid=9 next_prio=120
3659 rcu_preempt-9 [003] d..3 136.676 3437 rcu_preempt-9 [003] d..3 136.676916: sched_switch: prev_comm=rcu_preempt prev_pid=9 prev_prio=120 prev_state=S ==> next_comm=swapper/3 next_pid=0 next_prio=120
3660 bash-1998 [000] d..4 136.677 3438 bash-1998 [000] d..4 136.677014: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000
3661 bash-1998 [000] dN.4 136.677 3439 bash-1998 [000] dN.4 136.677016: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000
3662 bash-1998 [000] d..3 136.677 3440 bash-1998 [000] d..3 136.677018: sched_switch: prev_comm=bash prev_pid=1998 prev_prio=120 prev_state=R+ ==> next_comm=kworker/0:1 next_pid=59 next_prio=120
3663 kworker/0:1-59 [000] d..4 136.677 3441 kworker/0:1-59 [000] d..4 136.677022: sched_wakeup: comm=sshd pid=1995 prio=120 success=1 target_cpu=001
3664 kworker/0:1-59 [000] d..3 136.677 3442 kworker/0:1-59 [000] d..3 136.677025: sched_switch: prev_comm=kworker/0:1 prev_pid=59 prev_prio=120 prev_state=S ==> next_comm=bash next_pid=1998 next_prio=120
3665 [...] 3443 [...]
3666 3444
3667 # cat instances/bar/trace_pipe 3445 # cat instances/bar/trace_pipe
3668 migration/1-14 [001] d.h3 138.732 3446 migration/1-14 [001] d.h3 138.732674: softirq_raise: vec=3 [action=NET_RX]
3669 <idle>-0 [001] dNh3 138.732 3447 <idle>-0 [001] dNh3 138.732725: softirq_raise: vec=3 [action=NET_RX]
3670 bash-1998 [000] d.h1 138.733 3448 bash-1998 [000] d.h1 138.733101: softirq_raise: vec=1 [action=TIMER]
3671 bash-1998 [000] d.h1 138.733 3449 bash-1998 [000] d.h1 138.733102: softirq_raise: vec=9 [action=RCU]
3672 bash-1998 [000] ..s2 138.733 3450 bash-1998 [000] ..s2 138.733105: softirq_entry: vec=1 [action=TIMER]
3673 bash-1998 [000] ..s2 138.733 3451 bash-1998 [000] ..s2 138.733106: softirq_exit: vec=1 [action=TIMER]
3674 bash-1998 [000] ..s2 138.733 3452 bash-1998 [000] ..s2 138.733106: softirq_entry: vec=9 [action=RCU]
3675 bash-1998 [000] ..s2 138.733 3453 bash-1998 [000] ..s2 138.733109: softirq_exit: vec=9 [action=RCU]
3676 sshd-1995 [001] d.h1 138.733 3454 sshd-1995 [001] d.h1 138.733278: irq_handler_entry: irq=21 name=uhci_hcd:usb4
3677 sshd-1995 [001] d.h1 138.733 3455 sshd-1995 [001] d.h1 138.733280: irq_handler_exit: irq=21 ret=unhandled
3678 sshd-1995 [001] d.h1 138.733 3456 sshd-1995 [001] d.h1 138.733281: irq_handler_entry: irq=21 name=eth0
3679 sshd-1995 [001] d.h1 138.733 3457 sshd-1995 [001] d.h1 138.733283: irq_handler_exit: irq=21 ret=handled
3680 [...] 3458 [...]
3681 3459
3682 # cat instances/zoot/trace 3460 # cat instances/zoot/trace
3683 # tracer: nop 3461 # tracer: nop
3684 # 3462 #
3685 # entries-in-buffer/entries-written: 18996/ 3463 # entries-in-buffer/entries-written: 18996/18996 #P:4
3686 # 3464 #
3687 # _-----=> irq 3465 # _-----=> irqs-off
3688 # / _----=> nee 3466 # / _----=> need-resched
3689 # | / _---=> har 3467 # | / _---=> hardirq/softirq
3690 # || / _--=> pre 3468 # || / _--=> preempt-depth
3691 # ||| / dela 3469 # ||| / delay
3692 # TASK-PID CPU# |||| TIMEST 3470 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3693 # | | | |||| | 3471 # | | | |||| | |
3694 bash-1998 [000] d... 140.733 3472 bash-1998 [000] d... 140.733501: sys_write -> 0x2
3695 bash-1998 [000] d... 140.733 3473 bash-1998 [000] d... 140.733504: sys_dup2(oldfd: a, newfd: 1)
3696 bash-1998 [000] d... 140.733 3474 bash-1998 [000] d... 140.733506: sys_dup2 -> 0x1
3697 bash-1998 [000] d... 140.733 3475 bash-1998 [000] d... 140.733508: sys_fcntl(fd: a, cmd: 1, arg: 0)
3698 bash-1998 [000] d... 140.733 3476 bash-1998 [000] d... 140.733509: sys_fcntl -> 0x1
3699 bash-1998 [000] d... 140.733 3477 bash-1998 [000] d... 140.733510: sys_close(fd: a)
3700 bash-1998 [000] d... 140.733 3478 bash-1998 [000] d... 140.733510: sys_close -> 0x0
3701 bash-1998 [000] d... 140.733 3479 bash-1998 [000] d... 140.733514: sys_rt_sigprocmask(how: 0, nset: 0, oset: 6e2768, sigsetsize: 8)
3702 bash-1998 [000] d... 140.733 3480 bash-1998 [000] d... 140.733515: sys_rt_sigprocmask -> 0x0
3703 bash-1998 [000] d... 140.733 3481 bash-1998 [000] d... 140.733516: sys_rt_sigaction(sig: 2, act: 7fff718846f0, oact: 7fff71884650, sigsetsize: 8)
3704 bash-1998 [000] d... 140.733 3482 bash-1998 [000] d... 140.733516: sys_rt_sigaction -> 0x0
3705 3483
3706 You can see that the trace of the top most tr 3484 You can see that the trace of the top most trace buffer shows only
3707 the function tracing. The foo instance displa 3485 the function tracing. The foo instance displays wakeups and task
3708 switches. 3486 switches.
3709 3487
3710 To remove the instances, simply delete their 3488 To remove the instances, simply delete their directories:
3711 :: 3489 ::
3712 3490
3713 # rmdir instances/foo 3491 # rmdir instances/foo
3714 # rmdir instances/bar 3492 # rmdir instances/bar
3715 # rmdir instances/zoot 3493 # rmdir instances/zoot
3716 3494
3717 Note, if a process has a trace file open in o 3495 Note, if a process has a trace file open in one of the instance
3718 directories, the rmdir will fail with EBUSY. 3496 directories, the rmdir will fail with EBUSY.
3719 3497
3720 3498
3721 Stack trace 3499 Stack trace
3722 ----------- 3500 -----------
3723 Since the kernel has a fixed sized stack, it 3501 Since the kernel has a fixed sized stack, it is important not to
3724 waste it in functions. A kernel developer mus !! 3502 waste it in functions. A kernel developer must be conscience of
3725 what they allocate on the stack. If they add 3503 what they allocate on the stack. If they add too much, the system
3726 can be in danger of a stack overflow, and cor 3504 can be in danger of a stack overflow, and corruption will occur,
3727 usually leading to a system panic. 3505 usually leading to a system panic.
3728 3506
3729 There are some tools that check this, usually 3507 There are some tools that check this, usually with interrupts
3730 periodically checking usage. But if you can p 3508 periodically checking usage. But if you can perform a check
3731 at every function call that will become very 3509 at every function call that will become very useful. As ftrace provides
3732 a function tracer, it makes it convenient to 3510 a function tracer, it makes it convenient to check the stack size
3733 at every function call. This is enabled via t 3511 at every function call. This is enabled via the stack tracer.
3734 3512
3735 CONFIG_STACK_TRACER enables the ftrace stack 3513 CONFIG_STACK_TRACER enables the ftrace stack tracing functionality.
3736 To enable it, write a '1' into /proc/sys/kern 3514 To enable it, write a '1' into /proc/sys/kernel/stack_tracer_enabled.
3737 :: 3515 ::
3738 3516
3739 # echo 1 > /proc/sys/kernel/stack_tracer_ena 3517 # echo 1 > /proc/sys/kernel/stack_tracer_enabled
3740 3518
3741 You can also enable it from the kernel comman 3519 You can also enable it from the kernel command line to trace
3742 the stack size of the kernel during boot up, 3520 the stack size of the kernel during boot up, by adding "stacktrace"
3743 to the kernel command line parameter. 3521 to the kernel command line parameter.
3744 3522
3745 After running it for a few minutes, the outpu 3523 After running it for a few minutes, the output looks like:
3746 :: 3524 ::
3747 3525
3748 # cat stack_max_size 3526 # cat stack_max_size
3749 2928 3527 2928
3750 3528
3751 # cat stack_trace 3529 # cat stack_trace
3752 Depth Size Location (18 ent 3530 Depth Size Location (18 entries)
3753 ----- ---- -------- 3531 ----- ---- --------
3754 0) 2928 224 update_sd_lb_stats+ 3532 0) 2928 224 update_sd_lb_stats+0xbc/0x4ac
3755 1) 2704 160 find_busiest_group+ 3533 1) 2704 160 find_busiest_group+0x31/0x1f1
3756 2) 2544 256 load_balance+0xd9/0 3534 2) 2544 256 load_balance+0xd9/0x662
3757 3) 2288 80 idle_balance+0xbb/0 3535 3) 2288 80 idle_balance+0xbb/0x130
3758 4) 2208 128 __schedule+0x26e/0x 3536 4) 2208 128 __schedule+0x26e/0x5b9
3759 5) 2080 16 schedule+0x64/0x66 3537 5) 2080 16 schedule+0x64/0x66
3760 6) 2064 128 schedule_timeout+0x 3538 6) 2064 128 schedule_timeout+0x34/0xe0
3761 7) 1936 112 wait_for_common+0x9 3539 7) 1936 112 wait_for_common+0x97/0xf1
3762 8) 1824 16 wait_for_completion 3540 8) 1824 16 wait_for_completion+0x1d/0x1f
3763 9) 1808 128 flush_work+0xfe/0x1 3541 9) 1808 128 flush_work+0xfe/0x119
3764 10) 1680 16 tty_flush_to_ldisc+ 3542 10) 1680 16 tty_flush_to_ldisc+0x1e/0x20
3765 11) 1664 48 input_available_p+0 3543 11) 1664 48 input_available_p+0x1d/0x5c
3766 12) 1616 48 n_tty_poll+0x6d/0x1 3544 12) 1616 48 n_tty_poll+0x6d/0x134
3767 13) 1568 64 tty_poll+0x64/0x7f 3545 13) 1568 64 tty_poll+0x64/0x7f
3768 14) 1504 880 do_select+0x31e/0x5 3546 14) 1504 880 do_select+0x31e/0x511
3769 15) 624 400 core_sys_select+0x1 3547 15) 624 400 core_sys_select+0x177/0x216
3770 16) 224 96 sys_select+0x91/0xb 3548 16) 224 96 sys_select+0x91/0xb9
3771 17) 128 128 system_call_fastpat 3549 17) 128 128 system_call_fastpath+0x16/0x1b
3772 3550
3773 Note, if -mfentry is being used by gcc, funct 3551 Note, if -mfentry is being used by gcc, functions get traced before
3774 they set up the stack frame. This means that 3552 they set up the stack frame. This means that leaf level functions
3775 are not tested by the stack tracer when -mfen 3553 are not tested by the stack tracer when -mfentry is used.
3776 3554
3777 Currently, -mfentry is used by gcc 4.6.0 and 3555 Currently, -mfentry is used by gcc 4.6.0 and above on x86 only.
3778 3556
3779 More 3557 More
3780 ---- 3558 ----
3781 More details can be found in the source code, 3559 More details can be found in the source code, in the `kernel/trace/*.c` files.
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.