Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1 | ftrace - Function Tracer |
| 2 | ======================== |
| 3 | |
| 4 | Copyright 2008 Red Hat Inc. |
Steven Rostedt | a41eeba | 2008-07-14 16:41:12 -0400 | [diff] [blame] | 5 | Author: Steven Rostedt <srostedt@redhat.com> |
| 6 | License: The GNU Free Documentation License, Version 1.2 |
Steven Rostedt | a97762a | 2008-07-31 12:40:52 -0400 | [diff] [blame] | 7 | (dual licensed under the GPL v2) |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 8 | Reviewers: Elias Oltmanns, Randy Dunlap, Andrew Morton, |
| 9 | John Kacur, and David Teigland. |
Steven Rostedt | 42ec632 | 2008-11-03 15:18:56 -0500 | [diff] [blame] | 10 | Written for: 2.6.28-rc2 |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 11 | Updated for: 3.10 |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 12 | |
| 13 | Introduction |
| 14 | ------------ |
| 15 | |
| 16 | Ftrace is an internal tracer designed to help out developers and |
| 17 | designers of systems to find what is going on inside the kernel. |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 18 | It can be used for debugging or analyzing latencies and |
| 19 | performance issues that take place outside of user-space. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 20 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 21 | Although ftrace is typically considered the function tracer, it |
| 22 | is really a frame work of several assorted tracing utilities. |
| 23 | There's latency tracing to examine what occurs between interrupts |
| 24 | disabled and enabled, as well as for preemption and from a time |
| 25 | a task is woken to the task is actually scheduled in. |
| 26 | |
| 27 | One of the most common uses of ftrace is the event tracing. |
| 28 | Through out the kernel is hundreds of static event points that |
| 29 | can be enabled via the debugfs file system to see what is |
| 30 | going on in certain parts of the kernel. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 31 | |
| 32 | |
Mike Frysinger | 555f386 | 2009-09-14 20:10:15 -0400 | [diff] [blame] | 33 | Implementation Details |
| 34 | ---------------------- |
| 35 | |
| 36 | See ftrace-design.txt for details for arch porters and such. |
| 37 | |
| 38 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 39 | The File System |
| 40 | --------------- |
| 41 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 42 | Ftrace uses the debugfs file system to hold the control files as |
| 43 | well as the files to display output. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 44 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 45 | When debugfs is configured into the kernel (which selecting any ftrace |
| 46 | option will do) the directory /sys/kernel/debug will be created. To mount |
| 47 | this directory, you can add to your /etc/fstab file: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 48 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 49 | debugfs /sys/kernel/debug debugfs defaults 0 0 |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 50 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 51 | Or you can mount it at run time with: |
| 52 | |
| 53 | mount -t debugfs nodev /sys/kernel/debug |
| 54 | |
| 55 | For quicker access to that directory you may want to make a soft link to |
| 56 | it: |
| 57 | |
| 58 | ln -s /sys/kernel/debug /debug |
| 59 | |
| 60 | Any selected ftrace option will also create a directory called tracing |
| 61 | within the debugfs. The rest of the document will assume that you are in |
| 62 | the ftrace directory (cd /sys/kernel/debug/tracing) and will only concentrate |
| 63 | on the files within that directory and not distract from the content with |
| 64 | the extended "/sys/kernel/debug/tracing" path name. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 65 | |
| 66 | That's it! (assuming that you have ftrace configured into your kernel) |
| 67 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 68 | After mounting debugfs, you can see a directory called |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 69 | "tracing". This directory contains the control and output files |
| 70 | of ftrace. Here is a list of some of the key files: |
| 71 | |
| 72 | |
| 73 | Note: all time values are in microseconds. |
| 74 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 75 | current_tracer: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 76 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 77 | This is used to set or display the current tracer |
| 78 | that is configured. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 79 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 80 | available_tracers: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 81 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 82 | 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 Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 86 | |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 87 | tracing_on: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 88 | |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 89 | 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) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 91 | 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 Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 94 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 95 | trace: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 96 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 97 | This file holds the output of the trace in a human |
| 98 | readable format (described below). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 99 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 100 | trace_pipe: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 101 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 102 | The output is the same as the "trace" file but this |
| 103 | file is meant to be streamed with live tracing. |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 104 | 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 Vincent | 8547aa1 | 2015-05-04 19:48:54 +0200 | [diff] [blame] | 111 | adding more data, it will display the same |
| 112 | information every time it is read. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 113 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 114 | trace_options: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 115 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 116 | This file lets the user control the amount of data |
| 117 | that is displayed in one of the above output |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 118 | 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 Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 127 | |
KOSAKI Motohiro | 42b40b3 | 2009-03-07 23:55:09 +0900 | [diff] [blame] | 128 | tracing_max_latency: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 129 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 130 | 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 Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 133 | 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 Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 137 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 138 | 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 Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 145 | buffer_size_kb: |
| 146 | |
| 147 | This sets or displays the number of kilobytes each CPU |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 148 | buffer holds. By default, the trace buffers are the same size |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 149 | 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) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 157 | due to buffer management meta-data. ) |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 158 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 159 | 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 Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 175 | |
| 176 | tracing_cpumask: |
| 177 | |
| 178 | This is a mask that lets the user only trace |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 179 | on specified CPUs. The format is a hex string |
| 180 | representing the CPUs. |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 181 | |
| 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 Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 193 | This interface also allows for commands to be used. See the |
| 194 | "Filter commands" section for more details. |
| 195 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 196 | 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 | |
Steven Rostedt (Red Hat) | 627645f | 2015-11-03 16:37:15 -0500 | [diff] [blame] | 207 | set_event_pid: |
| 208 | |
| 209 | Have the events only trace a task with a PID listed in this file. |
| 210 | Note, sched_switch and sched_wake_up will also trace events |
| 211 | listed in this file. |
| 212 | |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 213 | To have the PIDs of children of tasks with their PID in this file |
| 214 | added on fork, enable the "event-fork" option. That option will also |
| 215 | cause the PIDs of tasks to be removed from this file when the task |
| 216 | exits. |
| 217 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 218 | set_graph_function: |
| 219 | |
| 220 | Set a "trigger" function where tracing should start |
| 221 | with the function graph tracer (See the section |
| 222 | "dynamic ftrace" for more details). |
| 223 | |
| 224 | available_filter_functions: |
| 225 | |
| 226 | This lists the functions that ftrace |
| 227 | has processed and can trace. These are the function |
| 228 | names that you can pass to "set_ftrace_filter" or |
| 229 | "set_ftrace_notrace". (See the section "dynamic ftrace" |
| 230 | below for more details.) |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 231 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 232 | enabled_functions: |
| 233 | |
| 234 | This file is more for debugging ftrace, but can also be useful |
| 235 | in seeing if any function has a callback attached to it. |
| 236 | Not only does the trace infrastructure use ftrace function |
| 237 | trace utility, but other subsystems might too. This file |
| 238 | displays all functions that have a callback attached to them |
| 239 | as well as the number of callbacks that have been attached. |
| 240 | Note, a callback may also call multiple functions which will |
| 241 | not be listed in this count. |
| 242 | |
| 243 | If the callback registered to be traced by a function with |
| 244 | the "save regs" attribute (thus even more overhead), a 'R' |
| 245 | will be displayed on the same line as the function that |
| 246 | is returning registers. |
| 247 | |
Masami Hiramatsu | f8b8be8 | 2014-11-21 05:25:16 -0500 | [diff] [blame] | 248 | If the callback registered to be traced by a function with |
| 249 | the "ip modify" attribute (thus the regs->ip can be changed), |
| 250 | an 'I' will be displayed on the same line as the function that |
| 251 | can be overridden. |
| 252 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 253 | function_profile_enabled: |
| 254 | |
| 255 | When set it will enable all functions with either the function |
| 256 | tracer, or if enabled, the function graph tracer. It will |
| 257 | keep a histogram of the number of functions that were called |
| 258 | and if run with the function graph tracer, it will also keep |
| 259 | track of the time spent in those functions. The histogram |
| 260 | content can be displayed in the files: |
| 261 | |
| 262 | trace_stats/function<cpu> ( function0, function1, etc). |
| 263 | |
| 264 | trace_stats: |
| 265 | |
| 266 | A directory that holds different tracing stats. |
| 267 | |
| 268 | kprobe_events: |
| 269 | |
| 270 | Enable dynamic trace points. See kprobetrace.txt. |
| 271 | |
| 272 | kprobe_profile: |
| 273 | |
| 274 | Dynamic trace points stats. See kprobetrace.txt. |
| 275 | |
| 276 | max_graph_depth: |
| 277 | |
| 278 | Used with the function graph tracer. This is the max depth |
| 279 | it will trace into a function. Setting this to a value of |
| 280 | one will show only the first kernel function that is called |
| 281 | from user space. |
| 282 | |
| 283 | printk_formats: |
| 284 | |
| 285 | This is for tools that read the raw format files. If an event in |
| 286 | the ring buffer references a string (currently only trace_printk() |
| 287 | does this), only a pointer to the string is recorded into the buffer |
| 288 | and not the string itself. This prevents tools from knowing what |
| 289 | that string was. This file displays the string and address for |
| 290 | the string allowing tools to map the pointers to what the |
| 291 | strings were. |
| 292 | |
| 293 | saved_cmdlines: |
| 294 | |
| 295 | Only the pid of the task is recorded in a trace event unless |
| 296 | the event specifically saves the task comm as well. Ftrace |
| 297 | makes a cache of pid mappings to comms to try to display |
| 298 | comms for events. If a pid for a comm is not listed, then |
| 299 | "<...>" is displayed in the output. |
| 300 | |
| 301 | snapshot: |
| 302 | |
| 303 | This displays the "snapshot" buffer and also lets the user |
| 304 | take a snapshot of the current running trace. |
| 305 | See the "Snapshot" section below for more details. |
| 306 | |
| 307 | stack_max_size: |
| 308 | |
| 309 | When the stack tracer is activated, this will display the |
| 310 | maximum stack size it has encountered. |
| 311 | See the "Stack Trace" section below. |
| 312 | |
| 313 | stack_trace: |
| 314 | |
| 315 | This displays the stack back trace of the largest stack |
| 316 | that was encountered when the stack tracer is activated. |
| 317 | See the "Stack Trace" section below. |
| 318 | |
| 319 | stack_trace_filter: |
| 320 | |
| 321 | This is similar to "set_ftrace_filter" but it limits what |
| 322 | functions the stack tracer will check. |
| 323 | |
| 324 | trace_clock: |
| 325 | |
| 326 | Whenever an event is recorded into the ring buffer, a |
| 327 | "timestamp" is added. This stamp comes from a specified |
| 328 | clock. By default, ftrace uses the "local" clock. This |
| 329 | clock is very fast and strictly per cpu, but on some |
| 330 | systems it may not be monotonic with respect to other |
| 331 | CPUs. In other words, the local clocks may not be in sync |
| 332 | with local clocks on other CPUs. |
| 333 | |
| 334 | Usual clocks for tracing: |
| 335 | |
| 336 | # cat trace_clock |
| 337 | [local] global counter x86-tsc |
| 338 | |
| 339 | local: Default clock, but may not be in sync across CPUs |
| 340 | |
| 341 | global: This clock is in sync with all CPUs but may |
| 342 | be a bit slower than the local clock. |
| 343 | |
| 344 | counter: This is not a clock at all, but literally an atomic |
| 345 | counter. It counts up one by one, but is in sync |
| 346 | with all CPUs. This is useful when you need to |
| 347 | know exactly the order events occurred with respect to |
| 348 | each other on different CPUs. |
| 349 | |
| 350 | uptime: This uses the jiffies counter and the time stamp |
| 351 | is relative to the time since boot up. |
| 352 | |
| 353 | perf: This makes ftrace use the same clock that perf uses. |
| 354 | Eventually perf will be able to read ftrace buffers |
| 355 | and this will help out in interleaving the data. |
| 356 | |
| 357 | x86-tsc: Architectures may define their own clocks. For |
| 358 | example, x86 uses its own TSC cycle clock here. |
| 359 | |
Naveen N. Rao | 197165d | 2015-04-24 14:24:44 +0530 | [diff] [blame] | 360 | ppc-tb: This uses the powerpc timebase register value. |
| 361 | This is in sync across CPUs and can also be used |
| 362 | to correlate events across hypervisor/guest if |
| 363 | tb_offset is known. |
| 364 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 365 | To set a clock, simply echo the clock name into this file. |
| 366 | |
| 367 | echo global > trace_clock |
| 368 | |
| 369 | trace_marker: |
| 370 | |
| 371 | This is a very useful file for synchronizing user space |
| 372 | with events happening in the kernel. Writing strings into |
| 373 | this file will be written into the ftrace buffer. |
| 374 | |
| 375 | It is useful in applications to open this file at the start |
| 376 | of the application and just reference the file descriptor |
| 377 | for the file. |
| 378 | |
| 379 | void trace_write(const char *fmt, ...) |
| 380 | { |
| 381 | va_list ap; |
| 382 | char buf[256]; |
| 383 | int n; |
| 384 | |
| 385 | if (trace_fd < 0) |
| 386 | return; |
| 387 | |
| 388 | va_start(ap, fmt); |
| 389 | n = vsnprintf(buf, 256, fmt, ap); |
| 390 | va_end(ap); |
| 391 | |
| 392 | write(trace_fd, buf, n); |
| 393 | } |
| 394 | |
| 395 | start: |
| 396 | |
| 397 | trace_fd = open("trace_marker", WR_ONLY); |
| 398 | |
| 399 | uprobe_events: |
| 400 | |
| 401 | Add dynamic tracepoints in programs. |
| 402 | See uprobetracer.txt |
| 403 | |
| 404 | uprobe_profile: |
| 405 | |
| 406 | Uprobe statistics. See uprobetrace.txt |
| 407 | |
| 408 | instances: |
| 409 | |
| 410 | This is a way to make multiple trace buffers where different |
| 411 | events can be recorded in different buffers. |
| 412 | See "Instances" section below. |
| 413 | |
| 414 | events: |
| 415 | |
| 416 | This is the trace event directory. It holds event tracepoints |
| 417 | (also known as static tracepoints) that have been compiled |
| 418 | into the kernel. It shows what event tracepoints exist |
| 419 | and how they are grouped by system. There are "enable" |
| 420 | files at various levels that can enable the tracepoints |
| 421 | when a "1" is written to them. |
| 422 | |
| 423 | See events.txt for more information. |
| 424 | |
| 425 | per_cpu: |
| 426 | |
| 427 | This is a directory that contains the trace per_cpu information. |
| 428 | |
| 429 | per_cpu/cpu0/buffer_size_kb: |
| 430 | |
| 431 | The ftrace buffer is defined per_cpu. That is, there's a separate |
| 432 | buffer for each CPU to allow writes to be done atomically, |
| 433 | and free from cache bouncing. These buffers may have different |
| 434 | size buffers. This file is similar to the buffer_size_kb |
| 435 | file, but it only displays or sets the buffer size for the |
| 436 | specific CPU. (here cpu0). |
| 437 | |
| 438 | per_cpu/cpu0/trace: |
| 439 | |
| 440 | This is similar to the "trace" file, but it will only display |
| 441 | the data specific for the CPU. If written to, it only clears |
| 442 | the specific CPU buffer. |
| 443 | |
| 444 | per_cpu/cpu0/trace_pipe |
| 445 | |
| 446 | This is similar to the "trace_pipe" file, and is a consuming |
| 447 | read, but it will only display (and consume) the data specific |
| 448 | for the CPU. |
| 449 | |
| 450 | per_cpu/cpu0/trace_pipe_raw |
| 451 | |
| 452 | For tools that can parse the ftrace ring buffer binary format, |
| 453 | the trace_pipe_raw file can be used to extract the data |
| 454 | from the ring buffer directly. With the use of the splice() |
| 455 | system call, the buffer data can be quickly transferred to |
| 456 | a file or to the network where a server is collecting the |
| 457 | data. |
| 458 | |
| 459 | Like trace_pipe, this is a consuming reader, where multiple |
| 460 | reads will always produce different data. |
| 461 | |
| 462 | per_cpu/cpu0/snapshot: |
| 463 | |
| 464 | This is similar to the main "snapshot" file, but will only |
| 465 | snapshot the current CPU (if supported). It only displays |
| 466 | the content of the snapshot for a given CPU, and if |
| 467 | written to, only clears this CPU buffer. |
| 468 | |
| 469 | per_cpu/cpu0/snapshot_raw: |
| 470 | |
| 471 | Similar to the trace_pipe_raw, but will read the binary format |
| 472 | from the snapshot buffer for the given CPU. |
| 473 | |
| 474 | per_cpu/cpu0/stats: |
| 475 | |
| 476 | This displays certain stats about the ring buffer: |
| 477 | |
| 478 | entries: The number of events that are still in the buffer. |
| 479 | |
| 480 | overrun: The number of lost events due to overwriting when |
| 481 | the buffer was full. |
| 482 | |
| 483 | commit overrun: Should always be zero. |
| 484 | This gets set if so many events happened within a nested |
| 485 | event (ring buffer is re-entrant), that it fills the |
| 486 | buffer and starts dropping events. |
| 487 | |
| 488 | bytes: Bytes actually read (not overwritten). |
| 489 | |
| 490 | oldest event ts: The oldest timestamp in the buffer |
| 491 | |
| 492 | now ts: The current timestamp |
| 493 | |
| 494 | dropped events: Events lost due to overwrite option being off. |
| 495 | |
| 496 | read events: The number of events read. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 497 | |
| 498 | The Tracers |
| 499 | ----------- |
| 500 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 501 | Here is the list of current tracers that may be configured. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 502 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 503 | "function" |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 504 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 505 | Function call tracer to trace all kernel functions. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 506 | |
Mike Frysinger | bc5c6c0 | 2009-06-10 04:48:41 -0400 | [diff] [blame] | 507 | "function_graph" |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 508 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 509 | Similar to the function tracer except that the |
| 510 | function tracer probes the functions on their entry |
| 511 | whereas the function graph tracer traces on both entry |
| 512 | and exit of the functions. It then provides the ability |
| 513 | to draw a graph of function calls similar to C code |
| 514 | source. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 515 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 516 | "irqsoff" |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 517 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 518 | Traces the areas that disable interrupts and saves |
| 519 | the trace with the longest max latency. |
| 520 | See tracing_max_latency. When a new max is recorded, |
| 521 | it replaces the old trace. It is best to view this |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 522 | trace with the latency-format option enabled. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 523 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 524 | "preemptoff" |
| 525 | |
| 526 | Similar to irqsoff but traces and records the amount of |
| 527 | time for which preemption is disabled. |
| 528 | |
| 529 | "preemptirqsoff" |
| 530 | |
| 531 | Similar to irqsoff and preemptoff, but traces and |
| 532 | records the largest time for which irqs and/or preemption |
| 533 | is disabled. |
| 534 | |
| 535 | "wakeup" |
| 536 | |
| 537 | Traces and records the max latency that it takes for |
| 538 | the highest priority task to get scheduled after |
| 539 | it has been woken up. |
Geunsik Lim | cdfb0d3 | 2012-02-08 19:05:37 +0900 | [diff] [blame] | 540 | Traces all tasks as an average developer would expect. |
| 541 | |
| 542 | "wakeup_rt" |
| 543 | |
| 544 | Traces and records the max latency that it takes for just |
| 545 | RT tasks (as the current "wakeup" does). This is useful |
| 546 | for those interested in wake up timings of RT tasks. |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 547 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 548 | "nop" |
| 549 | |
| 550 | This is the "trace nothing" tracer. To remove all |
| 551 | tracers from tracing simply echo "nop" into |
| 552 | current_tracer. |
Markus Metzger | e2ea539 | 2009-01-19 10:35:58 +0100 | [diff] [blame] | 553 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 554 | |
| 555 | Examples of using the tracer |
| 556 | ---------------------------- |
| 557 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 558 | Here are typical examples of using the tracers when controlling |
| 559 | them only with the debugfs interface (without using any |
| 560 | user-land utilities). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 561 | |
| 562 | Output format: |
| 563 | -------------- |
| 564 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 565 | Here is an example of the output format of the file "trace" |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 566 | |
| 567 | -------- |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 568 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 569 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 570 | # entries-in-buffer/entries-written: 140080/250280 #P:4 |
| 571 | # |
| 572 | # _-----=> irqs-off |
| 573 | # / _----=> need-resched |
| 574 | # | / _---=> hardirq/softirq |
| 575 | # || / _--=> preempt-depth |
| 576 | # ||| / delay |
| 577 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 578 | # | | | |||| | | |
| 579 | bash-1977 [000] .... 17284.993652: sys_close <-system_call_fastpath |
| 580 | bash-1977 [000] .... 17284.993653: __close_fd <-sys_close |
| 581 | bash-1977 [000] .... 17284.993653: _raw_spin_lock <-__close_fd |
| 582 | sshd-1974 [003] .... 17284.993653: __srcu_read_unlock <-fsnotify |
| 583 | bash-1977 [000] .... 17284.993654: add_preempt_count <-_raw_spin_lock |
| 584 | bash-1977 [000] ...1 17284.993655: _raw_spin_unlock <-__close_fd |
| 585 | bash-1977 [000] ...1 17284.993656: sub_preempt_count <-_raw_spin_unlock |
| 586 | bash-1977 [000] .... 17284.993657: filp_close <-__close_fd |
| 587 | bash-1977 [000] .... 17284.993657: dnotify_flush <-filp_close |
| 588 | sshd-1974 [003] .... 17284.993658: sys_select <-system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 589 | -------- |
| 590 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 591 | A header is printed with the tracer name that is represented by |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 592 | the trace. In this case the tracer is "function". Then it shows the |
| 593 | number of events in the buffer as well as the total number of entries |
| 594 | that were written. The difference is the number of entries that were |
| 595 | lost due to the buffer filling up (250280 - 140080 = 110200 events |
| 596 | lost). |
| 597 | |
| 598 | The header explains the content of the events. Task name "bash", the task |
| 599 | PID "1977", the CPU that it was running on "000", the latency format |
| 600 | (explained below), the timestamp in <secs>.<usecs> format, the |
| 601 | function name that was traced "sys_close" and the parent function that |
| 602 | called this function "system_call_fastpath". The timestamp is the time |
| 603 | at which the function was entered. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 604 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 605 | Latency trace format |
| 606 | -------------------- |
| 607 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 608 | When the latency-format option is enabled or when one of the latency |
| 609 | tracers is set, the trace file gives somewhat more information to see |
| 610 | why a latency happened. Here is a typical trace. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 611 | |
| 612 | # tracer: irqsoff |
| 613 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 614 | # irqsoff latency trace v1.1.5 on 3.8.0-test+ |
| 615 | # -------------------------------------------------------------------- |
| 616 | # latency: 259 us, #4/4, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 617 | # ----------------- |
| 618 | # | task: ps-6143 (uid:0 nice:0 policy:0 rt_prio:0) |
| 619 | # ----------------- |
| 620 | # => started at: __lock_task_sighand |
| 621 | # => ended at: _raw_spin_unlock_irqrestore |
| 622 | # |
| 623 | # |
| 624 | # _------=> CPU# |
| 625 | # / _-----=> irqs-off |
| 626 | # | / _----=> need-resched |
| 627 | # || / _---=> hardirq/softirq |
| 628 | # ||| / _--=> preempt-depth |
| 629 | # |||| / delay |
| 630 | # cmd pid ||||| time | caller |
| 631 | # \ / ||||| \ | / |
| 632 | ps-6143 2d... 0us!: trace_hardirqs_off <-__lock_task_sighand |
| 633 | ps-6143 2d..1 259us+: trace_hardirqs_on <-_raw_spin_unlock_irqrestore |
| 634 | ps-6143 2d..1 263us+: time_hardirqs_on <-_raw_spin_unlock_irqrestore |
| 635 | ps-6143 2d..1 306us : <stack trace> |
| 636 | => trace_hardirqs_on_caller |
| 637 | => trace_hardirqs_on |
| 638 | => _raw_spin_unlock_irqrestore |
| 639 | => do_task_stat |
| 640 | => proc_tgid_stat |
| 641 | => proc_single_show |
| 642 | => seq_read |
| 643 | => vfs_read |
| 644 | => sys_read |
| 645 | => system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 646 | |
| 647 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 648 | This shows that the current tracer is "irqsoff" tracing the time |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 649 | for which interrupts were disabled. It gives the trace version (which |
| 650 | never changes) and the version of the kernel upon which this was executed on |
| 651 | (3.10). Then it displays the max latency in microseconds (259 us). The number |
| 652 | of trace entries displayed and the total number (both are four: #4/4). |
| 653 | VP, KP, SP, and HP are always zero and are reserved for later use. |
| 654 | #P is the number of online CPUs (#P:4). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 655 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 656 | The task is the process that was running when the latency |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 657 | occurred. (ps pid: 6143). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 658 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 659 | The start and stop (the functions in which the interrupts were |
| 660 | disabled and enabled respectively) that caused the latencies: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 661 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 662 | __lock_task_sighand is where the interrupts were disabled. |
| 663 | _raw_spin_unlock_irqrestore is where they were enabled again. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 664 | |
| 665 | The next lines after the header are the trace itself. The header |
| 666 | explains which is which. |
| 667 | |
| 668 | cmd: The name of the process in the trace. |
| 669 | |
| 670 | pid: The PID of that process. |
| 671 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 672 | CPU#: The CPU which the process was running on. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 673 | |
| 674 | irqs-off: 'd' interrupts are disabled. '.' otherwise. |
Steven Rostedt | 9244489 | 2008-10-24 09:42:59 -0400 | [diff] [blame] | 675 | Note: If the architecture does not support a way to |
| 676 | read the irq flags variable, an 'X' will always |
| 677 | be printed here. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 678 | |
Peter Zijlstra | e5137b5 | 2013-10-04 17:28:26 +0200 | [diff] [blame] | 679 | need-resched: |
| 680 | 'N' both TIF_NEED_RESCHED and PREEMPT_NEED_RESCHED is set, |
| 681 | 'n' only TIF_NEED_RESCHED is set, |
| 682 | 'p' only PREEMPT_NEED_RESCHED is set, |
| 683 | '.' otherwise. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 684 | |
| 685 | hardirq/softirq: |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 686 | 'H' - hard irq occurred inside a softirq. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 687 | 'h' - hard irq is running |
| 688 | 's' - soft irq is running |
| 689 | '.' - normal context. |
| 690 | |
| 691 | preempt-depth: The level of preempt_disabled |
| 692 | |
| 693 | The above is mostly meaningful for kernel developers. |
| 694 | |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 695 | time: When the latency-format option is enabled, the trace file |
| 696 | output includes a timestamp relative to the start of the |
| 697 | trace. This differs from the output when latency-format |
| 698 | is disabled, which includes an absolute timestamp. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 699 | |
| 700 | delay: This is just to help catch your eye a bit better. And |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 701 | needs to be fixed to be only relative to the same CPU. |
| 702 | The marks are determined by the difference between this |
| 703 | current trace and the next trace. |
Byungchul Park | 8e1e1df | 2014-11-24 09:34:19 +0900 | [diff] [blame] | 704 | '$' - greater than 1 second |
Jungseok Lee | b838e1d | 2015-07-11 14:51:40 +0000 | [diff] [blame] | 705 | '@' - greater than 100 milisecond |
| 706 | '*' - greater than 10 milisecond |
Byungchul Park | 8e1e1df | 2014-11-24 09:34:19 +0900 | [diff] [blame] | 707 | '#' - greater than 1000 microsecond |
| 708 | '!' - greater than 100 microsecond |
| 709 | '+' - greater than 10 microsecond |
| 710 | ' ' - less than or equal to 10 microsecond. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 711 | |
| 712 | The rest is the same as the 'trace' file. |
| 713 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 714 | Note, the latency tracers will usually end with a back trace |
| 715 | to easily find where the latency occurred. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 716 | |
Steven Rostedt | ee6bce5 | 2008-11-12 17:52:37 -0500 | [diff] [blame] | 717 | trace_options |
| 718 | ------------- |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 719 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 720 | The trace_options file (or the options directory) is used to control |
| 721 | what gets printed in the trace output, or manipulate the tracers. |
| 722 | To see what is available, simply cat the file: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 723 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 724 | cat trace_options |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 725 | print-parent |
| 726 | nosym-offset |
| 727 | nosym-addr |
| 728 | noverbose |
| 729 | noraw |
| 730 | nohex |
| 731 | nobin |
| 732 | noblock |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 733 | trace_printk |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 734 | nobranch |
| 735 | annotate |
| 736 | nouserstacktrace |
| 737 | nosym-userobj |
| 738 | noprintk-msg-only |
| 739 | context-info |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 740 | nolatency-format |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 741 | sleep-time |
| 742 | graph-time |
| 743 | record-cmd |
| 744 | overwrite |
| 745 | nodisable_on_free |
| 746 | irq-info |
| 747 | markers |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 748 | noevent-fork |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 749 | function-trace |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 750 | nodisplay-graph |
| 751 | nostacktrace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 752 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 753 | To disable one of the options, echo in the option prepended with |
| 754 | "no". |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 755 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 756 | echo noprint-parent > trace_options |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 757 | |
| 758 | To enable an option, leave off the "no". |
| 759 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 760 | echo sym-offset > trace_options |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 761 | |
| 762 | Here are the available options: |
| 763 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 764 | print-parent - On function traces, display the calling (parent) |
| 765 | function as well as the function being traced. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 766 | |
| 767 | print-parent: |
Jingoo Han | 0ab943b | 2013-07-31 14:59:16 -0700 | [diff] [blame] | 768 | bash-4000 [01] 1477.606694: simple_strtoul <-kstrtoul |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 769 | |
| 770 | noprint-parent: |
| 771 | bash-4000 [01] 1477.606694: simple_strtoul |
| 772 | |
| 773 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 774 | sym-offset - Display not only the function name, but also the |
| 775 | offset in the function. For example, instead of |
| 776 | seeing just "ktime_get", you will see |
| 777 | "ktime_get+0xb/0x20". |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 778 | |
| 779 | sym-offset: |
| 780 | bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0 |
| 781 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 782 | sym-addr - this will also display the function address as well |
| 783 | as the function name. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 784 | |
| 785 | sym-addr: |
| 786 | bash-4000 [01] 1477.606694: simple_strtoul <c0339346> |
| 787 | |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 788 | verbose - This deals with the trace file when the |
| 789 | latency-format option is enabled. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 790 | |
| 791 | bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \ |
Jingoo Han | 0ab943b | 2013-07-31 14:59:16 -0700 | [diff] [blame] | 792 | (+0.000ms): simple_strtoul (kstrtoul) |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 793 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 794 | raw - This will display raw numbers. This option is best for |
| 795 | use with user applications that can translate the raw |
| 796 | numbers better than having it done in the kernel. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 797 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 798 | hex - Similar to raw, but the numbers will be in a hexadecimal |
| 799 | format. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 800 | |
| 801 | bin - This will print out the formats in raw binary. |
| 802 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 803 | block - When set, reading trace_pipe will not block when polled. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 804 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 805 | trace_printk - Can disable trace_printk() from writing into the buffer. |
| 806 | |
| 807 | branch - Enable branch tracing with the tracer. |
| 808 | |
| 809 | annotate - It is sometimes confusing when the CPU buffers are full |
| 810 | and one CPU buffer had a lot of events recently, thus |
| 811 | a shorter time frame, were another CPU may have only had |
| 812 | a few events, which lets it have older events. When |
| 813 | the trace is reported, it shows the oldest events first, |
| 814 | and it may look like only one CPU ran (the one with the |
| 815 | oldest events). When the annotate option is set, it will |
| 816 | display when a new CPU buffer started: |
| 817 | |
| 818 | <idle>-0 [001] dNs4 21169.031481: wake_up_idle_cpu <-add_timer_on |
| 819 | <idle>-0 [001] dNs4 21169.031482: _raw_spin_unlock_irqrestore <-add_timer_on |
| 820 | <idle>-0 [001] .Ns4 21169.031484: sub_preempt_count <-_raw_spin_unlock_irqrestore |
| 821 | ##### CPU 2 buffer started #### |
| 822 | <idle>-0 [002] .N.1 21169.031484: rcu_idle_exit <-cpu_idle |
| 823 | <idle>-0 [001] .Ns3 21169.031484: _raw_spin_unlock <-clocksource_watchdog |
| 824 | <idle>-0 [001] .Ns3 21169.031485: sub_preempt_count <-_raw_spin_unlock |
| 825 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 826 | userstacktrace - This option changes the trace. It records a |
| 827 | stacktrace of the current userspace thread. |
Török Edwin | 02b6751 | 2008-11-22 13:28:47 +0200 | [diff] [blame] | 828 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 829 | sym-userobj - when user stacktrace are enabled, look up which |
| 830 | object the address belongs to, and print a |
| 831 | relative address. This is especially useful when |
| 832 | ASLR is on, otherwise you don't get a chance to |
| 833 | resolve the address to object/file/line after |
| 834 | the app is no longer running |
Török Edwin | b54d3de | 2008-11-22 13:28:48 +0200 | [diff] [blame] | 835 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 836 | The lookup is performed when you read |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 837 | trace,trace_pipe. Example: |
Török Edwin | b54d3de | 2008-11-22 13:28:48 +0200 | [diff] [blame] | 838 | |
| 839 | a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0 |
| 840 | x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6] |
| 841 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 842 | |
| 843 | printk-msg-only - When set, trace_printk()s will only show the format |
| 844 | and not their parameters (if trace_bprintk() or |
| 845 | trace_bputs() was used to save the trace_printk()). |
| 846 | |
| 847 | context-info - Show only the event data. Hides the comm, PID, |
| 848 | timestamp, CPU, and other useful data. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 849 | |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 850 | latency-format - This option changes the trace. When |
| 851 | it is enabled, the trace displays |
| 852 | additional information about the |
| 853 | latencies, as described in "Latency |
| 854 | trace format". |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 855 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 856 | sleep-time - When running function graph tracer, to include |
| 857 | the time a task schedules out in its function. |
| 858 | When enabled, it will account time the task has been |
| 859 | scheduled out as part of the function call. |
| 860 | |
| 861 | graph-time - When running function graph tracer, to include the |
| 862 | time to call nested functions. When this is not set, |
| 863 | the time reported for the function will only include |
| 864 | the time the function itself executed for, not the time |
| 865 | for functions that it called. |
| 866 | |
| 867 | record-cmd - When any event or tracer is enabled, a hook is enabled |
| 868 | in the sched_switch trace point to fill comm cache |
| 869 | with mapped pids and comms. But this may cause some |
| 870 | overhead, and if you only care about pids, and not the |
| 871 | name of the task, disabling this option can lower the |
| 872 | impact of tracing. |
| 873 | |
David Sharp | 750912f | 2010-12-08 13:46:47 -0800 | [diff] [blame] | 874 | overwrite - This controls what happens when the trace buffer is |
| 875 | full. If "1" (default), the oldest events are |
| 876 | discarded and overwritten. If "0", then the newest |
| 877 | events are discarded. |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 878 | (see per_cpu/cpu0/stats for overrun and dropped) |
David Sharp | 750912f | 2010-12-08 13:46:47 -0800 | [diff] [blame] | 879 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 880 | disable_on_free - When the free_buffer is closed, tracing will |
| 881 | stop (tracing_on set to 0). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 882 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 883 | irq-info - Shows the interrupt, preempt count, need resched data. |
| 884 | When disabled, the trace looks like: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 885 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 886 | # tracer: function |
| 887 | # |
| 888 | # entries-in-buffer/entries-written: 144405/9452052 #P:4 |
| 889 | # |
| 890 | # TASK-PID CPU# TIMESTAMP FUNCTION |
| 891 | # | | | | | |
| 892 | <idle>-0 [002] 23636.756054: ttwu_do_activate.constprop.89 <-try_to_wake_up |
| 893 | <idle>-0 [002] 23636.756054: activate_task <-ttwu_do_activate.constprop.89 |
| 894 | <idle>-0 [002] 23636.756055: enqueue_task <-activate_task |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 895 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 896 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 897 | markers - When set, the trace_marker is writable (only by root). |
| 898 | When disabled, the trace_marker will error with EINVAL |
| 899 | on write. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 900 | |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 901 | event-fork - When set, tasks with PIDs listed in set_event_pid will have |
| 902 | the PIDs of their children added to set_event_pid when those |
| 903 | tasks fork. Also, when tasks with PIDs in set_event_pid exit, |
| 904 | their PIDs will be removed from the file. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 905 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 906 | function-trace - The latency tracers will enable function tracing |
| 907 | if this option is enabled (default it is). When |
| 908 | it is disabled, the latency tracers do not trace |
| 909 | functions. This keeps the overhead of the tracer down |
| 910 | when performing latency tests. |
| 911 | |
Steven Rostedt (Red Hat) | 540b589 | 2016-04-19 10:24:38 -0400 | [diff] [blame] | 912 | display-graph - When set, the latency tracers (irqsoff, wakeup, etc) will |
| 913 | use function graph tracing instead of function tracing. |
| 914 | |
| 915 | stacktrace - This is one of the options that changes the trace |
| 916 | itself. When a trace is recorded, so is the stack |
| 917 | of functions. This allows for back traces of |
| 918 | trace sites. |
| 919 | |
| 920 | Note: Some tracers have their own options. They only appear in this |
| 921 | file when the tracer is active. They always appear in the |
| 922 | options directory. |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 923 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 924 | |
| 925 | |
| 926 | irqsoff |
| 927 | ------- |
| 928 | |
| 929 | When interrupts are disabled, the CPU can not react to any other |
| 930 | external event (besides NMIs and SMIs). This prevents the timer |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 931 | interrupt from triggering or the mouse interrupt from letting |
| 932 | the kernel know of a new mouse event. The result is a latency |
| 933 | with the reaction time. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 934 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 935 | The irqsoff tracer tracks the time for which interrupts are |
| 936 | disabled. When a new maximum latency is hit, the tracer saves |
| 937 | the trace leading up to that latency point so that every time a |
| 938 | new maximum is reached, the old saved trace is discarded and the |
| 939 | new trace is saved. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 940 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 941 | To reset the maximum, echo 0 into tracing_max_latency. Here is |
| 942 | an example: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 943 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 944 | # echo 0 > options/function-trace |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 945 | # echo irqsoff > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 946 | # echo 1 > tracing_on |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 947 | # echo 0 > tracing_max_latency |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 948 | # ls -ltr |
| 949 | [...] |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 950 | # echo 0 > tracing_on |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 951 | # cat trace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 952 | # tracer: irqsoff |
| 953 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 954 | # irqsoff latency trace v1.1.5 on 3.8.0-test+ |
| 955 | # -------------------------------------------------------------------- |
| 956 | # latency: 16 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 957 | # ----------------- |
| 958 | # | task: swapper/0-0 (uid:0 nice:0 policy:0 rt_prio:0) |
| 959 | # ----------------- |
| 960 | # => started at: run_timer_softirq |
| 961 | # => ended at: run_timer_softirq |
| 962 | # |
| 963 | # |
| 964 | # _------=> CPU# |
| 965 | # / _-----=> irqs-off |
| 966 | # | / _----=> need-resched |
| 967 | # || / _---=> hardirq/softirq |
| 968 | # ||| / _--=> preempt-depth |
| 969 | # |||| / delay |
| 970 | # cmd pid ||||| time | caller |
| 971 | # \ / ||||| \ | / |
| 972 | <idle>-0 0d.s2 0us+: _raw_spin_lock_irq <-run_timer_softirq |
| 973 | <idle>-0 0dNs3 17us : _raw_spin_unlock_irq <-run_timer_softirq |
| 974 | <idle>-0 0dNs3 17us+: trace_hardirqs_on <-run_timer_softirq |
| 975 | <idle>-0 0dNs3 25us : <stack trace> |
| 976 | => _raw_spin_unlock_irq |
| 977 | => run_timer_softirq |
| 978 | => __do_softirq |
| 979 | => call_softirq |
| 980 | => do_softirq |
| 981 | => irq_exit |
| 982 | => smp_apic_timer_interrupt |
| 983 | => apic_timer_interrupt |
| 984 | => rcu_idle_exit |
| 985 | => cpu_idle |
| 986 | => rest_init |
| 987 | => start_kernel |
| 988 | => x86_64_start_reservations |
| 989 | => x86_64_start_kernel |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 990 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 991 | Here we see that that we had a latency of 16 microseconds (which is |
| 992 | very good). The _raw_spin_lock_irq in run_timer_softirq disabled |
| 993 | interrupts. The difference between the 16 and the displayed |
| 994 | timestamp 25us occurred because the clock was incremented |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 995 | between the time of recording the max latency and the time of |
| 996 | recording the function that had that latency. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 997 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 998 | Note the above example had function-trace not set. If we set |
| 999 | function-trace, we get a much larger output: |
| 1000 | |
| 1001 | with echo 1 > options/function-trace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1002 | |
| 1003 | # tracer: irqsoff |
| 1004 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1005 | # irqsoff latency trace v1.1.5 on 3.8.0-test+ |
| 1006 | # -------------------------------------------------------------------- |
| 1007 | # latency: 71 us, #168/168, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1008 | # ----------------- |
| 1009 | # | task: bash-2042 (uid:0 nice:0 policy:0 rt_prio:0) |
| 1010 | # ----------------- |
| 1011 | # => started at: ata_scsi_queuecmd |
| 1012 | # => ended at: ata_scsi_queuecmd |
| 1013 | # |
| 1014 | # |
| 1015 | # _------=> CPU# |
| 1016 | # / _-----=> irqs-off |
| 1017 | # | / _----=> need-resched |
| 1018 | # || / _---=> hardirq/softirq |
| 1019 | # ||| / _--=> preempt-depth |
| 1020 | # |||| / delay |
| 1021 | # cmd pid ||||| time | caller |
| 1022 | # \ / ||||| \ | / |
| 1023 | bash-2042 3d... 0us : _raw_spin_lock_irqsave <-ata_scsi_queuecmd |
| 1024 | bash-2042 3d... 0us : add_preempt_count <-_raw_spin_lock_irqsave |
| 1025 | bash-2042 3d..1 1us : ata_scsi_find_dev <-ata_scsi_queuecmd |
| 1026 | bash-2042 3d..1 1us : __ata_scsi_find_dev <-ata_scsi_find_dev |
| 1027 | bash-2042 3d..1 2us : ata_find_dev.part.14 <-__ata_scsi_find_dev |
| 1028 | bash-2042 3d..1 2us : ata_qc_new_init <-__ata_scsi_queuecmd |
| 1029 | bash-2042 3d..1 3us : ata_sg_init <-__ata_scsi_queuecmd |
| 1030 | bash-2042 3d..1 4us : ata_scsi_rw_xlat <-__ata_scsi_queuecmd |
| 1031 | bash-2042 3d..1 4us : ata_build_rw_tf <-ata_scsi_rw_xlat |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1032 | [...] |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1033 | bash-2042 3d..1 67us : delay_tsc <-__delay |
| 1034 | bash-2042 3d..1 67us : add_preempt_count <-delay_tsc |
| 1035 | bash-2042 3d..2 67us : sub_preempt_count <-delay_tsc |
| 1036 | bash-2042 3d..1 67us : add_preempt_count <-delay_tsc |
| 1037 | bash-2042 3d..2 68us : sub_preempt_count <-delay_tsc |
| 1038 | bash-2042 3d..1 68us+: ata_bmdma_start <-ata_bmdma_qc_issue |
| 1039 | bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd |
| 1040 | bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd |
| 1041 | bash-2042 3d..1 72us+: trace_hardirqs_on <-ata_scsi_queuecmd |
| 1042 | bash-2042 3d..1 120us : <stack trace> |
| 1043 | => _raw_spin_unlock_irqrestore |
| 1044 | => ata_scsi_queuecmd |
| 1045 | => scsi_dispatch_cmd |
| 1046 | => scsi_request_fn |
| 1047 | => __blk_run_queue_uncond |
| 1048 | => __blk_run_queue |
| 1049 | => blk_queue_bio |
| 1050 | => generic_make_request |
| 1051 | => submit_bio |
| 1052 | => submit_bh |
| 1053 | => __ext3_get_inode_loc |
| 1054 | => ext3_iget |
| 1055 | => ext3_lookup |
| 1056 | => lookup_real |
| 1057 | => __lookup_hash |
| 1058 | => walk_component |
| 1059 | => lookup_last |
| 1060 | => path_lookupat |
| 1061 | => filename_lookup |
| 1062 | => user_path_at_empty |
| 1063 | => user_path_at |
| 1064 | => vfs_fstatat |
| 1065 | => vfs_stat |
| 1066 | => sys_newstat |
| 1067 | => system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1068 | |
| 1069 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1070 | Here we traced a 71 microsecond latency. But we also see all the |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1071 | functions that were called during that time. Note that by |
| 1072 | enabling function tracing, we incur an added overhead. This |
| 1073 | overhead may extend the latency times. But nevertheless, this |
| 1074 | trace has provided some very helpful debugging information. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1075 | |
| 1076 | |
| 1077 | preemptoff |
| 1078 | ---------- |
| 1079 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1080 | When preemption is disabled, we may be able to receive |
| 1081 | interrupts but the task cannot be preempted and a higher |
| 1082 | priority task must wait for preemption to be enabled again |
| 1083 | before it can preempt a lower priority task. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1084 | |
Steven Rostedt | a41eeba | 2008-07-14 16:41:12 -0400 | [diff] [blame] | 1085 | The preemptoff tracer traces the places that disable preemption. |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1086 | Like the irqsoff tracer, it records the maximum latency for |
| 1087 | which preemption was disabled. The control of preemptoff tracer |
| 1088 | is much like the irqsoff tracer. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1089 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1090 | # echo 0 > options/function-trace |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1091 | # echo preemptoff > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1092 | # echo 1 > tracing_on |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1093 | # echo 0 > tracing_max_latency |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1094 | # ls -ltr |
| 1095 | [...] |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1096 | # echo 0 > tracing_on |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 1097 | # cat trace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1098 | # tracer: preemptoff |
| 1099 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1100 | # preemptoff latency trace v1.1.5 on 3.8.0-test+ |
| 1101 | # -------------------------------------------------------------------- |
| 1102 | # latency: 46 us, #4/4, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1103 | # ----------------- |
| 1104 | # | task: sshd-1991 (uid:0 nice:0 policy:0 rt_prio:0) |
| 1105 | # ----------------- |
| 1106 | # => started at: do_IRQ |
| 1107 | # => ended at: do_IRQ |
| 1108 | # |
| 1109 | # |
| 1110 | # _------=> CPU# |
| 1111 | # / _-----=> irqs-off |
| 1112 | # | / _----=> need-resched |
| 1113 | # || / _---=> hardirq/softirq |
| 1114 | # ||| / _--=> preempt-depth |
| 1115 | # |||| / delay |
| 1116 | # cmd pid ||||| time | caller |
| 1117 | # \ / ||||| \ | / |
| 1118 | sshd-1991 1d.h. 0us+: irq_enter <-do_IRQ |
| 1119 | sshd-1991 1d..1 46us : irq_exit <-do_IRQ |
| 1120 | sshd-1991 1d..1 47us+: trace_preempt_on <-do_IRQ |
| 1121 | sshd-1991 1d..1 52us : <stack trace> |
| 1122 | => sub_preempt_count |
| 1123 | => irq_exit |
| 1124 | => do_IRQ |
| 1125 | => ret_from_intr |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1126 | |
| 1127 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1128 | This has some more changes. Preemption was disabled when an |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1129 | interrupt came in (notice the 'h'), and was enabled on exit. |
| 1130 | But we also see that interrupts have been disabled when entering |
| 1131 | the preempt off section and leaving it (the 'd'). We do not know if |
| 1132 | interrupts were enabled in the mean time or shortly after this |
| 1133 | was over. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1134 | |
| 1135 | # tracer: preemptoff |
| 1136 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1137 | # preemptoff latency trace v1.1.5 on 3.8.0-test+ |
| 1138 | # -------------------------------------------------------------------- |
| 1139 | # latency: 83 us, #241/241, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1140 | # ----------------- |
| 1141 | # | task: bash-1994 (uid:0 nice:0 policy:0 rt_prio:0) |
| 1142 | # ----------------- |
| 1143 | # => started at: wake_up_new_task |
| 1144 | # => ended at: task_rq_unlock |
| 1145 | # |
| 1146 | # |
| 1147 | # _------=> CPU# |
| 1148 | # / _-----=> irqs-off |
| 1149 | # | / _----=> need-resched |
| 1150 | # || / _---=> hardirq/softirq |
| 1151 | # ||| / _--=> preempt-depth |
| 1152 | # |||| / delay |
| 1153 | # cmd pid ||||| time | caller |
| 1154 | # \ / ||||| \ | / |
| 1155 | bash-1994 1d..1 0us : _raw_spin_lock_irqsave <-wake_up_new_task |
| 1156 | bash-1994 1d..1 0us : select_task_rq_fair <-select_task_rq |
| 1157 | bash-1994 1d..1 1us : __rcu_read_lock <-select_task_rq_fair |
| 1158 | bash-1994 1d..1 1us : source_load <-select_task_rq_fair |
| 1159 | bash-1994 1d..1 1us : source_load <-select_task_rq_fair |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1160 | [...] |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1161 | bash-1994 1d..1 12us : irq_enter <-smp_apic_timer_interrupt |
| 1162 | bash-1994 1d..1 12us : rcu_irq_enter <-irq_enter |
| 1163 | bash-1994 1d..1 13us : add_preempt_count <-irq_enter |
| 1164 | bash-1994 1d.h1 13us : exit_idle <-smp_apic_timer_interrupt |
| 1165 | bash-1994 1d.h1 13us : hrtimer_interrupt <-smp_apic_timer_interrupt |
| 1166 | bash-1994 1d.h1 13us : _raw_spin_lock <-hrtimer_interrupt |
| 1167 | bash-1994 1d.h1 14us : add_preempt_count <-_raw_spin_lock |
| 1168 | bash-1994 1d.h2 14us : ktime_get_update_offsets <-hrtimer_interrupt |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1169 | [...] |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1170 | bash-1994 1d.h1 35us : lapic_next_event <-clockevents_program_event |
| 1171 | bash-1994 1d.h1 35us : irq_exit <-smp_apic_timer_interrupt |
| 1172 | bash-1994 1d.h1 36us : sub_preempt_count <-irq_exit |
| 1173 | bash-1994 1d..2 36us : do_softirq <-irq_exit |
| 1174 | bash-1994 1d..2 36us : __do_softirq <-call_softirq |
| 1175 | bash-1994 1d..2 36us : __local_bh_disable <-__do_softirq |
| 1176 | bash-1994 1d.s2 37us : add_preempt_count <-_raw_spin_lock_irq |
| 1177 | bash-1994 1d.s3 38us : _raw_spin_unlock <-run_timer_softirq |
| 1178 | bash-1994 1d.s3 39us : sub_preempt_count <-_raw_spin_unlock |
| 1179 | bash-1994 1d.s2 39us : call_timer_fn <-run_timer_softirq |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1180 | [...] |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1181 | bash-1994 1dNs2 81us : cpu_needs_another_gp <-rcu_process_callbacks |
| 1182 | bash-1994 1dNs2 82us : __local_bh_enable <-__do_softirq |
| 1183 | bash-1994 1dNs2 82us : sub_preempt_count <-__local_bh_enable |
| 1184 | bash-1994 1dN.2 82us : idle_cpu <-irq_exit |
| 1185 | bash-1994 1dN.2 83us : rcu_irq_exit <-irq_exit |
| 1186 | bash-1994 1dN.2 83us : sub_preempt_count <-irq_exit |
| 1187 | bash-1994 1.N.1 84us : _raw_spin_unlock_irqrestore <-task_rq_unlock |
| 1188 | bash-1994 1.N.1 84us+: trace_preempt_on <-task_rq_unlock |
| 1189 | bash-1994 1.N.1 104us : <stack trace> |
| 1190 | => sub_preempt_count |
| 1191 | => _raw_spin_unlock_irqrestore |
| 1192 | => task_rq_unlock |
| 1193 | => wake_up_new_task |
| 1194 | => do_fork |
| 1195 | => sys_clone |
| 1196 | => stub_clone |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1197 | |
| 1198 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1199 | The above is an example of the preemptoff trace with |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1200 | function-trace set. Here we see that interrupts were not disabled |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1201 | the entire time. The irq_enter code lets us know that we entered |
| 1202 | an interrupt 'h'. Before that, the functions being traced still |
| 1203 | show that it is not in an interrupt, but we can see from the |
| 1204 | functions themselves that this is not the case. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1205 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1206 | preemptirqsoff |
| 1207 | -------------- |
| 1208 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1209 | Knowing the locations that have interrupts disabled or |
| 1210 | preemption disabled for the longest times is helpful. But |
| 1211 | sometimes we would like to know when either preemption and/or |
| 1212 | interrupts are disabled. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1213 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 1214 | Consider the following code: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1215 | |
| 1216 | local_irq_disable(); |
| 1217 | call_function_with_irqs_off(); |
| 1218 | preempt_disable(); |
| 1219 | call_function_with_irqs_and_preemption_off(); |
| 1220 | local_irq_enable(); |
| 1221 | call_function_with_preemption_off(); |
| 1222 | preempt_enable(); |
| 1223 | |
| 1224 | The irqsoff tracer will record the total length of |
| 1225 | call_function_with_irqs_off() and |
| 1226 | call_function_with_irqs_and_preemption_off(). |
| 1227 | |
| 1228 | The preemptoff tracer will record the total length of |
| 1229 | call_function_with_irqs_and_preemption_off() and |
| 1230 | call_function_with_preemption_off(). |
| 1231 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1232 | But neither will trace the time that interrupts and/or |
| 1233 | preemption is disabled. This total time is the time that we can |
| 1234 | not schedule. To record this time, use the preemptirqsoff |
| 1235 | tracer. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1236 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1237 | Again, using this trace is much like the irqsoff and preemptoff |
| 1238 | tracers. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1239 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1240 | # echo 0 > options/function-trace |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1241 | # echo preemptirqsoff > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1242 | # echo 1 > tracing_on |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1243 | # echo 0 > tracing_max_latency |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1244 | # ls -ltr |
| 1245 | [...] |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1246 | # echo 0 > tracing_on |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 1247 | # cat trace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1248 | # tracer: preemptirqsoff |
| 1249 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1250 | # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+ |
| 1251 | # -------------------------------------------------------------------- |
| 1252 | # latency: 100 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1253 | # ----------------- |
| 1254 | # | task: ls-2230 (uid:0 nice:0 policy:0 rt_prio:0) |
| 1255 | # ----------------- |
| 1256 | # => started at: ata_scsi_queuecmd |
| 1257 | # => ended at: ata_scsi_queuecmd |
| 1258 | # |
| 1259 | # |
| 1260 | # _------=> CPU# |
| 1261 | # / _-----=> irqs-off |
| 1262 | # | / _----=> need-resched |
| 1263 | # || / _---=> hardirq/softirq |
| 1264 | # ||| / _--=> preempt-depth |
| 1265 | # |||| / delay |
| 1266 | # cmd pid ||||| time | caller |
| 1267 | # \ / ||||| \ | / |
| 1268 | ls-2230 3d... 0us+: _raw_spin_lock_irqsave <-ata_scsi_queuecmd |
| 1269 | ls-2230 3...1 100us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd |
| 1270 | ls-2230 3...1 101us+: trace_preempt_on <-ata_scsi_queuecmd |
| 1271 | ls-2230 3...1 111us : <stack trace> |
| 1272 | => sub_preempt_count |
| 1273 | => _raw_spin_unlock_irqrestore |
| 1274 | => ata_scsi_queuecmd |
| 1275 | => scsi_dispatch_cmd |
| 1276 | => scsi_request_fn |
| 1277 | => __blk_run_queue_uncond |
| 1278 | => __blk_run_queue |
| 1279 | => blk_queue_bio |
| 1280 | => generic_make_request |
| 1281 | => submit_bio |
| 1282 | => submit_bh |
| 1283 | => ext3_bread |
| 1284 | => ext3_dir_bread |
| 1285 | => htree_dirblock_to_tree |
| 1286 | => ext3_htree_fill_tree |
| 1287 | => ext3_readdir |
| 1288 | => vfs_readdir |
| 1289 | => sys_getdents |
| 1290 | => system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1291 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1292 | |
| 1293 | The trace_hardirqs_off_thunk is called from assembly on x86 when |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1294 | interrupts are disabled in the assembly code. Without the |
| 1295 | function tracing, we do not know if interrupts were enabled |
| 1296 | within the preemption points. We do see that it started with |
| 1297 | preemption enabled. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1298 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1299 | Here is a trace with function-trace set: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1300 | |
| 1301 | # tracer: preemptirqsoff |
| 1302 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1303 | # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+ |
| 1304 | # -------------------------------------------------------------------- |
| 1305 | # latency: 161 us, #339/339, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1306 | # ----------------- |
| 1307 | # | task: ls-2269 (uid:0 nice:0 policy:0 rt_prio:0) |
| 1308 | # ----------------- |
| 1309 | # => started at: schedule |
| 1310 | # => ended at: mutex_unlock |
| 1311 | # |
| 1312 | # |
| 1313 | # _------=> CPU# |
| 1314 | # / _-----=> irqs-off |
| 1315 | # | / _----=> need-resched |
| 1316 | # || / _---=> hardirq/softirq |
| 1317 | # ||| / _--=> preempt-depth |
| 1318 | # |||| / delay |
| 1319 | # cmd pid ||||| time | caller |
| 1320 | # \ / ||||| \ | / |
| 1321 | kworker/-59 3...1 0us : __schedule <-schedule |
| 1322 | kworker/-59 3d..1 0us : rcu_preempt_qs <-rcu_note_context_switch |
| 1323 | kworker/-59 3d..1 1us : add_preempt_count <-_raw_spin_lock_irq |
| 1324 | kworker/-59 3d..2 1us : deactivate_task <-__schedule |
| 1325 | kworker/-59 3d..2 1us : dequeue_task <-deactivate_task |
| 1326 | kworker/-59 3d..2 2us : update_rq_clock <-dequeue_task |
| 1327 | kworker/-59 3d..2 2us : dequeue_task_fair <-dequeue_task |
| 1328 | kworker/-59 3d..2 2us : update_curr <-dequeue_task_fair |
| 1329 | kworker/-59 3d..2 2us : update_min_vruntime <-update_curr |
| 1330 | kworker/-59 3d..2 3us : cpuacct_charge <-update_curr |
| 1331 | kworker/-59 3d..2 3us : __rcu_read_lock <-cpuacct_charge |
| 1332 | kworker/-59 3d..2 3us : __rcu_read_unlock <-cpuacct_charge |
| 1333 | kworker/-59 3d..2 3us : update_cfs_rq_blocked_load <-dequeue_task_fair |
| 1334 | kworker/-59 3d..2 4us : clear_buddies <-dequeue_task_fair |
| 1335 | kworker/-59 3d..2 4us : account_entity_dequeue <-dequeue_task_fair |
| 1336 | kworker/-59 3d..2 4us : update_min_vruntime <-dequeue_task_fair |
| 1337 | kworker/-59 3d..2 4us : update_cfs_shares <-dequeue_task_fair |
| 1338 | kworker/-59 3d..2 5us : hrtick_update <-dequeue_task_fair |
| 1339 | kworker/-59 3d..2 5us : wq_worker_sleeping <-__schedule |
| 1340 | kworker/-59 3d..2 5us : kthread_data <-wq_worker_sleeping |
| 1341 | kworker/-59 3d..2 5us : put_prev_task_fair <-__schedule |
| 1342 | kworker/-59 3d..2 6us : pick_next_task_fair <-pick_next_task |
| 1343 | kworker/-59 3d..2 6us : clear_buddies <-pick_next_task_fair |
| 1344 | kworker/-59 3d..2 6us : set_next_entity <-pick_next_task_fair |
| 1345 | kworker/-59 3d..2 6us : update_stats_wait_end <-set_next_entity |
| 1346 | ls-2269 3d..2 7us : finish_task_switch <-__schedule |
| 1347 | ls-2269 3d..2 7us : _raw_spin_unlock_irq <-finish_task_switch |
| 1348 | ls-2269 3d..2 8us : do_IRQ <-ret_from_intr |
| 1349 | ls-2269 3d..2 8us : irq_enter <-do_IRQ |
| 1350 | ls-2269 3d..2 8us : rcu_irq_enter <-irq_enter |
| 1351 | ls-2269 3d..2 9us : add_preempt_count <-irq_enter |
| 1352 | ls-2269 3d.h2 9us : exit_idle <-do_IRQ |
| 1353 | [...] |
| 1354 | ls-2269 3d.h3 20us : sub_preempt_count <-_raw_spin_unlock |
| 1355 | ls-2269 3d.h2 20us : irq_exit <-do_IRQ |
| 1356 | ls-2269 3d.h2 21us : sub_preempt_count <-irq_exit |
| 1357 | ls-2269 3d..3 21us : do_softirq <-irq_exit |
| 1358 | ls-2269 3d..3 21us : __do_softirq <-call_softirq |
| 1359 | ls-2269 3d..3 21us+: __local_bh_disable <-__do_softirq |
| 1360 | ls-2269 3d.s4 29us : sub_preempt_count <-_local_bh_enable_ip |
| 1361 | ls-2269 3d.s5 29us : sub_preempt_count <-_local_bh_enable_ip |
| 1362 | ls-2269 3d.s5 31us : do_IRQ <-ret_from_intr |
| 1363 | ls-2269 3d.s5 31us : irq_enter <-do_IRQ |
| 1364 | ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter |
| 1365 | [...] |
| 1366 | ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter |
| 1367 | ls-2269 3d.s5 32us : add_preempt_count <-irq_enter |
| 1368 | ls-2269 3d.H5 32us : exit_idle <-do_IRQ |
| 1369 | ls-2269 3d.H5 32us : handle_irq <-do_IRQ |
| 1370 | ls-2269 3d.H5 32us : irq_to_desc <-handle_irq |
| 1371 | ls-2269 3d.H5 33us : handle_fasteoi_irq <-handle_irq |
| 1372 | [...] |
| 1373 | ls-2269 3d.s5 158us : _raw_spin_unlock_irqrestore <-rtl8139_poll |
| 1374 | ls-2269 3d.s3 158us : net_rps_action_and_irq_enable.isra.65 <-net_rx_action |
| 1375 | ls-2269 3d.s3 159us : __local_bh_enable <-__do_softirq |
| 1376 | ls-2269 3d.s3 159us : sub_preempt_count <-__local_bh_enable |
| 1377 | ls-2269 3d..3 159us : idle_cpu <-irq_exit |
| 1378 | ls-2269 3d..3 159us : rcu_irq_exit <-irq_exit |
| 1379 | ls-2269 3d..3 160us : sub_preempt_count <-irq_exit |
| 1380 | ls-2269 3d... 161us : __mutex_unlock_slowpath <-mutex_unlock |
| 1381 | ls-2269 3d... 162us+: trace_hardirqs_on <-mutex_unlock |
| 1382 | ls-2269 3d... 186us : <stack trace> |
| 1383 | => __mutex_unlock_slowpath |
| 1384 | => mutex_unlock |
| 1385 | => process_output |
| 1386 | => n_tty_write |
| 1387 | => tty_write |
| 1388 | => vfs_write |
| 1389 | => sys_write |
| 1390 | => system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1391 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1392 | This is an interesting trace. It started with kworker running and |
| 1393 | scheduling out and ls taking over. But as soon as ls released the |
| 1394 | rq lock and enabled interrupts (but not preemption) an interrupt |
| 1395 | triggered. When the interrupt finished, it started running softirqs. |
| 1396 | But while the softirq was running, another interrupt triggered. |
| 1397 | When an interrupt is running inside a softirq, the annotation is 'H'. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1398 | |
| 1399 | |
| 1400 | wakeup |
| 1401 | ------ |
| 1402 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1403 | One common case that people are interested in tracing is the |
| 1404 | time it takes for a task that is woken to actually wake up. |
| 1405 | Now for non Real-Time tasks, this can be arbitrary. But tracing |
| 1406 | it none the less can be interesting. |
| 1407 | |
| 1408 | Without function tracing: |
| 1409 | |
| 1410 | # echo 0 > options/function-trace |
| 1411 | # echo wakeup > current_tracer |
| 1412 | # echo 1 > tracing_on |
| 1413 | # echo 0 > tracing_max_latency |
| 1414 | # chrt -f 5 sleep 1 |
| 1415 | # echo 0 > tracing_on |
| 1416 | # cat trace |
| 1417 | # tracer: wakeup |
| 1418 | # |
| 1419 | # wakeup latency trace v1.1.5 on 3.8.0-test+ |
| 1420 | # -------------------------------------------------------------------- |
| 1421 | # latency: 15 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1422 | # ----------------- |
| 1423 | # | task: kworker/3:1H-312 (uid:0 nice:-20 policy:0 rt_prio:0) |
| 1424 | # ----------------- |
| 1425 | # |
| 1426 | # _------=> CPU# |
| 1427 | # / _-----=> irqs-off |
| 1428 | # | / _----=> need-resched |
| 1429 | # || / _---=> hardirq/softirq |
| 1430 | # ||| / _--=> preempt-depth |
| 1431 | # |||| / delay |
| 1432 | # cmd pid ||||| time | caller |
| 1433 | # \ / ||||| \ | / |
| 1434 | <idle>-0 3dNs7 0us : 0:120:R + [003] 312:100:R kworker/3:1H |
| 1435 | <idle>-0 3dNs7 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up |
| 1436 | <idle>-0 3d..3 15us : __schedule <-schedule |
| 1437 | <idle>-0 3d..3 15us : 0:120:R ==> [003] 312:100:R kworker/3:1H |
| 1438 | |
| 1439 | The tracer only traces the highest priority task in the system |
| 1440 | to avoid tracing the normal circumstances. Here we see that |
| 1441 | the kworker with a nice priority of -20 (not very nice), took |
| 1442 | just 15 microseconds from the time it woke up, to the time it |
| 1443 | ran. |
| 1444 | |
| 1445 | Non Real-Time tasks are not that interesting. A more interesting |
| 1446 | trace is to concentrate only on Real-Time tasks. |
| 1447 | |
| 1448 | wakeup_rt |
| 1449 | --------- |
| 1450 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1451 | In a Real-Time environment it is very important to know the |
| 1452 | wakeup time it takes for the highest priority task that is woken |
| 1453 | up to the time that it executes. This is also known as "schedule |
| 1454 | latency". I stress the point that this is about RT tasks. It is |
| 1455 | also important to know the scheduling latency of non-RT tasks, |
| 1456 | but the average schedule latency is better for non-RT tasks. |
| 1457 | Tools like LatencyTop are more appropriate for such |
| 1458 | measurements. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1459 | |
Steven Rostedt | a41eeba | 2008-07-14 16:41:12 -0400 | [diff] [blame] | 1460 | Real-Time environments are interested in the worst case latency. |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1461 | That is the longest latency it takes for something to happen, |
| 1462 | and not the average. We can have a very fast scheduler that may |
| 1463 | only have a large latency once in a while, but that would not |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1464 | work well with Real-Time tasks. The wakeup_rt tracer was designed |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1465 | to record the worst case wakeups of RT tasks. Non-RT tasks are |
| 1466 | not recorded because the tracer only records one worst case and |
| 1467 | tracing non-RT tasks that are unpredictable will overwrite the |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1468 | worst case latency of RT tasks (just run the normal wakeup |
| 1469 | tracer for a while to see that effect). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1470 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1471 | Since this tracer only deals with RT tasks, we will run this |
| 1472 | slightly differently than we did with the previous tracers. |
| 1473 | Instead of performing an 'ls', we will run 'sleep 1' under |
| 1474 | 'chrt' which changes the priority of the task. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1475 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1476 | # echo 0 > options/function-trace |
| 1477 | # echo wakeup_rt > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1478 | # echo 1 > tracing_on |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1479 | # echo 0 > tracing_max_latency |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1480 | # chrt -f 5 sleep 1 |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1481 | # echo 0 > tracing_on |
Albin Tonnerre | 4a88d44 | 2009-08-31 22:40:08 +0200 | [diff] [blame] | 1482 | # cat trace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1483 | # tracer: wakeup |
| 1484 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1485 | # tracer: wakeup_rt |
| 1486 | # |
| 1487 | # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ |
| 1488 | # -------------------------------------------------------------------- |
| 1489 | # latency: 5 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1490 | # ----------------- |
| 1491 | # | task: sleep-2389 (uid:0 nice:0 policy:1 rt_prio:5) |
| 1492 | # ----------------- |
| 1493 | # |
| 1494 | # _------=> CPU# |
| 1495 | # / _-----=> irqs-off |
| 1496 | # | / _----=> need-resched |
| 1497 | # || / _---=> hardirq/softirq |
| 1498 | # ||| / _--=> preempt-depth |
| 1499 | # |||| / delay |
| 1500 | # cmd pid ||||| time | caller |
| 1501 | # \ / ||||| \ | / |
| 1502 | <idle>-0 3d.h4 0us : 0:120:R + [003] 2389: 94:R sleep |
| 1503 | <idle>-0 3d.h4 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up |
| 1504 | <idle>-0 3d..3 5us : __schedule <-schedule |
| 1505 | <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1506 | |
| 1507 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1508 | Running this on an idle system, we see that it only took 5 microseconds |
| 1509 | to perform the task switch. Note, since the trace point in the schedule |
| 1510 | is before the actual "switch", we stop the tracing when the recorded task |
| 1511 | is about to schedule in. This may change if we add a new marker at the |
| 1512 | end of the scheduler. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1513 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1514 | Notice that the recorded task is 'sleep' with the PID of 2389 |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1515 | and it has an rt_prio of 5. This priority is user-space priority |
| 1516 | and not the internal kernel priority. The policy is 1 for |
| 1517 | SCHED_FIFO and 2 for SCHED_RR. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1518 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1519 | Note, that the trace data shows the internal priority (99 - rtprio). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1520 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1521 | <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep |
| 1522 | |
| 1523 | The 0:120:R means idle was running with a nice priority of 0 (120 - 20) |
| 1524 | and in the running state 'R'. The sleep task was scheduled in with |
| 1525 | 2389: 94:R. That is the priority is the kernel rtprio (99 - 5 = 94) |
| 1526 | and it too is in the running state. |
| 1527 | |
| 1528 | Doing the same with chrt -r 5 and function-trace set. |
| 1529 | |
| 1530 | echo 1 > options/function-trace |
| 1531 | |
| 1532 | # tracer: wakeup_rt |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1533 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1534 | # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ |
| 1535 | # -------------------------------------------------------------------- |
| 1536 | # latency: 29 us, #85/85, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1537 | # ----------------- |
| 1538 | # | task: sleep-2448 (uid:0 nice:0 policy:1 rt_prio:5) |
| 1539 | # ----------------- |
| 1540 | # |
| 1541 | # _------=> CPU# |
| 1542 | # / _-----=> irqs-off |
| 1543 | # | / _----=> need-resched |
| 1544 | # || / _---=> hardirq/softirq |
| 1545 | # ||| / _--=> preempt-depth |
| 1546 | # |||| / delay |
| 1547 | # cmd pid ||||| time | caller |
| 1548 | # \ / ||||| \ | / |
| 1549 | <idle>-0 3d.h4 1us+: 0:120:R + [003] 2448: 94:R sleep |
| 1550 | <idle>-0 3d.h4 2us : ttwu_do_activate.constprop.87 <-try_to_wake_up |
| 1551 | <idle>-0 3d.h3 3us : check_preempt_curr <-ttwu_do_wakeup |
Kirill Tkhai | 8875125 | 2014-06-29 00:03:57 +0400 | [diff] [blame] | 1552 | <idle>-0 3d.h3 3us : resched_curr <-check_preempt_curr |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1553 | <idle>-0 3dNh3 4us : task_woken_rt <-ttwu_do_wakeup |
| 1554 | <idle>-0 3dNh3 4us : _raw_spin_unlock <-try_to_wake_up |
| 1555 | <idle>-0 3dNh3 4us : sub_preempt_count <-_raw_spin_unlock |
| 1556 | <idle>-0 3dNh2 5us : ttwu_stat <-try_to_wake_up |
| 1557 | <idle>-0 3dNh2 5us : _raw_spin_unlock_irqrestore <-try_to_wake_up |
| 1558 | <idle>-0 3dNh2 6us : sub_preempt_count <-_raw_spin_unlock_irqrestore |
| 1559 | <idle>-0 3dNh1 6us : _raw_spin_lock <-__run_hrtimer |
| 1560 | <idle>-0 3dNh1 6us : add_preempt_count <-_raw_spin_lock |
| 1561 | <idle>-0 3dNh2 7us : _raw_spin_unlock <-hrtimer_interrupt |
| 1562 | <idle>-0 3dNh2 7us : sub_preempt_count <-_raw_spin_unlock |
| 1563 | <idle>-0 3dNh1 7us : tick_program_event <-hrtimer_interrupt |
| 1564 | <idle>-0 3dNh1 7us : clockevents_program_event <-tick_program_event |
| 1565 | <idle>-0 3dNh1 8us : ktime_get <-clockevents_program_event |
| 1566 | <idle>-0 3dNh1 8us : lapic_next_event <-clockevents_program_event |
| 1567 | <idle>-0 3dNh1 8us : irq_exit <-smp_apic_timer_interrupt |
| 1568 | <idle>-0 3dNh1 9us : sub_preempt_count <-irq_exit |
| 1569 | <idle>-0 3dN.2 9us : idle_cpu <-irq_exit |
| 1570 | <idle>-0 3dN.2 9us : rcu_irq_exit <-irq_exit |
| 1571 | <idle>-0 3dN.2 10us : rcu_eqs_enter_common.isra.45 <-rcu_irq_exit |
| 1572 | <idle>-0 3dN.2 10us : sub_preempt_count <-irq_exit |
| 1573 | <idle>-0 3.N.1 11us : rcu_idle_exit <-cpu_idle |
| 1574 | <idle>-0 3dN.1 11us : rcu_eqs_exit_common.isra.43 <-rcu_idle_exit |
| 1575 | <idle>-0 3.N.1 11us : tick_nohz_idle_exit <-cpu_idle |
| 1576 | <idle>-0 3dN.1 12us : menu_hrtimer_cancel <-tick_nohz_idle_exit |
| 1577 | <idle>-0 3dN.1 12us : ktime_get <-tick_nohz_idle_exit |
| 1578 | <idle>-0 3dN.1 12us : tick_do_update_jiffies64 <-tick_nohz_idle_exit |
Frederic Weisbecker | cee1afc | 2016-04-13 15:56:50 +0200 | [diff] [blame] | 1579 | <idle>-0 3dN.1 13us : cpu_load_update_nohz <-tick_nohz_idle_exit |
| 1580 | <idle>-0 3dN.1 13us : _raw_spin_lock <-cpu_load_update_nohz |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1581 | <idle>-0 3dN.1 13us : add_preempt_count <-_raw_spin_lock |
Frederic Weisbecker | cee1afc | 2016-04-13 15:56:50 +0200 | [diff] [blame] | 1582 | <idle>-0 3dN.2 13us : __cpu_load_update <-cpu_load_update_nohz |
| 1583 | <idle>-0 3dN.2 14us : sched_avg_update <-__cpu_load_update |
| 1584 | <idle>-0 3dN.2 14us : _raw_spin_unlock <-cpu_load_update_nohz |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1585 | <idle>-0 3dN.2 14us : sub_preempt_count <-_raw_spin_unlock |
| 1586 | <idle>-0 3dN.1 15us : calc_load_exit_idle <-tick_nohz_idle_exit |
| 1587 | <idle>-0 3dN.1 15us : touch_softlockup_watchdog <-tick_nohz_idle_exit |
| 1588 | <idle>-0 3dN.1 15us : hrtimer_cancel <-tick_nohz_idle_exit |
| 1589 | <idle>-0 3dN.1 15us : hrtimer_try_to_cancel <-hrtimer_cancel |
| 1590 | <idle>-0 3dN.1 16us : lock_hrtimer_base.isra.18 <-hrtimer_try_to_cancel |
| 1591 | <idle>-0 3dN.1 16us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18 |
| 1592 | <idle>-0 3dN.1 16us : add_preempt_count <-_raw_spin_lock_irqsave |
| 1593 | <idle>-0 3dN.2 17us : __remove_hrtimer <-remove_hrtimer.part.16 |
| 1594 | <idle>-0 3dN.2 17us : hrtimer_force_reprogram <-__remove_hrtimer |
| 1595 | <idle>-0 3dN.2 17us : tick_program_event <-hrtimer_force_reprogram |
| 1596 | <idle>-0 3dN.2 18us : clockevents_program_event <-tick_program_event |
| 1597 | <idle>-0 3dN.2 18us : ktime_get <-clockevents_program_event |
| 1598 | <idle>-0 3dN.2 18us : lapic_next_event <-clockevents_program_event |
| 1599 | <idle>-0 3dN.2 19us : _raw_spin_unlock_irqrestore <-hrtimer_try_to_cancel |
| 1600 | <idle>-0 3dN.2 19us : sub_preempt_count <-_raw_spin_unlock_irqrestore |
| 1601 | <idle>-0 3dN.1 19us : hrtimer_forward <-tick_nohz_idle_exit |
| 1602 | <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward |
| 1603 | <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward |
| 1604 | <idle>-0 3dN.1 20us : hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11 |
| 1605 | <idle>-0 3dN.1 20us : __hrtimer_start_range_ns <-hrtimer_start_range_ns |
| 1606 | <idle>-0 3dN.1 21us : lock_hrtimer_base.isra.18 <-__hrtimer_start_range_ns |
| 1607 | <idle>-0 3dN.1 21us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18 |
| 1608 | <idle>-0 3dN.1 21us : add_preempt_count <-_raw_spin_lock_irqsave |
| 1609 | <idle>-0 3dN.2 22us : ktime_add_safe <-__hrtimer_start_range_ns |
| 1610 | <idle>-0 3dN.2 22us : enqueue_hrtimer <-__hrtimer_start_range_ns |
| 1611 | <idle>-0 3dN.2 22us : tick_program_event <-__hrtimer_start_range_ns |
| 1612 | <idle>-0 3dN.2 23us : clockevents_program_event <-tick_program_event |
| 1613 | <idle>-0 3dN.2 23us : ktime_get <-clockevents_program_event |
| 1614 | <idle>-0 3dN.2 23us : lapic_next_event <-clockevents_program_event |
| 1615 | <idle>-0 3dN.2 24us : _raw_spin_unlock_irqrestore <-__hrtimer_start_range_ns |
| 1616 | <idle>-0 3dN.2 24us : sub_preempt_count <-_raw_spin_unlock_irqrestore |
| 1617 | <idle>-0 3dN.1 24us : account_idle_ticks <-tick_nohz_idle_exit |
| 1618 | <idle>-0 3dN.1 24us : account_idle_time <-account_idle_ticks |
| 1619 | <idle>-0 3.N.1 25us : sub_preempt_count <-cpu_idle |
| 1620 | <idle>-0 3.N.. 25us : schedule <-cpu_idle |
| 1621 | <idle>-0 3.N.. 25us : __schedule <-preempt_schedule |
| 1622 | <idle>-0 3.N.. 26us : add_preempt_count <-__schedule |
| 1623 | <idle>-0 3.N.1 26us : rcu_note_context_switch <-__schedule |
| 1624 | <idle>-0 3.N.1 26us : rcu_sched_qs <-rcu_note_context_switch |
| 1625 | <idle>-0 3dN.1 27us : rcu_preempt_qs <-rcu_note_context_switch |
| 1626 | <idle>-0 3.N.1 27us : _raw_spin_lock_irq <-__schedule |
| 1627 | <idle>-0 3dN.1 27us : add_preempt_count <-_raw_spin_lock_irq |
| 1628 | <idle>-0 3dN.2 28us : put_prev_task_idle <-__schedule |
| 1629 | <idle>-0 3dN.2 28us : pick_next_task_stop <-pick_next_task |
| 1630 | <idle>-0 3dN.2 28us : pick_next_task_rt <-pick_next_task |
| 1631 | <idle>-0 3dN.2 29us : dequeue_pushable_task <-pick_next_task_rt |
| 1632 | <idle>-0 3d..3 29us : __schedule <-preempt_schedule |
| 1633 | <idle>-0 3d..3 30us : 0:120:R ==> [003] 2448: 94:R sleep |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1634 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1635 | This isn't that big of a trace, even with function tracing enabled, |
| 1636 | so I included the entire trace. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1637 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1638 | The interrupt went off while when the system was idle. Somewhere |
| 1639 | before task_woken_rt() was called, the NEED_RESCHED flag was set, |
| 1640 | this is indicated by the first occurrence of the 'N' flag. |
| 1641 | |
| 1642 | Latency tracing and events |
| 1643 | -------------------------- |
| 1644 | As function tracing can induce a much larger latency, but without |
| 1645 | seeing what happens within the latency it is hard to know what |
| 1646 | caused it. There is a middle ground, and that is with enabling |
| 1647 | events. |
| 1648 | |
| 1649 | # echo 0 > options/function-trace |
| 1650 | # echo wakeup_rt > current_tracer |
| 1651 | # echo 1 > events/enable |
| 1652 | # echo 1 > tracing_on |
| 1653 | # echo 0 > tracing_max_latency |
| 1654 | # chrt -f 5 sleep 1 |
| 1655 | # echo 0 > tracing_on |
| 1656 | # cat trace |
| 1657 | # tracer: wakeup_rt |
| 1658 | # |
| 1659 | # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ |
| 1660 | # -------------------------------------------------------------------- |
| 1661 | # latency: 6 us, #12/12, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) |
| 1662 | # ----------------- |
| 1663 | # | task: sleep-5882 (uid:0 nice:0 policy:1 rt_prio:5) |
| 1664 | # ----------------- |
| 1665 | # |
| 1666 | # _------=> CPU# |
| 1667 | # / _-----=> irqs-off |
| 1668 | # | / _----=> need-resched |
| 1669 | # || / _---=> hardirq/softirq |
| 1670 | # ||| / _--=> preempt-depth |
| 1671 | # |||| / delay |
| 1672 | # cmd pid ||||| time | caller |
| 1673 | # \ / ||||| \ | / |
| 1674 | <idle>-0 2d.h4 0us : 0:120:R + [002] 5882: 94:R sleep |
| 1675 | <idle>-0 2d.h4 0us : ttwu_do_activate.constprop.87 <-try_to_wake_up |
| 1676 | <idle>-0 2d.h4 1us : sched_wakeup: comm=sleep pid=5882 prio=94 success=1 target_cpu=002 |
| 1677 | <idle>-0 2dNh2 1us : hrtimer_expire_exit: hrtimer=ffff88007796feb8 |
| 1678 | <idle>-0 2.N.2 2us : power_end: cpu_id=2 |
| 1679 | <idle>-0 2.N.2 3us : cpu_idle: state=4294967295 cpu_id=2 |
| 1680 | <idle>-0 2dN.3 4us : hrtimer_cancel: hrtimer=ffff88007d50d5e0 |
| 1681 | <idle>-0 2dN.3 4us : hrtimer_start: hrtimer=ffff88007d50d5e0 function=tick_sched_timer expires=34311211000000 softexpires=34311211000000 |
| 1682 | <idle>-0 2.N.2 5us : rcu_utilization: Start context switch |
| 1683 | <idle>-0 2.N.2 5us : rcu_utilization: End context switch |
| 1684 | <idle>-0 2d..3 6us : __schedule <-schedule |
| 1685 | <idle>-0 2d..3 6us : 0:120:R ==> [002] 5882: 94:R sleep |
| 1686 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1687 | |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 1688 | function |
| 1689 | -------- |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1690 | |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 1691 | This tracer is the function tracer. Enabling the function tracer |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1692 | can be done from the debug file system. Make sure the |
| 1693 | ftrace_enabled is set; otherwise this tracer is a nop. |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1694 | See the "ftrace_enabled" section below. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1695 | |
| 1696 | # sysctl kernel.ftrace_enabled=1 |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1697 | # echo function > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1698 | # echo 1 > tracing_on |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1699 | # usleep 1 |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1700 | # echo 0 > tracing_on |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1701 | # cat trace |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 1702 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1703 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1704 | # entries-in-buffer/entries-written: 24799/24799 #P:4 |
| 1705 | # |
| 1706 | # _-----=> irqs-off |
| 1707 | # / _----=> need-resched |
| 1708 | # | / _---=> hardirq/softirq |
| 1709 | # || / _--=> preempt-depth |
| 1710 | # ||| / delay |
| 1711 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 1712 | # | | | |||| | | |
| 1713 | bash-1994 [002] .... 3082.063030: mutex_unlock <-rb_simple_write |
| 1714 | bash-1994 [002] .... 3082.063031: __mutex_unlock_slowpath <-mutex_unlock |
| 1715 | bash-1994 [002] .... 3082.063031: __fsnotify_parent <-fsnotify_modify |
| 1716 | bash-1994 [002] .... 3082.063032: fsnotify <-fsnotify_modify |
| 1717 | bash-1994 [002] .... 3082.063032: __srcu_read_lock <-fsnotify |
| 1718 | bash-1994 [002] .... 3082.063032: add_preempt_count <-__srcu_read_lock |
| 1719 | bash-1994 [002] ...1 3082.063032: sub_preempt_count <-__srcu_read_lock |
| 1720 | bash-1994 [002] .... 3082.063033: __srcu_read_unlock <-fsnotify |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1721 | [...] |
| 1722 | |
| 1723 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1724 | Note: function tracer uses ring buffers to store the above |
| 1725 | entries. The newest data may overwrite the oldest data. |
| 1726 | Sometimes using echo to stop the trace is not sufficient because |
| 1727 | the tracing could have overwritten the data that you wanted to |
| 1728 | record. For this reason, it is sometimes better to disable |
| 1729 | tracing directly from a program. This allows you to stop the |
| 1730 | tracing at the point that you hit the part that you are |
| 1731 | interested in. To disable the tracing directly from a C program, |
| 1732 | something like following code snippet can be used: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1733 | |
| 1734 | int trace_fd; |
| 1735 | [...] |
| 1736 | int main(int argc, char *argv[]) { |
| 1737 | [...] |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 1738 | trace_fd = open(tracing_file("tracing_on"), O_WRONLY); |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1739 | [...] |
| 1740 | if (condition_hit()) { |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 1741 | write(trace_fd, "0", 1); |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 1742 | } |
| 1743 | [...] |
| 1744 | } |
| 1745 | |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1746 | |
| 1747 | Single thread tracing |
| 1748 | --------------------- |
| 1749 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1750 | By writing into set_ftrace_pid you can trace a |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1751 | single thread. For example: |
| 1752 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1753 | # cat set_ftrace_pid |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1754 | no pid |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1755 | # echo 3111 > set_ftrace_pid |
| 1756 | # cat set_ftrace_pid |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1757 | 3111 |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1758 | # echo function > current_tracer |
| 1759 | # cat trace | head |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1760 | # tracer: function |
| 1761 | # |
| 1762 | # TASK-PID CPU# TIMESTAMP FUNCTION |
| 1763 | # | | | | | |
| 1764 | yum-updatesd-3111 [003] 1637.254676: finish_task_switch <-thread_return |
| 1765 | yum-updatesd-3111 [003] 1637.254681: hrtimer_cancel <-schedule_hrtimeout_range |
| 1766 | yum-updatesd-3111 [003] 1637.254682: hrtimer_try_to_cancel <-hrtimer_cancel |
| 1767 | yum-updatesd-3111 [003] 1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel |
| 1768 | yum-updatesd-3111 [003] 1637.254685: fget_light <-do_sys_poll |
| 1769 | yum-updatesd-3111 [003] 1637.254686: pipe_poll <-do_sys_poll |
Wang Long | 52e6892 | 2015-02-05 05:16:14 +0000 | [diff] [blame] | 1770 | # echo > set_ftrace_pid |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1771 | # cat trace |head |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1772 | # tracer: function |
| 1773 | # |
| 1774 | # TASK-PID CPU# TIMESTAMP FUNCTION |
| 1775 | # | | | | | |
| 1776 | ##### CPU 3 buffer started #### |
| 1777 | yum-updatesd-3111 [003] 1701.957688: free_poll_entry <-poll_freewait |
| 1778 | yum-updatesd-3111 [003] 1701.957689: remove_wait_queue <-free_poll_entry |
| 1779 | yum-updatesd-3111 [003] 1701.957691: fput <-free_poll_entry |
| 1780 | yum-updatesd-3111 [003] 1701.957692: audit_syscall_exit <-sysret_audit |
| 1781 | yum-updatesd-3111 [003] 1701.957693: path_put <-audit_syscall_exit |
| 1782 | |
| 1783 | If you want to trace a function when executing, you could use |
| 1784 | something like this simple program: |
| 1785 | |
| 1786 | #include <stdio.h> |
| 1787 | #include <stdlib.h> |
| 1788 | #include <sys/types.h> |
| 1789 | #include <sys/stat.h> |
| 1790 | #include <fcntl.h> |
| 1791 | #include <unistd.h> |
Jiri Olsa | 67b394f | 2009-10-23 19:36:18 -0400 | [diff] [blame] | 1792 | #include <string.h> |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1793 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1794 | #define _STR(x) #x |
| 1795 | #define STR(x) _STR(x) |
| 1796 | #define MAX_PATH 256 |
| 1797 | |
| 1798 | const char *find_debugfs(void) |
| 1799 | { |
| 1800 | static char debugfs[MAX_PATH+1]; |
| 1801 | static int debugfs_found; |
| 1802 | char type[100]; |
| 1803 | FILE *fp; |
| 1804 | |
| 1805 | if (debugfs_found) |
| 1806 | return debugfs; |
| 1807 | |
| 1808 | if ((fp = fopen("/proc/mounts","r")) == NULL) { |
| 1809 | perror("/proc/mounts"); |
| 1810 | return NULL; |
| 1811 | } |
| 1812 | |
| 1813 | while (fscanf(fp, "%*s %" |
| 1814 | STR(MAX_PATH) |
| 1815 | "s %99s %*s %*d %*d\n", |
| 1816 | debugfs, type) == 2) { |
| 1817 | if (strcmp(type, "debugfs") == 0) |
| 1818 | break; |
| 1819 | } |
| 1820 | fclose(fp); |
| 1821 | |
| 1822 | if (strcmp(type, "debugfs") != 0) { |
| 1823 | fprintf(stderr, "debugfs not mounted"); |
| 1824 | return NULL; |
| 1825 | } |
| 1826 | |
Jiri Olsa | 67b394f | 2009-10-23 19:36:18 -0400 | [diff] [blame] | 1827 | strcat(debugfs, "/tracing/"); |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1828 | debugfs_found = 1; |
| 1829 | |
| 1830 | return debugfs; |
| 1831 | } |
| 1832 | |
| 1833 | const char *tracing_file(const char *file_name) |
| 1834 | { |
| 1835 | static char trace_file[MAX_PATH+1]; |
| 1836 | snprintf(trace_file, MAX_PATH, "%s/%s", find_debugfs(), file_name); |
| 1837 | return trace_file; |
| 1838 | } |
| 1839 | |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1840 | int main (int argc, char **argv) |
| 1841 | { |
| 1842 | if (argc < 1) |
| 1843 | exit(-1); |
| 1844 | |
| 1845 | if (fork() > 0) { |
| 1846 | int fd, ffd; |
| 1847 | char line[64]; |
| 1848 | int s; |
| 1849 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1850 | ffd = open(tracing_file("current_tracer"), O_WRONLY); |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1851 | if (ffd < 0) |
| 1852 | exit(-1); |
| 1853 | write(ffd, "nop", 3); |
| 1854 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1855 | fd = open(tracing_file("set_ftrace_pid"), O_WRONLY); |
Steven Rostedt | df4fc31 | 2008-11-26 00:16:23 -0500 | [diff] [blame] | 1856 | s = sprintf(line, "%d\n", getpid()); |
| 1857 | write(fd, line, s); |
| 1858 | |
| 1859 | write(ffd, "function", 8); |
| 1860 | |
| 1861 | close(fd); |
| 1862 | close(ffd); |
| 1863 | |
| 1864 | execvp(argv[1], argv+1); |
| 1865 | } |
| 1866 | |
| 1867 | return 0; |
| 1868 | } |
| 1869 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1870 | Or this simple script! |
Markus Metzger | e2ea539 | 2009-01-19 10:35:58 +0100 | [diff] [blame] | 1871 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1872 | ------ |
| 1873 | #!/bin/bash |
Markus Metzger | e2ea539 | 2009-01-19 10:35:58 +0100 | [diff] [blame] | 1874 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 1875 | debugfs=`sed -ne 's/^debugfs \(.*\) debugfs.*/\1/p' /proc/mounts` |
| 1876 | echo nop > $debugfs/tracing/current_tracer |
| 1877 | echo 0 > $debugfs/tracing/tracing_on |
| 1878 | echo $$ > $debugfs/tracing/set_ftrace_pid |
| 1879 | echo function > $debugfs/tracing/current_tracer |
| 1880 | echo 1 > $debugfs/tracing/tracing_on |
| 1881 | exec "$@" |
| 1882 | ------ |
Markus Metzger | e2ea539 | 2009-01-19 10:35:58 +0100 | [diff] [blame] | 1883 | |
| 1884 | |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1885 | function graph tracer |
| 1886 | --------------------------- |
| 1887 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1888 | This tracer is similar to the function tracer except that it |
| 1889 | probes a function on its entry and its exit. This is done by |
| 1890 | using a dynamically allocated stack of return addresses in each |
| 1891 | task_struct. On function entry the tracer overwrites the return |
| 1892 | address of each function traced to set a custom probe. Thus the |
| 1893 | original return address is stored on the stack of return address |
| 1894 | in the task_struct. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1895 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1896 | Probing on both ends of a function leads to special features |
| 1897 | such as: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1898 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1899 | - measure of a function's time execution |
| 1900 | - having a reliable call stack to draw function calls graph |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1901 | |
| 1902 | This tracer is useful in several situations: |
| 1903 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1904 | - you want to find the reason of a strange kernel behavior and |
| 1905 | need to see what happens in detail on any areas (or specific |
| 1906 | ones). |
| 1907 | |
| 1908 | - you are experiencing weird latencies but it's difficult to |
| 1909 | find its origin. |
| 1910 | |
| 1911 | - you want to find quickly which path is taken by a specific |
| 1912 | function |
| 1913 | |
| 1914 | - you just want to peek inside a working kernel and want to see |
| 1915 | what happens there. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1916 | |
| 1917 | # tracer: function_graph |
| 1918 | # |
| 1919 | # CPU DURATION FUNCTION CALLS |
| 1920 | # | | | | | | | |
| 1921 | |
| 1922 | 0) | sys_open() { |
| 1923 | 0) | do_sys_open() { |
| 1924 | 0) | getname() { |
| 1925 | 0) | kmem_cache_alloc() { |
| 1926 | 0) 1.382 us | __might_sleep(); |
| 1927 | 0) 2.478 us | } |
| 1928 | 0) | strncpy_from_user() { |
| 1929 | 0) | might_fault() { |
| 1930 | 0) 1.389 us | __might_sleep(); |
| 1931 | 0) 2.553 us | } |
| 1932 | 0) 3.807 us | } |
| 1933 | 0) 7.876 us | } |
| 1934 | 0) | alloc_fd() { |
| 1935 | 0) 0.668 us | _spin_lock(); |
| 1936 | 0) 0.570 us | expand_files(); |
| 1937 | 0) 0.586 us | _spin_unlock(); |
| 1938 | |
| 1939 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1940 | There are several columns that can be dynamically |
| 1941 | enabled/disabled. You can use every combination of options you |
| 1942 | want, depending on your needs. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1943 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1944 | - The cpu number on which the function executed is default |
| 1945 | enabled. It is sometimes better to only trace one cpu (see |
| 1946 | tracing_cpu_mask file) or you might sometimes see unordered |
| 1947 | function calls while cpu tracing switch. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1948 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1949 | hide: echo nofuncgraph-cpu > trace_options |
| 1950 | show: echo funcgraph-cpu > trace_options |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1951 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1952 | - The duration (function's time of execution) is displayed on |
| 1953 | the closing bracket line of a function or on the same line |
| 1954 | than the current function in case of a leaf one. It is default |
| 1955 | enabled. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1956 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1957 | hide: echo nofuncgraph-duration > trace_options |
| 1958 | show: echo funcgraph-duration > trace_options |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1959 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 1960 | - The overhead field precedes the duration field in case of |
| 1961 | reached duration thresholds. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1962 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 1963 | hide: echo nofuncgraph-overhead > trace_options |
| 1964 | show: echo funcgraph-overhead > trace_options |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1965 | depends on: funcgraph-duration |
| 1966 | |
| 1967 | ie: |
| 1968 | |
Jungseok Lee | b838e1d | 2015-07-11 14:51:40 +0000 | [diff] [blame] | 1969 | 3) # 1837.709 us | } /* __switch_to */ |
| 1970 | 3) | finish_task_switch() { |
| 1971 | 3) 0.313 us | _raw_spin_unlock_irq(); |
| 1972 | 3) 3.177 us | } |
| 1973 | 3) # 1889.063 us | } /* __schedule */ |
| 1974 | 3) ! 140.417 us | } /* __schedule */ |
| 1975 | 3) # 2034.948 us | } /* schedule */ |
| 1976 | 3) * 33998.59 us | } /* schedule_preempt_disabled */ |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 1977 | |
| 1978 | [...] |
| 1979 | |
Jungseok Lee | b838e1d | 2015-07-11 14:51:40 +0000 | [diff] [blame] | 1980 | 1) 0.260 us | msecs_to_jiffies(); |
| 1981 | 1) 0.313 us | __rcu_read_unlock(); |
| 1982 | 1) + 61.770 us | } |
| 1983 | 1) + 64.479 us | } |
| 1984 | 1) 0.313 us | rcu_bh_qs(); |
| 1985 | 1) 0.313 us | __local_bh_enable(); |
| 1986 | 1) ! 217.240 us | } |
| 1987 | 1) 0.365 us | idle_cpu(); |
| 1988 | 1) | rcu_irq_exit() { |
| 1989 | 1) 0.417 us | rcu_eqs_enter_common.isra.47(); |
| 1990 | 1) 3.125 us | } |
| 1991 | 1) ! 227.812 us | } |
| 1992 | 1) ! 457.395 us | } |
| 1993 | 1) @ 119760.2 us | } |
| 1994 | |
| 1995 | [...] |
| 1996 | |
| 1997 | 2) | handle_IPI() { |
| 1998 | 1) 6.979 us | } |
| 1999 | 2) 0.417 us | scheduler_ipi(); |
| 2000 | 1) 9.791 us | } |
| 2001 | 1) + 12.917 us | } |
| 2002 | 2) 3.490 us | } |
| 2003 | 1) + 15.729 us | } |
| 2004 | 1) + 18.542 us | } |
| 2005 | 2) $ 3594274 us | } |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2006 | |
| 2007 | + means that the function exceeded 10 usecs. |
| 2008 | ! means that the function exceeded 100 usecs. |
Byungchul Park | 8e1e1df | 2014-11-24 09:34:19 +0900 | [diff] [blame] | 2009 | # means that the function exceeded 1000 usecs. |
Jungseok Lee | b838e1d | 2015-07-11 14:51:40 +0000 | [diff] [blame] | 2010 | * means that the function exceeded 10 msecs. |
| 2011 | @ means that the function exceeded 100 msecs. |
Byungchul Park | 8e1e1df | 2014-11-24 09:34:19 +0900 | [diff] [blame] | 2012 | $ means that the function exceeded 1 sec. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2013 | |
| 2014 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2015 | - The task/pid field displays the thread cmdline and pid which |
| 2016 | executed the function. It is default disabled. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2017 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2018 | hide: echo nofuncgraph-proc > trace_options |
| 2019 | show: echo funcgraph-proc > trace_options |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2020 | |
| 2021 | ie: |
| 2022 | |
| 2023 | # tracer: function_graph |
| 2024 | # |
| 2025 | # CPU TASK/PID DURATION FUNCTION CALLS |
| 2026 | # | | | | | | | | | |
| 2027 | 0) sh-4802 | | d_free() { |
| 2028 | 0) sh-4802 | | call_rcu() { |
| 2029 | 0) sh-4802 | | __call_rcu() { |
| 2030 | 0) sh-4802 | 0.616 us | rcu_process_gp_end(); |
| 2031 | 0) sh-4802 | 0.586 us | check_for_new_grace_period(); |
| 2032 | 0) sh-4802 | 2.899 us | } |
| 2033 | 0) sh-4802 | 4.040 us | } |
| 2034 | 0) sh-4802 | 5.151 us | } |
| 2035 | 0) sh-4802 | + 49.370 us | } |
| 2036 | |
| 2037 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2038 | - The absolute time field is an absolute timestamp given by the |
| 2039 | system clock since it started. A snapshot of this time is |
| 2040 | given on each entry/exit of functions |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2041 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2042 | hide: echo nofuncgraph-abstime > trace_options |
| 2043 | show: echo funcgraph-abstime > trace_options |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2044 | |
| 2045 | ie: |
| 2046 | |
| 2047 | # |
| 2048 | # TIME CPU DURATION FUNCTION CALLS |
| 2049 | # | | | | | | | | |
| 2050 | 360.774522 | 1) 0.541 us | } |
| 2051 | 360.774522 | 1) 4.663 us | } |
| 2052 | 360.774523 | 1) 0.541 us | __wake_up_bit(); |
| 2053 | 360.774524 | 1) 6.796 us | } |
| 2054 | 360.774524 | 1) 7.952 us | } |
| 2055 | 360.774525 | 1) 9.063 us | } |
| 2056 | 360.774525 | 1) 0.615 us | journal_mark_dirty(); |
| 2057 | 360.774527 | 1) 0.578 us | __brelse(); |
| 2058 | 360.774528 | 1) | reiserfs_prepare_for_journal() { |
| 2059 | 360.774528 | 1) | unlock_buffer() { |
| 2060 | 360.774529 | 1) | wake_up_bit() { |
| 2061 | 360.774529 | 1) | bit_waitqueue() { |
| 2062 | 360.774530 | 1) 0.594 us | __phys_addr(); |
| 2063 | |
| 2064 | |
Robert Elliott | 607e3a2 | 2014-05-20 17:10:51 -0500 | [diff] [blame] | 2065 | The function name is always displayed after the closing bracket |
| 2066 | for a function if the start of that function is not in the |
| 2067 | trace buffer. |
| 2068 | |
| 2069 | Display of the function name after the closing bracket may be |
| 2070 | enabled for functions whose start is in the trace buffer, |
| 2071 | allowing easier searching with grep for function durations. |
| 2072 | It is default disabled. |
| 2073 | |
| 2074 | hide: echo nofuncgraph-tail > trace_options |
| 2075 | show: echo funcgraph-tail > trace_options |
| 2076 | |
| 2077 | Example with nofuncgraph-tail (default): |
| 2078 | 0) | putname() { |
| 2079 | 0) | kmem_cache_free() { |
| 2080 | 0) 0.518 us | __phys_addr(); |
| 2081 | 0) 1.757 us | } |
| 2082 | 0) 2.861 us | } |
| 2083 | |
| 2084 | Example with funcgraph-tail: |
| 2085 | 0) | putname() { |
| 2086 | 0) | kmem_cache_free() { |
| 2087 | 0) 0.518 us | __phys_addr(); |
| 2088 | 0) 1.757 us | } /* kmem_cache_free() */ |
| 2089 | 0) 2.861 us | } /* putname() */ |
| 2090 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2091 | You can put some comments on specific functions by using |
Ingo Molnar | 5e1607a | 2009-03-05 10:24:48 +0100 | [diff] [blame] | 2092 | trace_printk() For example, if you want to put a comment inside |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2093 | the __might_sleep() function, you just have to include |
Ingo Molnar | 5e1607a | 2009-03-05 10:24:48 +0100 | [diff] [blame] | 2094 | <linux/ftrace.h> and call trace_printk() inside __might_sleep() |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2095 | |
Ingo Molnar | 5e1607a | 2009-03-05 10:24:48 +0100 | [diff] [blame] | 2096 | trace_printk("I'm a comment!\n") |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2097 | |
| 2098 | will produce: |
| 2099 | |
| 2100 | 1) | __might_sleep() { |
| 2101 | 1) | /* I'm a comment! */ |
| 2102 | 1) 1.449 us | } |
| 2103 | |
| 2104 | |
Jamie Gennis | 26779b3 | 2012-11-21 20:14:09 -0800 | [diff] [blame] | 2105 | You can disable the hierarchical function call formatting and instead print a |
| 2106 | flat list of function entry and return events. This uses the format described |
| 2107 | in the Output Formatting section and respects all the trace options that |
| 2108 | control that formatting. Hierarchical formatting is the default. |
| 2109 | |
| 2110 | hierachical: echo nofuncgraph-flat > trace_options |
| 2111 | flat: echo funcgraph-flat > trace_options |
| 2112 | |
| 2113 | ie: |
| 2114 | |
| 2115 | # tracer: function_graph |
| 2116 | # |
| 2117 | # entries-in-buffer/entries-written: 68355/68355 #P:2 |
| 2118 | # |
| 2119 | # _-----=> irqs-off |
| 2120 | # / _----=> need-resched |
| 2121 | # | / _---=> hardirq/softirq |
| 2122 | # || / _--=> preempt-depth |
| 2123 | # ||| / delay |
| 2124 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2125 | # | | | |||| | | |
| 2126 | sh-1806 [001] d... 198.843443: graph_ent: func=_raw_spin_lock |
| 2127 | sh-1806 [001] d... 198.843445: graph_ent: func=__raw_spin_lock |
| 2128 | sh-1806 [001] d..1 198.843447: graph_ret: func=__raw_spin_lock |
| 2129 | sh-1806 [001] d..1 198.843449: graph_ret: func=_raw_spin_lock |
| 2130 | sh-1806 [001] d..1 198.843451: graph_ent: func=_raw_spin_unlock_irqrestore |
| 2131 | sh-1806 [001] d... 198.843453: graph_ret: func=_raw_spin_unlock_irqrestore |
| 2132 | |
| 2133 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2134 | You might find other useful features for this tracer in the |
| 2135 | following "dynamic ftrace" section such as tracing only specific |
| 2136 | functions or tasks. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2137 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2138 | dynamic ftrace |
| 2139 | -------------- |
| 2140 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 2141 | If CONFIG_DYNAMIC_FTRACE is set, the system will run with |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2142 | virtually no overhead when function tracing is disabled. The way |
| 2143 | this works is the mcount function call (placed at the start of |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2144 | every kernel function, produced by the -pg switch in gcc), |
| 2145 | starts of pointing to a simple return. (Enabling FTRACE will |
| 2146 | include the -pg switch in the compiling of the kernel.) |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2147 | |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 2148 | At compile time every C file object is run through the |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2149 | recordmcount program (located in the scripts directory). This |
| 2150 | program will parse the ELF headers in the C object to find all |
| 2151 | the locations in the .text section that call mcount. (Note, only |
| 2152 | white listed .text sections are processed, since processing other |
| 2153 | sections like .init.text may cause races due to those sections |
| 2154 | being freed unexpectedly). |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2155 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2156 | A new section called "__mcount_loc" is created that holds |
| 2157 | references to all the mcount call sites in the .text section. |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2158 | The recordmcount program re-links this section back into the |
| 2159 | original object. The final linking stage of the kernel will add all these |
| 2160 | references into a single table. |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 2161 | |
| 2162 | On boot up, before SMP is initialized, the dynamic ftrace code |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2163 | scans this table and updates all the locations into nops. It |
| 2164 | also records the locations, which are added to the |
| 2165 | available_filter_functions list. Modules are processed as they |
| 2166 | are loaded and before they are executed. When a module is |
| 2167 | unloaded, it also removes its functions from the ftrace function |
| 2168 | list. This is automatic in the module unload code, and the |
| 2169 | module author does not need to worry about it. |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 2170 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2171 | When tracing is enabled, the process of modifying the function |
| 2172 | tracepoints is dependent on architecture. The old method is to use |
| 2173 | kstop_machine to prevent races with the CPUs executing code being |
| 2174 | modified (which can cause the CPU to do undesirable things, especially |
| 2175 | if the modified code crosses cache (or page) boundaries), and the nops are |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2176 | patched back to calls. But this time, they do not call mcount |
| 2177 | (which is just a function stub). They now call into the ftrace |
| 2178 | infrastructure. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2179 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2180 | The new method of modifying the function tracepoints is to place |
| 2181 | a breakpoint at the location to be modified, sync all CPUs, modify |
| 2182 | the rest of the instruction not covered by the breakpoint. Sync |
| 2183 | all CPUs again, and then remove the breakpoint with the finished |
| 2184 | version to the ftrace call site. |
| 2185 | |
| 2186 | Some archs do not even need to monkey around with the synchronization, |
| 2187 | and can just slap the new code on top of the old without any |
| 2188 | problems with other CPUs executing it at the same time. |
| 2189 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2190 | One special side-effect to the recording of the functions being |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 2191 | traced is that we can now selectively choose which functions we |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2192 | wish to trace and which ones we want the mcount calls to remain |
| 2193 | as nops. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2194 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2195 | Two files are used, one for enabling and one for disabling the |
| 2196 | tracing of specified functions. They are: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2197 | |
| 2198 | set_ftrace_filter |
| 2199 | |
| 2200 | and |
| 2201 | |
| 2202 | set_ftrace_notrace |
| 2203 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2204 | A list of available functions that you can add to these files is |
| 2205 | listed in: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2206 | |
| 2207 | available_filter_functions |
| 2208 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2209 | # cat available_filter_functions |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2210 | put_prev_task_idle |
| 2211 | kmem_cache_create |
| 2212 | pick_next_task_rt |
| 2213 | get_online_cpus |
| 2214 | pick_next_task_fair |
| 2215 | mutex_lock |
| 2216 | [...] |
| 2217 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 2218 | If I am only interested in sys_nanosleep and hrtimer_interrupt: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2219 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2220 | # echo sys_nanosleep hrtimer_interrupt > set_ftrace_filter |
Yang Hongyang | 6993b1b | 2010-01-25 11:10:32 +0800 | [diff] [blame] | 2221 | # echo function > current_tracer |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 2222 | # echo 1 > tracing_on |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2223 | # usleep 1 |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 2224 | # echo 0 > tracing_on |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2225 | # cat trace |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2226 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2227 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2228 | # entries-in-buffer/entries-written: 5/5 #P:4 |
| 2229 | # |
| 2230 | # _-----=> irqs-off |
| 2231 | # / _----=> need-resched |
| 2232 | # | / _---=> hardirq/softirq |
| 2233 | # || / _--=> preempt-depth |
| 2234 | # ||| / delay |
| 2235 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2236 | # | | | |||| | | |
| 2237 | usleep-2665 [001] .... 4186.475355: sys_nanosleep <-system_call_fastpath |
| 2238 | <idle>-0 [001] d.h1 4186.475409: hrtimer_interrupt <-smp_apic_timer_interrupt |
| 2239 | usleep-2665 [001] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt |
| 2240 | <idle>-0 [003] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt |
| 2241 | <idle>-0 [002] d.h1 4186.475427: hrtimer_interrupt <-smp_apic_timer_interrupt |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2242 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 2243 | To see which functions are being traced, you can cat the file: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2244 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2245 | # cat set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2246 | hrtimer_interrupt |
| 2247 | sys_nanosleep |
| 2248 | |
| 2249 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2250 | Perhaps this is not enough. The filters also allow simple wild |
| 2251 | cards. Only the following are currently available |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2252 | |
Steven Rostedt | a41eeba | 2008-07-14 16:41:12 -0400 | [diff] [blame] | 2253 | <match>* - will match functions that begin with <match> |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2254 | *<match> - will match functions that end with <match> |
| 2255 | *<match>* - will match functions that have <match> in it |
| 2256 | |
Steven Rostedt | f2d9c74 | 2008-07-15 10:57:33 -0400 | [diff] [blame] | 2257 | These are the only wild cards which are supported. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2258 | |
| 2259 | <match>*<match> will not work. |
| 2260 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2261 | Note: It is better to use quotes to enclose the wild cards, |
| 2262 | otherwise the shell may expand the parameters into names |
| 2263 | of files in the local directory. |
walimis | c072c24 | 2008-11-28 12:21:19 +0800 | [diff] [blame] | 2264 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2265 | # echo 'hrtimer_*' > set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2266 | |
| 2267 | Produces: |
| 2268 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2269 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2270 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2271 | # entries-in-buffer/entries-written: 897/897 #P:4 |
| 2272 | # |
| 2273 | # _-----=> irqs-off |
| 2274 | # / _----=> need-resched |
| 2275 | # | / _---=> hardirq/softirq |
| 2276 | # || / _--=> preempt-depth |
| 2277 | # ||| / delay |
| 2278 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2279 | # | | | |||| | | |
| 2280 | <idle>-0 [003] dN.1 4228.547803: hrtimer_cancel <-tick_nohz_idle_exit |
| 2281 | <idle>-0 [003] dN.1 4228.547804: hrtimer_try_to_cancel <-hrtimer_cancel |
| 2282 | <idle>-0 [003] dN.2 4228.547805: hrtimer_force_reprogram <-__remove_hrtimer |
| 2283 | <idle>-0 [003] dN.1 4228.547805: hrtimer_forward <-tick_nohz_idle_exit |
| 2284 | <idle>-0 [003] dN.1 4228.547805: hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11 |
| 2285 | <idle>-0 [003] d..1 4228.547858: hrtimer_get_next_event <-get_next_timer_interrupt |
| 2286 | <idle>-0 [003] d..1 4228.547859: hrtimer_start <-__tick_nohz_idle_enter |
| 2287 | <idle>-0 [003] d..2 4228.547860: hrtimer_force_reprogram <-__rem |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2288 | |
| 2289 | Notice that we lost the sys_nanosleep. |
| 2290 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2291 | # cat set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2292 | hrtimer_run_queues |
| 2293 | hrtimer_run_pending |
| 2294 | hrtimer_init |
| 2295 | hrtimer_cancel |
| 2296 | hrtimer_try_to_cancel |
| 2297 | hrtimer_forward |
| 2298 | hrtimer_start |
| 2299 | hrtimer_reprogram |
| 2300 | hrtimer_force_reprogram |
| 2301 | hrtimer_get_next_event |
| 2302 | hrtimer_interrupt |
| 2303 | hrtimer_nanosleep |
| 2304 | hrtimer_wakeup |
| 2305 | hrtimer_get_remaining |
| 2306 | hrtimer_get_res |
| 2307 | hrtimer_init_sleeper |
| 2308 | |
| 2309 | |
| 2310 | This is because the '>' and '>>' act just like they do in bash. |
| 2311 | To rewrite the filters, use '>' |
| 2312 | To append to the filters, use '>>' |
| 2313 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2314 | To clear out a filter so that all functions will be recorded |
| 2315 | again: |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2316 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2317 | # echo > set_ftrace_filter |
| 2318 | # cat set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2319 | # |
| 2320 | |
| 2321 | Again, now we want to append. |
| 2322 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2323 | # echo sys_nanosleep > set_ftrace_filter |
| 2324 | # cat set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2325 | sys_nanosleep |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2326 | # echo 'hrtimer_*' >> set_ftrace_filter |
| 2327 | # cat set_ftrace_filter |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2328 | hrtimer_run_queues |
| 2329 | hrtimer_run_pending |
| 2330 | hrtimer_init |
| 2331 | hrtimer_cancel |
| 2332 | hrtimer_try_to_cancel |
| 2333 | hrtimer_forward |
| 2334 | hrtimer_start |
| 2335 | hrtimer_reprogram |
| 2336 | hrtimer_force_reprogram |
| 2337 | hrtimer_get_next_event |
| 2338 | hrtimer_interrupt |
| 2339 | sys_nanosleep |
| 2340 | hrtimer_nanosleep |
| 2341 | hrtimer_wakeup |
| 2342 | hrtimer_get_remaining |
| 2343 | hrtimer_get_res |
| 2344 | hrtimer_init_sleeper |
| 2345 | |
| 2346 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2347 | The set_ftrace_notrace prevents those functions from being |
| 2348 | traced. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2349 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2350 | # echo '*preempt*' '*lock*' > set_ftrace_notrace |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2351 | |
| 2352 | Produces: |
| 2353 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2354 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2355 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2356 | # entries-in-buffer/entries-written: 39608/39608 #P:4 |
| 2357 | # |
| 2358 | # _-----=> irqs-off |
| 2359 | # / _----=> need-resched |
| 2360 | # | / _---=> hardirq/softirq |
| 2361 | # || / _--=> preempt-depth |
| 2362 | # ||| / delay |
| 2363 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2364 | # | | | |||| | | |
| 2365 | bash-1994 [000] .... 4342.324896: file_ra_state_init <-do_dentry_open |
| 2366 | bash-1994 [000] .... 4342.324897: open_check_o_direct <-do_last |
| 2367 | bash-1994 [000] .... 4342.324897: ima_file_check <-do_last |
| 2368 | bash-1994 [000] .... 4342.324898: process_measurement <-ima_file_check |
| 2369 | bash-1994 [000] .... 4342.324898: ima_get_action <-process_measurement |
| 2370 | bash-1994 [000] .... 4342.324898: ima_match_policy <-ima_get_action |
| 2371 | bash-1994 [000] .... 4342.324899: do_truncate <-do_last |
| 2372 | bash-1994 [000] .... 4342.324899: should_remove_suid <-do_truncate |
| 2373 | bash-1994 [000] .... 4342.324899: notify_change <-do_truncate |
| 2374 | bash-1994 [000] .... 4342.324900: current_fs_time <-notify_change |
| 2375 | bash-1994 [000] .... 4342.324900: current_kernel_time <-current_fs_time |
| 2376 | bash-1994 [000] .... 4342.324900: timespec_trunc <-current_fs_time |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2377 | |
| 2378 | We can see that there's no more lock or preempt tracing. |
| 2379 | |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2380 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2381 | Dynamic ftrace with the function graph tracer |
| 2382 | --------------------------------------------- |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2383 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2384 | Although what has been explained above concerns both the |
| 2385 | function tracer and the function-graph-tracer, there are some |
| 2386 | special features only available in the function-graph tracer. |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2387 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2388 | If you want to trace only one function and all of its children, |
| 2389 | you just have to echo its name into set_graph_function: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2390 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2391 | echo __do_fault > set_graph_function |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2392 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2393 | will produce the following "expanded" trace of the __do_fault() |
| 2394 | function: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2395 | |
| 2396 | 0) | __do_fault() { |
| 2397 | 0) | filemap_fault() { |
| 2398 | 0) | find_lock_page() { |
| 2399 | 0) 0.804 us | find_get_page(); |
| 2400 | 0) | __might_sleep() { |
| 2401 | 0) 1.329 us | } |
| 2402 | 0) 3.904 us | } |
| 2403 | 0) 4.979 us | } |
| 2404 | 0) 0.653 us | _spin_lock(); |
| 2405 | 0) 0.578 us | page_add_file_rmap(); |
| 2406 | 0) 0.525 us | native_set_pte_at(); |
| 2407 | 0) 0.585 us | _spin_unlock(); |
| 2408 | 0) | unlock_page() { |
| 2409 | 0) 0.541 us | page_waitqueue(); |
| 2410 | 0) 0.639 us | __wake_up_bit(); |
| 2411 | 0) 2.786 us | } |
| 2412 | 0) + 14.237 us | } |
| 2413 | 0) | __do_fault() { |
| 2414 | 0) | filemap_fault() { |
| 2415 | 0) | find_lock_page() { |
| 2416 | 0) 0.698 us | find_get_page(); |
| 2417 | 0) | __might_sleep() { |
| 2418 | 0) 1.412 us | } |
| 2419 | 0) 3.950 us | } |
| 2420 | 0) 5.098 us | } |
| 2421 | 0) 0.631 us | _spin_lock(); |
| 2422 | 0) 0.571 us | page_add_file_rmap(); |
| 2423 | 0) 0.526 us | native_set_pte_at(); |
| 2424 | 0) 0.586 us | _spin_unlock(); |
| 2425 | 0) | unlock_page() { |
| 2426 | 0) 0.533 us | page_waitqueue(); |
| 2427 | 0) 0.638 us | __wake_up_bit(); |
| 2428 | 0) 2.793 us | } |
| 2429 | 0) + 14.012 us | } |
| 2430 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2431 | You can also expand several functions at once: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2432 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2433 | echo sys_open > set_graph_function |
| 2434 | echo sys_close >> set_graph_function |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2435 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2436 | Now if you want to go back to trace all functions you can clear |
| 2437 | this special filter via: |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2438 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2439 | echo > set_graph_function |
Frederic Weisbecker | 985ec20 | 2009-02-18 06:35:34 +0100 | [diff] [blame] | 2440 | |
| 2441 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2442 | ftrace_enabled |
| 2443 | -------------- |
| 2444 | |
| 2445 | Note, the proc sysctl ftrace_enable is a big on/off switch for the |
| 2446 | function tracer. By default it is enabled (when function tracing is |
| 2447 | enabled in the kernel). If it is disabled, all function tracing is |
| 2448 | disabled. This includes not only the function tracers for ftrace, but |
| 2449 | also for any other uses (perf, kprobes, stack tracing, profiling, etc). |
| 2450 | |
| 2451 | Please disable this with care. |
| 2452 | |
| 2453 | This can be disable (and enabled) with: |
| 2454 | |
| 2455 | sysctl kernel.ftrace_enabled=0 |
| 2456 | sysctl kernel.ftrace_enabled=1 |
| 2457 | |
| 2458 | or |
| 2459 | |
| 2460 | echo 0 > /proc/sys/kernel/ftrace_enabled |
| 2461 | echo 1 > /proc/sys/kernel/ftrace_enabled |
| 2462 | |
| 2463 | |
Chase Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 2464 | Filter commands |
| 2465 | --------------- |
| 2466 | |
| 2467 | A few commands are supported by the set_ftrace_filter interface. |
| 2468 | Trace commands have the following format: |
| 2469 | |
| 2470 | <function>:<command>:<parameter> |
| 2471 | |
| 2472 | The following commands are supported: |
| 2473 | |
| 2474 | - mod |
| 2475 | This command enables function filtering per module. The |
| 2476 | parameter defines the module. For example, if only the write* |
| 2477 | functions in the ext3 module are desired, run: |
| 2478 | |
| 2479 | echo 'write*:mod:ext3' > set_ftrace_filter |
| 2480 | |
| 2481 | This command interacts with the filter in the same way as |
| 2482 | filtering based on function names. Thus, adding more functions |
| 2483 | in a different module is accomplished by appending (>>) to the |
| 2484 | filter file. Remove specific module functions by prepending |
| 2485 | '!': |
| 2486 | |
| 2487 | echo '!writeback*:mod:ext3' >> set_ftrace_filter |
| 2488 | |
Dmitry Safonov | b86d937 | 2015-09-29 19:46:16 +0300 | [diff] [blame] | 2489 | Mod command supports module globbing. Disable tracing for all |
| 2490 | functions except a specific module: |
| 2491 | |
| 2492 | echo '!*:mod:!ext3' >> set_ftrace_filter |
| 2493 | |
| 2494 | Disable tracing for all modules, but still trace kernel: |
| 2495 | |
| 2496 | echo '!*:mod:*' >> set_ftrace_filter |
| 2497 | |
| 2498 | Enable filter only for kernel: |
| 2499 | |
| 2500 | echo '*write*:mod:!*' >> set_ftrace_filter |
| 2501 | |
| 2502 | Enable filter for module globbing: |
| 2503 | |
| 2504 | echo '*write*:mod:*snd*' >> set_ftrace_filter |
| 2505 | |
Chase Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 2506 | - traceon/traceoff |
| 2507 | These commands turn tracing on and off when the specified |
| 2508 | functions are hit. The parameter determines how many times the |
| 2509 | tracing system is turned on and off. If unspecified, there is |
| 2510 | no limit. For example, to disable tracing when a schedule bug |
| 2511 | is hit the first 5 times, run: |
| 2512 | |
| 2513 | echo '__schedule_bug:traceoff:5' > set_ftrace_filter |
| 2514 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2515 | To always disable tracing when __schedule_bug is hit: |
| 2516 | |
| 2517 | echo '__schedule_bug:traceoff' > set_ftrace_filter |
| 2518 | |
Chase Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 2519 | These commands are cumulative whether or not they are appended |
| 2520 | to set_ftrace_filter. To remove a command, prepend it by '!' |
| 2521 | and drop the parameter: |
| 2522 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2523 | echo '!__schedule_bug:traceoff:0' > set_ftrace_filter |
| 2524 | |
| 2525 | The above removes the traceoff command for __schedule_bug |
| 2526 | that have a counter. To remove commands without counters: |
| 2527 | |
Chase Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 2528 | echo '!__schedule_bug:traceoff' > set_ftrace_filter |
| 2529 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2530 | - snapshot |
| 2531 | Will cause a snapshot to be triggered when the function is hit. |
| 2532 | |
| 2533 | echo 'native_flush_tlb_others:snapshot' > set_ftrace_filter |
| 2534 | |
| 2535 | To only snapshot once: |
| 2536 | |
| 2537 | echo 'native_flush_tlb_others:snapshot:1' > set_ftrace_filter |
| 2538 | |
| 2539 | To remove the above commands: |
| 2540 | |
| 2541 | echo '!native_flush_tlb_others:snapshot' > set_ftrace_filter |
| 2542 | echo '!native_flush_tlb_others:snapshot:0' > set_ftrace_filter |
| 2543 | |
| 2544 | - enable_event/disable_event |
| 2545 | These commands can enable or disable a trace event. Note, because |
| 2546 | function tracing callbacks are very sensitive, when these commands |
| 2547 | are registered, the trace point is activated, but disabled in |
| 2548 | a "soft" mode. That is, the tracepoint will be called, but |
| 2549 | just will not be traced. The event tracepoint stays in this mode |
| 2550 | as long as there's a command that triggers it. |
| 2551 | |
| 2552 | echo 'try_to_wake_up:enable_event:sched:sched_switch:2' > \ |
| 2553 | set_ftrace_filter |
| 2554 | |
| 2555 | The format is: |
| 2556 | |
| 2557 | <function>:enable_event:<system>:<event>[:count] |
| 2558 | <function>:disable_event:<system>:<event>[:count] |
| 2559 | |
| 2560 | To remove the events commands: |
| 2561 | |
| 2562 | |
| 2563 | echo '!try_to_wake_up:enable_event:sched:sched_switch:0' > \ |
| 2564 | set_ftrace_filter |
| 2565 | echo '!schedule:disable_event:sched:sched_switch' > \ |
| 2566 | set_ftrace_filter |
Chase Douglas | 07271aa | 2010-04-23 14:02:39 -0400 | [diff] [blame] | 2567 | |
Steven Rostedt (Red Hat) | ad71d88 | 2013-04-30 15:46:14 -0400 | [diff] [blame] | 2568 | - dump |
| 2569 | When the function is hit, it will dump the contents of the ftrace |
| 2570 | ring buffer to the console. This is useful if you need to debug |
| 2571 | something, and want to dump the trace when a certain function |
| 2572 | is hit. Perhaps its a function that is called before a tripple |
| 2573 | fault happens and does not allow you to get a regular dump. |
| 2574 | |
Steven Rostedt (Red Hat) | 90e3c03 | 2013-04-30 19:00:46 -0400 | [diff] [blame] | 2575 | - cpudump |
| 2576 | When the function is hit, it will dump the contents of the ftrace |
| 2577 | ring buffer for the current CPU to the console. Unlike the "dump" |
| 2578 | command, it only prints out the contents of the ring buffer for the |
| 2579 | CPU that executed the function that triggered the dump. |
| 2580 | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2581 | trace_pipe |
| 2582 | ---------- |
| 2583 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2584 | The trace_pipe outputs the same content as the trace file, but |
| 2585 | the effect on the tracing is different. Every read from |
| 2586 | trace_pipe is consumed. This means that subsequent reads will be |
| 2587 | different. The trace is live. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2588 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2589 | # echo function > current_tracer |
| 2590 | # cat trace_pipe > /tmp/trace.out & |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2591 | [1] 4153 |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 2592 | # echo 1 > tracing_on |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2593 | # usleep 1 |
Steven Rostedt | 6752ab4 | 2011-02-08 13:54:06 -0500 | [diff] [blame] | 2594 | # echo 0 > tracing_on |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2595 | # cat trace |
Steven Rostedt | 9b803c0 | 2008-11-03 15:15:08 -0500 | [diff] [blame] | 2596 | # tracer: function |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2597 | # |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2598 | # entries-in-buffer/entries-written: 0/0 #P:4 |
| 2599 | # |
| 2600 | # _-----=> irqs-off |
| 2601 | # / _----=> need-resched |
| 2602 | # | / _---=> hardirq/softirq |
| 2603 | # || / _--=> preempt-depth |
| 2604 | # ||| / delay |
| 2605 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2606 | # | | | |||| | | |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2607 | |
| 2608 | # |
| 2609 | # cat /tmp/trace.out |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2610 | bash-1994 [000] .... 5281.568961: mutex_unlock <-rb_simple_write |
| 2611 | bash-1994 [000] .... 5281.568963: __mutex_unlock_slowpath <-mutex_unlock |
| 2612 | bash-1994 [000] .... 5281.568963: __fsnotify_parent <-fsnotify_modify |
| 2613 | bash-1994 [000] .... 5281.568964: fsnotify <-fsnotify_modify |
| 2614 | bash-1994 [000] .... 5281.568964: __srcu_read_lock <-fsnotify |
| 2615 | bash-1994 [000] .... 5281.568964: add_preempt_count <-__srcu_read_lock |
| 2616 | bash-1994 [000] ...1 5281.568965: sub_preempt_count <-__srcu_read_lock |
| 2617 | bash-1994 [000] .... 5281.568965: __srcu_read_unlock <-fsnotify |
| 2618 | bash-1994 [000] .... 5281.568967: sys_dup2 <-system_call_fastpath |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2619 | |
| 2620 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2621 | Note, reading the trace_pipe file will block until more input is |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2622 | added. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2623 | |
| 2624 | trace entries |
| 2625 | ------------- |
| 2626 | |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2627 | Having too much or not enough data can be troublesome in |
| 2628 | diagnosing an issue in the kernel. The file buffer_size_kb is |
| 2629 | used to modify the size of the internal trace buffers. The |
| 2630 | number listed is the number of entries that can be recorded per |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2631 | CPU. To know the full size, multiply the number of possible CPUs |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2632 | with the number of entries. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2633 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2634 | # cat buffer_size_kb |
Steven Rostedt | 1696b2b | 2008-11-13 00:09:35 -0500 | [diff] [blame] | 2635 | 1408 (units kilobytes) |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2636 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2637 | Or simply read buffer_total_size_kb |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2638 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2639 | # cat buffer_total_size_kb |
| 2640 | 5632 |
| 2641 | |
| 2642 | To modify the buffer, simple echo in a number (in 1024 byte segments). |
| 2643 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2644 | # echo 10000 > buffer_size_kb |
| 2645 | # cat buffer_size_kb |
Steven Rostedt | 1696b2b | 2008-11-13 00:09:35 -0500 | [diff] [blame] | 2646 | 10000 (units kilobytes) |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2647 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2648 | It will try to allocate as much as possible. If you allocate too |
| 2649 | much, it can cause Out-Of-Memory to trigger. |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2650 | |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2651 | # echo 1000000000000 > buffer_size_kb |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2652 | -bash: echo: write error: Cannot allocate memory |
GeunSik Lim | 156f5a7 | 2009-06-02 15:01:37 +0900 | [diff] [blame] | 2653 | # cat buffer_size_kb |
Steven Rostedt | eb6d42e | 2008-07-10 12:46:01 -0400 | [diff] [blame] | 2654 | 85 |
| 2655 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2656 | The per_cpu buffers can be changed individually as well: |
| 2657 | |
| 2658 | # echo 10000 > per_cpu/cpu0/buffer_size_kb |
| 2659 | # echo 100 > per_cpu/cpu1/buffer_size_kb |
| 2660 | |
| 2661 | When the per_cpu buffers are not the same, the buffer_size_kb |
| 2662 | at the top level will just show an X |
| 2663 | |
| 2664 | # cat buffer_size_kb |
| 2665 | X |
| 2666 | |
| 2667 | This is where the buffer_total_size_kb is useful: |
| 2668 | |
| 2669 | # cat buffer_total_size_kb |
| 2670 | 12916 |
| 2671 | |
| 2672 | Writing to the top level buffer_size_kb will reset all the buffers |
| 2673 | to be the same again. |
| 2674 | |
Hiraku Toyooka | c1043fc | 2012-12-26 11:53:09 +0900 | [diff] [blame] | 2675 | Snapshot |
| 2676 | -------- |
| 2677 | CONFIG_TRACER_SNAPSHOT makes a generic snapshot feature |
| 2678 | available to all non latency tracers. (Latency tracers which |
| 2679 | record max latency, such as "irqsoff" or "wakeup", can't use |
| 2680 | this feature, since those are already using the snapshot |
| 2681 | mechanism internally.) |
| 2682 | |
| 2683 | Snapshot preserves a current trace buffer at a particular point |
| 2684 | in time without stopping tracing. Ftrace swaps the current |
| 2685 | buffer with a spare buffer, and tracing continues in the new |
| 2686 | current (=previous spare) buffer. |
| 2687 | |
| 2688 | The following debugfs files in "tracing" are related to this |
| 2689 | feature: |
| 2690 | |
| 2691 | snapshot: |
| 2692 | |
| 2693 | This is used to take a snapshot and to read the output |
| 2694 | of the snapshot. Echo 1 into this file to allocate a |
| 2695 | spare buffer and to take a snapshot (swap), then read |
| 2696 | the snapshot from this file in the same format as |
| 2697 | "trace" (described above in the section "The File |
| 2698 | System"). Both reads snapshot and tracing are executable |
| 2699 | in parallel. When the spare buffer is allocated, echoing |
| 2700 | 0 frees it, and echoing else (positive) values clear the |
| 2701 | snapshot contents. |
| 2702 | More details are shown in the table below. |
| 2703 | |
| 2704 | status\input | 0 | 1 | else | |
| 2705 | --------------+------------+------------+------------+ |
Hiraku Toyooka | 1abccd7 | 2013-03-08 16:32:25 +0900 | [diff] [blame] | 2706 | not allocated |(do nothing)| alloc+swap |(do nothing)| |
Hiraku Toyooka | c1043fc | 2012-12-26 11:53:09 +0900 | [diff] [blame] | 2707 | --------------+------------+------------+------------+ |
| 2708 | allocated | free | swap | clear | |
| 2709 | --------------+------------+------------+------------+ |
| 2710 | |
| 2711 | Here is an example of using the snapshot feature. |
| 2712 | |
| 2713 | # echo 1 > events/sched/enable |
| 2714 | # echo 1 > snapshot |
| 2715 | # cat snapshot |
| 2716 | # tracer: nop |
| 2717 | # |
| 2718 | # entries-in-buffer/entries-written: 71/71 #P:8 |
| 2719 | # |
| 2720 | # _-----=> irqs-off |
| 2721 | # / _----=> need-resched |
| 2722 | # | / _---=> hardirq/softirq |
| 2723 | # || / _--=> preempt-depth |
| 2724 | # ||| / delay |
| 2725 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2726 | # | | | |||| | | |
| 2727 | <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 |
| 2728 | 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 |
| 2729 | [...] |
| 2730 | <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 |
| 2731 | |
| 2732 | # cat trace |
| 2733 | # tracer: nop |
| 2734 | # |
| 2735 | # entries-in-buffer/entries-written: 77/77 #P:8 |
| 2736 | # |
| 2737 | # _-----=> irqs-off |
| 2738 | # / _----=> need-resched |
| 2739 | # | / _---=> hardirq/softirq |
| 2740 | # || / _--=> preempt-depth |
| 2741 | # ||| / delay |
| 2742 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2743 | # | | | |||| | | |
| 2744 | <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 |
| 2745 | 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 |
| 2746 | [...] |
| 2747 | |
| 2748 | |
| 2749 | If you try to use this snapshot feature when current tracer is |
| 2750 | one of the latency tracers, you will get the following results. |
| 2751 | |
| 2752 | # echo wakeup > current_tracer |
| 2753 | # echo 1 > snapshot |
| 2754 | bash: echo: write error: Device or resource busy |
| 2755 | # cat snapshot |
| 2756 | cat: snapshot: Device or resource busy |
| 2757 | |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2758 | |
| 2759 | Instances |
| 2760 | --------- |
| 2761 | In the debugfs tracing directory is a directory called "instances". |
| 2762 | This directory can have new directories created inside of it using |
| 2763 | mkdir, and removing directories with rmdir. The directory created |
| 2764 | with mkdir in this directory will already contain files and other |
| 2765 | directories after it is created. |
| 2766 | |
| 2767 | # mkdir instances/foo |
| 2768 | # ls instances/foo |
| 2769 | buffer_size_kb buffer_total_size_kb events free_buffer per_cpu |
| 2770 | set_event snapshot trace trace_clock trace_marker trace_options |
| 2771 | trace_pipe tracing_on |
| 2772 | |
| 2773 | As you can see, the new directory looks similar to the tracing directory |
| 2774 | itself. In fact, it is very similar, except that the buffer and |
| 2775 | events are agnostic from the main director, or from any other |
| 2776 | instances that are created. |
| 2777 | |
| 2778 | The files in the new directory work just like the files with the |
| 2779 | same name in the tracing directory except the buffer that is used |
| 2780 | is a separate and new buffer. The files affect that buffer but do not |
| 2781 | affect the main buffer with the exception of trace_options. Currently, |
| 2782 | the trace_options affect all instances and the top level buffer |
| 2783 | the same, but this may change in future releases. That is, options |
| 2784 | may become specific to the instance they reside in. |
| 2785 | |
| 2786 | Notice that none of the function tracer files are there, nor is |
| 2787 | current_tracer and available_tracers. This is because the buffers |
| 2788 | can currently only have events enabled for them. |
| 2789 | |
| 2790 | # mkdir instances/foo |
| 2791 | # mkdir instances/bar |
| 2792 | # mkdir instances/zoot |
| 2793 | # echo 100000 > buffer_size_kb |
| 2794 | # echo 1000 > instances/foo/buffer_size_kb |
| 2795 | # echo 5000 > instances/bar/per_cpu/cpu1/buffer_size_kb |
| 2796 | # echo function > current_trace |
| 2797 | # echo 1 > instances/foo/events/sched/sched_wakeup/enable |
| 2798 | # echo 1 > instances/foo/events/sched/sched_wakeup_new/enable |
| 2799 | # echo 1 > instances/foo/events/sched/sched_switch/enable |
| 2800 | # echo 1 > instances/bar/events/irq/enable |
| 2801 | # echo 1 > instances/zoot/events/syscalls/enable |
| 2802 | # cat trace_pipe |
| 2803 | CPU:2 [LOST 11745 EVENTS] |
| 2804 | bash-2044 [002] .... 10594.481032: _raw_spin_lock_irqsave <-get_page_from_freelist |
| 2805 | bash-2044 [002] d... 10594.481032: add_preempt_count <-_raw_spin_lock_irqsave |
| 2806 | bash-2044 [002] d..1 10594.481032: __rmqueue <-get_page_from_freelist |
| 2807 | bash-2044 [002] d..1 10594.481033: _raw_spin_unlock <-get_page_from_freelist |
| 2808 | bash-2044 [002] d..1 10594.481033: sub_preempt_count <-_raw_spin_unlock |
| 2809 | bash-2044 [002] d... 10594.481033: get_pageblock_flags_group <-get_pageblock_migratetype |
| 2810 | bash-2044 [002] d... 10594.481034: __mod_zone_page_state <-get_page_from_freelist |
| 2811 | bash-2044 [002] d... 10594.481034: zone_statistics <-get_page_from_freelist |
| 2812 | bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics |
| 2813 | bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics |
| 2814 | bash-2044 [002] .... 10594.481035: arch_dup_task_struct <-copy_process |
| 2815 | [...] |
| 2816 | |
| 2817 | # cat instances/foo/trace_pipe |
| 2818 | bash-1998 [000] d..4 136.676759: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000 |
| 2819 | bash-1998 [000] dN.4 136.676760: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000 |
| 2820 | <idle>-0 [003] d.h3 136.676906: sched_wakeup: comm=rcu_preempt pid=9 prio=120 success=1 target_cpu=003 |
| 2821 | <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 |
| 2822 | 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 |
| 2823 | bash-1998 [000] d..4 136.677014: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000 |
| 2824 | bash-1998 [000] dN.4 136.677016: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000 |
| 2825 | 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 |
| 2826 | kworker/0:1-59 [000] d..4 136.677022: sched_wakeup: comm=sshd pid=1995 prio=120 success=1 target_cpu=001 |
| 2827 | 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 |
| 2828 | [...] |
| 2829 | |
| 2830 | # cat instances/bar/trace_pipe |
| 2831 | migration/1-14 [001] d.h3 138.732674: softirq_raise: vec=3 [action=NET_RX] |
| 2832 | <idle>-0 [001] dNh3 138.732725: softirq_raise: vec=3 [action=NET_RX] |
| 2833 | bash-1998 [000] d.h1 138.733101: softirq_raise: vec=1 [action=TIMER] |
| 2834 | bash-1998 [000] d.h1 138.733102: softirq_raise: vec=9 [action=RCU] |
| 2835 | bash-1998 [000] ..s2 138.733105: softirq_entry: vec=1 [action=TIMER] |
| 2836 | bash-1998 [000] ..s2 138.733106: softirq_exit: vec=1 [action=TIMER] |
| 2837 | bash-1998 [000] ..s2 138.733106: softirq_entry: vec=9 [action=RCU] |
| 2838 | bash-1998 [000] ..s2 138.733109: softirq_exit: vec=9 [action=RCU] |
| 2839 | sshd-1995 [001] d.h1 138.733278: irq_handler_entry: irq=21 name=uhci_hcd:usb4 |
| 2840 | sshd-1995 [001] d.h1 138.733280: irq_handler_exit: irq=21 ret=unhandled |
| 2841 | sshd-1995 [001] d.h1 138.733281: irq_handler_entry: irq=21 name=eth0 |
| 2842 | sshd-1995 [001] d.h1 138.733283: irq_handler_exit: irq=21 ret=handled |
| 2843 | [...] |
| 2844 | |
| 2845 | # cat instances/zoot/trace |
| 2846 | # tracer: nop |
| 2847 | # |
| 2848 | # entries-in-buffer/entries-written: 18996/18996 #P:4 |
| 2849 | # |
| 2850 | # _-----=> irqs-off |
| 2851 | # / _----=> need-resched |
| 2852 | # | / _---=> hardirq/softirq |
| 2853 | # || / _--=> preempt-depth |
| 2854 | # ||| / delay |
| 2855 | # TASK-PID CPU# |||| TIMESTAMP FUNCTION |
| 2856 | # | | | |||| | | |
| 2857 | bash-1998 [000] d... 140.733501: sys_write -> 0x2 |
| 2858 | bash-1998 [000] d... 140.733504: sys_dup2(oldfd: a, newfd: 1) |
| 2859 | bash-1998 [000] d... 140.733506: sys_dup2 -> 0x1 |
| 2860 | bash-1998 [000] d... 140.733508: sys_fcntl(fd: a, cmd: 1, arg: 0) |
| 2861 | bash-1998 [000] d... 140.733509: sys_fcntl -> 0x1 |
| 2862 | bash-1998 [000] d... 140.733510: sys_close(fd: a) |
| 2863 | bash-1998 [000] d... 140.733510: sys_close -> 0x0 |
| 2864 | bash-1998 [000] d... 140.733514: sys_rt_sigprocmask(how: 0, nset: 0, oset: 6e2768, sigsetsize: 8) |
| 2865 | bash-1998 [000] d... 140.733515: sys_rt_sigprocmask -> 0x0 |
| 2866 | bash-1998 [000] d... 140.733516: sys_rt_sigaction(sig: 2, act: 7fff718846f0, oact: 7fff71884650, sigsetsize: 8) |
| 2867 | bash-1998 [000] d... 140.733516: sys_rt_sigaction -> 0x0 |
| 2868 | |
| 2869 | You can see that the trace of the top most trace buffer shows only |
| 2870 | the function tracing. The foo instance displays wakeups and task |
| 2871 | switches. |
| 2872 | |
| 2873 | To remove the instances, simply delete their directories: |
| 2874 | |
| 2875 | # rmdir instances/foo |
| 2876 | # rmdir instances/bar |
| 2877 | # rmdir instances/zoot |
| 2878 | |
| 2879 | Note, if a process has a trace file open in one of the instance |
| 2880 | directories, the rmdir will fail with EBUSY. |
| 2881 | |
| 2882 | |
| 2883 | Stack trace |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2884 | ----------- |
Steven Rostedt (Red Hat) | 8d01609 | 2013-03-13 11:05:11 -0400 | [diff] [blame] | 2885 | Since the kernel has a fixed sized stack, it is important not to |
| 2886 | waste it in functions. A kernel developer must be conscience of |
| 2887 | what they allocate on the stack. If they add too much, the system |
| 2888 | can be in danger of a stack overflow, and corruption will occur, |
| 2889 | usually leading to a system panic. |
| 2890 | |
| 2891 | There are some tools that check this, usually with interrupts |
| 2892 | periodically checking usage. But if you can perform a check |
| 2893 | at every function call that will become very useful. As ftrace provides |
| 2894 | a function tracer, it makes it convenient to check the stack size |
| 2895 | at every function call. This is enabled via the stack tracer. |
| 2896 | |
| 2897 | CONFIG_STACK_TRACER enables the ftrace stack tracing functionality. |
| 2898 | To enable it, write a '1' into /proc/sys/kernel/stack_tracer_enabled. |
| 2899 | |
| 2900 | # echo 1 > /proc/sys/kernel/stack_tracer_enabled |
| 2901 | |
| 2902 | You can also enable it from the kernel command line to trace |
| 2903 | the stack size of the kernel during boot up, by adding "stacktrace" |
| 2904 | to the kernel command line parameter. |
| 2905 | |
| 2906 | After running it for a few minutes, the output looks like: |
| 2907 | |
| 2908 | # cat stack_max_size |
| 2909 | 2928 |
| 2910 | |
| 2911 | # cat stack_trace |
| 2912 | Depth Size Location (18 entries) |
| 2913 | ----- ---- -------- |
| 2914 | 0) 2928 224 update_sd_lb_stats+0xbc/0x4ac |
| 2915 | 1) 2704 160 find_busiest_group+0x31/0x1f1 |
| 2916 | 2) 2544 256 load_balance+0xd9/0x662 |
| 2917 | 3) 2288 80 idle_balance+0xbb/0x130 |
| 2918 | 4) 2208 128 __schedule+0x26e/0x5b9 |
| 2919 | 5) 2080 16 schedule+0x64/0x66 |
| 2920 | 6) 2064 128 schedule_timeout+0x34/0xe0 |
| 2921 | 7) 1936 112 wait_for_common+0x97/0xf1 |
| 2922 | 8) 1824 16 wait_for_completion+0x1d/0x1f |
| 2923 | 9) 1808 128 flush_work+0xfe/0x119 |
| 2924 | 10) 1680 16 tty_flush_to_ldisc+0x1e/0x20 |
| 2925 | 11) 1664 48 input_available_p+0x1d/0x5c |
| 2926 | 12) 1616 48 n_tty_poll+0x6d/0x134 |
| 2927 | 13) 1568 64 tty_poll+0x64/0x7f |
| 2928 | 14) 1504 880 do_select+0x31e/0x511 |
| 2929 | 15) 624 400 core_sys_select+0x177/0x216 |
| 2930 | 16) 224 96 sys_select+0x91/0xb9 |
| 2931 | 17) 128 128 system_call_fastpath+0x16/0x1b |
| 2932 | |
| 2933 | Note, if -mfentry is being used by gcc, functions get traced before |
| 2934 | they set up the stack frame. This means that leaf level functions |
| 2935 | are not tested by the stack tracer when -mfentry is used. |
| 2936 | |
| 2937 | Currently, -mfentry is used by gcc 4.6.0 and above on x86 only. |
| 2938 | |
| 2939 | --------- |
Ingo Molnar | 5752674 | 2009-02-19 12:54:10 +0100 | [diff] [blame] | 2940 | |
| 2941 | More details can be found in the source code, in the |
GeunSik Lim | baf20b3 | 2009-06-01 10:49:41 +0200 | [diff] [blame] | 2942 | kernel/trace/*.c files. |