blob: ef621d34ba5bf7a5a3c2d570de2ebaecf0615cc0 [file] [log] [blame]
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001 ftrace - Function Tracer
2 ========================
3
4Copyright 2008 Red Hat Inc.
Steven Rostedta41eeba2008-07-14 16:41:12 -04005 Author: Steven Rostedt <srostedt@redhat.com>
6 License: The GNU Free Documentation License, Version 1.2
Steven Rostedta97762a2008-07-31 12:40:52 -04007 (dual licensed under the GPL v2)
Steven Rostedtf2d9c742008-07-15 10:57:33 -04008Reviewers: Elias Oltmanns, Randy Dunlap, Andrew Morton,
9 John Kacur, and David Teigland.
Steven Rostedt42ec6322008-11-03 15:18:56 -050010Written for: 2.6.28-rc2
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -040011Updated for: 3.10
Steven Rostedteb6d42e2008-07-10 12:46:01 -040012
13Introduction
14------------
15
16Ftrace is an internal tracer designed to help out developers and
17designers of systems to find what is going on inside the kernel.
Ingo Molnar57526742009-02-19 12:54:10 +010018It can be used for debugging or analyzing latencies and
19performance issues that take place outside of user-space.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040020
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -040021Although ftrace is typically considered the function tracer, it
22is really a frame work of several assorted tracing utilities.
23There's latency tracing to examine what occurs between interrupts
24disabled and enabled, as well as for preemption and from a time
25a task is woken to the task is actually scheduled in.
26
27One of the most common uses of ftrace is the event tracing.
28Through out the kernel is hundreds of static event points that
29can be enabled via the debugfs file system to see what is
30going on in certain parts of the kernel.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040031
32
Mike Frysinger555f3862009-09-14 20:10:15 -040033Implementation Details
34----------------------
35
36See ftrace-design.txt for details for arch porters and such.
37
38
Steven Rostedteb6d42e2008-07-10 12:46:01 -040039The File System
40---------------
41
Ingo Molnar57526742009-02-19 12:54:10 +010042Ftrace uses the debugfs file system to hold the control files as
43well as the files to display output.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040044
GeunSik Lim156f5a72009-06-02 15:01:37 +090045When debugfs is configured into the kernel (which selecting any ftrace
46option will do) the directory /sys/kernel/debug will be created. To mount
47this directory, you can add to your /etc/fstab file:
Steven Rostedteb6d42e2008-07-10 12:46:01 -040048
GeunSik Lim156f5a72009-06-02 15:01:37 +090049 debugfs /sys/kernel/debug debugfs defaults 0 0
Steven Rostedteb6d42e2008-07-10 12:46:01 -040050
GeunSik Lim156f5a72009-06-02 15:01:37 +090051Or you can mount it at run time with:
52
53 mount -t debugfs nodev /sys/kernel/debug
54
55For quicker access to that directory you may want to make a soft link to
56it:
57
58 ln -s /sys/kernel/debug /debug
59
60Any selected ftrace option will also create a directory called tracing
61within the debugfs. The rest of the document will assume that you are in
62the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate
63on the files within that directory and not distract from the content with
64the extended "/sys/kernel/debug/tracing" path name.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040065
66That's it! (assuming that you have ftrace configured into your kernel)
67
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -040068After mounting debugfs, you can see a directory called
Steven Rostedteb6d42e2008-07-10 12:46:01 -040069"tracing". This directory contains the control and output files
70of ftrace. Here is a list of some of the key files:
71
72
73 Note: all time values are in microseconds.
74
Ingo Molnar57526742009-02-19 12:54:10 +010075 current_tracer:
Steven Rostedteb6d42e2008-07-10 12:46:01 -040076
Ingo Molnar57526742009-02-19 12:54:10 +010077 This is used to set or display the current tracer
78 that is configured.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040079
Ingo Molnar57526742009-02-19 12:54:10 +010080 available_tracers:
Steven Rostedteb6d42e2008-07-10 12:46:01 -040081
Ingo Molnar57526742009-02-19 12:54:10 +010082 This holds the different types of tracers that
83 have been compiled into the kernel. The
84 tracers listed here can be configured by
85 echoing their name into current_tracer.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040086
Steven Rostedt6752ab42011-02-08 13:54:06 -050087 tracing_on:
Steven Rostedteb6d42e2008-07-10 12:46:01 -040088
Steven Rostedt6752ab42011-02-08 13:54:06 -050089 This sets or displays whether writing to the trace
90 ring buffer is enabled. Echo 0 into this file to disable
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -040091 the tracer or 1 to enable it. Note, this only disables
92 writing to the ring buffer, the tracing overhead may
93 still be occurring.
Steven Rostedteb6d42e2008-07-10 12:46:01 -040094
Ingo Molnar57526742009-02-19 12:54:10 +010095 trace:
Steven Rostedteb6d42e2008-07-10 12:46:01 -040096
Ingo Molnar57526742009-02-19 12:54:10 +010097 This file holds the output of the trace in a human
98 readable format (described below).
Steven Rostedteb6d42e2008-07-10 12:46:01 -040099
Ingo Molnar57526742009-02-19 12:54:10 +0100100 trace_pipe:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400101
Ingo Molnar57526742009-02-19 12:54:10 +0100102 The output is the same as the "trace" file but this
103 file is meant to be streamed with live tracing.
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200104 Reads from this file will block until new data is
105 retrieved. Unlike the "trace" file, this file is a
106 consumer. This means reading from this file causes
107 sequential reads to display more current data. Once
108 data is read from this file, it is consumed, and
109 will not be read again with a sequential read. The
110 "trace" file is static, and if the tracer is not
Rabin Vincent8547aa12015-05-04 19:48:54 +0200111 adding more data, it will display the same
112 information every time it is read.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400113
Ingo Molnar57526742009-02-19 12:54:10 +0100114 trace_options:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400115
Ingo Molnar57526742009-02-19 12:54:10 +0100116 This file lets the user control the amount of data
117 that is displayed in one of the above output
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400118 files. Options also exist to modify how a tracer
119 or events work (stack traces, timestamps, etc).
120
121 options:
122
123 This is a directory that has a file for every available
124 trace option (also in trace_options). Options may also be set
125 or cleared by writing a "1" or "0" respectively into the
126 corresponding file with the option name.
Steven Rostedtdf4fc312008-11-26 00:16:23 -0500127
KOSAKI Motohiro42b40b32009-03-07 23:55:09 +0900128 tracing_max_latency:
Frederic Weisbecker985ec202009-02-18 06:35:34 +0100129
Ingo Molnar57526742009-02-19 12:54:10 +0100130 Some of the tracers record the max latency.
131 For example, the time interrupts are disabled.
132 This time is saved in this file. The max trace
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200133 will also be stored, and displayed by "trace".
134 A new max trace will only be recorded if the
135 latency is greater than the value in this
136 file. (in microseconds)
Ingo Molnar57526742009-02-19 12:54:10 +0100137
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400138 tracing_thresh:
139
140 Some latency tracers will record a trace whenever the
141 latency is greater than the number in this file.
142 Only active when the file contains a number greater than 0.
143 (in microseconds)
144
Ingo Molnar57526742009-02-19 12:54:10 +0100145 buffer_size_kb:
146
147 This sets or displays the number of kilobytes each CPU
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400148 buffer holds. By default, the trace buffers are the same size
Ingo Molnar57526742009-02-19 12:54:10 +0100149 for each CPU. The displayed number is the size of the
150 CPU buffer and not total size of all buffers. The
151 trace buffers are allocated in pages (blocks of memory
152 that the kernel uses for allocation, usually 4 KB in size).
153 If the last page allocated has room for more bytes
154 than requested, the rest of the page will be used,
155 making the actual allocation bigger than requested.
156 ( Note, the size may not be a multiple of the page size
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400157 due to buffer management meta-data. )
Ingo Molnar57526742009-02-19 12:54:10 +0100158
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400159 buffer_total_size_kb:
160
161 This displays the total combined size of all the trace buffers.
162
163 free_buffer:
164
165 If a process is performing the tracing, and the ring buffer
166 should be shrunk "freed" when the process is finished, even
167 if it were to be killed by a signal, this file can be used
168 for that purpose. On close of this file, the ring buffer will
169 be resized to its minimum size. Having a process that is tracing
170 also open this file, when the process exits its file descriptor
171 for this file will be closed, and in doing so, the ring buffer
172 will be "freed".
173
174 It may also stop tracing if disable_on_free option is set.
Ingo Molnar57526742009-02-19 12:54:10 +0100175
176 tracing_cpumask:
177
178 This is a mask that lets the user only trace
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400179 on specified CPUs. The format is a hex string
180 representing the CPUs.
Ingo Molnar57526742009-02-19 12:54:10 +0100181
182 set_ftrace_filter:
183
184 When dynamic ftrace is configured in (see the
185 section below "dynamic ftrace"), the code is dynamically
186 modified (code text rewrite) to disable calling of the
187 function profiler (mcount). This lets tracing be configured
188 in with practically no overhead in performance. This also
189 has a side effect of enabling or disabling specific functions
190 to be traced. Echoing names of functions into this file
191 will limit the trace to only those functions.
192
Chase Douglas07271aa2010-04-23 14:02:39 -0400193 This interface also allows for commands to be used. See the
194 "Filter commands" section for more details.
195
Ingo Molnar57526742009-02-19 12:54:10 +0100196 set_ftrace_notrace:
197
198 This has an effect opposite to that of
199 set_ftrace_filter. Any function that is added here will not
200 be traced. If a function exists in both set_ftrace_filter
201 and set_ftrace_notrace, the function will _not_ be traced.
202
203 set_ftrace_pid:
204
205 Have the function tracer only trace a single thread.
206
207 set_graph_function:
208
209 Set a "trigger" function where tracing should start
210 with the function graph tracer (See the section
211 "dynamic ftrace" for more details).
212
213 available_filter_functions:
214
215 This lists the functions that ftrace
216 has processed and can trace. These are the function
217 names that you can pass to "set_ftrace_filter" or
218 "set_ftrace_notrace". (See the section "dynamic ftrace"
219 below for more details.)
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400220
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400221 enabled_functions:
222
223 This file is more for debugging ftrace, but can also be useful
224 in seeing if any function has a callback attached to it.
225 Not only does the trace infrastructure use ftrace function
226 trace utility, but other subsystems might too. This file
227 displays all functions that have a callback attached to them
228 as well as the number of callbacks that have been attached.
229 Note, a callback may also call multiple functions which will
230 not be listed in this count.
231
232 If the callback registered to be traced by a function with
233 the "save regs" attribute (thus even more overhead), a 'R'
234 will be displayed on the same line as the function that
235 is returning registers.
236
Masami Hiramatsuf8b8be82014-11-21 05:25:16 -0500237 If the callback registered to be traced by a function with
238 the "ip modify" attribute (thus the regs->ip can be changed),
239 an 'I' will be displayed on the same line as the function that
240 can be overridden.
241
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400242 function_profile_enabled:
243
244 When set it will enable all functions with either the function
245 tracer, or if enabled, the function graph tracer. It will
246 keep a histogram of the number of functions that were called
247 and if run with the function graph tracer, it will also keep
248 track of the time spent in those functions. The histogram
249 content can be displayed in the files:
250
251 trace_stats/function<cpu> ( function0, function1, etc).
252
253 trace_stats:
254
255 A directory that holds different tracing stats.
256
257 kprobe_events:
258
259 Enable dynamic trace points. See kprobetrace.txt.
260
261 kprobe_profile:
262
263 Dynamic trace points stats. See kprobetrace.txt.
264
265 max_graph_depth:
266
267 Used with the function graph tracer. This is the max depth
268 it will trace into a function. Setting this to a value of
269 one will show only the first kernel function that is called
270 from user space.
271
272 printk_formats:
273
274 This is for tools that read the raw format files. If an event in
275 the ring buffer references a string (currently only trace_printk()
276 does this), only a pointer to the string is recorded into the buffer
277 and not the string itself. This prevents tools from knowing what
278 that string was. This file displays the string and address for
279 the string allowing tools to map the pointers to what the
280 strings were.
281
282 saved_cmdlines:
283
284 Only the pid of the task is recorded in a trace event unless
285 the event specifically saves the task comm as well. Ftrace
286 makes a cache of pid mappings to comms to try to display
287 comms for events. If a pid for a comm is not listed, then
288 "<...>" is displayed in the output.
289
290 snapshot:
291
292 This displays the "snapshot" buffer and also lets the user
293 take a snapshot of the current running trace.
294 See the "Snapshot" section below for more details.
295
296 stack_max_size:
297
298 When the stack tracer is activated, this will display the
299 maximum stack size it has encountered.
300 See the "Stack Trace" section below.
301
302 stack_trace:
303
304 This displays the stack back trace of the largest stack
305 that was encountered when the stack tracer is activated.
306 See the "Stack Trace" section below.
307
308 stack_trace_filter:
309
310 This is similar to "set_ftrace_filter" but it limits what
311 functions the stack tracer will check.
312
313 trace_clock:
314
315 Whenever an event is recorded into the ring buffer, a
316 "timestamp" is added. This stamp comes from a specified
317 clock. By default, ftrace uses the "local" clock. This
318 clock is very fast and strictly per cpu, but on some
319 systems it may not be monotonic with respect to other
320 CPUs. In other words, the local clocks may not be in sync
321 with local clocks on other CPUs.
322
323 Usual clocks for tracing:
324
325 # cat trace_clock
326 [local] global counter x86-tsc
327
328 local: Default clock, but may not be in sync across CPUs
329
330 global: This clock is in sync with all CPUs but may
331 be a bit slower than the local clock.
332
333 counter: This is not a clock at all, but literally an atomic
334 counter. It counts up one by one, but is in sync
335 with all CPUs. This is useful when you need to
336 know exactly the order events occurred with respect to
337 each other on different CPUs.
338
339 uptime: This uses the jiffies counter and the time stamp
340 is relative to the time since boot up.
341
342 perf: This makes ftrace use the same clock that perf uses.
343 Eventually perf will be able to read ftrace buffers
344 and this will help out in interleaving the data.
345
346 x86-tsc: Architectures may define their own clocks. For
347 example, x86 uses its own TSC cycle clock here.
348
Naveen N. Rao197165d2015-04-24 14:24:44 +0530349 ppc-tb: This uses the powerpc timebase register value.
350 This is in sync across CPUs and can also be used
351 to correlate events across hypervisor/guest if
352 tb_offset is known.
353
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400354 To set a clock, simply echo the clock name into this file.
355
356 echo global > trace_clock
357
358 trace_marker:
359
360 This is a very useful file for synchronizing user space
361 with events happening in the kernel. Writing strings into
362 this file will be written into the ftrace buffer.
363
364 It is useful in applications to open this file at the start
365 of the application and just reference the file descriptor
366 for the file.
367
368 void trace_write(const char *fmt, ...)
369 {
370 va_list ap;
371 char buf[256];
372 int n;
373
374 if (trace_fd < 0)
375 return;
376
377 va_start(ap, fmt);
378 n = vsnprintf(buf, 256, fmt, ap);
379 va_end(ap);
380
381 write(trace_fd, buf, n);
382 }
383
384 start:
385
386 trace_fd = open("trace_marker", WR_ONLY);
387
388 uprobe_events:
389
390 Add dynamic tracepoints in programs.
391 See uprobetracer.txt
392
393 uprobe_profile:
394
395 Uprobe statistics. See uprobetrace.txt
396
397 instances:
398
399 This is a way to make multiple trace buffers where different
400 events can be recorded in different buffers.
401 See "Instances" section below.
402
403 events:
404
405 This is the trace event directory. It holds event tracepoints
406 (also known as static tracepoints) that have been compiled
407 into the kernel. It shows what event tracepoints exist
408 and how they are grouped by system. There are "enable"
409 files at various levels that can enable the tracepoints
410 when a "1" is written to them.
411
412 See events.txt for more information.
413
414 per_cpu:
415
416 This is a directory that contains the trace per_cpu information.
417
418 per_cpu/cpu0/buffer_size_kb:
419
420 The ftrace buffer is defined per_cpu. That is, there's a separate
421 buffer for each CPU to allow writes to be done atomically,
422 and free from cache bouncing. These buffers may have different
423 size buffers. This file is similar to the buffer_size_kb
424 file, but it only displays or sets the buffer size for the
425 specific CPU. (here cpu0).
426
427 per_cpu/cpu0/trace:
428
429 This is similar to the "trace" file, but it will only display
430 the data specific for the CPU. If written to, it only clears
431 the specific CPU buffer.
432
433 per_cpu/cpu0/trace_pipe
434
435 This is similar to the "trace_pipe" file, and is a consuming
436 read, but it will only display (and consume) the data specific
437 for the CPU.
438
439 per_cpu/cpu0/trace_pipe_raw
440
441 For tools that can parse the ftrace ring buffer binary format,
442 the trace_pipe_raw file can be used to extract the data
443 from the ring buffer directly. With the use of the splice()
444 system call, the buffer data can be quickly transferred to
445 a file or to the network where a server is collecting the
446 data.
447
448 Like trace_pipe, this is a consuming reader, where multiple
449 reads will always produce different data.
450
451 per_cpu/cpu0/snapshot:
452
453 This is similar to the main "snapshot" file, but will only
454 snapshot the current CPU (if supported). It only displays
455 the content of the snapshot for a given CPU, and if
456 written to, only clears this CPU buffer.
457
458 per_cpu/cpu0/snapshot_raw:
459
460 Similar to the trace_pipe_raw, but will read the binary format
461 from the snapshot buffer for the given CPU.
462
463 per_cpu/cpu0/stats:
464
465 This displays certain stats about the ring buffer:
466
467 entries: The number of events that are still in the buffer.
468
469 overrun: The number of lost events due to overwriting when
470 the buffer was full.
471
472 commit overrun: Should always be zero.
473 This gets set if so many events happened within a nested
474 event (ring buffer is re-entrant), that it fills the
475 buffer and starts dropping events.
476
477 bytes: Bytes actually read (not overwritten).
478
479 oldest event ts: The oldest timestamp in the buffer
480
481 now ts: The current timestamp
482
483 dropped events: Events lost due to overwrite option being off.
484
485 read events: The number of events read.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400486
487The Tracers
488-----------
489
Steven Rostedtf2d9c742008-07-15 10:57:33 -0400490Here is the list of current tracers that may be configured.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400491
Ingo Molnar57526742009-02-19 12:54:10 +0100492 "function"
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400493
Ingo Molnar57526742009-02-19 12:54:10 +0100494 Function call tracer to trace all kernel functions.
Frederic Weisbecker985ec202009-02-18 06:35:34 +0100495
Mike Frysingerbc5c6c02009-06-10 04:48:41 -0400496 "function_graph"
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400497
Ingo Molnar57526742009-02-19 12:54:10 +0100498 Similar to the function tracer except that the
499 function tracer probes the functions on their entry
500 whereas the function graph tracer traces on both entry
501 and exit of the functions. It then provides the ability
502 to draw a graph of function calls similar to C code
503 source.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400504
Ingo Molnar57526742009-02-19 12:54:10 +0100505 "irqsoff"
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400506
Ingo Molnar57526742009-02-19 12:54:10 +0100507 Traces the areas that disable interrupts and saves
508 the trace with the longest max latency.
509 See tracing_max_latency. When a new max is recorded,
510 it replaces the old trace. It is best to view this
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200511 trace with the latency-format option enabled.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400512
Ingo Molnar57526742009-02-19 12:54:10 +0100513 "preemptoff"
514
515 Similar to irqsoff but traces and records the amount of
516 time for which preemption is disabled.
517
518 "preemptirqsoff"
519
520 Similar to irqsoff and preemptoff, but traces and
521 records the largest time for which irqs and/or preemption
522 is disabled.
523
524 "wakeup"
525
526 Traces and records the max latency that it takes for
527 the highest priority task to get scheduled after
528 it has been woken up.
Geunsik Limcdfb0d32012-02-08 19:05:37 +0900529 Traces all tasks as an average developer would expect.
530
531 "wakeup_rt"
532
533 Traces and records the max latency that it takes for just
534 RT tasks (as the current "wakeup" does). This is useful
535 for those interested in wake up timings of RT tasks.
Ingo Molnar57526742009-02-19 12:54:10 +0100536
Ingo Molnar57526742009-02-19 12:54:10 +0100537 "nop"
538
539 This is the "trace nothing" tracer. To remove all
540 tracers from tracing simply echo "nop" into
541 current_tracer.
Markus Metzgere2ea5392009-01-19 10:35:58 +0100542
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400543
544Examples of using the tracer
545----------------------------
546
Ingo Molnar57526742009-02-19 12:54:10 +0100547Here are typical examples of using the tracers when controlling
548them only with the debugfs interface (without using any
549user-land utilities).
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400550
551Output format:
552--------------
553
Steven Rostedtf2d9c742008-07-15 10:57:33 -0400554Here is an example of the output format of the file "trace"
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400555
556 --------
Steven Rostedt9b803c02008-11-03 15:15:08 -0500557# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400558#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400559# entries-in-buffer/entries-written: 140080/250280 #P:4
560#
561# _-----=> irqs-off
562# / _----=> need-resched
563# | / _---=> hardirq/softirq
564# || / _--=> preempt-depth
565# ||| / delay
566# TASK-PID CPU# |||| TIMESTAMP FUNCTION
567# | | | |||| | |
568 bash-1977 [000] .... 17284.993652: sys_close <-system_call_fastpath
569 bash-1977 [000] .... 17284.993653: __close_fd <-sys_close
570 bash-1977 [000] .... 17284.993653: _raw_spin_lock <-__close_fd
571 sshd-1974 [003] .... 17284.993653: __srcu_read_unlock <-fsnotify
572 bash-1977 [000] .... 17284.993654: add_preempt_count <-_raw_spin_lock
573 bash-1977 [000] ...1 17284.993655: _raw_spin_unlock <-__close_fd
574 bash-1977 [000] ...1 17284.993656: sub_preempt_count <-_raw_spin_unlock
575 bash-1977 [000] .... 17284.993657: filp_close <-__close_fd
576 bash-1977 [000] .... 17284.993657: dnotify_flush <-filp_close
577 sshd-1974 [003] .... 17284.993658: sys_select <-system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400578 --------
579
Ingo Molnar57526742009-02-19 12:54:10 +0100580A header is printed with the tracer name that is represented by
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400581the trace. In this case the tracer is "function". Then it shows the
582number of events in the buffer as well as the total number of entries
583that were written. The difference is the number of entries that were
584lost due to the buffer filling up (250280 - 140080 = 110200 events
585lost).
586
587The header explains the content of the events. Task name "bash", the task
588PID "1977", the CPU that it was running on "000", the latency format
589(explained below), the timestamp in <secs>.<usecs> format, the
590function name that was traced "sys_close" and the parent function that
591called this function "system_call_fastpath". The timestamp is the time
592at which the function was entered.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400593
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400594Latency trace format
595--------------------
596
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400597When the latency-format option is enabled or when one of the latency
598tracers is set, the trace file gives somewhat more information to see
599why a latency happened. Here is a typical trace.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400600
601# tracer: irqsoff
602#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400603# irqsoff latency trace v1.1.5 on 3.8.0-test+
604# --------------------------------------------------------------------
605# latency: 259 us, #4/4, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
606# -----------------
607# | task: ps-6143 (uid:0 nice:0 policy:0 rt_prio:0)
608# -----------------
609# => started at: __lock_task_sighand
610# => ended at: _raw_spin_unlock_irqrestore
611#
612#
613# _------=> CPU#
614# / _-----=> irqs-off
615# | / _----=> need-resched
616# || / _---=> hardirq/softirq
617# ||| / _--=> preempt-depth
618# |||| / delay
619# cmd pid ||||| time | caller
620# \ / ||||| \ | /
621 ps-6143 2d... 0us!: trace_hardirqs_off <-__lock_task_sighand
622 ps-6143 2d..1 259us+: trace_hardirqs_on <-_raw_spin_unlock_irqrestore
623 ps-6143 2d..1 263us+: time_hardirqs_on <-_raw_spin_unlock_irqrestore
624 ps-6143 2d..1 306us : <stack trace>
625 => trace_hardirqs_on_caller
626 => trace_hardirqs_on
627 => _raw_spin_unlock_irqrestore
628 => do_task_stat
629 => proc_tgid_stat
630 => proc_single_show
631 => seq_read
632 => vfs_read
633 => sys_read
634 => system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400635
636
Ingo Molnar57526742009-02-19 12:54:10 +0100637This shows that the current tracer is "irqsoff" tracing the time
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400638for which interrupts were disabled. It gives the trace version (which
639never changes) and the version of the kernel upon which this was executed on
640(3.10). Then it displays the max latency in microseconds (259 us). The number
641of trace entries displayed and the total number (both are four: #4/4).
642VP, KP, SP, and HP are always zero and are reserved for later use.
643#P is the number of online CPUs (#P:4).
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400644
Ingo Molnar57526742009-02-19 12:54:10 +0100645The task is the process that was running when the latency
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400646occurred. (ps pid: 6143).
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400647
Ingo Molnar57526742009-02-19 12:54:10 +0100648The start and stop (the functions in which the interrupts were
649disabled and enabled respectively) that caused the latencies:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400650
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400651 __lock_task_sighand is where the interrupts were disabled.
652 _raw_spin_unlock_irqrestore is where they were enabled again.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400653
654The next lines after the header are the trace itself. The header
655explains which is which.
656
657 cmd: The name of the process in the trace.
658
659 pid: The PID of that process.
660
Steven Rostedtf2d9c742008-07-15 10:57:33 -0400661 CPU#: The CPU which the process was running on.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400662
663 irqs-off: 'd' interrupts are disabled. '.' otherwise.
Steven Rostedt92444892008-10-24 09:42:59 -0400664 Note: If the architecture does not support a way to
665 read the irq flags variable, an 'X' will always
666 be printed here.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400667
Peter Zijlstrae5137b52013-10-04 17:28:26 +0200668 need-resched:
669 'N' both TIF_NEED_RESCHED and PREEMPT_NEED_RESCHED is set,
670 'n' only TIF_NEED_RESCHED is set,
671 'p' only PREEMPT_NEED_RESCHED is set,
672 '.' otherwise.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400673
674 hardirq/softirq:
Steven Rostedtf2d9c742008-07-15 10:57:33 -0400675 'H' - hard irq occurred inside a softirq.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400676 'h' - hard irq is running
677 's' - soft irq is running
678 '.' - normal context.
679
680 preempt-depth: The level of preempt_disabled
681
682The above is mostly meaningful for kernel developers.
683
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200684 time: When the latency-format option is enabled, the trace file
685 output includes a timestamp relative to the start of the
686 trace. This differs from the output when latency-format
687 is disabled, which includes an absolute timestamp.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400688
689 delay: This is just to help catch your eye a bit better. And
Ingo Molnar57526742009-02-19 12:54:10 +0100690 needs to be fixed to be only relative to the same CPU.
691 The marks are determined by the difference between this
692 current trace and the next trace.
Byungchul Park8e1e1df2014-11-24 09:34:19 +0900693 '$' - greater than 1 second
Jungseok Leeb838e1d2015-07-11 14:51:40 +0000694 '@' - greater than 100 milisecond
695 '*' - greater than 10 milisecond
Byungchul Park8e1e1df2014-11-24 09:34:19 +0900696 '#' - greater than 1000 microsecond
697 '!' - greater than 100 microsecond
698 '+' - greater than 10 microsecond
699 ' ' - less than or equal to 10 microsecond.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400700
701 The rest is the same as the 'trace' file.
702
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400703 Note, the latency tracers will usually end with a back trace
704 to easily find where the latency occurred.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400705
Steven Rostedtee6bce52008-11-12 17:52:37 -0500706trace_options
707-------------
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400708
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400709The trace_options file (or the options directory) is used to control
710what gets printed in the trace output, or manipulate the tracers.
711To see what is available, simply cat the file:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400712
GeunSik Lim156f5a72009-06-02 15:01:37 +0900713 cat trace_options
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400714print-parent
715nosym-offset
716nosym-addr
717noverbose
718noraw
719nohex
720nobin
721noblock
722nostacktrace
723trace_printk
724noftrace_preempt
725nobranch
726annotate
727nouserstacktrace
728nosym-userobj
729noprintk-msg-only
730context-info
731latency-format
732sleep-time
733graph-time
734record-cmd
735overwrite
736nodisable_on_free
737irq-info
738markers
739function-trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400740
Ingo Molnar57526742009-02-19 12:54:10 +0100741To disable one of the options, echo in the option prepended with
742"no".
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400743
GeunSik Lim156f5a72009-06-02 15:01:37 +0900744 echo noprint-parent > trace_options
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400745
746To enable an option, leave off the "no".
747
GeunSik Lim156f5a72009-06-02 15:01:37 +0900748 echo sym-offset > trace_options
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400749
750Here are the available options:
751
Ingo Molnar57526742009-02-19 12:54:10 +0100752 print-parent - On function traces, display the calling (parent)
753 function as well as the function being traced.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400754
755 print-parent:
Jingoo Han0ab943b2013-07-31 14:59:16 -0700756 bash-4000 [01] 1477.606694: simple_strtoul <-kstrtoul
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400757
758 noprint-parent:
759 bash-4000 [01] 1477.606694: simple_strtoul
760
761
Ingo Molnar57526742009-02-19 12:54:10 +0100762 sym-offset - Display not only the function name, but also the
763 offset in the function. For example, instead of
764 seeing just "ktime_get", you will see
765 "ktime_get+0xb/0x20".
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400766
767 sym-offset:
768 bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0
769
Ingo Molnar57526742009-02-19 12:54:10 +0100770 sym-addr - this will also display the function address as well
771 as the function name.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400772
773 sym-addr:
774 bash-4000 [01] 1477.606694: simple_strtoul <c0339346>
775
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200776 verbose - This deals with the trace file when the
777 latency-format option is enabled.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400778
779 bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \
Jingoo Han0ab943b2013-07-31 14:59:16 -0700780 (+0.000ms): simple_strtoul (kstrtoul)
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400781
Ingo Molnar57526742009-02-19 12:54:10 +0100782 raw - This will display raw numbers. This option is best for
783 use with user applications that can translate the raw
784 numbers better than having it done in the kernel.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400785
Ingo Molnar57526742009-02-19 12:54:10 +0100786 hex - Similar to raw, but the numbers will be in a hexadecimal
787 format.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400788
789 bin - This will print out the formats in raw binary.
790
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400791 block - When set, reading trace_pipe will not block when polled.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400792
Ingo Molnar57526742009-02-19 12:54:10 +0100793 stacktrace - This is one of the options that changes the trace
794 itself. When a trace is recorded, so is the stack
795 of functions. This allows for back traces of
796 trace sites.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400797
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400798 trace_printk - Can disable trace_printk() from writing into the buffer.
799
800 branch - Enable branch tracing with the tracer.
801
802 annotate - It is sometimes confusing when the CPU buffers are full
803 and one CPU buffer had a lot of events recently, thus
804 a shorter time frame, were another CPU may have only had
805 a few events, which lets it have older events. When
806 the trace is reported, it shows the oldest events first,
807 and it may look like only one CPU ran (the one with the
808 oldest events). When the annotate option is set, it will
809 display when a new CPU buffer started:
810
811 <idle>-0 [001] dNs4 21169.031481: wake_up_idle_cpu <-add_timer_on
812 <idle>-0 [001] dNs4 21169.031482: _raw_spin_unlock_irqrestore <-add_timer_on
813 <idle>-0 [001] .Ns4 21169.031484: sub_preempt_count <-_raw_spin_unlock_irqrestore
814##### CPU 2 buffer started ####
815 <idle>-0 [002] .N.1 21169.031484: rcu_idle_exit <-cpu_idle
816 <idle>-0 [001] .Ns3 21169.031484: _raw_spin_unlock <-clocksource_watchdog
817 <idle>-0 [001] .Ns3 21169.031485: sub_preempt_count <-_raw_spin_unlock
818
Ingo Molnar57526742009-02-19 12:54:10 +0100819 userstacktrace - This option changes the trace. It records a
820 stacktrace of the current userspace thread.
Török Edwin02b67512008-11-22 13:28:47 +0200821
Ingo Molnar57526742009-02-19 12:54:10 +0100822 sym-userobj - when user stacktrace are enabled, look up which
823 object the address belongs to, and print a
824 relative address. This is especially useful when
825 ASLR is on, otherwise you don't get a chance to
826 resolve the address to object/file/line after
827 the app is no longer running
Török Edwinb54d3de2008-11-22 13:28:48 +0200828
Ingo Molnar57526742009-02-19 12:54:10 +0100829 The lookup is performed when you read
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200830 trace,trace_pipe. Example:
Török Edwinb54d3de2008-11-22 13:28:48 +0200831
832 a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0
833x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6]
834
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400835
836 printk-msg-only - When set, trace_printk()s will only show the format
837 and not their parameters (if trace_bprintk() or
838 trace_bputs() was used to save the trace_printk()).
839
840 context-info - Show only the event data. Hides the comm, PID,
841 timestamp, CPU, and other useful data.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400842
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200843 latency-format - This option changes the trace. When
844 it is enabled, the trace displays
845 additional information about the
846 latencies, as described in "Latency
847 trace format".
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400848
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400849 sleep-time - When running function graph tracer, to include
850 the time a task schedules out in its function.
851 When enabled, it will account time the task has been
852 scheduled out as part of the function call.
853
854 graph-time - When running function graph tracer, to include the
855 time to call nested functions. When this is not set,
856 the time reported for the function will only include
857 the time the function itself executed for, not the time
858 for functions that it called.
859
860 record-cmd - When any event or tracer is enabled, a hook is enabled
861 in the sched_switch trace point to fill comm cache
862 with mapped pids and comms. But this may cause some
863 overhead, and if you only care about pids, and not the
864 name of the task, disabling this option can lower the
865 impact of tracing.
866
David Sharp750912f2010-12-08 13:46:47 -0800867 overwrite - This controls what happens when the trace buffer is
868 full. If "1" (default), the oldest events are
869 discarded and overwritten. If "0", then the newest
870 events are discarded.
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400871 (see per_cpu/cpu0/stats for overrun and dropped)
David Sharp750912f2010-12-08 13:46:47 -0800872
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400873 disable_on_free - When the free_buffer is closed, tracing will
874 stop (tracing_on set to 0).
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400875
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400876 irq-info - Shows the interrupt, preempt count, need resched data.
877 When disabled, the trace looks like:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400878
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400879# tracer: function
880#
881# entries-in-buffer/entries-written: 144405/9452052 #P:4
882#
883# TASK-PID CPU# TIMESTAMP FUNCTION
884# | | | | |
885 <idle>-0 [002] 23636.756054: ttwu_do_activate.constprop.89 <-try_to_wake_up
886 <idle>-0 [002] 23636.756054: activate_task <-ttwu_do_activate.constprop.89
887 <idle>-0 [002] 23636.756055: enqueue_task <-activate_task
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400888
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400889
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400890 markers - When set, the trace_marker is writable (only by root).
891 When disabled, the trace_marker will error with EINVAL
892 on write.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400893
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400894
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400895 function-trace - The latency tracers will enable function tracing
896 if this option is enabled (default it is). When
897 it is disabled, the latency tracers do not trace
898 functions. This keeps the overhead of the tracer down
899 when performing latency tests.
900
901 Note: Some tracers have their own options. They only appear
902 when the tracer is active.
903
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400904
905
906irqsoff
907-------
908
909When interrupts are disabled, the CPU can not react to any other
910external event (besides NMIs and SMIs). This prevents the timer
Ingo Molnar57526742009-02-19 12:54:10 +0100911interrupt from triggering or the mouse interrupt from letting
912the kernel know of a new mouse event. The result is a latency
913with the reaction time.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400914
Ingo Molnar57526742009-02-19 12:54:10 +0100915The irqsoff tracer tracks the time for which interrupts are
916disabled. When a new maximum latency is hit, the tracer saves
917the trace leading up to that latency point so that every time a
918new maximum is reached, the old saved trace is discarded and the
919new trace is saved.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400920
Ingo Molnar57526742009-02-19 12:54:10 +0100921To reset the maximum, echo 0 into tracing_max_latency. Here is
922an example:
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400923
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400924 # echo 0 > options/function-trace
GeunSik Lim156f5a72009-06-02 15:01:37 +0900925 # echo irqsoff > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -0500926 # echo 1 > tracing_on
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400927 # echo 0 > tracing_max_latency
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400928 # ls -ltr
929 [...]
Steven Rostedt6752ab42011-02-08 13:54:06 -0500930 # echo 0 > tracing_on
Albin Tonnerre4a88d442009-08-31 22:40:08 +0200931 # cat trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400932# tracer: irqsoff
933#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400934# irqsoff latency trace v1.1.5 on 3.8.0-test+
935# --------------------------------------------------------------------
936# latency: 16 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
937# -----------------
938# | task: swapper/0-0 (uid:0 nice:0 policy:0 rt_prio:0)
939# -----------------
940# => started at: run_timer_softirq
941# => ended at: run_timer_softirq
942#
943#
944# _------=> CPU#
945# / _-----=> irqs-off
946# | / _----=> need-resched
947# || / _---=> hardirq/softirq
948# ||| / _--=> preempt-depth
949# |||| / delay
950# cmd pid ||||| time | caller
951# \ / ||||| \ | /
952 <idle>-0 0d.s2 0us+: _raw_spin_lock_irq <-run_timer_softirq
953 <idle>-0 0dNs3 17us : _raw_spin_unlock_irq <-run_timer_softirq
954 <idle>-0 0dNs3 17us+: trace_hardirqs_on <-run_timer_softirq
955 <idle>-0 0dNs3 25us : <stack trace>
956 => _raw_spin_unlock_irq
957 => run_timer_softirq
958 => __do_softirq
959 => call_softirq
960 => do_softirq
961 => irq_exit
962 => smp_apic_timer_interrupt
963 => apic_timer_interrupt
964 => rcu_idle_exit
965 => cpu_idle
966 => rest_init
967 => start_kernel
968 => x86_64_start_reservations
969 => x86_64_start_kernel
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400970
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400971Here we see that that we had a latency of 16 microseconds (which is
972very good). The _raw_spin_lock_irq in run_timer_softirq disabled
973interrupts. The difference between the 16 and the displayed
974timestamp 25us occurred because the clock was incremented
Ingo Molnar57526742009-02-19 12:54:10 +0100975between the time of recording the max latency and the time of
976recording the function that had that latency.
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400977
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400978Note the above example had function-trace not set. If we set
979function-trace, we get a much larger output:
980
981 with echo 1 > options/function-trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400982
983# tracer: irqsoff
984#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -0400985# irqsoff latency trace v1.1.5 on 3.8.0-test+
986# --------------------------------------------------------------------
987# latency: 71 us, #168/168, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
988# -----------------
989# | task: bash-2042 (uid:0 nice:0 policy:0 rt_prio:0)
990# -----------------
991# => started at: ata_scsi_queuecmd
992# => ended at: ata_scsi_queuecmd
993#
994#
995# _------=> CPU#
996# / _-----=> irqs-off
997# | / _----=> need-resched
998# || / _---=> hardirq/softirq
999# ||| / _--=> preempt-depth
1000# |||| / delay
1001# cmd pid ||||| time | caller
1002# \ / ||||| \ | /
1003 bash-2042 3d... 0us : _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1004 bash-2042 3d... 0us : add_preempt_count <-_raw_spin_lock_irqsave
1005 bash-2042 3d..1 1us : ata_scsi_find_dev <-ata_scsi_queuecmd
1006 bash-2042 3d..1 1us : __ata_scsi_find_dev <-ata_scsi_find_dev
1007 bash-2042 3d..1 2us : ata_find_dev.part.14 <-__ata_scsi_find_dev
1008 bash-2042 3d..1 2us : ata_qc_new_init <-__ata_scsi_queuecmd
1009 bash-2042 3d..1 3us : ata_sg_init <-__ata_scsi_queuecmd
1010 bash-2042 3d..1 4us : ata_scsi_rw_xlat <-__ata_scsi_queuecmd
1011 bash-2042 3d..1 4us : ata_build_rw_tf <-ata_scsi_rw_xlat
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001012[...]
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001013 bash-2042 3d..1 67us : delay_tsc <-__delay
1014 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1015 bash-2042 3d..2 67us : sub_preempt_count <-delay_tsc
1016 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc
1017 bash-2042 3d..2 68us : sub_preempt_count <-delay_tsc
1018 bash-2042 3d..1 68us+: ata_bmdma_start <-ata_bmdma_qc_issue
1019 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1020 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1021 bash-2042 3d..1 72us+: trace_hardirqs_on <-ata_scsi_queuecmd
1022 bash-2042 3d..1 120us : <stack trace>
1023 => _raw_spin_unlock_irqrestore
1024 => ata_scsi_queuecmd
1025 => scsi_dispatch_cmd
1026 => scsi_request_fn
1027 => __blk_run_queue_uncond
1028 => __blk_run_queue
1029 => blk_queue_bio
1030 => generic_make_request
1031 => submit_bio
1032 => submit_bh
1033 => __ext3_get_inode_loc
1034 => ext3_iget
1035 => ext3_lookup
1036 => lookup_real
1037 => __lookup_hash
1038 => walk_component
1039 => lookup_last
1040 => path_lookupat
1041 => filename_lookup
1042 => user_path_at_empty
1043 => user_path_at
1044 => vfs_fstatat
1045 => vfs_stat
1046 => sys_newstat
1047 => system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001048
1049
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001050Here we traced a 71 microsecond latency. But we also see all the
Ingo Molnar57526742009-02-19 12:54:10 +01001051functions that were called during that time. Note that by
1052enabling function tracing, we incur an added overhead. This
1053overhead may extend the latency times. But nevertheless, this
1054trace has provided some very helpful debugging information.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001055
1056
1057preemptoff
1058----------
1059
Ingo Molnar57526742009-02-19 12:54:10 +01001060When preemption is disabled, we may be able to receive
1061interrupts but the task cannot be preempted and a higher
1062priority task must wait for preemption to be enabled again
1063before it can preempt a lower priority task.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001064
Steven Rostedta41eeba2008-07-14 16:41:12 -04001065The preemptoff tracer traces the places that disable preemption.
Ingo Molnar57526742009-02-19 12:54:10 +01001066Like the irqsoff tracer, it records the maximum latency for
1067which preemption was disabled. The control of preemptoff tracer
1068is much like the irqsoff tracer.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001069
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001070 # echo 0 > options/function-trace
GeunSik Lim156f5a72009-06-02 15:01:37 +09001071 # echo preemptoff > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -05001072 # echo 1 > tracing_on
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001073 # echo 0 > tracing_max_latency
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001074 # ls -ltr
1075 [...]
Steven Rostedt6752ab42011-02-08 13:54:06 -05001076 # echo 0 > tracing_on
Albin Tonnerre4a88d442009-08-31 22:40:08 +02001077 # cat trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001078# tracer: preemptoff
1079#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001080# preemptoff latency trace v1.1.5 on 3.8.0-test+
1081# --------------------------------------------------------------------
1082# latency: 46 us, #4/4, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1083# -----------------
1084# | task: sshd-1991 (uid:0 nice:0 policy:0 rt_prio:0)
1085# -----------------
1086# => started at: do_IRQ
1087# => ended at: do_IRQ
1088#
1089#
1090# _------=> CPU#
1091# / _-----=> irqs-off
1092# | / _----=> need-resched
1093# || / _---=> hardirq/softirq
1094# ||| / _--=> preempt-depth
1095# |||| / delay
1096# cmd pid ||||| time | caller
1097# \ / ||||| \ | /
1098 sshd-1991 1d.h. 0us+: irq_enter <-do_IRQ
1099 sshd-1991 1d..1 46us : irq_exit <-do_IRQ
1100 sshd-1991 1d..1 47us+: trace_preempt_on <-do_IRQ
1101 sshd-1991 1d..1 52us : <stack trace>
1102 => sub_preempt_count
1103 => irq_exit
1104 => do_IRQ
1105 => ret_from_intr
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001106
1107
Ingo Molnar57526742009-02-19 12:54:10 +01001108This has some more changes. Preemption was disabled when an
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001109interrupt came in (notice the 'h'), and was enabled on exit.
1110But we also see that interrupts have been disabled when entering
1111the preempt off section and leaving it (the 'd'). We do not know if
1112interrupts were enabled in the mean time or shortly after this
1113was over.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001114
1115# tracer: preemptoff
1116#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001117# preemptoff latency trace v1.1.5 on 3.8.0-test+
1118# --------------------------------------------------------------------
1119# latency: 83 us, #241/241, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1120# -----------------
1121# | task: bash-1994 (uid:0 nice:0 policy:0 rt_prio:0)
1122# -----------------
1123# => started at: wake_up_new_task
1124# => ended at: task_rq_unlock
1125#
1126#
1127# _------=> CPU#
1128# / _-----=> irqs-off
1129# | / _----=> need-resched
1130# || / _---=> hardirq/softirq
1131# ||| / _--=> preempt-depth
1132# |||| / delay
1133# cmd pid ||||| time | caller
1134# \ / ||||| \ | /
1135 bash-1994 1d..1 0us : _raw_spin_lock_irqsave <-wake_up_new_task
1136 bash-1994 1d..1 0us : select_task_rq_fair <-select_task_rq
1137 bash-1994 1d..1 1us : __rcu_read_lock <-select_task_rq_fair
1138 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
1139 bash-1994 1d..1 1us : source_load <-select_task_rq_fair
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001140[...]
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001141 bash-1994 1d..1 12us : irq_enter <-smp_apic_timer_interrupt
1142 bash-1994 1d..1 12us : rcu_irq_enter <-irq_enter
1143 bash-1994 1d..1 13us : add_preempt_count <-irq_enter
1144 bash-1994 1d.h1 13us : exit_idle <-smp_apic_timer_interrupt
1145 bash-1994 1d.h1 13us : hrtimer_interrupt <-smp_apic_timer_interrupt
1146 bash-1994 1d.h1 13us : _raw_spin_lock <-hrtimer_interrupt
1147 bash-1994 1d.h1 14us : add_preempt_count <-_raw_spin_lock
1148 bash-1994 1d.h2 14us : ktime_get_update_offsets <-hrtimer_interrupt
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001149[...]
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001150 bash-1994 1d.h1 35us : lapic_next_event <-clockevents_program_event
1151 bash-1994 1d.h1 35us : irq_exit <-smp_apic_timer_interrupt
1152 bash-1994 1d.h1 36us : sub_preempt_count <-irq_exit
1153 bash-1994 1d..2 36us : do_softirq <-irq_exit
1154 bash-1994 1d..2 36us : __do_softirq <-call_softirq
1155 bash-1994 1d..2 36us : __local_bh_disable <-__do_softirq
1156 bash-1994 1d.s2 37us : add_preempt_count <-_raw_spin_lock_irq
1157 bash-1994 1d.s3 38us : _raw_spin_unlock <-run_timer_softirq
1158 bash-1994 1d.s3 39us : sub_preempt_count <-_raw_spin_unlock
1159 bash-1994 1d.s2 39us : call_timer_fn <-run_timer_softirq
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001160[...]
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001161 bash-1994 1dNs2 81us : cpu_needs_another_gp <-rcu_process_callbacks
1162 bash-1994 1dNs2 82us : __local_bh_enable <-__do_softirq
1163 bash-1994 1dNs2 82us : sub_preempt_count <-__local_bh_enable
1164 bash-1994 1dN.2 82us : idle_cpu <-irq_exit
1165 bash-1994 1dN.2 83us : rcu_irq_exit <-irq_exit
1166 bash-1994 1dN.2 83us : sub_preempt_count <-irq_exit
1167 bash-1994 1.N.1 84us : _raw_spin_unlock_irqrestore <-task_rq_unlock
1168 bash-1994 1.N.1 84us+: trace_preempt_on <-task_rq_unlock
1169 bash-1994 1.N.1 104us : <stack trace>
1170 => sub_preempt_count
1171 => _raw_spin_unlock_irqrestore
1172 => task_rq_unlock
1173 => wake_up_new_task
1174 => do_fork
1175 => sys_clone
1176 => stub_clone
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001177
1178
Ingo Molnar57526742009-02-19 12:54:10 +01001179The above is an example of the preemptoff trace with
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001180function-trace set. Here we see that interrupts were not disabled
Ingo Molnar57526742009-02-19 12:54:10 +01001181the entire time. The irq_enter code lets us know that we entered
1182an interrupt 'h'. Before that, the functions being traced still
1183show that it is not in an interrupt, but we can see from the
1184functions themselves that this is not the case.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001185
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001186preemptirqsoff
1187--------------
1188
Ingo Molnar57526742009-02-19 12:54:10 +01001189Knowing the locations that have interrupts disabled or
1190preemption disabled for the longest times is helpful. But
1191sometimes we would like to know when either preemption and/or
1192interrupts are disabled.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001193
Steven Rostedtf2d9c742008-07-15 10:57:33 -04001194Consider the following code:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001195
1196 local_irq_disable();
1197 call_function_with_irqs_off();
1198 preempt_disable();
1199 call_function_with_irqs_and_preemption_off();
1200 local_irq_enable();
1201 call_function_with_preemption_off();
1202 preempt_enable();
1203
1204The irqsoff tracer will record the total length of
1205call_function_with_irqs_off() and
1206call_function_with_irqs_and_preemption_off().
1207
1208The preemptoff tracer will record the total length of
1209call_function_with_irqs_and_preemption_off() and
1210call_function_with_preemption_off().
1211
Ingo Molnar57526742009-02-19 12:54:10 +01001212But neither will trace the time that interrupts and/or
1213preemption is disabled. This total time is the time that we can
1214not schedule. To record this time, use the preemptirqsoff
1215tracer.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001216
Ingo Molnar57526742009-02-19 12:54:10 +01001217Again, using this trace is much like the irqsoff and preemptoff
1218tracers.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001219
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001220 # echo 0 > options/function-trace
GeunSik Lim156f5a72009-06-02 15:01:37 +09001221 # echo preemptirqsoff > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -05001222 # echo 1 > tracing_on
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001223 # echo 0 > tracing_max_latency
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001224 # ls -ltr
1225 [...]
Steven Rostedt6752ab42011-02-08 13:54:06 -05001226 # echo 0 > tracing_on
Albin Tonnerre4a88d442009-08-31 22:40:08 +02001227 # cat trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001228# tracer: preemptirqsoff
1229#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001230# preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1231# --------------------------------------------------------------------
1232# latency: 100 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1233# -----------------
1234# | task: ls-2230 (uid:0 nice:0 policy:0 rt_prio:0)
1235# -----------------
1236# => started at: ata_scsi_queuecmd
1237# => ended at: ata_scsi_queuecmd
1238#
1239#
1240# _------=> CPU#
1241# / _-----=> irqs-off
1242# | / _----=> need-resched
1243# || / _---=> hardirq/softirq
1244# ||| / _--=> preempt-depth
1245# |||| / delay
1246# cmd pid ||||| time | caller
1247# \ / ||||| \ | /
1248 ls-2230 3d... 0us+: _raw_spin_lock_irqsave <-ata_scsi_queuecmd
1249 ls-2230 3...1 100us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd
1250 ls-2230 3...1 101us+: trace_preempt_on <-ata_scsi_queuecmd
1251 ls-2230 3...1 111us : <stack trace>
1252 => sub_preempt_count
1253 => _raw_spin_unlock_irqrestore
1254 => ata_scsi_queuecmd
1255 => scsi_dispatch_cmd
1256 => scsi_request_fn
1257 => __blk_run_queue_uncond
1258 => __blk_run_queue
1259 => blk_queue_bio
1260 => generic_make_request
1261 => submit_bio
1262 => submit_bh
1263 => ext3_bread
1264 => ext3_dir_bread
1265 => htree_dirblock_to_tree
1266 => ext3_htree_fill_tree
1267 => ext3_readdir
1268 => vfs_readdir
1269 => sys_getdents
1270 => system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001271
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001272
1273The trace_hardirqs_off_thunk is called from assembly on x86 when
Ingo Molnar57526742009-02-19 12:54:10 +01001274interrupts are disabled in the assembly code. Without the
1275function tracing, we do not know if interrupts were enabled
1276within the preemption points. We do see that it started with
1277preemption enabled.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001278
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001279Here is a trace with function-trace set:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001280
1281# tracer: preemptirqsoff
1282#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001283# preemptirqsoff latency trace v1.1.5 on 3.8.0-test+
1284# --------------------------------------------------------------------
1285# latency: 161 us, #339/339, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1286# -----------------
1287# | task: ls-2269 (uid:0 nice:0 policy:0 rt_prio:0)
1288# -----------------
1289# => started at: schedule
1290# => ended at: mutex_unlock
1291#
1292#
1293# _------=> CPU#
1294# / _-----=> irqs-off
1295# | / _----=> need-resched
1296# || / _---=> hardirq/softirq
1297# ||| / _--=> preempt-depth
1298# |||| / delay
1299# cmd pid ||||| time | caller
1300# \ / ||||| \ | /
1301kworker/-59 3...1 0us : __schedule <-schedule
1302kworker/-59 3d..1 0us : rcu_preempt_qs <-rcu_note_context_switch
1303kworker/-59 3d..1 1us : add_preempt_count <-_raw_spin_lock_irq
1304kworker/-59 3d..2 1us : deactivate_task <-__schedule
1305kworker/-59 3d..2 1us : dequeue_task <-deactivate_task
1306kworker/-59 3d..2 2us : update_rq_clock <-dequeue_task
1307kworker/-59 3d..2 2us : dequeue_task_fair <-dequeue_task
1308kworker/-59 3d..2 2us : update_curr <-dequeue_task_fair
1309kworker/-59 3d..2 2us : update_min_vruntime <-update_curr
1310kworker/-59 3d..2 3us : cpuacct_charge <-update_curr
1311kworker/-59 3d..2 3us : __rcu_read_lock <-cpuacct_charge
1312kworker/-59 3d..2 3us : __rcu_read_unlock <-cpuacct_charge
1313kworker/-59 3d..2 3us : update_cfs_rq_blocked_load <-dequeue_task_fair
1314kworker/-59 3d..2 4us : clear_buddies <-dequeue_task_fair
1315kworker/-59 3d..2 4us : account_entity_dequeue <-dequeue_task_fair
1316kworker/-59 3d..2 4us : update_min_vruntime <-dequeue_task_fair
1317kworker/-59 3d..2 4us : update_cfs_shares <-dequeue_task_fair
1318kworker/-59 3d..2 5us : hrtick_update <-dequeue_task_fair
1319kworker/-59 3d..2 5us : wq_worker_sleeping <-__schedule
1320kworker/-59 3d..2 5us : kthread_data <-wq_worker_sleeping
1321kworker/-59 3d..2 5us : put_prev_task_fair <-__schedule
1322kworker/-59 3d..2 6us : pick_next_task_fair <-pick_next_task
1323kworker/-59 3d..2 6us : clear_buddies <-pick_next_task_fair
1324kworker/-59 3d..2 6us : set_next_entity <-pick_next_task_fair
1325kworker/-59 3d..2 6us : update_stats_wait_end <-set_next_entity
1326 ls-2269 3d..2 7us : finish_task_switch <-__schedule
1327 ls-2269 3d..2 7us : _raw_spin_unlock_irq <-finish_task_switch
1328 ls-2269 3d..2 8us : do_IRQ <-ret_from_intr
1329 ls-2269 3d..2 8us : irq_enter <-do_IRQ
1330 ls-2269 3d..2 8us : rcu_irq_enter <-irq_enter
1331 ls-2269 3d..2 9us : add_preempt_count <-irq_enter
1332 ls-2269 3d.h2 9us : exit_idle <-do_IRQ
1333[...]
1334 ls-2269 3d.h3 20us : sub_preempt_count <-_raw_spin_unlock
1335 ls-2269 3d.h2 20us : irq_exit <-do_IRQ
1336 ls-2269 3d.h2 21us : sub_preempt_count <-irq_exit
1337 ls-2269 3d..3 21us : do_softirq <-irq_exit
1338 ls-2269 3d..3 21us : __do_softirq <-call_softirq
1339 ls-2269 3d..3 21us+: __local_bh_disable <-__do_softirq
1340 ls-2269 3d.s4 29us : sub_preempt_count <-_local_bh_enable_ip
1341 ls-2269 3d.s5 29us : sub_preempt_count <-_local_bh_enable_ip
1342 ls-2269 3d.s5 31us : do_IRQ <-ret_from_intr
1343 ls-2269 3d.s5 31us : irq_enter <-do_IRQ
1344 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1345[...]
1346 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter
1347 ls-2269 3d.s5 32us : add_preempt_count <-irq_enter
1348 ls-2269 3d.H5 32us : exit_idle <-do_IRQ
1349 ls-2269 3d.H5 32us : handle_irq <-do_IRQ
1350 ls-2269 3d.H5 32us : irq_to_desc <-handle_irq
1351 ls-2269 3d.H5 33us : handle_fasteoi_irq <-handle_irq
1352[...]
1353 ls-2269 3d.s5 158us : _raw_spin_unlock_irqrestore <-rtl8139_poll
1354 ls-2269 3d.s3 158us : net_rps_action_and_irq_enable.isra.65 <-net_rx_action
1355 ls-2269 3d.s3 159us : __local_bh_enable <-__do_softirq
1356 ls-2269 3d.s3 159us : sub_preempt_count <-__local_bh_enable
1357 ls-2269 3d..3 159us : idle_cpu <-irq_exit
1358 ls-2269 3d..3 159us : rcu_irq_exit <-irq_exit
1359 ls-2269 3d..3 160us : sub_preempt_count <-irq_exit
1360 ls-2269 3d... 161us : __mutex_unlock_slowpath <-mutex_unlock
1361 ls-2269 3d... 162us+: trace_hardirqs_on <-mutex_unlock
1362 ls-2269 3d... 186us : <stack trace>
1363 => __mutex_unlock_slowpath
1364 => mutex_unlock
1365 => process_output
1366 => n_tty_write
1367 => tty_write
1368 => vfs_write
1369 => sys_write
1370 => system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001371
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001372This is an interesting trace. It started with kworker running and
1373scheduling out and ls taking over. But as soon as ls released the
1374rq lock and enabled interrupts (but not preemption) an interrupt
1375triggered. When the interrupt finished, it started running softirqs.
1376But while the softirq was running, another interrupt triggered.
1377When an interrupt is running inside a softirq, the annotation is 'H'.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001378
1379
1380wakeup
1381------
1382
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001383One common case that people are interested in tracing is the
1384time it takes for a task that is woken to actually wake up.
1385Now for non Real-Time tasks, this can be arbitrary. But tracing
1386it none the less can be interesting.
1387
1388Without function tracing:
1389
1390 # echo 0 > options/function-trace
1391 # echo wakeup > current_tracer
1392 # echo 1 > tracing_on
1393 # echo 0 > tracing_max_latency
1394 # chrt -f 5 sleep 1
1395 # echo 0 > tracing_on
1396 # cat trace
1397# tracer: wakeup
1398#
1399# wakeup latency trace v1.1.5 on 3.8.0-test+
1400# --------------------------------------------------------------------
1401# latency: 15 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1402# -----------------
1403# | task: kworker/3:1H-312 (uid:0 nice:-20 policy:0 rt_prio:0)
1404# -----------------
1405#
1406# _------=> CPU#
1407# / _-----=> irqs-off
1408# | / _----=> need-resched
1409# || / _---=> hardirq/softirq
1410# ||| / _--=> preempt-depth
1411# |||| / delay
1412# cmd pid ||||| time | caller
1413# \ / ||||| \ | /
1414 <idle>-0 3dNs7 0us : 0:120:R + [003] 312:100:R kworker/3:1H
1415 <idle>-0 3dNs7 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
1416 <idle>-0 3d..3 15us : __schedule <-schedule
1417 <idle>-0 3d..3 15us : 0:120:R ==> [003] 312:100:R kworker/3:1H
1418
1419The tracer only traces the highest priority task in the system
1420to avoid tracing the normal circumstances. Here we see that
1421the kworker with a nice priority of -20 (not very nice), took
1422just 15 microseconds from the time it woke up, to the time it
1423ran.
1424
1425Non Real-Time tasks are not that interesting. A more interesting
1426trace is to concentrate only on Real-Time tasks.
1427
1428wakeup_rt
1429---------
1430
Ingo Molnar57526742009-02-19 12:54:10 +01001431In a Real-Time environment it is very important to know the
1432wakeup time it takes for the highest priority task that is woken
1433up to the time that it executes. This is also known as "schedule
1434latency". I stress the point that this is about RT tasks. It is
1435also important to know the scheduling latency of non-RT tasks,
1436but the average schedule latency is better for non-RT tasks.
1437Tools like LatencyTop are more appropriate for such
1438measurements.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001439
Steven Rostedta41eeba2008-07-14 16:41:12 -04001440Real-Time environments are interested in the worst case latency.
Ingo Molnar57526742009-02-19 12:54:10 +01001441That is the longest latency it takes for something to happen,
1442and not the average. We can have a very fast scheduler that may
1443only have a large latency once in a while, but that would not
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001444work well with Real-Time tasks. The wakeup_rt tracer was designed
Ingo Molnar57526742009-02-19 12:54:10 +01001445to record the worst case wakeups of RT tasks. Non-RT tasks are
1446not recorded because the tracer only records one worst case and
1447tracing non-RT tasks that are unpredictable will overwrite the
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001448worst case latency of RT tasks (just run the normal wakeup
1449tracer for a while to see that effect).
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001450
Ingo Molnar57526742009-02-19 12:54:10 +01001451Since this tracer only deals with RT tasks, we will run this
1452slightly differently than we did with the previous tracers.
1453Instead of performing an 'ls', we will run 'sleep 1' under
1454'chrt' which changes the priority of the task.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001455
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001456 # echo 0 > options/function-trace
1457 # echo wakeup_rt > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -05001458 # echo 1 > tracing_on
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001459 # echo 0 > tracing_max_latency
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001460 # chrt -f 5 sleep 1
Steven Rostedt6752ab42011-02-08 13:54:06 -05001461 # echo 0 > tracing_on
Albin Tonnerre4a88d442009-08-31 22:40:08 +02001462 # cat trace
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001463# tracer: wakeup
1464#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001465# tracer: wakeup_rt
1466#
1467# wakeup_rt latency trace v1.1.5 on 3.8.0-test+
1468# --------------------------------------------------------------------
1469# latency: 5 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1470# -----------------
1471# | task: sleep-2389 (uid:0 nice:0 policy:1 rt_prio:5)
1472# -----------------
1473#
1474# _------=> CPU#
1475# / _-----=> irqs-off
1476# | / _----=> need-resched
1477# || / _---=> hardirq/softirq
1478# ||| / _--=> preempt-depth
1479# |||| / delay
1480# cmd pid ||||| time | caller
1481# \ / ||||| \ | /
1482 <idle>-0 3d.h4 0us : 0:120:R + [003] 2389: 94:R sleep
1483 <idle>-0 3d.h4 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up
1484 <idle>-0 3d..3 5us : __schedule <-schedule
1485 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001486
1487
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001488Running this on an idle system, we see that it only took 5 microseconds
1489to perform the task switch. Note, since the trace point in the schedule
1490is before the actual "switch", we stop the tracing when the recorded task
1491is about to schedule in. This may change if we add a new marker at the
1492end of the scheduler.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001493
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001494Notice that the recorded task is 'sleep' with the PID of 2389
Ingo Molnar57526742009-02-19 12:54:10 +01001495and it has an rt_prio of 5. This priority is user-space priority
1496and not the internal kernel priority. The policy is 1 for
1497SCHED_FIFO and 2 for SCHED_RR.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001498
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001499Note, that the trace data shows the internal priority (99 - rtprio).
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001500
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001501 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep
1502
1503The 0:120:R means idle was running with a nice priority of 0 (120 - 20)
1504and in the running state 'R'. The sleep task was scheduled in with
15052389: 94:R. That is the priority is the kernel rtprio (99 - 5 = 94)
1506and it too is in the running state.
1507
1508Doing the same with chrt -r 5 and function-trace set.
1509
1510 echo 1 > options/function-trace
1511
1512# tracer: wakeup_rt
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001513#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001514# wakeup_rt latency trace v1.1.5 on 3.8.0-test+
1515# --------------------------------------------------------------------
1516# latency: 29 us, #85/85, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1517# -----------------
1518# | task: sleep-2448 (uid:0 nice:0 policy:1 rt_prio:5)
1519# -----------------
1520#
1521# _------=> CPU#
1522# / _-----=> irqs-off
1523# | / _----=> need-resched
1524# || / _---=> hardirq/softirq
1525# ||| / _--=> preempt-depth
1526# |||| / delay
1527# cmd pid ||||| time | caller
1528# \ / ||||| \ | /
1529 <idle>-0 3d.h4 1us+: 0:120:R + [003] 2448: 94:R sleep
1530 <idle>-0 3d.h4 2us : ttwu_do_activate.constprop.87 <-try_to_wake_up
1531 <idle>-0 3d.h3 3us : check_preempt_curr <-ttwu_do_wakeup
Kirill Tkhai88751252014-06-29 00:03:57 +04001532 <idle>-0 3d.h3 3us : resched_curr <-check_preempt_curr
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001533 <idle>-0 3dNh3 4us : task_woken_rt <-ttwu_do_wakeup
1534 <idle>-0 3dNh3 4us : _raw_spin_unlock <-try_to_wake_up
1535 <idle>-0 3dNh3 4us : sub_preempt_count <-_raw_spin_unlock
1536 <idle>-0 3dNh2 5us : ttwu_stat <-try_to_wake_up
1537 <idle>-0 3dNh2 5us : _raw_spin_unlock_irqrestore <-try_to_wake_up
1538 <idle>-0 3dNh2 6us : sub_preempt_count <-_raw_spin_unlock_irqrestore
1539 <idle>-0 3dNh1 6us : _raw_spin_lock <-__run_hrtimer
1540 <idle>-0 3dNh1 6us : add_preempt_count <-_raw_spin_lock
1541 <idle>-0 3dNh2 7us : _raw_spin_unlock <-hrtimer_interrupt
1542 <idle>-0 3dNh2 7us : sub_preempt_count <-_raw_spin_unlock
1543 <idle>-0 3dNh1 7us : tick_program_event <-hrtimer_interrupt
1544 <idle>-0 3dNh1 7us : clockevents_program_event <-tick_program_event
1545 <idle>-0 3dNh1 8us : ktime_get <-clockevents_program_event
1546 <idle>-0 3dNh1 8us : lapic_next_event <-clockevents_program_event
1547 <idle>-0 3dNh1 8us : irq_exit <-smp_apic_timer_interrupt
1548 <idle>-0 3dNh1 9us : sub_preempt_count <-irq_exit
1549 <idle>-0 3dN.2 9us : idle_cpu <-irq_exit
1550 <idle>-0 3dN.2 9us : rcu_irq_exit <-irq_exit
1551 <idle>-0 3dN.2 10us : rcu_eqs_enter_common.isra.45 <-rcu_irq_exit
1552 <idle>-0 3dN.2 10us : sub_preempt_count <-irq_exit
1553 <idle>-0 3.N.1 11us : rcu_idle_exit <-cpu_idle
1554 <idle>-0 3dN.1 11us : rcu_eqs_exit_common.isra.43 <-rcu_idle_exit
1555 <idle>-0 3.N.1 11us : tick_nohz_idle_exit <-cpu_idle
1556 <idle>-0 3dN.1 12us : menu_hrtimer_cancel <-tick_nohz_idle_exit
1557 <idle>-0 3dN.1 12us : ktime_get <-tick_nohz_idle_exit
1558 <idle>-0 3dN.1 12us : tick_do_update_jiffies64 <-tick_nohz_idle_exit
1559 <idle>-0 3dN.1 13us : update_cpu_load_nohz <-tick_nohz_idle_exit
1560 <idle>-0 3dN.1 13us : _raw_spin_lock <-update_cpu_load_nohz
1561 <idle>-0 3dN.1 13us : add_preempt_count <-_raw_spin_lock
1562 <idle>-0 3dN.2 13us : __update_cpu_load <-update_cpu_load_nohz
1563 <idle>-0 3dN.2 14us : sched_avg_update <-__update_cpu_load
1564 <idle>-0 3dN.2 14us : _raw_spin_unlock <-update_cpu_load_nohz
1565 <idle>-0 3dN.2 14us : sub_preempt_count <-_raw_spin_unlock
1566 <idle>-0 3dN.1 15us : calc_load_exit_idle <-tick_nohz_idle_exit
1567 <idle>-0 3dN.1 15us : touch_softlockup_watchdog <-tick_nohz_idle_exit
1568 <idle>-0 3dN.1 15us : hrtimer_cancel <-tick_nohz_idle_exit
1569 <idle>-0 3dN.1 15us : hrtimer_try_to_cancel <-hrtimer_cancel
1570 <idle>-0 3dN.1 16us : lock_hrtimer_base.isra.18 <-hrtimer_try_to_cancel
1571 <idle>-0 3dN.1 16us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
1572 <idle>-0 3dN.1 16us : add_preempt_count <-_raw_spin_lock_irqsave
1573 <idle>-0 3dN.2 17us : __remove_hrtimer <-remove_hrtimer.part.16
1574 <idle>-0 3dN.2 17us : hrtimer_force_reprogram <-__remove_hrtimer
1575 <idle>-0 3dN.2 17us : tick_program_event <-hrtimer_force_reprogram
1576 <idle>-0 3dN.2 18us : clockevents_program_event <-tick_program_event
1577 <idle>-0 3dN.2 18us : ktime_get <-clockevents_program_event
1578 <idle>-0 3dN.2 18us : lapic_next_event <-clockevents_program_event
1579 <idle>-0 3dN.2 19us : _raw_spin_unlock_irqrestore <-hrtimer_try_to_cancel
1580 <idle>-0 3dN.2 19us : sub_preempt_count <-_raw_spin_unlock_irqrestore
1581 <idle>-0 3dN.1 19us : hrtimer_forward <-tick_nohz_idle_exit
1582 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
1583 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward
1584 <idle>-0 3dN.1 20us : hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
1585 <idle>-0 3dN.1 20us : __hrtimer_start_range_ns <-hrtimer_start_range_ns
1586 <idle>-0 3dN.1 21us : lock_hrtimer_base.isra.18 <-__hrtimer_start_range_ns
1587 <idle>-0 3dN.1 21us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18
1588 <idle>-0 3dN.1 21us : add_preempt_count <-_raw_spin_lock_irqsave
1589 <idle>-0 3dN.2 22us : ktime_add_safe <-__hrtimer_start_range_ns
1590 <idle>-0 3dN.2 22us : enqueue_hrtimer <-__hrtimer_start_range_ns
1591 <idle>-0 3dN.2 22us : tick_program_event <-__hrtimer_start_range_ns
1592 <idle>-0 3dN.2 23us : clockevents_program_event <-tick_program_event
1593 <idle>-0 3dN.2 23us : ktime_get <-clockevents_program_event
1594 <idle>-0 3dN.2 23us : lapic_next_event <-clockevents_program_event
1595 <idle>-0 3dN.2 24us : _raw_spin_unlock_irqrestore <-__hrtimer_start_range_ns
1596 <idle>-0 3dN.2 24us : sub_preempt_count <-_raw_spin_unlock_irqrestore
1597 <idle>-0 3dN.1 24us : account_idle_ticks <-tick_nohz_idle_exit
1598 <idle>-0 3dN.1 24us : account_idle_time <-account_idle_ticks
1599 <idle>-0 3.N.1 25us : sub_preempt_count <-cpu_idle
1600 <idle>-0 3.N.. 25us : schedule <-cpu_idle
1601 <idle>-0 3.N.. 25us : __schedule <-preempt_schedule
1602 <idle>-0 3.N.. 26us : add_preempt_count <-__schedule
1603 <idle>-0 3.N.1 26us : rcu_note_context_switch <-__schedule
1604 <idle>-0 3.N.1 26us : rcu_sched_qs <-rcu_note_context_switch
1605 <idle>-0 3dN.1 27us : rcu_preempt_qs <-rcu_note_context_switch
1606 <idle>-0 3.N.1 27us : _raw_spin_lock_irq <-__schedule
1607 <idle>-0 3dN.1 27us : add_preempt_count <-_raw_spin_lock_irq
1608 <idle>-0 3dN.2 28us : put_prev_task_idle <-__schedule
1609 <idle>-0 3dN.2 28us : pick_next_task_stop <-pick_next_task
1610 <idle>-0 3dN.2 28us : pick_next_task_rt <-pick_next_task
1611 <idle>-0 3dN.2 29us : dequeue_pushable_task <-pick_next_task_rt
1612 <idle>-0 3d..3 29us : __schedule <-preempt_schedule
1613 <idle>-0 3d..3 30us : 0:120:R ==> [003] 2448: 94:R sleep
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001614
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001615This isn't that big of a trace, even with function tracing enabled,
1616so I included the entire trace.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001617
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001618The interrupt went off while when the system was idle. Somewhere
1619before task_woken_rt() was called, the NEED_RESCHED flag was set,
1620this is indicated by the first occurrence of the 'N' flag.
1621
1622Latency tracing and events
1623--------------------------
1624As function tracing can induce a much larger latency, but without
1625seeing what happens within the latency it is hard to know what
1626caused it. There is a middle ground, and that is with enabling
1627events.
1628
1629 # echo 0 > options/function-trace
1630 # echo wakeup_rt > current_tracer
1631 # echo 1 > events/enable
1632 # echo 1 > tracing_on
1633 # echo 0 > tracing_max_latency
1634 # chrt -f 5 sleep 1
1635 # echo 0 > tracing_on
1636 # cat trace
1637# tracer: wakeup_rt
1638#
1639# wakeup_rt latency trace v1.1.5 on 3.8.0-test+
1640# --------------------------------------------------------------------
1641# latency: 6 us, #12/12, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4)
1642# -----------------
1643# | task: sleep-5882 (uid:0 nice:0 policy:1 rt_prio:5)
1644# -----------------
1645#
1646# _------=> CPU#
1647# / _-----=> irqs-off
1648# | / _----=> need-resched
1649# || / _---=> hardirq/softirq
1650# ||| / _--=> preempt-depth
1651# |||| / delay
1652# cmd pid ||||| time | caller
1653# \ / ||||| \ | /
1654 <idle>-0 2d.h4 0us : 0:120:R + [002] 5882: 94:R sleep
1655 <idle>-0 2d.h4 0us : ttwu_do_activate.constprop.87 <-try_to_wake_up
1656 <idle>-0 2d.h4 1us : sched_wakeup: comm=sleep pid=5882 prio=94 success=1 target_cpu=002
1657 <idle>-0 2dNh2 1us : hrtimer_expire_exit: hrtimer=ffff88007796feb8
1658 <idle>-0 2.N.2 2us : power_end: cpu_id=2
1659 <idle>-0 2.N.2 3us : cpu_idle: state=4294967295 cpu_id=2
1660 <idle>-0 2dN.3 4us : hrtimer_cancel: hrtimer=ffff88007d50d5e0
1661 <idle>-0 2dN.3 4us : hrtimer_start: hrtimer=ffff88007d50d5e0 function=tick_sched_timer expires=34311211000000 softexpires=34311211000000
1662 <idle>-0 2.N.2 5us : rcu_utilization: Start context switch
1663 <idle>-0 2.N.2 5us : rcu_utilization: End context switch
1664 <idle>-0 2d..3 6us : __schedule <-schedule
1665 <idle>-0 2d..3 6us : 0:120:R ==> [002] 5882: 94:R sleep
1666
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001667
Steven Rostedt9b803c02008-11-03 15:15:08 -05001668function
1669--------
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001670
Steven Rostedt9b803c02008-11-03 15:15:08 -05001671This tracer is the function tracer. Enabling the function tracer
Ingo Molnar57526742009-02-19 12:54:10 +01001672can be done from the debug file system. Make sure the
1673ftrace_enabled is set; otherwise this tracer is a nop.
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001674See the "ftrace_enabled" section below.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001675
1676 # sysctl kernel.ftrace_enabled=1
GeunSik Lim156f5a72009-06-02 15:01:37 +09001677 # echo function > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -05001678 # echo 1 > tracing_on
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001679 # usleep 1
Steven Rostedt6752ab42011-02-08 13:54:06 -05001680 # echo 0 > tracing_on
GeunSik Lim156f5a72009-06-02 15:01:37 +09001681 # cat trace
Steven Rostedt9b803c02008-11-03 15:15:08 -05001682# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001683#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001684# entries-in-buffer/entries-written: 24799/24799 #P:4
1685#
1686# _-----=> irqs-off
1687# / _----=> need-resched
1688# | / _---=> hardirq/softirq
1689# || / _--=> preempt-depth
1690# ||| / delay
1691# TASK-PID CPU# |||| TIMESTAMP FUNCTION
1692# | | | |||| | |
1693 bash-1994 [002] .... 3082.063030: mutex_unlock <-rb_simple_write
1694 bash-1994 [002] .... 3082.063031: __mutex_unlock_slowpath <-mutex_unlock
1695 bash-1994 [002] .... 3082.063031: __fsnotify_parent <-fsnotify_modify
1696 bash-1994 [002] .... 3082.063032: fsnotify <-fsnotify_modify
1697 bash-1994 [002] .... 3082.063032: __srcu_read_lock <-fsnotify
1698 bash-1994 [002] .... 3082.063032: add_preempt_count <-__srcu_read_lock
1699 bash-1994 [002] ...1 3082.063032: sub_preempt_count <-__srcu_read_lock
1700 bash-1994 [002] .... 3082.063033: __srcu_read_unlock <-fsnotify
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001701[...]
1702
1703
Ingo Molnar57526742009-02-19 12:54:10 +01001704Note: function tracer uses ring buffers to store the above
1705entries. The newest data may overwrite the oldest data.
1706Sometimes using echo to stop the trace is not sufficient because
1707the tracing could have overwritten the data that you wanted to
1708record. For this reason, it is sometimes better to disable
1709tracing directly from a program. This allows you to stop the
1710tracing at the point that you hit the part that you are
1711interested in. To disable the tracing directly from a C program,
1712something like following code snippet can be used:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001713
1714int trace_fd;
1715[...]
1716int main(int argc, char *argv[]) {
1717 [...]
Steven Rostedt6752ab42011-02-08 13:54:06 -05001718 trace_fd = open(tracing_file("tracing_on"), O_WRONLY);
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001719 [...]
1720 if (condition_hit()) {
Steven Rostedtf2d9c742008-07-15 10:57:33 -04001721 write(trace_fd, "0", 1);
Steven Rostedteb6d42e2008-07-10 12:46:01 -04001722 }
1723 [...]
1724}
1725
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001726
1727Single thread tracing
1728---------------------
1729
GeunSik Lim156f5a72009-06-02 15:01:37 +09001730By writing into set_ftrace_pid you can trace a
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001731single thread. For example:
1732
GeunSik Lim156f5a72009-06-02 15:01:37 +09001733# cat set_ftrace_pid
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001734no pid
GeunSik Lim156f5a72009-06-02 15:01:37 +09001735# echo 3111 > set_ftrace_pid
1736# cat set_ftrace_pid
Steven Rostedtdf4fc312008-11-26 00:16:23 -050017373111
GeunSik Lim156f5a72009-06-02 15:01:37 +09001738# echo function > current_tracer
1739# cat trace | head
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001740 # tracer: function
1741 #
1742 # TASK-PID CPU# TIMESTAMP FUNCTION
1743 # | | | | |
1744 yum-updatesd-3111 [003] 1637.254676: finish_task_switch <-thread_return
1745 yum-updatesd-3111 [003] 1637.254681: hrtimer_cancel <-schedule_hrtimeout_range
1746 yum-updatesd-3111 [003] 1637.254682: hrtimer_try_to_cancel <-hrtimer_cancel
1747 yum-updatesd-3111 [003] 1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel
1748 yum-updatesd-3111 [003] 1637.254685: fget_light <-do_sys_poll
1749 yum-updatesd-3111 [003] 1637.254686: pipe_poll <-do_sys_poll
Wang Long52e68922015-02-05 05:16:14 +00001750# echo > set_ftrace_pid
GeunSik Lim156f5a72009-06-02 15:01:37 +09001751# cat trace |head
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001752 # tracer: function
1753 #
1754 # TASK-PID CPU# TIMESTAMP FUNCTION
1755 # | | | | |
1756 ##### CPU 3 buffer started ####
1757 yum-updatesd-3111 [003] 1701.957688: free_poll_entry <-poll_freewait
1758 yum-updatesd-3111 [003] 1701.957689: remove_wait_queue <-free_poll_entry
1759 yum-updatesd-3111 [003] 1701.957691: fput <-free_poll_entry
1760 yum-updatesd-3111 [003] 1701.957692: audit_syscall_exit <-sysret_audit
1761 yum-updatesd-3111 [003] 1701.957693: path_put <-audit_syscall_exit
1762
1763If you want to trace a function when executing, you could use
1764something like this simple program:
1765
1766#include <stdio.h>
1767#include <stdlib.h>
1768#include <sys/types.h>
1769#include <sys/stat.h>
1770#include <fcntl.h>
1771#include <unistd.h>
Jiri Olsa67b394f2009-10-23 19:36:18 -04001772#include <string.h>
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001773
GeunSik Lim156f5a72009-06-02 15:01:37 +09001774#define _STR(x) #x
1775#define STR(x) _STR(x)
1776#define MAX_PATH 256
1777
1778const char *find_debugfs(void)
1779{
1780 static char debugfs[MAX_PATH+1];
1781 static int debugfs_found;
1782 char type[100];
1783 FILE *fp;
1784
1785 if (debugfs_found)
1786 return debugfs;
1787
1788 if ((fp = fopen("/proc/mounts","r")) == NULL) {
1789 perror("/proc/mounts");
1790 return NULL;
1791 }
1792
1793 while (fscanf(fp, "%*s %"
1794 STR(MAX_PATH)
1795 "s %99s %*s %*d %*d\n",
1796 debugfs, type) == 2) {
1797 if (strcmp(type, "debugfs") == 0)
1798 break;
1799 }
1800 fclose(fp);
1801
1802 if (strcmp(type, "debugfs") != 0) {
1803 fprintf(stderr, "debugfs not mounted");
1804 return NULL;
1805 }
1806
Jiri Olsa67b394f2009-10-23 19:36:18 -04001807 strcat(debugfs, "/tracing/");
GeunSik Lim156f5a72009-06-02 15:01:37 +09001808 debugfs_found = 1;
1809
1810 return debugfs;
1811}
1812
1813const char *tracing_file(const char *file_name)
1814{
1815 static char trace_file[MAX_PATH+1];
1816 snprintf(trace_file, MAX_PATH, "%s/%s", find_debugfs(), file_name);
1817 return trace_file;
1818}
1819
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001820int main (int argc, char **argv)
1821{
1822 if (argc < 1)
1823 exit(-1);
1824
1825 if (fork() > 0) {
1826 int fd, ffd;
1827 char line[64];
1828 int s;
1829
GeunSik Lim156f5a72009-06-02 15:01:37 +09001830 ffd = open(tracing_file("current_tracer"), O_WRONLY);
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001831 if (ffd < 0)
1832 exit(-1);
1833 write(ffd, "nop", 3);
1834
GeunSik Lim156f5a72009-06-02 15:01:37 +09001835 fd = open(tracing_file("set_ftrace_pid"), O_WRONLY);
Steven Rostedtdf4fc312008-11-26 00:16:23 -05001836 s = sprintf(line, "%d\n", getpid());
1837 write(fd, line, s);
1838
1839 write(ffd, "function", 8);
1840
1841 close(fd);
1842 close(ffd);
1843
1844 execvp(argv[1], argv+1);
1845 }
1846
1847 return 0;
1848}
1849
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001850Or this simple script!
Markus Metzgere2ea5392009-01-19 10:35:58 +01001851
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001852------
1853#!/bin/bash
Markus Metzgere2ea5392009-01-19 10:35:58 +01001854
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04001855debugfs=`sed -ne 's/^debugfs \(.*\) debugfs.*/\1/p' /proc/mounts`
1856echo nop > $debugfs/tracing/current_tracer
1857echo 0 > $debugfs/tracing/tracing_on
1858echo $$ > $debugfs/tracing/set_ftrace_pid
1859echo function > $debugfs/tracing/current_tracer
1860echo 1 > $debugfs/tracing/tracing_on
1861exec "$@"
1862------
Markus Metzgere2ea5392009-01-19 10:35:58 +01001863
1864
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001865function graph tracer
1866---------------------------
1867
Ingo Molnar57526742009-02-19 12:54:10 +01001868This tracer is similar to the function tracer except that it
1869probes a function on its entry and its exit. This is done by
1870using a dynamically allocated stack of return addresses in each
1871task_struct. On function entry the tracer overwrites the return
1872address of each function traced to set a custom probe. Thus the
1873original return address is stored on the stack of return address
1874in the task_struct.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001875
Ingo Molnar57526742009-02-19 12:54:10 +01001876Probing on both ends of a function leads to special features
1877such as:
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001878
Ingo Molnar57526742009-02-19 12:54:10 +01001879- measure of a function's time execution
1880- having a reliable call stack to draw function calls graph
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001881
1882This tracer is useful in several situations:
1883
Ingo Molnar57526742009-02-19 12:54:10 +01001884- you want to find the reason of a strange kernel behavior and
1885 need to see what happens in detail on any areas (or specific
1886 ones).
1887
1888- you are experiencing weird latencies but it's difficult to
1889 find its origin.
1890
1891- you want to find quickly which path is taken by a specific
1892 function
1893
1894- you just want to peek inside a working kernel and want to see
1895 what happens there.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001896
1897# tracer: function_graph
1898#
1899# CPU DURATION FUNCTION CALLS
1900# | | | | | | |
1901
1902 0) | sys_open() {
1903 0) | do_sys_open() {
1904 0) | getname() {
1905 0) | kmem_cache_alloc() {
1906 0) 1.382 us | __might_sleep();
1907 0) 2.478 us | }
1908 0) | strncpy_from_user() {
1909 0) | might_fault() {
1910 0) 1.389 us | __might_sleep();
1911 0) 2.553 us | }
1912 0) 3.807 us | }
1913 0) 7.876 us | }
1914 0) | alloc_fd() {
1915 0) 0.668 us | _spin_lock();
1916 0) 0.570 us | expand_files();
1917 0) 0.586 us | _spin_unlock();
1918
1919
Ingo Molnar57526742009-02-19 12:54:10 +01001920There are several columns that can be dynamically
1921enabled/disabled. You can use every combination of options you
1922want, depending on your needs.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001923
Ingo Molnar57526742009-02-19 12:54:10 +01001924- The cpu number on which the function executed is default
1925 enabled. It is sometimes better to only trace one cpu (see
1926 tracing_cpu_mask file) or you might sometimes see unordered
1927 function calls while cpu tracing switch.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001928
GeunSik Lim156f5a72009-06-02 15:01:37 +09001929 hide: echo nofuncgraph-cpu > trace_options
1930 show: echo funcgraph-cpu > trace_options
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001931
Ingo Molnar57526742009-02-19 12:54:10 +01001932- The duration (function's time of execution) is displayed on
1933 the closing bracket line of a function or on the same line
1934 than the current function in case of a leaf one. It is default
1935 enabled.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001936
GeunSik Lim156f5a72009-06-02 15:01:37 +09001937 hide: echo nofuncgraph-duration > trace_options
1938 show: echo funcgraph-duration > trace_options
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001939
Ingo Molnar57526742009-02-19 12:54:10 +01001940- The overhead field precedes the duration field in case of
1941 reached duration thresholds.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001942
GeunSik Lim156f5a72009-06-02 15:01:37 +09001943 hide: echo nofuncgraph-overhead > trace_options
1944 show: echo funcgraph-overhead > trace_options
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001945 depends on: funcgraph-duration
1946
1947 ie:
1948
Jungseok Leeb838e1d2015-07-11 14:51:40 +00001949 3) # 1837.709 us | } /* __switch_to */
1950 3) | finish_task_switch() {
1951 3) 0.313 us | _raw_spin_unlock_irq();
1952 3) 3.177 us | }
1953 3) # 1889.063 us | } /* __schedule */
1954 3) ! 140.417 us | } /* __schedule */
1955 3) # 2034.948 us | } /* schedule */
1956 3) * 33998.59 us | } /* schedule_preempt_disabled */
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001957
1958 [...]
1959
Jungseok Leeb838e1d2015-07-11 14:51:40 +00001960 1) 0.260 us | msecs_to_jiffies();
1961 1) 0.313 us | __rcu_read_unlock();
1962 1) + 61.770 us | }
1963 1) + 64.479 us | }
1964 1) 0.313 us | rcu_bh_qs();
1965 1) 0.313 us | __local_bh_enable();
1966 1) ! 217.240 us | }
1967 1) 0.365 us | idle_cpu();
1968 1) | rcu_irq_exit() {
1969 1) 0.417 us | rcu_eqs_enter_common.isra.47();
1970 1) 3.125 us | }
1971 1) ! 227.812 us | }
1972 1) ! 457.395 us | }
1973 1) @ 119760.2 us | }
1974
1975 [...]
1976
1977 2) | handle_IPI() {
1978 1) 6.979 us | }
1979 2) 0.417 us | scheduler_ipi();
1980 1) 9.791 us | }
1981 1) + 12.917 us | }
1982 2) 3.490 us | }
1983 1) + 15.729 us | }
1984 1) + 18.542 us | }
1985 2) $ 3594274 us | }
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001986
1987 + means that the function exceeded 10 usecs.
1988 ! means that the function exceeded 100 usecs.
Byungchul Park8e1e1df2014-11-24 09:34:19 +09001989 # means that the function exceeded 1000 usecs.
Jungseok Leeb838e1d2015-07-11 14:51:40 +00001990 * means that the function exceeded 10 msecs.
1991 @ means that the function exceeded 100 msecs.
Byungchul Park8e1e1df2014-11-24 09:34:19 +09001992 $ means that the function exceeded 1 sec.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001993
1994
Ingo Molnar57526742009-02-19 12:54:10 +01001995- The task/pid field displays the thread cmdline and pid which
1996 executed the function. It is default disabled.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01001997
GeunSik Lim156f5a72009-06-02 15:01:37 +09001998 hide: echo nofuncgraph-proc > trace_options
1999 show: echo funcgraph-proc > trace_options
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002000
2001 ie:
2002
2003 # tracer: function_graph
2004 #
2005 # CPU TASK/PID DURATION FUNCTION CALLS
2006 # | | | | | | | | |
2007 0) sh-4802 | | d_free() {
2008 0) sh-4802 | | call_rcu() {
2009 0) sh-4802 | | __call_rcu() {
2010 0) sh-4802 | 0.616 us | rcu_process_gp_end();
2011 0) sh-4802 | 0.586 us | check_for_new_grace_period();
2012 0) sh-4802 | 2.899 us | }
2013 0) sh-4802 | 4.040 us | }
2014 0) sh-4802 | 5.151 us | }
2015 0) sh-4802 | + 49.370 us | }
2016
2017
Ingo Molnar57526742009-02-19 12:54:10 +01002018- The absolute time field is an absolute timestamp given by the
2019 system clock since it started. A snapshot of this time is
2020 given on each entry/exit of functions
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002021
GeunSik Lim156f5a72009-06-02 15:01:37 +09002022 hide: echo nofuncgraph-abstime > trace_options
2023 show: echo funcgraph-abstime > trace_options
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002024
2025 ie:
2026
2027 #
2028 # TIME CPU DURATION FUNCTION CALLS
2029 # | | | | | | | |
2030 360.774522 | 1) 0.541 us | }
2031 360.774522 | 1) 4.663 us | }
2032 360.774523 | 1) 0.541 us | __wake_up_bit();
2033 360.774524 | 1) 6.796 us | }
2034 360.774524 | 1) 7.952 us | }
2035 360.774525 | 1) 9.063 us | }
2036 360.774525 | 1) 0.615 us | journal_mark_dirty();
2037 360.774527 | 1) 0.578 us | __brelse();
2038 360.774528 | 1) | reiserfs_prepare_for_journal() {
2039 360.774528 | 1) | unlock_buffer() {
2040 360.774529 | 1) | wake_up_bit() {
2041 360.774529 | 1) | bit_waitqueue() {
2042 360.774530 | 1) 0.594 us | __phys_addr();
2043
2044
Robert Elliott607e3a22014-05-20 17:10:51 -05002045The function name is always displayed after the closing bracket
2046for a function if the start of that function is not in the
2047trace buffer.
2048
2049Display of the function name after the closing bracket may be
2050enabled for functions whose start is in the trace buffer,
2051allowing easier searching with grep for function durations.
2052It is default disabled.
2053
2054 hide: echo nofuncgraph-tail > trace_options
2055 show: echo funcgraph-tail > trace_options
2056
2057 Example with nofuncgraph-tail (default):
2058 0) | putname() {
2059 0) | kmem_cache_free() {
2060 0) 0.518 us | __phys_addr();
2061 0) 1.757 us | }
2062 0) 2.861 us | }
2063
2064 Example with funcgraph-tail:
2065 0) | putname() {
2066 0) | kmem_cache_free() {
2067 0) 0.518 us | __phys_addr();
2068 0) 1.757 us | } /* kmem_cache_free() */
2069 0) 2.861 us | } /* putname() */
2070
Ingo Molnar57526742009-02-19 12:54:10 +01002071You can put some comments on specific functions by using
Ingo Molnar5e1607a2009-03-05 10:24:48 +01002072trace_printk() For example, if you want to put a comment inside
Ingo Molnar57526742009-02-19 12:54:10 +01002073the __might_sleep() function, you just have to include
Ingo Molnar5e1607a2009-03-05 10:24:48 +01002074<linux/ftrace.h> and call trace_printk() inside __might_sleep()
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002075
Ingo Molnar5e1607a2009-03-05 10:24:48 +01002076trace_printk("I'm a comment!\n")
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002077
2078will produce:
2079
2080 1) | __might_sleep() {
2081 1) | /* I'm a comment! */
2082 1) 1.449 us | }
2083
2084
Ingo Molnar57526742009-02-19 12:54:10 +01002085You might find other useful features for this tracer in the
2086following "dynamic ftrace" section such as tracing only specific
2087functions or tasks.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002088
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002089dynamic ftrace
2090--------------
2091
Steven Rostedtf2d9c742008-07-15 10:57:33 -04002092If CONFIG_DYNAMIC_FTRACE is set, the system will run with
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002093virtually no overhead when function tracing is disabled. The way
2094this works is the mcount function call (placed at the start of
Ingo Molnar57526742009-02-19 12:54:10 +01002095every kernel function, produced by the -pg switch in gcc),
2096starts of pointing to a simple return. (Enabling FTRACE will
2097include the -pg switch in the compiling of the kernel.)
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002098
Steven Rostedt9b803c02008-11-03 15:15:08 -05002099At compile time every C file object is run through the
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002100recordmcount program (located in the scripts directory). This
2101program will parse the ELF headers in the C object to find all
2102the locations in the .text section that call mcount. (Note, only
2103white listed .text sections are processed, since processing other
2104sections like .init.text may cause races due to those sections
2105being freed unexpectedly).
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002106
Ingo Molnar57526742009-02-19 12:54:10 +01002107A new section called "__mcount_loc" is created that holds
2108references to all the mcount call sites in the .text section.
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002109The recordmcount program re-links this section back into the
2110original object. The final linking stage of the kernel will add all these
2111references into a single table.
Steven Rostedt9b803c02008-11-03 15:15:08 -05002112
2113On boot up, before SMP is initialized, the dynamic ftrace code
Ingo Molnar57526742009-02-19 12:54:10 +01002114scans this table and updates all the locations into nops. It
2115also records the locations, which are added to the
2116available_filter_functions list. Modules are processed as they
2117are loaded and before they are executed. When a module is
2118unloaded, it also removes its functions from the ftrace function
2119list. This is automatic in the module unload code, and the
2120module author does not need to worry about it.
Steven Rostedt9b803c02008-11-03 15:15:08 -05002121
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002122When tracing is enabled, the process of modifying the function
2123tracepoints is dependent on architecture. The old method is to use
2124kstop_machine to prevent races with the CPUs executing code being
2125modified (which can cause the CPU to do undesirable things, especially
2126if the modified code crosses cache (or page) boundaries), and the nops are
Ingo Molnar57526742009-02-19 12:54:10 +01002127patched back to calls. But this time, they do not call mcount
2128(which is just a function stub). They now call into the ftrace
2129infrastructure.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002130
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002131The new method of modifying the function tracepoints is to place
2132a breakpoint at the location to be modified, sync all CPUs, modify
2133the rest of the instruction not covered by the breakpoint. Sync
2134all CPUs again, and then remove the breakpoint with the finished
2135version to the ftrace call site.
2136
2137Some archs do not even need to monkey around with the synchronization,
2138and can just slap the new code on top of the old without any
2139problems with other CPUs executing it at the same time.
2140
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002141One special side-effect to the recording of the functions being
Steven Rostedtf2d9c742008-07-15 10:57:33 -04002142traced is that we can now selectively choose which functions we
Ingo Molnar57526742009-02-19 12:54:10 +01002143wish to trace and which ones we want the mcount calls to remain
2144as nops.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002145
Ingo Molnar57526742009-02-19 12:54:10 +01002146Two files are used, one for enabling and one for disabling the
2147tracing of specified functions. They are:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002148
2149 set_ftrace_filter
2150
2151and
2152
2153 set_ftrace_notrace
2154
Ingo Molnar57526742009-02-19 12:54:10 +01002155A list of available functions that you can add to these files is
2156listed in:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002157
2158 available_filter_functions
2159
GeunSik Lim156f5a72009-06-02 15:01:37 +09002160 # cat available_filter_functions
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002161put_prev_task_idle
2162kmem_cache_create
2163pick_next_task_rt
2164get_online_cpus
2165pick_next_task_fair
2166mutex_lock
2167[...]
2168
Steven Rostedtf2d9c742008-07-15 10:57:33 -04002169If I am only interested in sys_nanosleep and hrtimer_interrupt:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002170
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002171 # echo sys_nanosleep hrtimer_interrupt > set_ftrace_filter
Yang Hongyang6993b1b2010-01-25 11:10:32 +08002172 # echo function > current_tracer
Steven Rostedt6752ab42011-02-08 13:54:06 -05002173 # echo 1 > tracing_on
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002174 # usleep 1
Steven Rostedt6752ab42011-02-08 13:54:06 -05002175 # echo 0 > tracing_on
GeunSik Lim156f5a72009-06-02 15:01:37 +09002176 # cat trace
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002177# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002178#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002179# entries-in-buffer/entries-written: 5/5 #P:4
2180#
2181# _-----=> irqs-off
2182# / _----=> need-resched
2183# | / _---=> hardirq/softirq
2184# || / _--=> preempt-depth
2185# ||| / delay
2186# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2187# | | | |||| | |
2188 usleep-2665 [001] .... 4186.475355: sys_nanosleep <-system_call_fastpath
2189 <idle>-0 [001] d.h1 4186.475409: hrtimer_interrupt <-smp_apic_timer_interrupt
2190 usleep-2665 [001] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
2191 <idle>-0 [003] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt
2192 <idle>-0 [002] d.h1 4186.475427: hrtimer_interrupt <-smp_apic_timer_interrupt
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002193
Steven Rostedtf2d9c742008-07-15 10:57:33 -04002194To see which functions are being traced, you can cat the file:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002195
GeunSik Lim156f5a72009-06-02 15:01:37 +09002196 # cat set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002197hrtimer_interrupt
2198sys_nanosleep
2199
2200
Ingo Molnar57526742009-02-19 12:54:10 +01002201Perhaps this is not enough. The filters also allow simple wild
2202cards. Only the following are currently available
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002203
Steven Rostedta41eeba2008-07-14 16:41:12 -04002204 <match>* - will match functions that begin with <match>
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002205 *<match> - will match functions that end with <match>
2206 *<match>* - will match functions that have <match> in it
2207
Steven Rostedtf2d9c742008-07-15 10:57:33 -04002208These are the only wild cards which are supported.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002209
2210 <match>*<match> will not work.
2211
Ingo Molnar57526742009-02-19 12:54:10 +01002212Note: It is better to use quotes to enclose the wild cards,
2213 otherwise the shell may expand the parameters into names
2214 of files in the local directory.
walimisc072c242008-11-28 12:21:19 +08002215
GeunSik Lim156f5a72009-06-02 15:01:37 +09002216 # echo 'hrtimer_*' > set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002217
2218Produces:
2219
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002220# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002221#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002222# entries-in-buffer/entries-written: 897/897 #P:4
2223#
2224# _-----=> irqs-off
2225# / _----=> need-resched
2226# | / _---=> hardirq/softirq
2227# || / _--=> preempt-depth
2228# ||| / delay
2229# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2230# | | | |||| | |
2231 <idle>-0 [003] dN.1 4228.547803: hrtimer_cancel <-tick_nohz_idle_exit
2232 <idle>-0 [003] dN.1 4228.547804: hrtimer_try_to_cancel <-hrtimer_cancel
2233 <idle>-0 [003] dN.2 4228.547805: hrtimer_force_reprogram <-__remove_hrtimer
2234 <idle>-0 [003] dN.1 4228.547805: hrtimer_forward <-tick_nohz_idle_exit
2235 <idle>-0 [003] dN.1 4228.547805: hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11
2236 <idle>-0 [003] d..1 4228.547858: hrtimer_get_next_event <-get_next_timer_interrupt
2237 <idle>-0 [003] d..1 4228.547859: hrtimer_start <-__tick_nohz_idle_enter
2238 <idle>-0 [003] d..2 4228.547860: hrtimer_force_reprogram <-__rem
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002239
2240Notice that we lost the sys_nanosleep.
2241
GeunSik Lim156f5a72009-06-02 15:01:37 +09002242 # cat set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002243hrtimer_run_queues
2244hrtimer_run_pending
2245hrtimer_init
2246hrtimer_cancel
2247hrtimer_try_to_cancel
2248hrtimer_forward
2249hrtimer_start
2250hrtimer_reprogram
2251hrtimer_force_reprogram
2252hrtimer_get_next_event
2253hrtimer_interrupt
2254hrtimer_nanosleep
2255hrtimer_wakeup
2256hrtimer_get_remaining
2257hrtimer_get_res
2258hrtimer_init_sleeper
2259
2260
2261This is because the '>' and '>>' act just like they do in bash.
2262To rewrite the filters, use '>'
2263To append to the filters, use '>>'
2264
Ingo Molnar57526742009-02-19 12:54:10 +01002265To clear out a filter so that all functions will be recorded
2266again:
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002267
GeunSik Lim156f5a72009-06-02 15:01:37 +09002268 # echo > set_ftrace_filter
2269 # cat set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002270 #
2271
2272Again, now we want to append.
2273
GeunSik Lim156f5a72009-06-02 15:01:37 +09002274 # echo sys_nanosleep > set_ftrace_filter
2275 # cat set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002276sys_nanosleep
GeunSik Lim156f5a72009-06-02 15:01:37 +09002277 # echo 'hrtimer_*' >> set_ftrace_filter
2278 # cat set_ftrace_filter
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002279hrtimer_run_queues
2280hrtimer_run_pending
2281hrtimer_init
2282hrtimer_cancel
2283hrtimer_try_to_cancel
2284hrtimer_forward
2285hrtimer_start
2286hrtimer_reprogram
2287hrtimer_force_reprogram
2288hrtimer_get_next_event
2289hrtimer_interrupt
2290sys_nanosleep
2291hrtimer_nanosleep
2292hrtimer_wakeup
2293hrtimer_get_remaining
2294hrtimer_get_res
2295hrtimer_init_sleeper
2296
2297
Ingo Molnar57526742009-02-19 12:54:10 +01002298The set_ftrace_notrace prevents those functions from being
2299traced.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002300
GeunSik Lim156f5a72009-06-02 15:01:37 +09002301 # echo '*preempt*' '*lock*' > set_ftrace_notrace
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002302
2303Produces:
2304
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002305# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002306#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002307# entries-in-buffer/entries-written: 39608/39608 #P:4
2308#
2309# _-----=> irqs-off
2310# / _----=> need-resched
2311# | / _---=> hardirq/softirq
2312# || / _--=> preempt-depth
2313# ||| / delay
2314# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2315# | | | |||| | |
2316 bash-1994 [000] .... 4342.324896: file_ra_state_init <-do_dentry_open
2317 bash-1994 [000] .... 4342.324897: open_check_o_direct <-do_last
2318 bash-1994 [000] .... 4342.324897: ima_file_check <-do_last
2319 bash-1994 [000] .... 4342.324898: process_measurement <-ima_file_check
2320 bash-1994 [000] .... 4342.324898: ima_get_action <-process_measurement
2321 bash-1994 [000] .... 4342.324898: ima_match_policy <-ima_get_action
2322 bash-1994 [000] .... 4342.324899: do_truncate <-do_last
2323 bash-1994 [000] .... 4342.324899: should_remove_suid <-do_truncate
2324 bash-1994 [000] .... 4342.324899: notify_change <-do_truncate
2325 bash-1994 [000] .... 4342.324900: current_fs_time <-notify_change
2326 bash-1994 [000] .... 4342.324900: current_kernel_time <-current_fs_time
2327 bash-1994 [000] .... 4342.324900: timespec_trunc <-current_fs_time
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002328
2329We can see that there's no more lock or preempt tracing.
2330
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002331
Ingo Molnar57526742009-02-19 12:54:10 +01002332Dynamic ftrace with the function graph tracer
2333---------------------------------------------
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002334
Ingo Molnar57526742009-02-19 12:54:10 +01002335Although what has been explained above concerns both the
2336function tracer and the function-graph-tracer, there are some
2337special features only available in the function-graph tracer.
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002338
Ingo Molnar57526742009-02-19 12:54:10 +01002339If you want to trace only one function and all of its children,
2340you just have to echo its name into set_graph_function:
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002341
Ingo Molnar57526742009-02-19 12:54:10 +01002342 echo __do_fault > set_graph_function
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002343
Ingo Molnar57526742009-02-19 12:54:10 +01002344will produce the following "expanded" trace of the __do_fault()
2345function:
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002346
2347 0) | __do_fault() {
2348 0) | filemap_fault() {
2349 0) | find_lock_page() {
2350 0) 0.804 us | find_get_page();
2351 0) | __might_sleep() {
2352 0) 1.329 us | }
2353 0) 3.904 us | }
2354 0) 4.979 us | }
2355 0) 0.653 us | _spin_lock();
2356 0) 0.578 us | page_add_file_rmap();
2357 0) 0.525 us | native_set_pte_at();
2358 0) 0.585 us | _spin_unlock();
2359 0) | unlock_page() {
2360 0) 0.541 us | page_waitqueue();
2361 0) 0.639 us | __wake_up_bit();
2362 0) 2.786 us | }
2363 0) + 14.237 us | }
2364 0) | __do_fault() {
2365 0) | filemap_fault() {
2366 0) | find_lock_page() {
2367 0) 0.698 us | find_get_page();
2368 0) | __might_sleep() {
2369 0) 1.412 us | }
2370 0) 3.950 us | }
2371 0) 5.098 us | }
2372 0) 0.631 us | _spin_lock();
2373 0) 0.571 us | page_add_file_rmap();
2374 0) 0.526 us | native_set_pte_at();
2375 0) 0.586 us | _spin_unlock();
2376 0) | unlock_page() {
2377 0) 0.533 us | page_waitqueue();
2378 0) 0.638 us | __wake_up_bit();
2379 0) 2.793 us | }
2380 0) + 14.012 us | }
2381
Ingo Molnar57526742009-02-19 12:54:10 +01002382You can also expand several functions at once:
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002383
Ingo Molnar57526742009-02-19 12:54:10 +01002384 echo sys_open > set_graph_function
2385 echo sys_close >> set_graph_function
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002386
Ingo Molnar57526742009-02-19 12:54:10 +01002387Now if you want to go back to trace all functions you can clear
2388this special filter via:
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002389
Ingo Molnar57526742009-02-19 12:54:10 +01002390 echo > set_graph_function
Frederic Weisbecker985ec202009-02-18 06:35:34 +01002391
2392
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002393ftrace_enabled
2394--------------
2395
2396Note, the proc sysctl ftrace_enable is a big on/off switch for the
2397function tracer. By default it is enabled (when function tracing is
2398enabled in the kernel). If it is disabled, all function tracing is
2399disabled. This includes not only the function tracers for ftrace, but
2400also for any other uses (perf, kprobes, stack tracing, profiling, etc).
2401
2402Please disable this with care.
2403
2404This can be disable (and enabled) with:
2405
2406 sysctl kernel.ftrace_enabled=0
2407 sysctl kernel.ftrace_enabled=1
2408
2409 or
2410
2411 echo 0 > /proc/sys/kernel/ftrace_enabled
2412 echo 1 > /proc/sys/kernel/ftrace_enabled
2413
2414
Chase Douglas07271aa2010-04-23 14:02:39 -04002415Filter commands
2416---------------
2417
2418A few commands are supported by the set_ftrace_filter interface.
2419Trace commands have the following format:
2420
2421<function>:<command>:<parameter>
2422
2423The following commands are supported:
2424
2425- mod
2426 This command enables function filtering per module. The
2427 parameter defines the module. For example, if only the write*
2428 functions in the ext3 module are desired, run:
2429
2430 echo 'write*:mod:ext3' > set_ftrace_filter
2431
2432 This command interacts with the filter in the same way as
2433 filtering based on function names. Thus, adding more functions
2434 in a different module is accomplished by appending (>>) to the
2435 filter file. Remove specific module functions by prepending
2436 '!':
2437
2438 echo '!writeback*:mod:ext3' >> set_ftrace_filter
2439
2440- traceon/traceoff
2441 These commands turn tracing on and off when the specified
2442 functions are hit. The parameter determines how many times the
2443 tracing system is turned on and off. If unspecified, there is
2444 no limit. For example, to disable tracing when a schedule bug
2445 is hit the first 5 times, run:
2446
2447 echo '__schedule_bug:traceoff:5' > set_ftrace_filter
2448
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002449 To always disable tracing when __schedule_bug is hit:
2450
2451 echo '__schedule_bug:traceoff' > set_ftrace_filter
2452
Chase Douglas07271aa2010-04-23 14:02:39 -04002453 These commands are cumulative whether or not they are appended
2454 to set_ftrace_filter. To remove a command, prepend it by '!'
2455 and drop the parameter:
2456
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002457 echo '!__schedule_bug:traceoff:0' > set_ftrace_filter
2458
2459 The above removes the traceoff command for __schedule_bug
2460 that have a counter. To remove commands without counters:
2461
Chase Douglas07271aa2010-04-23 14:02:39 -04002462 echo '!__schedule_bug:traceoff' > set_ftrace_filter
2463
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002464- snapshot
2465 Will cause a snapshot to be triggered when the function is hit.
2466
2467 echo 'native_flush_tlb_others:snapshot' > set_ftrace_filter
2468
2469 To only snapshot once:
2470
2471 echo 'native_flush_tlb_others:snapshot:1' > set_ftrace_filter
2472
2473 To remove the above commands:
2474
2475 echo '!native_flush_tlb_others:snapshot' > set_ftrace_filter
2476 echo '!native_flush_tlb_others:snapshot:0' > set_ftrace_filter
2477
2478- enable_event/disable_event
2479 These commands can enable or disable a trace event. Note, because
2480 function tracing callbacks are very sensitive, when these commands
2481 are registered, the trace point is activated, but disabled in
2482 a "soft" mode. That is, the tracepoint will be called, but
2483 just will not be traced. The event tracepoint stays in this mode
2484 as long as there's a command that triggers it.
2485
2486 echo 'try_to_wake_up:enable_event:sched:sched_switch:2' > \
2487 set_ftrace_filter
2488
2489 The format is:
2490
2491 <function>:enable_event:<system>:<event>[:count]
2492 <function>:disable_event:<system>:<event>[:count]
2493
2494 To remove the events commands:
2495
2496
2497 echo '!try_to_wake_up:enable_event:sched:sched_switch:0' > \
2498 set_ftrace_filter
2499 echo '!schedule:disable_event:sched:sched_switch' > \
2500 set_ftrace_filter
Chase Douglas07271aa2010-04-23 14:02:39 -04002501
Steven Rostedt (Red Hat)ad71d882013-04-30 15:46:14 -04002502- dump
2503 When the function is hit, it will dump the contents of the ftrace
2504 ring buffer to the console. This is useful if you need to debug
2505 something, and want to dump the trace when a certain function
2506 is hit. Perhaps its a function that is called before a tripple
2507 fault happens and does not allow you to get a regular dump.
2508
Steven Rostedt (Red Hat)90e3c032013-04-30 19:00:46 -04002509- cpudump
2510 When the function is hit, it will dump the contents of the ftrace
2511 ring buffer for the current CPU to the console. Unlike the "dump"
2512 command, it only prints out the contents of the ring buffer for the
2513 CPU that executed the function that triggered the dump.
2514
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002515trace_pipe
2516----------
2517
Ingo Molnar57526742009-02-19 12:54:10 +01002518The trace_pipe outputs the same content as the trace file, but
2519the effect on the tracing is different. Every read from
2520trace_pipe is consumed. This means that subsequent reads will be
2521different. The trace is live.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002522
GeunSik Lim156f5a72009-06-02 15:01:37 +09002523 # echo function > current_tracer
2524 # cat trace_pipe > /tmp/trace.out &
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002525[1] 4153
Steven Rostedt6752ab42011-02-08 13:54:06 -05002526 # echo 1 > tracing_on
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002527 # usleep 1
Steven Rostedt6752ab42011-02-08 13:54:06 -05002528 # echo 0 > tracing_on
GeunSik Lim156f5a72009-06-02 15:01:37 +09002529 # cat trace
Steven Rostedt9b803c02008-11-03 15:15:08 -05002530# tracer: function
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002531#
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002532# entries-in-buffer/entries-written: 0/0 #P:4
2533#
2534# _-----=> irqs-off
2535# / _----=> need-resched
2536# | / _---=> hardirq/softirq
2537# || / _--=> preempt-depth
2538# ||| / delay
2539# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2540# | | | |||| | |
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002541
2542 #
2543 # cat /tmp/trace.out
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002544 bash-1994 [000] .... 5281.568961: mutex_unlock <-rb_simple_write
2545 bash-1994 [000] .... 5281.568963: __mutex_unlock_slowpath <-mutex_unlock
2546 bash-1994 [000] .... 5281.568963: __fsnotify_parent <-fsnotify_modify
2547 bash-1994 [000] .... 5281.568964: fsnotify <-fsnotify_modify
2548 bash-1994 [000] .... 5281.568964: __srcu_read_lock <-fsnotify
2549 bash-1994 [000] .... 5281.568964: add_preempt_count <-__srcu_read_lock
2550 bash-1994 [000] ...1 5281.568965: sub_preempt_count <-__srcu_read_lock
2551 bash-1994 [000] .... 5281.568965: __srcu_read_unlock <-fsnotify
2552 bash-1994 [000] .... 5281.568967: sys_dup2 <-system_call_fastpath
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002553
2554
Ingo Molnar57526742009-02-19 12:54:10 +01002555Note, reading the trace_pipe file will block until more input is
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002556added.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002557
2558trace entries
2559-------------
2560
Ingo Molnar57526742009-02-19 12:54:10 +01002561Having too much or not enough data can be troublesome in
2562diagnosing an issue in the kernel. The file buffer_size_kb is
2563used to modify the size of the internal trace buffers. The
2564number listed is the number of entries that can be recorded per
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002565CPU. To know the full size, multiply the number of possible CPUs
Ingo Molnar57526742009-02-19 12:54:10 +01002566with the number of entries.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002567
GeunSik Lim156f5a72009-06-02 15:01:37 +09002568 # cat buffer_size_kb
Steven Rostedt1696b2b2008-11-13 00:09:35 -050025691408 (units kilobytes)
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002570
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002571Or simply read buffer_total_size_kb
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002572
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002573 # cat buffer_total_size_kb
25745632
2575
2576To modify the buffer, simple echo in a number (in 1024 byte segments).
2577
GeunSik Lim156f5a72009-06-02 15:01:37 +09002578 # echo 10000 > buffer_size_kb
2579 # cat buffer_size_kb
Steven Rostedt1696b2b2008-11-13 00:09:35 -0500258010000 (units kilobytes)
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002581
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002582It will try to allocate as much as possible. If you allocate too
2583much, it can cause Out-Of-Memory to trigger.
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002584
GeunSik Lim156f5a72009-06-02 15:01:37 +09002585 # echo 1000000000000 > buffer_size_kb
Steven Rostedteb6d42e2008-07-10 12:46:01 -04002586-bash: echo: write error: Cannot allocate memory
GeunSik Lim156f5a72009-06-02 15:01:37 +09002587 # cat buffer_size_kb
Steven Rostedteb6d42e2008-07-10 12:46:01 -0400258885
2589
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002590The per_cpu buffers can be changed individually as well:
2591
2592 # echo 10000 > per_cpu/cpu0/buffer_size_kb
2593 # echo 100 > per_cpu/cpu1/buffer_size_kb
2594
2595When the per_cpu buffers are not the same, the buffer_size_kb
2596at the top level will just show an X
2597
2598 # cat buffer_size_kb
2599X
2600
2601This is where the buffer_total_size_kb is useful:
2602
2603 # cat buffer_total_size_kb
260412916
2605
2606Writing to the top level buffer_size_kb will reset all the buffers
2607to be the same again.
2608
Hiraku Toyookac1043fc2012-12-26 11:53:09 +09002609Snapshot
2610--------
2611CONFIG_TRACER_SNAPSHOT makes a generic snapshot feature
2612available to all non latency tracers. (Latency tracers which
2613record max latency, such as "irqsoff" or "wakeup", can't use
2614this feature, since those are already using the snapshot
2615mechanism internally.)
2616
2617Snapshot preserves a current trace buffer at a particular point
2618in time without stopping tracing. Ftrace swaps the current
2619buffer with a spare buffer, and tracing continues in the new
2620current (=previous spare) buffer.
2621
2622The following debugfs files in "tracing" are related to this
2623feature:
2624
2625 snapshot:
2626
2627 This is used to take a snapshot and to read the output
2628 of the snapshot. Echo 1 into this file to allocate a
2629 spare buffer and to take a snapshot (swap), then read
2630 the snapshot from this file in the same format as
2631 "trace" (described above in the section "The File
2632 System"). Both reads snapshot and tracing are executable
2633 in parallel. When the spare buffer is allocated, echoing
2634 0 frees it, and echoing else (positive) values clear the
2635 snapshot contents.
2636 More details are shown in the table below.
2637
2638 status\input | 0 | 1 | else |
2639 --------------+------------+------------+------------+
Hiraku Toyooka1abccd72013-03-08 16:32:25 +09002640 not allocated |(do nothing)| alloc+swap |(do nothing)|
Hiraku Toyookac1043fc2012-12-26 11:53:09 +09002641 --------------+------------+------------+------------+
2642 allocated | free | swap | clear |
2643 --------------+------------+------------+------------+
2644
2645Here is an example of using the snapshot feature.
2646
2647 # echo 1 > events/sched/enable
2648 # echo 1 > snapshot
2649 # cat snapshot
2650# tracer: nop
2651#
2652# entries-in-buffer/entries-written: 71/71 #P:8
2653#
2654# _-----=> irqs-off
2655# / _----=> need-resched
2656# | / _---=> hardirq/softirq
2657# || / _--=> preempt-depth
2658# ||| / delay
2659# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2660# | | | |||| | |
2661 <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
2662 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
2663[...]
2664 <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
2665
2666 # cat trace
2667# tracer: nop
2668#
2669# entries-in-buffer/entries-written: 77/77 #P:8
2670#
2671# _-----=> irqs-off
2672# / _----=> need-resched
2673# | / _---=> hardirq/softirq
2674# || / _--=> preempt-depth
2675# ||| / delay
2676# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2677# | | | |||| | |
2678 <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
2679 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
2680[...]
2681
2682
2683If you try to use this snapshot feature when current tracer is
2684one of the latency tracers, you will get the following results.
2685
2686 # echo wakeup > current_tracer
2687 # echo 1 > snapshot
2688bash: echo: write error: Device or resource busy
2689 # cat snapshot
2690cat: snapshot: Device or resource busy
2691
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002692
2693Instances
2694---------
2695In the debugfs tracing directory is a directory called "instances".
2696This directory can have new directories created inside of it using
2697mkdir, and removing directories with rmdir. The directory created
2698with mkdir in this directory will already contain files and other
2699directories after it is created.
2700
2701 # mkdir instances/foo
2702 # ls instances/foo
2703buffer_size_kb buffer_total_size_kb events free_buffer per_cpu
2704set_event snapshot trace trace_clock trace_marker trace_options
2705trace_pipe tracing_on
2706
2707As you can see, the new directory looks similar to the tracing directory
2708itself. In fact, it is very similar, except that the buffer and
2709events are agnostic from the main director, or from any other
2710instances that are created.
2711
2712The files in the new directory work just like the files with the
2713same name in the tracing directory except the buffer that is used
2714is a separate and new buffer. The files affect that buffer but do not
2715affect the main buffer with the exception of trace_options. Currently,
2716the trace_options affect all instances and the top level buffer
2717the same, but this may change in future releases. That is, options
2718may become specific to the instance they reside in.
2719
2720Notice that none of the function tracer files are there, nor is
2721current_tracer and available_tracers. This is because the buffers
2722can currently only have events enabled for them.
2723
2724 # mkdir instances/foo
2725 # mkdir instances/bar
2726 # mkdir instances/zoot
2727 # echo 100000 > buffer_size_kb
2728 # echo 1000 > instances/foo/buffer_size_kb
2729 # echo 5000 > instances/bar/per_cpu/cpu1/buffer_size_kb
2730 # echo function > current_trace
2731 # echo 1 > instances/foo/events/sched/sched_wakeup/enable
2732 # echo 1 > instances/foo/events/sched/sched_wakeup_new/enable
2733 # echo 1 > instances/foo/events/sched/sched_switch/enable
2734 # echo 1 > instances/bar/events/irq/enable
2735 # echo 1 > instances/zoot/events/syscalls/enable
2736 # cat trace_pipe
2737CPU:2 [LOST 11745 EVENTS]
2738 bash-2044 [002] .... 10594.481032: _raw_spin_lock_irqsave <-get_page_from_freelist
2739 bash-2044 [002] d... 10594.481032: add_preempt_count <-_raw_spin_lock_irqsave
2740 bash-2044 [002] d..1 10594.481032: __rmqueue <-get_page_from_freelist
2741 bash-2044 [002] d..1 10594.481033: _raw_spin_unlock <-get_page_from_freelist
2742 bash-2044 [002] d..1 10594.481033: sub_preempt_count <-_raw_spin_unlock
2743 bash-2044 [002] d... 10594.481033: get_pageblock_flags_group <-get_pageblock_migratetype
2744 bash-2044 [002] d... 10594.481034: __mod_zone_page_state <-get_page_from_freelist
2745 bash-2044 [002] d... 10594.481034: zone_statistics <-get_page_from_freelist
2746 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
2747 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics
2748 bash-2044 [002] .... 10594.481035: arch_dup_task_struct <-copy_process
2749[...]
2750
2751 # cat instances/foo/trace_pipe
2752 bash-1998 [000] d..4 136.676759: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000
2753 bash-1998 [000] dN.4 136.676760: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000
2754 <idle>-0 [003] d.h3 136.676906: sched_wakeup: comm=rcu_preempt pid=9 prio=120 success=1 target_cpu=003
2755 <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
2756 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
2757 bash-1998 [000] d..4 136.677014: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000
2758 bash-1998 [000] dN.4 136.677016: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000
2759 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
2760 kworker/0:1-59 [000] d..4 136.677022: sched_wakeup: comm=sshd pid=1995 prio=120 success=1 target_cpu=001
2761 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
2762[...]
2763
2764 # cat instances/bar/trace_pipe
2765 migration/1-14 [001] d.h3 138.732674: softirq_raise: vec=3 [action=NET_RX]
2766 <idle>-0 [001] dNh3 138.732725: softirq_raise: vec=3 [action=NET_RX]
2767 bash-1998 [000] d.h1 138.733101: softirq_raise: vec=1 [action=TIMER]
2768 bash-1998 [000] d.h1 138.733102: softirq_raise: vec=9 [action=RCU]
2769 bash-1998 [000] ..s2 138.733105: softirq_entry: vec=1 [action=TIMER]
2770 bash-1998 [000] ..s2 138.733106: softirq_exit: vec=1 [action=TIMER]
2771 bash-1998 [000] ..s2 138.733106: softirq_entry: vec=9 [action=RCU]
2772 bash-1998 [000] ..s2 138.733109: softirq_exit: vec=9 [action=RCU]
2773 sshd-1995 [001] d.h1 138.733278: irq_handler_entry: irq=21 name=uhci_hcd:usb4
2774 sshd-1995 [001] d.h1 138.733280: irq_handler_exit: irq=21 ret=unhandled
2775 sshd-1995 [001] d.h1 138.733281: irq_handler_entry: irq=21 name=eth0
2776 sshd-1995 [001] d.h1 138.733283: irq_handler_exit: irq=21 ret=handled
2777[...]
2778
2779 # cat instances/zoot/trace
2780# tracer: nop
2781#
2782# entries-in-buffer/entries-written: 18996/18996 #P:4
2783#
2784# _-----=> irqs-off
2785# / _----=> need-resched
2786# | / _---=> hardirq/softirq
2787# || / _--=> preempt-depth
2788# ||| / delay
2789# TASK-PID CPU# |||| TIMESTAMP FUNCTION
2790# | | | |||| | |
2791 bash-1998 [000] d... 140.733501: sys_write -> 0x2
2792 bash-1998 [000] d... 140.733504: sys_dup2(oldfd: a, newfd: 1)
2793 bash-1998 [000] d... 140.733506: sys_dup2 -> 0x1
2794 bash-1998 [000] d... 140.733508: sys_fcntl(fd: a, cmd: 1, arg: 0)
2795 bash-1998 [000] d... 140.733509: sys_fcntl -> 0x1
2796 bash-1998 [000] d... 140.733510: sys_close(fd: a)
2797 bash-1998 [000] d... 140.733510: sys_close -> 0x0
2798 bash-1998 [000] d... 140.733514: sys_rt_sigprocmask(how: 0, nset: 0, oset: 6e2768, sigsetsize: 8)
2799 bash-1998 [000] d... 140.733515: sys_rt_sigprocmask -> 0x0
2800 bash-1998 [000] d... 140.733516: sys_rt_sigaction(sig: 2, act: 7fff718846f0, oact: 7fff71884650, sigsetsize: 8)
2801 bash-1998 [000] d... 140.733516: sys_rt_sigaction -> 0x0
2802
2803You can see that the trace of the top most trace buffer shows only
2804the function tracing. The foo instance displays wakeups and task
2805switches.
2806
2807To remove the instances, simply delete their directories:
2808
2809 # rmdir instances/foo
2810 # rmdir instances/bar
2811 # rmdir instances/zoot
2812
2813Note, if a process has a trace file open in one of the instance
2814directories, the rmdir will fail with EBUSY.
2815
2816
2817Stack trace
Ingo Molnar57526742009-02-19 12:54:10 +01002818-----------
Steven Rostedt (Red Hat)8d016092013-03-13 11:05:11 -04002819Since the kernel has a fixed sized stack, it is important not to
2820waste it in functions. A kernel developer must be conscience of
2821what they allocate on the stack. If they add too much, the system
2822can be in danger of a stack overflow, and corruption will occur,
2823usually leading to a system panic.
2824
2825There are some tools that check this, usually with interrupts
2826periodically checking usage. But if you can perform a check
2827at every function call that will become very useful. As ftrace provides
2828a function tracer, it makes it convenient to check the stack size
2829at every function call. This is enabled via the stack tracer.
2830
2831CONFIG_STACK_TRACER enables the ftrace stack tracing functionality.
2832To enable it, write a '1' into /proc/sys/kernel/stack_tracer_enabled.
2833
2834 # echo 1 > /proc/sys/kernel/stack_tracer_enabled
2835
2836You can also enable it from the kernel command line to trace
2837the stack size of the kernel during boot up, by adding "stacktrace"
2838to the kernel command line parameter.
2839
2840After running it for a few minutes, the output looks like:
2841
2842 # cat stack_max_size
28432928
2844
2845 # cat stack_trace
2846 Depth Size Location (18 entries)
2847 ----- ---- --------
2848 0) 2928 224 update_sd_lb_stats+0xbc/0x4ac
2849 1) 2704 160 find_busiest_group+0x31/0x1f1
2850 2) 2544 256 load_balance+0xd9/0x662
2851 3) 2288 80 idle_balance+0xbb/0x130
2852 4) 2208 128 __schedule+0x26e/0x5b9
2853 5) 2080 16 schedule+0x64/0x66
2854 6) 2064 128 schedule_timeout+0x34/0xe0
2855 7) 1936 112 wait_for_common+0x97/0xf1
2856 8) 1824 16 wait_for_completion+0x1d/0x1f
2857 9) 1808 128 flush_work+0xfe/0x119
2858 10) 1680 16 tty_flush_to_ldisc+0x1e/0x20
2859 11) 1664 48 input_available_p+0x1d/0x5c
2860 12) 1616 48 n_tty_poll+0x6d/0x134
2861 13) 1568 64 tty_poll+0x64/0x7f
2862 14) 1504 880 do_select+0x31e/0x511
2863 15) 624 400 core_sys_select+0x177/0x216
2864 16) 224 96 sys_select+0x91/0xb9
2865 17) 128 128 system_call_fastpath+0x16/0x1b
2866
2867Note, if -mfentry is being used by gcc, functions get traced before
2868they set up the stack frame. This means that leaf level functions
2869are not tested by the stack tracer when -mfentry is used.
2870
2871Currently, -mfentry is used by gcc 4.6.0 and above on x86 only.
2872
2873---------
Ingo Molnar57526742009-02-19 12:54:10 +01002874
2875More details can be found in the source code, in the
GeunSik Limbaf20b32009-06-01 10:49:41 +02002876kernel/trace/*.c files.