1# SPDX-License-Identifier: GPL-2.0-only 2# 3# Architectures that offer an FUNCTION_TRACER implementation should 4# select HAVE_FUNCTION_TRACER: 5# 6 7config USER_STACKTRACE_SUPPORT 8 bool 9 10config NOP_TRACER 11 bool 12 13config HAVE_FUNCTION_TRACER 14 bool 15 help 16 See Documentation/trace/ftrace-design.rst 17 18config HAVE_FUNCTION_GRAPH_TRACER 19 bool 20 help 21 See Documentation/trace/ftrace-design.rst 22 23config HAVE_DYNAMIC_FTRACE 24 bool 25 help 26 See Documentation/trace/ftrace-design.rst 27 28config HAVE_DYNAMIC_FTRACE_WITH_REGS 29 bool 30 31config HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS 32 bool 33 34config HAVE_DYNAMIC_FTRACE_WITH_ARGS 35 bool 36 help 37 If this is set, then arguments and stack can be found from 38 the pt_regs passed into the function callback regs parameter 39 by default, even without setting the REGS flag in the ftrace_ops. 40 This allows for use of regs_get_kernel_argument() and 41 kernel_stack_pointer(). 42 43config HAVE_FTRACE_MCOUNT_RECORD 44 bool 45 help 46 See Documentation/trace/ftrace-design.rst 47 48config HAVE_SYSCALL_TRACEPOINTS 49 bool 50 help 51 See Documentation/trace/ftrace-design.rst 52 53config HAVE_FENTRY 54 bool 55 help 56 Arch supports the gcc options -pg with -mfentry 57 58config HAVE_NOP_MCOUNT 59 bool 60 help 61 Arch supports the gcc options -pg with -mrecord-mcount and -nop-mcount 62 63config HAVE_OBJTOOL_MCOUNT 64 bool 65 help 66 Arch supports objtool --mcount 67 68config HAVE_C_RECORDMCOUNT 69 bool 70 help 71 C version of recordmcount available? 72 73config TRACER_MAX_TRACE 74 bool 75 76config TRACE_CLOCK 77 bool 78 79config RING_BUFFER 80 bool 81 select TRACE_CLOCK 82 select IRQ_WORK 83 84config EVENT_TRACING 85 select CONTEXT_SWITCH_TRACER 86 select GLOB 87 bool 88 89config CONTEXT_SWITCH_TRACER 90 bool 91 92config RING_BUFFER_ALLOW_SWAP 93 bool 94 help 95 Allow the use of ring_buffer_swap_cpu. 96 Adds a very slight overhead to tracing when enabled. 97 98config PREEMPTIRQ_TRACEPOINTS 99 bool 100 depends on TRACE_PREEMPT_TOGGLE || TRACE_IRQFLAGS 101 select TRACING 102 default y 103 help 104 Create preempt/irq toggle tracepoints if needed, so that other parts 105 of the kernel can use them to generate or add hooks to them. 106 107# All tracer options should select GENERIC_TRACER. For those options that are 108# enabled by all tracers (context switch and event tracer) they select TRACING. 109# This allows those options to appear when no other tracer is selected. But the 110# options do not appear when something else selects it. We need the two options 111# GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the 112# hiding of the automatic options. 113 114config TRACING 115 bool 116 select RING_BUFFER 117 select STACKTRACE if STACKTRACE_SUPPORT 118 select TRACEPOINTS 119 select NOP_TRACER 120 select BINARY_PRINTF 121 select EVENT_TRACING 122 select TRACE_CLOCK 123 124config GENERIC_TRACER 125 bool 126 select TRACING 127 128# 129# Minimum requirements an architecture has to meet for us to 130# be able to offer generic tracing facilities: 131# 132config TRACING_SUPPORT 133 bool 134 depends on TRACE_IRQFLAGS_SUPPORT 135 depends on STACKTRACE_SUPPORT 136 default y 137 138if TRACING_SUPPORT 139 140menuconfig FTRACE 141 bool "Tracers" 142 default y if DEBUG_KERNEL 143 help 144 Enable the kernel tracing infrastructure. 145 146if FTRACE 147 148config BOOTTIME_TRACING 149 bool "Boot-time Tracing support" 150 depends on TRACING 151 select BOOT_CONFIG 152 help 153 Enable developer to setup ftrace subsystem via supplemental 154 kernel cmdline at boot time for debugging (tracing) driver 155 initialization and boot process. 156 157config FUNCTION_TRACER 158 bool "Kernel Function Tracer" 159 depends on HAVE_FUNCTION_TRACER 160 select KALLSYMS 161 select GENERIC_TRACER 162 select CONTEXT_SWITCH_TRACER 163 select GLOB 164 select TASKS_RCU if PREEMPTION 165 select TASKS_RUDE_RCU 166 help 167 Enable the kernel to trace every kernel function. This is done 168 by using a compiler feature to insert a small, 5-byte No-Operation 169 instruction at the beginning of every kernel function, which NOP 170 sequence is then dynamically patched into a tracer call when 171 tracing is enabled by the administrator. If it's runtime disabled 172 (the bootup default), then the overhead of the instructions is very 173 small and not measurable even in micro-benchmarks. 174 175config FUNCTION_GRAPH_TRACER 176 bool "Kernel Function Graph Tracer" 177 depends on HAVE_FUNCTION_GRAPH_TRACER 178 depends on FUNCTION_TRACER 179 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE 180 default y 181 help 182 Enable the kernel to trace a function at both its return 183 and its entry. 184 Its first purpose is to trace the duration of functions and 185 draw a call graph for each thread with some information like 186 the return value. This is done by setting the current return 187 address on the current task structure into a stack of calls. 188 189config DYNAMIC_FTRACE 190 bool "enable/disable function tracing dynamically" 191 depends on FUNCTION_TRACER 192 depends on HAVE_DYNAMIC_FTRACE 193 default y 194 help 195 This option will modify all the calls to function tracing 196 dynamically (will patch them out of the binary image and 197 replace them with a No-Op instruction) on boot up. During 198 compile time, a table is made of all the locations that ftrace 199 can function trace, and this table is linked into the kernel 200 image. When this is enabled, functions can be individually 201 enabled, and the functions not enabled will not affect 202 performance of the system. 203 204 See the files in /sys/kernel/debug/tracing: 205 available_filter_functions 206 set_ftrace_filter 207 set_ftrace_notrace 208 209 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but 210 otherwise has native performance as long as no tracing is active. 211 212config DYNAMIC_FTRACE_WITH_REGS 213 def_bool y 214 depends on DYNAMIC_FTRACE 215 depends on HAVE_DYNAMIC_FTRACE_WITH_REGS 216 217config DYNAMIC_FTRACE_WITH_DIRECT_CALLS 218 def_bool y 219 depends on DYNAMIC_FTRACE_WITH_REGS 220 depends on HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS 221 222config FUNCTION_PROFILER 223 bool "Kernel function profiler" 224 depends on FUNCTION_TRACER 225 default n 226 help 227 This option enables the kernel function profiler. A file is created 228 in debugfs called function_profile_enabled which defaults to zero. 229 When a 1 is echoed into this file profiling begins, and when a 230 zero is entered, profiling stops. A "functions" file is created in 231 the trace_stat directory; this file shows the list of functions that 232 have been hit and their counters. 233 234 If in doubt, say N. 235 236config STACK_TRACER 237 bool "Trace max stack" 238 depends on HAVE_FUNCTION_TRACER 239 select FUNCTION_TRACER 240 select STACKTRACE 241 select KALLSYMS 242 help 243 This special tracer records the maximum stack footprint of the 244 kernel and displays it in /sys/kernel/debug/tracing/stack_trace. 245 246 This tracer works by hooking into every function call that the 247 kernel executes, and keeping a maximum stack depth value and 248 stack-trace saved. If this is configured with DYNAMIC_FTRACE 249 then it will not have any overhead while the stack tracer 250 is disabled. 251 252 To enable the stack tracer on bootup, pass in 'stacktrace' 253 on the kernel command line. 254 255 The stack tracer can also be enabled or disabled via the 256 sysctl kernel.stack_tracer_enabled 257 258 Say N if unsure. 259 260config TRACE_PREEMPT_TOGGLE 261 bool 262 help 263 Enables hooks which will be called when preemption is first disabled, 264 and last enabled. 265 266config IRQSOFF_TRACER 267 bool "Interrupts-off Latency Tracer" 268 default n 269 depends on TRACE_IRQFLAGS_SUPPORT 270 select TRACE_IRQFLAGS 271 select GENERIC_TRACER 272 select TRACER_MAX_TRACE 273 select RING_BUFFER_ALLOW_SWAP 274 select TRACER_SNAPSHOT 275 select TRACER_SNAPSHOT_PER_CPU_SWAP 276 help 277 This option measures the time spent in irqs-off critical 278 sections, with microsecond accuracy. 279 280 The default measurement method is a maximum search, which is 281 disabled by default and can be runtime (re-)started 282 via: 283 284 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency 285 286 (Note that kernel size and overhead increase with this option 287 enabled. This option and the preempt-off timing option can be 288 used together or separately.) 289 290config PREEMPT_TRACER 291 bool "Preemption-off Latency Tracer" 292 default n 293 depends on PREEMPTION 294 select GENERIC_TRACER 295 select TRACER_MAX_TRACE 296 select RING_BUFFER_ALLOW_SWAP 297 select TRACER_SNAPSHOT 298 select TRACER_SNAPSHOT_PER_CPU_SWAP 299 select TRACE_PREEMPT_TOGGLE 300 help 301 This option measures the time spent in preemption-off critical 302 sections, with microsecond accuracy. 303 304 The default measurement method is a maximum search, which is 305 disabled by default and can be runtime (re-)started 306 via: 307 308 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency 309 310 (Note that kernel size and overhead increase with this option 311 enabled. This option and the irqs-off timing option can be 312 used together or separately.) 313 314config SCHED_TRACER 315 bool "Scheduling Latency Tracer" 316 select GENERIC_TRACER 317 select CONTEXT_SWITCH_TRACER 318 select TRACER_MAX_TRACE 319 select TRACER_SNAPSHOT 320 help 321 This tracer tracks the latency of the highest priority task 322 to be scheduled in, starting from the point it has woken up. 323 324config HWLAT_TRACER 325 bool "Tracer to detect hardware latencies (like SMIs)" 326 select GENERIC_TRACER 327 help 328 This tracer, when enabled will create one or more kernel threads, 329 depending on what the cpumask file is set to, which each thread 330 spinning in a loop looking for interruptions caused by 331 something other than the kernel. For example, if a 332 System Management Interrupt (SMI) takes a noticeable amount of 333 time, this tracer will detect it. This is useful for testing 334 if a system is reliable for Real Time tasks. 335 336 Some files are created in the tracing directory when this 337 is enabled: 338 339 hwlat_detector/width - time in usecs for how long to spin for 340 hwlat_detector/window - time in usecs between the start of each 341 iteration 342 343 A kernel thread is created that will spin with interrupts disabled 344 for "width" microseconds in every "window" cycle. It will not spin 345 for "window - width" microseconds, where the system can 346 continue to operate. 347 348 The output will appear in the trace and trace_pipe files. 349 350 When the tracer is not running, it has no affect on the system, 351 but when it is running, it can cause the system to be 352 periodically non responsive. Do not run this tracer on a 353 production system. 354 355 To enable this tracer, echo in "hwlat" into the current_tracer 356 file. Every time a latency is greater than tracing_thresh, it will 357 be recorded into the ring buffer. 358 359config OSNOISE_TRACER 360 bool "OS Noise tracer" 361 select GENERIC_TRACER 362 help 363 In the context of high-performance computing (HPC), the Operating 364 System Noise (osnoise) refers to the interference experienced by an 365 application due to activities inside the operating system. In the 366 context of Linux, NMIs, IRQs, SoftIRQs, and any other system thread 367 can cause noise to the system. Moreover, hardware-related jobs can 368 also cause noise, for example, via SMIs. 369 370 The osnoise tracer leverages the hwlat_detector by running a similar 371 loop with preemption, SoftIRQs and IRQs enabled, thus allowing all 372 the sources of osnoise during its execution. The osnoise tracer takes 373 note of the entry and exit point of any source of interferences, 374 increasing a per-cpu interference counter. It saves an interference 375 counter for each source of interference. The interference counter for 376 NMI, IRQs, SoftIRQs, and threads is increased anytime the tool 377 observes these interferences' entry events. When a noise happens 378 without any interference from the operating system level, the 379 hardware noise counter increases, pointing to a hardware-related 380 noise. In this way, osnoise can account for any source of 381 interference. At the end of the period, the osnoise tracer prints 382 the sum of all noise, the max single noise, the percentage of CPU 383 available for the thread, and the counters for the noise sources. 384 385 In addition to the tracer, a set of tracepoints were added to 386 facilitate the identification of the osnoise source. 387 388 The output will appear in the trace and trace_pipe files. 389 390 To enable this tracer, echo in "osnoise" into the current_tracer 391 file. 392 393config TIMERLAT_TRACER 394 bool "Timerlat tracer" 395 select OSNOISE_TRACER 396 select GENERIC_TRACER 397 help 398 The timerlat tracer aims to help the preemptive kernel developers 399 to find sources of wakeup latencies of real-time threads. 400 401 The tracer creates a per-cpu kernel thread with real-time priority. 402 The tracer thread sets a periodic timer to wakeup itself, and goes 403 to sleep waiting for the timer to fire. At the wakeup, the thread 404 then computes a wakeup latency value as the difference between 405 the current time and the absolute time that the timer was set 406 to expire. 407 408 The tracer prints two lines at every activation. The first is the 409 timer latency observed at the hardirq context before the 410 activation of the thread. The second is the timer latency observed 411 by the thread, which is the same level that cyclictest reports. The 412 ACTIVATION ID field serves to relate the irq execution to its 413 respective thread execution. 414 415 The tracer is build on top of osnoise tracer, and the osnoise: 416 events can be used to trace the source of interference from NMI, 417 IRQs and other threads. It also enables the capture of the 418 stacktrace at the IRQ context, which helps to identify the code 419 path that can cause thread delay. 420 421config MMIOTRACE 422 bool "Memory mapped IO tracing" 423 depends on HAVE_MMIOTRACE_SUPPORT && PCI 424 select GENERIC_TRACER 425 help 426 Mmiotrace traces Memory Mapped I/O access and is meant for 427 debugging and reverse engineering. It is called from the ioremap 428 implementation and works via page faults. Tracing is disabled by 429 default and can be enabled at run-time. 430 431 See Documentation/trace/mmiotrace.rst. 432 If you are not helping to develop drivers, say N. 433 434config ENABLE_DEFAULT_TRACERS 435 bool "Trace process context switches and events" 436 depends on !GENERIC_TRACER 437 select TRACING 438 help 439 This tracer hooks to various trace points in the kernel, 440 allowing the user to pick and choose which trace point they 441 want to trace. It also includes the sched_switch tracer plugin. 442 443config FTRACE_SYSCALLS 444 bool "Trace syscalls" 445 depends on HAVE_SYSCALL_TRACEPOINTS 446 select GENERIC_TRACER 447 select KALLSYMS 448 help 449 Basic tracer to catch the syscall entry and exit events. 450 451config TRACER_SNAPSHOT 452 bool "Create a snapshot trace buffer" 453 select TRACER_MAX_TRACE 454 help 455 Allow tracing users to take snapshot of the current buffer using the 456 ftrace interface, e.g.: 457 458 echo 1 > /sys/kernel/debug/tracing/snapshot 459 cat snapshot 460 461config TRACER_SNAPSHOT_PER_CPU_SWAP 462 bool "Allow snapshot to swap per CPU" 463 depends on TRACER_SNAPSHOT 464 select RING_BUFFER_ALLOW_SWAP 465 help 466 Allow doing a snapshot of a single CPU buffer instead of a 467 full swap (all buffers). If this is set, then the following is 468 allowed: 469 470 echo 1 > /sys/kernel/debug/tracing/per_cpu/cpu2/snapshot 471 472 After which, only the tracing buffer for CPU 2 was swapped with 473 the main tracing buffer, and the other CPU buffers remain the same. 474 475 When this is enabled, this adds a little more overhead to the 476 trace recording, as it needs to add some checks to synchronize 477 recording with swaps. But this does not affect the performance 478 of the overall system. This is enabled by default when the preempt 479 or irq latency tracers are enabled, as those need to swap as well 480 and already adds the overhead (plus a lot more). 481 482config TRACE_BRANCH_PROFILING 483 bool 484 select GENERIC_TRACER 485 486choice 487 prompt "Branch Profiling" 488 default BRANCH_PROFILE_NONE 489 help 490 The branch profiling is a software profiler. It will add hooks 491 into the C conditionals to test which path a branch takes. 492 493 The likely/unlikely profiler only looks at the conditions that 494 are annotated with a likely or unlikely macro. 495 496 The "all branch" profiler will profile every if-statement in the 497 kernel. This profiler will also enable the likely/unlikely 498 profiler. 499 500 Either of the above profilers adds a bit of overhead to the system. 501 If unsure, choose "No branch profiling". 502 503config BRANCH_PROFILE_NONE 504 bool "No branch profiling" 505 help 506 No branch profiling. Branch profiling adds a bit of overhead. 507 Only enable it if you want to analyse the branching behavior. 508 Otherwise keep it disabled. 509 510config PROFILE_ANNOTATED_BRANCHES 511 bool "Trace likely/unlikely profiler" 512 select TRACE_BRANCH_PROFILING 513 help 514 This tracer profiles all likely and unlikely macros 515 in the kernel. It will display the results in: 516 517 /sys/kernel/debug/tracing/trace_stat/branch_annotated 518 519 Note: this will add a significant overhead; only turn this 520 on if you need to profile the system's use of these macros. 521 522config PROFILE_ALL_BRANCHES 523 bool "Profile all if conditionals" if !FORTIFY_SOURCE 524 select TRACE_BRANCH_PROFILING 525 help 526 This tracer profiles all branch conditions. Every if () 527 taken in the kernel is recorded whether it hit or miss. 528 The results will be displayed in: 529 530 /sys/kernel/debug/tracing/trace_stat/branch_all 531 532 This option also enables the likely/unlikely profiler. 533 534 This configuration, when enabled, will impose a great overhead 535 on the system. This should only be enabled when the system 536 is to be analyzed in much detail. 537endchoice 538 539config TRACING_BRANCHES 540 bool 541 help 542 Selected by tracers that will trace the likely and unlikely 543 conditions. This prevents the tracers themselves from being 544 profiled. Profiling the tracing infrastructure can only happen 545 when the likelys and unlikelys are not being traced. 546 547config BRANCH_TRACER 548 bool "Trace likely/unlikely instances" 549 depends on TRACE_BRANCH_PROFILING 550 select TRACING_BRANCHES 551 help 552 This traces the events of likely and unlikely condition 553 calls in the kernel. The difference between this and the 554 "Trace likely/unlikely profiler" is that this is not a 555 histogram of the callers, but actually places the calling 556 events into a running trace buffer to see when and where the 557 events happened, as well as their results. 558 559 Say N if unsure. 560 561config BLK_DEV_IO_TRACE 562 bool "Support for tracing block IO actions" 563 depends on SYSFS 564 depends on BLOCK 565 select RELAY 566 select DEBUG_FS 567 select TRACEPOINTS 568 select GENERIC_TRACER 569 select STACKTRACE 570 help 571 Say Y here if you want to be able to trace the block layer actions 572 on a given queue. Tracing allows you to see any traffic happening 573 on a block device queue. For more information (and the userspace 574 support tools needed), fetch the blktrace tools from: 575 576 git://git.kernel.dk/blktrace.git 577 578 Tracing also is possible using the ftrace interface, e.g.: 579 580 echo 1 > /sys/block/sda/sda1/trace/enable 581 echo blk > /sys/kernel/debug/tracing/current_tracer 582 cat /sys/kernel/debug/tracing/trace_pipe 583 584 If unsure, say N. 585 586config KPROBE_EVENTS 587 depends on KPROBES 588 depends on HAVE_REGS_AND_STACK_ACCESS_API 589 bool "Enable kprobes-based dynamic events" 590 select TRACING 591 select PROBE_EVENTS 592 select DYNAMIC_EVENTS 593 default y 594 help 595 This allows the user to add tracing events (similar to tracepoints) 596 on the fly via the ftrace interface. See 597 Documentation/trace/kprobetrace.rst for more details. 598 599 Those events can be inserted wherever kprobes can probe, and record 600 various register and memory values. 601 602 This option is also required by perf-probe subcommand of perf tools. 603 If you want to use perf tools, this option is strongly recommended. 604 605config KPROBE_EVENTS_ON_NOTRACE 606 bool "Do NOT protect notrace function from kprobe events" 607 depends on KPROBE_EVENTS 608 depends on DYNAMIC_FTRACE 609 default n 610 help 611 This is only for the developers who want to debug ftrace itself 612 using kprobe events. 613 614 If kprobes can use ftrace instead of breakpoint, ftrace related 615 functions are protected from kprobe-events to prevent an infinite 616 recursion or any unexpected execution path which leads to a kernel 617 crash. 618 619 This option disables such protection and allows you to put kprobe 620 events on ftrace functions for debugging ftrace by itself. 621 Note that this might let you shoot yourself in the foot. 622 623 If unsure, say N. 624 625config UPROBE_EVENTS 626 bool "Enable uprobes-based dynamic events" 627 depends on ARCH_SUPPORTS_UPROBES 628 depends on MMU 629 depends on PERF_EVENTS 630 select UPROBES 631 select PROBE_EVENTS 632 select DYNAMIC_EVENTS 633 select TRACING 634 default y 635 help 636 This allows the user to add tracing events on top of userspace 637 dynamic events (similar to tracepoints) on the fly via the trace 638 events interface. Those events can be inserted wherever uprobes 639 can probe, and record various registers. 640 This option is required if you plan to use perf-probe subcommand 641 of perf tools on user space applications. 642 643config BPF_EVENTS 644 depends on BPF_SYSCALL 645 depends on (KPROBE_EVENTS || UPROBE_EVENTS) && PERF_EVENTS 646 bool 647 default y 648 help 649 This allows the user to attach BPF programs to kprobe, uprobe, and 650 tracepoint events. 651 652config DYNAMIC_EVENTS 653 def_bool n 654 655config PROBE_EVENTS 656 def_bool n 657 658config BPF_KPROBE_OVERRIDE 659 bool "Enable BPF programs to override a kprobed function" 660 depends on BPF_EVENTS 661 depends on FUNCTION_ERROR_INJECTION 662 default n 663 help 664 Allows BPF to override the execution of a probed function and 665 set a different return value. This is used for error injection. 666 667config FTRACE_MCOUNT_RECORD 668 def_bool y 669 depends on DYNAMIC_FTRACE 670 depends on HAVE_FTRACE_MCOUNT_RECORD 671 672config FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 673 bool 674 depends on FTRACE_MCOUNT_RECORD 675 676config FTRACE_MCOUNT_USE_CC 677 def_bool y 678 depends on $(cc-option,-mrecord-mcount) 679 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 680 depends on FTRACE_MCOUNT_RECORD 681 682config FTRACE_MCOUNT_USE_OBJTOOL 683 def_bool y 684 depends on HAVE_OBJTOOL_MCOUNT 685 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 686 depends on !FTRACE_MCOUNT_USE_CC 687 depends on FTRACE_MCOUNT_RECORD 688 689config FTRACE_MCOUNT_USE_RECORDMCOUNT 690 def_bool y 691 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 692 depends on !FTRACE_MCOUNT_USE_CC 693 depends on !FTRACE_MCOUNT_USE_OBJTOOL 694 depends on FTRACE_MCOUNT_RECORD 695 696config TRACING_MAP 697 bool 698 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG 699 help 700 tracing_map is a special-purpose lock-free map for tracing, 701 separated out as a stand-alone facility in order to allow it 702 to be shared between multiple tracers. It isn't meant to be 703 generally used outside of that context, and is normally 704 selected by tracers that use it. 705 706config SYNTH_EVENTS 707 bool "Synthetic trace events" 708 select TRACING 709 select DYNAMIC_EVENTS 710 default n 711 help 712 Synthetic events are user-defined trace events that can be 713 used to combine data from other trace events or in fact any 714 data source. Synthetic events can be generated indirectly 715 via the trace() action of histogram triggers or directly 716 by way of an in-kernel API. 717 718 See Documentation/trace/events.rst or 719 Documentation/trace/histogram.rst for details and examples. 720 721 If in doubt, say N. 722 723config HIST_TRIGGERS 724 bool "Histogram triggers" 725 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG 726 select TRACING_MAP 727 select TRACING 728 select DYNAMIC_EVENTS 729 select SYNTH_EVENTS 730 default n 731 help 732 Hist triggers allow one or more arbitrary trace event fields 733 to be aggregated into hash tables and dumped to stdout by 734 reading a debugfs/tracefs file. They're useful for 735 gathering quick and dirty (though precise) summaries of 736 event activity as an initial guide for further investigation 737 using more advanced tools. 738 739 Inter-event tracing of quantities such as latencies is also 740 supported using hist triggers under this option. 741 742 See Documentation/trace/histogram.rst. 743 If in doubt, say N. 744 745config TRACE_EVENT_INJECT 746 bool "Trace event injection" 747 depends on TRACING 748 help 749 Allow user-space to inject a specific trace event into the ring 750 buffer. This is mainly used for testing purpose. 751 752 If unsure, say N. 753 754config TRACEPOINT_BENCHMARK 755 bool "Add tracepoint that benchmarks tracepoints" 756 help 757 This option creates the tracepoint "benchmark:benchmark_event". 758 When the tracepoint is enabled, it kicks off a kernel thread that 759 goes into an infinite loop (calling cond_resched() to let other tasks 760 run), and calls the tracepoint. Each iteration will record the time 761 it took to write to the tracepoint and the next iteration that 762 data will be passed to the tracepoint itself. That is, the tracepoint 763 will report the time it took to do the previous tracepoint. 764 The string written to the tracepoint is a static string of 128 bytes 765 to keep the time the same. The initial string is simply a write of 766 "START". The second string records the cold cache time of the first 767 write which is not added to the rest of the calculations. 768 769 As it is a tight loop, it benchmarks as hot cache. That's fine because 770 we care most about hot paths that are probably in cache already. 771 772 An example of the output: 773 774 START 775 first=3672 [COLD CACHED] 776 last=632 first=3672 max=632 min=632 avg=316 std=446 std^2=199712 777 last=278 first=3672 max=632 min=278 avg=303 std=316 std^2=100337 778 last=277 first=3672 max=632 min=277 avg=296 std=258 std^2=67064 779 last=273 first=3672 max=632 min=273 avg=292 std=224 std^2=50411 780 last=273 first=3672 max=632 min=273 avg=288 std=200 std^2=40389 781 last=281 first=3672 max=632 min=273 avg=287 std=183 std^2=33666 782 783 784config RING_BUFFER_BENCHMARK 785 tristate "Ring buffer benchmark stress tester" 786 depends on RING_BUFFER 787 help 788 This option creates a test to stress the ring buffer and benchmark it. 789 It creates its own ring buffer such that it will not interfere with 790 any other users of the ring buffer (such as ftrace). It then creates 791 a producer and consumer that will run for 10 seconds and sleep for 792 10 seconds. Each interval it will print out the number of events 793 it recorded and give a rough estimate of how long each iteration took. 794 795 It does not disable interrupts or raise its priority, so it may be 796 affected by processes that are running. 797 798 If unsure, say N. 799 800config TRACE_EVAL_MAP_FILE 801 bool "Show eval mappings for trace events" 802 depends on TRACING 803 help 804 The "print fmt" of the trace events will show the enum/sizeof names 805 instead of their values. This can cause problems for user space tools 806 that use this string to parse the raw data as user space does not know 807 how to convert the string to its value. 808 809 To fix this, there's a special macro in the kernel that can be used 810 to convert an enum/sizeof into its value. If this macro is used, then 811 the print fmt strings will be converted to their values. 812 813 If something does not get converted properly, this option can be 814 used to show what enums/sizeof the kernel tried to convert. 815 816 This option is for debugging the conversions. A file is created 817 in the tracing directory called "eval_map" that will show the 818 names matched with their values and what trace event system they 819 belong too. 820 821 Normally, the mapping of the strings to values will be freed after 822 boot up or module load. With this option, they will not be freed, as 823 they are needed for the "eval_map" file. Enabling this option will 824 increase the memory footprint of the running kernel. 825 826 If unsure, say N. 827 828config FTRACE_RECORD_RECURSION 829 bool "Record functions that recurse in function tracing" 830 depends on FUNCTION_TRACER 831 help 832 All callbacks that attach to the function tracing have some sort 833 of protection against recursion. Even though the protection exists, 834 it adds overhead. This option will create a file in the tracefs 835 file system called "recursed_functions" that will list the functions 836 that triggered a recursion. 837 838 This will add more overhead to cases that have recursion. 839 840 If unsure, say N 841 842config FTRACE_RECORD_RECURSION_SIZE 843 int "Max number of recursed functions to record" 844 default 128 845 depends on FTRACE_RECORD_RECURSION 846 help 847 This defines the limit of number of functions that can be 848 listed in the "recursed_functions" file, that lists all 849 the functions that caused a recursion to happen. 850 This file can be reset, but the limit can not change in 851 size at runtime. 852 853config RING_BUFFER_RECORD_RECURSION 854 bool "Record functions that recurse in the ring buffer" 855 depends on FTRACE_RECORD_RECURSION 856 # default y, because it is coupled with FTRACE_RECORD_RECURSION 857 default y 858 help 859 The ring buffer has its own internal recursion. Although when 860 recursion happens it wont cause harm because of the protection, 861 but it does cause an unwanted overhead. Enabling this option will 862 place where recursion was detected into the ftrace "recursed_functions" 863 file. 864 865 This will add more overhead to cases that have recursion. 866 867config GCOV_PROFILE_FTRACE 868 bool "Enable GCOV profiling on ftrace subsystem" 869 depends on GCOV_KERNEL 870 help 871 Enable GCOV profiling on ftrace subsystem for checking 872 which functions/lines are tested. 873 874 If unsure, say N. 875 876 Note that on a kernel compiled with this config, ftrace will 877 run significantly slower. 878 879config FTRACE_SELFTEST 880 bool 881 882config FTRACE_STARTUP_TEST 883 bool "Perform a startup test on ftrace" 884 depends on GENERIC_TRACER 885 select FTRACE_SELFTEST 886 help 887 This option performs a series of startup tests on ftrace. On bootup 888 a series of tests are made to verify that the tracer is 889 functioning properly. It will do tests on all the configured 890 tracers of ftrace. 891 892config EVENT_TRACE_STARTUP_TEST 893 bool "Run selftest on trace events" 894 depends on FTRACE_STARTUP_TEST 895 default y 896 help 897 This option performs a test on all trace events in the system. 898 It basically just enables each event and runs some code that 899 will trigger events (not necessarily the event it enables) 900 This may take some time run as there are a lot of events. 901 902config EVENT_TRACE_TEST_SYSCALLS 903 bool "Run selftest on syscall events" 904 depends on EVENT_TRACE_STARTUP_TEST 905 help 906 This option will also enable testing every syscall event. 907 It only enables the event and disables it and runs various loads 908 with the event enabled. This adds a bit more time for kernel boot 909 up since it runs this on every system call defined. 910 911 TBD - enable a way to actually call the syscalls as we test their 912 events 913 914config RING_BUFFER_STARTUP_TEST 915 bool "Ring buffer startup self test" 916 depends on RING_BUFFER 917 help 918 Run a simple self test on the ring buffer on boot up. Late in the 919 kernel boot sequence, the test will start that kicks off 920 a thread per cpu. Each thread will write various size events 921 into the ring buffer. Another thread is created to send IPIs 922 to each of the threads, where the IPI handler will also write 923 to the ring buffer, to test/stress the nesting ability. 924 If any anomalies are discovered, a warning will be displayed 925 and all ring buffers will be disabled. 926 927 The test runs for 10 seconds. This will slow your boot time 928 by at least 10 more seconds. 929 930 At the end of the test, statics and more checks are done. 931 It will output the stats of each per cpu buffer. What 932 was written, the sizes, what was read, what was lost, and 933 other similar details. 934 935 If unsure, say N 936 937config RING_BUFFER_VALIDATE_TIME_DELTAS 938 bool "Verify ring buffer time stamp deltas" 939 depends on RING_BUFFER 940 help 941 This will audit the time stamps on the ring buffer sub 942 buffer to make sure that all the time deltas for the 943 events on a sub buffer matches the current time stamp. 944 This audit is performed for every event that is not 945 interrupted, or interrupting another event. A check 946 is also made when traversing sub buffers to make sure 947 that all the deltas on the previous sub buffer do not 948 add up to be greater than the current time stamp. 949 950 NOTE: This adds significant overhead to recording of events, 951 and should only be used to test the logic of the ring buffer. 952 Do not use it on production systems. 953 954 Only say Y if you understand what this does, and you 955 still want it enabled. Otherwise say N 956 957config MMIOTRACE_TEST 958 tristate "Test module for mmiotrace" 959 depends on MMIOTRACE && m 960 help 961 This is a dumb module for testing mmiotrace. It is very dangerous 962 as it will write garbage to IO memory starting at a given address. 963 However, it should be safe to use on e.g. unused portion of VRAM. 964 965 Say N, unless you absolutely know what you are doing. 966 967config PREEMPTIRQ_DELAY_TEST 968 tristate "Test module to create a preempt / IRQ disable delay thread to test latency tracers" 969 depends on m 970 help 971 Select this option to build a test module that can help test latency 972 tracers by executing a preempt or irq disable section with a user 973 configurable delay. The module busy waits for the duration of the 974 critical section. 975 976 For example, the following invocation generates a burst of three 977 irq-disabled critical sections for 500us: 978 modprobe preemptirq_delay_test test_mode=irq delay=500 burst_size=3 979 980 What's more, if you want to attach the test on the cpu which the latency 981 tracer is running on, specify cpu_affinity=cpu_num at the end of the 982 command. 983 984 If unsure, say N 985 986config SYNTH_EVENT_GEN_TEST 987 tristate "Test module for in-kernel synthetic event generation" 988 depends on SYNTH_EVENTS 989 help 990 This option creates a test module to check the base 991 functionality of in-kernel synthetic event definition and 992 generation. 993 994 To test, insert the module, and then check the trace buffer 995 for the generated sample events. 996 997 If unsure, say N. 998 999config KPROBE_EVENT_GEN_TEST 1000 tristate "Test module for in-kernel kprobe event generation" 1001 depends on KPROBE_EVENTS 1002 help 1003 This option creates a test module to check the base 1004 functionality of in-kernel kprobe event definition. 1005 1006 To test, insert the module, and then check the trace buffer 1007 for the generated kprobe events. 1008 1009 If unsure, say N. 1010 1011config HIST_TRIGGERS_DEBUG 1012 bool "Hist trigger debug support" 1013 depends on HIST_TRIGGERS 1014 help 1015 Add "hist_debug" file for each event, which when read will 1016 dump out a bunch of internal details about the hist triggers 1017 defined on that event. 1018 1019 The hist_debug file serves a couple of purposes: 1020 1021 - Helps developers verify that nothing is broken. 1022 1023 - Provides educational information to support the details 1024 of the hist trigger internals as described by 1025 Documentation/trace/histogram-design.rst. 1026 1027 The hist_debug output only covers the data structures 1028 related to the histogram definitions themselves and doesn't 1029 display the internals of map buckets or variable values of 1030 running histograms. 1031 1032 If unsure, say N. 1033 1034endif # FTRACE 1035 1036endif # TRACING_SUPPORT 1037 1038