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.txt 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.
99 the ring buffer content as well as the <<
100 99
101 available_tracers: 100 available_tracers:
102 101
103 This holds the different types of trac 102 This holds the different types of tracers that
104 have been compiled into the kernel. Th 103 have been compiled into the kernel. The
105 tracers listed here can be configured 104 tracers listed here can be configured by
106 echoing their name into current_tracer 105 echoing their name into current_tracer.
107 106
108 tracing_on: 107 tracing_on:
109 108
110 This sets or displays whether writing 109 This sets or displays whether writing to the trace
111 ring buffer is enabled. Echo 0 into th 110 ring buffer is enabled. Echo 0 into this file to disable
112 the tracer or 1 to enable it. Note, th 111 the tracer or 1 to enable it. Note, this only disables
113 writing to the ring buffer, the tracin 112 writing to the ring buffer, the tracing overhead may
114 still be occurring. 113 still be occurring.
115 114
116 The kernel function tracing_off() can 115 The kernel function tracing_off() can be used within the
117 kernel to disable writing to the ring 116 kernel to disable writing to the ring buffer, which will
118 set this file to "0". User space can r 117 set this file to "0". User space can re-enable tracing by
119 echoing "1" into the file. 118 echoing "1" into the file.
120 119
121 Note, the function and event trigger " 120 Note, the function and event trigger "traceoff" will also
122 set this file to zero and stop tracing 121 set this file to zero and stop tracing. Which can also
123 be re-enabled by user space using this 122 be re-enabled by user space using this file.
124 123
125 trace: 124 trace:
126 125
127 This file holds the output of the trac 126 This file holds the output of the trace in a human
128 readable format (described below). Ope !! 127 readable format (described below). Note, tracing is temporarily
129 writing with the O_TRUNC flag clears t !! 128 disabled when the file is open for reading. Once all readers
130 Note, this file is not a consumer. If !! 129 are closed, tracing is re-enabled.
131 (no tracer running, or tracing_on is z <<
132 the same output each time it is read. <<
133 it may produce inconsistent results as <<
134 the entire buffer without consuming it <<
135 130
136 trace_pipe: 131 trace_pipe:
137 132
138 The output is the same as the "trace" 133 The output is the same as the "trace" file but this
139 file is meant to be streamed with live 134 file is meant to be streamed with live tracing.
140 Reads from this file will block until 135 Reads from this file will block until new data is
141 retrieved. Unlike the "trace" file, t 136 retrieved. Unlike the "trace" file, this file is a
142 consumer. This means reading from this 137 consumer. This means reading from this file causes
143 sequential reads to display more curre 138 sequential reads to display more current data. Once
144 data is read from this file, it is con 139 data is read from this file, it is consumed, and
145 will not be read again with a sequenti 140 will not be read again with a sequential read. The
146 "trace" file is static, and if the tra 141 "trace" file is static, and if the tracer is not
147 adding more data, it will display the 142 adding more data, it will display the same
148 information every time it is read. !! 143 information every time it is read. Unlike the
>> 144 "trace" file, opening this file for reading will not
>> 145 temporarily disable tracing.
149 146
150 trace_options: 147 trace_options:
151 148
152 This file lets the user control the am 149 This file lets the user control the amount of data
153 that is displayed in one of the above 150 that is displayed in one of the above output
154 files. Options also exist to modify ho 151 files. Options also exist to modify how a tracer
155 or events work (stack traces, timestam 152 or events work (stack traces, timestamps, etc).
156 153
157 options: 154 options:
158 155
159 This is a directory that has a file fo 156 This is a directory that has a file for every available
160 trace option (also in trace_options). 157 trace option (also in trace_options). Options may also be set
161 or cleared by writing a "1" or "0" res 158 or cleared by writing a "1" or "0" respectively into the
162 corresponding file with the option nam 159 corresponding file with the option name.
163 160
164 tracing_max_latency: 161 tracing_max_latency:
165 162
166 Some of the tracers record the max lat 163 Some of the tracers record the max latency.
167 For example, the maximum time that int 164 For example, the maximum time that interrupts are disabled.
168 The maximum time is saved in this file 165 The maximum time is saved in this file. The max trace will also be
169 stored, and displayed by "trace". A ne 166 stored, and displayed by "trace". A new max trace will only be
170 recorded if the latency is greater tha 167 recorded if the latency is greater than the value in this file
171 (in microseconds). 168 (in microseconds).
172 169
173 By echoing in a time into this file, n 170 By echoing in a time into this file, no latency will be recorded
174 unless it is greater than the time in 171 unless it is greater than the time in this file.
175 172
176 tracing_thresh: 173 tracing_thresh:
177 174
178 Some latency tracers will record a tra 175 Some latency tracers will record a trace whenever the
179 latency is greater than the number in 176 latency is greater than the number in this file.
180 Only active when the file contains a n 177 Only active when the file contains a number greater than 0.
181 (in microseconds) 178 (in microseconds)
182 179
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: 180 buffer_size_kb:
199 181
200 This sets or displays the number of ki 182 This sets or displays the number of kilobytes each CPU
201 buffer holds. By default, the trace bu 183 buffer holds. By default, the trace buffers are the same size
202 for each CPU. The displayed number is 184 for each CPU. The displayed number is the size of the
203 CPU buffer and not total size of all b 185 CPU buffer and not total size of all buffers. The
204 trace buffers are allocated in pages ( 186 trace buffers are allocated in pages (blocks of memory
205 that the kernel uses for allocation, u 187 that the kernel uses for allocation, usually 4 KB in size).
206 A few extra pages may be allocated to !! 188 If the last page allocated has room for more bytes
207 meta-data. If the last page allocated <<
208 than requested, the rest of the page w 189 than requested, the rest of the page will be used,
209 making the actual allocation bigger th 190 making the actual allocation bigger than requested or shown.
210 ( Note, the size may not be a multiple 191 ( Note, the size may not be a multiple of the page size
211 due to buffer management meta-data. ) 192 due to buffer management meta-data. )
212 193
213 Buffer sizes for individual CPUs may v 194 Buffer sizes for individual CPUs may vary
214 (see "per_cpu/cpu0/buffer_size_kb" bel 195 (see "per_cpu/cpu0/buffer_size_kb" below), and if they do
215 this file will show "X". 196 this file will show "X".
216 197
217 buffer_total_size_kb: 198 buffer_total_size_kb:
218 199
219 This displays the total combined size 200 This displays the total combined size of all the trace buffers.
220 201
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: 202 free_buffer:
243 203
244 If a process is performing tracing, an 204 If a process is performing tracing, and the ring buffer should be
245 shrunk "freed" when the process is fin 205 shrunk "freed" when the process is finished, even if it were to be
246 killed by a signal, this file can be u 206 killed by a signal, this file can be used for that purpose. On close
247 of this file, the ring buffer will be 207 of this file, the ring buffer will be resized to its minimum size.
248 Having a process that is tracing also 208 Having a process that is tracing also open this file, when the process
249 exits its file descriptor for this fil 209 exits its file descriptor for this file will be closed, and in doing so,
250 the ring buffer will be "freed". 210 the ring buffer will be "freed".
251 211
252 It may also stop tracing if disable_on 212 It may also stop tracing if disable_on_free option is set.
253 213
254 tracing_cpumask: 214 tracing_cpumask:
255 215
256 This is a mask that lets the user only 216 This is a mask that lets the user only trace on specified CPUs.
257 The format is a hex string representin 217 The format is a hex string representing the CPUs.
258 218
259 set_ftrace_filter: 219 set_ftrace_filter:
260 220
261 When dynamic ftrace is configured in ( 221 When dynamic ftrace is configured in (see the
262 section below "dynamic ftrace"), the c 222 section below "dynamic ftrace"), the code is dynamically
263 modified (code text rewrite) to disabl 223 modified (code text rewrite) to disable calling of the
264 function profiler (mcount). This lets 224 function profiler (mcount). This lets tracing be configured
265 in with practically no overhead in per 225 in with practically no overhead in performance. This also
266 has a side effect of enabling or disab 226 has a side effect of enabling or disabling specific functions
267 to be traced. Echoing names of functio 227 to be traced. Echoing names of functions into this file
268 will limit the trace to only those fun 228 will limit the trace to only those functions.
269 This influences the tracers "function" 229 This influences the tracers "function" and "function_graph"
270 and thus also function profiling (see 230 and thus also function profiling (see "function_profile_enabled").
271 231
272 The functions listed in "available_fil 232 The functions listed in "available_filter_functions" are what
273 can be written into this file. 233 can be written into this file.
274 234
275 This interface also allows for command 235 This interface also allows for commands to be used. See the
276 "Filter commands" section for more det 236 "Filter commands" section for more details.
277 237
278 As a speed up, since processing string !! 238 As a speed up, since processing strings can't be quite expensive
279 and requires a check of all functions 239 and requires a check of all functions registered to tracing, instead
280 an index can be written into this file 240 an index can be written into this file. A number (starting with "1")
281 written will instead select the same c 241 written will instead select the same corresponding at the line position
282 of the "available_filter_functions" fi 242 of the "available_filter_functions" file.
283 243
284 set_ftrace_notrace: 244 set_ftrace_notrace:
285 245
286 This has an effect opposite to that of 246 This has an effect opposite to that of
287 set_ftrace_filter. Any function that i 247 set_ftrace_filter. Any function that is added here will not
288 be traced. If a function exists in bot 248 be traced. If a function exists in both set_ftrace_filter
289 and set_ftrace_notrace, the function w 249 and set_ftrace_notrace, the function will _not_ be traced.
290 250
291 set_ftrace_pid: 251 set_ftrace_pid:
292 252
293 Have the function tracer only trace th 253 Have the function tracer only trace the threads whose PID are
294 listed in this file. 254 listed in this file.
295 255
296 If the "function-fork" option is set, 256 If the "function-fork" option is set, then when a task whose
297 PID is listed in this file forks, the 257 PID is listed in this file forks, the child's PID will
298 automatically be added to this file, a 258 automatically be added to this file, and the child will be
299 traced by the function tracer as well. 259 traced by the function tracer as well. This option will also
300 cause PIDs of tasks that exit to be re 260 cause PIDs of tasks that exit to be removed from the file.
301 261
302 set_ftrace_notrace_pid: <<
303 <<
304 Have the function tracer ignore thread <<
305 this file. <<
306 <<
307 If the "function-fork" option is set, <<
308 PID is listed in this file forks, the <<
309 automatically be added to this file, a <<
310 traced by the function tracer as well. <<
311 cause PIDs of tasks that exit to be re <<
312 <<
313 If a PID is in both this file and "set <<
314 file takes precedence, and the thread <<
315 <<
316 set_event_pid: 262 set_event_pid:
317 263
318 Have the events only trace a task with 264 Have the events only trace a task with a PID listed in this file.
319 Note, sched_switch and sched_wake_up w 265 Note, sched_switch and sched_wake_up will also trace events
320 listed in this file. 266 listed in this file.
321 267
322 To have the PIDs of children of tasks 268 To have the PIDs of children of tasks with their PID in this file
323 added on fork, enable the "event-fork" 269 added on fork, enable the "event-fork" option. That option will also
324 cause the PIDs of tasks to be removed 270 cause the PIDs of tasks to be removed from this file when the task
325 exits. 271 exits.
326 272
327 set_event_notrace_pid: <<
328 <<
329 Have the events not trace a task with <<
330 Note, sched_switch and sched_wakeup wi <<
331 in this file, even if a thread's PID i <<
332 sched_switch or sched_wakeup events al <<
333 be traced. <<
334 <<
335 To have the PIDs of children of tasks <<
336 added on fork, enable the "event-fork" <<
337 cause the PIDs of tasks to be removed <<
338 exits. <<
339 <<
340 set_graph_function: 273 set_graph_function:
341 274
342 Functions listed in this file will cau 275 Functions listed in this file will cause the function graph
343 tracer to only trace these functions a 276 tracer to only trace these functions and the functions that
344 they call. (See the section "dynamic f 277 they call. (See the section "dynamic ftrace" for more details).
345 Note, set_ftrace_filter and set_ftrace 278 Note, set_ftrace_filter and set_ftrace_notrace still affects
346 what functions are being traced. 279 what functions are being traced.
347 280
348 set_graph_notrace: 281 set_graph_notrace:
349 282
350 Similar to set_graph_function, but wil 283 Similar to set_graph_function, but will disable function graph
351 tracing when the function is hit until 284 tracing when the function is hit until it exits the function.
352 This makes it possible to ignore traci 285 This makes it possible to ignore tracing functions that are called
353 by a specific function. 286 by a specific function.
354 287
355 available_filter_functions: 288 available_filter_functions:
356 289
357 This lists the functions that ftrace h 290 This lists the functions that ftrace has processed and can trace.
358 These are the function names that you 291 These are the function names that you can pass to
359 "set_ftrace_filter", "set_ftrace_notra 292 "set_ftrace_filter", "set_ftrace_notrace",
360 "set_graph_function", or "set_graph_no 293 "set_graph_function", or "set_graph_notrace".
361 (See the section "dynamic ftrace" belo 294 (See the section "dynamic ftrace" below for more details.)
362 295
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: 296 dyn_ftrace_total_info:
370 297
371 This file is for debugging purposes. T 298 This file is for debugging purposes. The number of functions that
372 have been converted to nops and are av 299 have been converted to nops and are available to be traced.
373 300
374 enabled_functions: 301 enabled_functions:
375 302
376 This file is more for debugging ftrace 303 This file is more for debugging ftrace, but can also be useful
377 in seeing if any function has a callba 304 in seeing if any function has a callback attached to it.
378 Not only does the trace infrastructure 305 Not only does the trace infrastructure use ftrace function
379 trace utility, but other subsystems mi 306 trace utility, but other subsystems might too. This file
380 displays all functions that have a cal 307 displays all functions that have a callback attached to them
381 as well as the number of callbacks tha 308 as well as the number of callbacks that have been attached.
382 Note, a callback may also call multipl 309 Note, a callback may also call multiple functions which will
383 not be listed in this count. 310 not be listed in this count.
384 311
385 If the callback registered to be trace 312 If the callback registered to be traced by a function with
386 the "save regs" attribute (thus even m 313 the "save regs" attribute (thus even more overhead), a 'R'
387 will be displayed on the same line as 314 will be displayed on the same line as the function that
388 is returning registers. 315 is returning registers.
389 316
390 If the callback registered to be trace 317 If the callback registered to be traced by a function with
391 the "ip modify" attribute (thus the re 318 the "ip modify" attribute (thus the regs->ip can be changed),
392 an 'I' will be displayed on the same l 319 an 'I' will be displayed on the same line as the function that
393 can be overridden. 320 can be overridden.
394 321
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 322 If the architecture supports it, it will also show what callback
409 is being directly called by the functi 323 is being directly called by the function. If the count is greater
410 than 1 it most likely will be ftrace_o 324 than 1 it most likely will be ftrace_ops_list_func().
411 325
412 If the callback of a function jumps to !! 326 If the callback of the function jumps to a trampoline that is
413 specific to the callback and which is !! 327 specific to a the callback and not the standard trampoline,
414 its address will be printed as well as 328 its address will be printed as well as the function that the
415 trampoline calls. 329 trampoline calls.
416 330
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: 331 function_profile_enabled:
430 332
431 When set it will enable all functions 333 When set it will enable all functions with either the function
432 tracer, or if configured, the function 334 tracer, or if configured, the function graph tracer. It will
433 keep a histogram of the number of func 335 keep a histogram of the number of functions that were called
434 and if the function graph tracer was c 336 and if the function graph tracer was configured, it will also keep
435 track of the time spent in those funct 337 track of the time spent in those functions. The histogram
436 content can be displayed in the files: 338 content can be displayed in the files:
437 339
438 trace_stat/function<cpu> ( function0, 340 trace_stat/function<cpu> ( function0, function1, etc).
439 341
440 trace_stat: 342 trace_stat:
441 343
442 A directory that holds different traci 344 A directory that holds different tracing stats.
443 345
444 kprobe_events: 346 kprobe_events:
445 347
446 Enable dynamic trace points. See kprob !! 348 Enable dynamic trace points. See kprobetrace.txt.
447 349
448 kprobe_profile: 350 kprobe_profile:
449 351
450 Dynamic trace points stats. See kprobe !! 352 Dynamic trace points stats. See kprobetrace.txt.
451 353
452 max_graph_depth: 354 max_graph_depth:
453 355
454 Used with the function graph tracer. T 356 Used with the function graph tracer. This is the max depth
455 it will trace into a function. Setting 357 it will trace into a function. Setting this to a value of
456 one will show only the first kernel fu 358 one will show only the first kernel function that is called
457 from user space. 359 from user space.
458 360
459 printk_formats: 361 printk_formats:
460 362
461 This is for tools that read the raw fo 363 This is for tools that read the raw format files. If an event in
462 the ring buffer references a string, o 364 the ring buffer references a string, only a pointer to the string
463 is recorded into the buffer and not th 365 is recorded into the buffer and not the string itself. This prevents
464 tools from knowing what that string wa 366 tools from knowing what that string was. This file displays the string
465 and address for the string allowing to 367 and address for the string allowing tools to map the pointers to what
466 the strings were. 368 the strings were.
467 369
468 saved_cmdlines: 370 saved_cmdlines:
469 371
470 Only the pid of the task is recorded i 372 Only the pid of the task is recorded in a trace event unless
471 the event specifically saves the task 373 the event specifically saves the task comm as well. Ftrace
472 makes a cache of pid mappings to comms 374 makes a cache of pid mappings to comms to try to display
473 comms for events. If a pid for a comm 375 comms for events. If a pid for a comm is not listed, then
474 "<...>" is displayed in the output. 376 "<...>" is displayed in the output.
475 377
476 If the option "record-cmd" is set to " 378 If the option "record-cmd" is set to "0", then comms of tasks
477 will not be saved during recording. By 379 will not be saved during recording. By default, it is enabled.
478 380
479 saved_cmdlines_size: 381 saved_cmdlines_size:
480 382
481 By default, 128 comms are saved (see " 383 By default, 128 comms are saved (see "saved_cmdlines" above). To
482 increase or decrease the amount of com 384 increase or decrease the amount of comms that are cached, echo
483 the number of comms to cache into this !! 385 in a the number of comms to cache, into this file.
484 386
485 saved_tgids: 387 saved_tgids:
486 388
487 If the option "record-tgid" is set, on 389 If the option "record-tgid" is set, on each scheduling context switch
488 the Task Group ID of a task is saved i 390 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 391 the thread to its TGID. By default, the "record-tgid" option is
490 disabled. 392 disabled.
491 393
492 snapshot: 394 snapshot:
493 395
494 This displays the "snapshot" buffer an 396 This displays the "snapshot" buffer and also lets the user
495 take a snapshot of the current running 397 take a snapshot of the current running trace.
496 See the "Snapshot" section below for m 398 See the "Snapshot" section below for more details.
497 399
498 stack_max_size: 400 stack_max_size:
499 401
500 When the stack tracer is activated, th 402 When the stack tracer is activated, this will display the
501 maximum stack size it has encountered. 403 maximum stack size it has encountered.
502 See the "Stack Trace" section below. 404 See the "Stack Trace" section below.
503 405
504 stack_trace: 406 stack_trace:
505 407
506 This displays the stack back trace of 408 This displays the stack back trace of the largest stack
507 that was encountered when the stack tr 409 that was encountered when the stack tracer is activated.
508 See the "Stack Trace" section below. 410 See the "Stack Trace" section below.
509 411
510 stack_trace_filter: 412 stack_trace_filter:
511 413
512 This is similar to "set_ftrace_filter" 414 This is similar to "set_ftrace_filter" but it limits what
513 functions the stack tracer will check. 415 functions the stack tracer will check.
514 416
515 trace_clock: 417 trace_clock:
516 418
517 Whenever an event is recorded into the 419 Whenever an event is recorded into the ring buffer, a
518 "timestamp" is added. This stamp comes 420 "timestamp" is added. This stamp comes from a specified
519 clock. By default, ftrace uses the "lo 421 clock. By default, ftrace uses the "local" clock. This
520 clock is very fast and strictly per cp 422 clock is very fast and strictly per cpu, but on some
521 systems it may not be monotonic with r 423 systems it may not be monotonic with respect to other
522 CPUs. In other words, the local clocks 424 CPUs. In other words, the local clocks may not be in sync
523 with local clocks on other CPUs. 425 with local clocks on other CPUs.
524 426
525 Usual clocks for tracing:: 427 Usual clocks for tracing::
526 428
527 # cat trace_clock 429 # cat trace_clock
528 [local] global counter x86-tsc 430 [local] global counter x86-tsc
529 431
530 The clock with the square brackets aro 432 The clock with the square brackets around it is the one in effect.
531 433
532 local: 434 local:
533 Default clock, but may not be 435 Default clock, but may not be in sync across CPUs
534 436
535 global: 437 global:
536 This clock is in sync with all 438 This clock is in sync with all CPUs but may
537 be a bit slower than the local 439 be a bit slower than the local clock.
538 440
539 counter: 441 counter:
540 This is not a clock at all, bu 442 This is not a clock at all, but literally an atomic
541 counter. It counts up one by o 443 counter. It counts up one by one, but is in sync
542 with all CPUs. This is useful 444 with all CPUs. This is useful when you need to
543 know exactly the order events 445 know exactly the order events occurred with respect to
544 each other on different CPUs. 446 each other on different CPUs.
545 447
546 uptime: 448 uptime:
547 This uses the jiffies counter 449 This uses the jiffies counter and the time stamp
548 is relative to the time since 450 is relative to the time since boot up.
549 451
550 perf: 452 perf:
551 This makes ftrace use the same 453 This makes ftrace use the same clock that perf uses.
552 Eventually perf will be able t 454 Eventually perf will be able to read ftrace buffers
553 and this will help out in inte 455 and this will help out in interleaving the data.
554 456
555 x86-tsc: 457 x86-tsc:
556 Architectures may define their 458 Architectures may define their own clocks. For
557 example, x86 uses its own TSC 459 example, x86 uses its own TSC cycle clock here.
558 460
559 ppc-tb: 461 ppc-tb:
560 This uses the powerpc timebase 462 This uses the powerpc timebase register value.
561 This is in sync across CPUs an 463 This is in sync across CPUs and can also be used
562 to correlate events across hyp 464 to correlate events across hypervisor/guest if
563 tb_offset is known. 465 tb_offset is known.
564 466
565 mono: 467 mono:
566 This uses the fast monotonic c 468 This uses the fast monotonic clock (CLOCK_MONOTONIC)
567 which is monotonic and is subj 469 which is monotonic and is subject to NTP rate adjustments.
568 470
569 mono_raw: 471 mono_raw:
570 This is the raw monotonic cloc 472 This is the raw monotonic clock (CLOCK_MONOTONIC_RAW)
571 which is monotonic but is not 473 which is monotonic but is not subject to any rate adjustments
572 and ticks at the same rate as 474 and ticks at the same rate as the hardware clocksource.
573 475
574 boot: 476 boot:
575 This is the boot clock (CLOCK_ 477 This is the boot clock (CLOCK_BOOTTIME) and is based on the
576 fast monotonic clock, but also 478 fast monotonic clock, but also accounts for time spent in
577 suspend. Since the clock acces 479 suspend. Since the clock access is designed for use in
578 tracing in the suspend path, s 480 tracing in the suspend path, some side effects are possible
579 if clock is accessed after the 481 if clock is accessed after the suspend time is accounted before
580 the fast mono clock is updated 482 the fast mono clock is updated. In this case, the clock update
581 appears to happen slightly soo 483 appears to happen slightly sooner than it normally would have.
582 Also on 32-bit systems, it's p 484 Also on 32-bit systems, it's possible that the 64-bit boot offset
583 sees a partial update. These e 485 sees a partial update. These effects are rare and post
584 processing should be able to h 486 processing should be able to handle them. See comments in the
585 ktime_get_boot_fast_ns() funct 487 ktime_get_boot_fast_ns() function for more information.
586 488
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 489 To set a clock, simply echo the clock name into this file::
600 490
601 # echo global > trace_clock 491 # echo global > trace_clock
602 492
603 Setting a clock clears the ring buffer <<
604 "snapshot" buffer. <<
605 <<
606 trace_marker: 493 trace_marker:
607 494
608 This is a very useful file for synchro 495 This is a very useful file for synchronizing user space
609 with events happening in the kernel. W 496 with events happening in the kernel. Writing strings into
610 this file will be written into the ftr 497 this file will be written into the ftrace buffer.
611 498
612 It is useful in applications to open t 499 It is useful in applications to open this file at the start
613 of the application and just reference 500 of the application and just reference the file descriptor
614 for the file:: 501 for the file::
615 502
616 void trace_write(const char *f 503 void trace_write(const char *fmt, ...)
617 { 504 {
618 va_list ap; 505 va_list ap;
619 char buf[256]; 506 char buf[256];
620 int n; 507 int n;
621 508
622 if (trace_fd < 0) 509 if (trace_fd < 0)
623 return; 510 return;
624 511
625 va_start(ap, fmt); 512 va_start(ap, fmt);
626 n = vsnprintf(buf, 256 513 n = vsnprintf(buf, 256, fmt, ap);
627 va_end(ap); 514 va_end(ap);
628 515
629 write(trace_fd, buf, n 516 write(trace_fd, buf, n);
630 } 517 }
631 518
632 start:: 519 start::
633 520
634 trace_fd = open("trace_marker" !! 521 trace_fd = open("trace_marker", WR_ONLY);
635 522
636 Note: Writing into the trace_marker fi 523 Note: Writing into the trace_marker file can also initiate triggers
637 that are written into /sys/kerne 524 that are written into /sys/kernel/tracing/events/ftrace/print/trigger
638 See "Event triggers" in Document 525 See "Event triggers" in Documentation/trace/events.rst and an
639 example in Documentation/trace/h 526 example in Documentation/trace/histogram.rst (Section 3.)
640 527
641 trace_marker_raw: 528 trace_marker_raw:
642 529
643 This is similar to trace_marker above, !! 530 This is similar to trace_marker above, but is meant for for binary data
644 to be written to it, where a tool can 531 to be written to it, where a tool can be used to parse the data
645 from trace_pipe_raw. 532 from trace_pipe_raw.
646 533
647 uprobe_events: 534 uprobe_events:
648 535
649 Add dynamic tracepoints in programs. 536 Add dynamic tracepoints in programs.
650 See uprobetracer.rst !! 537 See uprobetracer.txt
651 538
652 uprobe_profile: 539 uprobe_profile:
653 540
654 Uprobe statistics. See uprobetrace.txt 541 Uprobe statistics. See uprobetrace.txt
655 542
656 instances: 543 instances:
657 544
658 This is a way to make multiple trace b 545 This is a way to make multiple trace buffers where different
659 events can be recorded in different bu 546 events can be recorded in different buffers.
660 See "Instances" section below. 547 See "Instances" section below.
661 548
662 events: 549 events:
663 550
664 This is the trace event directory. It 551 This is the trace event directory. It holds event tracepoints
665 (also known as static tracepoints) tha 552 (also known as static tracepoints) that have been compiled
666 into the kernel. It shows what event t 553 into the kernel. It shows what event tracepoints exist
667 and how they are grouped by system. Th 554 and how they are grouped by system. There are "enable"
668 files at various levels that can enabl 555 files at various levels that can enable the tracepoints
669 when a "1" is written to them. 556 when a "1" is written to them.
670 557
671 See events.rst for more information. !! 558 See events.txt for more information.
672 559
673 set_event: 560 set_event:
674 561
675 By echoing in the event into this file 562 By echoing in the event into this file, will enable that event.
676 563
677 See events.rst for more information. !! 564 See events.txt for more information.
678 565
679 available_events: 566 available_events:
680 567
681 A list of events that can be enabled i 568 A list of events that can be enabled in tracing.
682 569
683 See events.rst for more information. !! 570 See events.txt for more information.
684 571
685 timestamp_mode: 572 timestamp_mode:
686 573
687 Certain tracers may change the timesta 574 Certain tracers may change the timestamp mode used when
688 logging trace events into the event bu 575 logging trace events into the event buffer. Events with
689 different modes can coexist within a b 576 different modes can coexist within a buffer but the mode in
690 effect when an event is logged determi 577 effect when an event is logged determines which timestamp mode
691 is used for that event. The default t 578 is used for that event. The default timestamp mode is
692 'delta'. 579 'delta'.
693 580
694 Usual timestamp modes for tracing: 581 Usual timestamp modes for tracing:
695 582
696 # cat timestamp_mode 583 # cat timestamp_mode
697 [delta] absolute 584 [delta] absolute
698 585
699 The timestamp mode with the square b 586 The timestamp mode with the square brackets around it is the
700 one in effect. 587 one in effect.
701 588
702 delta: Default timestamp mode - time 589 delta: Default timestamp mode - timestamp is a delta against
703 a per-buffer timestamp. 590 a per-buffer timestamp.
704 591
705 absolute: The timestamp is a full ti 592 absolute: The timestamp is a full timestamp, not a delta
706 against some other value. As 593 against some other value. As such it takes up more
707 space and is less efficient. 594 space and is less efficient.
708 595
709 hwlat_detector: 596 hwlat_detector:
710 597
711 Directory for the Hardware Latency Det 598 Directory for the Hardware Latency Detector.
712 See "Hardware Latency Detector" sectio 599 See "Hardware Latency Detector" section below.
713 600
714 per_cpu: 601 per_cpu:
715 602
716 This is a directory that contains the 603 This is a directory that contains the trace per_cpu information.
717 604
718 per_cpu/cpu0/buffer_size_kb: 605 per_cpu/cpu0/buffer_size_kb:
719 606
720 The ftrace buffer is defined per_cpu. 607 The ftrace buffer is defined per_cpu. That is, there's a separate
721 buffer for each CPU to allow writes to 608 buffer for each CPU to allow writes to be done atomically,
722 and free from cache bouncing. These bu 609 and free from cache bouncing. These buffers may have different
723 size buffers. This file is similar to 610 size buffers. This file is similar to the buffer_size_kb
724 file, but it only displays or sets the 611 file, but it only displays or sets the buffer size for the
725 specific CPU. (here cpu0). 612 specific CPU. (here cpu0).
726 613
727 per_cpu/cpu0/trace: 614 per_cpu/cpu0/trace:
728 615
729 This is similar to the "trace" file, b 616 This is similar to the "trace" file, but it will only display
730 the data specific for the CPU. If writ 617 the data specific for the CPU. If written to, it only clears
731 the specific CPU buffer. 618 the specific CPU buffer.
732 619
733 per_cpu/cpu0/trace_pipe 620 per_cpu/cpu0/trace_pipe
734 621
735 This is similar to the "trace_pipe" fi 622 This is similar to the "trace_pipe" file, and is a consuming
736 read, but it will only display (and co 623 read, but it will only display (and consume) the data specific
737 for the CPU. 624 for the CPU.
738 625
739 per_cpu/cpu0/trace_pipe_raw 626 per_cpu/cpu0/trace_pipe_raw
740 627
741 For tools that can parse the ftrace ri 628 For tools that can parse the ftrace ring buffer binary format,
742 the trace_pipe_raw file can be used to 629 the trace_pipe_raw file can be used to extract the data
743 from the ring buffer directly. With th 630 from the ring buffer directly. With the use of the splice()
744 system call, the buffer data can be qu 631 system call, the buffer data can be quickly transferred to
745 a file or to the network where a serve 632 a file or to the network where a server is collecting the
746 data. 633 data.
747 634
748 Like trace_pipe, this is a consuming r 635 Like trace_pipe, this is a consuming reader, where multiple
749 reads will always produce different da 636 reads will always produce different data.
750 637
751 per_cpu/cpu0/snapshot: 638 per_cpu/cpu0/snapshot:
752 639
753 This is similar to the main "snapshot" 640 This is similar to the main "snapshot" file, but will only
754 snapshot the current CPU (if supported 641 snapshot the current CPU (if supported). It only displays
755 the content of the snapshot for a give 642 the content of the snapshot for a given CPU, and if
756 written to, only clears this CPU buffe 643 written to, only clears this CPU buffer.
757 644
758 per_cpu/cpu0/snapshot_raw: 645 per_cpu/cpu0/snapshot_raw:
759 646
760 Similar to the trace_pipe_raw, but wil 647 Similar to the trace_pipe_raw, but will read the binary format
761 from the snapshot buffer for the given 648 from the snapshot buffer for the given CPU.
762 649
763 per_cpu/cpu0/stats: 650 per_cpu/cpu0/stats:
764 651
765 This displays certain stats about the 652 This displays certain stats about the ring buffer:
766 653
767 entries: 654 entries:
768 The number of events that are 655 The number of events that are still in the buffer.
769 656
770 overrun: 657 overrun:
771 The number of lost events due 658 The number of lost events due to overwriting when
772 the buffer was full. 659 the buffer was full.
773 660
774 commit overrun: 661 commit overrun:
775 Should always be zero. 662 Should always be zero.
776 This gets set if so many event 663 This gets set if so many events happened within a nested
777 event (ring buffer is re-entra 664 event (ring buffer is re-entrant), that it fills the
778 buffer and starts dropping eve 665 buffer and starts dropping events.
779 666
780 bytes: 667 bytes:
781 Bytes actually read (not overw 668 Bytes actually read (not overwritten).
782 669
783 oldest event ts: 670 oldest event ts:
784 The oldest timestamp in the bu 671 The oldest timestamp in the buffer
785 672
786 now ts: 673 now ts:
787 The current timestamp 674 The current timestamp
788 675
789 dropped events: 676 dropped events:
790 Events lost due to overwrite o 677 Events lost due to overwrite option being off.
791 678
792 read events: 679 read events:
793 The number of events read. 680 The number of events read.
794 681
795 The Tracers 682 The Tracers
796 ----------- 683 -----------
797 684
798 Here is the list of current tracers that may b 685 Here is the list of current tracers that may be configured.
799 686
800 "function" 687 "function"
801 688
802 Function call tracer to trace all kern 689 Function call tracer to trace all kernel functions.
803 690
804 "function_graph" 691 "function_graph"
805 692
806 Similar to the function tracer except 693 Similar to the function tracer except that the
807 function tracer probes the functions o 694 function tracer probes the functions on their entry
808 whereas the function graph tracer trac 695 whereas the function graph tracer traces on both entry
809 and exit of the functions. It then pro 696 and exit of the functions. It then provides the ability
810 to draw a graph of function calls simi 697 to draw a graph of function calls similar to C code
811 source. 698 source.
812 699
813 "blk" 700 "blk"
814 701
815 The block tracer. The tracer used by t 702 The block tracer. The tracer used by the blktrace user
816 application. 703 application.
817 704
818 "hwlat" 705 "hwlat"
819 706
820 The Hardware Latency tracer is used to 707 The Hardware Latency tracer is used to detect if the hardware
821 produces any latency. See "Hardware La 708 produces any latency. See "Hardware Latency Detector" section
822 below. 709 below.
823 710
824 "irqsoff" 711 "irqsoff"
825 712
826 Traces the areas that disable interrup 713 Traces the areas that disable interrupts and saves
827 the trace with the longest max latency 714 the trace with the longest max latency.
828 See tracing_max_latency. When a new ma 715 See tracing_max_latency. When a new max is recorded,
829 it replaces the old trace. It is best 716 it replaces the old trace. It is best to view this
830 trace with the latency-format option e 717 trace with the latency-format option enabled, which
831 happens automatically when the tracer 718 happens automatically when the tracer is selected.
832 719
833 "preemptoff" 720 "preemptoff"
834 721
835 Similar to irqsoff but traces and reco 722 Similar to irqsoff but traces and records the amount of
836 time for which preemption is disabled. 723 time for which preemption is disabled.
837 724
838 "preemptirqsoff" 725 "preemptirqsoff"
839 726
840 Similar to irqsoff and preemptoff, but 727 Similar to irqsoff and preemptoff, but traces and
841 records the largest time for which irq 728 records the largest time for which irqs and/or preemption
842 is disabled. 729 is disabled.
843 730
844 "wakeup" 731 "wakeup"
845 732
846 Traces and records the max latency tha 733 Traces and records the max latency that it takes for
847 the highest priority task to get sched 734 the highest priority task to get scheduled after
848 it has been woken up. 735 it has been woken up.
849 Traces all tasks as an average develop 736 Traces all tasks as an average developer would expect.
850 737
851 "wakeup_rt" 738 "wakeup_rt"
852 739
853 Traces and records the max latency tha 740 Traces and records the max latency that it takes for just
854 RT tasks (as the current "wakeup" does 741 RT tasks (as the current "wakeup" does). This is useful
855 for those interested in wake up timing 742 for those interested in wake up timings of RT tasks.
856 743
857 "wakeup_dl" 744 "wakeup_dl"
858 745
859 Traces and records the max latency tha 746 Traces and records the max latency that it takes for
860 a SCHED_DEADLINE task to be woken (as 747 a SCHED_DEADLINE task to be woken (as the "wakeup" and
861 "wakeup_rt" does). 748 "wakeup_rt" does).
862 749
863 "mmiotrace" 750 "mmiotrace"
864 751
865 A special tracer that is used to trace 752 A special tracer that is used to trace binary module.
866 It will trace all the calls that a mod 753 It will trace all the calls that a module makes to the
867 hardware. Everything it writes and rea 754 hardware. Everything it writes and reads from the I/O
868 as well. 755 as well.
869 756
870 "branch" 757 "branch"
871 758
872 This tracer can be configured when tra 759 This tracer can be configured when tracing likely/unlikely
873 calls within the kernel. It will trace 760 calls within the kernel. It will trace when a likely and
874 unlikely branch is hit and if it was c 761 unlikely branch is hit and if it was correct in its prediction
875 of being correct. 762 of being correct.
876 763
877 "nop" 764 "nop"
878 765
879 This is the "trace nothing" tracer. To 766 This is the "trace nothing" tracer. To remove all
880 tracers from tracing simply echo "nop" 767 tracers from tracing simply echo "nop" into
881 current_tracer. 768 current_tracer.
882 769
883 Error conditions 770 Error conditions
884 ---------------- 771 ----------------
885 772
886 For most ftrace commands, failure modes are 773 For most ftrace commands, failure modes are obvious and communicated
887 using standard return codes. 774 using standard return codes.
888 775
889 For other more involved commands, extended e 776 For other more involved commands, extended error information may be
890 available via the tracing/error_log file. F 777 available via the tracing/error_log file. For the commands that
891 support it, reading the tracing/error_log fi 778 support it, reading the tracing/error_log file after an error will
892 display more detailed information about what 779 display more detailed information about what went wrong, if
893 information is available. The tracing/error 780 information is available. The tracing/error_log file is a circular
894 error log displaying a small number (current 781 error log displaying a small number (currently, 8) of ftrace errors
895 for the last (8) failed commands. 782 for the last (8) failed commands.
896 783
897 The extended error information and usage tak 784 The extended error information and usage takes the form shown in
898 this example:: 785 this example::
899 786
900 # echo xxx > /sys/kernel/tracing/events/sc !! 787 # echo xxx > /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger
901 echo: write error: Invalid argument 788 echo: write error: Invalid argument
902 789
903 # cat /sys/kernel/tracing/error_log !! 790 # cat /sys/kernel/debug/tracing/error_log
904 [ 5348.887237] location: error: Couldn't y 791 [ 5348.887237] location: error: Couldn't yyy: zzz
905 Command: xxx 792 Command: xxx
906 ^ 793 ^
907 [ 7517.023364] location: error: Bad rrr: s 794 [ 7517.023364] location: error: Bad rrr: sss
908 Command: ppp qqq 795 Command: ppp qqq
909 ^ 796 ^
910 797
911 To clear the error log, echo the empty strin 798 To clear the error log, echo the empty string into it::
912 799
913 # echo > /sys/kernel/tracing/error_log !! 800 # echo > /sys/kernel/debug/tracing/error_log
914 801
915 Examples of using the tracer 802 Examples of using the tracer
916 ---------------------------- 803 ----------------------------
917 804
918 Here are typical examples of using the tracers 805 Here are typical examples of using the tracers when controlling
919 them only with the tracefs interface (without 806 them only with the tracefs interface (without using any
920 user-land utilities). 807 user-land utilities).
921 808
922 Output format: 809 Output format:
923 -------------- 810 --------------
924 811
925 Here is an example of the output format of the 812 Here is an example of the output format of the file "trace"::
926 813
927 # tracer: function 814 # tracer: function
928 # 815 #
929 # entries-in-buffer/entries-written: 140080/ 816 # entries-in-buffer/entries-written: 140080/250280 #P:4
930 # 817 #
931 # _-----=> irqs 818 # _-----=> irqs-off
932 # / _----=> need 819 # / _----=> need-resched
933 # | / _---=> hard 820 # | / _---=> hardirq/softirq
934 # || / _--=> pree 821 # || / _--=> preempt-depth
935 # ||| / delay 822 # ||| / delay
936 # TASK-PID CPU# |||| TIMESTA 823 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
937 # | | | |||| | 824 # | | | |||| | |
938 bash-1977 [000] .... 17284.9936 825 bash-1977 [000] .... 17284.993652: sys_close <-system_call_fastpath
939 bash-1977 [000] .... 17284.9936 826 bash-1977 [000] .... 17284.993653: __close_fd <-sys_close
940 bash-1977 [000] .... 17284.9936 827 bash-1977 [000] .... 17284.993653: _raw_spin_lock <-__close_fd
941 sshd-1974 [003] .... 17284.9936 828 sshd-1974 [003] .... 17284.993653: __srcu_read_unlock <-fsnotify
942 bash-1977 [000] .... 17284.9936 829 bash-1977 [000] .... 17284.993654: add_preempt_count <-_raw_spin_lock
943 bash-1977 [000] ...1 17284.9936 830 bash-1977 [000] ...1 17284.993655: _raw_spin_unlock <-__close_fd
944 bash-1977 [000] ...1 17284.9936 831 bash-1977 [000] ...1 17284.993656: sub_preempt_count <-_raw_spin_unlock
945 bash-1977 [000] .... 17284.9936 832 bash-1977 [000] .... 17284.993657: filp_close <-__close_fd
946 bash-1977 [000] .... 17284.9936 833 bash-1977 [000] .... 17284.993657: dnotify_flush <-filp_close
947 sshd-1974 [003] .... 17284.9936 834 sshd-1974 [003] .... 17284.993658: sys_select <-system_call_fastpath
948 .... 835 ....
949 836
950 A header is printed with the tracer name that 837 A header is printed with the tracer name that is represented by
951 the trace. In this case the tracer is "functio 838 the trace. In this case the tracer is "function". Then it shows the
952 number of events in the buffer as well as the 839 number of events in the buffer as well as the total number of entries
953 that were written. The difference is the numbe 840 that were written. The difference is the number of entries that were
954 lost due to the buffer filling up (250280 - 14 841 lost due to the buffer filling up (250280 - 140080 = 110200 events
955 lost). 842 lost).
956 843
957 The header explains the content of the events. 844 The header explains the content of the events. Task name "bash", the task
958 PID "1977", the CPU that it was running on "00 845 PID "1977", the CPU that it was running on "000", the latency format
959 (explained below), the timestamp in <secs>.<us 846 (explained below), the timestamp in <secs>.<usecs> format, the
960 function name that was traced "sys_close" and 847 function name that was traced "sys_close" and the parent function that
961 called this function "system_call_fastpath". T 848 called this function "system_call_fastpath". The timestamp is the time
962 at which the function was entered. 849 at which the function was entered.
963 850
964 Latency trace format 851 Latency trace format
965 -------------------- 852 --------------------
966 853
967 When the latency-format option is enabled or w 854 When the latency-format option is enabled or when one of the latency
968 tracers is set, the trace file gives somewhat 855 tracers is set, the trace file gives somewhat more information to see
969 why a latency happened. Here is a typical trac 856 why a latency happened. Here is a typical trace::
970 857
971 # tracer: irqsoff 858 # tracer: irqsoff
972 # 859 #
973 # irqsoff latency trace v1.1.5 on 3.8.0-test 860 # irqsoff latency trace v1.1.5 on 3.8.0-test+
974 # ------------------------------------------ 861 # --------------------------------------------------------------------
975 # latency: 259 us, #4/4, CPU#2 | (M:preempt 862 # latency: 259 us, #4/4, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
976 # ----------------- 863 # -----------------
977 # | task: ps-6143 (uid:0 nice:0 policy:0 864 # | task: ps-6143 (uid:0 nice:0 policy:0 rt_prio:0)
978 # ----------------- 865 # -----------------
979 # => started at: __lock_task_sighand 866 # => started at: __lock_task_sighand
980 # => ended at: _raw_spin_unlock_irqrestor 867 # => ended at: _raw_spin_unlock_irqrestore
981 # 868 #
982 # 869 #
983 # _------=> CPU# 870 # _------=> CPU#
984 # / _-----=> irqs-off 871 # / _-----=> irqs-off
985 # | / _----=> need-resched 872 # | / _----=> need-resched
986 # || / _---=> hardirq/softirq 873 # || / _---=> hardirq/softirq
987 # ||| / _--=> preempt-depth 874 # ||| / _--=> preempt-depth
988 # |||| / delay 875 # |||| / delay
989 # cmd pid ||||| time | caller 876 # cmd pid ||||| time | caller
990 # \ / ||||| \ | / 877 # \ / ||||| \ | /
991 ps-6143 2d... 0us!: trace_hardir 878 ps-6143 2d... 0us!: trace_hardirqs_off <-__lock_task_sighand
992 ps-6143 2d..1 259us+: trace_hardir 879 ps-6143 2d..1 259us+: trace_hardirqs_on <-_raw_spin_unlock_irqrestore
993 ps-6143 2d..1 263us+: time_hardirq 880 ps-6143 2d..1 263us+: time_hardirqs_on <-_raw_spin_unlock_irqrestore
994 ps-6143 2d..1 306us : <stack trace 881 ps-6143 2d..1 306us : <stack trace>
995 => trace_hardirqs_on_caller 882 => trace_hardirqs_on_caller
996 => trace_hardirqs_on 883 => trace_hardirqs_on
997 => _raw_spin_unlock_irqrestore 884 => _raw_spin_unlock_irqrestore
998 => do_task_stat 885 => do_task_stat
999 => proc_tgid_stat 886 => proc_tgid_stat
1000 => proc_single_show 887 => proc_single_show
1001 => seq_read 888 => seq_read
1002 => vfs_read 889 => vfs_read
1003 => sys_read 890 => sys_read
1004 => system_call_fastpath 891 => system_call_fastpath
1005 892
1006 893
1007 This shows that the current tracer is "irqsof 894 This shows that the current tracer is "irqsoff" tracing the time
1008 for which interrupts were disabled. It gives 895 for which interrupts were disabled. It gives the trace version (which
1009 never changes) and the version of the kernel 896 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 897 (3.8). Then it displays the max latency in microseconds (259 us). The number
1011 of trace entries displayed and the total numb 898 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 899 VP, KP, SP, and HP are always zero and are reserved for later use.
1013 #P is the number of online CPUs (#P:4). 900 #P is the number of online CPUs (#P:4).
1014 901
1015 The task is the process that was running when 902 The task is the process that was running when the latency
1016 occurred. (ps pid: 6143). 903 occurred. (ps pid: 6143).
1017 904
1018 The start and stop (the functions in which th 905 The start and stop (the functions in which the interrupts were
1019 disabled and enabled respectively) that cause 906 disabled and enabled respectively) that caused the latencies:
1020 907
1021 - __lock_task_sighand is where the interrup 908 - __lock_task_sighand is where the interrupts were disabled.
1022 - _raw_spin_unlock_irqrestore is where they 909 - _raw_spin_unlock_irqrestore is where they were enabled again.
1023 910
1024 The next lines after the header are the trace 911 The next lines after the header are the trace itself. The header
1025 explains which is which. 912 explains which is which.
1026 913
1027 cmd: The name of the process in the trace. 914 cmd: The name of the process in the trace.
1028 915
1029 pid: The PID of that process. 916 pid: The PID of that process.
1030 917
1031 CPU#: The CPU which the process was running 918 CPU#: The CPU which the process was running on.
1032 919
1033 irqs-off: 'd' interrupts are disabled. '.' 920 irqs-off: 'd' interrupts are disabled. '.' otherwise.
1034 .. caution:: If the architecture does 921 .. caution:: If the architecture does not support a way to
1035 read the irq flags variable, 922 read the irq flags variable, an 'X' will always
1036 be printed here. 923 be printed here.
1037 924
1038 need-resched: 925 need-resched:
1039 - 'N' both TIF_NEED_RESCHED and PREEM 926 - 'N' both TIF_NEED_RESCHED and PREEMPT_NEED_RESCHED is set,
1040 - 'n' only TIF_NEED_RESCHED is set, 927 - 'n' only TIF_NEED_RESCHED is set,
1041 - 'p' only PREEMPT_NEED_RESCHED is se 928 - 'p' only PREEMPT_NEED_RESCHED is set,
1042 - '.' otherwise. 929 - '.' otherwise.
1043 930
1044 hardirq/softirq: 931 hardirq/softirq:
1045 - 'Z' - NMI occurred inside a hardirq 932 - 'Z' - NMI occurred inside a hardirq
1046 - 'z' - NMI is running 933 - 'z' - NMI is running
1047 - 'H' - hard irq occurred inside a so 934 - 'H' - hard irq occurred inside a softirq.
1048 - 'h' - hard irq is running 935 - 'h' - hard irq is running
1049 - 's' - soft irq is running 936 - 's' - soft irq is running
1050 - '.' - normal context. 937 - '.' - normal context.
1051 938
1052 preempt-depth: The level of preempt_disable 939 preempt-depth: The level of preempt_disabled
1053 940
1054 The above is mostly meaningful for kernel dev 941 The above is mostly meaningful for kernel developers.
1055 942
1056 time: 943 time:
1057 When the latency-format option is ena 944 When the latency-format option is enabled, the trace file
1058 output includes a timestamp relative 945 output includes a timestamp relative to the start of the
1059 trace. This differs from the output w 946 trace. This differs from the output when latency-format
1060 is disabled, which includes an absolu 947 is disabled, which includes an absolute timestamp.
1061 948
1062 delay: 949 delay:
1063 This is just to help catch your eye a 950 This is just to help catch your eye a bit better. And
1064 needs to be fixed to be only relative 951 needs to be fixed to be only relative to the same CPU.
1065 The marks are determined by the diffe 952 The marks are determined by the difference between this
1066 current trace and the next trace. 953 current trace and the next trace.
1067 954
1068 - '$' - greater than 1 second 955 - '$' - greater than 1 second
1069 - '@' - greater than 100 millisecon 956 - '@' - greater than 100 millisecond
1070 - '*' - greater than 10 millisecond 957 - '*' - greater than 10 millisecond
1071 - '#' - greater than 1000 microseco 958 - '#' - greater than 1000 microsecond
1072 - '!' - greater than 100 microsecon 959 - '!' - greater than 100 microsecond
1073 - '+' - greater than 10 microsecond 960 - '+' - greater than 10 microsecond
1074 - ' ' - less than or equal to 10 mi 961 - ' ' - less than or equal to 10 microsecond.
1075 962
1076 The rest is the same as the 'trace' file. 963 The rest is the same as the 'trace' file.
1077 964
1078 Note, the latency tracers will usually end 965 Note, the latency tracers will usually end with a back trace
1079 to easily find where the latency occurred. 966 to easily find where the latency occurred.
1080 967
1081 trace_options 968 trace_options
1082 ------------- 969 -------------
1083 970
1084 The trace_options file (or the options direct 971 The trace_options file (or the options directory) is used to control
1085 what gets printed in the trace output, or man 972 what gets printed in the trace output, or manipulate the tracers.
1086 To see what is available, simply cat the file 973 To see what is available, simply cat the file::
1087 974
1088 cat trace_options 975 cat trace_options
1089 print-parent 976 print-parent
1090 nosym-offset 977 nosym-offset
1091 nosym-addr 978 nosym-addr
1092 noverbose 979 noverbose
1093 noraw 980 noraw
1094 nohex 981 nohex
1095 nobin 982 nobin
1096 noblock 983 noblock
1097 nofields <<
1098 trace_printk 984 trace_printk
1099 annotate 985 annotate
1100 nouserstacktrace 986 nouserstacktrace
1101 nosym-userobj 987 nosym-userobj
1102 noprintk-msg-only 988 noprintk-msg-only
1103 context-info 989 context-info
1104 nolatency-format 990 nolatency-format
1105 record-cmd 991 record-cmd
1106 norecord-tgid 992 norecord-tgid
1107 overwrite 993 overwrite
1108 nodisable_on_free 994 nodisable_on_free
1109 irq-info 995 irq-info
1110 markers 996 markers
1111 noevent-fork 997 noevent-fork
1112 function-trace 998 function-trace
1113 nofunction-fork 999 nofunction-fork
1114 nodisplay-graph 1000 nodisplay-graph
1115 nostacktrace 1001 nostacktrace
1116 nobranch 1002 nobranch
1117 1003
1118 To disable one of the options, echo in the op 1004 To disable one of the options, echo in the option prepended with
1119 "no":: 1005 "no"::
1120 1006
1121 echo noprint-parent > trace_options 1007 echo noprint-parent > trace_options
1122 1008
1123 To enable an option, leave off the "no":: 1009 To enable an option, leave off the "no"::
1124 1010
1125 echo sym-offset > trace_options 1011 echo sym-offset > trace_options
1126 1012
1127 Here are the available options: 1013 Here are the available options:
1128 1014
1129 print-parent 1015 print-parent
1130 On function traces, display the calli 1016 On function traces, display the calling (parent)
1131 function as well as the function bein 1017 function as well as the function being traced.
1132 :: 1018 ::
1133 1019
1134 print-parent: 1020 print-parent:
1135 bash-4000 [01] 1477.606694: simp 1021 bash-4000 [01] 1477.606694: simple_strtoul <-kstrtoul
1136 1022
1137 noprint-parent: 1023 noprint-parent:
1138 bash-4000 [01] 1477.606694: simp 1024 bash-4000 [01] 1477.606694: simple_strtoul
1139 1025
1140 1026
1141 sym-offset 1027 sym-offset
1142 Display not only the function name, b 1028 Display not only the function name, but also the
1143 offset in the function. For example, 1029 offset in the function. For example, instead of
1144 seeing just "ktime_get", you will see 1030 seeing just "ktime_get", you will see
1145 "ktime_get+0xb/0x20". 1031 "ktime_get+0xb/0x20".
1146 :: 1032 ::
1147 1033
1148 sym-offset: 1034 sym-offset:
1149 bash-4000 [01] 1477.606694: simp 1035 bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0
1150 1036
1151 sym-addr 1037 sym-addr
1152 This will also display the function a 1038 This will also display the function address as well
1153 as the function name. 1039 as the function name.
1154 :: 1040 ::
1155 1041
1156 sym-addr: 1042 sym-addr:
1157 bash-4000 [01] 1477.606694: simp 1043 bash-4000 [01] 1477.606694: simple_strtoul <c0339346>
1158 1044
1159 verbose 1045 verbose
1160 This deals with the trace file when t 1046 This deals with the trace file when the
1161 latency-format option is enabled. 1047 latency-format option is enabled.
1162 :: 1048 ::
1163 1049
1164 bash 4000 1 0 00000000 00010a95 1050 bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \
1165 (+0.000ms): simple_strtoul (kstrt 1051 (+0.000ms): simple_strtoul (kstrtoul)
1166 1052
1167 raw 1053 raw
1168 This will display raw numbers. This o 1054 This will display raw numbers. This option is best for
1169 use with user applications that can t 1055 use with user applications that can translate the raw
1170 numbers better than having it done in 1056 numbers better than having it done in the kernel.
1171 1057
1172 hex 1058 hex
1173 Similar to raw, but the numbers will 1059 Similar to raw, but the numbers will be in a hexadecimal format.
1174 1060
1175 bin 1061 bin
1176 This will print out the formats in ra 1062 This will print out the formats in raw binary.
1177 1063
1178 block 1064 block
1179 When set, reading trace_pipe will not 1065 When set, reading trace_pipe will not block when polled.
1180 1066
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 1067 trace_printk
1187 Can disable trace_printk() from writi 1068 Can disable trace_printk() from writing into the buffer.
1188 1069
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 1070 annotate
1202 It is sometimes confusing when the CP 1071 It is sometimes confusing when the CPU buffers are full
1203 and one CPU buffer had a lot of event 1072 and one CPU buffer had a lot of events recently, thus
1204 a shorter time frame, were another CP 1073 a shorter time frame, were another CPU may have only had
1205 a few events, which lets it have olde 1074 a few events, which lets it have older events. When
1206 the trace is reported, it shows the o 1075 the trace is reported, it shows the oldest events first,
1207 and it may look like only one CPU ran 1076 and it may look like only one CPU ran (the one with the
1208 oldest events). When the annotate opt 1077 oldest events). When the annotate option is set, it will
1209 display when a new CPU buffer started 1078 display when a new CPU buffer started::
1210 1079
1211 <idle>-0 [001] 1080 <idle>-0 [001] dNs4 21169.031481: wake_up_idle_cpu <-add_timer_on
1212 <idle>-0 [001] 1081 <idle>-0 [001] dNs4 21169.031482: _raw_spin_unlock_irqrestore <-add_timer_on
1213 <idle>-0 [001] 1082 <idle>-0 [001] .Ns4 21169.031484: sub_preempt_count <-_raw_spin_unlock_irqrestore
1214 ##### CPU 2 buffer started ## 1083 ##### CPU 2 buffer started ####
1215 <idle>-0 [002] 1084 <idle>-0 [002] .N.1 21169.031484: rcu_idle_exit <-cpu_idle
1216 <idle>-0 [001] 1085 <idle>-0 [001] .Ns3 21169.031484: _raw_spin_unlock <-clocksource_watchdog
1217 <idle>-0 [001] 1086 <idle>-0 [001] .Ns3 21169.031485: sub_preempt_count <-_raw_spin_unlock
1218 1087
1219 userstacktrace 1088 userstacktrace
1220 This option changes the trace. It rec 1089 This option changes the trace. It records a
1221 stacktrace of the current user space 1090 stacktrace of the current user space thread after
1222 each trace event. 1091 each trace event.
1223 1092
1224 sym-userobj 1093 sym-userobj
1225 when user stacktrace are enabled, loo 1094 when user stacktrace are enabled, look up which
1226 object the address belongs to, and pr 1095 object the address belongs to, and print a
1227 relative address. This is especially 1096 relative address. This is especially useful when
1228 ASLR is on, otherwise you don't get a 1097 ASLR is on, otherwise you don't get a chance to
1229 resolve the address to object/file/li 1098 resolve the address to object/file/line after
1230 the app is no longer running 1099 the app is no longer running
1231 1100
1232 The lookup is performed when you read 1101 The lookup is performed when you read
1233 trace,trace_pipe. Example:: 1102 trace,trace_pipe. Example::
1234 1103
1235 a.out-1623 [000] 40874.465 1104 a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0
1236 x494] <- /root/a.out[+0x4a8 1105 x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6]
1237 1106
1238 1107
1239 printk-msg-only 1108 printk-msg-only
1240 When set, trace_printk()s will only s 1109 When set, trace_printk()s will only show the format
1241 and not their parameters (if trace_bp 1110 and not their parameters (if trace_bprintk() or
1242 trace_bputs() was used to save the tr 1111 trace_bputs() was used to save the trace_printk()).
1243 1112
1244 context-info 1113 context-info
1245 Show only the event data. Hides the c 1114 Show only the event data. Hides the comm, PID,
1246 timestamp, CPU, and other useful data 1115 timestamp, CPU, and other useful data.
1247 1116
1248 latency-format 1117 latency-format
1249 This option changes the trace output. 1118 This option changes the trace output. When it is enabled,
1250 the trace displays additional informa 1119 the trace displays additional information about the
1251 latency, as described in "Latency tra 1120 latency, as described in "Latency trace format".
1252 1121
1253 pause-on-trace <<
1254 When set, opening the trace file for <<
1255 writing to the ring buffer (as if tra <<
1256 This simulates the original behavior <<
1257 When the file is closed, tracing will <<
1258 <<
1259 hash-ptr <<
1260 When set, "%p" in the event printk fo <<
1261 hashed pointer value instead of real <<
1262 This will be useful if you want to fi <<
1263 value is corresponding to the real va <<
1264 <<
1265 record-cmd 1122 record-cmd
1266 When any event or tracer is enabled, 1123 When any event or tracer is enabled, a hook is enabled
1267 in the sched_switch trace point to fi 1124 in the sched_switch trace point to fill comm cache
1268 with mapped pids and comms. But this 1125 with mapped pids and comms. But this may cause some
1269 overhead, and if you only care about 1126 overhead, and if you only care about pids, and not the
1270 name of the task, disabling this opti 1127 name of the task, disabling this option can lower the
1271 impact of tracing. See "saved_cmdline 1128 impact of tracing. See "saved_cmdlines".
1272 1129
1273 record-tgid 1130 record-tgid
1274 When any event or tracer is enabled, 1131 When any event or tracer is enabled, a hook is enabled
1275 in the sched_switch trace point to fi 1132 in the sched_switch trace point to fill the cache of
1276 mapped Thread Group IDs (TGID) mappin 1133 mapped Thread Group IDs (TGID) mapping to pids. See
1277 "saved_tgids". 1134 "saved_tgids".
1278 1135
1279 overwrite 1136 overwrite
1280 This controls what happens when the t 1137 This controls what happens when the trace buffer is
1281 full. If "1" (default), the oldest ev 1138 full. If "1" (default), the oldest events are
1282 discarded and overwritten. If "0", th 1139 discarded and overwritten. If "0", then the newest
1283 events are discarded. 1140 events are discarded.
1284 (see per_cpu/cpu0/stats for overrun a 1141 (see per_cpu/cpu0/stats for overrun and dropped)
1285 1142
1286 disable_on_free 1143 disable_on_free
1287 When the free_buffer is closed, traci 1144 When the free_buffer is closed, tracing will
1288 stop (tracing_on set to 0). 1145 stop (tracing_on set to 0).
1289 1146
1290 irq-info 1147 irq-info
1291 Shows the interrupt, preempt count, n 1148 Shows the interrupt, preempt count, need resched data.
1292 When disabled, the trace looks like:: 1149 When disabled, the trace looks like::
1293 1150
1294 # tracer: function 1151 # tracer: function
1295 # 1152 #
1296 # entries-in-buffer/entries-w 1153 # entries-in-buffer/entries-written: 144405/9452052 #P:4
1297 # 1154 #
1298 # TASK-PID CPU# 1155 # TASK-PID CPU# TIMESTAMP FUNCTION
1299 # | | | 1156 # | | | | |
1300 <idle>-0 [002] 1157 <idle>-0 [002] 23636.756054: ttwu_do_activate.constprop.89 <-try_to_wake_up
1301 <idle>-0 [002] 1158 <idle>-0 [002] 23636.756054: activate_task <-ttwu_do_activate.constprop.89
1302 <idle>-0 [002] 1159 <idle>-0 [002] 23636.756055: enqueue_task <-activate_task
1303 1160
1304 1161
1305 markers 1162 markers
1306 When set, the trace_marker is writabl 1163 When set, the trace_marker is writable (only by root).
1307 When disabled, the trace_marker will 1164 When disabled, the trace_marker will error with EINVAL
1308 on write. 1165 on write.
1309 1166
1310 event-fork 1167 event-fork
1311 When set, tasks with PIDs listed in s 1168 When set, tasks with PIDs listed in set_event_pid will have
1312 the PIDs of their children added to s 1169 the PIDs of their children added to set_event_pid when those
1313 tasks fork. Also, when tasks with PID 1170 tasks fork. Also, when tasks with PIDs in set_event_pid exit,
1314 their PIDs will be removed from the f 1171 their PIDs will be removed from the file.
1315 1172
1316 This affects PIDs listed in set_event <<
1317 <<
1318 function-trace 1173 function-trace
1319 The latency tracers will enable funct 1174 The latency tracers will enable function tracing
1320 if this option is enabled (default it 1175 if this option is enabled (default it is). When
1321 it is disabled, the latency tracers d 1176 it is disabled, the latency tracers do not trace
1322 functions. This keeps the overhead of 1177 functions. This keeps the overhead of the tracer down
1323 when performing latency tests. 1178 when performing latency tests.
1324 1179
1325 function-fork 1180 function-fork
1326 When set, tasks with PIDs listed in s 1181 When set, tasks with PIDs listed in set_ftrace_pid will
1327 have the PIDs of their children added 1182 have the PIDs of their children added to set_ftrace_pid
1328 when those tasks fork. Also, when tas 1183 when those tasks fork. Also, when tasks with PIDs in
1329 set_ftrace_pid exit, their PIDs will 1184 set_ftrace_pid exit, their PIDs will be removed from the
1330 file. 1185 file.
1331 1186
1332 This affects PIDs in set_ftrace_notra <<
1333 <<
1334 display-graph 1187 display-graph
1335 When set, the latency tracers (irqsof 1188 When set, the latency tracers (irqsoff, wakeup, etc) will
1336 use function graph tracing instead of 1189 use function graph tracing instead of function tracing.
1337 1190
1338 stacktrace 1191 stacktrace
1339 When set, a stack trace is recorded a 1192 When set, a stack trace is recorded after any trace event
1340 is recorded. 1193 is recorded.
1341 1194
1342 branch 1195 branch
1343 Enable branch tracing with the tracer 1196 Enable branch tracing with the tracer. This enables branch
1344 tracer along with the currently set t 1197 tracer along with the currently set tracer. Enabling this
1345 with the "nop" tracer is the same as 1198 with the "nop" tracer is the same as just enabling the
1346 "branch" tracer. 1199 "branch" tracer.
1347 1200
1348 .. tip:: Some tracers have their own options. 1201 .. tip:: Some tracers have their own options. They only appear in this
1349 file when the tracer is active. They a 1202 file when the tracer is active. They always appear in the
1350 options directory. 1203 options directory.
1351 1204
1352 1205
1353 Here are the per tracer options: 1206 Here are the per tracer options:
1354 1207
1355 Options for function tracer: 1208 Options for function tracer:
1356 1209
1357 func_stack_trace 1210 func_stack_trace
1358 When set, a stack trace is recorded a 1211 When set, a stack trace is recorded after every
1359 function that is recorded. NOTE! Limi 1212 function that is recorded. NOTE! Limit the functions
1360 that are recorded before enabling thi 1213 that are recorded before enabling this, with
1361 "set_ftrace_filter" otherwise the sys 1214 "set_ftrace_filter" otherwise the system performance
1362 will be critically degraded. Remember 1215 will be critically degraded. Remember to disable
1363 this option before clearing the funct 1216 this option before clearing the function filter.
1364 1217
1365 Options for function_graph tracer: 1218 Options for function_graph tracer:
1366 1219
1367 Since the function_graph tracer has a slight 1220 Since the function_graph tracer has a slightly different output
1368 it has its own options to control what is di 1221 it has its own options to control what is displayed.
1369 1222
1370 funcgraph-overrun 1223 funcgraph-overrun
1371 When set, the "overrun" of the graph 1224 When set, the "overrun" of the graph stack is
1372 displayed after each function traced. 1225 displayed after each function traced. The
1373 overrun, is when the stack depth of t 1226 overrun, is when the stack depth of the calls
1374 is greater than what is reserved for 1227 is greater than what is reserved for each task.
1375 Each task has a fixed array of functi 1228 Each task has a fixed array of functions to
1376 trace in the call graph. If the depth 1229 trace in the call graph. If the depth of the
1377 calls exceeds that, the function is n 1230 calls exceeds that, the function is not traced.
1378 The overrun is the number of function 1231 The overrun is the number of functions missed
1379 due to exceeding this array. 1232 due to exceeding this array.
1380 1233
1381 funcgraph-cpu 1234 funcgraph-cpu
1382 When set, the CPU number of the CPU w 1235 When set, the CPU number of the CPU where the trace
1383 occurred is displayed. 1236 occurred is displayed.
1384 1237
1385 funcgraph-overhead 1238 funcgraph-overhead
1386 When set, if the function takes longe 1239 When set, if the function takes longer than
1387 A certain amount, then a delay marker 1240 A certain amount, then a delay marker is
1388 displayed. See "delay" above, under t 1241 displayed. See "delay" above, under the
1389 header description. 1242 header description.
1390 1243
1391 funcgraph-proc 1244 funcgraph-proc
1392 Unlike other tracers, the process' co 1245 Unlike other tracers, the process' command line
1393 is not displayed by default, but inst 1246 is not displayed by default, but instead only
1394 when a task is traced in and out duri 1247 when a task is traced in and out during a context
1395 switch. Enabling this options has the 1248 switch. Enabling this options has the command
1396 of each process displayed at every li 1249 of each process displayed at every line.
1397 1250
1398 funcgraph-duration 1251 funcgraph-duration
1399 At the end of each function (the retu 1252 At the end of each function (the return)
1400 the duration of the amount of time in 1253 the duration of the amount of time in the
1401 function is displayed in microseconds 1254 function is displayed in microseconds.
1402 1255
1403 funcgraph-abstime 1256 funcgraph-abstime
1404 When set, the timestamp is displayed 1257 When set, the timestamp is displayed at each line.
1405 1258
1406 funcgraph-irqs 1259 funcgraph-irqs
1407 When disabled, functions that happen 1260 When disabled, functions that happen inside an
1408 interrupt will not be traced. 1261 interrupt will not be traced.
1409 1262
1410 funcgraph-tail 1263 funcgraph-tail
1411 When set, the return event will inclu 1264 When set, the return event will include the function
1412 that it represents. By default this i 1265 that it represents. By default this is off, and
1413 only a closing curly bracket "}" is d 1266 only a closing curly bracket "}" is displayed for
1414 the return of a function. 1267 the return of a function.
1415 1268
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 1269 sleep-time
1430 When running function graph tracer, t 1270 When running function graph tracer, to include
1431 the time a task schedules out in its 1271 the time a task schedules out in its function.
1432 When enabled, it will account time th 1272 When enabled, it will account time the task has been
1433 scheduled out as part of the function 1273 scheduled out as part of the function call.
1434 1274
1435 graph-time 1275 graph-time
1436 When running function profiler with f 1276 When running function profiler with function graph tracer,
1437 to include the time to call nested fu 1277 to include the time to call nested functions. When this is
1438 not set, the time reported for the fu 1278 not set, the time reported for the function will only
1439 include the time the function itself 1279 include the time the function itself executed for, not the
1440 time for functions that it called. 1280 time for functions that it called.
1441 1281
1442 Options for blk tracer: 1282 Options for blk tracer:
1443 1283
1444 blk_classic 1284 blk_classic
1445 Shows a more minimalistic output. 1285 Shows a more minimalistic output.
1446 1286
1447 1287
1448 irqsoff 1288 irqsoff
1449 ------- 1289 -------
1450 1290
1451 When interrupts are disabled, the CPU can not 1291 When interrupts are disabled, the CPU can not react to any other
1452 external event (besides NMIs and SMIs). This 1292 external event (besides NMIs and SMIs). This prevents the timer
1453 interrupt from triggering or the mouse interr 1293 interrupt from triggering or the mouse interrupt from letting
1454 the kernel know of a new mouse event. The res 1294 the kernel know of a new mouse event. The result is a latency
1455 with the reaction time. 1295 with the reaction time.
1456 1296
1457 The irqsoff tracer tracks the time for which 1297 The irqsoff tracer tracks the time for which interrupts are
1458 disabled. When a new maximum latency is hit, 1298 disabled. When a new maximum latency is hit, the tracer saves
1459 the trace leading up to that latency point so 1299 the trace leading up to that latency point so that every time a
1460 new maximum is reached, the old saved trace i 1300 new maximum is reached, the old saved trace is discarded and the
1461 new trace is saved. 1301 new trace is saved.
1462 1302
1463 To reset the maximum, echo 0 into tracing_max 1303 To reset the maximum, echo 0 into tracing_max_latency. Here is
1464 an example:: 1304 an example::
1465 1305
1466 # echo 0 > options/function-trace 1306 # echo 0 > options/function-trace
1467 # echo irqsoff > current_tracer 1307 # echo irqsoff > current_tracer
1468 # echo 1 > tracing_on 1308 # echo 1 > tracing_on
1469 # echo 0 > tracing_max_latency 1309 # echo 0 > tracing_max_latency
1470 # ls -ltr 1310 # ls -ltr
1471 [...] 1311 [...]
1472 # echo 0 > tracing_on 1312 # echo 0 > tracing_on
1473 # cat trace 1313 # cat trace
1474 # tracer: irqsoff 1314 # tracer: irqsoff
1475 # 1315 #
1476 # irqsoff latency trace v1.1.5 on 3.8.0-tes 1316 # irqsoff latency trace v1.1.5 on 3.8.0-test+
1477 # ----------------------------------------- 1317 # --------------------------------------------------------------------
1478 # latency: 16 us, #4/4, CPU#0 | (M:preempt 1318 # latency: 16 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1479 # ----------------- 1319 # -----------------
1480 # | task: swapper/0-0 (uid:0 nice:0 poli 1320 # | task: swapper/0-0 (uid:0 nice:0 policy:0 rt_prio:0)
1481 # ----------------- 1321 # -----------------
1482 # => started at: run_timer_softirq 1322 # => started at: run_timer_softirq
1483 # => ended at: run_timer_softirq 1323 # => ended at: run_timer_softirq
1484 # 1324 #
1485 # 1325 #
1486 # _------=> CPU# 1326 # _------=> CPU#
1487 # / _-----=> irqs-off 1327 # / _-----=> irqs-off
1488 # | / _----=> need-resched 1328 # | / _----=> need-resched
1489 # || / _---=> hardirq/softir 1329 # || / _---=> hardirq/softirq
1490 # ||| / _--=> preempt-depth 1330 # ||| / _--=> preempt-depth
1491 # |||| / delay 1331 # |||| / delay
1492 # cmd pid ||||| time | caller 1332 # cmd pid ||||| time | caller
1493 # \ / ||||| \ | / 1333 # \ / ||||| \ | /
1494 <idle>-0 0d.s2 0us+: _raw_spin_l 1334 <idle>-0 0d.s2 0us+: _raw_spin_lock_irq <-run_timer_softirq
1495 <idle>-0 0dNs3 17us : _raw_spin_u 1335 <idle>-0 0dNs3 17us : _raw_spin_unlock_irq <-run_timer_softirq
1496 <idle>-0 0dNs3 17us+: trace_hardi 1336 <idle>-0 0dNs3 17us+: trace_hardirqs_on <-run_timer_softirq
1497 <idle>-0 0dNs3 25us : <stack trac 1337 <idle>-0 0dNs3 25us : <stack trace>
1498 => _raw_spin_unlock_irq 1338 => _raw_spin_unlock_irq
1499 => run_timer_softirq 1339 => run_timer_softirq
1500 => __do_softirq 1340 => __do_softirq
1501 => call_softirq 1341 => call_softirq
1502 => do_softirq 1342 => do_softirq
1503 => irq_exit 1343 => irq_exit
1504 => smp_apic_timer_interrupt 1344 => smp_apic_timer_interrupt
1505 => apic_timer_interrupt 1345 => apic_timer_interrupt
1506 => rcu_idle_exit 1346 => rcu_idle_exit
1507 => cpu_idle 1347 => cpu_idle
1508 => rest_init 1348 => rest_init
1509 => start_kernel 1349 => start_kernel
1510 => x86_64_start_reservations 1350 => x86_64_start_reservations
1511 => x86_64_start_kernel 1351 => x86_64_start_kernel
1512 1352
1513 Here we see that we had a latency of 16 micro !! 1353 Here we see that that we had a latency of 16 microseconds (which is
1514 very good). The _raw_spin_lock_irq in run_tim 1354 very good). The _raw_spin_lock_irq in run_timer_softirq disabled
1515 interrupts. The difference between the 16 and 1355 interrupts. The difference between the 16 and the displayed
1516 timestamp 25us occurred because the clock was 1356 timestamp 25us occurred because the clock was incremented
1517 between the time of recording the max latency 1357 between the time of recording the max latency and the time of
1518 recording the function that had that latency. 1358 recording the function that had that latency.
1519 1359
1520 Note the above example had function-trace not 1360 Note the above example had function-trace not set. If we set
1521 function-trace, we get a much larger output:: 1361 function-trace, we get a much larger output::
1522 1362
1523 with echo 1 > options/function-trace 1363 with echo 1 > options/function-trace
1524 1364
1525 # tracer: irqsoff 1365 # tracer: irqsoff
1526 # 1366 #
1527 # irqsoff latency trace v1.1.5 on 3.8.0-tes 1367 # irqsoff latency trace v1.1.5 on 3.8.0-test+
1528 # ----------------------------------------- 1368 # --------------------------------------------------------------------
1529 # latency: 71 us, #168/168, CPU#3 | (M:pree 1369 # latency: 71 us, #168/168, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1530 # ----------------- 1370 # -----------------
1531 # | task: bash-2042 (uid:0 nice:0 policy 1371 # | task: bash-2042 (uid:0 nice:0 policy:0 rt_prio:0)
1532 # ----------------- 1372 # -----------------
1533 # => started at: ata_scsi_queuecmd 1373 # => started at: ata_scsi_queuecmd
1534 # => ended at: ata_scsi_queuecmd 1374 # => ended at: ata_scsi_queuecmd
1535 # 1375 #
1536 # 1376 #
1537 # _------=> CPU# 1377 # _------=> CPU#
1538 # / _-----=> irqs-off 1378 # / _-----=> irqs-off
1539 # | / _----=> need-resched 1379 # | / _----=> need-resched
1540 # || / _---=> hardirq/softir 1380 # || / _---=> hardirq/softirq
1541 # ||| / _--=> preempt-depth 1381 # ||| / _--=> preempt-depth
1542 # |||| / delay 1382 # |||| / delay
1543 # cmd pid ||||| time | caller 1383 # cmd pid ||||| time | caller
1544 # \ / ||||| \ | / 1384 # \ / ||||| \ | /
1545 bash-2042 3d... 0us : _raw_spin_l 1385 bash-2042 3d... 0us : _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1546 bash-2042 3d... 0us : add_preempt 1386 bash-2042 3d... 0us : add_preempt_count <-_raw_spin_lock_irqsave
1547 bash-2042 3d..1 1us : ata_scsi_fi 1387 bash-2042 3d..1 1us : ata_scsi_find_dev <-ata_scsi_queuecmd
1548 bash-2042 3d..1 1us : __ata_scsi_ 1388 bash-2042 3d..1 1us : __ata_scsi_find_dev <-ata_scsi_find_dev
1549 bash-2042 3d..1 2us : ata_find_de 1389 bash-2042 3d..1 2us : ata_find_dev.part.14 <-__ata_scsi_find_dev
1550 bash-2042 3d..1 2us : ata_qc_new_ 1390 bash-2042 3d..1 2us : ata_qc_new_init <-__ata_scsi_queuecmd
1551 bash-2042 3d..1 3us : ata_sg_init 1391 bash-2042 3d..1 3us : ata_sg_init <-__ata_scsi_queuecmd
1552 bash-2042 3d..1 4us : ata_scsi_rw 1392 bash-2042 3d..1 4us : ata_scsi_rw_xlat <-__ata_scsi_queuecmd
1553 bash-2042 3d..1 4us : ata_build_r 1393 bash-2042 3d..1 4us : ata_build_rw_tf <-ata_scsi_rw_xlat
1554 [...] 1394 [...]
1555 bash-2042 3d..1 67us : delay_tsc < 1395 bash-2042 3d..1 67us : delay_tsc <-__delay
1556 bash-2042 3d..1 67us : add_preempt 1396 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1557 bash-2042 3d..2 67us : sub_preempt 1397 bash-2042 3d..2 67us : sub_preempt_count <-delay_tsc
1558 bash-2042 3d..1 67us : add_preempt 1398 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1559 bash-2042 3d..2 68us : sub_preempt 1399 bash-2042 3d..2 68us : sub_preempt_count <-delay_tsc
1560 bash-2042 3d..1 68us+: ata_bmdma_s 1400 bash-2042 3d..1 68us+: ata_bmdma_start <-ata_bmdma_qc_issue
1561 bash-2042 3d..1 71us : _raw_spin_u 1401 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1562 bash-2042 3d..1 71us : _raw_spin_u 1402 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1563 bash-2042 3d..1 72us+: trace_hardi 1403 bash-2042 3d..1 72us+: trace_hardirqs_on <-ata_scsi_queuecmd
1564 bash-2042 3d..1 120us : <stack trac 1404 bash-2042 3d..1 120us : <stack trace>
1565 => _raw_spin_unlock_irqrestore 1405 => _raw_spin_unlock_irqrestore
1566 => ata_scsi_queuecmd 1406 => ata_scsi_queuecmd
1567 => scsi_dispatch_cmd 1407 => scsi_dispatch_cmd
1568 => scsi_request_fn 1408 => scsi_request_fn
1569 => __blk_run_queue_uncond 1409 => __blk_run_queue_uncond
1570 => __blk_run_queue 1410 => __blk_run_queue
1571 => blk_queue_bio 1411 => blk_queue_bio
1572 => submit_bio_noacct !! 1412 => generic_make_request
1573 => submit_bio 1413 => submit_bio
1574 => submit_bh 1414 => submit_bh
1575 => __ext3_get_inode_loc 1415 => __ext3_get_inode_loc
1576 => ext3_iget 1416 => ext3_iget
1577 => ext3_lookup 1417 => ext3_lookup
1578 => lookup_real 1418 => lookup_real
1579 => __lookup_hash 1419 => __lookup_hash
1580 => walk_component 1420 => walk_component
1581 => lookup_last 1421 => lookup_last
1582 => path_lookupat 1422 => path_lookupat
1583 => filename_lookup 1423 => filename_lookup
1584 => user_path_at_empty 1424 => user_path_at_empty
1585 => user_path_at 1425 => user_path_at
1586 => vfs_fstatat 1426 => vfs_fstatat
1587 => vfs_stat 1427 => vfs_stat
1588 => sys_newstat 1428 => sys_newstat
1589 => system_call_fastpath 1429 => system_call_fastpath
1590 1430
1591 1431
1592 Here we traced a 71 microsecond latency. But 1432 Here we traced a 71 microsecond latency. But we also see all the
1593 functions that were called during that time. 1433 functions that were called during that time. Note that by
1594 enabling function tracing, we incur an added 1434 enabling function tracing, we incur an added overhead. This
1595 overhead may extend the latency times. But ne 1435 overhead may extend the latency times. But nevertheless, this
1596 trace has provided some very helpful debuggin 1436 trace has provided some very helpful debugging information.
1597 1437
1598 If we prefer function graph output instead of 1438 If we prefer function graph output instead of function, we can set
1599 display-graph option:: 1439 display-graph option::
1600 1440
1601 with echo 1 > options/display-graph 1441 with echo 1 > options/display-graph
1602 1442
1603 # tracer: irqsoff 1443 # tracer: irqsoff
1604 # 1444 #
1605 # irqsoff latency trace v1.1.5 on 4.20.0-rc 1445 # irqsoff latency trace v1.1.5 on 4.20.0-rc6+
1606 # ----------------------------------------- 1446 # --------------------------------------------------------------------
1607 # latency: 3751 us, #274/274, CPU#0 | (M:de 1447 # latency: 3751 us, #274/274, CPU#0 | (M:desktop VP:0, KP:0, SP:0 HP:0 #P:4)
1608 # ----------------- 1448 # -----------------
1609 # | task: bash-1507 (uid:0 nice:0 policy 1449 # | task: bash-1507 (uid:0 nice:0 policy:0 rt_prio:0)
1610 # ----------------- 1450 # -----------------
1611 # => started at: free_debug_processing 1451 # => started at: free_debug_processing
1612 # => ended at: return_to_handler 1452 # => ended at: return_to_handler
1613 # 1453 #
1614 # 1454 #
1615 # _-- 1455 # _-----=> irqs-off
1616 # / _- 1456 # / _----=> need-resched
1617 # | / _ 1457 # | / _---=> hardirq/softirq
1618 # || / 1458 # || / _--=> preempt-depth
1619 # ||| / 1459 # ||| /
1620 # REL TIME CPU TASK/PID |||| 1460 # REL TIME CPU TASK/PID |||| DURATION FUNCTION CALLS
1621 # | | | | |||| 1461 # | | | | |||| | | | | | |
1622 0 us | 0) bash-1507 | d... 1462 0 us | 0) bash-1507 | d... | 0.000 us | _raw_spin_lock_irqsave();
1623 0 us | 0) bash-1507 | d..1 1463 0 us | 0) bash-1507 | d..1 | 0.378 us | do_raw_spin_trylock();
1624 1 us | 0) bash-1507 | d..2 1464 1 us | 0) bash-1507 | d..2 | | set_track() {
1625 2 us | 0) bash-1507 | d..2 1465 2 us | 0) bash-1507 | d..2 | | save_stack_trace() {
1626 2 us | 0) bash-1507 | d..2 1466 2 us | 0) bash-1507 | d..2 | | __save_stack_trace() {
1627 3 us | 0) bash-1507 | d..2 1467 3 us | 0) bash-1507 | d..2 | | __unwind_start() {
1628 3 us | 0) bash-1507 | d..2 1468 3 us | 0) bash-1507 | d..2 | | get_stack_info() {
1629 3 us | 0) bash-1507 | d..2 1469 3 us | 0) bash-1507 | d..2 | 0.351 us | in_task_stack();
1630 4 us | 0) bash-1507 | d..2 1470 4 us | 0) bash-1507 | d..2 | 1.107 us | }
1631 [...] 1471 [...]
1632 3750 us | 0) bash-1507 | d..1 1472 3750 us | 0) bash-1507 | d..1 | 0.516 us | do_raw_spin_unlock();
1633 3750 us | 0) bash-1507 | d..1 1473 3750 us | 0) bash-1507 | d..1 | 0.000 us | _raw_spin_unlock_irqrestore();
1634 3764 us | 0) bash-1507 | d..1 1474 3764 us | 0) bash-1507 | d..1 | 0.000 us | tracer_hardirqs_on();
1635 bash-1507 0d..1 3792us : <stack trac 1475 bash-1507 0d..1 3792us : <stack trace>
1636 => free_debug_processing 1476 => free_debug_processing
1637 => __slab_free 1477 => __slab_free
1638 => kmem_cache_free 1478 => kmem_cache_free
1639 => vm_area_free 1479 => vm_area_free
1640 => remove_vma 1480 => remove_vma
1641 => exit_mmap 1481 => exit_mmap
1642 => mmput 1482 => mmput
1643 => begin_new_exec !! 1483 => flush_old_exec
1644 => load_elf_binary 1484 => load_elf_binary
1645 => search_binary_handler 1485 => search_binary_handler
1646 => __do_execve_file.isra.32 1486 => __do_execve_file.isra.32
1647 => __x64_sys_execve 1487 => __x64_sys_execve
1648 => do_syscall_64 1488 => do_syscall_64
1649 => entry_SYSCALL_64_after_hwframe 1489 => entry_SYSCALL_64_after_hwframe
1650 1490
1651 preemptoff 1491 preemptoff
1652 ---------- 1492 ----------
1653 1493
1654 When preemption is disabled, we may be able t 1494 When preemption is disabled, we may be able to receive
1655 interrupts but the task cannot be preempted a 1495 interrupts but the task cannot be preempted and a higher
1656 priority task must wait for preemption to be 1496 priority task must wait for preemption to be enabled again
1657 before it can preempt a lower priority task. 1497 before it can preempt a lower priority task.
1658 1498
1659 The preemptoff tracer traces the places that 1499 The preemptoff tracer traces the places that disable preemption.
1660 Like the irqsoff tracer, it records the maxim 1500 Like the irqsoff tracer, it records the maximum latency for
1661 which preemption was disabled. The control of 1501 which preemption was disabled. The control of preemptoff tracer
1662 is much like the irqsoff tracer. 1502 is much like the irqsoff tracer.
1663 :: 1503 ::
1664 1504
1665 # echo 0 > options/function-trace 1505 # echo 0 > options/function-trace
1666 # echo preemptoff > current_tracer 1506 # echo preemptoff > current_tracer
1667 # echo 1 > tracing_on 1507 # echo 1 > tracing_on
1668 # echo 0 > tracing_max_latency 1508 # echo 0 > tracing_max_latency
1669 # ls -ltr 1509 # ls -ltr
1670 [...] 1510 [...]
1671 # echo 0 > tracing_on 1511 # echo 0 > tracing_on
1672 # cat trace 1512 # cat trace
1673 # tracer: preemptoff 1513 # tracer: preemptoff
1674 # 1514 #
1675 # preemptoff latency trace v1.1.5 on 3.8.0- 1515 # preemptoff latency trace v1.1.5 on 3.8.0-test+
1676 # ----------------------------------------- 1516 # --------------------------------------------------------------------
1677 # latency: 46 us, #4/4, CPU#1 | (M:preempt 1517 # latency: 46 us, #4/4, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1678 # ----------------- 1518 # -----------------
1679 # | task: sshd-1991 (uid:0 nice:0 policy 1519 # | task: sshd-1991 (uid:0 nice:0 policy:0 rt_prio:0)
1680 # ----------------- 1520 # -----------------
1681 # => started at: do_IRQ 1521 # => started at: do_IRQ
1682 # => ended at: do_IRQ 1522 # => ended at: do_IRQ
1683 # 1523 #
1684 # 1524 #
1685 # _------=> CPU# 1525 # _------=> CPU#
1686 # / _-----=> irqs-off 1526 # / _-----=> irqs-off
1687 # | / _----=> need-resched 1527 # | / _----=> need-resched
1688 # || / _---=> hardirq/softir 1528 # || / _---=> hardirq/softirq
1689 # ||| / _--=> preempt-depth 1529 # ||| / _--=> preempt-depth
1690 # |||| / delay 1530 # |||| / delay
1691 # cmd pid ||||| time | caller 1531 # cmd pid ||||| time | caller
1692 # \ / ||||| \ | / 1532 # \ / ||||| \ | /
1693 sshd-1991 1d.h. 0us+: irq_enter < 1533 sshd-1991 1d.h. 0us+: irq_enter <-do_IRQ
1694 sshd-1991 1d..1 46us : irq_exit <- 1534 sshd-1991 1d..1 46us : irq_exit <-do_IRQ
1695 sshd-1991 1d..1 47us+: trace_preem 1535 sshd-1991 1d..1 47us+: trace_preempt_on <-do_IRQ
1696 sshd-1991 1d..1 52us : <stack trac 1536 sshd-1991 1d..1 52us : <stack trace>
1697 => sub_preempt_count 1537 => sub_preempt_count
1698 => irq_exit 1538 => irq_exit
1699 => do_IRQ 1539 => do_IRQ
1700 => ret_from_intr 1540 => ret_from_intr
1701 1541
1702 1542
1703 This has some more changes. Preemption was di 1543 This has some more changes. Preemption was disabled when an
1704 interrupt came in (notice the 'h'), and was e 1544 interrupt came in (notice the 'h'), and was enabled on exit.
1705 But we also see that interrupts have been dis 1545 But we also see that interrupts have been disabled when entering
1706 the preempt off section and leaving it (the ' 1546 the preempt off section and leaving it (the 'd'). We do not know if
1707 interrupts were enabled in the mean time or s 1547 interrupts were enabled in the mean time or shortly after this
1708 was over. 1548 was over.
1709 :: 1549 ::
1710 1550
1711 # tracer: preemptoff 1551 # tracer: preemptoff
1712 # 1552 #
1713 # preemptoff latency trace v1.1.5 on 3.8.0- 1553 # preemptoff latency trace v1.1.5 on 3.8.0-test+
1714 # ----------------------------------------- 1554 # --------------------------------------------------------------------
1715 # latency: 83 us, #241/241, CPU#1 | (M:pree 1555 # latency: 83 us, #241/241, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1716 # ----------------- 1556 # -----------------
1717 # | task: bash-1994 (uid:0 nice:0 policy 1557 # | task: bash-1994 (uid:0 nice:0 policy:0 rt_prio:0)
1718 # ----------------- 1558 # -----------------
1719 # => started at: wake_up_new_task 1559 # => started at: wake_up_new_task
1720 # => ended at: task_rq_unlock 1560 # => ended at: task_rq_unlock
1721 # 1561 #
1722 # 1562 #
1723 # _------=> CPU# 1563 # _------=> CPU#
1724 # / _-----=> irqs-off 1564 # / _-----=> irqs-off
1725 # | / _----=> need-resched 1565 # | / _----=> need-resched
1726 # || / _---=> hardirq/softir 1566 # || / _---=> hardirq/softirq
1727 # ||| / _--=> preempt-depth 1567 # ||| / _--=> preempt-depth
1728 # |||| / delay 1568 # |||| / delay
1729 # cmd pid ||||| time | caller 1569 # cmd pid ||||| time | caller
1730 # \ / ||||| \ | / 1570 # \ / ||||| \ | /
1731 bash-1994 1d..1 0us : _raw_spin_l 1571 bash-1994 1d..1 0us : _raw_spin_lock_irqsave <-wake_up_new_task
1732 bash-1994 1d..1 0us : select_task 1572 bash-1994 1d..1 0us : select_task_rq_fair <-select_task_rq
1733 bash-1994 1d..1 1us : __rcu_read_ 1573 bash-1994 1d..1 1us : __rcu_read_lock <-select_task_rq_fair
1734 bash-1994 1d..1 1us : source_load 1574 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
1735 bash-1994 1d..1 1us : source_load 1575 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
1736 [...] 1576 [...]
1737 bash-1994 1d..1 12us : irq_enter < 1577 bash-1994 1d..1 12us : irq_enter <-smp_apic_timer_interrupt
1738 bash-1994 1d..1 12us : rcu_irq_ent 1578 bash-1994 1d..1 12us : rcu_irq_enter <-irq_enter
1739 bash-1994 1d..1 13us : add_preempt 1579 bash-1994 1d..1 13us : add_preempt_count <-irq_enter
1740 bash-1994 1d.h1 13us : exit_idle < 1580 bash-1994 1d.h1 13us : exit_idle <-smp_apic_timer_interrupt
1741 bash-1994 1d.h1 13us : hrtimer_int 1581 bash-1994 1d.h1 13us : hrtimer_interrupt <-smp_apic_timer_interrupt
1742 bash-1994 1d.h1 13us : _raw_spin_l 1582 bash-1994 1d.h1 13us : _raw_spin_lock <-hrtimer_interrupt
1743 bash-1994 1d.h1 14us : add_preempt 1583 bash-1994 1d.h1 14us : add_preempt_count <-_raw_spin_lock
1744 bash-1994 1d.h2 14us : ktime_get_u 1584 bash-1994 1d.h2 14us : ktime_get_update_offsets <-hrtimer_interrupt
1745 [...] 1585 [...]
1746 bash-1994 1d.h1 35us : lapic_next_ 1586 bash-1994 1d.h1 35us : lapic_next_event <-clockevents_program_event
1747 bash-1994 1d.h1 35us : irq_exit <- 1587 bash-1994 1d.h1 35us : irq_exit <-smp_apic_timer_interrupt
1748 bash-1994 1d.h1 36us : sub_preempt 1588 bash-1994 1d.h1 36us : sub_preempt_count <-irq_exit
1749 bash-1994 1d..2 36us : do_softirq 1589 bash-1994 1d..2 36us : do_softirq <-irq_exit
1750 bash-1994 1d..2 36us : __do_softir 1590 bash-1994 1d..2 36us : __do_softirq <-call_softirq
1751 bash-1994 1d..2 36us : __local_bh_ 1591 bash-1994 1d..2 36us : __local_bh_disable <-__do_softirq
1752 bash-1994 1d.s2 37us : add_preempt 1592 bash-1994 1d.s2 37us : add_preempt_count <-_raw_spin_lock_irq
1753 bash-1994 1d.s3 38us : _raw_spin_u 1593 bash-1994 1d.s3 38us : _raw_spin_unlock <-run_timer_softirq
1754 bash-1994 1d.s3 39us : sub_preempt 1594 bash-1994 1d.s3 39us : sub_preempt_count <-_raw_spin_unlock
1755 bash-1994 1d.s2 39us : call_timer_ 1595 bash-1994 1d.s2 39us : call_timer_fn <-run_timer_softirq
1756 [...] 1596 [...]
1757 bash-1994 1dNs2 81us : cpu_needs_a 1597 bash-1994 1dNs2 81us : cpu_needs_another_gp <-rcu_process_callbacks
1758 bash-1994 1dNs2 82us : __local_bh_ 1598 bash-1994 1dNs2 82us : __local_bh_enable <-__do_softirq
1759 bash-1994 1dNs2 82us : sub_preempt 1599 bash-1994 1dNs2 82us : sub_preempt_count <-__local_bh_enable
1760 bash-1994 1dN.2 82us : idle_cpu <- 1600 bash-1994 1dN.2 82us : idle_cpu <-irq_exit
1761 bash-1994 1dN.2 83us : rcu_irq_exi 1601 bash-1994 1dN.2 83us : rcu_irq_exit <-irq_exit
1762 bash-1994 1dN.2 83us : sub_preempt 1602 bash-1994 1dN.2 83us : sub_preempt_count <-irq_exit
1763 bash-1994 1.N.1 84us : _raw_spin_u 1603 bash-1994 1.N.1 84us : _raw_spin_unlock_irqrestore <-task_rq_unlock
1764 bash-1994 1.N.1 84us+: trace_preem 1604 bash-1994 1.N.1 84us+: trace_preempt_on <-task_rq_unlock
1765 bash-1994 1.N.1 104us : <stack trac 1605 bash-1994 1.N.1 104us : <stack trace>
1766 => sub_preempt_count 1606 => sub_preempt_count
1767 => _raw_spin_unlock_irqrestore 1607 => _raw_spin_unlock_irqrestore
1768 => task_rq_unlock 1608 => task_rq_unlock
1769 => wake_up_new_task 1609 => wake_up_new_task
1770 => do_fork 1610 => do_fork
1771 => sys_clone 1611 => sys_clone
1772 => stub_clone 1612 => stub_clone
1773 1613
1774 1614
1775 The above is an example of the preemptoff tra 1615 The above is an example of the preemptoff trace with
1776 function-trace set. Here we see that interrup 1616 function-trace set. Here we see that interrupts were not disabled
1777 the entire time. The irq_enter code lets us k 1617 the entire time. The irq_enter code lets us know that we entered
1778 an interrupt 'h'. Before that, the functions 1618 an interrupt 'h'. Before that, the functions being traced still
1779 show that it is not in an interrupt, but we c 1619 show that it is not in an interrupt, but we can see from the
1780 functions themselves that this is not the cas 1620 functions themselves that this is not the case.
1781 1621
1782 preemptirqsoff 1622 preemptirqsoff
1783 -------------- 1623 --------------
1784 1624
1785 Knowing the locations that have interrupts di 1625 Knowing the locations that have interrupts disabled or
1786 preemption disabled for the longest times is 1626 preemption disabled for the longest times is helpful. But
1787 sometimes we would like to know when either p 1627 sometimes we would like to know when either preemption and/or
1788 interrupts are disabled. 1628 interrupts are disabled.
1789 1629
1790 Consider the following code:: 1630 Consider the following code::
1791 1631
1792 local_irq_disable(); 1632 local_irq_disable();
1793 call_function_with_irqs_off(); 1633 call_function_with_irqs_off();
1794 preempt_disable(); 1634 preempt_disable();
1795 call_function_with_irqs_and_preemption_of 1635 call_function_with_irqs_and_preemption_off();
1796 local_irq_enable(); 1636 local_irq_enable();
1797 call_function_with_preemption_off(); 1637 call_function_with_preemption_off();
1798 preempt_enable(); 1638 preempt_enable();
1799 1639
1800 The irqsoff tracer will record the total leng 1640 The irqsoff tracer will record the total length of
1801 call_function_with_irqs_off() and 1641 call_function_with_irqs_off() and
1802 call_function_with_irqs_and_preemption_off(). 1642 call_function_with_irqs_and_preemption_off().
1803 1643
1804 The preemptoff tracer will record the total l 1644 The preemptoff tracer will record the total length of
1805 call_function_with_irqs_and_preemption_off() 1645 call_function_with_irqs_and_preemption_off() and
1806 call_function_with_preemption_off(). 1646 call_function_with_preemption_off().
1807 1647
1808 But neither will trace the time that interrup 1648 But neither will trace the time that interrupts and/or
1809 preemption is disabled. This total time is th 1649 preemption is disabled. This total time is the time that we can
1810 not schedule. To record this time, use the pr 1650 not schedule. To record this time, use the preemptirqsoff
1811 tracer. 1651 tracer.
1812 1652
1813 Again, using this trace is much like the irqs 1653 Again, using this trace is much like the irqsoff and preemptoff
1814 tracers. 1654 tracers.
1815 :: 1655 ::
1816 1656
1817 # echo 0 > options/function-trace 1657 # echo 0 > options/function-trace
1818 # echo preemptirqsoff > current_tracer 1658 # echo preemptirqsoff > current_tracer
1819 # echo 1 > tracing_on 1659 # echo 1 > tracing_on
1820 # echo 0 > tracing_max_latency 1660 # echo 0 > tracing_max_latency
1821 # ls -ltr 1661 # ls -ltr
1822 [...] 1662 [...]
1823 # echo 0 > tracing_on 1663 # echo 0 > tracing_on
1824 # cat trace 1664 # cat trace
1825 # tracer: preemptirqsoff 1665 # tracer: preemptirqsoff
1826 # 1666 #
1827 # preemptirqsoff latency trace v1.1.5 on 3. 1667 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1828 # ----------------------------------------- 1668 # --------------------------------------------------------------------
1829 # latency: 100 us, #4/4, CPU#3 | (M:preempt 1669 # latency: 100 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1830 # ----------------- 1670 # -----------------
1831 # | task: ls-2230 (uid:0 nice:0 policy:0 1671 # | task: ls-2230 (uid:0 nice:0 policy:0 rt_prio:0)
1832 # ----------------- 1672 # -----------------
1833 # => started at: ata_scsi_queuecmd 1673 # => started at: ata_scsi_queuecmd
1834 # => ended at: ata_scsi_queuecmd 1674 # => ended at: ata_scsi_queuecmd
1835 # 1675 #
1836 # 1676 #
1837 # _------=> CPU# 1677 # _------=> CPU#
1838 # / _-----=> irqs-off 1678 # / _-----=> irqs-off
1839 # | / _----=> need-resched 1679 # | / _----=> need-resched
1840 # || / _---=> hardirq/softir 1680 # || / _---=> hardirq/softirq
1841 # ||| / _--=> preempt-depth 1681 # ||| / _--=> preempt-depth
1842 # |||| / delay 1682 # |||| / delay
1843 # cmd pid ||||| time | caller 1683 # cmd pid ||||| time | caller
1844 # \ / ||||| \ | / 1684 # \ / ||||| \ | /
1845 ls-2230 3d... 0us+: _raw_spin_l 1685 ls-2230 3d... 0us+: _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1846 ls-2230 3...1 100us : _raw_spin_u 1686 ls-2230 3...1 100us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1847 ls-2230 3...1 101us+: trace_preem 1687 ls-2230 3...1 101us+: trace_preempt_on <-ata_scsi_queuecmd
1848 ls-2230 3...1 111us : <stack trac 1688 ls-2230 3...1 111us : <stack trace>
1849 => sub_preempt_count 1689 => sub_preempt_count
1850 => _raw_spin_unlock_irqrestore 1690 => _raw_spin_unlock_irqrestore
1851 => ata_scsi_queuecmd 1691 => ata_scsi_queuecmd
1852 => scsi_dispatch_cmd 1692 => scsi_dispatch_cmd
1853 => scsi_request_fn 1693 => scsi_request_fn
1854 => __blk_run_queue_uncond 1694 => __blk_run_queue_uncond
1855 => __blk_run_queue 1695 => __blk_run_queue
1856 => blk_queue_bio 1696 => blk_queue_bio
1857 => submit_bio_noacct !! 1697 => generic_make_request
1858 => submit_bio 1698 => submit_bio
1859 => submit_bh 1699 => submit_bh
1860 => ext3_bread 1700 => ext3_bread
1861 => ext3_dir_bread 1701 => ext3_dir_bread
1862 => htree_dirblock_to_tree 1702 => htree_dirblock_to_tree
1863 => ext3_htree_fill_tree 1703 => ext3_htree_fill_tree
1864 => ext3_readdir 1704 => ext3_readdir
1865 => vfs_readdir 1705 => vfs_readdir
1866 => sys_getdents 1706 => sys_getdents
1867 => system_call_fastpath 1707 => system_call_fastpath
1868 1708
1869 1709
1870 The trace_hardirqs_off_thunk is called from a 1710 The trace_hardirqs_off_thunk is called from assembly on x86 when
1871 interrupts are disabled in the assembly code. 1711 interrupts are disabled in the assembly code. Without the
1872 function tracing, we do not know if interrupt 1712 function tracing, we do not know if interrupts were enabled
1873 within the preemption points. We do see that 1713 within the preemption points. We do see that it started with
1874 preemption enabled. 1714 preemption enabled.
1875 1715
1876 Here is a trace with function-trace set:: 1716 Here is a trace with function-trace set::
1877 1717
1878 # tracer: preemptirqsoff 1718 # tracer: preemptirqsoff
1879 # 1719 #
1880 # preemptirqsoff latency trace v1.1.5 on 3. 1720 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1881 # ----------------------------------------- 1721 # --------------------------------------------------------------------
1882 # latency: 161 us, #339/339, CPU#3 | (M:pre 1722 # latency: 161 us, #339/339, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1883 # ----------------- 1723 # -----------------
1884 # | task: ls-2269 (uid:0 nice:0 policy:0 1724 # | task: ls-2269 (uid:0 nice:0 policy:0 rt_prio:0)
1885 # ----------------- 1725 # -----------------
1886 # => started at: schedule 1726 # => started at: schedule
1887 # => ended at: mutex_unlock 1727 # => ended at: mutex_unlock
1888 # 1728 #
1889 # 1729 #
1890 # _------=> CPU# 1730 # _------=> CPU#
1891 # / _-----=> irqs-off 1731 # / _-----=> irqs-off
1892 # | / _----=> need-resched 1732 # | / _----=> need-resched
1893 # || / _---=> hardirq/softir 1733 # || / _---=> hardirq/softirq
1894 # ||| / _--=> preempt-depth 1734 # ||| / _--=> preempt-depth
1895 # |||| / delay 1735 # |||| / delay
1896 # cmd pid ||||| time | caller 1736 # cmd pid ||||| time | caller
1897 # \ / ||||| \ | / 1737 # \ / ||||| \ | /
1898 kworker/-59 3...1 0us : __schedule 1738 kworker/-59 3...1 0us : __schedule <-schedule
1899 kworker/-59 3d..1 0us : rcu_preempt 1739 kworker/-59 3d..1 0us : rcu_preempt_qs <-rcu_note_context_switch
1900 kworker/-59 3d..1 1us : add_preempt 1740 kworker/-59 3d..1 1us : add_preempt_count <-_raw_spin_lock_irq
1901 kworker/-59 3d..2 1us : deactivate_ 1741 kworker/-59 3d..2 1us : deactivate_task <-__schedule
1902 kworker/-59 3d..2 1us : dequeue_tas 1742 kworker/-59 3d..2 1us : dequeue_task <-deactivate_task
1903 kworker/-59 3d..2 2us : update_rq_c 1743 kworker/-59 3d..2 2us : update_rq_clock <-dequeue_task
1904 kworker/-59 3d..2 2us : dequeue_tas 1744 kworker/-59 3d..2 2us : dequeue_task_fair <-dequeue_task
1905 kworker/-59 3d..2 2us : update_curr 1745 kworker/-59 3d..2 2us : update_curr <-dequeue_task_fair
1906 kworker/-59 3d..2 2us : update_min_ 1746 kworker/-59 3d..2 2us : update_min_vruntime <-update_curr
1907 kworker/-59 3d..2 3us : cpuacct_cha 1747 kworker/-59 3d..2 3us : cpuacct_charge <-update_curr
1908 kworker/-59 3d..2 3us : __rcu_read_ 1748 kworker/-59 3d..2 3us : __rcu_read_lock <-cpuacct_charge
1909 kworker/-59 3d..2 3us : __rcu_read_ 1749 kworker/-59 3d..2 3us : __rcu_read_unlock <-cpuacct_charge
1910 kworker/-59 3d..2 3us : update_cfs_ 1750 kworker/-59 3d..2 3us : update_cfs_rq_blocked_load <-dequeue_task_fair
1911 kworker/-59 3d..2 4us : clear_buddi 1751 kworker/-59 3d..2 4us : clear_buddies <-dequeue_task_fair
1912 kworker/-59 3d..2 4us : account_ent 1752 kworker/-59 3d..2 4us : account_entity_dequeue <-dequeue_task_fair
1913 kworker/-59 3d..2 4us : update_min_ 1753 kworker/-59 3d..2 4us : update_min_vruntime <-dequeue_task_fair
1914 kworker/-59 3d..2 4us : update_cfs_ 1754 kworker/-59 3d..2 4us : update_cfs_shares <-dequeue_task_fair
1915 kworker/-59 3d..2 5us : hrtick_upda 1755 kworker/-59 3d..2 5us : hrtick_update <-dequeue_task_fair
1916 kworker/-59 3d..2 5us : wq_worker_s 1756 kworker/-59 3d..2 5us : wq_worker_sleeping <-__schedule
1917 kworker/-59 3d..2 5us : kthread_dat 1757 kworker/-59 3d..2 5us : kthread_data <-wq_worker_sleeping
1918 kworker/-59 3d..2 5us : put_prev_ta 1758 kworker/-59 3d..2 5us : put_prev_task_fair <-__schedule
1919 kworker/-59 3d..2 6us : pick_next_t 1759 kworker/-59 3d..2 6us : pick_next_task_fair <-pick_next_task
1920 kworker/-59 3d..2 6us : clear_buddi 1760 kworker/-59 3d..2 6us : clear_buddies <-pick_next_task_fair
1921 kworker/-59 3d..2 6us : set_next_en 1761 kworker/-59 3d..2 6us : set_next_entity <-pick_next_task_fair
1922 kworker/-59 3d..2 6us : update_stat 1762 kworker/-59 3d..2 6us : update_stats_wait_end <-set_next_entity
1923 ls-2269 3d..2 7us : finish_task 1763 ls-2269 3d..2 7us : finish_task_switch <-__schedule
1924 ls-2269 3d..2 7us : _raw_spin_u 1764 ls-2269 3d..2 7us : _raw_spin_unlock_irq <-finish_task_switch
1925 ls-2269 3d..2 8us : do_IRQ <-re 1765 ls-2269 3d..2 8us : do_IRQ <-ret_from_intr
1926 ls-2269 3d..2 8us : irq_enter < 1766 ls-2269 3d..2 8us : irq_enter <-do_IRQ
1927 ls-2269 3d..2 8us : rcu_irq_ent 1767 ls-2269 3d..2 8us : rcu_irq_enter <-irq_enter
1928 ls-2269 3d..2 9us : add_preempt 1768 ls-2269 3d..2 9us : add_preempt_count <-irq_enter
1929 ls-2269 3d.h2 9us : exit_idle < 1769 ls-2269 3d.h2 9us : exit_idle <-do_IRQ
1930 [...] 1770 [...]
1931 ls-2269 3d.h3 20us : sub_preempt 1771 ls-2269 3d.h3 20us : sub_preempt_count <-_raw_spin_unlock
1932 ls-2269 3d.h2 20us : irq_exit <- 1772 ls-2269 3d.h2 20us : irq_exit <-do_IRQ
1933 ls-2269 3d.h2 21us : sub_preempt 1773 ls-2269 3d.h2 21us : sub_preempt_count <-irq_exit
1934 ls-2269 3d..3 21us : do_softirq 1774 ls-2269 3d..3 21us : do_softirq <-irq_exit
1935 ls-2269 3d..3 21us : __do_softir 1775 ls-2269 3d..3 21us : __do_softirq <-call_softirq
1936 ls-2269 3d..3 21us+: __local_bh_ 1776 ls-2269 3d..3 21us+: __local_bh_disable <-__do_softirq
1937 ls-2269 3d.s4 29us : sub_preempt 1777 ls-2269 3d.s4 29us : sub_preempt_count <-_local_bh_enable_ip
1938 ls-2269 3d.s5 29us : sub_preempt 1778 ls-2269 3d.s5 29us : sub_preempt_count <-_local_bh_enable_ip
1939 ls-2269 3d.s5 31us : do_IRQ <-re 1779 ls-2269 3d.s5 31us : do_IRQ <-ret_from_intr
1940 ls-2269 3d.s5 31us : irq_enter < 1780 ls-2269 3d.s5 31us : irq_enter <-do_IRQ
1941 ls-2269 3d.s5 31us : rcu_irq_ent 1781 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1942 [...] 1782 [...]
1943 ls-2269 3d.s5 31us : rcu_irq_ent 1783 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1944 ls-2269 3d.s5 32us : add_preempt 1784 ls-2269 3d.s5 32us : add_preempt_count <-irq_enter
1945 ls-2269 3d.H5 32us : exit_idle < 1785 ls-2269 3d.H5 32us : exit_idle <-do_IRQ
1946 ls-2269 3d.H5 32us : handle_irq 1786 ls-2269 3d.H5 32us : handle_irq <-do_IRQ
1947 ls-2269 3d.H5 32us : irq_to_desc 1787 ls-2269 3d.H5 32us : irq_to_desc <-handle_irq
1948 ls-2269 3d.H5 33us : handle_fast 1788 ls-2269 3d.H5 33us : handle_fasteoi_irq <-handle_irq
1949 [...] 1789 [...]
1950 ls-2269 3d.s5 158us : _raw_spin_u 1790 ls-2269 3d.s5 158us : _raw_spin_unlock_irqrestore <-rtl8139_poll
1951 ls-2269 3d.s3 158us : net_rps_act 1791 ls-2269 3d.s3 158us : net_rps_action_and_irq_enable.isra.65 <-net_rx_action
1952 ls-2269 3d.s3 159us : __local_bh_ 1792 ls-2269 3d.s3 159us : __local_bh_enable <-__do_softirq
1953 ls-2269 3d.s3 159us : sub_preempt 1793 ls-2269 3d.s3 159us : sub_preempt_count <-__local_bh_enable
1954 ls-2269 3d..3 159us : idle_cpu <- 1794 ls-2269 3d..3 159us : idle_cpu <-irq_exit
1955 ls-2269 3d..3 159us : rcu_irq_exi 1795 ls-2269 3d..3 159us : rcu_irq_exit <-irq_exit
1956 ls-2269 3d..3 160us : sub_preempt 1796 ls-2269 3d..3 160us : sub_preempt_count <-irq_exit
1957 ls-2269 3d... 161us : __mutex_unl 1797 ls-2269 3d... 161us : __mutex_unlock_slowpath <-mutex_unlock
1958 ls-2269 3d... 162us+: trace_hardi 1798 ls-2269 3d... 162us+: trace_hardirqs_on <-mutex_unlock
1959 ls-2269 3d... 186us : <stack trac 1799 ls-2269 3d... 186us : <stack trace>
1960 => __mutex_unlock_slowpath 1800 => __mutex_unlock_slowpath
1961 => mutex_unlock 1801 => mutex_unlock
1962 => process_output 1802 => process_output
1963 => n_tty_write 1803 => n_tty_write
1964 => tty_write 1804 => tty_write
1965 => vfs_write 1805 => vfs_write
1966 => sys_write 1806 => sys_write
1967 => system_call_fastpath 1807 => system_call_fastpath
1968 1808
1969 This is an interesting trace. It started with 1809 This is an interesting trace. It started with kworker running and
1970 scheduling out and ls taking over. But as soo 1810 scheduling out and ls taking over. But as soon as ls released the
1971 rq lock and enabled interrupts (but not preem 1811 rq lock and enabled interrupts (but not preemption) an interrupt
1972 triggered. When the interrupt finished, it st 1812 triggered. When the interrupt finished, it started running softirqs.
1973 But while the softirq was running, another in 1813 But while the softirq was running, another interrupt triggered.
1974 When an interrupt is running inside a softirq 1814 When an interrupt is running inside a softirq, the annotation is 'H'.
1975 1815
1976 1816
1977 wakeup 1817 wakeup
1978 ------ 1818 ------
1979 1819
1980 One common case that people are interested in 1820 One common case that people are interested in tracing is the
1981 time it takes for a task that is woken to act 1821 time it takes for a task that is woken to actually wake up.
1982 Now for non Real-Time tasks, this can be arbi 1822 Now for non Real-Time tasks, this can be arbitrary. But tracing
1983 it nonetheless can be interesting. !! 1823 it none the less can be interesting.
1984 1824
1985 Without function tracing:: 1825 Without function tracing::
1986 1826
1987 # echo 0 > options/function-trace 1827 # echo 0 > options/function-trace
1988 # echo wakeup > current_tracer 1828 # echo wakeup > current_tracer
1989 # echo 1 > tracing_on 1829 # echo 1 > tracing_on
1990 # echo 0 > tracing_max_latency 1830 # echo 0 > tracing_max_latency
1991 # chrt -f 5 sleep 1 1831 # chrt -f 5 sleep 1
1992 # echo 0 > tracing_on 1832 # echo 0 > tracing_on
1993 # cat trace 1833 # cat trace
1994 # tracer: wakeup 1834 # tracer: wakeup
1995 # 1835 #
1996 # wakeup latency trace v1.1.5 on 3.8.0-test 1836 # wakeup latency trace v1.1.5 on 3.8.0-test+
1997 # ----------------------------------------- 1837 # --------------------------------------------------------------------
1998 # latency: 15 us, #4/4, CPU#3 | (M:preempt 1838 # latency: 15 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1999 # ----------------- 1839 # -----------------
2000 # | task: kworker/3:1H-312 (uid:0 nice:- 1840 # | task: kworker/3:1H-312 (uid:0 nice:-20 policy:0 rt_prio:0)
2001 # ----------------- 1841 # -----------------
2002 # 1842 #
2003 # _------=> CPU# 1843 # _------=> CPU#
2004 # / _-----=> irqs-off 1844 # / _-----=> irqs-off
2005 # | / _----=> need-resched 1845 # | / _----=> need-resched
2006 # || / _---=> hardirq/softir 1846 # || / _---=> hardirq/softirq
2007 # ||| / _--=> preempt-depth 1847 # ||| / _--=> preempt-depth
2008 # |||| / delay 1848 # |||| / delay
2009 # cmd pid ||||| time | caller 1849 # cmd pid ||||| time | caller
2010 # \ / ||||| \ | / 1850 # \ / ||||| \ | /
2011 <idle>-0 3dNs7 0us : 0:120: 1851 <idle>-0 3dNs7 0us : 0:120:R + [003] 312:100:R kworker/3:1H
2012 <idle>-0 3dNs7 1us+: ttwu_do_act 1852 <idle>-0 3dNs7 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
2013 <idle>-0 3d..3 15us : __schedule 1853 <idle>-0 3d..3 15us : __schedule <-schedule
2014 <idle>-0 3d..3 15us : 0:120: 1854 <idle>-0 3d..3 15us : 0:120:R ==> [003] 312:100:R kworker/3:1H
2015 1855
2016 The tracer only traces the highest priority t 1856 The tracer only traces the highest priority task in the system
2017 to avoid tracing the normal circumstances. He 1857 to avoid tracing the normal circumstances. Here we see that
2018 the kworker with a nice priority of -20 (not 1858 the kworker with a nice priority of -20 (not very nice), took
2019 just 15 microseconds from the time it woke up 1859 just 15 microseconds from the time it woke up, to the time it
2020 ran. 1860 ran.
2021 1861
2022 Non Real-Time tasks are not that interesting. 1862 Non Real-Time tasks are not that interesting. A more interesting
2023 trace is to concentrate only on Real-Time tas 1863 trace is to concentrate only on Real-Time tasks.
2024 1864
2025 wakeup_rt 1865 wakeup_rt
2026 --------- 1866 ---------
2027 1867
2028 In a Real-Time environment it is very importa 1868 In a Real-Time environment it is very important to know the
2029 wakeup time it takes for the highest priority 1869 wakeup time it takes for the highest priority task that is woken
2030 up to the time that it executes. This is also 1870 up to the time that it executes. This is also known as "schedule
2031 latency". I stress the point that this is abo 1871 latency". I stress the point that this is about RT tasks. It is
2032 also important to know the scheduling latency 1872 also important to know the scheduling latency of non-RT tasks,
2033 but the average schedule latency is better fo 1873 but the average schedule latency is better for non-RT tasks.
2034 Tools like LatencyTop are more appropriate fo 1874 Tools like LatencyTop are more appropriate for such
2035 measurements. 1875 measurements.
2036 1876
2037 Real-Time environments are interested in the 1877 Real-Time environments are interested in the worst case latency.
2038 That is the longest latency it takes for some 1878 That is the longest latency it takes for something to happen,
2039 and not the average. We can have a very fast 1879 and not the average. We can have a very fast scheduler that may
2040 only have a large latency once in a while, bu 1880 only have a large latency once in a while, but that would not
2041 work well with Real-Time tasks. The wakeup_r 1881 work well with Real-Time tasks. The wakeup_rt tracer was designed
2042 to record the worst case wakeups of RT tasks. 1882 to record the worst case wakeups of RT tasks. Non-RT tasks are
2043 not recorded because the tracer only records 1883 not recorded because the tracer only records one worst case and
2044 tracing non-RT tasks that are unpredictable w 1884 tracing non-RT tasks that are unpredictable will overwrite the
2045 worst case latency of RT tasks (just run the 1885 worst case latency of RT tasks (just run the normal wakeup
2046 tracer for a while to see that effect). 1886 tracer for a while to see that effect).
2047 1887
2048 Since this tracer only deals with RT tasks, w 1888 Since this tracer only deals with RT tasks, we will run this
2049 slightly differently than we did with the pre 1889 slightly differently than we did with the previous tracers.
2050 Instead of performing an 'ls', we will run 's 1890 Instead of performing an 'ls', we will run 'sleep 1' under
2051 'chrt' which changes the priority of the task 1891 'chrt' which changes the priority of the task.
2052 :: 1892 ::
2053 1893
2054 # echo 0 > options/function-trace 1894 # echo 0 > options/function-trace
2055 # echo wakeup_rt > current_tracer 1895 # echo wakeup_rt > current_tracer
2056 # echo 1 > tracing_on 1896 # echo 1 > tracing_on
2057 # echo 0 > tracing_max_latency 1897 # echo 0 > tracing_max_latency
2058 # chrt -f 5 sleep 1 1898 # chrt -f 5 sleep 1
2059 # echo 0 > tracing_on 1899 # echo 0 > tracing_on
2060 # cat trace 1900 # cat trace
2061 # tracer: wakeup 1901 # tracer: wakeup
2062 # 1902 #
2063 # tracer: wakeup_rt 1903 # tracer: wakeup_rt
2064 # 1904 #
2065 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 1905 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2066 # ----------------------------------------- 1906 # --------------------------------------------------------------------
2067 # latency: 5 us, #4/4, CPU#3 | (M:preempt V 1907 # latency: 5 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2068 # ----------------- 1908 # -----------------
2069 # | task: sleep-2389 (uid:0 nice:0 polic 1909 # | task: sleep-2389 (uid:0 nice:0 policy:1 rt_prio:5)
2070 # ----------------- 1910 # -----------------
2071 # 1911 #
2072 # _------=> CPU# 1912 # _------=> CPU#
2073 # / _-----=> irqs-off 1913 # / _-----=> irqs-off
2074 # | / _----=> need-resched 1914 # | / _----=> need-resched
2075 # || / _---=> hardirq/softir 1915 # || / _---=> hardirq/softirq
2076 # ||| / _--=> preempt-depth 1916 # ||| / _--=> preempt-depth
2077 # |||| / delay 1917 # |||| / delay
2078 # cmd pid ||||| time | caller 1918 # cmd pid ||||| time | caller
2079 # \ / ||||| \ | / 1919 # \ / ||||| \ | /
2080 <idle>-0 3d.h4 0us : 0:120: 1920 <idle>-0 3d.h4 0us : 0:120:R + [003] 2389: 94:R sleep
2081 <idle>-0 3d.h4 1us+: ttwu_do_act 1921 <idle>-0 3d.h4 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
2082 <idle>-0 3d..3 5us : __schedule 1922 <idle>-0 3d..3 5us : __schedule <-schedule
2083 <idle>-0 3d..3 5us : 0:120: 1923 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
2084 1924
2085 1925
2086 Running this on an idle system, we see that i 1926 Running this on an idle system, we see that it only took 5 microseconds
2087 to perform the task switch. Note, since the 1927 to perform the task switch. Note, since the trace point in the schedule
2088 is before the actual "switch", we stop the tr 1928 is before the actual "switch", we stop the tracing when the recorded task
2089 is about to schedule in. This may change if w 1929 is about to schedule in. This may change if we add a new marker at the
2090 end of the scheduler. 1930 end of the scheduler.
2091 1931
2092 Notice that the recorded task is 'sleep' with 1932 Notice that the recorded task is 'sleep' with the PID of 2389
2093 and it has an rt_prio of 5. This priority is 1933 and it has an rt_prio of 5. This priority is user-space priority
2094 and not the internal kernel priority. The pol 1934 and not the internal kernel priority. The policy is 1 for
2095 SCHED_FIFO and 2 for SCHED_RR. 1935 SCHED_FIFO and 2 for SCHED_RR.
2096 1936
2097 Note, that the trace data shows the internal 1937 Note, that the trace data shows the internal priority (99 - rtprio).
2098 :: 1938 ::
2099 1939
2100 <idle>-0 3d..3 5us : 0:120:R 1940 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
2101 1941
2102 The 0:120:R means idle was running with a nic 1942 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 1943 and in the running state 'R'. The sleep task was scheduled in with
2104 2389: 94:R. That is the priority is the kerne 1944 2389: 94:R. That is the priority is the kernel rtprio (99 - 5 = 94)
2105 and it too is in the running state. 1945 and it too is in the running state.
2106 1946
2107 Doing the same with chrt -r 5 and function-tr 1947 Doing the same with chrt -r 5 and function-trace set.
2108 :: 1948 ::
2109 1949
2110 echo 1 > options/function-trace 1950 echo 1 > options/function-trace
2111 1951
2112 # tracer: wakeup_rt 1952 # tracer: wakeup_rt
2113 # 1953 #
2114 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 1954 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2115 # ----------------------------------------- 1955 # --------------------------------------------------------------------
2116 # latency: 29 us, #85/85, CPU#3 | (M:preemp 1956 # latency: 29 us, #85/85, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2117 # ----------------- 1957 # -----------------
2118 # | task: sleep-2448 (uid:0 nice:0 polic 1958 # | task: sleep-2448 (uid:0 nice:0 policy:1 rt_prio:5)
2119 # ----------------- 1959 # -----------------
2120 # 1960 #
2121 # _------=> CPU# 1961 # _------=> CPU#
2122 # / _-----=> irqs-off 1962 # / _-----=> irqs-off
2123 # | / _----=> need-resched 1963 # | / _----=> need-resched
2124 # || / _---=> hardirq/softir 1964 # || / _---=> hardirq/softirq
2125 # ||| / _--=> preempt-depth 1965 # ||| / _--=> preempt-depth
2126 # |||| / delay 1966 # |||| / delay
2127 # cmd pid ||||| time | caller 1967 # cmd pid ||||| time | caller
2128 # \ / ||||| \ | / 1968 # \ / ||||| \ | /
2129 <idle>-0 3d.h4 1us+: 0:120: 1969 <idle>-0 3d.h4 1us+: 0:120:R + [003] 2448: 94:R sleep
2130 <idle>-0 3d.h4 2us : ttwu_do_act 1970 <idle>-0 3d.h4 2us : ttwu_do_activate.constprop.87 <-try_to_wake_up
2131 <idle>-0 3d.h3 3us : check_preem 1971 <idle>-0 3d.h3 3us : check_preempt_curr <-ttwu_do_wakeup
2132 <idle>-0 3d.h3 3us : resched_cur 1972 <idle>-0 3d.h3 3us : resched_curr <-check_preempt_curr
2133 <idle>-0 3dNh3 4us : task_woken_ 1973 <idle>-0 3dNh3 4us : task_woken_rt <-ttwu_do_wakeup
2134 <idle>-0 3dNh3 4us : _raw_spin_u 1974 <idle>-0 3dNh3 4us : _raw_spin_unlock <-try_to_wake_up
2135 <idle>-0 3dNh3 4us : sub_preempt 1975 <idle>-0 3dNh3 4us : sub_preempt_count <-_raw_spin_unlock
2136 <idle>-0 3dNh2 5us : ttwu_stat < 1976 <idle>-0 3dNh2 5us : ttwu_stat <-try_to_wake_up
2137 <idle>-0 3dNh2 5us : _raw_spin_u 1977 <idle>-0 3dNh2 5us : _raw_spin_unlock_irqrestore <-try_to_wake_up
2138 <idle>-0 3dNh2 6us : sub_preempt 1978 <idle>-0 3dNh2 6us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2139 <idle>-0 3dNh1 6us : _raw_spin_l 1979 <idle>-0 3dNh1 6us : _raw_spin_lock <-__run_hrtimer
2140 <idle>-0 3dNh1 6us : add_preempt 1980 <idle>-0 3dNh1 6us : add_preempt_count <-_raw_spin_lock
2141 <idle>-0 3dNh2 7us : _raw_spin_u 1981 <idle>-0 3dNh2 7us : _raw_spin_unlock <-hrtimer_interrupt
2142 <idle>-0 3dNh2 7us : sub_preempt 1982 <idle>-0 3dNh2 7us : sub_preempt_count <-_raw_spin_unlock
2143 <idle>-0 3dNh1 7us : tick_progra 1983 <idle>-0 3dNh1 7us : tick_program_event <-hrtimer_interrupt
2144 <idle>-0 3dNh1 7us : clockevents 1984 <idle>-0 3dNh1 7us : clockevents_program_event <-tick_program_event
2145 <idle>-0 3dNh1 8us : ktime_get < 1985 <idle>-0 3dNh1 8us : ktime_get <-clockevents_program_event
2146 <idle>-0 3dNh1 8us : lapic_next_ 1986 <idle>-0 3dNh1 8us : lapic_next_event <-clockevents_program_event
2147 <idle>-0 3dNh1 8us : irq_exit <- 1987 <idle>-0 3dNh1 8us : irq_exit <-smp_apic_timer_interrupt
2148 <idle>-0 3dNh1 9us : sub_preempt 1988 <idle>-0 3dNh1 9us : sub_preempt_count <-irq_exit
2149 <idle>-0 3dN.2 9us : idle_cpu <- 1989 <idle>-0 3dN.2 9us : idle_cpu <-irq_exit
2150 <idle>-0 3dN.2 9us : rcu_irq_exi 1990 <idle>-0 3dN.2 9us : rcu_irq_exit <-irq_exit
2151 <idle>-0 3dN.2 10us : rcu_eqs_ent 1991 <idle>-0 3dN.2 10us : rcu_eqs_enter_common.isra.45 <-rcu_irq_exit
2152 <idle>-0 3dN.2 10us : sub_preempt 1992 <idle>-0 3dN.2 10us : sub_preempt_count <-irq_exit
2153 <idle>-0 3.N.1 11us : rcu_idle_ex 1993 <idle>-0 3.N.1 11us : rcu_idle_exit <-cpu_idle
2154 <idle>-0 3dN.1 11us : rcu_eqs_exi 1994 <idle>-0 3dN.1 11us : rcu_eqs_exit_common.isra.43 <-rcu_idle_exit
2155 <idle>-0 3.N.1 11us : tick_nohz_i 1995 <idle>-0 3.N.1 11us : tick_nohz_idle_exit <-cpu_idle
2156 <idle>-0 3dN.1 12us : menu_hrtime 1996 <idle>-0 3dN.1 12us : menu_hrtimer_cancel <-tick_nohz_idle_exit
2157 <idle>-0 3dN.1 12us : ktime_get < 1997 <idle>-0 3dN.1 12us : ktime_get <-tick_nohz_idle_exit
2158 <idle>-0 3dN.1 12us : tick_do_upd 1998 <idle>-0 3dN.1 12us : tick_do_update_jiffies64 <-tick_nohz_idle_exit
2159 <idle>-0 3dN.1 13us : cpu_load_up 1999 <idle>-0 3dN.1 13us : cpu_load_update_nohz <-tick_nohz_idle_exit
2160 <idle>-0 3dN.1 13us : _raw_spin_l 2000 <idle>-0 3dN.1 13us : _raw_spin_lock <-cpu_load_update_nohz
2161 <idle>-0 3dN.1 13us : add_preempt 2001 <idle>-0 3dN.1 13us : add_preempt_count <-_raw_spin_lock
2162 <idle>-0 3dN.2 13us : __cpu_load_ 2002 <idle>-0 3dN.2 13us : __cpu_load_update <-cpu_load_update_nohz
2163 <idle>-0 3dN.2 14us : sched_avg_u 2003 <idle>-0 3dN.2 14us : sched_avg_update <-__cpu_load_update
2164 <idle>-0 3dN.2 14us : _raw_spin_u 2004 <idle>-0 3dN.2 14us : _raw_spin_unlock <-cpu_load_update_nohz
2165 <idle>-0 3dN.2 14us : sub_preempt 2005 <idle>-0 3dN.2 14us : sub_preempt_count <-_raw_spin_unlock
2166 <idle>-0 3dN.1 15us : calc_load_n 2006 <idle>-0 3dN.1 15us : calc_load_nohz_stop <-tick_nohz_idle_exit
2167 <idle>-0 3dN.1 15us : touch_softl 2007 <idle>-0 3dN.1 15us : touch_softlockup_watchdog <-tick_nohz_idle_exit
2168 <idle>-0 3dN.1 15us : hrtimer_can 2008 <idle>-0 3dN.1 15us : hrtimer_cancel <-tick_nohz_idle_exit
2169 <idle>-0 3dN.1 15us : hrtimer_try 2009 <idle>-0 3dN.1 15us : hrtimer_try_to_cancel <-hrtimer_cancel
2170 <idle>-0 3dN.1 16us : lock_hrtime 2010 <idle>-0 3dN.1 16us : lock_hrtimer_base.isra.18 <-hrtimer_try_to_cancel
2171 <idle>-0 3dN.1 16us : _raw_spin_l 2011 <idle>-0 3dN.1 16us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
2172 <idle>-0 3dN.1 16us : add_preempt 2012 <idle>-0 3dN.1 16us : add_preempt_count <-_raw_spin_lock_irqsave
2173 <idle>-0 3dN.2 17us : __remove_hr 2013 <idle>-0 3dN.2 17us : __remove_hrtimer <-remove_hrtimer.part.16
2174 <idle>-0 3dN.2 17us : hrtimer_for 2014 <idle>-0 3dN.2 17us : hrtimer_force_reprogram <-__remove_hrtimer
2175 <idle>-0 3dN.2 17us : tick_progra 2015 <idle>-0 3dN.2 17us : tick_program_event <-hrtimer_force_reprogram
2176 <idle>-0 3dN.2 18us : clockevents 2016 <idle>-0 3dN.2 18us : clockevents_program_event <-tick_program_event
2177 <idle>-0 3dN.2 18us : ktime_get < 2017 <idle>-0 3dN.2 18us : ktime_get <-clockevents_program_event
2178 <idle>-0 3dN.2 18us : lapic_next_ 2018 <idle>-0 3dN.2 18us : lapic_next_event <-clockevents_program_event
2179 <idle>-0 3dN.2 19us : _raw_spin_u 2019 <idle>-0 3dN.2 19us : _raw_spin_unlock_irqrestore <-hrtimer_try_to_cancel
2180 <idle>-0 3dN.2 19us : sub_preempt 2020 <idle>-0 3dN.2 19us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2181 <idle>-0 3dN.1 19us : hrtimer_for 2021 <idle>-0 3dN.1 19us : hrtimer_forward <-tick_nohz_idle_exit
2182 <idle>-0 3dN.1 20us : ktime_add_s 2022 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
2183 <idle>-0 3dN.1 20us : ktime_add_s 2023 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
2184 <idle>-0 3dN.1 20us : hrtimer_sta 2024 <idle>-0 3dN.1 20us : hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
2185 <idle>-0 3dN.1 20us : __hrtimer_s 2025 <idle>-0 3dN.1 20us : __hrtimer_start_range_ns <-hrtimer_start_range_ns
2186 <idle>-0 3dN.1 21us : lock_hrtime 2026 <idle>-0 3dN.1 21us : lock_hrtimer_base.isra.18 <-__hrtimer_start_range_ns
2187 <idle>-0 3dN.1 21us : _raw_spin_l 2027 <idle>-0 3dN.1 21us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
2188 <idle>-0 3dN.1 21us : add_preempt 2028 <idle>-0 3dN.1 21us : add_preempt_count <-_raw_spin_lock_irqsave
2189 <idle>-0 3dN.2 22us : ktime_add_s 2029 <idle>-0 3dN.2 22us : ktime_add_safe <-__hrtimer_start_range_ns
2190 <idle>-0 3dN.2 22us : enqueue_hrt 2030 <idle>-0 3dN.2 22us : enqueue_hrtimer <-__hrtimer_start_range_ns
2191 <idle>-0 3dN.2 22us : tick_progra 2031 <idle>-0 3dN.2 22us : tick_program_event <-__hrtimer_start_range_ns
2192 <idle>-0 3dN.2 23us : clockevents 2032 <idle>-0 3dN.2 23us : clockevents_program_event <-tick_program_event
2193 <idle>-0 3dN.2 23us : ktime_get < 2033 <idle>-0 3dN.2 23us : ktime_get <-clockevents_program_event
2194 <idle>-0 3dN.2 23us : lapic_next_ 2034 <idle>-0 3dN.2 23us : lapic_next_event <-clockevents_program_event
2195 <idle>-0 3dN.2 24us : _raw_spin_u 2035 <idle>-0 3dN.2 24us : _raw_spin_unlock_irqrestore <-__hrtimer_start_range_ns
2196 <idle>-0 3dN.2 24us : sub_preempt 2036 <idle>-0 3dN.2 24us : sub_preempt_count <-_raw_spin_unlock_irqrestore
2197 <idle>-0 3dN.1 24us : account_idl 2037 <idle>-0 3dN.1 24us : account_idle_ticks <-tick_nohz_idle_exit
2198 <idle>-0 3dN.1 24us : account_idl 2038 <idle>-0 3dN.1 24us : account_idle_time <-account_idle_ticks
2199 <idle>-0 3.N.1 25us : sub_preempt 2039 <idle>-0 3.N.1 25us : sub_preempt_count <-cpu_idle
2200 <idle>-0 3.N.. 25us : schedule <- 2040 <idle>-0 3.N.. 25us : schedule <-cpu_idle
2201 <idle>-0 3.N.. 25us : __schedule 2041 <idle>-0 3.N.. 25us : __schedule <-preempt_schedule
2202 <idle>-0 3.N.. 26us : add_preempt 2042 <idle>-0 3.N.. 26us : add_preempt_count <-__schedule
2203 <idle>-0 3.N.1 26us : rcu_note_co 2043 <idle>-0 3.N.1 26us : rcu_note_context_switch <-__schedule
2204 <idle>-0 3.N.1 26us : rcu_sched_q 2044 <idle>-0 3.N.1 26us : rcu_sched_qs <-rcu_note_context_switch
2205 <idle>-0 3dN.1 27us : rcu_preempt 2045 <idle>-0 3dN.1 27us : rcu_preempt_qs <-rcu_note_context_switch
2206 <idle>-0 3.N.1 27us : _raw_spin_l 2046 <idle>-0 3.N.1 27us : _raw_spin_lock_irq <-__schedule
2207 <idle>-0 3dN.1 27us : add_preempt 2047 <idle>-0 3dN.1 27us : add_preempt_count <-_raw_spin_lock_irq
2208 <idle>-0 3dN.2 28us : put_prev_ta 2048 <idle>-0 3dN.2 28us : put_prev_task_idle <-__schedule
2209 <idle>-0 3dN.2 28us : pick_next_t 2049 <idle>-0 3dN.2 28us : pick_next_task_stop <-pick_next_task
2210 <idle>-0 3dN.2 28us : pick_next_t 2050 <idle>-0 3dN.2 28us : pick_next_task_rt <-pick_next_task
2211 <idle>-0 3dN.2 29us : dequeue_pus 2051 <idle>-0 3dN.2 29us : dequeue_pushable_task <-pick_next_task_rt
2212 <idle>-0 3d..3 29us : __schedule 2052 <idle>-0 3d..3 29us : __schedule <-preempt_schedule
2213 <idle>-0 3d..3 30us : 0:120: 2053 <idle>-0 3d..3 30us : 0:120:R ==> [003] 2448: 94:R sleep
2214 2054
2215 This isn't that big of a trace, even with fun 2055 This isn't that big of a trace, even with function tracing enabled,
2216 so I included the entire trace. 2056 so I included the entire trace.
2217 2057
2218 The interrupt went off while when the system 2058 The interrupt went off while when the system was idle. Somewhere
2219 before task_woken_rt() was called, the NEED_R 2059 before task_woken_rt() was called, the NEED_RESCHED flag was set,
2220 this is indicated by the first occurrence of 2060 this is indicated by the first occurrence of the 'N' flag.
2221 2061
2222 Latency tracing and events 2062 Latency tracing and events
2223 -------------------------- 2063 --------------------------
2224 As function tracing can induce a much larger 2064 As function tracing can induce a much larger latency, but without
2225 seeing what happens within the latency it is 2065 seeing what happens within the latency it is hard to know what
2226 caused it. There is a middle ground, and that 2066 caused it. There is a middle ground, and that is with enabling
2227 events. 2067 events.
2228 :: 2068 ::
2229 2069
2230 # echo 0 > options/function-trace 2070 # echo 0 > options/function-trace
2231 # echo wakeup_rt > current_tracer 2071 # echo wakeup_rt > current_tracer
2232 # echo 1 > events/enable 2072 # echo 1 > events/enable
2233 # echo 1 > tracing_on 2073 # echo 1 > tracing_on
2234 # echo 0 > tracing_max_latency 2074 # echo 0 > tracing_max_latency
2235 # chrt -f 5 sleep 1 2075 # chrt -f 5 sleep 1
2236 # echo 0 > tracing_on 2076 # echo 0 > tracing_on
2237 # cat trace 2077 # cat trace
2238 # tracer: wakeup_rt 2078 # tracer: wakeup_rt
2239 # 2079 #
2240 # wakeup_rt latency trace v1.1.5 on 3.8.0-t 2080 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+
2241 # ----------------------------------------- 2081 # --------------------------------------------------------------------
2242 # latency: 6 us, #12/12, CPU#2 | (M:preempt 2082 # latency: 6 us, #12/12, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
2243 # ----------------- 2083 # -----------------
2244 # | task: sleep-5882 (uid:0 nice:0 polic 2084 # | task: sleep-5882 (uid:0 nice:0 policy:1 rt_prio:5)
2245 # ----------------- 2085 # -----------------
2246 # 2086 #
2247 # _------=> CPU# 2087 # _------=> CPU#
2248 # / _-----=> irqs-off 2088 # / _-----=> irqs-off
2249 # | / _----=> need-resched 2089 # | / _----=> need-resched
2250 # || / _---=> hardirq/softir 2090 # || / _---=> hardirq/softirq
2251 # ||| / _--=> preempt-depth 2091 # ||| / _--=> preempt-depth
2252 # |||| / delay 2092 # |||| / delay
2253 # cmd pid ||||| time | caller 2093 # cmd pid ||||| time | caller
2254 # \ / ||||| \ | / 2094 # \ / ||||| \ | /
2255 <idle>-0 2d.h4 0us : 0:120: 2095 <idle>-0 2d.h4 0us : 0:120:R + [002] 5882: 94:R sleep
2256 <idle>-0 2d.h4 0us : ttwu_do_act 2096 <idle>-0 2d.h4 0us : ttwu_do_activate.constprop.87 <-try_to_wake_up
2257 <idle>-0 2d.h4 1us : sched_wakeu 2097 <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 2098 <idle>-0 2dNh2 1us : hrtimer_expire_exit: hrtimer=ffff88007796feb8
2259 <idle>-0 2.N.2 2us : power_end: 2099 <idle>-0 2.N.2 2us : power_end: cpu_id=2
2260 <idle>-0 2.N.2 3us : cpu_idle: s 2100 <idle>-0 2.N.2 3us : cpu_idle: state=4294967295 cpu_id=2
2261 <idle>-0 2dN.3 4us : hrtimer_can 2101 <idle>-0 2dN.3 4us : hrtimer_cancel: hrtimer=ffff88007d50d5e0
2262 <idle>-0 2dN.3 4us : hrtimer_sta 2102 <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 2103 <idle>-0 2.N.2 5us : rcu_utilization: Start context switch
2264 <idle>-0 2.N.2 5us : rcu_utiliza 2104 <idle>-0 2.N.2 5us : rcu_utilization: End context switch
2265 <idle>-0 2d..3 6us : __schedule 2105 <idle>-0 2d..3 6us : __schedule <-schedule
2266 <idle>-0 2d..3 6us : 0:120: 2106 <idle>-0 2d..3 6us : 0:120:R ==> [002] 5882: 94:R sleep
2267 2107
2268 2108
2269 Hardware Latency Detector 2109 Hardware Latency Detector
2270 ------------------------- 2110 -------------------------
2271 2111
2272 The hardware latency detector is executed by 2112 The hardware latency detector is executed by enabling the "hwlat" tracer.
2273 2113
2274 NOTE, this tracer will affect the performance 2114 NOTE, this tracer will affect the performance of the system as it will
2275 periodically make a CPU constantly busy with 2115 periodically make a CPU constantly busy with interrupts disabled.
2276 :: 2116 ::
2277 2117
2278 # echo hwlat > current_tracer 2118 # echo hwlat > current_tracer
2279 # sleep 100 2119 # sleep 100
2280 # cat trace 2120 # cat trace
2281 # tracer: hwlat 2121 # tracer: hwlat
2282 # 2122 #
2283 # entries-in-buffer/entries-written: 13/13 <<
2284 # <<
2285 # _-----=> irq 2123 # _-----=> irqs-off
2286 # / _----=> nee 2124 # / _----=> need-resched
2287 # | / _---=> har 2125 # | / _---=> hardirq/softirq
2288 # || / _--=> pre 2126 # || / _--=> preempt-depth
2289 # ||| / dela 2127 # ||| / delay
2290 # TASK-PID CPU# |||| TIMEST 2128 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
2291 # | | | |||| | 2129 # | | | |||| | |
2292 <...>-1729 [001] d... 678.473 !! 2130 <...>-3638 [001] d... 19452.055471: #1 inner/outer(us): 12/14 ts:1499801089.066141940
2293 <...>-1729 [004] d... 689.556 !! 2131 <...>-3638 [003] d... 19454.071354: #2 inner/outer(us): 11/9 ts:1499801091.082164365
2294 <...>-1729 [005] d... 714.756 !! 2132 <...>-3638 [002] dn.. 19461.126852: #3 inner/outer(us): 12/9 ts:1499801098.138150062
2295 <...>-1729 [001] d... 718.788 !! 2133 <...>-3638 [001] d... 19488.340960: #4 inner/outer(us): 8/12 ts:1499801125.354139633
2296 <...>-1729 [002] d... 719.796 !! 2134 <...>-3638 [003] d... 19494.388553: #5 inner/outer(us): 8/12 ts:1499801131.402150961
2297 <...>-1729 [006] d... 844.787 !! 2135 <...>-3638 [003] d... 19501.283419: #6 inner/outer(us): 0/12 ts:1499801138.297435289 nmi-total:4 nmi-count:1
2298 <...>-1729 [003] d... 849.827 <<
2299 <...>-1729 [007] d... 853.859 <<
2300 <...>-1729 [001] d... 855.874 <<
2301 <...>-1729 [001] d... 863.938 <<
2302 <...>-1729 [007] d... 878.050 <<
2303 <...>-1729 [007] d... 886.114 <<
2304 2136
2305 2137
2306 The above output is somewhat the same in the 2138 The above output is somewhat the same in the header. All events will have
2307 interrupts disabled 'd'. Under the FUNCTION t 2139 interrupts disabled 'd'. Under the FUNCTION title there is:
2308 2140
2309 #1 2141 #1
2310 This is the count of events recorded 2142 This is the count of events recorded that were greater than the
2311 tracing_threshold (See below). 2143 tracing_threshold (See below).
2312 2144
2313 inner/outer(us): 11/11 !! 2145 inner/outer(us): 12/14
2314 2146
2315 This shows two numbers as "inner latenc 2147 This shows two numbers as "inner latency" and "outer latency". The test
2316 runs in a loop checking a timestamp twi 2148 runs in a loop checking a timestamp twice. The latency detected within
2317 the two timestamps is the "inner latenc 2149 the two timestamps is the "inner latency" and the latency detected
2318 after the previous timestamp and the ne 2150 after the previous timestamp and the next timestamp in the loop is
2319 the "outer latency". 2151 the "outer latency".
2320 2152
2321 ts:1581527483.343962693 !! 2153 ts:1499801089.066141940
2322 2154
2323 The absolute timestamp that the first l !! 2155 The absolute timestamp that the event happened.
2324 2156
2325 count:6 !! 2157 nmi-total:4 nmi-count:1
2326 <<
2327 The number of times a latency was detec <<
2328 <<
2329 nmi-total:7 nmi-count:1 <<
2330 2158
2331 On architectures that support it, if an 2159 On architectures that support it, if an NMI comes in during the
2332 test, the time spent in NMI is reported 2160 test, the time spent in NMI is reported in "nmi-total" (in
2333 microseconds). 2161 microseconds).
2334 2162
2335 All architectures that have NMIs will s 2163 All architectures that have NMIs will show the "nmi-count" if an
2336 NMI comes in during the test. 2164 NMI comes in during the test.
2337 2165
2338 hwlat files: 2166 hwlat files:
2339 2167
2340 tracing_threshold 2168 tracing_threshold
2341 This gets automatically set to "10" t 2169 This gets automatically set to "10" to represent 10
2342 microseconds. This is the threshold o 2170 microseconds. This is the threshold of latency that
2343 needs to be detected before the trace 2171 needs to be detected before the trace will be recorded.
2344 2172
2345 Note, when hwlat tracer is finished ( 2173 Note, when hwlat tracer is finished (another tracer is
2346 written into "current_tracer"), the o 2174 written into "current_tracer"), the original value for
2347 tracing_threshold is placed back into 2175 tracing_threshold is placed back into this file.
2348 2176
2349 hwlat_detector/width 2177 hwlat_detector/width
2350 The length of time the test runs with 2178 The length of time the test runs with interrupts disabled.
2351 2179
2352 hwlat_detector/window 2180 hwlat_detector/window
2353 The length of time of the window whic 2181 The length of time of the window which the test
2354 runs. That is, the test will run for 2182 runs. That is, the test will run for "width"
2355 microseconds per "window" microsecond 2183 microseconds per "window" microseconds
2356 2184
2357 tracing_cpumask 2185 tracing_cpumask
2358 When the test is started. A kernel th 2186 When the test is started. A kernel thread is created that
2359 runs the test. This thread will alter 2187 runs the test. This thread will alternate between CPUs
2360 listed in the tracing_cpumask between 2188 listed in the tracing_cpumask between each period
2361 (one "window"). To limit the test to 2189 (one "window"). To limit the test to specific CPUs
2362 set the mask in this file to only the 2190 set the mask in this file to only the CPUs that the test
2363 should run on. 2191 should run on.
2364 2192
2365 function 2193 function
2366 -------- 2194 --------
2367 2195
2368 This tracer is the function tracer. Enabling 2196 This tracer is the function tracer. Enabling the function tracer
2369 can be done from the debug file system. Make 2197 can be done from the debug file system. Make sure the
2370 ftrace_enabled is set; otherwise this tracer 2198 ftrace_enabled is set; otherwise this tracer is a nop.
2371 See the "ftrace_enabled" section below. 2199 See the "ftrace_enabled" section below.
2372 :: 2200 ::
2373 2201
2374 # sysctl kernel.ftrace_enabled=1 2202 # sysctl kernel.ftrace_enabled=1
2375 # echo function > current_tracer 2203 # echo function > current_tracer
2376 # echo 1 > tracing_on 2204 # echo 1 > tracing_on
2377 # usleep 1 2205 # usleep 1
2378 # echo 0 > tracing_on 2206 # echo 0 > tracing_on
2379 # cat trace 2207 # cat trace
2380 # tracer: function 2208 # tracer: function
2381 # 2209 #
2382 # entries-in-buffer/entries-written: 24799/ 2210 # entries-in-buffer/entries-written: 24799/24799 #P:4
2383 # 2211 #
2384 # _-----=> irq 2212 # _-----=> irqs-off
2385 # / _----=> nee 2213 # / _----=> need-resched
2386 # | / _---=> har 2214 # | / _---=> hardirq/softirq
2387 # || / _--=> pre 2215 # || / _--=> preempt-depth
2388 # ||| / dela 2216 # ||| / delay
2389 # TASK-PID CPU# |||| TIMEST 2217 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
2390 # | | | |||| | 2218 # | | | |||| | |
2391 bash-1994 [002] .... 3082.063 2219 bash-1994 [002] .... 3082.063030: mutex_unlock <-rb_simple_write
2392 bash-1994 [002] .... 3082.063 2220 bash-1994 [002] .... 3082.063031: __mutex_unlock_slowpath <-mutex_unlock
2393 bash-1994 [002] .... 3082.063 2221 bash-1994 [002] .... 3082.063031: __fsnotify_parent <-fsnotify_modify
2394 bash-1994 [002] .... 3082.063 2222 bash-1994 [002] .... 3082.063032: fsnotify <-fsnotify_modify
2395 bash-1994 [002] .... 3082.063 2223 bash-1994 [002] .... 3082.063032: __srcu_read_lock <-fsnotify
2396 bash-1994 [002] .... 3082.063 2224 bash-1994 [002] .... 3082.063032: add_preempt_count <-__srcu_read_lock
2397 bash-1994 [002] ...1 3082.063 2225 bash-1994 [002] ...1 3082.063032: sub_preempt_count <-__srcu_read_lock
2398 bash-1994 [002] .... 3082.063 2226 bash-1994 [002] .... 3082.063033: __srcu_read_unlock <-fsnotify
2399 [...] 2227 [...]
2400 2228
2401 2229
2402 Note: function tracer uses ring buffers to st 2230 Note: function tracer uses ring buffers to store the above
2403 entries. The newest data may overwrite the ol 2231 entries. The newest data may overwrite the oldest data.
2404 Sometimes using echo to stop the trace is not 2232 Sometimes using echo to stop the trace is not sufficient because
2405 the tracing could have overwritten the data t 2233 the tracing could have overwritten the data that you wanted to
2406 record. For this reason, it is sometimes bett 2234 record. For this reason, it is sometimes better to disable
2407 tracing directly from a program. This allows 2235 tracing directly from a program. This allows you to stop the
2408 tracing at the point that you hit the part th 2236 tracing at the point that you hit the part that you are
2409 interested in. To disable the tracing directl 2237 interested in. To disable the tracing directly from a C program,
2410 something like following code snippet can be 2238 something like following code snippet can be used::
2411 2239
2412 int trace_fd; 2240 int trace_fd;
2413 [...] 2241 [...]
2414 int main(int argc, char *argv[]) { 2242 int main(int argc, char *argv[]) {
2415 [...] 2243 [...]
2416 trace_fd = open(tracing_file( 2244 trace_fd = open(tracing_file("tracing_on"), O_WRONLY);
2417 [...] 2245 [...]
2418 if (condition_hit()) { 2246 if (condition_hit()) {
2419 write(trace_fd, "0", 2247 write(trace_fd, "0", 1);
2420 } 2248 }
2421 [...] 2249 [...]
2422 } 2250 }
2423 2251
2424 2252
2425 Single thread tracing 2253 Single thread tracing
2426 --------------------- 2254 ---------------------
2427 2255
2428 By writing into set_ftrace_pid you can trace 2256 By writing into set_ftrace_pid you can trace a
2429 single thread. For example:: 2257 single thread. For example::
2430 2258
2431 # cat set_ftrace_pid 2259 # cat set_ftrace_pid
2432 no pid 2260 no pid
2433 # echo 3111 > set_ftrace_pid 2261 # echo 3111 > set_ftrace_pid
2434 # cat set_ftrace_pid 2262 # cat set_ftrace_pid
2435 3111 2263 3111
2436 # echo function > current_tracer 2264 # echo function > current_tracer
2437 # cat trace | head 2265 # cat trace | head
2438 # tracer: function 2266 # tracer: function
2439 # 2267 #
2440 # TASK-PID CPU# TIMESTAMP 2268 # TASK-PID CPU# TIMESTAMP FUNCTION
2441 # | | | | 2269 # | | | | |
2442 yum-updatesd-3111 [003] 1637.254676: 2270 yum-updatesd-3111 [003] 1637.254676: finish_task_switch <-thread_return
2443 yum-updatesd-3111 [003] 1637.254681: 2271 yum-updatesd-3111 [003] 1637.254681: hrtimer_cancel <-schedule_hrtimeout_range
2444 yum-updatesd-3111 [003] 1637.254682: 2272 yum-updatesd-3111 [003] 1637.254682: hrtimer_try_to_cancel <-hrtimer_cancel
2445 yum-updatesd-3111 [003] 1637.254683: 2273 yum-updatesd-3111 [003] 1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel
2446 yum-updatesd-3111 [003] 1637.254685: 2274 yum-updatesd-3111 [003] 1637.254685: fget_light <-do_sys_poll
2447 yum-updatesd-3111 [003] 1637.254686: 2275 yum-updatesd-3111 [003] 1637.254686: pipe_poll <-do_sys_poll
2448 # echo > set_ftrace_pid 2276 # echo > set_ftrace_pid
2449 # cat trace |head 2277 # cat trace |head
2450 # tracer: function 2278 # tracer: function
2451 # 2279 #
2452 # TASK-PID CPU# TIMESTAMP 2280 # TASK-PID CPU# TIMESTAMP FUNCTION
2453 # | | | | 2281 # | | | | |
2454 ##### CPU 3 buffer started #### 2282 ##### CPU 3 buffer started ####
2455 yum-updatesd-3111 [003] 1701.957688: 2283 yum-updatesd-3111 [003] 1701.957688: free_poll_entry <-poll_freewait
2456 yum-updatesd-3111 [003] 1701.957689: 2284 yum-updatesd-3111 [003] 1701.957689: remove_wait_queue <-free_poll_entry
2457 yum-updatesd-3111 [003] 1701.957691: 2285 yum-updatesd-3111 [003] 1701.957691: fput <-free_poll_entry
2458 yum-updatesd-3111 [003] 1701.957692: 2286 yum-updatesd-3111 [003] 1701.957692: audit_syscall_exit <-sysret_audit
2459 yum-updatesd-3111 [003] 1701.957693: 2287 yum-updatesd-3111 [003] 1701.957693: path_put <-audit_syscall_exit
2460 2288
2461 If you want to trace a function when executin 2289 If you want to trace a function when executing, you could use
2462 something like this simple program. 2290 something like this simple program.
2463 :: 2291 ::
2464 2292
2465 #include <stdio.h> 2293 #include <stdio.h>
2466 #include <stdlib.h> 2294 #include <stdlib.h>
2467 #include <sys/types.h> 2295 #include <sys/types.h>
2468 #include <sys/stat.h> 2296 #include <sys/stat.h>
2469 #include <fcntl.h> 2297 #include <fcntl.h>
2470 #include <unistd.h> 2298 #include <unistd.h>
2471 #include <string.h> 2299 #include <string.h>
2472 2300
2473 #define _STR(x) #x 2301 #define _STR(x) #x
2474 #define STR(x) _STR(x) 2302 #define STR(x) _STR(x)
2475 #define MAX_PATH 256 2303 #define MAX_PATH 256
2476 2304
2477 const char *find_tracefs(void) 2305 const char *find_tracefs(void)
2478 { 2306 {
2479 static char tracefs[MAX_PATH+1 2307 static char tracefs[MAX_PATH+1];
2480 static int tracefs_found; 2308 static int tracefs_found;
2481 char type[100]; 2309 char type[100];
2482 FILE *fp; 2310 FILE *fp;
2483 2311
2484 if (tracefs_found) 2312 if (tracefs_found)
2485 return tracefs; 2313 return tracefs;
2486 2314
2487 if ((fp = fopen("/proc/mounts" 2315 if ((fp = fopen("/proc/mounts","r")) == NULL) {
2488 perror("/proc/mounts") 2316 perror("/proc/mounts");
2489 return NULL; 2317 return NULL;
2490 } 2318 }
2491 2319
2492 while (fscanf(fp, "%*s %" 2320 while (fscanf(fp, "%*s %"
2493 STR(MAX_PATH) 2321 STR(MAX_PATH)
2494 "s %99s %*s %*d 2322 "s %99s %*s %*d %*d\n",
2495 tracefs, type) = 2323 tracefs, type) == 2) {
2496 if (strcmp(type, "trac 2324 if (strcmp(type, "tracefs") == 0)
2497 break; 2325 break;
2498 } 2326 }
2499 fclose(fp); 2327 fclose(fp);
2500 2328
2501 if (strcmp(type, "tracefs") != 2329 if (strcmp(type, "tracefs") != 0) {
2502 fprintf(stderr, "trace 2330 fprintf(stderr, "tracefs not mounted");
2503 return NULL; 2331 return NULL;
2504 } 2332 }
2505 2333
2506 strcat(tracefs, "/tracing/"); 2334 strcat(tracefs, "/tracing/");
2507 tracefs_found = 1; 2335 tracefs_found = 1;
2508 2336
2509 return tracefs; 2337 return tracefs;
2510 } 2338 }
2511 2339
2512 const char *tracing_file(const char * 2340 const char *tracing_file(const char *file_name)
2513 { 2341 {
2514 static char trace_file[MAX_PAT 2342 static char trace_file[MAX_PATH+1];
2515 snprintf(trace_file, MAX_PATH, 2343 snprintf(trace_file, MAX_PATH, "%s/%s", find_tracefs(), file_name);
2516 return trace_file; 2344 return trace_file;
2517 } 2345 }
2518 2346
2519 int main (int argc, char **argv) 2347 int main (int argc, char **argv)
2520 { 2348 {
2521 if (argc < 1) 2349 if (argc < 1)
2522 exit(-1); 2350 exit(-1);
2523 2351
2524 if (fork() > 0) { 2352 if (fork() > 0) {
2525 int fd, ffd; 2353 int fd, ffd;
2526 char line[64]; 2354 char line[64];
2527 int s; 2355 int s;
2528 2356
2529 ffd = open(tracing_fi 2357 ffd = open(tracing_file("current_tracer"), O_WRONLY);
2530 if (ffd < 0) 2358 if (ffd < 0)
2531 exit(-1); 2359 exit(-1);
2532 write(ffd, "nop", 3); 2360 write(ffd, "nop", 3);
2533 2361
2534 fd = open(tracing_fil 2362 fd = open(tracing_file("set_ftrace_pid"), O_WRONLY);
2535 s = sprintf(line, "%d 2363 s = sprintf(line, "%d\n", getpid());
2536 write(fd, line, s); 2364 write(fd, line, s);
2537 2365
2538 write(ffd, "function" 2366 write(ffd, "function", 8);
2539 2367
2540 close(fd); 2368 close(fd);
2541 close(ffd); 2369 close(ffd);
2542 2370
2543 execvp(argv[1], argv+ 2371 execvp(argv[1], argv+1);
2544 } 2372 }
2545 2373
2546 return 0; 2374 return 0;
2547 } 2375 }
2548 2376
2549 Or this simple script! 2377 Or this simple script!
2550 :: 2378 ::
2551 2379
2552 #!/bin/bash 2380 #!/bin/bash
2553 2381
2554 tracefs=`sed -ne 's/^tracefs \(.*\) tracefs 2382 tracefs=`sed -ne 's/^tracefs \(.*\) tracefs.*/\1/p' /proc/mounts`
2555 echo 0 > $tracefs/tracing_on !! 2383 echo nop > $tracefs/tracing/current_tracer
2556 echo $$ > $tracefs/set_ftrace_pid !! 2384 echo 0 > $tracefs/tracing/tracing_on
2557 echo function > $tracefs/current_tracer !! 2385 echo $$ > $tracefs/tracing/set_ftrace_pid
2558 echo 1 > $tracefs/tracing_on !! 2386 echo function > $tracefs/tracing/current_tracer
>> 2387 echo 1 > $tracefs/tracing/tracing_on
2559 exec "$@" 2388 exec "$@"
2560 2389
2561 2390
2562 function graph tracer 2391 function graph tracer
2563 --------------------------- 2392 ---------------------------
2564 2393
2565 This tracer is similar to the function tracer 2394 This tracer is similar to the function tracer except that it
2566 probes a function on its entry and its exit. 2395 probes a function on its entry and its exit. This is done by
2567 using a dynamically allocated stack of return 2396 using a dynamically allocated stack of return addresses in each
2568 task_struct. On function entry the tracer ove 2397 task_struct. On function entry the tracer overwrites the return
2569 address of each function traced to set a cust 2398 address of each function traced to set a custom probe. Thus the
2570 original return address is stored on the stac 2399 original return address is stored on the stack of return address
2571 in the task_struct. 2400 in the task_struct.
2572 2401
2573 Probing on both ends of a function leads to s 2402 Probing on both ends of a function leads to special features
2574 such as: 2403 such as:
2575 2404
2576 - measure of a function's time execution 2405 - measure of a function's time execution
2577 - having a reliable call stack to draw functi 2406 - having a reliable call stack to draw function calls graph
2578 2407
2579 This tracer is useful in several situations: 2408 This tracer is useful in several situations:
2580 2409
2581 - you want to find the reason of a strange ke 2410 - you want to find the reason of a strange kernel behavior and
2582 need to see what happens in detail on any a 2411 need to see what happens in detail on any areas (or specific
2583 ones). 2412 ones).
2584 2413
2585 - you are experiencing weird latencies but it 2414 - you are experiencing weird latencies but it's difficult to
2586 find its origin. 2415 find its origin.
2587 2416
2588 - you want to find quickly which path is take 2417 - you want to find quickly which path is taken by a specific
2589 function 2418 function
2590 2419
2591 - you just want to peek inside a working kern 2420 - you just want to peek inside a working kernel and want to see
2592 what happens there. 2421 what happens there.
2593 2422
2594 :: 2423 ::
2595 2424
2596 # tracer: function_graph 2425 # tracer: function_graph
2597 # 2426 #
2598 # CPU DURATION FUNCTION C 2427 # CPU DURATION FUNCTION CALLS
2599 # | | | | | | 2428 # | | | | | | |
2600 2429
2601 0) | sys_open() { 2430 0) | sys_open() {
2602 0) | do_sys_open() { 2431 0) | do_sys_open() {
2603 0) | getname() { 2432 0) | getname() {
2604 0) | kmem_cache_alloc 2433 0) | kmem_cache_alloc() {
2605 0) 1.382 us | __might_sleep( 2434 0) 1.382 us | __might_sleep();
2606 0) 2.478 us | } 2435 0) 2.478 us | }
2607 0) | strncpy_from_use 2436 0) | strncpy_from_user() {
2608 0) | might_fault() 2437 0) | might_fault() {
2609 0) 1.389 us | __might_slee 2438 0) 1.389 us | __might_sleep();
2610 0) 2.553 us | } 2439 0) 2.553 us | }
2611 0) 3.807 us | } 2440 0) 3.807 us | }
2612 0) 7.876 us | } 2441 0) 7.876 us | }
2613 0) | alloc_fd() { 2442 0) | alloc_fd() {
2614 0) 0.668 us | _spin_lock(); 2443 0) 0.668 us | _spin_lock();
2615 0) 0.570 us | expand_files(); 2444 0) 0.570 us | expand_files();
2616 0) 0.586 us | _spin_unlock(); 2445 0) 0.586 us | _spin_unlock();
2617 2446
2618 2447
2619 There are several columns that can be dynamic 2448 There are several columns that can be dynamically
2620 enabled/disabled. You can use every combinati 2449 enabled/disabled. You can use every combination of options you
2621 want, depending on your needs. 2450 want, depending on your needs.
2622 2451
2623 - The cpu number on which the function execut 2452 - The cpu number on which the function executed is default
2624 enabled. It is sometimes better to only tr 2453 enabled. It is sometimes better to only trace one cpu (see
2625 tracing_cpumask file) or you might sometime !! 2454 tracing_cpu_mask file) or you might sometimes see unordered
2626 function calls while cpu tracing switch. 2455 function calls while cpu tracing switch.
2627 2456
2628 - hide: echo nofuncgraph-cpu > trace_ 2457 - hide: echo nofuncgraph-cpu > trace_options
2629 - show: echo funcgraph-cpu > trace_op 2458 - show: echo funcgraph-cpu > trace_options
2630 2459
2631 - The duration (function's time of execution) 2460 - The duration (function's time of execution) is displayed on
2632 the closing bracket line of a function or o 2461 the closing bracket line of a function or on the same line
2633 than the current function in case of a leaf 2462 than the current function in case of a leaf one. It is default
2634 enabled. 2463 enabled.
2635 2464
2636 - hide: echo nofuncgraph-duration > t 2465 - hide: echo nofuncgraph-duration > trace_options
2637 - show: echo funcgraph-duration > tra 2466 - show: echo funcgraph-duration > trace_options
2638 2467
2639 - The overhead field precedes the duration fi 2468 - The overhead field precedes the duration field in case of
2640 reached duration thresholds. 2469 reached duration thresholds.
2641 2470
2642 - hide: echo nofuncgraph-overhead > t 2471 - hide: echo nofuncgraph-overhead > trace_options
2643 - show: echo funcgraph-overhead > tra 2472 - show: echo funcgraph-overhead > trace_options
2644 - depends on: funcgraph-duration 2473 - depends on: funcgraph-duration
2645 2474
2646 ie:: 2475 ie::
2647 2476
2648 3) # 1837.709 us | } /* __switch 2477 3) # 1837.709 us | } /* __switch_to */
2649 3) | finish_task_s 2478 3) | finish_task_switch() {
2650 3) 0.313 us | _raw_spin_u 2479 3) 0.313 us | _raw_spin_unlock_irq();
2651 3) 3.177 us | } 2480 3) 3.177 us | }
2652 3) # 1889.063 us | } /* __schedule 2481 3) # 1889.063 us | } /* __schedule */
2653 3) ! 140.417 us | } /* __schedule * 2482 3) ! 140.417 us | } /* __schedule */
2654 3) # 2034.948 us | } /* schedule */ 2483 3) # 2034.948 us | } /* schedule */
2655 3) * 33998.59 us | } /* schedule_preempt 2484 3) * 33998.59 us | } /* schedule_preempt_disabled */
2656 2485
2657 [...] 2486 [...]
2658 2487
2659 1) 0.260 us | msecs_to_ 2488 1) 0.260 us | msecs_to_jiffies();
2660 1) 0.313 us | __rcu_rea 2489 1) 0.313 us | __rcu_read_unlock();
2661 1) + 61.770 us | } 2490 1) + 61.770 us | }
2662 1) + 64.479 us | } 2491 1) + 64.479 us | }
2663 1) 0.313 us | rcu_bh_qs(); 2492 1) 0.313 us | rcu_bh_qs();
2664 1) 0.313 us | __local_bh_en 2493 1) 0.313 us | __local_bh_enable();
2665 1) ! 217.240 us | } 2494 1) ! 217.240 us | }
2666 1) 0.365 us | idle_cpu(); 2495 1) 0.365 us | idle_cpu();
2667 1) | rcu_irq_exit() 2496 1) | rcu_irq_exit() {
2668 1) 0.417 us | rcu_eqs_enter 2497 1) 0.417 us | rcu_eqs_enter_common.isra.47();
2669 1) 3.125 us | } 2498 1) 3.125 us | }
2670 1) ! 227.812 us | } 2499 1) ! 227.812 us | }
2671 1) ! 457.395 us | } 2500 1) ! 457.395 us | }
2672 1) @ 119760.2 us | } 2501 1) @ 119760.2 us | }
2673 2502
2674 [...] 2503 [...]
2675 2504
2676 2) | handle_IPI() { 2505 2) | handle_IPI() {
2677 1) 6.979 us | } 2506 1) 6.979 us | }
2678 2) 0.417 us | scheduler_ipi(); 2507 2) 0.417 us | scheduler_ipi();
2679 1) 9.791 us | } 2508 1) 9.791 us | }
2680 1) + 12.917 us | } 2509 1) + 12.917 us | }
2681 2) 3.490 us | } 2510 2) 3.490 us | }
2682 1) + 15.729 us | } 2511 1) + 15.729 us | }
2683 1) + 18.542 us | } 2512 1) + 18.542 us | }
2684 2) $ 3594274 us | } 2513 2) $ 3594274 us | }
2685 2514
2686 Flags:: 2515 Flags::
2687 2516
2688 + means that the function exceeded 10 usecs 2517 + means that the function exceeded 10 usecs.
2689 ! means that the function exceeded 100 usec 2518 ! means that the function exceeded 100 usecs.
2690 # means that the function exceeded 1000 use 2519 # means that the function exceeded 1000 usecs.
2691 * means that the function exceeded 10 msecs 2520 * means that the function exceeded 10 msecs.
2692 @ means that the function exceeded 100 msec 2521 @ means that the function exceeded 100 msecs.
2693 $ means that the function exceeded 1 sec. 2522 $ means that the function exceeded 1 sec.
2694 2523
2695 2524
2696 - The task/pid field displays the thread cmdl 2525 - The task/pid field displays the thread cmdline and pid which
2697 executed the function. It is default disabl 2526 executed the function. It is default disabled.
2698 2527
2699 - hide: echo nofuncgraph-proc > trace 2528 - hide: echo nofuncgraph-proc > trace_options
2700 - show: echo funcgraph-proc > trace_o 2529 - show: echo funcgraph-proc > trace_options
2701 2530
2702 ie:: 2531 ie::
2703 2532
2704 # tracer: function_graph 2533 # tracer: function_graph
2705 # 2534 #
2706 # CPU TASK/PID DURATION 2535 # CPU TASK/PID DURATION FUNCTION CALLS
2707 # | | | | | 2536 # | | | | | | | | |
2708 0) sh-4802 | | 2537 0) sh-4802 | | d_free() {
2709 0) sh-4802 | | 2538 0) sh-4802 | | call_rcu() {
2710 0) sh-4802 | | 2539 0) sh-4802 | | __call_rcu() {
2711 0) sh-4802 | 0.616 us | 2540 0) sh-4802 | 0.616 us | rcu_process_gp_end();
2712 0) sh-4802 | 0.586 us | 2541 0) sh-4802 | 0.586 us | check_for_new_grace_period();
2713 0) sh-4802 | 2.899 us | 2542 0) sh-4802 | 2.899 us | }
2714 0) sh-4802 | 4.040 us | 2543 0) sh-4802 | 4.040 us | }
2715 0) sh-4802 | 5.151 us | 2544 0) sh-4802 | 5.151 us | }
2716 0) sh-4802 | + 49.370 us | 2545 0) sh-4802 | + 49.370 us | }
2717 2546
2718 2547
2719 - The absolute time field is an absolute time 2548 - The absolute time field is an absolute timestamp given by the
2720 system clock since it started. A snapshot o 2549 system clock since it started. A snapshot of this time is
2721 given on each entry/exit of functions 2550 given on each entry/exit of functions
2722 2551
2723 - hide: echo nofuncgraph-abstime > tr 2552 - hide: echo nofuncgraph-abstime > trace_options
2724 - show: echo funcgraph-abstime > trac 2553 - show: echo funcgraph-abstime > trace_options
2725 2554
2726 ie:: 2555 ie::
2727 2556
2728 # 2557 #
2729 # TIME CPU DURATION 2558 # TIME CPU DURATION FUNCTION CALLS
2730 # | | | | 2559 # | | | | | | | |
2731 360.774522 | 1) 0.541 us | 2560 360.774522 | 1) 0.541 us | }
2732 360.774522 | 1) 4.663 us | 2561 360.774522 | 1) 4.663 us | }
2733 360.774523 | 1) 0.541 us | 2562 360.774523 | 1) 0.541 us | __wake_up_bit();
2734 360.774524 | 1) 6.796 us | 2563 360.774524 | 1) 6.796 us | }
2735 360.774524 | 1) 7.952 us | 2564 360.774524 | 1) 7.952 us | }
2736 360.774525 | 1) 9.063 us | 2565 360.774525 | 1) 9.063 us | }
2737 360.774525 | 1) 0.615 us | 2566 360.774525 | 1) 0.615 us | journal_mark_dirty();
2738 360.774527 | 1) 0.578 us | 2567 360.774527 | 1) 0.578 us | __brelse();
2739 360.774528 | 1) | 2568 360.774528 | 1) | reiserfs_prepare_for_journal() {
2740 360.774528 | 1) | 2569 360.774528 | 1) | unlock_buffer() {
2741 360.774529 | 1) | 2570 360.774529 | 1) | wake_up_bit() {
2742 360.774529 | 1) | 2571 360.774529 | 1) | bit_waitqueue() {
2743 360.774530 | 1) 0.594 us | 2572 360.774530 | 1) 0.594 us | __phys_addr();
2744 2573
2745 2574
2746 The function name is always displayed after t 2575 The function name is always displayed after the closing bracket
2747 for a function if the start of that function 2576 for a function if the start of that function is not in the
2748 trace buffer. 2577 trace buffer.
2749 2578
2750 Display of the function name after the closin 2579 Display of the function name after the closing bracket may be
2751 enabled for functions whose start is in the t 2580 enabled for functions whose start is in the trace buffer,
2752 allowing easier searching with grep for funct 2581 allowing easier searching with grep for function durations.
2753 It is default disabled. 2582 It is default disabled.
2754 2583
2755 - hide: echo nofuncgraph-tail > trace 2584 - hide: echo nofuncgraph-tail > trace_options
2756 - show: echo funcgraph-tail > trace_o 2585 - show: echo funcgraph-tail > trace_options
2757 2586
2758 Example with nofuncgraph-tail (default):: 2587 Example with nofuncgraph-tail (default)::
2759 2588
2760 0) | putname() { 2589 0) | putname() {
2761 0) | kmem_cache_free 2590 0) | kmem_cache_free() {
2762 0) 0.518 us | __phys_addr() 2591 0) 0.518 us | __phys_addr();
2763 0) 1.757 us | } 2592 0) 1.757 us | }
2764 0) 2.861 us | } 2593 0) 2.861 us | }
2765 2594
2766 Example with funcgraph-tail:: 2595 Example with funcgraph-tail::
2767 2596
2768 0) | putname() { 2597 0) | putname() {
2769 0) | kmem_cache_free 2598 0) | kmem_cache_free() {
2770 0) 0.518 us | __phys_addr() 2599 0) 0.518 us | __phys_addr();
2771 0) 1.757 us | } /* kmem_cache 2600 0) 1.757 us | } /* kmem_cache_free() */
2772 0) 2.861 us | } /* putname() */ 2601 0) 2.861 us | } /* putname() */
2773 2602
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 2603 You can put some comments on specific functions by using
2888 trace_printk() For example, if you want to pu 2604 trace_printk() For example, if you want to put a comment inside
2889 the __might_sleep() function, you just have t 2605 the __might_sleep() function, you just have to include
2890 <linux/ftrace.h> and call trace_printk() insi 2606 <linux/ftrace.h> and call trace_printk() inside __might_sleep()::
2891 2607
2892 trace_printk("I'm a comment!\n") 2608 trace_printk("I'm a comment!\n")
2893 2609
2894 will produce:: 2610 will produce::
2895 2611
2896 1) | __might_sle 2612 1) | __might_sleep() {
2897 1) | /* I'm a 2613 1) | /* I'm a comment! */
2898 1) 1.449 us | } 2614 1) 1.449 us | }
2899 2615
2900 2616
2901 You might find other useful features for this 2617 You might find other useful features for this tracer in the
2902 following "dynamic ftrace" section such as tr 2618 following "dynamic ftrace" section such as tracing only specific
2903 functions or tasks. 2619 functions or tasks.
2904 2620
2905 dynamic ftrace 2621 dynamic ftrace
2906 -------------- 2622 --------------
2907 2623
2908 If CONFIG_DYNAMIC_FTRACE is set, the system w 2624 If CONFIG_DYNAMIC_FTRACE is set, the system will run with
2909 virtually no overhead when function tracing i 2625 virtually no overhead when function tracing is disabled. The way
2910 this works is the mcount function call (place 2626 this works is the mcount function call (placed at the start of
2911 every kernel function, produced by the -pg sw 2627 every kernel function, produced by the -pg switch in gcc),
2912 starts of pointing to a simple return. (Enabl 2628 starts of pointing to a simple return. (Enabling FTRACE will
2913 include the -pg switch in the compiling of th 2629 include the -pg switch in the compiling of the kernel.)
2914 2630
2915 At compile time every C file object is run th 2631 At compile time every C file object is run through the
2916 recordmcount program (located in the scripts 2632 recordmcount program (located in the scripts directory). This
2917 program will parse the ELF headers in the C o 2633 program will parse the ELF headers in the C object to find all
2918 the locations in the .text section that call 2634 the locations in the .text section that call mcount. Starting
2919 with gcc version 4.6, the -mfentry has been a 2635 with gcc version 4.6, the -mfentry has been added for x86, which
2920 calls "__fentry__" instead of "mcount". Which 2636 calls "__fentry__" instead of "mcount". Which is called before
2921 the creation of the stack frame. 2637 the creation of the stack frame.
2922 2638
2923 Note, not all sections are traced. They may b 2639 Note, not all sections are traced. They may be prevented by either
2924 a notrace, or blocked another way and all inl 2640 a notrace, or blocked another way and all inline functions are not
2925 traced. Check the "available_filter_functions 2641 traced. Check the "available_filter_functions" file to see what functions
2926 can be traced. 2642 can be traced.
2927 2643
2928 A section called "__mcount_loc" is created th 2644 A section called "__mcount_loc" is created that holds
2929 references to all the mcount/fentry call site 2645 references to all the mcount/fentry call sites in the .text section.
2930 The recordmcount program re-links this sectio 2646 The recordmcount program re-links this section back into the
2931 original object. The final linking stage of t 2647 original object. The final linking stage of the kernel will add all these
2932 references into a single table. 2648 references into a single table.
2933 2649
2934 On boot up, before SMP is initialized, the dy 2650 On boot up, before SMP is initialized, the dynamic ftrace code
2935 scans this table and updates all the location 2651 scans this table and updates all the locations into nops. It
2936 also records the locations, which are added t 2652 also records the locations, which are added to the
2937 available_filter_functions list. Modules are 2653 available_filter_functions list. Modules are processed as they
2938 are loaded and before they are executed. Whe 2654 are loaded and before they are executed. When a module is
2939 unloaded, it also removes its functions from 2655 unloaded, it also removes its functions from the ftrace function
2940 list. This is automatic in the module unload 2656 list. This is automatic in the module unload code, and the
2941 module author does not need to worry about it 2657 module author does not need to worry about it.
2942 2658
2943 When tracing is enabled, the process of modif 2659 When tracing is enabled, the process of modifying the function
2944 tracepoints is dependent on architecture. The 2660 tracepoints is dependent on architecture. The old method is to use
2945 kstop_machine to prevent races with the CPUs 2661 kstop_machine to prevent races with the CPUs executing code being
2946 modified (which can cause the CPU to do undes 2662 modified (which can cause the CPU to do undesirable things, especially
2947 if the modified code crosses cache (or page) 2663 if the modified code crosses cache (or page) boundaries), and the nops are
2948 patched back to calls. But this time, they do 2664 patched back to calls. But this time, they do not call mcount
2949 (which is just a function stub). They now cal 2665 (which is just a function stub). They now call into the ftrace
2950 infrastructure. 2666 infrastructure.
2951 2667
2952 The new method of modifying the function trac 2668 The new method of modifying the function tracepoints is to place
2953 a breakpoint at the location to be modified, 2669 a breakpoint at the location to be modified, sync all CPUs, modify
2954 the rest of the instruction not covered by th 2670 the rest of the instruction not covered by the breakpoint. Sync
2955 all CPUs again, and then remove the breakpoin 2671 all CPUs again, and then remove the breakpoint with the finished
2956 version to the ftrace call site. 2672 version to the ftrace call site.
2957 2673
2958 Some archs do not even need to monkey around 2674 Some archs do not even need to monkey around with the synchronization,
2959 and can just slap the new code on top of the 2675 and can just slap the new code on top of the old without any
2960 problems with other CPUs executing it at the 2676 problems with other CPUs executing it at the same time.
2961 2677
2962 One special side-effect to the recording of t 2678 One special side-effect to the recording of the functions being
2963 traced is that we can now selectively choose 2679 traced is that we can now selectively choose which functions we
2964 wish to trace and which ones we want the mcou 2680 wish to trace and which ones we want the mcount calls to remain
2965 as nops. 2681 as nops.
2966 2682
2967 Two files are used, one for enabling and one 2683 Two files are used, one for enabling and one for disabling the
2968 tracing of specified functions. They are: 2684 tracing of specified functions. They are:
2969 2685
2970 set_ftrace_filter 2686 set_ftrace_filter
2971 2687
2972 and 2688 and
2973 2689
2974 set_ftrace_notrace 2690 set_ftrace_notrace
2975 2691
2976 A list of available functions that you can ad 2692 A list of available functions that you can add to these files is
2977 listed in: 2693 listed in:
2978 2694
2979 available_filter_functions 2695 available_filter_functions
2980 2696
2981 :: 2697 ::
2982 2698
2983 # cat available_filter_functions 2699 # cat available_filter_functions
2984 put_prev_task_idle 2700 put_prev_task_idle
2985 kmem_cache_create 2701 kmem_cache_create
2986 pick_next_task_rt 2702 pick_next_task_rt
2987 cpus_read_lock !! 2703 get_online_cpus
2988 pick_next_task_fair 2704 pick_next_task_fair
2989 mutex_lock 2705 mutex_lock
2990 [...] 2706 [...]
2991 2707
2992 If I am only interested in sys_nanosleep and 2708 If I am only interested in sys_nanosleep and hrtimer_interrupt::
2993 2709
2994 # echo sys_nanosleep hrtimer_interrupt > se 2710 # echo sys_nanosleep hrtimer_interrupt > set_ftrace_filter
2995 # echo function > current_tracer 2711 # echo function > current_tracer
2996 # echo 1 > tracing_on 2712 # echo 1 > tracing_on
2997 # usleep 1 2713 # usleep 1
2998 # echo 0 > tracing_on 2714 # echo 0 > tracing_on
2999 # cat trace 2715 # cat trace
3000 # tracer: function 2716 # tracer: function
3001 # 2717 #
3002 # entries-in-buffer/entries-written: 5/5 2718 # entries-in-buffer/entries-written: 5/5 #P:4
3003 # 2719 #
3004 # _-----=> irq 2720 # _-----=> irqs-off
3005 # / _----=> nee 2721 # / _----=> need-resched
3006 # | / _---=> har 2722 # | / _---=> hardirq/softirq
3007 # || / _--=> pre 2723 # || / _--=> preempt-depth
3008 # ||| / dela 2724 # ||| / delay
3009 # TASK-PID CPU# |||| TIMEST 2725 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3010 # | | | |||| | 2726 # | | | |||| | |
3011 usleep-2665 [001] .... 4186.475 2727 usleep-2665 [001] .... 4186.475355: sys_nanosleep <-system_call_fastpath
3012 <idle>-0 [001] d.h1 4186.475 2728 <idle>-0 [001] d.h1 4186.475409: hrtimer_interrupt <-smp_apic_timer_interrupt
3013 usleep-2665 [001] d.h1 4186.475 2729 usleep-2665 [001] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
3014 <idle>-0 [003] d.h1 4186.475 2730 <idle>-0 [003] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
3015 <idle>-0 [002] d.h1 4186.475 2731 <idle>-0 [002] d.h1 4186.475427: hrtimer_interrupt <-smp_apic_timer_interrupt
3016 2732
3017 To see which functions are being traced, you 2733 To see which functions are being traced, you can cat the file:
3018 :: 2734 ::
3019 2735
3020 # cat set_ftrace_filter 2736 # cat set_ftrace_filter
3021 hrtimer_interrupt 2737 hrtimer_interrupt
3022 sys_nanosleep 2738 sys_nanosleep
3023 2739
3024 2740
3025 Perhaps this is not enough. The filters also 2741 Perhaps this is not enough. The filters also allow glob(7) matching.
3026 2742
3027 ``<match>*`` 2743 ``<match>*``
3028 will match functions that begin with 2744 will match functions that begin with <match>
3029 ``*<match>`` 2745 ``*<match>``
3030 will match functions that end with <m 2746 will match functions that end with <match>
3031 ``*<match>*`` 2747 ``*<match>*``
3032 will match functions that have <match 2748 will match functions that have <match> in it
3033 ``<match1>*<match2>`` 2749 ``<match1>*<match2>``
3034 will match functions that begin with 2750 will match functions that begin with <match1> and end with <match2>
3035 2751
3036 .. note:: 2752 .. note::
3037 It is better to use quotes to enclose t 2753 It is better to use quotes to enclose the wild cards,
3038 otherwise the shell may expand the para 2754 otherwise the shell may expand the parameters into names
3039 of files in the local directory. 2755 of files in the local directory.
3040 2756
3041 :: 2757 ::
3042 2758
3043 # echo 'hrtimer_*' > set_ftrace_filter 2759 # echo 'hrtimer_*' > set_ftrace_filter
3044 2760
3045 Produces:: 2761 Produces::
3046 2762
3047 # tracer: function 2763 # tracer: function
3048 # 2764 #
3049 # entries-in-buffer/entries-written: 897/89 2765 # entries-in-buffer/entries-written: 897/897 #P:4
3050 # 2766 #
3051 # _-----=> irq 2767 # _-----=> irqs-off
3052 # / _----=> nee 2768 # / _----=> need-resched
3053 # | / _---=> har 2769 # | / _---=> hardirq/softirq
3054 # || / _--=> pre 2770 # || / _--=> preempt-depth
3055 # ||| / dela 2771 # ||| / delay
3056 # TASK-PID CPU# |||| TIMEST 2772 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3057 # | | | |||| | 2773 # | | | |||| | |
3058 <idle>-0 [003] dN.1 4228.547 2774 <idle>-0 [003] dN.1 4228.547803: hrtimer_cancel <-tick_nohz_idle_exit
3059 <idle>-0 [003] dN.1 4228.547 2775 <idle>-0 [003] dN.1 4228.547804: hrtimer_try_to_cancel <-hrtimer_cancel
3060 <idle>-0 [003] dN.2 4228.547 2776 <idle>-0 [003] dN.2 4228.547805: hrtimer_force_reprogram <-__remove_hrtimer
3061 <idle>-0 [003] dN.1 4228.547 2777 <idle>-0 [003] dN.1 4228.547805: hrtimer_forward <-tick_nohz_idle_exit
3062 <idle>-0 [003] dN.1 4228.547 2778 <idle>-0 [003] dN.1 4228.547805: hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
3063 <idle>-0 [003] d..1 4228.547 2779 <idle>-0 [003] d..1 4228.547858: hrtimer_get_next_event <-get_next_timer_interrupt
3064 <idle>-0 [003] d..1 4228.547 2780 <idle>-0 [003] d..1 4228.547859: hrtimer_start <-__tick_nohz_idle_enter
3065 <idle>-0 [003] d..2 4228.547 2781 <idle>-0 [003] d..2 4228.547860: hrtimer_force_reprogram <-__rem
3066 2782
3067 Notice that we lost the sys_nanosleep. 2783 Notice that we lost the sys_nanosleep.
3068 :: 2784 ::
3069 2785
3070 # cat set_ftrace_filter 2786 # cat set_ftrace_filter
3071 hrtimer_run_queues 2787 hrtimer_run_queues
3072 hrtimer_run_pending 2788 hrtimer_run_pending
3073 hrtimer_init 2789 hrtimer_init
3074 hrtimer_cancel 2790 hrtimer_cancel
3075 hrtimer_try_to_cancel 2791 hrtimer_try_to_cancel
3076 hrtimer_forward 2792 hrtimer_forward
3077 hrtimer_start 2793 hrtimer_start
3078 hrtimer_reprogram 2794 hrtimer_reprogram
3079 hrtimer_force_reprogram 2795 hrtimer_force_reprogram
3080 hrtimer_get_next_event 2796 hrtimer_get_next_event
3081 hrtimer_interrupt 2797 hrtimer_interrupt
3082 hrtimer_nanosleep 2798 hrtimer_nanosleep
3083 hrtimer_wakeup 2799 hrtimer_wakeup
3084 hrtimer_get_remaining 2800 hrtimer_get_remaining
3085 hrtimer_get_res 2801 hrtimer_get_res
3086 hrtimer_init_sleeper 2802 hrtimer_init_sleeper
3087 2803
3088 2804
3089 This is because the '>' and '>>' act just lik 2805 This is because the '>' and '>>' act just like they do in bash.
3090 To rewrite the filters, use '>' 2806 To rewrite the filters, use '>'
3091 To append to the filters, use '>>' 2807 To append to the filters, use '>>'
3092 2808
3093 To clear out a filter so that all functions w 2809 To clear out a filter so that all functions will be recorded
3094 again:: 2810 again::
3095 2811
3096 # echo > set_ftrace_filter 2812 # echo > set_ftrace_filter
3097 # cat set_ftrace_filter 2813 # cat set_ftrace_filter
3098 # 2814 #
3099 2815
3100 Again, now we want to append. 2816 Again, now we want to append.
3101 2817
3102 :: 2818 ::
3103 2819
3104 # echo sys_nanosleep > set_ftrace_filter 2820 # echo sys_nanosleep > set_ftrace_filter
3105 # cat set_ftrace_filter 2821 # cat set_ftrace_filter
3106 sys_nanosleep 2822 sys_nanosleep
3107 # echo 'hrtimer_*' >> set_ftrace_filter 2823 # echo 'hrtimer_*' >> set_ftrace_filter
3108 # cat set_ftrace_filter 2824 # cat set_ftrace_filter
3109 hrtimer_run_queues 2825 hrtimer_run_queues
3110 hrtimer_run_pending 2826 hrtimer_run_pending
3111 hrtimer_init 2827 hrtimer_init
3112 hrtimer_cancel 2828 hrtimer_cancel
3113 hrtimer_try_to_cancel 2829 hrtimer_try_to_cancel
3114 hrtimer_forward 2830 hrtimer_forward
3115 hrtimer_start 2831 hrtimer_start
3116 hrtimer_reprogram 2832 hrtimer_reprogram
3117 hrtimer_force_reprogram 2833 hrtimer_force_reprogram
3118 hrtimer_get_next_event 2834 hrtimer_get_next_event
3119 hrtimer_interrupt 2835 hrtimer_interrupt
3120 sys_nanosleep 2836 sys_nanosleep
3121 hrtimer_nanosleep 2837 hrtimer_nanosleep
3122 hrtimer_wakeup 2838 hrtimer_wakeup
3123 hrtimer_get_remaining 2839 hrtimer_get_remaining
3124 hrtimer_get_res 2840 hrtimer_get_res
3125 hrtimer_init_sleeper 2841 hrtimer_init_sleeper
3126 2842
3127 2843
3128 The set_ftrace_notrace prevents those functio 2844 The set_ftrace_notrace prevents those functions from being
3129 traced. 2845 traced.
3130 :: 2846 ::
3131 2847
3132 # echo '*preempt*' '*lock*' > set_ftrace_no 2848 # echo '*preempt*' '*lock*' > set_ftrace_notrace
3133 2849
3134 Produces:: 2850 Produces::
3135 2851
3136 # tracer: function 2852 # tracer: function
3137 # 2853 #
3138 # entries-in-buffer/entries-written: 39608/ 2854 # entries-in-buffer/entries-written: 39608/39608 #P:4
3139 # 2855 #
3140 # _-----=> irq 2856 # _-----=> irqs-off
3141 # / _----=> nee 2857 # / _----=> need-resched
3142 # | / _---=> har 2858 # | / _---=> hardirq/softirq
3143 # || / _--=> pre 2859 # || / _--=> preempt-depth
3144 # ||| / dela 2860 # ||| / delay
3145 # TASK-PID CPU# |||| TIMEST 2861 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3146 # | | | |||| | 2862 # | | | |||| | |
3147 bash-1994 [000] .... 4342.324 2863 bash-1994 [000] .... 4342.324896: file_ra_state_init <-do_dentry_open
3148 bash-1994 [000] .... 4342.324 2864 bash-1994 [000] .... 4342.324897: open_check_o_direct <-do_last
3149 bash-1994 [000] .... 4342.324 2865 bash-1994 [000] .... 4342.324897: ima_file_check <-do_last
3150 bash-1994 [000] .... 4342.324 2866 bash-1994 [000] .... 4342.324898: process_measurement <-ima_file_check
3151 bash-1994 [000] .... 4342.324 2867 bash-1994 [000] .... 4342.324898: ima_get_action <-process_measurement
3152 bash-1994 [000] .... 4342.324 2868 bash-1994 [000] .... 4342.324898: ima_match_policy <-ima_get_action
3153 bash-1994 [000] .... 4342.324 2869 bash-1994 [000] .... 4342.324899: do_truncate <-do_last
3154 bash-1994 [000] .... 4342.324 !! 2870 bash-1994 [000] .... 4342.324899: should_remove_suid <-do_truncate
3155 bash-1994 [000] .... 4342.324 2871 bash-1994 [000] .... 4342.324899: notify_change <-do_truncate
3156 bash-1994 [000] .... 4342.324 2872 bash-1994 [000] .... 4342.324900: current_fs_time <-notify_change
3157 bash-1994 [000] .... 4342.324 2873 bash-1994 [000] .... 4342.324900: current_kernel_time <-current_fs_time
3158 bash-1994 [000] .... 4342.324 2874 bash-1994 [000] .... 4342.324900: timespec_trunc <-current_fs_time
3159 2875
3160 We can see that there's no more lock or preem 2876 We can see that there's no more lock or preempt tracing.
3161 2877
3162 Selecting function filters via index 2878 Selecting function filters via index
3163 ------------------------------------ 2879 ------------------------------------
3164 2880
3165 Because processing of strings is expensive (t 2881 Because processing of strings is expensive (the address of the function
3166 needs to be looked up before comparing to the 2882 needs to be looked up before comparing to the string being passed in),
3167 an index can be used as well to enable functi 2883 an index can be used as well to enable functions. This is useful in the
3168 case of setting thousands of specific functio 2884 case of setting thousands of specific functions at a time. By passing
3169 in a list of numbers, no string processing wi 2885 in a list of numbers, no string processing will occur. Instead, the function
3170 at the specific location in the internal arra 2886 at the specific location in the internal array (which corresponds to the
3171 functions in the "available_filter_functions" 2887 functions in the "available_filter_functions" file), is selected.
3172 2888
3173 :: 2889 ::
3174 2890
3175 # echo 1 > set_ftrace_filter 2891 # echo 1 > set_ftrace_filter
3176 2892
3177 Will select the first function listed in "ava 2893 Will select the first function listed in "available_filter_functions"
3178 2894
3179 :: 2895 ::
3180 2896
3181 # head -1 available_filter_functions 2897 # head -1 available_filter_functions
3182 trace_initcall_finish_cb 2898 trace_initcall_finish_cb
3183 2899
3184 # cat set_ftrace_filter 2900 # cat set_ftrace_filter
3185 trace_initcall_finish_cb 2901 trace_initcall_finish_cb
3186 2902
3187 # head -50 available_filter_functions | tai 2903 # head -50 available_filter_functions | tail -1
3188 x86_pmu_commit_txn 2904 x86_pmu_commit_txn
3189 2905
3190 # echo 1 50 > set_ftrace_filter 2906 # echo 1 50 > set_ftrace_filter
3191 # cat set_ftrace_filter 2907 # cat set_ftrace_filter
3192 trace_initcall_finish_cb 2908 trace_initcall_finish_cb
3193 x86_pmu_commit_txn 2909 x86_pmu_commit_txn
3194 2910
3195 Dynamic ftrace with the function graph tracer 2911 Dynamic ftrace with the function graph tracer
3196 --------------------------------------------- 2912 ---------------------------------------------
3197 2913
3198 Although what has been explained above concer 2914 Although what has been explained above concerns both the
3199 function tracer and the function-graph-tracer 2915 function tracer and the function-graph-tracer, there are some
3200 special features only available in the functi 2916 special features only available in the function-graph tracer.
3201 2917
3202 If you want to trace only one function and al 2918 If you want to trace only one function and all of its children,
3203 you just have to echo its name into set_graph 2919 you just have to echo its name into set_graph_function::
3204 2920
3205 echo __do_fault > set_graph_function 2921 echo __do_fault > set_graph_function
3206 2922
3207 will produce the following "expanded" trace o 2923 will produce the following "expanded" trace of the __do_fault()
3208 function:: 2924 function::
3209 2925
3210 0) | __do_fault() { 2926 0) | __do_fault() {
3211 0) | filemap_fault() { 2927 0) | filemap_fault() {
3212 0) | find_lock_page() { 2928 0) | find_lock_page() {
3213 0) 0.804 us | find_get_page(); 2929 0) 0.804 us | find_get_page();
3214 0) | __might_sleep() 2930 0) | __might_sleep() {
3215 0) 1.329 us | } 2931 0) 1.329 us | }
3216 0) 3.904 us | } 2932 0) 3.904 us | }
3217 0) 4.979 us | } 2933 0) 4.979 us | }
3218 0) 0.653 us | _spin_lock(); 2934 0) 0.653 us | _spin_lock();
3219 0) 0.578 us | page_add_file_rmap() 2935 0) 0.578 us | page_add_file_rmap();
3220 0) 0.525 us | native_set_pte_at(); 2936 0) 0.525 us | native_set_pte_at();
3221 0) 0.585 us | _spin_unlock(); 2937 0) 0.585 us | _spin_unlock();
3222 0) | unlock_page() { 2938 0) | unlock_page() {
3223 0) 0.541 us | page_waitqueue(); 2939 0) 0.541 us | page_waitqueue();
3224 0) 0.639 us | __wake_up_bit(); 2940 0) 0.639 us | __wake_up_bit();
3225 0) 2.786 us | } 2941 0) 2.786 us | }
3226 0) + 14.237 us | } 2942 0) + 14.237 us | }
3227 0) | __do_fault() { 2943 0) | __do_fault() {
3228 0) | filemap_fault() { 2944 0) | filemap_fault() {
3229 0) | find_lock_page() { 2945 0) | find_lock_page() {
3230 0) 0.698 us | find_get_page(); 2946 0) 0.698 us | find_get_page();
3231 0) | __might_sleep() 2947 0) | __might_sleep() {
3232 0) 1.412 us | } 2948 0) 1.412 us | }
3233 0) 3.950 us | } 2949 0) 3.950 us | }
3234 0) 5.098 us | } 2950 0) 5.098 us | }
3235 0) 0.631 us | _spin_lock(); 2951 0) 0.631 us | _spin_lock();
3236 0) 0.571 us | page_add_file_rmap() 2952 0) 0.571 us | page_add_file_rmap();
3237 0) 0.526 us | native_set_pte_at(); 2953 0) 0.526 us | native_set_pte_at();
3238 0) 0.586 us | _spin_unlock(); 2954 0) 0.586 us | _spin_unlock();
3239 0) | unlock_page() { 2955 0) | unlock_page() {
3240 0) 0.533 us | page_waitqueue(); 2956 0) 0.533 us | page_waitqueue();
3241 0) 0.638 us | __wake_up_bit(); 2957 0) 0.638 us | __wake_up_bit();
3242 0) 2.793 us | } 2958 0) 2.793 us | }
3243 0) + 14.012 us | } 2959 0) + 14.012 us | }
3244 2960
3245 You can also expand several functions at once 2961 You can also expand several functions at once::
3246 2962
3247 echo sys_open > set_graph_function 2963 echo sys_open > set_graph_function
3248 echo sys_close >> set_graph_function 2964 echo sys_close >> set_graph_function
3249 2965
3250 Now if you want to go back to trace all funct 2966 Now if you want to go back to trace all functions you can clear
3251 this special filter via:: 2967 this special filter via::
3252 2968
3253 echo > set_graph_function 2969 echo > set_graph_function
3254 2970
3255 2971
3256 ftrace_enabled 2972 ftrace_enabled
3257 -------------- 2973 --------------
3258 2974
3259 Note, the proc sysctl ftrace_enable is a big 2975 Note, the proc sysctl ftrace_enable is a big on/off switch for the
3260 function tracer. By default it is enabled (wh 2976 function tracer. By default it is enabled (when function tracing is
3261 enabled in the kernel). If it is disabled, al 2977 enabled in the kernel). If it is disabled, all function tracing is
3262 disabled. This includes not only the function 2978 disabled. This includes not only the function tracers for ftrace, but
3263 also for any other uses (perf, kprobes, stack 2979 also for any other uses (perf, kprobes, stack tracing, profiling, etc). It
3264 cannot be disabled if there is a callback wit 2980 cannot be disabled if there is a callback with FTRACE_OPS_FL_PERMANENT set
3265 registered. 2981 registered.
3266 2982
3267 Please disable this with care. 2983 Please disable this with care.
3268 2984
3269 This can be disable (and enabled) with:: 2985 This can be disable (and enabled) with::
3270 2986
3271 sysctl kernel.ftrace_enabled=0 2987 sysctl kernel.ftrace_enabled=0
3272 sysctl kernel.ftrace_enabled=1 2988 sysctl kernel.ftrace_enabled=1
3273 2989
3274 or 2990 or
3275 2991
3276 echo 0 > /proc/sys/kernel/ftrace_enabled 2992 echo 0 > /proc/sys/kernel/ftrace_enabled
3277 echo 1 > /proc/sys/kernel/ftrace_enabled 2993 echo 1 > /proc/sys/kernel/ftrace_enabled
3278 2994
3279 2995
3280 Filter commands 2996 Filter commands
3281 --------------- 2997 ---------------
3282 2998
3283 A few commands are supported by the set_ftrac 2999 A few commands are supported by the set_ftrace_filter interface.
3284 Trace commands have the following format:: 3000 Trace commands have the following format::
3285 3001
3286 <function>:<command>:<parameter> 3002 <function>:<command>:<parameter>
3287 3003
3288 The following commands are supported: 3004 The following commands are supported:
3289 3005
3290 - mod: 3006 - mod:
3291 This command enables function filtering per 3007 This command enables function filtering per module. The
3292 parameter defines the module. For example, 3008 parameter defines the module. For example, if only the write*
3293 functions in the ext3 module are desired, r 3009 functions in the ext3 module are desired, run:
3294 3010
3295 echo 'write*:mod:ext3' > set_ftrace_filter 3011 echo 'write*:mod:ext3' > set_ftrace_filter
3296 3012
3297 This command interacts with the filter in t 3013 This command interacts with the filter in the same way as
3298 filtering based on function names. Thus, ad 3014 filtering based on function names. Thus, adding more functions
3299 in a different module is accomplished by ap 3015 in a different module is accomplished by appending (>>) to the
3300 filter file. Remove specific module functio 3016 filter file. Remove specific module functions by prepending
3301 '!':: 3017 '!'::
3302 3018
3303 echo '!writeback*:mod:ext3' >> set_ftrace_ 3019 echo '!writeback*:mod:ext3' >> set_ftrace_filter
3304 3020
3305 Mod command supports module globbing. Disab 3021 Mod command supports module globbing. Disable tracing for all
3306 functions except a specific module:: 3022 functions except a specific module::
3307 3023
3308 echo '!*:mod:!ext3' >> set_ftrace_filter 3024 echo '!*:mod:!ext3' >> set_ftrace_filter
3309 3025
3310 Disable tracing for all modules, but still 3026 Disable tracing for all modules, but still trace kernel::
3311 3027
3312 echo '!*:mod:*' >> set_ftrace_filter 3028 echo '!*:mod:*' >> set_ftrace_filter
3313 3029
3314 Enable filter only for kernel:: 3030 Enable filter only for kernel::
3315 3031
3316 echo '*write*:mod:!*' >> set_ftrace_filter 3032 echo '*write*:mod:!*' >> set_ftrace_filter
3317 3033
3318 Enable filter for module globbing:: 3034 Enable filter for module globbing::
3319 3035
3320 echo '*write*:mod:*snd*' >> set_ftrace_fil 3036 echo '*write*:mod:*snd*' >> set_ftrace_filter
3321 3037
3322 - traceon/traceoff: 3038 - traceon/traceoff:
3323 These commands turn tracing on and off when 3039 These commands turn tracing on and off when the specified
3324 functions are hit. The parameter determines 3040 functions are hit. The parameter determines how many times the
3325 tracing system is turned on and off. If uns 3041 tracing system is turned on and off. If unspecified, there is
3326 no limit. For example, to disable tracing w 3042 no limit. For example, to disable tracing when a schedule bug
3327 is hit the first 5 times, run:: 3043 is hit the first 5 times, run::
3328 3044
3329 echo '__schedule_bug:traceoff:5' > set_ftr 3045 echo '__schedule_bug:traceoff:5' > set_ftrace_filter
3330 3046
3331 To always disable tracing when __schedule_b 3047 To always disable tracing when __schedule_bug is hit::
3332 3048
3333 echo '__schedule_bug:traceoff' > set_ftrac 3049 echo '__schedule_bug:traceoff' > set_ftrace_filter
3334 3050
3335 These commands are cumulative whether or no 3051 These commands are cumulative whether or not they are appended
3336 to set_ftrace_filter. To remove a command, 3052 to set_ftrace_filter. To remove a command, prepend it by '!'
3337 and drop the parameter:: 3053 and drop the parameter::
3338 3054
3339 echo '!__schedule_bug:traceoff:0' > set_ft 3055 echo '!__schedule_bug:traceoff:0' > set_ftrace_filter
3340 3056
3341 The above removes the traceoff command for 3057 The above removes the traceoff command for __schedule_bug
3342 that have a counter. To remove commands wit 3058 that have a counter. To remove commands without counters::
3343 3059
3344 echo '!__schedule_bug:traceoff' > set_ftra 3060 echo '!__schedule_bug:traceoff' > set_ftrace_filter
3345 3061
3346 - snapshot: 3062 - snapshot:
3347 Will cause a snapshot to be triggered when 3063 Will cause a snapshot to be triggered when the function is hit.
3348 :: 3064 ::
3349 3065
3350 echo 'native_flush_tlb_others:snapshot' > 3066 echo 'native_flush_tlb_others:snapshot' > set_ftrace_filter
3351 3067
3352 To only snapshot once: 3068 To only snapshot once:
3353 :: 3069 ::
3354 3070
3355 echo 'native_flush_tlb_others:snapshot:1' 3071 echo 'native_flush_tlb_others:snapshot:1' > set_ftrace_filter
3356 3072
3357 To remove the above commands:: 3073 To remove the above commands::
3358 3074
3359 echo '!native_flush_tlb_others:snapshot' > 3075 echo '!native_flush_tlb_others:snapshot' > set_ftrace_filter
3360 echo '!native_flush_tlb_others:snapshot:0' 3076 echo '!native_flush_tlb_others:snapshot:0' > set_ftrace_filter
3361 3077
3362 - enable_event/disable_event: 3078 - enable_event/disable_event:
3363 These commands can enable or disable a trac 3079 These commands can enable or disable a trace event. Note, because
3364 function tracing callbacks are very sensiti 3080 function tracing callbacks are very sensitive, when these commands
3365 are registered, the trace point is activate 3081 are registered, the trace point is activated, but disabled in
3366 a "soft" mode. That is, the tracepoint will 3082 a "soft" mode. That is, the tracepoint will be called, but
3367 just will not be traced. The event tracepoi 3083 just will not be traced. The event tracepoint stays in this mode
3368 as long as there's a command that triggers 3084 as long as there's a command that triggers it.
3369 :: 3085 ::
3370 3086
3371 echo 'try_to_wake_up:enable_event:sched:sc 3087 echo 'try_to_wake_up:enable_event:sched:sched_switch:2' > \
3372 set_ftrace_filter 3088 set_ftrace_filter
3373 3089
3374 The format is:: 3090 The format is::
3375 3091
3376 <function>:enable_event:<system>:<event>[ 3092 <function>:enable_event:<system>:<event>[:count]
3377 <function>:disable_event:<system>:<event> 3093 <function>:disable_event:<system>:<event>[:count]
3378 3094
3379 To remove the events commands:: 3095 To remove the events commands::
3380 3096
3381 echo '!try_to_wake_up:enable_event:sched:s 3097 echo '!try_to_wake_up:enable_event:sched:sched_switch:0' > \
3382 set_ftrace_filter 3098 set_ftrace_filter
3383 echo '!schedule:disable_event:sched:sched_ 3099 echo '!schedule:disable_event:sched:sched_switch' > \
3384 set_ftrace_filter 3100 set_ftrace_filter
3385 3101
3386 - dump: 3102 - dump:
3387 When the function is hit, it will dump the 3103 When the function is hit, it will dump the contents of the ftrace
3388 ring buffer to the console. This is useful 3104 ring buffer to the console. This is useful if you need to debug
3389 something, and want to dump the trace when 3105 something, and want to dump the trace when a certain function
3390 is hit. Perhaps it's a function that is cal 3106 is hit. Perhaps it's a function that is called before a triple
3391 fault happens and does not allow you to get 3107 fault happens and does not allow you to get a regular dump.
3392 3108
3393 - cpudump: 3109 - cpudump:
3394 When the function is hit, it will dump the 3110 When the function is hit, it will dump the contents of the ftrace
3395 ring buffer for the current CPU to the cons 3111 ring buffer for the current CPU to the console. Unlike the "dump"
3396 command, it only prints out the contents of 3112 command, it only prints out the contents of the ring buffer for the
3397 CPU that executed the function that trigger 3113 CPU that executed the function that triggered the dump.
3398 3114
3399 - stacktrace: 3115 - stacktrace:
3400 When the function is hit, a stack trace is 3116 When the function is hit, a stack trace is recorded.
3401 3117
3402 trace_pipe 3118 trace_pipe
3403 ---------- 3119 ----------
3404 3120
3405 The trace_pipe outputs the same content as th 3121 The trace_pipe outputs the same content as the trace file, but
3406 the effect on the tracing is different. Every 3122 the effect on the tracing is different. Every read from
3407 trace_pipe is consumed. This means that subse 3123 trace_pipe is consumed. This means that subsequent reads will be
3408 different. The trace is live. 3124 different. The trace is live.
3409 :: 3125 ::
3410 3126
3411 # echo function > current_tracer 3127 # echo function > current_tracer
3412 # cat trace_pipe > /tmp/trace.out & 3128 # cat trace_pipe > /tmp/trace.out &
3413 [1] 4153 3129 [1] 4153
3414 # echo 1 > tracing_on 3130 # echo 1 > tracing_on
3415 # usleep 1 3131 # usleep 1
3416 # echo 0 > tracing_on 3132 # echo 0 > tracing_on
3417 # cat trace 3133 # cat trace
3418 # tracer: function 3134 # tracer: function
3419 # 3135 #
3420 # entries-in-buffer/entries-written: 0/0 3136 # entries-in-buffer/entries-written: 0/0 #P:4
3421 # 3137 #
3422 # _-----=> irq 3138 # _-----=> irqs-off
3423 # / _----=> nee 3139 # / _----=> need-resched
3424 # | / _---=> har 3140 # | / _---=> hardirq/softirq
3425 # || / _--=> pre 3141 # || / _--=> preempt-depth
3426 # ||| / dela 3142 # ||| / delay
3427 # TASK-PID CPU# |||| TIMEST 3143 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3428 # | | | |||| | 3144 # | | | |||| | |
3429 3145
3430 # 3146 #
3431 # cat /tmp/trace.out 3147 # cat /tmp/trace.out
3432 bash-1994 [000] .... 5281.5689 3148 bash-1994 [000] .... 5281.568961: mutex_unlock <-rb_simple_write
3433 bash-1994 [000] .... 5281.5689 3149 bash-1994 [000] .... 5281.568963: __mutex_unlock_slowpath <-mutex_unlock
3434 bash-1994 [000] .... 5281.5689 3150 bash-1994 [000] .... 5281.568963: __fsnotify_parent <-fsnotify_modify
3435 bash-1994 [000] .... 5281.5689 3151 bash-1994 [000] .... 5281.568964: fsnotify <-fsnotify_modify
3436 bash-1994 [000] .... 5281.5689 3152 bash-1994 [000] .... 5281.568964: __srcu_read_lock <-fsnotify
3437 bash-1994 [000] .... 5281.5689 3153 bash-1994 [000] .... 5281.568964: add_preempt_count <-__srcu_read_lock
3438 bash-1994 [000] ...1 5281.5689 3154 bash-1994 [000] ...1 5281.568965: sub_preempt_count <-__srcu_read_lock
3439 bash-1994 [000] .... 5281.5689 3155 bash-1994 [000] .... 5281.568965: __srcu_read_unlock <-fsnotify
3440 bash-1994 [000] .... 5281.5689 3156 bash-1994 [000] .... 5281.568967: sys_dup2 <-system_call_fastpath
3441 3157
3442 3158
3443 Note, reading the trace_pipe file will block 3159 Note, reading the trace_pipe file will block until more input is
3444 added. This is contrary to the trace file. If 3160 added. This is contrary to the trace file. If any process opened
3445 the trace file for reading, it will actually 3161 the trace file for reading, it will actually disable tracing and
3446 prevent new entries from being added. The tra 3162 prevent new entries from being added. The trace_pipe file does
3447 not have this limitation. 3163 not have this limitation.
3448 3164
3449 trace entries 3165 trace entries
3450 ------------- 3166 -------------
3451 3167
3452 Having too much or not enough data can be tro 3168 Having too much or not enough data can be troublesome in
3453 diagnosing an issue in the kernel. The file b 3169 diagnosing an issue in the kernel. The file buffer_size_kb is
3454 used to modify the size of the internal trace 3170 used to modify the size of the internal trace buffers. The
3455 number listed is the number of entries that c 3171 number listed is the number of entries that can be recorded per
3456 CPU. To know the full size, multiply the numb 3172 CPU. To know the full size, multiply the number of possible CPUs
3457 with the number of entries. 3173 with the number of entries.
3458 :: 3174 ::
3459 3175
3460 # cat buffer_size_kb 3176 # cat buffer_size_kb
3461 1408 (units kilobytes) 3177 1408 (units kilobytes)
3462 3178
3463 Or simply read buffer_total_size_kb 3179 Or simply read buffer_total_size_kb
3464 :: 3180 ::
3465 3181
3466 # cat buffer_total_size_kb 3182 # cat buffer_total_size_kb
3467 5632 3183 5632
3468 3184
3469 To modify the buffer, simple echo in a number 3185 To modify the buffer, simple echo in a number (in 1024 byte segments).
3470 :: 3186 ::
3471 3187
3472 # echo 10000 > buffer_size_kb 3188 # echo 10000 > buffer_size_kb
3473 # cat buffer_size_kb 3189 # cat buffer_size_kb
3474 10000 (units kilobytes) 3190 10000 (units kilobytes)
3475 3191
3476 It will try to allocate as much as possible. 3192 It will try to allocate as much as possible. If you allocate too
3477 much, it can cause Out-Of-Memory to trigger. 3193 much, it can cause Out-Of-Memory to trigger.
3478 :: 3194 ::
3479 3195
3480 # echo 1000000000000 > buffer_size_kb 3196 # echo 1000000000000 > buffer_size_kb
3481 -bash: echo: write error: Cannot allocate m 3197 -bash: echo: write error: Cannot allocate memory
3482 # cat buffer_size_kb 3198 # cat buffer_size_kb
3483 85 3199 85
3484 3200
3485 The per_cpu buffers can be changed individual 3201 The per_cpu buffers can be changed individually as well:
3486 :: 3202 ::
3487 3203
3488 # echo 10000 > per_cpu/cpu0/buffer_size_kb 3204 # echo 10000 > per_cpu/cpu0/buffer_size_kb
3489 # echo 100 > per_cpu/cpu1/buffer_size_kb 3205 # echo 100 > per_cpu/cpu1/buffer_size_kb
3490 3206
3491 When the per_cpu buffers are not the same, th 3207 When the per_cpu buffers are not the same, the buffer_size_kb
3492 at the top level will just show an X 3208 at the top level will just show an X
3493 :: 3209 ::
3494 3210
3495 # cat buffer_size_kb 3211 # cat buffer_size_kb
3496 X 3212 X
3497 3213
3498 This is where the buffer_total_size_kb is use 3214 This is where the buffer_total_size_kb is useful:
3499 :: 3215 ::
3500 3216
3501 # cat buffer_total_size_kb 3217 # cat buffer_total_size_kb
3502 12916 3218 12916
3503 3219
3504 Writing to the top level buffer_size_kb will 3220 Writing to the top level buffer_size_kb will reset all the buffers
3505 to be the same again. 3221 to be the same again.
3506 3222
3507 Snapshot 3223 Snapshot
3508 -------- 3224 --------
3509 CONFIG_TRACER_SNAPSHOT makes a generic snapsh 3225 CONFIG_TRACER_SNAPSHOT makes a generic snapshot feature
3510 available to all non latency tracers. (Latenc 3226 available to all non latency tracers. (Latency tracers which
3511 record max latency, such as "irqsoff" or "wak 3227 record max latency, such as "irqsoff" or "wakeup", can't use
3512 this feature, since those are already using t 3228 this feature, since those are already using the snapshot
3513 mechanism internally.) 3229 mechanism internally.)
3514 3230
3515 Snapshot preserves a current trace buffer at 3231 Snapshot preserves a current trace buffer at a particular point
3516 in time without stopping tracing. Ftrace swap 3232 in time without stopping tracing. Ftrace swaps the current
3517 buffer with a spare buffer, and tracing conti 3233 buffer with a spare buffer, and tracing continues in the new
3518 current (=previous spare) buffer. 3234 current (=previous spare) buffer.
3519 3235
3520 The following tracefs files in "tracing" are 3236 The following tracefs files in "tracing" are related to this
3521 feature: 3237 feature:
3522 3238
3523 snapshot: 3239 snapshot:
3524 3240
3525 This is used to take a snapshot and t 3241 This is used to take a snapshot and to read the output
3526 of the snapshot. Echo 1 into this fil 3242 of the snapshot. Echo 1 into this file to allocate a
3527 spare buffer and to take a snapshot ( 3243 spare buffer and to take a snapshot (swap), then read
3528 the snapshot from this file in the sa 3244 the snapshot from this file in the same format as
3529 "trace" (described above in the secti 3245 "trace" (described above in the section "The File
3530 System"). Both reads snapshot and tra 3246 System"). Both reads snapshot and tracing are executable
3531 in parallel. When the spare buffer is 3247 in parallel. When the spare buffer is allocated, echoing
3532 0 frees it, and echoing else (positiv 3248 0 frees it, and echoing else (positive) values clear the
3533 snapshot contents. 3249 snapshot contents.
3534 More details are shown in the table b 3250 More details are shown in the table below.
3535 3251
3536 +--------------+------------+-------- 3252 +--------------+------------+------------+------------+
3537 |status\\input | 0 | 1 3253 |status\\input | 0 | 1 | else |
3538 +==============+============+======== 3254 +==============+============+============+============+
3539 |not allocated |(do nothing)| alloc+s 3255 |not allocated |(do nothing)| alloc+swap |(do nothing)|
3540 +--------------+------------+-------- 3256 +--------------+------------+------------+------------+
3541 |allocated | free | swap 3257 |allocated | free | swap | clear |
3542 +--------------+------------+-------- 3258 +--------------+------------+------------+------------+
3543 3259
3544 Here is an example of using the snapshot feat 3260 Here is an example of using the snapshot feature.
3545 :: 3261 ::
3546 3262
3547 # echo 1 > events/sched/enable 3263 # echo 1 > events/sched/enable
3548 # echo 1 > snapshot 3264 # echo 1 > snapshot
3549 # cat snapshot 3265 # cat snapshot
3550 # tracer: nop 3266 # tracer: nop
3551 # 3267 #
3552 # entries-in-buffer/entries-written: 71/71 3268 # entries-in-buffer/entries-written: 71/71 #P:8
3553 # 3269 #
3554 # _-----=> irq 3270 # _-----=> irqs-off
3555 # / _----=> nee 3271 # / _----=> need-resched
3556 # | / _---=> har 3272 # | / _---=> hardirq/softirq
3557 # || / _--=> pre 3273 # || / _--=> preempt-depth
3558 # ||| / dela 3274 # ||| / delay
3559 # TASK-PID CPU# |||| TIMEST 3275 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3560 # | | | |||| | 3276 # | | | |||| | |
3561 <idle>-0 [005] d... 2440.603 3277 <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 3278 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 [...] 3279 [...]
3564 <idle>-0 [002] d... 2440.70723 3280 <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 3281
3566 # cat trace 3282 # cat trace
3567 # tracer: nop 3283 # tracer: nop
3568 # 3284 #
3569 # entries-in-buffer/entries-written: 77/77 3285 # entries-in-buffer/entries-written: 77/77 #P:8
3570 # 3286 #
3571 # _-----=> irq 3287 # _-----=> irqs-off
3572 # / _----=> nee 3288 # / _----=> need-resched
3573 # | / _---=> har 3289 # | / _---=> hardirq/softirq
3574 # || / _--=> pre 3290 # || / _--=> preempt-depth
3575 # ||| / dela 3291 # ||| / delay
3576 # TASK-PID CPU# |||| TIMEST 3292 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3577 # | | | |||| | 3293 # | | | |||| | |
3578 <idle>-0 [007] d... 2440.707 3294 <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 3295 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 [...] 3296 [...]
3581 3297
3582 3298
3583 If you try to use this snapshot feature when 3299 If you try to use this snapshot feature when current tracer is
3584 one of the latency tracers, you will get the 3300 one of the latency tracers, you will get the following results.
3585 :: 3301 ::
3586 3302
3587 # echo wakeup > current_tracer 3303 # echo wakeup > current_tracer
3588 # echo 1 > snapshot 3304 # echo 1 > snapshot
3589 bash: echo: write error: Device or resource 3305 bash: echo: write error: Device or resource busy
3590 # cat snapshot 3306 # cat snapshot
3591 cat: snapshot: Device or resource busy 3307 cat: snapshot: Device or resource busy
3592 3308
3593 3309
3594 Instances 3310 Instances
3595 --------- 3311 ---------
3596 In the tracefs tracing directory, there is a !! 3312 In the tracefs tracing directory is a directory called "instances".
3597 This directory can have new directories creat 3313 This directory can have new directories created inside of it using
3598 mkdir, and removing directories with rmdir. T 3314 mkdir, and removing directories with rmdir. The directory created
3599 with mkdir in this directory will already con 3315 with mkdir in this directory will already contain files and other
3600 directories after it is created. 3316 directories after it is created.
3601 :: 3317 ::
3602 3318
3603 # mkdir instances/foo 3319 # mkdir instances/foo
3604 # ls instances/foo 3320 # ls instances/foo
3605 buffer_size_kb buffer_total_size_kb event 3321 buffer_size_kb buffer_total_size_kb events free_buffer per_cpu
3606 set_event snapshot trace trace_clock tr 3322 set_event snapshot trace trace_clock trace_marker trace_options
3607 trace_pipe tracing_on 3323 trace_pipe tracing_on
3608 3324
3609 As you can see, the new directory looks simil 3325 As you can see, the new directory looks similar to the tracing directory
3610 itself. In fact, it is very similar, except t 3326 itself. In fact, it is very similar, except that the buffer and
3611 events are agnostic from the main directory, !! 3327 events are agnostic from the main director, or from any other
3612 instances that are created. 3328 instances that are created.
3613 3329
3614 The files in the new directory work just like 3330 The files in the new directory work just like the files with the
3615 same name in the tracing directory except the 3331 same name in the tracing directory except the buffer that is used
3616 is a separate and new buffer. The files affec 3332 is a separate and new buffer. The files affect that buffer but do not
3617 affect the main buffer with the exception of 3333 affect the main buffer with the exception of trace_options. Currently,
3618 the trace_options affect all instances and th 3334 the trace_options affect all instances and the top level buffer
3619 the same, but this may change in future relea 3335 the same, but this may change in future releases. That is, options
3620 may become specific to the instance they resi 3336 may become specific to the instance they reside in.
3621 3337
3622 Notice that none of the function tracer files 3338 Notice that none of the function tracer files are there, nor is
3623 current_tracer and available_tracers. This is 3339 current_tracer and available_tracers. This is because the buffers
3624 can currently only have events enabled for th 3340 can currently only have events enabled for them.
3625 :: 3341 ::
3626 3342
3627 # mkdir instances/foo 3343 # mkdir instances/foo
3628 # mkdir instances/bar 3344 # mkdir instances/bar
3629 # mkdir instances/zoot 3345 # mkdir instances/zoot
3630 # echo 100000 > buffer_size_kb 3346 # echo 100000 > buffer_size_kb
3631 # echo 1000 > instances/foo/buffer_size_kb 3347 # echo 1000 > instances/foo/buffer_size_kb
3632 # echo 5000 > instances/bar/per_cpu/cpu1/bu 3348 # echo 5000 > instances/bar/per_cpu/cpu1/buffer_size_kb
3633 # echo function > current_trace 3349 # echo function > current_trace
3634 # echo 1 > instances/foo/events/sched/sched 3350 # echo 1 > instances/foo/events/sched/sched_wakeup/enable
3635 # echo 1 > instances/foo/events/sched/sched 3351 # echo 1 > instances/foo/events/sched/sched_wakeup_new/enable
3636 # echo 1 > instances/foo/events/sched/sched 3352 # echo 1 > instances/foo/events/sched/sched_switch/enable
3637 # echo 1 > instances/bar/events/irq/enable 3353 # echo 1 > instances/bar/events/irq/enable
3638 # echo 1 > instances/zoot/events/syscalls/e 3354 # echo 1 > instances/zoot/events/syscalls/enable
3639 # cat trace_pipe 3355 # cat trace_pipe
3640 CPU:2 [LOST 11745 EVENTS] 3356 CPU:2 [LOST 11745 EVENTS]
3641 bash-2044 [002] .... 10594.481 3357 bash-2044 [002] .... 10594.481032: _raw_spin_lock_irqsave <-get_page_from_freelist
3642 bash-2044 [002] d... 10594.481 3358 bash-2044 [002] d... 10594.481032: add_preempt_count <-_raw_spin_lock_irqsave
3643 bash-2044 [002] d..1 10594.481 3359 bash-2044 [002] d..1 10594.481032: __rmqueue <-get_page_from_freelist
3644 bash-2044 [002] d..1 10594.481 3360 bash-2044 [002] d..1 10594.481033: _raw_spin_unlock <-get_page_from_freelist
3645 bash-2044 [002] d..1 10594.481 3361 bash-2044 [002] d..1 10594.481033: sub_preempt_count <-_raw_spin_unlock
3646 bash-2044 [002] d... 10594.481 3362 bash-2044 [002] d... 10594.481033: get_pageblock_flags_group <-get_pageblock_migratetype
3647 bash-2044 [002] d... 10594.481 3363 bash-2044 [002] d... 10594.481034: __mod_zone_page_state <-get_page_from_freelist
3648 bash-2044 [002] d... 10594.481 3364 bash-2044 [002] d... 10594.481034: zone_statistics <-get_page_from_freelist
3649 bash-2044 [002] d... 10594.481 3365 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
3650 bash-2044 [002] d... 10594.481 3366 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
3651 bash-2044 [002] .... 10594.481 3367 bash-2044 [002] .... 10594.481035: arch_dup_task_struct <-copy_process
3652 [...] 3368 [...]
3653 3369
3654 # cat instances/foo/trace_pipe 3370 # cat instances/foo/trace_pipe
3655 bash-1998 [000] d..4 136.676 3371 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 3372 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 3373 <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 3374 <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 3375 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 3376 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 3377 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 3378 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 3379 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 3380 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 [...] 3381 [...]
3666 3382
3667 # cat instances/bar/trace_pipe 3383 # cat instances/bar/trace_pipe
3668 migration/1-14 [001] d.h3 138.732 3384 migration/1-14 [001] d.h3 138.732674: softirq_raise: vec=3 [action=NET_RX]
3669 <idle>-0 [001] dNh3 138.732 3385 <idle>-0 [001] dNh3 138.732725: softirq_raise: vec=3 [action=NET_RX]
3670 bash-1998 [000] d.h1 138.733 3386 bash-1998 [000] d.h1 138.733101: softirq_raise: vec=1 [action=TIMER]
3671 bash-1998 [000] d.h1 138.733 3387 bash-1998 [000] d.h1 138.733102: softirq_raise: vec=9 [action=RCU]
3672 bash-1998 [000] ..s2 138.733 3388 bash-1998 [000] ..s2 138.733105: softirq_entry: vec=1 [action=TIMER]
3673 bash-1998 [000] ..s2 138.733 3389 bash-1998 [000] ..s2 138.733106: softirq_exit: vec=1 [action=TIMER]
3674 bash-1998 [000] ..s2 138.733 3390 bash-1998 [000] ..s2 138.733106: softirq_entry: vec=9 [action=RCU]
3675 bash-1998 [000] ..s2 138.733 3391 bash-1998 [000] ..s2 138.733109: softirq_exit: vec=9 [action=RCU]
3676 sshd-1995 [001] d.h1 138.733 3392 sshd-1995 [001] d.h1 138.733278: irq_handler_entry: irq=21 name=uhci_hcd:usb4
3677 sshd-1995 [001] d.h1 138.733 3393 sshd-1995 [001] d.h1 138.733280: irq_handler_exit: irq=21 ret=unhandled
3678 sshd-1995 [001] d.h1 138.733 3394 sshd-1995 [001] d.h1 138.733281: irq_handler_entry: irq=21 name=eth0
3679 sshd-1995 [001] d.h1 138.733 3395 sshd-1995 [001] d.h1 138.733283: irq_handler_exit: irq=21 ret=handled
3680 [...] 3396 [...]
3681 3397
3682 # cat instances/zoot/trace 3398 # cat instances/zoot/trace
3683 # tracer: nop 3399 # tracer: nop
3684 # 3400 #
3685 # entries-in-buffer/entries-written: 18996/ 3401 # entries-in-buffer/entries-written: 18996/18996 #P:4
3686 # 3402 #
3687 # _-----=> irq 3403 # _-----=> irqs-off
3688 # / _----=> nee 3404 # / _----=> need-resched
3689 # | / _---=> har 3405 # | / _---=> hardirq/softirq
3690 # || / _--=> pre 3406 # || / _--=> preempt-depth
3691 # ||| / dela 3407 # ||| / delay
3692 # TASK-PID CPU# |||| TIMEST 3408 # TASK-PID CPU# |||| TIMESTAMP FUNCTION
3693 # | | | |||| | 3409 # | | | |||| | |
3694 bash-1998 [000] d... 140.733 3410 bash-1998 [000] d... 140.733501: sys_write -> 0x2
3695 bash-1998 [000] d... 140.733 3411 bash-1998 [000] d... 140.733504: sys_dup2(oldfd: a, newfd: 1)
3696 bash-1998 [000] d... 140.733 3412 bash-1998 [000] d... 140.733506: sys_dup2 -> 0x1
3697 bash-1998 [000] d... 140.733 3413 bash-1998 [000] d... 140.733508: sys_fcntl(fd: a, cmd: 1, arg: 0)
3698 bash-1998 [000] d... 140.733 3414 bash-1998 [000] d... 140.733509: sys_fcntl -> 0x1
3699 bash-1998 [000] d... 140.733 3415 bash-1998 [000] d... 140.733510: sys_close(fd: a)
3700 bash-1998 [000] d... 140.733 3416 bash-1998 [000] d... 140.733510: sys_close -> 0x0
3701 bash-1998 [000] d... 140.733 3417 bash-1998 [000] d... 140.733514: sys_rt_sigprocmask(how: 0, nset: 0, oset: 6e2768, sigsetsize: 8)
3702 bash-1998 [000] d... 140.733 3418 bash-1998 [000] d... 140.733515: sys_rt_sigprocmask -> 0x0
3703 bash-1998 [000] d... 140.733 3419 bash-1998 [000] d... 140.733516: sys_rt_sigaction(sig: 2, act: 7fff718846f0, oact: 7fff71884650, sigsetsize: 8)
3704 bash-1998 [000] d... 140.733 3420 bash-1998 [000] d... 140.733516: sys_rt_sigaction -> 0x0
3705 3421
3706 You can see that the trace of the top most tr 3422 You can see that the trace of the top most trace buffer shows only
3707 the function tracing. The foo instance displa 3423 the function tracing. The foo instance displays wakeups and task
3708 switches. 3424 switches.
3709 3425
3710 To remove the instances, simply delete their 3426 To remove the instances, simply delete their directories:
3711 :: 3427 ::
3712 3428
3713 # rmdir instances/foo 3429 # rmdir instances/foo
3714 # rmdir instances/bar 3430 # rmdir instances/bar
3715 # rmdir instances/zoot 3431 # rmdir instances/zoot
3716 3432
3717 Note, if a process has a trace file open in o 3433 Note, if a process has a trace file open in one of the instance
3718 directories, the rmdir will fail with EBUSY. 3434 directories, the rmdir will fail with EBUSY.
3719 3435
3720 3436
3721 Stack trace 3437 Stack trace
3722 ----------- 3438 -----------
3723 Since the kernel has a fixed sized stack, it 3439 Since the kernel has a fixed sized stack, it is important not to
3724 waste it in functions. A kernel developer mus !! 3440 waste it in functions. A kernel developer must be conscience of
3725 what they allocate on the stack. If they add 3441 what they allocate on the stack. If they add too much, the system
3726 can be in danger of a stack overflow, and cor 3442 can be in danger of a stack overflow, and corruption will occur,
3727 usually leading to a system panic. 3443 usually leading to a system panic.
3728 3444
3729 There are some tools that check this, usually 3445 There are some tools that check this, usually with interrupts
3730 periodically checking usage. But if you can p 3446 periodically checking usage. But if you can perform a check
3731 at every function call that will become very 3447 at every function call that will become very useful. As ftrace provides
3732 a function tracer, it makes it convenient to 3448 a function tracer, it makes it convenient to check the stack size
3733 at every function call. This is enabled via t 3449 at every function call. This is enabled via the stack tracer.
3734 3450
3735 CONFIG_STACK_TRACER enables the ftrace stack 3451 CONFIG_STACK_TRACER enables the ftrace stack tracing functionality.
3736 To enable it, write a '1' into /proc/sys/kern 3452 To enable it, write a '1' into /proc/sys/kernel/stack_tracer_enabled.
3737 :: 3453 ::
3738 3454
3739 # echo 1 > /proc/sys/kernel/stack_tracer_ena 3455 # echo 1 > /proc/sys/kernel/stack_tracer_enabled
3740 3456
3741 You can also enable it from the kernel comman 3457 You can also enable it from the kernel command line to trace
3742 the stack size of the kernel during boot up, 3458 the stack size of the kernel during boot up, by adding "stacktrace"
3743 to the kernel command line parameter. 3459 to the kernel command line parameter.
3744 3460
3745 After running it for a few minutes, the outpu 3461 After running it for a few minutes, the output looks like:
3746 :: 3462 ::
3747 3463
3748 # cat stack_max_size 3464 # cat stack_max_size
3749 2928 3465 2928
3750 3466
3751 # cat stack_trace 3467 # cat stack_trace
3752 Depth Size Location (18 ent 3468 Depth Size Location (18 entries)
3753 ----- ---- -------- 3469 ----- ---- --------
3754 0) 2928 224 update_sd_lb_stats+ 3470 0) 2928 224 update_sd_lb_stats+0xbc/0x4ac
3755 1) 2704 160 find_busiest_group+ 3471 1) 2704 160 find_busiest_group+0x31/0x1f1
3756 2) 2544 256 load_balance+0xd9/0 3472 2) 2544 256 load_balance+0xd9/0x662
3757 3) 2288 80 idle_balance+0xbb/0 3473 3) 2288 80 idle_balance+0xbb/0x130
3758 4) 2208 128 __schedule+0x26e/0x 3474 4) 2208 128 __schedule+0x26e/0x5b9
3759 5) 2080 16 schedule+0x64/0x66 3475 5) 2080 16 schedule+0x64/0x66
3760 6) 2064 128 schedule_timeout+0x 3476 6) 2064 128 schedule_timeout+0x34/0xe0
3761 7) 1936 112 wait_for_common+0x9 3477 7) 1936 112 wait_for_common+0x97/0xf1
3762 8) 1824 16 wait_for_completion 3478 8) 1824 16 wait_for_completion+0x1d/0x1f
3763 9) 1808 128 flush_work+0xfe/0x1 3479 9) 1808 128 flush_work+0xfe/0x119
3764 10) 1680 16 tty_flush_to_ldisc+ 3480 10) 1680 16 tty_flush_to_ldisc+0x1e/0x20
3765 11) 1664 48 input_available_p+0 3481 11) 1664 48 input_available_p+0x1d/0x5c
3766 12) 1616 48 n_tty_poll+0x6d/0x1 3482 12) 1616 48 n_tty_poll+0x6d/0x134
3767 13) 1568 64 tty_poll+0x64/0x7f 3483 13) 1568 64 tty_poll+0x64/0x7f
3768 14) 1504 880 do_select+0x31e/0x5 3484 14) 1504 880 do_select+0x31e/0x511
3769 15) 624 400 core_sys_select+0x1 3485 15) 624 400 core_sys_select+0x177/0x216
3770 16) 224 96 sys_select+0x91/0xb 3486 16) 224 96 sys_select+0x91/0xb9
3771 17) 128 128 system_call_fastpat 3487 17) 128 128 system_call_fastpath+0x16/0x1b
3772 3488
3773 Note, if -mfentry is being used by gcc, funct 3489 Note, if -mfentry is being used by gcc, functions get traced before
3774 they set up the stack frame. This means that 3490 they set up the stack frame. This means that leaf level functions
3775 are not tested by the stack tracer when -mfen 3491 are not tested by the stack tracer when -mfentry is used.
3776 3492
3777 Currently, -mfentry is used by gcc 4.6.0 and 3493 Currently, -mfentry is used by gcc 4.6.0 and above on x86 only.
3778 3494
3779 More 3495 More
3780 ---- 3496 ----
3781 More details can be found in the source code, 3497 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.