1======================== 2ftrace - Function Tracer 3======================== 4 5Copyright 2008 Red Hat Inc. 6 7:Author: Steven Rostedt <srostedt@redhat.com> 8:License: The GNU Free Documentation License, Version 1.2 9 (dual licensed under the GPL v2) 10:Original Reviewers: Elias Oltmanns, Randy Dunlap, Andrew Morton, 11 John Kacur, and David Teigland. 12 13- Written for: 2.6.28-rc2 14- Updated for: 3.10 15- Updated for: 4.13 - Copyright 2017 VMware Inc. Steven Rostedt 16- Converted to rst format - Changbin Du <changbin.du@intel.com> 17 18Introduction 19------------ 20 21Ftrace is an internal tracer designed to help out developers and 22designers of systems to find what is going on inside the kernel. 23It can be used for debugging or analyzing latencies and 24performance issues that take place outside of user-space. 25 26Although ftrace is typically considered the function tracer, it 27is really a framework of several assorted tracing utilities. 28There's latency tracing to examine what occurs between interrupts 29disabled and enabled, as well as for preemption and from a time 30a task is woken to the task is actually scheduled in. 31 32One of the most common uses of ftrace is the event tracing. 33Throughout the kernel is hundreds of static event points that 34can be enabled via the tracefs file system to see what is 35going on in certain parts of the kernel. 36 37See events.rst for more information. 38 39 40Implementation Details 41---------------------- 42 43See :doc:`ftrace-design` for details for arch porters and such. 44 45 46The File System 47--------------- 48 49Ftrace uses the tracefs file system to hold the control files as 50well as the files to display output. 51 52When tracefs is configured into the kernel (which selecting any ftrace 53option will do) the directory /sys/kernel/tracing will be created. To mount 54this directory, you can add to your /etc/fstab file:: 55 56 tracefs /sys/kernel/tracing tracefs defaults 0 0 57 58Or you can mount it at run time with:: 59 60 mount -t tracefs nodev /sys/kernel/tracing 61 62For quicker access to that directory you may want to make a soft link to 63it:: 64 65 ln -s /sys/kernel/tracing /tracing 66 67.. attention:: 68 69 Before 4.1, all ftrace tracing control files were within the debugfs 70 file system, which is typically located at /sys/kernel/debug/tracing. 71 For backward compatibility, when mounting the debugfs file system, 72 the tracefs file system will be automatically mounted at: 73 74 /sys/kernel/debug/tracing 75 76 All files located in the tracefs file system will be located in that 77 debugfs file system directory as well. 78 79.. attention:: 80 81 Any selected ftrace option will also create the tracefs file system. 82 The rest of the document will assume that you are in the ftrace directory 83 (cd /sys/kernel/tracing) and will only concentrate on the files within that 84 directory and not distract from the content with the extended 85 "/sys/kernel/tracing" path name. 86 87That's it! (assuming that you have ftrace configured into your kernel) 88 89After mounting tracefs you will have access to the control and output files 90of ftrace. Here is a list of some of the key files: 91 92 93 Note: all time values are in microseconds. 94 95 current_tracer: 96 97 This is used to set or display the current tracer 98 that is configured. Changing the current tracer clears 99 the ring buffer content as well as the "snapshot" buffer. 100 101 available_tracers: 102 103 This holds the different types of tracers that 104 have been compiled into the kernel. The 105 tracers listed here can be configured by 106 echoing their name into current_tracer. 107 108 tracing_on: 109 110 This sets or displays whether writing to the trace 111 ring buffer is enabled. Echo 0 into this file to disable 112 the tracer or 1 to enable it. Note, this only disables 113 writing to the ring buffer, the tracing overhead may 114 still be occurring. 115 116 The kernel function tracing_off() can be used within the 117 kernel to disable writing to the ring buffer, which will 118 set this file to "0". User space can re-enable tracing by 119 echoing "1" into the file. 120 121 Note, the function and event trigger "traceoff" will also 122 set this file to zero and stop tracing. Which can also 123 be re-enabled by user space using this file. 124 125 trace: 126 127 This file holds the output of the trace in a human 128 readable format (described below). Opening this file for 129 writing with the O_TRUNC flag clears the ring buffer content. 130 Note, this file is not a consumer. If tracing is off 131 (no tracer running, or tracing_on is zero), it will produce 132 the same output each time it is read. When tracing is on, 133 it may produce inconsistent results as it tries to read 134 the entire buffer without consuming it. 135 136 trace_pipe: 137 138 The output is the same as the "trace" file but this 139 file is meant to be streamed with live tracing. 140 Reads from this file will block until new data is 141 retrieved. Unlike the "trace" file, this file is a 142 consumer. This means reading from this file causes 143 sequential reads to display more current data. Once 144 data is read from this file, it is consumed, and 145 will not be read again with a sequential read. The 146 "trace" file is static, and if the tracer is not 147 adding more data, it will display the same 148 information every time it is read. 149 150 trace_options: 151 152 This file lets the user control the amount of data 153 that is displayed in one of the above output 154 files. Options also exist to modify how a tracer 155 or events work (stack traces, timestamps, etc). 156 157 options: 158 159 This is a directory that has a file for every available 160 trace option (also in trace_options). Options may also be set 161 or cleared by writing a "1" or "0" respectively into the 162 corresponding file with the option name. 163 164 tracing_max_latency: 165 166 Some of the tracers record the max latency. 167 For example, the maximum time that interrupts are disabled. 168 The maximum time is saved in this file. The max trace will also be 169 stored, and displayed by "trace". A new max trace will only be 170 recorded if the latency is greater than the value in this file 171 (in microseconds). 172 173 By echoing in a time into this file, no latency will be recorded 174 unless it is greater than the time in this file. 175 176 tracing_thresh: 177 178 Some latency tracers will record a trace whenever the 179 latency is greater than the number in this file. 180 Only active when the file contains a number greater than 0. 181 (in microseconds) 182 183 buffer_size_kb: 184 185 This sets or displays the number of kilobytes each CPU 186 buffer holds. By default, the trace buffers are the same size 187 for each CPU. The displayed number is the size of the 188 CPU buffer and not total size of all buffers. The 189 trace buffers are allocated in pages (blocks of memory 190 that the kernel uses for allocation, usually 4 KB in size). 191 A few extra pages may be allocated to accommodate buffer management 192 meta-data. If the last page allocated has room for more bytes 193 than requested, the rest of the page will be used, 194 making the actual allocation bigger than requested or shown. 195 ( Note, the size may not be a multiple of the page size 196 due to buffer management meta-data. ) 197 198 Buffer sizes for individual CPUs may vary 199 (see "per_cpu/cpu0/buffer_size_kb" below), and if they do 200 this file will show "X". 201 202 buffer_total_size_kb: 203 204 This displays the total combined size of all the trace buffers. 205 206 free_buffer: 207 208 If a process is performing tracing, and the ring buffer should be 209 shrunk "freed" when the process is finished, even if it were to be 210 killed by a signal, this file can be used for that purpose. On close 211 of this file, the ring buffer will be resized to its minimum size. 212 Having a process that is tracing also open this file, when the process 213 exits its file descriptor for this file will be closed, and in doing so, 214 the ring buffer will be "freed". 215 216 It may also stop tracing if disable_on_free option is set. 217 218 tracing_cpumask: 219 220 This is a mask that lets the user only trace on specified CPUs. 221 The format is a hex string representing the CPUs. 222 223 set_ftrace_filter: 224 225 When dynamic ftrace is configured in (see the 226 section below "dynamic ftrace"), the code is dynamically 227 modified (code text rewrite) to disable calling of the 228 function profiler (mcount). This lets tracing be configured 229 in with practically no overhead in performance. This also 230 has a side effect of enabling or disabling specific functions 231 to be traced. Echoing names of functions into this file 232 will limit the trace to only those functions. 233 This influences the tracers "function" and "function_graph" 234 and thus also function profiling (see "function_profile_enabled"). 235 236 The functions listed in "available_filter_functions" are what 237 can be written into this file. 238 239 This interface also allows for commands to be used. See the 240 "Filter commands" section for more details. 241 242 As a speed up, since processing strings can be quite expensive 243 and requires a check of all functions registered to tracing, instead 244 an index can be written into this file. A number (starting with "1") 245 written will instead select the same corresponding at the line position 246 of the "available_filter_functions" file. 247 248 set_ftrace_notrace: 249 250 This has an effect opposite to that of 251 set_ftrace_filter. Any function that is added here will not 252 be traced. If a function exists in both set_ftrace_filter 253 and set_ftrace_notrace, the function will _not_ be traced. 254 255 set_ftrace_pid: 256 257 Have the function tracer only trace the threads whose PID are 258 listed in this file. 259 260 If the "function-fork" option is set, then when a task whose 261 PID is listed in this file forks, the child's PID will 262 automatically be added to this file, and the child will be 263 traced by the function tracer as well. This option will also 264 cause PIDs of tasks that exit to be removed from the file. 265 266 set_ftrace_notrace_pid: 267 268 Have the function tracer ignore threads whose PID are listed in 269 this file. 270 271 If the "function-fork" option is set, then when a task whose 272 PID is listed in this file forks, the child's PID will 273 automatically be added to this file, and the child will not be 274 traced by the function tracer as well. This option will also 275 cause PIDs of tasks that exit to be removed from the file. 276 277 If a PID is in both this file and "set_ftrace_pid", then this 278 file takes precedence, and the thread will not be traced. 279 280 set_event_pid: 281 282 Have the events only trace a task with a PID listed in this file. 283 Note, sched_switch and sched_wake_up will also trace events 284 listed in this file. 285 286 To have the PIDs of children of tasks with their PID in this file 287 added on fork, enable the "event-fork" option. That option will also 288 cause the PIDs of tasks to be removed from this file when the task 289 exits. 290 291 set_event_notrace_pid: 292 293 Have the events not trace a task with a PID listed in this file. 294 Note, sched_switch and sched_wakeup will trace threads not listed 295 in this file, even if a thread's PID is in the file if the 296 sched_switch or sched_wakeup events also trace a thread that should 297 be traced. 298 299 To have the PIDs of children of tasks with their PID in this file 300 added on fork, enable the "event-fork" option. That option will also 301 cause the PIDs of tasks to be removed from this file when the task 302 exits. 303 304 set_graph_function: 305 306 Functions listed in this file will cause the function graph 307 tracer to only trace these functions and the functions that 308 they call. (See the section "dynamic ftrace" for more details). 309 Note, set_ftrace_filter and set_ftrace_notrace still affects 310 what functions are being traced. 311 312 set_graph_notrace: 313 314 Similar to set_graph_function, but will disable function graph 315 tracing when the function is hit until it exits the function. 316 This makes it possible to ignore tracing functions that are called 317 by a specific function. 318 319 available_filter_functions: 320 321 This lists the functions that ftrace has processed and can trace. 322 These are the function names that you can pass to 323 "set_ftrace_filter", "set_ftrace_notrace", 324 "set_graph_function", or "set_graph_notrace". 325 (See the section "dynamic ftrace" below for more details.) 326 327 dyn_ftrace_total_info: 328 329 This file is for debugging purposes. The number of functions that 330 have been converted to nops and are available to be traced. 331 332 enabled_functions: 333 334 This file is more for debugging ftrace, but can also be useful 335 in seeing if any function has a callback attached to it. 336 Not only does the trace infrastructure use ftrace function 337 trace utility, but other subsystems might too. This file 338 displays all functions that have a callback attached to them 339 as well as the number of callbacks that have been attached. 340 Note, a callback may also call multiple functions which will 341 not be listed in this count. 342 343 If the callback registered to be traced by a function with 344 the "save regs" attribute (thus even more overhead), a 'R' 345 will be displayed on the same line as the function that 346 is returning registers. 347 348 If the callback registered to be traced by a function with 349 the "ip modify" attribute (thus the regs->ip can be changed), 350 an 'I' will be displayed on the same line as the function that 351 can be overridden. 352 353 If the architecture supports it, it will also show what callback 354 is being directly called by the function. If the count is greater 355 than 1 it most likely will be ftrace_ops_list_func(). 356 357 If the callback of the function jumps to a trampoline that is 358 specific to a the callback and not the standard trampoline, 359 its address will be printed as well as the function that the 360 trampoline calls. 361 362 function_profile_enabled: 363 364 When set it will enable all functions with either the function 365 tracer, or if configured, the function graph tracer. It will 366 keep a histogram of the number of functions that were called 367 and if the function graph tracer was configured, it will also keep 368 track of the time spent in those functions. The histogram 369 content can be displayed in the files: 370 371 trace_stat/function<cpu> ( function0, function1, etc). 372 373 trace_stat: 374 375 A directory that holds different tracing stats. 376 377 kprobe_events: 378 379 Enable dynamic trace points. See kprobetrace.rst. 380 381 kprobe_profile: 382 383 Dynamic trace points stats. See kprobetrace.rst. 384 385 max_graph_depth: 386 387 Used with the function graph tracer. This is the max depth 388 it will trace into a function. Setting this to a value of 389 one will show only the first kernel function that is called 390 from user space. 391 392 printk_formats: 393 394 This is for tools that read the raw format files. If an event in 395 the ring buffer references a string, only a pointer to the string 396 is recorded into the buffer and not the string itself. This prevents 397 tools from knowing what that string was. This file displays the string 398 and address for the string allowing tools to map the pointers to what 399 the strings were. 400 401 saved_cmdlines: 402 403 Only the pid of the task is recorded in a trace event unless 404 the event specifically saves the task comm as well. Ftrace 405 makes a cache of pid mappings to comms to try to display 406 comms for events. If a pid for a comm is not listed, then 407 "<...>" is displayed in the output. 408 409 If the option "record-cmd" is set to "0", then comms of tasks 410 will not be saved during recording. By default, it is enabled. 411 412 saved_cmdlines_size: 413 414 By default, 128 comms are saved (see "saved_cmdlines" above). To 415 increase or decrease the amount of comms that are cached, echo 416 the number of comms to cache into this file. 417 418 saved_tgids: 419 420 If the option "record-tgid" is set, on each scheduling context switch 421 the Task Group ID of a task is saved in a table mapping the PID of 422 the thread to its TGID. By default, the "record-tgid" option is 423 disabled. 424 425 snapshot: 426 427 This displays the "snapshot" buffer and also lets the user 428 take a snapshot of the current running trace. 429 See the "Snapshot" section below for more details. 430 431 stack_max_size: 432 433 When the stack tracer is activated, this will display the 434 maximum stack size it has encountered. 435 See the "Stack Trace" section below. 436 437 stack_trace: 438 439 This displays the stack back trace of the largest stack 440 that was encountered when the stack tracer is activated. 441 See the "Stack Trace" section below. 442 443 stack_trace_filter: 444 445 This is similar to "set_ftrace_filter" but it limits what 446 functions the stack tracer will check. 447 448 trace_clock: 449 450 Whenever an event is recorded into the ring buffer, a 451 "timestamp" is added. This stamp comes from a specified 452 clock. By default, ftrace uses the "local" clock. This 453 clock is very fast and strictly per cpu, but on some 454 systems it may not be monotonic with respect to other 455 CPUs. In other words, the local clocks may not be in sync 456 with local clocks on other CPUs. 457 458 Usual clocks for tracing:: 459 460 # cat trace_clock 461 [local] global counter x86-tsc 462 463 The clock with the square brackets around it is the one in effect. 464 465 local: 466 Default clock, but may not be in sync across CPUs 467 468 global: 469 This clock is in sync with all CPUs but may 470 be a bit slower than the local clock. 471 472 counter: 473 This is not a clock at all, but literally an atomic 474 counter. It counts up one by one, but is in sync 475 with all CPUs. This is useful when you need to 476 know exactly the order events occurred with respect to 477 each other on different CPUs. 478 479 uptime: 480 This uses the jiffies counter and the time stamp 481 is relative to the time since boot up. 482 483 perf: 484 This makes ftrace use the same clock that perf uses. 485 Eventually perf will be able to read ftrace buffers 486 and this will help out in interleaving the data. 487 488 x86-tsc: 489 Architectures may define their own clocks. For 490 example, x86 uses its own TSC cycle clock here. 491 492 ppc-tb: 493 This uses the powerpc timebase register value. 494 This is in sync across CPUs and can also be used 495 to correlate events across hypervisor/guest if 496 tb_offset is known. 497 498 mono: 499 This uses the fast monotonic clock (CLOCK_MONOTONIC) 500 which is monotonic and is subject to NTP rate adjustments. 501 502 mono_raw: 503 This is the raw monotonic clock (CLOCK_MONOTONIC_RAW) 504 which is monotonic but is not subject to any rate adjustments 505 and ticks at the same rate as the hardware clocksource. 506 507 boot: 508 This is the boot clock (CLOCK_BOOTTIME) and is based on the 509 fast monotonic clock, but also accounts for time spent in 510 suspend. Since the clock access is designed for use in 511 tracing in the suspend path, some side effects are possible 512 if clock is accessed after the suspend time is accounted before 513 the fast mono clock is updated. In this case, the clock update 514 appears to happen slightly sooner than it normally would have. 515 Also on 32-bit systems, it's possible that the 64-bit boot offset 516 sees a partial update. These effects are rare and post 517 processing should be able to handle them. See comments in the 518 ktime_get_boot_fast_ns() function for more information. 519 520 To set a clock, simply echo the clock name into this file:: 521 522 # echo global > trace_clock 523 524 Setting a clock clears the ring buffer content as well as the 525 "snapshot" buffer. 526 527 trace_marker: 528 529 This is a very useful file for synchronizing user space 530 with events happening in the kernel. Writing strings into 531 this file will be written into the ftrace buffer. 532 533 It is useful in applications to open this file at the start 534 of the application and just reference the file descriptor 535 for the file:: 536 537 void trace_write(const char *fmt, ...) 538 { 539 va_list ap; 540 char buf[256]; 541 int n; 542 543 if (trace_fd < 0) 544 return; 545 546 va_start(ap, fmt); 547 n = vsnprintf(buf, 256, fmt, ap); 548 va_end(ap); 549 550 write(trace_fd, buf, n); 551 } 552 553 start:: 554 555 trace_fd = open("trace_marker", WR_ONLY); 556 557 Note: Writing into the trace_marker file can also initiate triggers 558 that are written into /sys/kernel/tracing/events/ftrace/print/trigger 559 See "Event triggers" in Documentation/trace/events.rst and an 560 example in Documentation/trace/histogram.rst (Section 3.) 561 562 trace_marker_raw: 563 564 This is similar to trace_marker above, but is meant for binary data 565 to be written to it, where a tool can be used to parse the data 566 from trace_pipe_raw. 567 568 uprobe_events: 569 570 Add dynamic tracepoints in programs. 571 See uprobetracer.rst 572 573 uprobe_profile: 574 575 Uprobe statistics. See uprobetrace.txt 576 577 instances: 578 579 This is a way to make multiple trace buffers where different 580 events can be recorded in different buffers. 581 See "Instances" section below. 582 583 events: 584 585 This is the trace event directory. It holds event tracepoints 586 (also known as static tracepoints) that have been compiled 587 into the kernel. It shows what event tracepoints exist 588 and how they are grouped by system. There are "enable" 589 files at various levels that can enable the tracepoints 590 when a "1" is written to them. 591 592 See events.rst for more information. 593 594 set_event: 595 596 By echoing in the event into this file, will enable that event. 597 598 See events.rst for more information. 599 600 available_events: 601 602 A list of events that can be enabled in tracing. 603 604 See events.rst for more information. 605 606 timestamp_mode: 607 608 Certain tracers may change the timestamp mode used when 609 logging trace events into the event buffer. Events with 610 different modes can coexist within a buffer but the mode in 611 effect when an event is logged determines which timestamp mode 612 is used for that event. The default timestamp mode is 613 'delta'. 614 615 Usual timestamp modes for tracing: 616 617 # cat timestamp_mode 618 [delta] absolute 619 620 The timestamp mode with the square brackets around it is the 621 one in effect. 622 623 delta: Default timestamp mode - timestamp is a delta against 624 a per-buffer timestamp. 625 626 absolute: The timestamp is a full timestamp, not a delta 627 against some other value. As such it takes up more 628 space and is less efficient. 629 630 hwlat_detector: 631 632 Directory for the Hardware Latency Detector. 633 See "Hardware Latency Detector" section below. 634 635 per_cpu: 636 637 This is a directory that contains the trace per_cpu information. 638 639 per_cpu/cpu0/buffer_size_kb: 640 641 The ftrace buffer is defined per_cpu. That is, there's a separate 642 buffer for each CPU to allow writes to be done atomically, 643 and free from cache bouncing. These buffers may have different 644 size buffers. This file is similar to the buffer_size_kb 645 file, but it only displays or sets the buffer size for the 646 specific CPU. (here cpu0). 647 648 per_cpu/cpu0/trace: 649 650 This is similar to the "trace" file, but it will only display 651 the data specific for the CPU. If written to, it only clears 652 the specific CPU buffer. 653 654 per_cpu/cpu0/trace_pipe 655 656 This is similar to the "trace_pipe" file, and is a consuming 657 read, but it will only display (and consume) the data specific 658 for the CPU. 659 660 per_cpu/cpu0/trace_pipe_raw 661 662 For tools that can parse the ftrace ring buffer binary format, 663 the trace_pipe_raw file can be used to extract the data 664 from the ring buffer directly. With the use of the splice() 665 system call, the buffer data can be quickly transferred to 666 a file or to the network where a server is collecting the 667 data. 668 669 Like trace_pipe, this is a consuming reader, where multiple 670 reads will always produce different data. 671 672 per_cpu/cpu0/snapshot: 673 674 This is similar to the main "snapshot" file, but will only 675 snapshot the current CPU (if supported). It only displays 676 the content of the snapshot for a given CPU, and if 677 written to, only clears this CPU buffer. 678 679 per_cpu/cpu0/snapshot_raw: 680 681 Similar to the trace_pipe_raw, but will read the binary format 682 from the snapshot buffer for the given CPU. 683 684 per_cpu/cpu0/stats: 685 686 This displays certain stats about the ring buffer: 687 688 entries: 689 The number of events that are still in the buffer. 690 691 overrun: 692 The number of lost events due to overwriting when 693 the buffer was full. 694 695 commit overrun: 696 Should always be zero. 697 This gets set if so many events happened within a nested 698 event (ring buffer is re-entrant), that it fills the 699 buffer and starts dropping events. 700 701 bytes: 702 Bytes actually read (not overwritten). 703 704 oldest event ts: 705 The oldest timestamp in the buffer 706 707 now ts: 708 The current timestamp 709 710 dropped events: 711 Events lost due to overwrite option being off. 712 713 read events: 714 The number of events read. 715 716The Tracers 717----------- 718 719Here is the list of current tracers that may be configured. 720 721 "function" 722 723 Function call tracer to trace all kernel functions. 724 725 "function_graph" 726 727 Similar to the function tracer except that the 728 function tracer probes the functions on their entry 729 whereas the function graph tracer traces on both entry 730 and exit of the functions. It then provides the ability 731 to draw a graph of function calls similar to C code 732 source. 733 734 "blk" 735 736 The block tracer. The tracer used by the blktrace user 737 application. 738 739 "hwlat" 740 741 The Hardware Latency tracer is used to detect if the hardware 742 produces any latency. See "Hardware Latency Detector" section 743 below. 744 745 "irqsoff" 746 747 Traces the areas that disable interrupts and saves 748 the trace with the longest max latency. 749 See tracing_max_latency. When a new max is recorded, 750 it replaces the old trace. It is best to view this 751 trace with the latency-format option enabled, which 752 happens automatically when the tracer is selected. 753 754 "preemptoff" 755 756 Similar to irqsoff but traces and records the amount of 757 time for which preemption is disabled. 758 759 "preemptirqsoff" 760 761 Similar to irqsoff and preemptoff, but traces and 762 records the largest time for which irqs and/or preemption 763 is disabled. 764 765 "wakeup" 766 767 Traces and records the max latency that it takes for 768 the highest priority task to get scheduled after 769 it has been woken up. 770 Traces all tasks as an average developer would expect. 771 772 "wakeup_rt" 773 774 Traces and records the max latency that it takes for just 775 RT tasks (as the current "wakeup" does). This is useful 776 for those interested in wake up timings of RT tasks. 777 778 "wakeup_dl" 779 780 Traces and records the max latency that it takes for 781 a SCHED_DEADLINE task to be woken (as the "wakeup" and 782 "wakeup_rt" does). 783 784 "mmiotrace" 785 786 A special tracer that is used to trace binary module. 787 It will trace all the calls that a module makes to the 788 hardware. Everything it writes and reads from the I/O 789 as well. 790 791 "branch" 792 793 This tracer can be configured when tracing likely/unlikely 794 calls within the kernel. It will trace when a likely and 795 unlikely branch is hit and if it was correct in its prediction 796 of being correct. 797 798 "nop" 799 800 This is the "trace nothing" tracer. To remove all 801 tracers from tracing simply echo "nop" into 802 current_tracer. 803 804Error conditions 805---------------- 806 807 For most ftrace commands, failure modes are obvious and communicated 808 using standard return codes. 809 810 For other more involved commands, extended error information may be 811 available via the tracing/error_log file. For the commands that 812 support it, reading the tracing/error_log file after an error will 813 display more detailed information about what went wrong, if 814 information is available. The tracing/error_log file is a circular 815 error log displaying a small number (currently, 8) of ftrace errors 816 for the last (8) failed commands. 817 818 The extended error information and usage takes the form shown in 819 this example:: 820 821 # echo xxx > /sys/kernel/debug/tracing/events/sched/sched_wakeup/trigger 822 echo: write error: Invalid argument 823 824 # cat /sys/kernel/debug/tracing/error_log 825 [ 5348.887237] location: error: Couldn't yyy: zzz 826 Command: xxx 827 ^ 828 [ 7517.023364] location: error: Bad rrr: sss 829 Command: ppp qqq 830 ^ 831 832 To clear the error log, echo the empty string into it:: 833 834 # echo > /sys/kernel/debug/tracing/error_log 835 836Examples of using the tracer 837---------------------------- 838 839Here are typical examples of using the tracers when controlling 840them only with the tracefs interface (without using any 841user-land utilities). 842 843Output format: 844-------------- 845 846Here is an example of the output format of the file "trace":: 847 848 # tracer: function 849 # 850 # entries-in-buffer/entries-written: 140080/250280 #P:4 851 # 852 # _-----=> irqs-off 853 # / _----=> need-resched 854 # | / _---=> hardirq/softirq 855 # || / _--=> preempt-depth 856 # ||| / delay 857 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 858 # | | | |||| | | 859 bash-1977 [000] .... 17284.993652: sys_close <-system_call_fastpath 860 bash-1977 [000] .... 17284.993653: __close_fd <-sys_close 861 bash-1977 [000] .... 17284.993653: _raw_spin_lock <-__close_fd 862 sshd-1974 [003] .... 17284.993653: __srcu_read_unlock <-fsnotify 863 bash-1977 [000] .... 17284.993654: add_preempt_count <-_raw_spin_lock 864 bash-1977 [000] ...1 17284.993655: _raw_spin_unlock <-__close_fd 865 bash-1977 [000] ...1 17284.993656: sub_preempt_count <-_raw_spin_unlock 866 bash-1977 [000] .... 17284.993657: filp_close <-__close_fd 867 bash-1977 [000] .... 17284.993657: dnotify_flush <-filp_close 868 sshd-1974 [003] .... 17284.993658: sys_select <-system_call_fastpath 869 .... 870 871A header is printed with the tracer name that is represented by 872the trace. In this case the tracer is "function". Then it shows the 873number of events in the buffer as well as the total number of entries 874that were written. The difference is the number of entries that were 875lost due to the buffer filling up (250280 - 140080 = 110200 events 876lost). 877 878The header explains the content of the events. Task name "bash", the task 879PID "1977", the CPU that it was running on "000", the latency format 880(explained below), the timestamp in <secs>.<usecs> format, the 881function name that was traced "sys_close" and the parent function that 882called this function "system_call_fastpath". The timestamp is the time 883at which the function was entered. 884 885Latency trace format 886-------------------- 887 888When the latency-format option is enabled or when one of the latency 889tracers is set, the trace file gives somewhat more information to see 890why a latency happened. Here is a typical trace:: 891 892 # tracer: irqsoff 893 # 894 # irqsoff latency trace v1.1.5 on 3.8.0-test+ 895 # -------------------------------------------------------------------- 896 # latency: 259 us, #4/4, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 897 # ----------------- 898 # | task: ps-6143 (uid:0 nice:0 policy:0 rt_prio:0) 899 # ----------------- 900 # => started at: __lock_task_sighand 901 # => ended at: _raw_spin_unlock_irqrestore 902 # 903 # 904 # _------=> CPU# 905 # / _-----=> irqs-off 906 # | / _----=> need-resched 907 # || / _---=> hardirq/softirq 908 # ||| / _--=> preempt-depth 909 # |||| / delay 910 # cmd pid ||||| time | caller 911 # \ / ||||| \ | / 912 ps-6143 2d... 0us!: trace_hardirqs_off <-__lock_task_sighand 913 ps-6143 2d..1 259us+: trace_hardirqs_on <-_raw_spin_unlock_irqrestore 914 ps-6143 2d..1 263us+: time_hardirqs_on <-_raw_spin_unlock_irqrestore 915 ps-6143 2d..1 306us : <stack trace> 916 => trace_hardirqs_on_caller 917 => trace_hardirqs_on 918 => _raw_spin_unlock_irqrestore 919 => do_task_stat 920 => proc_tgid_stat 921 => proc_single_show 922 => seq_read 923 => vfs_read 924 => sys_read 925 => system_call_fastpath 926 927 928This shows that the current tracer is "irqsoff" tracing the time 929for which interrupts were disabled. It gives the trace version (which 930never changes) and the version of the kernel upon which this was executed on 931(3.8). Then it displays the max latency in microseconds (259 us). The number 932of trace entries displayed and the total number (both are four: #4/4). 933VP, KP, SP, and HP are always zero and are reserved for later use. 934#P is the number of online CPUs (#P:4). 935 936The task is the process that was running when the latency 937occurred. (ps pid: 6143). 938 939The start and stop (the functions in which the interrupts were 940disabled and enabled respectively) that caused the latencies: 941 942 - __lock_task_sighand is where the interrupts were disabled. 943 - _raw_spin_unlock_irqrestore is where they were enabled again. 944 945The next lines after the header are the trace itself. The header 946explains which is which. 947 948 cmd: The name of the process in the trace. 949 950 pid: The PID of that process. 951 952 CPU#: The CPU which the process was running on. 953 954 irqs-off: 'd' interrupts are disabled. '.' otherwise. 955 .. caution:: If the architecture does not support a way to 956 read the irq flags variable, an 'X' will always 957 be printed here. 958 959 need-resched: 960 - 'N' both TIF_NEED_RESCHED and PREEMPT_NEED_RESCHED is set, 961 - 'n' only TIF_NEED_RESCHED is set, 962 - 'p' only PREEMPT_NEED_RESCHED is set, 963 - '.' otherwise. 964 965 hardirq/softirq: 966 - 'Z' - NMI occurred inside a hardirq 967 - 'z' - NMI is running 968 - 'H' - hard irq occurred inside a softirq. 969 - 'h' - hard irq is running 970 - 's' - soft irq is running 971 - '.' - normal context. 972 973 preempt-depth: The level of preempt_disabled 974 975The above is mostly meaningful for kernel developers. 976 977 time: 978 When the latency-format option is enabled, the trace file 979 output includes a timestamp relative to the start of the 980 trace. This differs from the output when latency-format 981 is disabled, which includes an absolute timestamp. 982 983 delay: 984 This is just to help catch your eye a bit better. And 985 needs to be fixed to be only relative to the same CPU. 986 The marks are determined by the difference between this 987 current trace and the next trace. 988 989 - '$' - greater than 1 second 990 - '@' - greater than 100 millisecond 991 - '*' - greater than 10 millisecond 992 - '#' - greater than 1000 microsecond 993 - '!' - greater than 100 microsecond 994 - '+' - greater than 10 microsecond 995 - ' ' - less than or equal to 10 microsecond. 996 997 The rest is the same as the 'trace' file. 998 999 Note, the latency tracers will usually end with a back trace 1000 to easily find where the latency occurred. 1001 1002trace_options 1003------------- 1004 1005The trace_options file (or the options directory) is used to control 1006what gets printed in the trace output, or manipulate the tracers. 1007To see what is available, simply cat the file:: 1008 1009 cat trace_options 1010 print-parent 1011 nosym-offset 1012 nosym-addr 1013 noverbose 1014 noraw 1015 nohex 1016 nobin 1017 noblock 1018 trace_printk 1019 annotate 1020 nouserstacktrace 1021 nosym-userobj 1022 noprintk-msg-only 1023 context-info 1024 nolatency-format 1025 record-cmd 1026 norecord-tgid 1027 overwrite 1028 nodisable_on_free 1029 irq-info 1030 markers 1031 noevent-fork 1032 function-trace 1033 nofunction-fork 1034 nodisplay-graph 1035 nostacktrace 1036 nobranch 1037 1038To disable one of the options, echo in the option prepended with 1039"no":: 1040 1041 echo noprint-parent > trace_options 1042 1043To enable an option, leave off the "no":: 1044 1045 echo sym-offset > trace_options 1046 1047Here are the available options: 1048 1049 print-parent 1050 On function traces, display the calling (parent) 1051 function as well as the function being traced. 1052 :: 1053 1054 print-parent: 1055 bash-4000 [01] 1477.606694: simple_strtoul <-kstrtoul 1056 1057 noprint-parent: 1058 bash-4000 [01] 1477.606694: simple_strtoul 1059 1060 1061 sym-offset 1062 Display not only the function name, but also the 1063 offset in the function. For example, instead of 1064 seeing just "ktime_get", you will see 1065 "ktime_get+0xb/0x20". 1066 :: 1067 1068 sym-offset: 1069 bash-4000 [01] 1477.606694: simple_strtoul+0x6/0xa0 1070 1071 sym-addr 1072 This will also display the function address as well 1073 as the function name. 1074 :: 1075 1076 sym-addr: 1077 bash-4000 [01] 1477.606694: simple_strtoul <c0339346> 1078 1079 verbose 1080 This deals with the trace file when the 1081 latency-format option is enabled. 1082 :: 1083 1084 bash 4000 1 0 00000000 00010a95 [58127d26] 1720.415ms \ 1085 (+0.000ms): simple_strtoul (kstrtoul) 1086 1087 raw 1088 This will display raw numbers. This option is best for 1089 use with user applications that can translate the raw 1090 numbers better than having it done in the kernel. 1091 1092 hex 1093 Similar to raw, but the numbers will be in a hexadecimal format. 1094 1095 bin 1096 This will print out the formats in raw binary. 1097 1098 block 1099 When set, reading trace_pipe will not block when polled. 1100 1101 trace_printk 1102 Can disable trace_printk() from writing into the buffer. 1103 1104 annotate 1105 It is sometimes confusing when the CPU buffers are full 1106 and one CPU buffer had a lot of events recently, thus 1107 a shorter time frame, were another CPU may have only had 1108 a few events, which lets it have older events. When 1109 the trace is reported, it shows the oldest events first, 1110 and it may look like only one CPU ran (the one with the 1111 oldest events). When the annotate option is set, it will 1112 display when a new CPU buffer started:: 1113 1114 <idle>-0 [001] dNs4 21169.031481: wake_up_idle_cpu <-add_timer_on 1115 <idle>-0 [001] dNs4 21169.031482: _raw_spin_unlock_irqrestore <-add_timer_on 1116 <idle>-0 [001] .Ns4 21169.031484: sub_preempt_count <-_raw_spin_unlock_irqrestore 1117 ##### CPU 2 buffer started #### 1118 <idle>-0 [002] .N.1 21169.031484: rcu_idle_exit <-cpu_idle 1119 <idle>-0 [001] .Ns3 21169.031484: _raw_spin_unlock <-clocksource_watchdog 1120 <idle>-0 [001] .Ns3 21169.031485: sub_preempt_count <-_raw_spin_unlock 1121 1122 userstacktrace 1123 This option changes the trace. It records a 1124 stacktrace of the current user space thread after 1125 each trace event. 1126 1127 sym-userobj 1128 when user stacktrace are enabled, look up which 1129 object the address belongs to, and print a 1130 relative address. This is especially useful when 1131 ASLR is on, otherwise you don't get a chance to 1132 resolve the address to object/file/line after 1133 the app is no longer running 1134 1135 The lookup is performed when you read 1136 trace,trace_pipe. Example:: 1137 1138 a.out-1623 [000] 40874.465068: /root/a.out[+0x480] <-/root/a.out[+0 1139 x494] <- /root/a.out[+0x4a8] <- /lib/libc-2.7.so[+0x1e1a6] 1140 1141 1142 printk-msg-only 1143 When set, trace_printk()s will only show the format 1144 and not their parameters (if trace_bprintk() or 1145 trace_bputs() was used to save the trace_printk()). 1146 1147 context-info 1148 Show only the event data. Hides the comm, PID, 1149 timestamp, CPU, and other useful data. 1150 1151 latency-format 1152 This option changes the trace output. When it is enabled, 1153 the trace displays additional information about the 1154 latency, as described in "Latency trace format". 1155 1156 pause-on-trace 1157 When set, opening the trace file for read, will pause 1158 writing to the ring buffer (as if tracing_on was set to zero). 1159 This simulates the original behavior of the trace file. 1160 When the file is closed, tracing will be enabled again. 1161 1162 hash-ptr 1163 When set, "%p" in the event printk format displays the 1164 hashed pointer value instead of real address. 1165 This will be useful if you want to find out which hashed 1166 value is corresponding to the real value in trace log. 1167 1168 record-cmd 1169 When any event or tracer is enabled, a hook is enabled 1170 in the sched_switch trace point to fill comm cache 1171 with mapped pids and comms. But this may cause some 1172 overhead, and if you only care about pids, and not the 1173 name of the task, disabling this option can lower the 1174 impact of tracing. See "saved_cmdlines". 1175 1176 record-tgid 1177 When any event or tracer is enabled, a hook is enabled 1178 in the sched_switch trace point to fill the cache of 1179 mapped Thread Group IDs (TGID) mapping to pids. See 1180 "saved_tgids". 1181 1182 overwrite 1183 This controls what happens when the trace buffer is 1184 full. If "1" (default), the oldest events are 1185 discarded and overwritten. If "0", then the newest 1186 events are discarded. 1187 (see per_cpu/cpu0/stats for overrun and dropped) 1188 1189 disable_on_free 1190 When the free_buffer is closed, tracing will 1191 stop (tracing_on set to 0). 1192 1193 irq-info 1194 Shows the interrupt, preempt count, need resched data. 1195 When disabled, the trace looks like:: 1196 1197 # tracer: function 1198 # 1199 # entries-in-buffer/entries-written: 144405/9452052 #P:4 1200 # 1201 # TASK-PID CPU# TIMESTAMP FUNCTION 1202 # | | | | | 1203 <idle>-0 [002] 23636.756054: ttwu_do_activate.constprop.89 <-try_to_wake_up 1204 <idle>-0 [002] 23636.756054: activate_task <-ttwu_do_activate.constprop.89 1205 <idle>-0 [002] 23636.756055: enqueue_task <-activate_task 1206 1207 1208 markers 1209 When set, the trace_marker is writable (only by root). 1210 When disabled, the trace_marker will error with EINVAL 1211 on write. 1212 1213 event-fork 1214 When set, tasks with PIDs listed in set_event_pid will have 1215 the PIDs of their children added to set_event_pid when those 1216 tasks fork. Also, when tasks with PIDs in set_event_pid exit, 1217 their PIDs will be removed from the file. 1218 1219 This affects PIDs listed in set_event_notrace_pid as well. 1220 1221 function-trace 1222 The latency tracers will enable function tracing 1223 if this option is enabled (default it is). When 1224 it is disabled, the latency tracers do not trace 1225 functions. This keeps the overhead of the tracer down 1226 when performing latency tests. 1227 1228 function-fork 1229 When set, tasks with PIDs listed in set_ftrace_pid will 1230 have the PIDs of their children added to set_ftrace_pid 1231 when those tasks fork. Also, when tasks with PIDs in 1232 set_ftrace_pid exit, their PIDs will be removed from the 1233 file. 1234 1235 This affects PIDs in set_ftrace_notrace_pid as well. 1236 1237 display-graph 1238 When set, the latency tracers (irqsoff, wakeup, etc) will 1239 use function graph tracing instead of function tracing. 1240 1241 stacktrace 1242 When set, a stack trace is recorded after any trace event 1243 is recorded. 1244 1245 branch 1246 Enable branch tracing with the tracer. This enables branch 1247 tracer along with the currently set tracer. Enabling this 1248 with the "nop" tracer is the same as just enabling the 1249 "branch" tracer. 1250 1251.. tip:: Some tracers have their own options. They only appear in this 1252 file when the tracer is active. They always appear in the 1253 options directory. 1254 1255 1256Here are the per tracer options: 1257 1258Options for function tracer: 1259 1260 func_stack_trace 1261 When set, a stack trace is recorded after every 1262 function that is recorded. NOTE! Limit the functions 1263 that are recorded before enabling this, with 1264 "set_ftrace_filter" otherwise the system performance 1265 will be critically degraded. Remember to disable 1266 this option before clearing the function filter. 1267 1268Options for function_graph tracer: 1269 1270 Since the function_graph tracer has a slightly different output 1271 it has its own options to control what is displayed. 1272 1273 funcgraph-overrun 1274 When set, the "overrun" of the graph stack is 1275 displayed after each function traced. The 1276 overrun, is when the stack depth of the calls 1277 is greater than what is reserved for each task. 1278 Each task has a fixed array of functions to 1279 trace in the call graph. If the depth of the 1280 calls exceeds that, the function is not traced. 1281 The overrun is the number of functions missed 1282 due to exceeding this array. 1283 1284 funcgraph-cpu 1285 When set, the CPU number of the CPU where the trace 1286 occurred is displayed. 1287 1288 funcgraph-overhead 1289 When set, if the function takes longer than 1290 A certain amount, then a delay marker is 1291 displayed. See "delay" above, under the 1292 header description. 1293 1294 funcgraph-proc 1295 Unlike other tracers, the process' command line 1296 is not displayed by default, but instead only 1297 when a task is traced in and out during a context 1298 switch. Enabling this options has the command 1299 of each process displayed at every line. 1300 1301 funcgraph-duration 1302 At the end of each function (the return) 1303 the duration of the amount of time in the 1304 function is displayed in microseconds. 1305 1306 funcgraph-abstime 1307 When set, the timestamp is displayed at each line. 1308 1309 funcgraph-irqs 1310 When disabled, functions that happen inside an 1311 interrupt will not be traced. 1312 1313 funcgraph-tail 1314 When set, the return event will include the function 1315 that it represents. By default this is off, and 1316 only a closing curly bracket "}" is displayed for 1317 the return of a function. 1318 1319 sleep-time 1320 When running function graph tracer, to include 1321 the time a task schedules out in its function. 1322 When enabled, it will account time the task has been 1323 scheduled out as part of the function call. 1324 1325 graph-time 1326 When running function profiler with function graph tracer, 1327 to include the time to call nested functions. When this is 1328 not set, the time reported for the function will only 1329 include the time the function itself executed for, not the 1330 time for functions that it called. 1331 1332Options for blk tracer: 1333 1334 blk_classic 1335 Shows a more minimalistic output. 1336 1337 1338irqsoff 1339------- 1340 1341When interrupts are disabled, the CPU can not react to any other 1342external event (besides NMIs and SMIs). This prevents the timer 1343interrupt from triggering or the mouse interrupt from letting 1344the kernel know of a new mouse event. The result is a latency 1345with the reaction time. 1346 1347The irqsoff tracer tracks the time for which interrupts are 1348disabled. When a new maximum latency is hit, the tracer saves 1349the trace leading up to that latency point so that every time a 1350new maximum is reached, the old saved trace is discarded and the 1351new trace is saved. 1352 1353To reset the maximum, echo 0 into tracing_max_latency. Here is 1354an example:: 1355 1356 # echo 0 > options/function-trace 1357 # echo irqsoff > current_tracer 1358 # echo 1 > tracing_on 1359 # echo 0 > tracing_max_latency 1360 # ls -ltr 1361 [...] 1362 # echo 0 > tracing_on 1363 # cat trace 1364 # tracer: irqsoff 1365 # 1366 # irqsoff latency trace v1.1.5 on 3.8.0-test+ 1367 # -------------------------------------------------------------------- 1368 # latency: 16 us, #4/4, CPU#0 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1369 # ----------------- 1370 # | task: swapper/0-0 (uid:0 nice:0 policy:0 rt_prio:0) 1371 # ----------------- 1372 # => started at: run_timer_softirq 1373 # => ended at: run_timer_softirq 1374 # 1375 # 1376 # _------=> CPU# 1377 # / _-----=> irqs-off 1378 # | / _----=> need-resched 1379 # || / _---=> hardirq/softirq 1380 # ||| / _--=> preempt-depth 1381 # |||| / delay 1382 # cmd pid ||||| time | caller 1383 # \ / ||||| \ | / 1384 <idle>-0 0d.s2 0us+: _raw_spin_lock_irq <-run_timer_softirq 1385 <idle>-0 0dNs3 17us : _raw_spin_unlock_irq <-run_timer_softirq 1386 <idle>-0 0dNs3 17us+: trace_hardirqs_on <-run_timer_softirq 1387 <idle>-0 0dNs3 25us : <stack trace> 1388 => _raw_spin_unlock_irq 1389 => run_timer_softirq 1390 => __do_softirq 1391 => call_softirq 1392 => do_softirq 1393 => irq_exit 1394 => smp_apic_timer_interrupt 1395 => apic_timer_interrupt 1396 => rcu_idle_exit 1397 => cpu_idle 1398 => rest_init 1399 => start_kernel 1400 => x86_64_start_reservations 1401 => x86_64_start_kernel 1402 1403Here we see that we had a latency of 16 microseconds (which is 1404very good). The _raw_spin_lock_irq in run_timer_softirq disabled 1405interrupts. The difference between the 16 and the displayed 1406timestamp 25us occurred because the clock was incremented 1407between the time of recording the max latency and the time of 1408recording the function that had that latency. 1409 1410Note the above example had function-trace not set. If we set 1411function-trace, we get a much larger output:: 1412 1413 with echo 1 > options/function-trace 1414 1415 # tracer: irqsoff 1416 # 1417 # irqsoff latency trace v1.1.5 on 3.8.0-test+ 1418 # -------------------------------------------------------------------- 1419 # latency: 71 us, #168/168, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1420 # ----------------- 1421 # | task: bash-2042 (uid:0 nice:0 policy:0 rt_prio:0) 1422 # ----------------- 1423 # => started at: ata_scsi_queuecmd 1424 # => ended at: ata_scsi_queuecmd 1425 # 1426 # 1427 # _------=> CPU# 1428 # / _-----=> irqs-off 1429 # | / _----=> need-resched 1430 # || / _---=> hardirq/softirq 1431 # ||| / _--=> preempt-depth 1432 # |||| / delay 1433 # cmd pid ||||| time | caller 1434 # \ / ||||| \ | / 1435 bash-2042 3d... 0us : _raw_spin_lock_irqsave <-ata_scsi_queuecmd 1436 bash-2042 3d... 0us : add_preempt_count <-_raw_spin_lock_irqsave 1437 bash-2042 3d..1 1us : ata_scsi_find_dev <-ata_scsi_queuecmd 1438 bash-2042 3d..1 1us : __ata_scsi_find_dev <-ata_scsi_find_dev 1439 bash-2042 3d..1 2us : ata_find_dev.part.14 <-__ata_scsi_find_dev 1440 bash-2042 3d..1 2us : ata_qc_new_init <-__ata_scsi_queuecmd 1441 bash-2042 3d..1 3us : ata_sg_init <-__ata_scsi_queuecmd 1442 bash-2042 3d..1 4us : ata_scsi_rw_xlat <-__ata_scsi_queuecmd 1443 bash-2042 3d..1 4us : ata_build_rw_tf <-ata_scsi_rw_xlat 1444 [...] 1445 bash-2042 3d..1 67us : delay_tsc <-__delay 1446 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc 1447 bash-2042 3d..2 67us : sub_preempt_count <-delay_tsc 1448 bash-2042 3d..1 67us : add_preempt_count <-delay_tsc 1449 bash-2042 3d..2 68us : sub_preempt_count <-delay_tsc 1450 bash-2042 3d..1 68us+: ata_bmdma_start <-ata_bmdma_qc_issue 1451 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd 1452 bash-2042 3d..1 71us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd 1453 bash-2042 3d..1 72us+: trace_hardirqs_on <-ata_scsi_queuecmd 1454 bash-2042 3d..1 120us : <stack trace> 1455 => _raw_spin_unlock_irqrestore 1456 => ata_scsi_queuecmd 1457 => scsi_dispatch_cmd 1458 => scsi_request_fn 1459 => __blk_run_queue_uncond 1460 => __blk_run_queue 1461 => blk_queue_bio 1462 => submit_bio_noacct 1463 => submit_bio 1464 => submit_bh 1465 => __ext3_get_inode_loc 1466 => ext3_iget 1467 => ext3_lookup 1468 => lookup_real 1469 => __lookup_hash 1470 => walk_component 1471 => lookup_last 1472 => path_lookupat 1473 => filename_lookup 1474 => user_path_at_empty 1475 => user_path_at 1476 => vfs_fstatat 1477 => vfs_stat 1478 => sys_newstat 1479 => system_call_fastpath 1480 1481 1482Here we traced a 71 microsecond latency. But we also see all the 1483functions that were called during that time. Note that by 1484enabling function tracing, we incur an added overhead. This 1485overhead may extend the latency times. But nevertheless, this 1486trace has provided some very helpful debugging information. 1487 1488If we prefer function graph output instead of function, we can set 1489display-graph option:: 1490 1491 with echo 1 > options/display-graph 1492 1493 # tracer: irqsoff 1494 # 1495 # irqsoff latency trace v1.1.5 on 4.20.0-rc6+ 1496 # -------------------------------------------------------------------- 1497 # latency: 3751 us, #274/274, CPU#0 | (M:desktop VP:0, KP:0, SP:0 HP:0 #P:4) 1498 # ----------------- 1499 # | task: bash-1507 (uid:0 nice:0 policy:0 rt_prio:0) 1500 # ----------------- 1501 # => started at: free_debug_processing 1502 # => ended at: return_to_handler 1503 # 1504 # 1505 # _-----=> irqs-off 1506 # / _----=> need-resched 1507 # | / _---=> hardirq/softirq 1508 # || / _--=> preempt-depth 1509 # ||| / 1510 # REL TIME CPU TASK/PID |||| DURATION FUNCTION CALLS 1511 # | | | | |||| | | | | | | 1512 0 us | 0) bash-1507 | d... | 0.000 us | _raw_spin_lock_irqsave(); 1513 0 us | 0) bash-1507 | d..1 | 0.378 us | do_raw_spin_trylock(); 1514 1 us | 0) bash-1507 | d..2 | | set_track() { 1515 2 us | 0) bash-1507 | d..2 | | save_stack_trace() { 1516 2 us | 0) bash-1507 | d..2 | | __save_stack_trace() { 1517 3 us | 0) bash-1507 | d..2 | | __unwind_start() { 1518 3 us | 0) bash-1507 | d..2 | | get_stack_info() { 1519 3 us | 0) bash-1507 | d..2 | 0.351 us | in_task_stack(); 1520 4 us | 0) bash-1507 | d..2 | 1.107 us | } 1521 [...] 1522 3750 us | 0) bash-1507 | d..1 | 0.516 us | do_raw_spin_unlock(); 1523 3750 us | 0) bash-1507 | d..1 | 0.000 us | _raw_spin_unlock_irqrestore(); 1524 3764 us | 0) bash-1507 | d..1 | 0.000 us | tracer_hardirqs_on(); 1525 bash-1507 0d..1 3792us : <stack trace> 1526 => free_debug_processing 1527 => __slab_free 1528 => kmem_cache_free 1529 => vm_area_free 1530 => remove_vma 1531 => exit_mmap 1532 => mmput 1533 => begin_new_exec 1534 => load_elf_binary 1535 => search_binary_handler 1536 => __do_execve_file.isra.32 1537 => __x64_sys_execve 1538 => do_syscall_64 1539 => entry_SYSCALL_64_after_hwframe 1540 1541preemptoff 1542---------- 1543 1544When preemption is disabled, we may be able to receive 1545interrupts but the task cannot be preempted and a higher 1546priority task must wait for preemption to be enabled again 1547before it can preempt a lower priority task. 1548 1549The preemptoff tracer traces the places that disable preemption. 1550Like the irqsoff tracer, it records the maximum latency for 1551which preemption was disabled. The control of preemptoff tracer 1552is much like the irqsoff tracer. 1553:: 1554 1555 # echo 0 > options/function-trace 1556 # echo preemptoff > current_tracer 1557 # echo 1 > tracing_on 1558 # echo 0 > tracing_max_latency 1559 # ls -ltr 1560 [...] 1561 # echo 0 > tracing_on 1562 # cat trace 1563 # tracer: preemptoff 1564 # 1565 # preemptoff latency trace v1.1.5 on 3.8.0-test+ 1566 # -------------------------------------------------------------------- 1567 # latency: 46 us, #4/4, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1568 # ----------------- 1569 # | task: sshd-1991 (uid:0 nice:0 policy:0 rt_prio:0) 1570 # ----------------- 1571 # => started at: do_IRQ 1572 # => ended at: do_IRQ 1573 # 1574 # 1575 # _------=> CPU# 1576 # / _-----=> irqs-off 1577 # | / _----=> need-resched 1578 # || / _---=> hardirq/softirq 1579 # ||| / _--=> preempt-depth 1580 # |||| / delay 1581 # cmd pid ||||| time | caller 1582 # \ / ||||| \ | / 1583 sshd-1991 1d.h. 0us+: irq_enter <-do_IRQ 1584 sshd-1991 1d..1 46us : irq_exit <-do_IRQ 1585 sshd-1991 1d..1 47us+: trace_preempt_on <-do_IRQ 1586 sshd-1991 1d..1 52us : <stack trace> 1587 => sub_preempt_count 1588 => irq_exit 1589 => do_IRQ 1590 => ret_from_intr 1591 1592 1593This has some more changes. Preemption was disabled when an 1594interrupt came in (notice the 'h'), and was enabled on exit. 1595But we also see that interrupts have been disabled when entering 1596the preempt off section and leaving it (the 'd'). We do not know if 1597interrupts were enabled in the mean time or shortly after this 1598was over. 1599:: 1600 1601 # tracer: preemptoff 1602 # 1603 # preemptoff latency trace v1.1.5 on 3.8.0-test+ 1604 # -------------------------------------------------------------------- 1605 # latency: 83 us, #241/241, CPU#1 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1606 # ----------------- 1607 # | task: bash-1994 (uid:0 nice:0 policy:0 rt_prio:0) 1608 # ----------------- 1609 # => started at: wake_up_new_task 1610 # => ended at: task_rq_unlock 1611 # 1612 # 1613 # _------=> CPU# 1614 # / _-----=> irqs-off 1615 # | / _----=> need-resched 1616 # || / _---=> hardirq/softirq 1617 # ||| / _--=> preempt-depth 1618 # |||| / delay 1619 # cmd pid ||||| time | caller 1620 # \ / ||||| \ | / 1621 bash-1994 1d..1 0us : _raw_spin_lock_irqsave <-wake_up_new_task 1622 bash-1994 1d..1 0us : select_task_rq_fair <-select_task_rq 1623 bash-1994 1d..1 1us : __rcu_read_lock <-select_task_rq_fair 1624 bash-1994 1d..1 1us : source_load <-select_task_rq_fair 1625 bash-1994 1d..1 1us : source_load <-select_task_rq_fair 1626 [...] 1627 bash-1994 1d..1 12us : irq_enter <-smp_apic_timer_interrupt 1628 bash-1994 1d..1 12us : rcu_irq_enter <-irq_enter 1629 bash-1994 1d..1 13us : add_preempt_count <-irq_enter 1630 bash-1994 1d.h1 13us : exit_idle <-smp_apic_timer_interrupt 1631 bash-1994 1d.h1 13us : hrtimer_interrupt <-smp_apic_timer_interrupt 1632 bash-1994 1d.h1 13us : _raw_spin_lock <-hrtimer_interrupt 1633 bash-1994 1d.h1 14us : add_preempt_count <-_raw_spin_lock 1634 bash-1994 1d.h2 14us : ktime_get_update_offsets <-hrtimer_interrupt 1635 [...] 1636 bash-1994 1d.h1 35us : lapic_next_event <-clockevents_program_event 1637 bash-1994 1d.h1 35us : irq_exit <-smp_apic_timer_interrupt 1638 bash-1994 1d.h1 36us : sub_preempt_count <-irq_exit 1639 bash-1994 1d..2 36us : do_softirq <-irq_exit 1640 bash-1994 1d..2 36us : __do_softirq <-call_softirq 1641 bash-1994 1d..2 36us : __local_bh_disable <-__do_softirq 1642 bash-1994 1d.s2 37us : add_preempt_count <-_raw_spin_lock_irq 1643 bash-1994 1d.s3 38us : _raw_spin_unlock <-run_timer_softirq 1644 bash-1994 1d.s3 39us : sub_preempt_count <-_raw_spin_unlock 1645 bash-1994 1d.s2 39us : call_timer_fn <-run_timer_softirq 1646 [...] 1647 bash-1994 1dNs2 81us : cpu_needs_another_gp <-rcu_process_callbacks 1648 bash-1994 1dNs2 82us : __local_bh_enable <-__do_softirq 1649 bash-1994 1dNs2 82us : sub_preempt_count <-__local_bh_enable 1650 bash-1994 1dN.2 82us : idle_cpu <-irq_exit 1651 bash-1994 1dN.2 83us : rcu_irq_exit <-irq_exit 1652 bash-1994 1dN.2 83us : sub_preempt_count <-irq_exit 1653 bash-1994 1.N.1 84us : _raw_spin_unlock_irqrestore <-task_rq_unlock 1654 bash-1994 1.N.1 84us+: trace_preempt_on <-task_rq_unlock 1655 bash-1994 1.N.1 104us : <stack trace> 1656 => sub_preempt_count 1657 => _raw_spin_unlock_irqrestore 1658 => task_rq_unlock 1659 => wake_up_new_task 1660 => do_fork 1661 => sys_clone 1662 => stub_clone 1663 1664 1665The above is an example of the preemptoff trace with 1666function-trace set. Here we see that interrupts were not disabled 1667the entire time. The irq_enter code lets us know that we entered 1668an interrupt 'h'. Before that, the functions being traced still 1669show that it is not in an interrupt, but we can see from the 1670functions themselves that this is not the case. 1671 1672preemptirqsoff 1673-------------- 1674 1675Knowing the locations that have interrupts disabled or 1676preemption disabled for the longest times is helpful. But 1677sometimes we would like to know when either preemption and/or 1678interrupts are disabled. 1679 1680Consider the following code:: 1681 1682 local_irq_disable(); 1683 call_function_with_irqs_off(); 1684 preempt_disable(); 1685 call_function_with_irqs_and_preemption_off(); 1686 local_irq_enable(); 1687 call_function_with_preemption_off(); 1688 preempt_enable(); 1689 1690The irqsoff tracer will record the total length of 1691call_function_with_irqs_off() and 1692call_function_with_irqs_and_preemption_off(). 1693 1694The preemptoff tracer will record the total length of 1695call_function_with_irqs_and_preemption_off() and 1696call_function_with_preemption_off(). 1697 1698But neither will trace the time that interrupts and/or 1699preemption is disabled. This total time is the time that we can 1700not schedule. To record this time, use the preemptirqsoff 1701tracer. 1702 1703Again, using this trace is much like the irqsoff and preemptoff 1704tracers. 1705:: 1706 1707 # echo 0 > options/function-trace 1708 # echo preemptirqsoff > current_tracer 1709 # echo 1 > tracing_on 1710 # echo 0 > tracing_max_latency 1711 # ls -ltr 1712 [...] 1713 # echo 0 > tracing_on 1714 # cat trace 1715 # tracer: preemptirqsoff 1716 # 1717 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+ 1718 # -------------------------------------------------------------------- 1719 # latency: 100 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1720 # ----------------- 1721 # | task: ls-2230 (uid:0 nice:0 policy:0 rt_prio:0) 1722 # ----------------- 1723 # => started at: ata_scsi_queuecmd 1724 # => ended at: ata_scsi_queuecmd 1725 # 1726 # 1727 # _------=> CPU# 1728 # / _-----=> irqs-off 1729 # | / _----=> need-resched 1730 # || / _---=> hardirq/softirq 1731 # ||| / _--=> preempt-depth 1732 # |||| / delay 1733 # cmd pid ||||| time | caller 1734 # \ / ||||| \ | / 1735 ls-2230 3d... 0us+: _raw_spin_lock_irqsave <-ata_scsi_queuecmd 1736 ls-2230 3...1 100us : _raw_spin_unlock_irqrestore <-ata_scsi_queuecmd 1737 ls-2230 3...1 101us+: trace_preempt_on <-ata_scsi_queuecmd 1738 ls-2230 3...1 111us : <stack trace> 1739 => sub_preempt_count 1740 => _raw_spin_unlock_irqrestore 1741 => ata_scsi_queuecmd 1742 => scsi_dispatch_cmd 1743 => scsi_request_fn 1744 => __blk_run_queue_uncond 1745 => __blk_run_queue 1746 => blk_queue_bio 1747 => submit_bio_noacct 1748 => submit_bio 1749 => submit_bh 1750 => ext3_bread 1751 => ext3_dir_bread 1752 => htree_dirblock_to_tree 1753 => ext3_htree_fill_tree 1754 => ext3_readdir 1755 => vfs_readdir 1756 => sys_getdents 1757 => system_call_fastpath 1758 1759 1760The trace_hardirqs_off_thunk is called from assembly on x86 when 1761interrupts are disabled in the assembly code. Without the 1762function tracing, we do not know if interrupts were enabled 1763within the preemption points. We do see that it started with 1764preemption enabled. 1765 1766Here is a trace with function-trace set:: 1767 1768 # tracer: preemptirqsoff 1769 # 1770 # preemptirqsoff latency trace v1.1.5 on 3.8.0-test+ 1771 # -------------------------------------------------------------------- 1772 # latency: 161 us, #339/339, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1773 # ----------------- 1774 # | task: ls-2269 (uid:0 nice:0 policy:0 rt_prio:0) 1775 # ----------------- 1776 # => started at: schedule 1777 # => ended at: mutex_unlock 1778 # 1779 # 1780 # _------=> CPU# 1781 # / _-----=> irqs-off 1782 # | / _----=> need-resched 1783 # || / _---=> hardirq/softirq 1784 # ||| / _--=> preempt-depth 1785 # |||| / delay 1786 # cmd pid ||||| time | caller 1787 # \ / ||||| \ | / 1788 kworker/-59 3...1 0us : __schedule <-schedule 1789 kworker/-59 3d..1 0us : rcu_preempt_qs <-rcu_note_context_switch 1790 kworker/-59 3d..1 1us : add_preempt_count <-_raw_spin_lock_irq 1791 kworker/-59 3d..2 1us : deactivate_task <-__schedule 1792 kworker/-59 3d..2 1us : dequeue_task <-deactivate_task 1793 kworker/-59 3d..2 2us : update_rq_clock <-dequeue_task 1794 kworker/-59 3d..2 2us : dequeue_task_fair <-dequeue_task 1795 kworker/-59 3d..2 2us : update_curr <-dequeue_task_fair 1796 kworker/-59 3d..2 2us : update_min_vruntime <-update_curr 1797 kworker/-59 3d..2 3us : cpuacct_charge <-update_curr 1798 kworker/-59 3d..2 3us : __rcu_read_lock <-cpuacct_charge 1799 kworker/-59 3d..2 3us : __rcu_read_unlock <-cpuacct_charge 1800 kworker/-59 3d..2 3us : update_cfs_rq_blocked_load <-dequeue_task_fair 1801 kworker/-59 3d..2 4us : clear_buddies <-dequeue_task_fair 1802 kworker/-59 3d..2 4us : account_entity_dequeue <-dequeue_task_fair 1803 kworker/-59 3d..2 4us : update_min_vruntime <-dequeue_task_fair 1804 kworker/-59 3d..2 4us : update_cfs_shares <-dequeue_task_fair 1805 kworker/-59 3d..2 5us : hrtick_update <-dequeue_task_fair 1806 kworker/-59 3d..2 5us : wq_worker_sleeping <-__schedule 1807 kworker/-59 3d..2 5us : kthread_data <-wq_worker_sleeping 1808 kworker/-59 3d..2 5us : put_prev_task_fair <-__schedule 1809 kworker/-59 3d..2 6us : pick_next_task_fair <-pick_next_task 1810 kworker/-59 3d..2 6us : clear_buddies <-pick_next_task_fair 1811 kworker/-59 3d..2 6us : set_next_entity <-pick_next_task_fair 1812 kworker/-59 3d..2 6us : update_stats_wait_end <-set_next_entity 1813 ls-2269 3d..2 7us : finish_task_switch <-__schedule 1814 ls-2269 3d..2 7us : _raw_spin_unlock_irq <-finish_task_switch 1815 ls-2269 3d..2 8us : do_IRQ <-ret_from_intr 1816 ls-2269 3d..2 8us : irq_enter <-do_IRQ 1817 ls-2269 3d..2 8us : rcu_irq_enter <-irq_enter 1818 ls-2269 3d..2 9us : add_preempt_count <-irq_enter 1819 ls-2269 3d.h2 9us : exit_idle <-do_IRQ 1820 [...] 1821 ls-2269 3d.h3 20us : sub_preempt_count <-_raw_spin_unlock 1822 ls-2269 3d.h2 20us : irq_exit <-do_IRQ 1823 ls-2269 3d.h2 21us : sub_preempt_count <-irq_exit 1824 ls-2269 3d..3 21us : do_softirq <-irq_exit 1825 ls-2269 3d..3 21us : __do_softirq <-call_softirq 1826 ls-2269 3d..3 21us+: __local_bh_disable <-__do_softirq 1827 ls-2269 3d.s4 29us : sub_preempt_count <-_local_bh_enable_ip 1828 ls-2269 3d.s5 29us : sub_preempt_count <-_local_bh_enable_ip 1829 ls-2269 3d.s5 31us : do_IRQ <-ret_from_intr 1830 ls-2269 3d.s5 31us : irq_enter <-do_IRQ 1831 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter 1832 [...] 1833 ls-2269 3d.s5 31us : rcu_irq_enter <-irq_enter 1834 ls-2269 3d.s5 32us : add_preempt_count <-irq_enter 1835 ls-2269 3d.H5 32us : exit_idle <-do_IRQ 1836 ls-2269 3d.H5 32us : handle_irq <-do_IRQ 1837 ls-2269 3d.H5 32us : irq_to_desc <-handle_irq 1838 ls-2269 3d.H5 33us : handle_fasteoi_irq <-handle_irq 1839 [...] 1840 ls-2269 3d.s5 158us : _raw_spin_unlock_irqrestore <-rtl8139_poll 1841 ls-2269 3d.s3 158us : net_rps_action_and_irq_enable.isra.65 <-net_rx_action 1842 ls-2269 3d.s3 159us : __local_bh_enable <-__do_softirq 1843 ls-2269 3d.s3 159us : sub_preempt_count <-__local_bh_enable 1844 ls-2269 3d..3 159us : idle_cpu <-irq_exit 1845 ls-2269 3d..3 159us : rcu_irq_exit <-irq_exit 1846 ls-2269 3d..3 160us : sub_preempt_count <-irq_exit 1847 ls-2269 3d... 161us : __mutex_unlock_slowpath <-mutex_unlock 1848 ls-2269 3d... 162us+: trace_hardirqs_on <-mutex_unlock 1849 ls-2269 3d... 186us : <stack trace> 1850 => __mutex_unlock_slowpath 1851 => mutex_unlock 1852 => process_output 1853 => n_tty_write 1854 => tty_write 1855 => vfs_write 1856 => sys_write 1857 => system_call_fastpath 1858 1859This is an interesting trace. It started with kworker running and 1860scheduling out and ls taking over. But as soon as ls released the 1861rq lock and enabled interrupts (but not preemption) an interrupt 1862triggered. When the interrupt finished, it started running softirqs. 1863But while the softirq was running, another interrupt triggered. 1864When an interrupt is running inside a softirq, the annotation is 'H'. 1865 1866 1867wakeup 1868------ 1869 1870One common case that people are interested in tracing is the 1871time it takes for a task that is woken to actually wake up. 1872Now for non Real-Time tasks, this can be arbitrary. But tracing 1873it none the less can be interesting. 1874 1875Without function tracing:: 1876 1877 # echo 0 > options/function-trace 1878 # echo wakeup > current_tracer 1879 # echo 1 > tracing_on 1880 # echo 0 > tracing_max_latency 1881 # chrt -f 5 sleep 1 1882 # echo 0 > tracing_on 1883 # cat trace 1884 # tracer: wakeup 1885 # 1886 # wakeup latency trace v1.1.5 on 3.8.0-test+ 1887 # -------------------------------------------------------------------- 1888 # latency: 15 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1889 # ----------------- 1890 # | task: kworker/3:1H-312 (uid:0 nice:-20 policy:0 rt_prio:0) 1891 # ----------------- 1892 # 1893 # _------=> CPU# 1894 # / _-----=> irqs-off 1895 # | / _----=> need-resched 1896 # || / _---=> hardirq/softirq 1897 # ||| / _--=> preempt-depth 1898 # |||| / delay 1899 # cmd pid ||||| time | caller 1900 # \ / ||||| \ | / 1901 <idle>-0 3dNs7 0us : 0:120:R + [003] 312:100:R kworker/3:1H 1902 <idle>-0 3dNs7 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up 1903 <idle>-0 3d..3 15us : __schedule <-schedule 1904 <idle>-0 3d..3 15us : 0:120:R ==> [003] 312:100:R kworker/3:1H 1905 1906The tracer only traces the highest priority task in the system 1907to avoid tracing the normal circumstances. Here we see that 1908the kworker with a nice priority of -20 (not very nice), took 1909just 15 microseconds from the time it woke up, to the time it 1910ran. 1911 1912Non Real-Time tasks are not that interesting. A more interesting 1913trace is to concentrate only on Real-Time tasks. 1914 1915wakeup_rt 1916--------- 1917 1918In a Real-Time environment it is very important to know the 1919wakeup time it takes for the highest priority task that is woken 1920up to the time that it executes. This is also known as "schedule 1921latency". I stress the point that this is about RT tasks. It is 1922also important to know the scheduling latency of non-RT tasks, 1923but the average schedule latency is better for non-RT tasks. 1924Tools like LatencyTop are more appropriate for such 1925measurements. 1926 1927Real-Time environments are interested in the worst case latency. 1928That is the longest latency it takes for something to happen, 1929and not the average. We can have a very fast scheduler that may 1930only have a large latency once in a while, but that would not 1931work well with Real-Time tasks. The wakeup_rt tracer was designed 1932to record the worst case wakeups of RT tasks. Non-RT tasks are 1933not recorded because the tracer only records one worst case and 1934tracing non-RT tasks that are unpredictable will overwrite the 1935worst case latency of RT tasks (just run the normal wakeup 1936tracer for a while to see that effect). 1937 1938Since this tracer only deals with RT tasks, we will run this 1939slightly differently than we did with the previous tracers. 1940Instead of performing an 'ls', we will run 'sleep 1' under 1941'chrt' which changes the priority of the task. 1942:: 1943 1944 # echo 0 > options/function-trace 1945 # echo wakeup_rt > current_tracer 1946 # echo 1 > tracing_on 1947 # echo 0 > tracing_max_latency 1948 # chrt -f 5 sleep 1 1949 # echo 0 > tracing_on 1950 # cat trace 1951 # tracer: wakeup 1952 # 1953 # tracer: wakeup_rt 1954 # 1955 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ 1956 # -------------------------------------------------------------------- 1957 # latency: 5 us, #4/4, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 1958 # ----------------- 1959 # | task: sleep-2389 (uid:0 nice:0 policy:1 rt_prio:5) 1960 # ----------------- 1961 # 1962 # _------=> CPU# 1963 # / _-----=> irqs-off 1964 # | / _----=> need-resched 1965 # || / _---=> hardirq/softirq 1966 # ||| / _--=> preempt-depth 1967 # |||| / delay 1968 # cmd pid ||||| time | caller 1969 # \ / ||||| \ | / 1970 <idle>-0 3d.h4 0us : 0:120:R + [003] 2389: 94:R sleep 1971 <idle>-0 3d.h4 1us+: ttwu_do_activate.constprop.87 <-try_to_wake_up 1972 <idle>-0 3d..3 5us : __schedule <-schedule 1973 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep 1974 1975 1976Running this on an idle system, we see that it only took 5 microseconds 1977to perform the task switch. Note, since the trace point in the schedule 1978is before the actual "switch", we stop the tracing when the recorded task 1979is about to schedule in. This may change if we add a new marker at the 1980end of the scheduler. 1981 1982Notice that the recorded task is 'sleep' with the PID of 2389 1983and it has an rt_prio of 5. This priority is user-space priority 1984and not the internal kernel priority. The policy is 1 for 1985SCHED_FIFO and 2 for SCHED_RR. 1986 1987Note, that the trace data shows the internal priority (99 - rtprio). 1988:: 1989 1990 <idle>-0 3d..3 5us : 0:120:R ==> [003] 2389: 94:R sleep 1991 1992The 0:120:R means idle was running with a nice priority of 0 (120 - 120) 1993and in the running state 'R'. The sleep task was scheduled in with 19942389: 94:R. That is the priority is the kernel rtprio (99 - 5 = 94) 1995and it too is in the running state. 1996 1997Doing the same with chrt -r 5 and function-trace set. 1998:: 1999 2000 echo 1 > options/function-trace 2001 2002 # tracer: wakeup_rt 2003 # 2004 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ 2005 # -------------------------------------------------------------------- 2006 # latency: 29 us, #85/85, CPU#3 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 2007 # ----------------- 2008 # | task: sleep-2448 (uid:0 nice:0 policy:1 rt_prio:5) 2009 # ----------------- 2010 # 2011 # _------=> CPU# 2012 # / _-----=> irqs-off 2013 # | / _----=> need-resched 2014 # || / _---=> hardirq/softirq 2015 # ||| / _--=> preempt-depth 2016 # |||| / delay 2017 # cmd pid ||||| time | caller 2018 # \ / ||||| \ | / 2019 <idle>-0 3d.h4 1us+: 0:120:R + [003] 2448: 94:R sleep 2020 <idle>-0 3d.h4 2us : ttwu_do_activate.constprop.87 <-try_to_wake_up 2021 <idle>-0 3d.h3 3us : check_preempt_curr <-ttwu_do_wakeup 2022 <idle>-0 3d.h3 3us : resched_curr <-check_preempt_curr 2023 <idle>-0 3dNh3 4us : task_woken_rt <-ttwu_do_wakeup 2024 <idle>-0 3dNh3 4us : _raw_spin_unlock <-try_to_wake_up 2025 <idle>-0 3dNh3 4us : sub_preempt_count <-_raw_spin_unlock 2026 <idle>-0 3dNh2 5us : ttwu_stat <-try_to_wake_up 2027 <idle>-0 3dNh2 5us : _raw_spin_unlock_irqrestore <-try_to_wake_up 2028 <idle>-0 3dNh2 6us : sub_preempt_count <-_raw_spin_unlock_irqrestore 2029 <idle>-0 3dNh1 6us : _raw_spin_lock <-__run_hrtimer 2030 <idle>-0 3dNh1 6us : add_preempt_count <-_raw_spin_lock 2031 <idle>-0 3dNh2 7us : _raw_spin_unlock <-hrtimer_interrupt 2032 <idle>-0 3dNh2 7us : sub_preempt_count <-_raw_spin_unlock 2033 <idle>-0 3dNh1 7us : tick_program_event <-hrtimer_interrupt 2034 <idle>-0 3dNh1 7us : clockevents_program_event <-tick_program_event 2035 <idle>-0 3dNh1 8us : ktime_get <-clockevents_program_event 2036 <idle>-0 3dNh1 8us : lapic_next_event <-clockevents_program_event 2037 <idle>-0 3dNh1 8us : irq_exit <-smp_apic_timer_interrupt 2038 <idle>-0 3dNh1 9us : sub_preempt_count <-irq_exit 2039 <idle>-0 3dN.2 9us : idle_cpu <-irq_exit 2040 <idle>-0 3dN.2 9us : rcu_irq_exit <-irq_exit 2041 <idle>-0 3dN.2 10us : rcu_eqs_enter_common.isra.45 <-rcu_irq_exit 2042 <idle>-0 3dN.2 10us : sub_preempt_count <-irq_exit 2043 <idle>-0 3.N.1 11us : rcu_idle_exit <-cpu_idle 2044 <idle>-0 3dN.1 11us : rcu_eqs_exit_common.isra.43 <-rcu_idle_exit 2045 <idle>-0 3.N.1 11us : tick_nohz_idle_exit <-cpu_idle 2046 <idle>-0 3dN.1 12us : menu_hrtimer_cancel <-tick_nohz_idle_exit 2047 <idle>-0 3dN.1 12us : ktime_get <-tick_nohz_idle_exit 2048 <idle>-0 3dN.1 12us : tick_do_update_jiffies64 <-tick_nohz_idle_exit 2049 <idle>-0 3dN.1 13us : cpu_load_update_nohz <-tick_nohz_idle_exit 2050 <idle>-0 3dN.1 13us : _raw_spin_lock <-cpu_load_update_nohz 2051 <idle>-0 3dN.1 13us : add_preempt_count <-_raw_spin_lock 2052 <idle>-0 3dN.2 13us : __cpu_load_update <-cpu_load_update_nohz 2053 <idle>-0 3dN.2 14us : sched_avg_update <-__cpu_load_update 2054 <idle>-0 3dN.2 14us : _raw_spin_unlock <-cpu_load_update_nohz 2055 <idle>-0 3dN.2 14us : sub_preempt_count <-_raw_spin_unlock 2056 <idle>-0 3dN.1 15us : calc_load_nohz_stop <-tick_nohz_idle_exit 2057 <idle>-0 3dN.1 15us : touch_softlockup_watchdog <-tick_nohz_idle_exit 2058 <idle>-0 3dN.1 15us : hrtimer_cancel <-tick_nohz_idle_exit 2059 <idle>-0 3dN.1 15us : hrtimer_try_to_cancel <-hrtimer_cancel 2060 <idle>-0 3dN.1 16us : lock_hrtimer_base.isra.18 <-hrtimer_try_to_cancel 2061 <idle>-0 3dN.1 16us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18 2062 <idle>-0 3dN.1 16us : add_preempt_count <-_raw_spin_lock_irqsave 2063 <idle>-0 3dN.2 17us : __remove_hrtimer <-remove_hrtimer.part.16 2064 <idle>-0 3dN.2 17us : hrtimer_force_reprogram <-__remove_hrtimer 2065 <idle>-0 3dN.2 17us : tick_program_event <-hrtimer_force_reprogram 2066 <idle>-0 3dN.2 18us : clockevents_program_event <-tick_program_event 2067 <idle>-0 3dN.2 18us : ktime_get <-clockevents_program_event 2068 <idle>-0 3dN.2 18us : lapic_next_event <-clockevents_program_event 2069 <idle>-0 3dN.2 19us : _raw_spin_unlock_irqrestore <-hrtimer_try_to_cancel 2070 <idle>-0 3dN.2 19us : sub_preempt_count <-_raw_spin_unlock_irqrestore 2071 <idle>-0 3dN.1 19us : hrtimer_forward <-tick_nohz_idle_exit 2072 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward 2073 <idle>-0 3dN.1 20us : ktime_add_safe <-hrtimer_forward 2074 <idle>-0 3dN.1 20us : hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11 2075 <idle>-0 3dN.1 20us : __hrtimer_start_range_ns <-hrtimer_start_range_ns 2076 <idle>-0 3dN.1 21us : lock_hrtimer_base.isra.18 <-__hrtimer_start_range_ns 2077 <idle>-0 3dN.1 21us : _raw_spin_lock_irqsave <-lock_hrtimer_base.isra.18 2078 <idle>-0 3dN.1 21us : add_preempt_count <-_raw_spin_lock_irqsave 2079 <idle>-0 3dN.2 22us : ktime_add_safe <-__hrtimer_start_range_ns 2080 <idle>-0 3dN.2 22us : enqueue_hrtimer <-__hrtimer_start_range_ns 2081 <idle>-0 3dN.2 22us : tick_program_event <-__hrtimer_start_range_ns 2082 <idle>-0 3dN.2 23us : clockevents_program_event <-tick_program_event 2083 <idle>-0 3dN.2 23us : ktime_get <-clockevents_program_event 2084 <idle>-0 3dN.2 23us : lapic_next_event <-clockevents_program_event 2085 <idle>-0 3dN.2 24us : _raw_spin_unlock_irqrestore <-__hrtimer_start_range_ns 2086 <idle>-0 3dN.2 24us : sub_preempt_count <-_raw_spin_unlock_irqrestore 2087 <idle>-0 3dN.1 24us : account_idle_ticks <-tick_nohz_idle_exit 2088 <idle>-0 3dN.1 24us : account_idle_time <-account_idle_ticks 2089 <idle>-0 3.N.1 25us : sub_preempt_count <-cpu_idle 2090 <idle>-0 3.N.. 25us : schedule <-cpu_idle 2091 <idle>-0 3.N.. 25us : __schedule <-preempt_schedule 2092 <idle>-0 3.N.. 26us : add_preempt_count <-__schedule 2093 <idle>-0 3.N.1 26us : rcu_note_context_switch <-__schedule 2094 <idle>-0 3.N.1 26us : rcu_sched_qs <-rcu_note_context_switch 2095 <idle>-0 3dN.1 27us : rcu_preempt_qs <-rcu_note_context_switch 2096 <idle>-0 3.N.1 27us : _raw_spin_lock_irq <-__schedule 2097 <idle>-0 3dN.1 27us : add_preempt_count <-_raw_spin_lock_irq 2098 <idle>-0 3dN.2 28us : put_prev_task_idle <-__schedule 2099 <idle>-0 3dN.2 28us : pick_next_task_stop <-pick_next_task 2100 <idle>-0 3dN.2 28us : pick_next_task_rt <-pick_next_task 2101 <idle>-0 3dN.2 29us : dequeue_pushable_task <-pick_next_task_rt 2102 <idle>-0 3d..3 29us : __schedule <-preempt_schedule 2103 <idle>-0 3d..3 30us : 0:120:R ==> [003] 2448: 94:R sleep 2104 2105This isn't that big of a trace, even with function tracing enabled, 2106so I included the entire trace. 2107 2108The interrupt went off while when the system was idle. Somewhere 2109before task_woken_rt() was called, the NEED_RESCHED flag was set, 2110this is indicated by the first occurrence of the 'N' flag. 2111 2112Latency tracing and events 2113-------------------------- 2114As function tracing can induce a much larger latency, but without 2115seeing what happens within the latency it is hard to know what 2116caused it. There is a middle ground, and that is with enabling 2117events. 2118:: 2119 2120 # echo 0 > options/function-trace 2121 # echo wakeup_rt > current_tracer 2122 # echo 1 > events/enable 2123 # echo 1 > tracing_on 2124 # echo 0 > tracing_max_latency 2125 # chrt -f 5 sleep 1 2126 # echo 0 > tracing_on 2127 # cat trace 2128 # tracer: wakeup_rt 2129 # 2130 # wakeup_rt latency trace v1.1.5 on 3.8.0-test+ 2131 # -------------------------------------------------------------------- 2132 # latency: 6 us, #12/12, CPU#2 | (M:preempt VP:0, KP:0, SP:0 HP:0 #P:4) 2133 # ----------------- 2134 # | task: sleep-5882 (uid:0 nice:0 policy:1 rt_prio:5) 2135 # ----------------- 2136 # 2137 # _------=> CPU# 2138 # / _-----=> irqs-off 2139 # | / _----=> need-resched 2140 # || / _---=> hardirq/softirq 2141 # ||| / _--=> preempt-depth 2142 # |||| / delay 2143 # cmd pid ||||| time | caller 2144 # \ / ||||| \ | / 2145 <idle>-0 2d.h4 0us : 0:120:R + [002] 5882: 94:R sleep 2146 <idle>-0 2d.h4 0us : ttwu_do_activate.constprop.87 <-try_to_wake_up 2147 <idle>-0 2d.h4 1us : sched_wakeup: comm=sleep pid=5882 prio=94 success=1 target_cpu=002 2148 <idle>-0 2dNh2 1us : hrtimer_expire_exit: hrtimer=ffff88007796feb8 2149 <idle>-0 2.N.2 2us : power_end: cpu_id=2 2150 <idle>-0 2.N.2 3us : cpu_idle: state=4294967295 cpu_id=2 2151 <idle>-0 2dN.3 4us : hrtimer_cancel: hrtimer=ffff88007d50d5e0 2152 <idle>-0 2dN.3 4us : hrtimer_start: hrtimer=ffff88007d50d5e0 function=tick_sched_timer expires=34311211000000 softexpires=34311211000000 2153 <idle>-0 2.N.2 5us : rcu_utilization: Start context switch 2154 <idle>-0 2.N.2 5us : rcu_utilization: End context switch 2155 <idle>-0 2d..3 6us : __schedule <-schedule 2156 <idle>-0 2d..3 6us : 0:120:R ==> [002] 5882: 94:R sleep 2157 2158 2159Hardware Latency Detector 2160------------------------- 2161 2162The hardware latency detector is executed by enabling the "hwlat" tracer. 2163 2164NOTE, this tracer will affect the performance of the system as it will 2165periodically make a CPU constantly busy with interrupts disabled. 2166:: 2167 2168 # echo hwlat > current_tracer 2169 # sleep 100 2170 # cat trace 2171 # tracer: hwlat 2172 # 2173 # entries-in-buffer/entries-written: 13/13 #P:8 2174 # 2175 # _-----=> irqs-off 2176 # / _----=> need-resched 2177 # | / _---=> hardirq/softirq 2178 # || / _--=> preempt-depth 2179 # ||| / delay 2180 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 2181 # | | | |||| | | 2182 <...>-1729 [001] d... 678.473449: #1 inner/outer(us): 11/12 ts:1581527483.343962693 count:6 2183 <...>-1729 [004] d... 689.556542: #2 inner/outer(us): 16/9 ts:1581527494.889008092 count:1 2184 <...>-1729 [005] d... 714.756290: #3 inner/outer(us): 16/16 ts:1581527519.678961629 count:5 2185 <...>-1729 [001] d... 718.788247: #4 inner/outer(us): 9/17 ts:1581527523.889012713 count:1 2186 <...>-1729 [002] d... 719.796341: #5 inner/outer(us): 13/9 ts:1581527524.912872606 count:1 2187 <...>-1729 [006] d... 844.787091: #6 inner/outer(us): 9/12 ts:1581527649.889048502 count:2 2188 <...>-1729 [003] d... 849.827033: #7 inner/outer(us): 18/9 ts:1581527654.889013793 count:1 2189 <...>-1729 [007] d... 853.859002: #8 inner/outer(us): 9/12 ts:1581527658.889065736 count:1 2190 <...>-1729 [001] d... 855.874978: #9 inner/outer(us): 9/11 ts:1581527660.861991877 count:1 2191 <...>-1729 [001] d... 863.938932: #10 inner/outer(us): 9/11 ts:1581527668.970010500 count:1 nmi-total:7 nmi-count:1 2192 <...>-1729 [007] d... 878.050780: #11 inner/outer(us): 9/12 ts:1581527683.385002600 count:1 nmi-total:5 nmi-count:1 2193 <...>-1729 [007] d... 886.114702: #12 inner/outer(us): 9/12 ts:1581527691.385001600 count:1 2194 2195 2196The above output is somewhat the same in the header. All events will have 2197interrupts disabled 'd'. Under the FUNCTION title there is: 2198 2199 #1 2200 This is the count of events recorded that were greater than the 2201 tracing_threshold (See below). 2202 2203 inner/outer(us): 11/11 2204 2205 This shows two numbers as "inner latency" and "outer latency". The test 2206 runs in a loop checking a timestamp twice. The latency detected within 2207 the two timestamps is the "inner latency" and the latency detected 2208 after the previous timestamp and the next timestamp in the loop is 2209 the "outer latency". 2210 2211 ts:1581527483.343962693 2212 2213 The absolute timestamp that the first latency was recorded in the window. 2214 2215 count:6 2216 2217 The number of times a latency was detected during the window. 2218 2219 nmi-total:7 nmi-count:1 2220 2221 On architectures that support it, if an NMI comes in during the 2222 test, the time spent in NMI is reported in "nmi-total" (in 2223 microseconds). 2224 2225 All architectures that have NMIs will show the "nmi-count" if an 2226 NMI comes in during the test. 2227 2228hwlat files: 2229 2230 tracing_threshold 2231 This gets automatically set to "10" to represent 10 2232 microseconds. This is the threshold of latency that 2233 needs to be detected before the trace will be recorded. 2234 2235 Note, when hwlat tracer is finished (another tracer is 2236 written into "current_tracer"), the original value for 2237 tracing_threshold is placed back into this file. 2238 2239 hwlat_detector/width 2240 The length of time the test runs with interrupts disabled. 2241 2242 hwlat_detector/window 2243 The length of time of the window which the test 2244 runs. That is, the test will run for "width" 2245 microseconds per "window" microseconds 2246 2247 tracing_cpumask 2248 When the test is started. A kernel thread is created that 2249 runs the test. This thread will alternate between CPUs 2250 listed in the tracing_cpumask between each period 2251 (one "window"). To limit the test to specific CPUs 2252 set the mask in this file to only the CPUs that the test 2253 should run on. 2254 2255function 2256-------- 2257 2258This tracer is the function tracer. Enabling the function tracer 2259can be done from the debug file system. Make sure the 2260ftrace_enabled is set; otherwise this tracer is a nop. 2261See the "ftrace_enabled" section below. 2262:: 2263 2264 # sysctl kernel.ftrace_enabled=1 2265 # echo function > current_tracer 2266 # echo 1 > tracing_on 2267 # usleep 1 2268 # echo 0 > tracing_on 2269 # cat trace 2270 # tracer: function 2271 # 2272 # entries-in-buffer/entries-written: 24799/24799 #P:4 2273 # 2274 # _-----=> irqs-off 2275 # / _----=> need-resched 2276 # | / _---=> hardirq/softirq 2277 # || / _--=> preempt-depth 2278 # ||| / delay 2279 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 2280 # | | | |||| | | 2281 bash-1994 [002] .... 3082.063030: mutex_unlock <-rb_simple_write 2282 bash-1994 [002] .... 3082.063031: __mutex_unlock_slowpath <-mutex_unlock 2283 bash-1994 [002] .... 3082.063031: __fsnotify_parent <-fsnotify_modify 2284 bash-1994 [002] .... 3082.063032: fsnotify <-fsnotify_modify 2285 bash-1994 [002] .... 3082.063032: __srcu_read_lock <-fsnotify 2286 bash-1994 [002] .... 3082.063032: add_preempt_count <-__srcu_read_lock 2287 bash-1994 [002] ...1 3082.063032: sub_preempt_count <-__srcu_read_lock 2288 bash-1994 [002] .... 3082.063033: __srcu_read_unlock <-fsnotify 2289 [...] 2290 2291 2292Note: function tracer uses ring buffers to store the above 2293entries. The newest data may overwrite the oldest data. 2294Sometimes using echo to stop the trace is not sufficient because 2295the tracing could have overwritten the data that you wanted to 2296record. For this reason, it is sometimes better to disable 2297tracing directly from a program. This allows you to stop the 2298tracing at the point that you hit the part that you are 2299interested in. To disable the tracing directly from a C program, 2300something like following code snippet can be used:: 2301 2302 int trace_fd; 2303 [...] 2304 int main(int argc, char *argv[]) { 2305 [...] 2306 trace_fd = open(tracing_file("tracing_on"), O_WRONLY); 2307 [...] 2308 if (condition_hit()) { 2309 write(trace_fd, "0", 1); 2310 } 2311 [...] 2312 } 2313 2314 2315Single thread tracing 2316--------------------- 2317 2318By writing into set_ftrace_pid you can trace a 2319single thread. For example:: 2320 2321 # cat set_ftrace_pid 2322 no pid 2323 # echo 3111 > set_ftrace_pid 2324 # cat set_ftrace_pid 2325 3111 2326 # echo function > current_tracer 2327 # cat trace | head 2328 # tracer: function 2329 # 2330 # TASK-PID CPU# TIMESTAMP FUNCTION 2331 # | | | | | 2332 yum-updatesd-3111 [003] 1637.254676: finish_task_switch <-thread_return 2333 yum-updatesd-3111 [003] 1637.254681: hrtimer_cancel <-schedule_hrtimeout_range 2334 yum-updatesd-3111 [003] 1637.254682: hrtimer_try_to_cancel <-hrtimer_cancel 2335 yum-updatesd-3111 [003] 1637.254683: lock_hrtimer_base <-hrtimer_try_to_cancel 2336 yum-updatesd-3111 [003] 1637.254685: fget_light <-do_sys_poll 2337 yum-updatesd-3111 [003] 1637.254686: pipe_poll <-do_sys_poll 2338 # echo > set_ftrace_pid 2339 # cat trace |head 2340 # tracer: function 2341 # 2342 # TASK-PID CPU# TIMESTAMP FUNCTION 2343 # | | | | | 2344 ##### CPU 3 buffer started #### 2345 yum-updatesd-3111 [003] 1701.957688: free_poll_entry <-poll_freewait 2346 yum-updatesd-3111 [003] 1701.957689: remove_wait_queue <-free_poll_entry 2347 yum-updatesd-3111 [003] 1701.957691: fput <-free_poll_entry 2348 yum-updatesd-3111 [003] 1701.957692: audit_syscall_exit <-sysret_audit 2349 yum-updatesd-3111 [003] 1701.957693: path_put <-audit_syscall_exit 2350 2351If you want to trace a function when executing, you could use 2352something like this simple program. 2353:: 2354 2355 #include <stdio.h> 2356 #include <stdlib.h> 2357 #include <sys/types.h> 2358 #include <sys/stat.h> 2359 #include <fcntl.h> 2360 #include <unistd.h> 2361 #include <string.h> 2362 2363 #define _STR(x) #x 2364 #define STR(x) _STR(x) 2365 #define MAX_PATH 256 2366 2367 const char *find_tracefs(void) 2368 { 2369 static char tracefs[MAX_PATH+1]; 2370 static int tracefs_found; 2371 char type[100]; 2372 FILE *fp; 2373 2374 if (tracefs_found) 2375 return tracefs; 2376 2377 if ((fp = fopen("/proc/mounts","r")) == NULL) { 2378 perror("/proc/mounts"); 2379 return NULL; 2380 } 2381 2382 while (fscanf(fp, "%*s %" 2383 STR(MAX_PATH) 2384 "s %99s %*s %*d %*d\n", 2385 tracefs, type) == 2) { 2386 if (strcmp(type, "tracefs") == 0) 2387 break; 2388 } 2389 fclose(fp); 2390 2391 if (strcmp(type, "tracefs") != 0) { 2392 fprintf(stderr, "tracefs not mounted"); 2393 return NULL; 2394 } 2395 2396 strcat(tracefs, "/tracing/"); 2397 tracefs_found = 1; 2398 2399 return tracefs; 2400 } 2401 2402 const char *tracing_file(const char *file_name) 2403 { 2404 static char trace_file[MAX_PATH+1]; 2405 snprintf(trace_file, MAX_PATH, "%s/%s", find_tracefs(), file_name); 2406 return trace_file; 2407 } 2408 2409 int main (int argc, char **argv) 2410 { 2411 if (argc < 1) 2412 exit(-1); 2413 2414 if (fork() > 0) { 2415 int fd, ffd; 2416 char line[64]; 2417 int s; 2418 2419 ffd = open(tracing_file("current_tracer"), O_WRONLY); 2420 if (ffd < 0) 2421 exit(-1); 2422 write(ffd, "nop", 3); 2423 2424 fd = open(tracing_file("set_ftrace_pid"), O_WRONLY); 2425 s = sprintf(line, "%d\n", getpid()); 2426 write(fd, line, s); 2427 2428 write(ffd, "function", 8); 2429 2430 close(fd); 2431 close(ffd); 2432 2433 execvp(argv[1], argv+1); 2434 } 2435 2436 return 0; 2437 } 2438 2439Or this simple script! 2440:: 2441 2442 #!/bin/bash 2443 2444 tracefs=`sed -ne 's/^tracefs \(.*\) tracefs.*/\1/p' /proc/mounts` 2445 echo nop > $tracefs/tracing/current_tracer 2446 echo 0 > $tracefs/tracing/tracing_on 2447 echo $$ > $tracefs/tracing/set_ftrace_pid 2448 echo function > $tracefs/tracing/current_tracer 2449 echo 1 > $tracefs/tracing/tracing_on 2450 exec "$@" 2451 2452 2453function graph tracer 2454--------------------------- 2455 2456This tracer is similar to the function tracer except that it 2457probes a function on its entry and its exit. This is done by 2458using a dynamically allocated stack of return addresses in each 2459task_struct. On function entry the tracer overwrites the return 2460address of each function traced to set a custom probe. Thus the 2461original return address is stored on the stack of return address 2462in the task_struct. 2463 2464Probing on both ends of a function leads to special features 2465such as: 2466 2467- measure of a function's time execution 2468- having a reliable call stack to draw function calls graph 2469 2470This tracer is useful in several situations: 2471 2472- you want to find the reason of a strange kernel behavior and 2473 need to see what happens in detail on any areas (or specific 2474 ones). 2475 2476- you are experiencing weird latencies but it's difficult to 2477 find its origin. 2478 2479- you want to find quickly which path is taken by a specific 2480 function 2481 2482- you just want to peek inside a working kernel and want to see 2483 what happens there. 2484 2485:: 2486 2487 # tracer: function_graph 2488 # 2489 # CPU DURATION FUNCTION CALLS 2490 # | | | | | | | 2491 2492 0) | sys_open() { 2493 0) | do_sys_open() { 2494 0) | getname() { 2495 0) | kmem_cache_alloc() { 2496 0) 1.382 us | __might_sleep(); 2497 0) 2.478 us | } 2498 0) | strncpy_from_user() { 2499 0) | might_fault() { 2500 0) 1.389 us | __might_sleep(); 2501 0) 2.553 us | } 2502 0) 3.807 us | } 2503 0) 7.876 us | } 2504 0) | alloc_fd() { 2505 0) 0.668 us | _spin_lock(); 2506 0) 0.570 us | expand_files(); 2507 0) 0.586 us | _spin_unlock(); 2508 2509 2510There are several columns that can be dynamically 2511enabled/disabled. You can use every combination of options you 2512want, depending on your needs. 2513 2514- The cpu number on which the function executed is default 2515 enabled. It is sometimes better to only trace one cpu (see 2516 tracing_cpu_mask file) or you might sometimes see unordered 2517 function calls while cpu tracing switch. 2518 2519 - hide: echo nofuncgraph-cpu > trace_options 2520 - show: echo funcgraph-cpu > trace_options 2521 2522- The duration (function's time of execution) is displayed on 2523 the closing bracket line of a function or on the same line 2524 than the current function in case of a leaf one. It is default 2525 enabled. 2526 2527 - hide: echo nofuncgraph-duration > trace_options 2528 - show: echo funcgraph-duration > trace_options 2529 2530- The overhead field precedes the duration field in case of 2531 reached duration thresholds. 2532 2533 - hide: echo nofuncgraph-overhead > trace_options 2534 - show: echo funcgraph-overhead > trace_options 2535 - depends on: funcgraph-duration 2536 2537 ie:: 2538 2539 3) # 1837.709 us | } /* __switch_to */ 2540 3) | finish_task_switch() { 2541 3) 0.313 us | _raw_spin_unlock_irq(); 2542 3) 3.177 us | } 2543 3) # 1889.063 us | } /* __schedule */ 2544 3) ! 140.417 us | } /* __schedule */ 2545 3) # 2034.948 us | } /* schedule */ 2546 3) * 33998.59 us | } /* schedule_preempt_disabled */ 2547 2548 [...] 2549 2550 1) 0.260 us | msecs_to_jiffies(); 2551 1) 0.313 us | __rcu_read_unlock(); 2552 1) + 61.770 us | } 2553 1) + 64.479 us | } 2554 1) 0.313 us | rcu_bh_qs(); 2555 1) 0.313 us | __local_bh_enable(); 2556 1) ! 217.240 us | } 2557 1) 0.365 us | idle_cpu(); 2558 1) | rcu_irq_exit() { 2559 1) 0.417 us | rcu_eqs_enter_common.isra.47(); 2560 1) 3.125 us | } 2561 1) ! 227.812 us | } 2562 1) ! 457.395 us | } 2563 1) @ 119760.2 us | } 2564 2565 [...] 2566 2567 2) | handle_IPI() { 2568 1) 6.979 us | } 2569 2) 0.417 us | scheduler_ipi(); 2570 1) 9.791 us | } 2571 1) + 12.917 us | } 2572 2) 3.490 us | } 2573 1) + 15.729 us | } 2574 1) + 18.542 us | } 2575 2) $ 3594274 us | } 2576 2577Flags:: 2578 2579 + means that the function exceeded 10 usecs. 2580 ! means that the function exceeded 100 usecs. 2581 # means that the function exceeded 1000 usecs. 2582 * means that the function exceeded 10 msecs. 2583 @ means that the function exceeded 100 msecs. 2584 $ means that the function exceeded 1 sec. 2585 2586 2587- The task/pid field displays the thread cmdline and pid which 2588 executed the function. It is default disabled. 2589 2590 - hide: echo nofuncgraph-proc > trace_options 2591 - show: echo funcgraph-proc > trace_options 2592 2593 ie:: 2594 2595 # tracer: function_graph 2596 # 2597 # CPU TASK/PID DURATION FUNCTION CALLS 2598 # | | | | | | | | | 2599 0) sh-4802 | | d_free() { 2600 0) sh-4802 | | call_rcu() { 2601 0) sh-4802 | | __call_rcu() { 2602 0) sh-4802 | 0.616 us | rcu_process_gp_end(); 2603 0) sh-4802 | 0.586 us | check_for_new_grace_period(); 2604 0) sh-4802 | 2.899 us | } 2605 0) sh-4802 | 4.040 us | } 2606 0) sh-4802 | 5.151 us | } 2607 0) sh-4802 | + 49.370 us | } 2608 2609 2610- The absolute time field is an absolute timestamp given by the 2611 system clock since it started. A snapshot of this time is 2612 given on each entry/exit of functions 2613 2614 - hide: echo nofuncgraph-abstime > trace_options 2615 - show: echo funcgraph-abstime > trace_options 2616 2617 ie:: 2618 2619 # 2620 # TIME CPU DURATION FUNCTION CALLS 2621 # | | | | | | | | 2622 360.774522 | 1) 0.541 us | } 2623 360.774522 | 1) 4.663 us | } 2624 360.774523 | 1) 0.541 us | __wake_up_bit(); 2625 360.774524 | 1) 6.796 us | } 2626 360.774524 | 1) 7.952 us | } 2627 360.774525 | 1) 9.063 us | } 2628 360.774525 | 1) 0.615 us | journal_mark_dirty(); 2629 360.774527 | 1) 0.578 us | __brelse(); 2630 360.774528 | 1) | reiserfs_prepare_for_journal() { 2631 360.774528 | 1) | unlock_buffer() { 2632 360.774529 | 1) | wake_up_bit() { 2633 360.774529 | 1) | bit_waitqueue() { 2634 360.774530 | 1) 0.594 us | __phys_addr(); 2635 2636 2637The function name is always displayed after the closing bracket 2638for a function if the start of that function is not in the 2639trace buffer. 2640 2641Display of the function name after the closing bracket may be 2642enabled for functions whose start is in the trace buffer, 2643allowing easier searching with grep for function durations. 2644It is default disabled. 2645 2646 - hide: echo nofuncgraph-tail > trace_options 2647 - show: echo funcgraph-tail > trace_options 2648 2649 Example with nofuncgraph-tail (default):: 2650 2651 0) | putname() { 2652 0) | kmem_cache_free() { 2653 0) 0.518 us | __phys_addr(); 2654 0) 1.757 us | } 2655 0) 2.861 us | } 2656 2657 Example with funcgraph-tail:: 2658 2659 0) | putname() { 2660 0) | kmem_cache_free() { 2661 0) 0.518 us | __phys_addr(); 2662 0) 1.757 us | } /* kmem_cache_free() */ 2663 0) 2.861 us | } /* putname() */ 2664 2665You can put some comments on specific functions by using 2666trace_printk() For example, if you want to put a comment inside 2667the __might_sleep() function, you just have to include 2668<linux/ftrace.h> and call trace_printk() inside __might_sleep():: 2669 2670 trace_printk("I'm a comment!\n") 2671 2672will produce:: 2673 2674 1) | __might_sleep() { 2675 1) | /* I'm a comment! */ 2676 1) 1.449 us | } 2677 2678 2679You might find other useful features for this tracer in the 2680following "dynamic ftrace" section such as tracing only specific 2681functions or tasks. 2682 2683dynamic ftrace 2684-------------- 2685 2686If CONFIG_DYNAMIC_FTRACE is set, the system will run with 2687virtually no overhead when function tracing is disabled. The way 2688this works is the mcount function call (placed at the start of 2689every kernel function, produced by the -pg switch in gcc), 2690starts of pointing to a simple return. (Enabling FTRACE will 2691include the -pg switch in the compiling of the kernel.) 2692 2693At compile time every C file object is run through the 2694recordmcount program (located in the scripts directory). This 2695program will parse the ELF headers in the C object to find all 2696the locations in the .text section that call mcount. Starting 2697with gcc version 4.6, the -mfentry has been added for x86, which 2698calls "__fentry__" instead of "mcount". Which is called before 2699the creation of the stack frame. 2700 2701Note, not all sections are traced. They may be prevented by either 2702a notrace, or blocked another way and all inline functions are not 2703traced. Check the "available_filter_functions" file to see what functions 2704can be traced. 2705 2706A section called "__mcount_loc" is created that holds 2707references to all the mcount/fentry call sites in the .text section. 2708The recordmcount program re-links this section back into the 2709original object. The final linking stage of the kernel will add all these 2710references into a single table. 2711 2712On boot up, before SMP is initialized, the dynamic ftrace code 2713scans this table and updates all the locations into nops. It 2714also records the locations, which are added to the 2715available_filter_functions list. Modules are processed as they 2716are loaded and before they are executed. When a module is 2717unloaded, it also removes its functions from the ftrace function 2718list. This is automatic in the module unload code, and the 2719module author does not need to worry about it. 2720 2721When tracing is enabled, the process of modifying the function 2722tracepoints is dependent on architecture. The old method is to use 2723kstop_machine to prevent races with the CPUs executing code being 2724modified (which can cause the CPU to do undesirable things, especially 2725if the modified code crosses cache (or page) boundaries), and the nops are 2726patched back to calls. But this time, they do not call mcount 2727(which is just a function stub). They now call into the ftrace 2728infrastructure. 2729 2730The new method of modifying the function tracepoints is to place 2731a breakpoint at the location to be modified, sync all CPUs, modify 2732the rest of the instruction not covered by the breakpoint. Sync 2733all CPUs again, and then remove the breakpoint with the finished 2734version to the ftrace call site. 2735 2736Some archs do not even need to monkey around with the synchronization, 2737and can just slap the new code on top of the old without any 2738problems with other CPUs executing it at the same time. 2739 2740One special side-effect to the recording of the functions being 2741traced is that we can now selectively choose which functions we 2742wish to trace and which ones we want the mcount calls to remain 2743as nops. 2744 2745Two files are used, one for enabling and one for disabling the 2746tracing of specified functions. They are: 2747 2748 set_ftrace_filter 2749 2750and 2751 2752 set_ftrace_notrace 2753 2754A list of available functions that you can add to these files is 2755listed in: 2756 2757 available_filter_functions 2758 2759:: 2760 2761 # cat available_filter_functions 2762 put_prev_task_idle 2763 kmem_cache_create 2764 pick_next_task_rt 2765 get_online_cpus 2766 pick_next_task_fair 2767 mutex_lock 2768 [...] 2769 2770If I am only interested in sys_nanosleep and hrtimer_interrupt:: 2771 2772 # echo sys_nanosleep hrtimer_interrupt > set_ftrace_filter 2773 # echo function > current_tracer 2774 # echo 1 > tracing_on 2775 # usleep 1 2776 # echo 0 > tracing_on 2777 # cat trace 2778 # tracer: function 2779 # 2780 # entries-in-buffer/entries-written: 5/5 #P:4 2781 # 2782 # _-----=> irqs-off 2783 # / _----=> need-resched 2784 # | / _---=> hardirq/softirq 2785 # || / _--=> preempt-depth 2786 # ||| / delay 2787 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 2788 # | | | |||| | | 2789 usleep-2665 [001] .... 4186.475355: sys_nanosleep <-system_call_fastpath 2790 <idle>-0 [001] d.h1 4186.475409: hrtimer_interrupt <-smp_apic_timer_interrupt 2791 usleep-2665 [001] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt 2792 <idle>-0 [003] d.h1 4186.475426: hrtimer_interrupt <-smp_apic_timer_interrupt 2793 <idle>-0 [002] d.h1 4186.475427: hrtimer_interrupt <-smp_apic_timer_interrupt 2794 2795To see which functions are being traced, you can cat the file: 2796:: 2797 2798 # cat set_ftrace_filter 2799 hrtimer_interrupt 2800 sys_nanosleep 2801 2802 2803Perhaps this is not enough. The filters also allow glob(7) matching. 2804 2805 ``<match>*`` 2806 will match functions that begin with <match> 2807 ``*<match>`` 2808 will match functions that end with <match> 2809 ``*<match>*`` 2810 will match functions that have <match> in it 2811 ``<match1>*<match2>`` 2812 will match functions that begin with <match1> and end with <match2> 2813 2814.. note:: 2815 It is better to use quotes to enclose the wild cards, 2816 otherwise the shell may expand the parameters into names 2817 of files in the local directory. 2818 2819:: 2820 2821 # echo 'hrtimer_*' > set_ftrace_filter 2822 2823Produces:: 2824 2825 # tracer: function 2826 # 2827 # entries-in-buffer/entries-written: 897/897 #P:4 2828 # 2829 # _-----=> irqs-off 2830 # / _----=> need-resched 2831 # | / _---=> hardirq/softirq 2832 # || / _--=> preempt-depth 2833 # ||| / delay 2834 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 2835 # | | | |||| | | 2836 <idle>-0 [003] dN.1 4228.547803: hrtimer_cancel <-tick_nohz_idle_exit 2837 <idle>-0 [003] dN.1 4228.547804: hrtimer_try_to_cancel <-hrtimer_cancel 2838 <idle>-0 [003] dN.2 4228.547805: hrtimer_force_reprogram <-__remove_hrtimer 2839 <idle>-0 [003] dN.1 4228.547805: hrtimer_forward <-tick_nohz_idle_exit 2840 <idle>-0 [003] dN.1 4228.547805: hrtimer_start_range_ns <-hrtimer_start_expires.constprop.11 2841 <idle>-0 [003] d..1 4228.547858: hrtimer_get_next_event <-get_next_timer_interrupt 2842 <idle>-0 [003] d..1 4228.547859: hrtimer_start <-__tick_nohz_idle_enter 2843 <idle>-0 [003] d..2 4228.547860: hrtimer_force_reprogram <-__rem 2844 2845Notice that we lost the sys_nanosleep. 2846:: 2847 2848 # cat set_ftrace_filter 2849 hrtimer_run_queues 2850 hrtimer_run_pending 2851 hrtimer_init 2852 hrtimer_cancel 2853 hrtimer_try_to_cancel 2854 hrtimer_forward 2855 hrtimer_start 2856 hrtimer_reprogram 2857 hrtimer_force_reprogram 2858 hrtimer_get_next_event 2859 hrtimer_interrupt 2860 hrtimer_nanosleep 2861 hrtimer_wakeup 2862 hrtimer_get_remaining 2863 hrtimer_get_res 2864 hrtimer_init_sleeper 2865 2866 2867This is because the '>' and '>>' act just like they do in bash. 2868To rewrite the filters, use '>' 2869To append to the filters, use '>>' 2870 2871To clear out a filter so that all functions will be recorded 2872again:: 2873 2874 # echo > set_ftrace_filter 2875 # cat set_ftrace_filter 2876 # 2877 2878Again, now we want to append. 2879 2880:: 2881 2882 # echo sys_nanosleep > set_ftrace_filter 2883 # cat set_ftrace_filter 2884 sys_nanosleep 2885 # echo 'hrtimer_*' >> set_ftrace_filter 2886 # cat set_ftrace_filter 2887 hrtimer_run_queues 2888 hrtimer_run_pending 2889 hrtimer_init 2890 hrtimer_cancel 2891 hrtimer_try_to_cancel 2892 hrtimer_forward 2893 hrtimer_start 2894 hrtimer_reprogram 2895 hrtimer_force_reprogram 2896 hrtimer_get_next_event 2897 hrtimer_interrupt 2898 sys_nanosleep 2899 hrtimer_nanosleep 2900 hrtimer_wakeup 2901 hrtimer_get_remaining 2902 hrtimer_get_res 2903 hrtimer_init_sleeper 2904 2905 2906The set_ftrace_notrace prevents those functions from being 2907traced. 2908:: 2909 2910 # echo '*preempt*' '*lock*' > set_ftrace_notrace 2911 2912Produces:: 2913 2914 # tracer: function 2915 # 2916 # entries-in-buffer/entries-written: 39608/39608 #P:4 2917 # 2918 # _-----=> irqs-off 2919 # / _----=> need-resched 2920 # | / _---=> hardirq/softirq 2921 # || / _--=> preempt-depth 2922 # ||| / delay 2923 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 2924 # | | | |||| | | 2925 bash-1994 [000] .... 4342.324896: file_ra_state_init <-do_dentry_open 2926 bash-1994 [000] .... 4342.324897: open_check_o_direct <-do_last 2927 bash-1994 [000] .... 4342.324897: ima_file_check <-do_last 2928 bash-1994 [000] .... 4342.324898: process_measurement <-ima_file_check 2929 bash-1994 [000] .... 4342.324898: ima_get_action <-process_measurement 2930 bash-1994 [000] .... 4342.324898: ima_match_policy <-ima_get_action 2931 bash-1994 [000] .... 4342.324899: do_truncate <-do_last 2932 bash-1994 [000] .... 4342.324899: should_remove_suid <-do_truncate 2933 bash-1994 [000] .... 4342.324899: notify_change <-do_truncate 2934 bash-1994 [000] .... 4342.324900: current_fs_time <-notify_change 2935 bash-1994 [000] .... 4342.324900: current_kernel_time <-current_fs_time 2936 bash-1994 [000] .... 4342.324900: timespec_trunc <-current_fs_time 2937 2938We can see that there's no more lock or preempt tracing. 2939 2940Selecting function filters via index 2941------------------------------------ 2942 2943Because processing of strings is expensive (the address of the function 2944needs to be looked up before comparing to the string being passed in), 2945an index can be used as well to enable functions. This is useful in the 2946case of setting thousands of specific functions at a time. By passing 2947in a list of numbers, no string processing will occur. Instead, the function 2948at the specific location in the internal array (which corresponds to the 2949functions in the "available_filter_functions" file), is selected. 2950 2951:: 2952 2953 # echo 1 > set_ftrace_filter 2954 2955Will select the first function listed in "available_filter_functions" 2956 2957:: 2958 2959 # head -1 available_filter_functions 2960 trace_initcall_finish_cb 2961 2962 # cat set_ftrace_filter 2963 trace_initcall_finish_cb 2964 2965 # head -50 available_filter_functions | tail -1 2966 x86_pmu_commit_txn 2967 2968 # echo 1 50 > set_ftrace_filter 2969 # cat set_ftrace_filter 2970 trace_initcall_finish_cb 2971 x86_pmu_commit_txn 2972 2973Dynamic ftrace with the function graph tracer 2974--------------------------------------------- 2975 2976Although what has been explained above concerns both the 2977function tracer and the function-graph-tracer, there are some 2978special features only available in the function-graph tracer. 2979 2980If you want to trace only one function and all of its children, 2981you just have to echo its name into set_graph_function:: 2982 2983 echo __do_fault > set_graph_function 2984 2985will produce the following "expanded" trace of the __do_fault() 2986function:: 2987 2988 0) | __do_fault() { 2989 0) | filemap_fault() { 2990 0) | find_lock_page() { 2991 0) 0.804 us | find_get_page(); 2992 0) | __might_sleep() { 2993 0) 1.329 us | } 2994 0) 3.904 us | } 2995 0) 4.979 us | } 2996 0) 0.653 us | _spin_lock(); 2997 0) 0.578 us | page_add_file_rmap(); 2998 0) 0.525 us | native_set_pte_at(); 2999 0) 0.585 us | _spin_unlock(); 3000 0) | unlock_page() { 3001 0) 0.541 us | page_waitqueue(); 3002 0) 0.639 us | __wake_up_bit(); 3003 0) 2.786 us | } 3004 0) + 14.237 us | } 3005 0) | __do_fault() { 3006 0) | filemap_fault() { 3007 0) | find_lock_page() { 3008 0) 0.698 us | find_get_page(); 3009 0) | __might_sleep() { 3010 0) 1.412 us | } 3011 0) 3.950 us | } 3012 0) 5.098 us | } 3013 0) 0.631 us | _spin_lock(); 3014 0) 0.571 us | page_add_file_rmap(); 3015 0) 0.526 us | native_set_pte_at(); 3016 0) 0.586 us | _spin_unlock(); 3017 0) | unlock_page() { 3018 0) 0.533 us | page_waitqueue(); 3019 0) 0.638 us | __wake_up_bit(); 3020 0) 2.793 us | } 3021 0) + 14.012 us | } 3022 3023You can also expand several functions at once:: 3024 3025 echo sys_open > set_graph_function 3026 echo sys_close >> set_graph_function 3027 3028Now if you want to go back to trace all functions you can clear 3029this special filter via:: 3030 3031 echo > set_graph_function 3032 3033 3034ftrace_enabled 3035-------------- 3036 3037Note, the proc sysctl ftrace_enable is a big on/off switch for the 3038function tracer. By default it is enabled (when function tracing is 3039enabled in the kernel). If it is disabled, all function tracing is 3040disabled. This includes not only the function tracers for ftrace, but 3041also for any other uses (perf, kprobes, stack tracing, profiling, etc). It 3042cannot be disabled if there is a callback with FTRACE_OPS_FL_PERMANENT set 3043registered. 3044 3045Please disable this with care. 3046 3047This can be disable (and enabled) with:: 3048 3049 sysctl kernel.ftrace_enabled=0 3050 sysctl kernel.ftrace_enabled=1 3051 3052 or 3053 3054 echo 0 > /proc/sys/kernel/ftrace_enabled 3055 echo 1 > /proc/sys/kernel/ftrace_enabled 3056 3057 3058Filter commands 3059--------------- 3060 3061A few commands are supported by the set_ftrace_filter interface. 3062Trace commands have the following format:: 3063 3064 <function>:<command>:<parameter> 3065 3066The following commands are supported: 3067 3068- mod: 3069 This command enables function filtering per module. The 3070 parameter defines the module. For example, if only the write* 3071 functions in the ext3 module are desired, run: 3072 3073 echo 'write*:mod:ext3' > set_ftrace_filter 3074 3075 This command interacts with the filter in the same way as 3076 filtering based on function names. Thus, adding more functions 3077 in a different module is accomplished by appending (>>) to the 3078 filter file. Remove specific module functions by prepending 3079 '!':: 3080 3081 echo '!writeback*:mod:ext3' >> set_ftrace_filter 3082 3083 Mod command supports module globbing. Disable tracing for all 3084 functions except a specific module:: 3085 3086 echo '!*:mod:!ext3' >> set_ftrace_filter 3087 3088 Disable tracing for all modules, but still trace kernel:: 3089 3090 echo '!*:mod:*' >> set_ftrace_filter 3091 3092 Enable filter only for kernel:: 3093 3094 echo '*write*:mod:!*' >> set_ftrace_filter 3095 3096 Enable filter for module globbing:: 3097 3098 echo '*write*:mod:*snd*' >> set_ftrace_filter 3099 3100- traceon/traceoff: 3101 These commands turn tracing on and off when the specified 3102 functions are hit. The parameter determines how many times the 3103 tracing system is turned on and off. If unspecified, there is 3104 no limit. For example, to disable tracing when a schedule bug 3105 is hit the first 5 times, run:: 3106 3107 echo '__schedule_bug:traceoff:5' > set_ftrace_filter 3108 3109 To always disable tracing when __schedule_bug is hit:: 3110 3111 echo '__schedule_bug:traceoff' > set_ftrace_filter 3112 3113 These commands are cumulative whether or not they are appended 3114 to set_ftrace_filter. To remove a command, prepend it by '!' 3115 and drop the parameter:: 3116 3117 echo '!__schedule_bug:traceoff:0' > set_ftrace_filter 3118 3119 The above removes the traceoff command for __schedule_bug 3120 that have a counter. To remove commands without counters:: 3121 3122 echo '!__schedule_bug:traceoff' > set_ftrace_filter 3123 3124- snapshot: 3125 Will cause a snapshot to be triggered when the function is hit. 3126 :: 3127 3128 echo 'native_flush_tlb_others:snapshot' > set_ftrace_filter 3129 3130 To only snapshot once: 3131 :: 3132 3133 echo 'native_flush_tlb_others:snapshot:1' > set_ftrace_filter 3134 3135 To remove the above commands:: 3136 3137 echo '!native_flush_tlb_others:snapshot' > set_ftrace_filter 3138 echo '!native_flush_tlb_others:snapshot:0' > set_ftrace_filter 3139 3140- enable_event/disable_event: 3141 These commands can enable or disable a trace event. Note, because 3142 function tracing callbacks are very sensitive, when these commands 3143 are registered, the trace point is activated, but disabled in 3144 a "soft" mode. That is, the tracepoint will be called, but 3145 just will not be traced. The event tracepoint stays in this mode 3146 as long as there's a command that triggers it. 3147 :: 3148 3149 echo 'try_to_wake_up:enable_event:sched:sched_switch:2' > \ 3150 set_ftrace_filter 3151 3152 The format is:: 3153 3154 <function>:enable_event:<system>:<event>[:count] 3155 <function>:disable_event:<system>:<event>[:count] 3156 3157 To remove the events commands:: 3158 3159 echo '!try_to_wake_up:enable_event:sched:sched_switch:0' > \ 3160 set_ftrace_filter 3161 echo '!schedule:disable_event:sched:sched_switch' > \ 3162 set_ftrace_filter 3163 3164- dump: 3165 When the function is hit, it will dump the contents of the ftrace 3166 ring buffer to the console. This is useful if you need to debug 3167 something, and want to dump the trace when a certain function 3168 is hit. Perhaps it's a function that is called before a triple 3169 fault happens and does not allow you to get a regular dump. 3170 3171- cpudump: 3172 When the function is hit, it will dump the contents of the ftrace 3173 ring buffer for the current CPU to the console. Unlike the "dump" 3174 command, it only prints out the contents of the ring buffer for the 3175 CPU that executed the function that triggered the dump. 3176 3177- stacktrace: 3178 When the function is hit, a stack trace is recorded. 3179 3180trace_pipe 3181---------- 3182 3183The trace_pipe outputs the same content as the trace file, but 3184the effect on the tracing is different. Every read from 3185trace_pipe is consumed. This means that subsequent reads will be 3186different. The trace is live. 3187:: 3188 3189 # echo function > current_tracer 3190 # cat trace_pipe > /tmp/trace.out & 3191 [1] 4153 3192 # echo 1 > tracing_on 3193 # usleep 1 3194 # echo 0 > tracing_on 3195 # cat trace 3196 # tracer: function 3197 # 3198 # entries-in-buffer/entries-written: 0/0 #P:4 3199 # 3200 # _-----=> irqs-off 3201 # / _----=> need-resched 3202 # | / _---=> hardirq/softirq 3203 # || / _--=> preempt-depth 3204 # ||| / delay 3205 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 3206 # | | | |||| | | 3207 3208 # 3209 # cat /tmp/trace.out 3210 bash-1994 [000] .... 5281.568961: mutex_unlock <-rb_simple_write 3211 bash-1994 [000] .... 5281.568963: __mutex_unlock_slowpath <-mutex_unlock 3212 bash-1994 [000] .... 5281.568963: __fsnotify_parent <-fsnotify_modify 3213 bash-1994 [000] .... 5281.568964: fsnotify <-fsnotify_modify 3214 bash-1994 [000] .... 5281.568964: __srcu_read_lock <-fsnotify 3215 bash-1994 [000] .... 5281.568964: add_preempt_count <-__srcu_read_lock 3216 bash-1994 [000] ...1 5281.568965: sub_preempt_count <-__srcu_read_lock 3217 bash-1994 [000] .... 5281.568965: __srcu_read_unlock <-fsnotify 3218 bash-1994 [000] .... 5281.568967: sys_dup2 <-system_call_fastpath 3219 3220 3221Note, reading the trace_pipe file will block until more input is 3222added. This is contrary to the trace file. If any process opened 3223the trace file for reading, it will actually disable tracing and 3224prevent new entries from being added. The trace_pipe file does 3225not have this limitation. 3226 3227trace entries 3228------------- 3229 3230Having too much or not enough data can be troublesome in 3231diagnosing an issue in the kernel. The file buffer_size_kb is 3232used to modify the size of the internal trace buffers. The 3233number listed is the number of entries that can be recorded per 3234CPU. To know the full size, multiply the number of possible CPUs 3235with the number of entries. 3236:: 3237 3238 # cat buffer_size_kb 3239 1408 (units kilobytes) 3240 3241Or simply read buffer_total_size_kb 3242:: 3243 3244 # cat buffer_total_size_kb 3245 5632 3246 3247To modify the buffer, simple echo in a number (in 1024 byte segments). 3248:: 3249 3250 # echo 10000 > buffer_size_kb 3251 # cat buffer_size_kb 3252 10000 (units kilobytes) 3253 3254It will try to allocate as much as possible. If you allocate too 3255much, it can cause Out-Of-Memory to trigger. 3256:: 3257 3258 # echo 1000000000000 > buffer_size_kb 3259 -bash: echo: write error: Cannot allocate memory 3260 # cat buffer_size_kb 3261 85 3262 3263The per_cpu buffers can be changed individually as well: 3264:: 3265 3266 # echo 10000 > per_cpu/cpu0/buffer_size_kb 3267 # echo 100 > per_cpu/cpu1/buffer_size_kb 3268 3269When the per_cpu buffers are not the same, the buffer_size_kb 3270at the top level will just show an X 3271:: 3272 3273 # cat buffer_size_kb 3274 X 3275 3276This is where the buffer_total_size_kb is useful: 3277:: 3278 3279 # cat buffer_total_size_kb 3280 12916 3281 3282Writing to the top level buffer_size_kb will reset all the buffers 3283to be the same again. 3284 3285Snapshot 3286-------- 3287CONFIG_TRACER_SNAPSHOT makes a generic snapshot feature 3288available to all non latency tracers. (Latency tracers which 3289record max latency, such as "irqsoff" or "wakeup", can't use 3290this feature, since those are already using the snapshot 3291mechanism internally.) 3292 3293Snapshot preserves a current trace buffer at a particular point 3294in time without stopping tracing. Ftrace swaps the current 3295buffer with a spare buffer, and tracing continues in the new 3296current (=previous spare) buffer. 3297 3298The following tracefs files in "tracing" are related to this 3299feature: 3300 3301 snapshot: 3302 3303 This is used to take a snapshot and to read the output 3304 of the snapshot. Echo 1 into this file to allocate a 3305 spare buffer and to take a snapshot (swap), then read 3306 the snapshot from this file in the same format as 3307 "trace" (described above in the section "The File 3308 System"). Both reads snapshot and tracing are executable 3309 in parallel. When the spare buffer is allocated, echoing 3310 0 frees it, and echoing else (positive) values clear the 3311 snapshot contents. 3312 More details are shown in the table below. 3313 3314 +--------------+------------+------------+------------+ 3315 |status\\input | 0 | 1 | else | 3316 +==============+============+============+============+ 3317 |not allocated |(do nothing)| alloc+swap |(do nothing)| 3318 +--------------+------------+------------+------------+ 3319 |allocated | free | swap | clear | 3320 +--------------+------------+------------+------------+ 3321 3322Here is an example of using the snapshot feature. 3323:: 3324 3325 # echo 1 > events/sched/enable 3326 # echo 1 > snapshot 3327 # cat snapshot 3328 # tracer: nop 3329 # 3330 # entries-in-buffer/entries-written: 71/71 #P:8 3331 # 3332 # _-----=> irqs-off 3333 # / _----=> need-resched 3334 # | / _---=> hardirq/softirq 3335 # || / _--=> preempt-depth 3336 # ||| / delay 3337 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 3338 # | | | |||| | | 3339 <idle>-0 [005] d... 2440.603828: sched_switch: prev_comm=swapper/5 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2242 next_prio=120 3340 sleep-2242 [005] d... 2440.603846: sched_switch: prev_comm=snapshot-test-2 prev_pid=2242 prev_prio=120 prev_state=R ==> next_comm=kworker/5:1 next_pid=60 next_prio=120 3341 [...] 3342 <idle>-0 [002] d... 2440.707230: sched_switch: prev_comm=swapper/2 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2229 next_prio=120 3343 3344 # cat trace 3345 # tracer: nop 3346 # 3347 # entries-in-buffer/entries-written: 77/77 #P:8 3348 # 3349 # _-----=> irqs-off 3350 # / _----=> need-resched 3351 # | / _---=> hardirq/softirq 3352 # || / _--=> preempt-depth 3353 # ||| / delay 3354 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 3355 # | | | |||| | | 3356 <idle>-0 [007] d... 2440.707395: sched_switch: prev_comm=swapper/7 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=snapshot-test-2 next_pid=2243 next_prio=120 3357 snapshot-test-2-2229 [002] d... 2440.707438: sched_switch: prev_comm=snapshot-test-2 prev_pid=2229 prev_prio=120 prev_state=S ==> next_comm=swapper/2 next_pid=0 next_prio=120 3358 [...] 3359 3360 3361If you try to use this snapshot feature when current tracer is 3362one of the latency tracers, you will get the following results. 3363:: 3364 3365 # echo wakeup > current_tracer 3366 # echo 1 > snapshot 3367 bash: echo: write error: Device or resource busy 3368 # cat snapshot 3369 cat: snapshot: Device or resource busy 3370 3371 3372Instances 3373--------- 3374In the tracefs tracing directory is a directory called "instances". 3375This directory can have new directories created inside of it using 3376mkdir, and removing directories with rmdir. The directory created 3377with mkdir in this directory will already contain files and other 3378directories after it is created. 3379:: 3380 3381 # mkdir instances/foo 3382 # ls instances/foo 3383 buffer_size_kb buffer_total_size_kb events free_buffer per_cpu 3384 set_event snapshot trace trace_clock trace_marker trace_options 3385 trace_pipe tracing_on 3386 3387As you can see, the new directory looks similar to the tracing directory 3388itself. In fact, it is very similar, except that the buffer and 3389events are agnostic from the main directory, or from any other 3390instances that are created. 3391 3392The files in the new directory work just like the files with the 3393same name in the tracing directory except the buffer that is used 3394is a separate and new buffer. The files affect that buffer but do not 3395affect the main buffer with the exception of trace_options. Currently, 3396the trace_options affect all instances and the top level buffer 3397the same, but this may change in future releases. That is, options 3398may become specific to the instance they reside in. 3399 3400Notice that none of the function tracer files are there, nor is 3401current_tracer and available_tracers. This is because the buffers 3402can currently only have events enabled for them. 3403:: 3404 3405 # mkdir instances/foo 3406 # mkdir instances/bar 3407 # mkdir instances/zoot 3408 # echo 100000 > buffer_size_kb 3409 # echo 1000 > instances/foo/buffer_size_kb 3410 # echo 5000 > instances/bar/per_cpu/cpu1/buffer_size_kb 3411 # echo function > current_trace 3412 # echo 1 > instances/foo/events/sched/sched_wakeup/enable 3413 # echo 1 > instances/foo/events/sched/sched_wakeup_new/enable 3414 # echo 1 > instances/foo/events/sched/sched_switch/enable 3415 # echo 1 > instances/bar/events/irq/enable 3416 # echo 1 > instances/zoot/events/syscalls/enable 3417 # cat trace_pipe 3418 CPU:2 [LOST 11745 EVENTS] 3419 bash-2044 [002] .... 10594.481032: _raw_spin_lock_irqsave <-get_page_from_freelist 3420 bash-2044 [002] d... 10594.481032: add_preempt_count <-_raw_spin_lock_irqsave 3421 bash-2044 [002] d..1 10594.481032: __rmqueue <-get_page_from_freelist 3422 bash-2044 [002] d..1 10594.481033: _raw_spin_unlock <-get_page_from_freelist 3423 bash-2044 [002] d..1 10594.481033: sub_preempt_count <-_raw_spin_unlock 3424 bash-2044 [002] d... 10594.481033: get_pageblock_flags_group <-get_pageblock_migratetype 3425 bash-2044 [002] d... 10594.481034: __mod_zone_page_state <-get_page_from_freelist 3426 bash-2044 [002] d... 10594.481034: zone_statistics <-get_page_from_freelist 3427 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics 3428 bash-2044 [002] d... 10594.481034: __inc_zone_state <-zone_statistics 3429 bash-2044 [002] .... 10594.481035: arch_dup_task_struct <-copy_process 3430 [...] 3431 3432 # cat instances/foo/trace_pipe 3433 bash-1998 [000] d..4 136.676759: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000 3434 bash-1998 [000] dN.4 136.676760: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000 3435 <idle>-0 [003] d.h3 136.676906: sched_wakeup: comm=rcu_preempt pid=9 prio=120 success=1 target_cpu=003 3436 <idle>-0 [003] d..3 136.676909: sched_switch: prev_comm=swapper/3 prev_pid=0 prev_prio=120 prev_state=R ==> next_comm=rcu_preempt next_pid=9 next_prio=120 3437 rcu_preempt-9 [003] d..3 136.676916: sched_switch: prev_comm=rcu_preempt prev_pid=9 prev_prio=120 prev_state=S ==> next_comm=swapper/3 next_pid=0 next_prio=120 3438 bash-1998 [000] d..4 136.677014: sched_wakeup: comm=kworker/0:1 pid=59 prio=120 success=1 target_cpu=000 3439 bash-1998 [000] dN.4 136.677016: sched_wakeup: comm=bash pid=1998 prio=120 success=1 target_cpu=000 3440 bash-1998 [000] d..3 136.677018: sched_switch: prev_comm=bash prev_pid=1998 prev_prio=120 prev_state=R+ ==> next_comm=kworker/0:1 next_pid=59 next_prio=120 3441 kworker/0:1-59 [000] d..4 136.677022: sched_wakeup: comm=sshd pid=1995 prio=120 success=1 target_cpu=001 3442 kworker/0:1-59 [000] d..3 136.677025: sched_switch: prev_comm=kworker/0:1 prev_pid=59 prev_prio=120 prev_state=S ==> next_comm=bash next_pid=1998 next_prio=120 3443 [...] 3444 3445 # cat instances/bar/trace_pipe 3446 migration/1-14 [001] d.h3 138.732674: softirq_raise: vec=3 [action=NET_RX] 3447 <idle>-0 [001] dNh3 138.732725: softirq_raise: vec=3 [action=NET_RX] 3448 bash-1998 [000] d.h1 138.733101: softirq_raise: vec=1 [action=TIMER] 3449 bash-1998 [000] d.h1 138.733102: softirq_raise: vec=9 [action=RCU] 3450 bash-1998 [000] ..s2 138.733105: softirq_entry: vec=1 [action=TIMER] 3451 bash-1998 [000] ..s2 138.733106: softirq_exit: vec=1 [action=TIMER] 3452 bash-1998 [000] ..s2 138.733106: softirq_entry: vec=9 [action=RCU] 3453 bash-1998 [000] ..s2 138.733109: softirq_exit: vec=9 [action=RCU] 3454 sshd-1995 [001] d.h1 138.733278: irq_handler_entry: irq=21 name=uhci_hcd:usb4 3455 sshd-1995 [001] d.h1 138.733280: irq_handler_exit: irq=21 ret=unhandled 3456 sshd-1995 [001] d.h1 138.733281: irq_handler_entry: irq=21 name=eth0 3457 sshd-1995 [001] d.h1 138.733283: irq_handler_exit: irq=21 ret=handled 3458 [...] 3459 3460 # cat instances/zoot/trace 3461 # tracer: nop 3462 # 3463 # entries-in-buffer/entries-written: 18996/18996 #P:4 3464 # 3465 # _-----=> irqs-off 3466 # / _----=> need-resched 3467 # | / _---=> hardirq/softirq 3468 # || / _--=> preempt-depth 3469 # ||| / delay 3470 # TASK-PID CPU# |||| TIMESTAMP FUNCTION 3471 # | | | |||| | | 3472 bash-1998 [000] d... 140.733501: sys_write -> 0x2 3473 bash-1998 [000] d... 140.733504: sys_dup2(oldfd: a, newfd: 1) 3474 bash-1998 [000] d... 140.733506: sys_dup2 -> 0x1 3475 bash-1998 [000] d... 140.733508: sys_fcntl(fd: a, cmd: 1, arg: 0) 3476 bash-1998 [000] d... 140.733509: sys_fcntl -> 0x1 3477 bash-1998 [000] d... 140.733510: sys_close(fd: a) 3478 bash-1998 [000] d... 140.733510: sys_close -> 0x0 3479 bash-1998 [000] d... 140.733514: sys_rt_sigprocmask(how: 0, nset: 0, oset: 6e2768, sigsetsize: 8) 3480 bash-1998 [000] d... 140.733515: sys_rt_sigprocmask -> 0x0 3481 bash-1998 [000] d... 140.733516: sys_rt_sigaction(sig: 2, act: 7fff718846f0, oact: 7fff71884650, sigsetsize: 8) 3482 bash-1998 [000] d... 140.733516: sys_rt_sigaction -> 0x0 3483 3484You can see that the trace of the top most trace buffer shows only 3485the function tracing. The foo instance displays wakeups and task 3486switches. 3487 3488To remove the instances, simply delete their directories: 3489:: 3490 3491 # rmdir instances/foo 3492 # rmdir instances/bar 3493 # rmdir instances/zoot 3494 3495Note, if a process has a trace file open in one of the instance 3496directories, the rmdir will fail with EBUSY. 3497 3498 3499Stack trace 3500----------- 3501Since the kernel has a fixed sized stack, it is important not to 3502waste it in functions. A kernel developer must be conscience of 3503what they allocate on the stack. If they add too much, the system 3504can be in danger of a stack overflow, and corruption will occur, 3505usually leading to a system panic. 3506 3507There are some tools that check this, usually with interrupts 3508periodically checking usage. But if you can perform a check 3509at every function call that will become very useful. As ftrace provides 3510a function tracer, it makes it convenient to check the stack size 3511at every function call. This is enabled via the stack tracer. 3512 3513CONFIG_STACK_TRACER enables the ftrace stack tracing functionality. 3514To enable it, write a '1' into /proc/sys/kernel/stack_tracer_enabled. 3515:: 3516 3517 # echo 1 > /proc/sys/kernel/stack_tracer_enabled 3518 3519You can also enable it from the kernel command line to trace 3520the stack size of the kernel during boot up, by adding "stacktrace" 3521to the kernel command line parameter. 3522 3523After running it for a few minutes, the output looks like: 3524:: 3525 3526 # cat stack_max_size 3527 2928 3528 3529 # cat stack_trace 3530 Depth Size Location (18 entries) 3531 ----- ---- -------- 3532 0) 2928 224 update_sd_lb_stats+0xbc/0x4ac 3533 1) 2704 160 find_busiest_group+0x31/0x1f1 3534 2) 2544 256 load_balance+0xd9/0x662 3535 3) 2288 80 idle_balance+0xbb/0x130 3536 4) 2208 128 __schedule+0x26e/0x5b9 3537 5) 2080 16 schedule+0x64/0x66 3538 6) 2064 128 schedule_timeout+0x34/0xe0 3539 7) 1936 112 wait_for_common+0x97/0xf1 3540 8) 1824 16 wait_for_completion+0x1d/0x1f 3541 9) 1808 128 flush_work+0xfe/0x119 3542 10) 1680 16 tty_flush_to_ldisc+0x1e/0x20 3543 11) 1664 48 input_available_p+0x1d/0x5c 3544 12) 1616 48 n_tty_poll+0x6d/0x134 3545 13) 1568 64 tty_poll+0x64/0x7f 3546 14) 1504 880 do_select+0x31e/0x511 3547 15) 624 400 core_sys_select+0x177/0x216 3548 16) 224 96 sys_select+0x91/0xb9 3549 17) 128 128 system_call_fastpath+0x16/0x1b 3550 3551Note, if -mfentry is being used by gcc, functions get traced before 3552they set up the stack frame. This means that leaf level functions 3553are not tested by the stack tracer when -mfentry is used. 3554 3555Currently, -mfentry is used by gcc 4.6.0 and above on x86 only. 3556 3557More 3558---- 3559More details can be found in the source code, in the `kernel/trace/*.c` files. 3560