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 DYNAMIC_FTRACE_WITH_ARGS 223 def_bool y 224 depends on DYNAMIC_FTRACE 225 depends on HAVE_DYNAMIC_FTRACE_WITH_ARGS 226 227config FUNCTION_PROFILER 228 bool "Kernel function profiler" 229 depends on FUNCTION_TRACER 230 default n 231 help 232 This option enables the kernel function profiler. A file is created 233 in debugfs called function_profile_enabled which defaults to zero. 234 When a 1 is echoed into this file profiling begins, and when a 235 zero is entered, profiling stops. A "functions" file is created in 236 the trace_stat directory; this file shows the list of functions that 237 have been hit and their counters. 238 239 If in doubt, say N. 240 241config STACK_TRACER 242 bool "Trace max stack" 243 depends on HAVE_FUNCTION_TRACER 244 select FUNCTION_TRACER 245 select STACKTRACE 246 select KALLSYMS 247 help 248 This special tracer records the maximum stack footprint of the 249 kernel and displays it in /sys/kernel/debug/tracing/stack_trace. 250 251 This tracer works by hooking into every function call that the 252 kernel executes, and keeping a maximum stack depth value and 253 stack-trace saved. If this is configured with DYNAMIC_FTRACE 254 then it will not have any overhead while the stack tracer 255 is disabled. 256 257 To enable the stack tracer on bootup, pass in 'stacktrace' 258 on the kernel command line. 259 260 The stack tracer can also be enabled or disabled via the 261 sysctl kernel.stack_tracer_enabled 262 263 Say N if unsure. 264 265config TRACE_PREEMPT_TOGGLE 266 bool 267 help 268 Enables hooks which will be called when preemption is first disabled, 269 and last enabled. 270 271config IRQSOFF_TRACER 272 bool "Interrupts-off Latency Tracer" 273 default n 274 depends on TRACE_IRQFLAGS_SUPPORT 275 select TRACE_IRQFLAGS 276 select GENERIC_TRACER 277 select TRACER_MAX_TRACE 278 select RING_BUFFER_ALLOW_SWAP 279 select TRACER_SNAPSHOT 280 select TRACER_SNAPSHOT_PER_CPU_SWAP 281 help 282 This option measures the time spent in irqs-off critical 283 sections, with microsecond accuracy. 284 285 The default measurement method is a maximum search, which is 286 disabled by default and can be runtime (re-)started 287 via: 288 289 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency 290 291 (Note that kernel size and overhead increase with this option 292 enabled. This option and the preempt-off timing option can be 293 used together or separately.) 294 295config PREEMPT_TRACER 296 bool "Preemption-off Latency Tracer" 297 default n 298 depends on PREEMPTION 299 select GENERIC_TRACER 300 select TRACER_MAX_TRACE 301 select RING_BUFFER_ALLOW_SWAP 302 select TRACER_SNAPSHOT 303 select TRACER_SNAPSHOT_PER_CPU_SWAP 304 select TRACE_PREEMPT_TOGGLE 305 help 306 This option measures the time spent in preemption-off critical 307 sections, with microsecond accuracy. 308 309 The default measurement method is a maximum search, which is 310 disabled by default and can be runtime (re-)started 311 via: 312 313 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency 314 315 (Note that kernel size and overhead increase with this option 316 enabled. This option and the irqs-off timing option can be 317 used together or separately.) 318 319config SCHED_TRACER 320 bool "Scheduling Latency Tracer" 321 select GENERIC_TRACER 322 select CONTEXT_SWITCH_TRACER 323 select TRACER_MAX_TRACE 324 select TRACER_SNAPSHOT 325 help 326 This tracer tracks the latency of the highest priority task 327 to be scheduled in, starting from the point it has woken up. 328 329config HWLAT_TRACER 330 bool "Tracer to detect hardware latencies (like SMIs)" 331 select GENERIC_TRACER 332 help 333 This tracer, when enabled will create one or more kernel threads, 334 depending on what the cpumask file is set to, which each thread 335 spinning in a loop looking for interruptions caused by 336 something other than the kernel. For example, if a 337 System Management Interrupt (SMI) takes a noticeable amount of 338 time, this tracer will detect it. This is useful for testing 339 if a system is reliable for Real Time tasks. 340 341 Some files are created in the tracing directory when this 342 is enabled: 343 344 hwlat_detector/width - time in usecs for how long to spin for 345 hwlat_detector/window - time in usecs between the start of each 346 iteration 347 348 A kernel thread is created that will spin with interrupts disabled 349 for "width" microseconds in every "window" cycle. It will not spin 350 for "window - width" microseconds, where the system can 351 continue to operate. 352 353 The output will appear in the trace and trace_pipe files. 354 355 When the tracer is not running, it has no affect on the system, 356 but when it is running, it can cause the system to be 357 periodically non responsive. Do not run this tracer on a 358 production system. 359 360 To enable this tracer, echo in "hwlat" into the current_tracer 361 file. Every time a latency is greater than tracing_thresh, it will 362 be recorded into the ring buffer. 363 364config OSNOISE_TRACER 365 bool "OS Noise tracer" 366 select GENERIC_TRACER 367 help 368 In the context of high-performance computing (HPC), the Operating 369 System Noise (osnoise) refers to the interference experienced by an 370 application due to activities inside the operating system. In the 371 context of Linux, NMIs, IRQs, SoftIRQs, and any other system thread 372 can cause noise to the system. Moreover, hardware-related jobs can 373 also cause noise, for example, via SMIs. 374 375 The osnoise tracer leverages the hwlat_detector by running a similar 376 loop with preemption, SoftIRQs and IRQs enabled, thus allowing all 377 the sources of osnoise during its execution. The osnoise tracer takes 378 note of the entry and exit point of any source of interferences, 379 increasing a per-cpu interference counter. It saves an interference 380 counter for each source of interference. The interference counter for 381 NMI, IRQs, SoftIRQs, and threads is increased anytime the tool 382 observes these interferences' entry events. When a noise happens 383 without any interference from the operating system level, the 384 hardware noise counter increases, pointing to a hardware-related 385 noise. In this way, osnoise can account for any source of 386 interference. At the end of the period, the osnoise tracer prints 387 the sum of all noise, the max single noise, the percentage of CPU 388 available for the thread, and the counters for the noise sources. 389 390 In addition to the tracer, a set of tracepoints were added to 391 facilitate the identification of the osnoise source. 392 393 The output will appear in the trace and trace_pipe files. 394 395 To enable this tracer, echo in "osnoise" into the current_tracer 396 file. 397 398config TIMERLAT_TRACER 399 bool "Timerlat tracer" 400 select OSNOISE_TRACER 401 select GENERIC_TRACER 402 help 403 The timerlat tracer aims to help the preemptive kernel developers 404 to find sources of wakeup latencies of real-time threads. 405 406 The tracer creates a per-cpu kernel thread with real-time priority. 407 The tracer thread sets a periodic timer to wakeup itself, and goes 408 to sleep waiting for the timer to fire. At the wakeup, the thread 409 then computes a wakeup latency value as the difference between 410 the current time and the absolute time that the timer was set 411 to expire. 412 413 The tracer prints two lines at every activation. The first is the 414 timer latency observed at the hardirq context before the 415 activation of the thread. The second is the timer latency observed 416 by the thread, which is the same level that cyclictest reports. The 417 ACTIVATION ID field serves to relate the irq execution to its 418 respective thread execution. 419 420 The tracer is build on top of osnoise tracer, and the osnoise: 421 events can be used to trace the source of interference from NMI, 422 IRQs and other threads. It also enables the capture of the 423 stacktrace at the IRQ context, which helps to identify the code 424 path that can cause thread delay. 425 426config MMIOTRACE 427 bool "Memory mapped IO tracing" 428 depends on HAVE_MMIOTRACE_SUPPORT && PCI 429 select GENERIC_TRACER 430 help 431 Mmiotrace traces Memory Mapped I/O access and is meant for 432 debugging and reverse engineering. It is called from the ioremap 433 implementation and works via page faults. Tracing is disabled by 434 default and can be enabled at run-time. 435 436 See Documentation/trace/mmiotrace.rst. 437 If you are not helping to develop drivers, say N. 438 439config ENABLE_DEFAULT_TRACERS 440 bool "Trace process context switches and events" 441 depends on !GENERIC_TRACER 442 select TRACING 443 help 444 This tracer hooks to various trace points in the kernel, 445 allowing the user to pick and choose which trace point they 446 want to trace. It also includes the sched_switch tracer plugin. 447 448config FTRACE_SYSCALLS 449 bool "Trace syscalls" 450 depends on HAVE_SYSCALL_TRACEPOINTS 451 select GENERIC_TRACER 452 select KALLSYMS 453 help 454 Basic tracer to catch the syscall entry and exit events. 455 456config TRACER_SNAPSHOT 457 bool "Create a snapshot trace buffer" 458 select TRACER_MAX_TRACE 459 help 460 Allow tracing users to take snapshot of the current buffer using the 461 ftrace interface, e.g.: 462 463 echo 1 > /sys/kernel/debug/tracing/snapshot 464 cat snapshot 465 466config TRACER_SNAPSHOT_PER_CPU_SWAP 467 bool "Allow snapshot to swap per CPU" 468 depends on TRACER_SNAPSHOT 469 select RING_BUFFER_ALLOW_SWAP 470 help 471 Allow doing a snapshot of a single CPU buffer instead of a 472 full swap (all buffers). If this is set, then the following is 473 allowed: 474 475 echo 1 > /sys/kernel/debug/tracing/per_cpu/cpu2/snapshot 476 477 After which, only the tracing buffer for CPU 2 was swapped with 478 the main tracing buffer, and the other CPU buffers remain the same. 479 480 When this is enabled, this adds a little more overhead to the 481 trace recording, as it needs to add some checks to synchronize 482 recording with swaps. But this does not affect the performance 483 of the overall system. This is enabled by default when the preempt 484 or irq latency tracers are enabled, as those need to swap as well 485 and already adds the overhead (plus a lot more). 486 487config TRACE_BRANCH_PROFILING 488 bool 489 select GENERIC_TRACER 490 491choice 492 prompt "Branch Profiling" 493 default BRANCH_PROFILE_NONE 494 help 495 The branch profiling is a software profiler. It will add hooks 496 into the C conditionals to test which path a branch takes. 497 498 The likely/unlikely profiler only looks at the conditions that 499 are annotated with a likely or unlikely macro. 500 501 The "all branch" profiler will profile every if-statement in the 502 kernel. This profiler will also enable the likely/unlikely 503 profiler. 504 505 Either of the above profilers adds a bit of overhead to the system. 506 If unsure, choose "No branch profiling". 507 508config BRANCH_PROFILE_NONE 509 bool "No branch profiling" 510 help 511 No branch profiling. Branch profiling adds a bit of overhead. 512 Only enable it if you want to analyse the branching behavior. 513 Otherwise keep it disabled. 514 515config PROFILE_ANNOTATED_BRANCHES 516 bool "Trace likely/unlikely profiler" 517 select TRACE_BRANCH_PROFILING 518 help 519 This tracer profiles all likely and unlikely macros 520 in the kernel. It will display the results in: 521 522 /sys/kernel/debug/tracing/trace_stat/branch_annotated 523 524 Note: this will add a significant overhead; only turn this 525 on if you need to profile the system's use of these macros. 526 527config PROFILE_ALL_BRANCHES 528 bool "Profile all if conditionals" if !FORTIFY_SOURCE 529 select TRACE_BRANCH_PROFILING 530 help 531 This tracer profiles all branch conditions. Every if () 532 taken in the kernel is recorded whether it hit or miss. 533 The results will be displayed in: 534 535 /sys/kernel/debug/tracing/trace_stat/branch_all 536 537 This option also enables the likely/unlikely profiler. 538 539 This configuration, when enabled, will impose a great overhead 540 on the system. This should only be enabled when the system 541 is to be analyzed in much detail. 542endchoice 543 544config TRACING_BRANCHES 545 bool 546 help 547 Selected by tracers that will trace the likely and unlikely 548 conditions. This prevents the tracers themselves from being 549 profiled. Profiling the tracing infrastructure can only happen 550 when the likelys and unlikelys are not being traced. 551 552config BRANCH_TRACER 553 bool "Trace likely/unlikely instances" 554 depends on TRACE_BRANCH_PROFILING 555 select TRACING_BRANCHES 556 help 557 This traces the events of likely and unlikely condition 558 calls in the kernel. The difference between this and the 559 "Trace likely/unlikely profiler" is that this is not a 560 histogram of the callers, but actually places the calling 561 events into a running trace buffer to see when and where the 562 events happened, as well as their results. 563 564 Say N if unsure. 565 566config BLK_DEV_IO_TRACE 567 bool "Support for tracing block IO actions" 568 depends on SYSFS 569 depends on BLOCK 570 select RELAY 571 select DEBUG_FS 572 select TRACEPOINTS 573 select GENERIC_TRACER 574 select STACKTRACE 575 help 576 Say Y here if you want to be able to trace the block layer actions 577 on a given queue. Tracing allows you to see any traffic happening 578 on a block device queue. For more information (and the userspace 579 support tools needed), fetch the blktrace tools from: 580 581 git://git.kernel.dk/blktrace.git 582 583 Tracing also is possible using the ftrace interface, e.g.: 584 585 echo 1 > /sys/block/sda/sda1/trace/enable 586 echo blk > /sys/kernel/debug/tracing/current_tracer 587 cat /sys/kernel/debug/tracing/trace_pipe 588 589 If unsure, say N. 590 591config KPROBE_EVENTS 592 depends on KPROBES 593 depends on HAVE_REGS_AND_STACK_ACCESS_API 594 bool "Enable kprobes-based dynamic events" 595 select TRACING 596 select PROBE_EVENTS 597 select DYNAMIC_EVENTS 598 default y 599 help 600 This allows the user to add tracing events (similar to tracepoints) 601 on the fly via the ftrace interface. See 602 Documentation/trace/kprobetrace.rst for more details. 603 604 Those events can be inserted wherever kprobes can probe, and record 605 various register and memory values. 606 607 This option is also required by perf-probe subcommand of perf tools. 608 If you want to use perf tools, this option is strongly recommended. 609 610config KPROBE_EVENTS_ON_NOTRACE 611 bool "Do NOT protect notrace function from kprobe events" 612 depends on KPROBE_EVENTS 613 depends on DYNAMIC_FTRACE 614 default n 615 help 616 This is only for the developers who want to debug ftrace itself 617 using kprobe events. 618 619 If kprobes can use ftrace instead of breakpoint, ftrace related 620 functions are protected from kprobe-events to prevent an infinite 621 recursion or any unexpected execution path which leads to a kernel 622 crash. 623 624 This option disables such protection and allows you to put kprobe 625 events on ftrace functions for debugging ftrace by itself. 626 Note that this might let you shoot yourself in the foot. 627 628 If unsure, say N. 629 630config UPROBE_EVENTS 631 bool "Enable uprobes-based dynamic events" 632 depends on ARCH_SUPPORTS_UPROBES 633 depends on MMU 634 depends on PERF_EVENTS 635 select UPROBES 636 select PROBE_EVENTS 637 select DYNAMIC_EVENTS 638 select TRACING 639 default y 640 help 641 This allows the user to add tracing events on top of userspace 642 dynamic events (similar to tracepoints) on the fly via the trace 643 events interface. Those events can be inserted wherever uprobes 644 can probe, and record various registers. 645 This option is required if you plan to use perf-probe subcommand 646 of perf tools on user space applications. 647 648config BPF_EVENTS 649 depends on BPF_SYSCALL 650 depends on (KPROBE_EVENTS || UPROBE_EVENTS) && PERF_EVENTS 651 bool 652 default y 653 help 654 This allows the user to attach BPF programs to kprobe, uprobe, and 655 tracepoint events. 656 657config DYNAMIC_EVENTS 658 def_bool n 659 660config PROBE_EVENTS 661 def_bool n 662 663config BPF_KPROBE_OVERRIDE 664 bool "Enable BPF programs to override a kprobed function" 665 depends on BPF_EVENTS 666 depends on FUNCTION_ERROR_INJECTION 667 default n 668 help 669 Allows BPF to override the execution of a probed function and 670 set a different return value. This is used for error injection. 671 672config FTRACE_MCOUNT_RECORD 673 def_bool y 674 depends on DYNAMIC_FTRACE 675 depends on HAVE_FTRACE_MCOUNT_RECORD 676 677config FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 678 bool 679 depends on FTRACE_MCOUNT_RECORD 680 681config FTRACE_MCOUNT_USE_CC 682 def_bool y 683 depends on $(cc-option,-mrecord-mcount) 684 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 685 depends on FTRACE_MCOUNT_RECORD 686 687config FTRACE_MCOUNT_USE_OBJTOOL 688 def_bool y 689 depends on HAVE_OBJTOOL_MCOUNT 690 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 691 depends on !FTRACE_MCOUNT_USE_CC 692 depends on FTRACE_MCOUNT_RECORD 693 694config FTRACE_MCOUNT_USE_RECORDMCOUNT 695 def_bool y 696 depends on !FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY 697 depends on !FTRACE_MCOUNT_USE_CC 698 depends on !FTRACE_MCOUNT_USE_OBJTOOL 699 depends on FTRACE_MCOUNT_RECORD 700 701config TRACING_MAP 702 bool 703 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG 704 help 705 tracing_map is a special-purpose lock-free map for tracing, 706 separated out as a stand-alone facility in order to allow it 707 to be shared between multiple tracers. It isn't meant to be 708 generally used outside of that context, and is normally 709 selected by tracers that use it. 710 711config SYNTH_EVENTS 712 bool "Synthetic trace events" 713 select TRACING 714 select DYNAMIC_EVENTS 715 default n 716 help 717 Synthetic events are user-defined trace events that can be 718 used to combine data from other trace events or in fact any 719 data source. Synthetic events can be generated indirectly 720 via the trace() action of histogram triggers or directly 721 by way of an in-kernel API. 722 723 See Documentation/trace/events.rst or 724 Documentation/trace/histogram.rst for details and examples. 725 726 If in doubt, say N. 727 728config HIST_TRIGGERS 729 bool "Histogram triggers" 730 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG 731 select TRACING_MAP 732 select TRACING 733 select DYNAMIC_EVENTS 734 select SYNTH_EVENTS 735 default n 736 help 737 Hist triggers allow one or more arbitrary trace event fields 738 to be aggregated into hash tables and dumped to stdout by 739 reading a debugfs/tracefs file. They're useful for 740 gathering quick and dirty (though precise) summaries of 741 event activity as an initial guide for further investigation 742 using more advanced tools. 743 744 Inter-event tracing of quantities such as latencies is also 745 supported using hist triggers under this option. 746 747 See Documentation/trace/histogram.rst. 748 If in doubt, say N. 749 750config TRACE_EVENT_INJECT 751 bool "Trace event injection" 752 depends on TRACING 753 help 754 Allow user-space to inject a specific trace event into the ring 755 buffer. This is mainly used for testing purpose. 756 757 If unsure, say N. 758 759config TRACEPOINT_BENCHMARK 760 bool "Add tracepoint that benchmarks tracepoints" 761 help 762 This option creates the tracepoint "benchmark:benchmark_event". 763 When the tracepoint is enabled, it kicks off a kernel thread that 764 goes into an infinite loop (calling cond_resched() to let other tasks 765 run), and calls the tracepoint. Each iteration will record the time 766 it took to write to the tracepoint and the next iteration that 767 data will be passed to the tracepoint itself. That is, the tracepoint 768 will report the time it took to do the previous tracepoint. 769 The string written to the tracepoint is a static string of 128 bytes 770 to keep the time the same. The initial string is simply a write of 771 "START". The second string records the cold cache time of the first 772 write which is not added to the rest of the calculations. 773 774 As it is a tight loop, it benchmarks as hot cache. That's fine because 775 we care most about hot paths that are probably in cache already. 776 777 An example of the output: 778 779 START 780 first=3672 [COLD CACHED] 781 last=632 first=3672 max=632 min=632 avg=316 std=446 std^2=199712 782 last=278 first=3672 max=632 min=278 avg=303 std=316 std^2=100337 783 last=277 first=3672 max=632 min=277 avg=296 std=258 std^2=67064 784 last=273 first=3672 max=632 min=273 avg=292 std=224 std^2=50411 785 last=273 first=3672 max=632 min=273 avg=288 std=200 std^2=40389 786 last=281 first=3672 max=632 min=273 avg=287 std=183 std^2=33666 787 788 789config RING_BUFFER_BENCHMARK 790 tristate "Ring buffer benchmark stress tester" 791 depends on RING_BUFFER 792 help 793 This option creates a test to stress the ring buffer and benchmark it. 794 It creates its own ring buffer such that it will not interfere with 795 any other users of the ring buffer (such as ftrace). It then creates 796 a producer and consumer that will run for 10 seconds and sleep for 797 10 seconds. Each interval it will print out the number of events 798 it recorded and give a rough estimate of how long each iteration took. 799 800 It does not disable interrupts or raise its priority, so it may be 801 affected by processes that are running. 802 803 If unsure, say N. 804 805config TRACE_EVAL_MAP_FILE 806 bool "Show eval mappings for trace events" 807 depends on TRACING 808 help 809 The "print fmt" of the trace events will show the enum/sizeof names 810 instead of their values. This can cause problems for user space tools 811 that use this string to parse the raw data as user space does not know 812 how to convert the string to its value. 813 814 To fix this, there's a special macro in the kernel that can be used 815 to convert an enum/sizeof into its value. If this macro is used, then 816 the print fmt strings will be converted to their values. 817 818 If something does not get converted properly, this option can be 819 used to show what enums/sizeof the kernel tried to convert. 820 821 This option is for debugging the conversions. A file is created 822 in the tracing directory called "eval_map" that will show the 823 names matched with their values and what trace event system they 824 belong too. 825 826 Normally, the mapping of the strings to values will be freed after 827 boot up or module load. With this option, they will not be freed, as 828 they are needed for the "eval_map" file. Enabling this option will 829 increase the memory footprint of the running kernel. 830 831 If unsure, say N. 832 833config FTRACE_RECORD_RECURSION 834 bool "Record functions that recurse in function tracing" 835 depends on FUNCTION_TRACER 836 help 837 All callbacks that attach to the function tracing have some sort 838 of protection against recursion. Even though the protection exists, 839 it adds overhead. This option will create a file in the tracefs 840 file system called "recursed_functions" that will list the functions 841 that triggered a recursion. 842 843 This will add more overhead to cases that have recursion. 844 845 If unsure, say N 846 847config FTRACE_RECORD_RECURSION_SIZE 848 int "Max number of recursed functions to record" 849 default 128 850 depends on FTRACE_RECORD_RECURSION 851 help 852 This defines the limit of number of functions that can be 853 listed in the "recursed_functions" file, that lists all 854 the functions that caused a recursion to happen. 855 This file can be reset, but the limit can not change in 856 size at runtime. 857 858config RING_BUFFER_RECORD_RECURSION 859 bool "Record functions that recurse in the ring buffer" 860 depends on FTRACE_RECORD_RECURSION 861 # default y, because it is coupled with FTRACE_RECORD_RECURSION 862 default y 863 help 864 The ring buffer has its own internal recursion. Although when 865 recursion happens it wont cause harm because of the protection, 866 but it does cause an unwanted overhead. Enabling this option will 867 place where recursion was detected into the ftrace "recursed_functions" 868 file. 869 870 This will add more overhead to cases that have recursion. 871 872config GCOV_PROFILE_FTRACE 873 bool "Enable GCOV profiling on ftrace subsystem" 874 depends on GCOV_KERNEL 875 help 876 Enable GCOV profiling on ftrace subsystem for checking 877 which functions/lines are tested. 878 879 If unsure, say N. 880 881 Note that on a kernel compiled with this config, ftrace will 882 run significantly slower. 883 884config FTRACE_SELFTEST 885 bool 886 887config FTRACE_STARTUP_TEST 888 bool "Perform a startup test on ftrace" 889 depends on GENERIC_TRACER 890 select FTRACE_SELFTEST 891 help 892 This option performs a series of startup tests on ftrace. On bootup 893 a series of tests are made to verify that the tracer is 894 functioning properly. It will do tests on all the configured 895 tracers of ftrace. 896 897config EVENT_TRACE_STARTUP_TEST 898 bool "Run selftest on trace events" 899 depends on FTRACE_STARTUP_TEST 900 default y 901 help 902 This option performs a test on all trace events in the system. 903 It basically just enables each event and runs some code that 904 will trigger events (not necessarily the event it enables) 905 This may take some time run as there are a lot of events. 906 907config EVENT_TRACE_TEST_SYSCALLS 908 bool "Run selftest on syscall events" 909 depends on EVENT_TRACE_STARTUP_TEST 910 help 911 This option will also enable testing every syscall event. 912 It only enables the event and disables it and runs various loads 913 with the event enabled. This adds a bit more time for kernel boot 914 up since it runs this on every system call defined. 915 916 TBD - enable a way to actually call the syscalls as we test their 917 events 918 919config RING_BUFFER_STARTUP_TEST 920 bool "Ring buffer startup self test" 921 depends on RING_BUFFER 922 help 923 Run a simple self test on the ring buffer on boot up. Late in the 924 kernel boot sequence, the test will start that kicks off 925 a thread per cpu. Each thread will write various size events 926 into the ring buffer. Another thread is created to send IPIs 927 to each of the threads, where the IPI handler will also write 928 to the ring buffer, to test/stress the nesting ability. 929 If any anomalies are discovered, a warning will be displayed 930 and all ring buffers will be disabled. 931 932 The test runs for 10 seconds. This will slow your boot time 933 by at least 10 more seconds. 934 935 At the end of the test, statics and more checks are done. 936 It will output the stats of each per cpu buffer. What 937 was written, the sizes, what was read, what was lost, and 938 other similar details. 939 940 If unsure, say N 941 942config RING_BUFFER_VALIDATE_TIME_DELTAS 943 bool "Verify ring buffer time stamp deltas" 944 depends on RING_BUFFER 945 help 946 This will audit the time stamps on the ring buffer sub 947 buffer to make sure that all the time deltas for the 948 events on a sub buffer matches the current time stamp. 949 This audit is performed for every event that is not 950 interrupted, or interrupting another event. A check 951 is also made when traversing sub buffers to make sure 952 that all the deltas on the previous sub buffer do not 953 add up to be greater than the current time stamp. 954 955 NOTE: This adds significant overhead to recording of events, 956 and should only be used to test the logic of the ring buffer. 957 Do not use it on production systems. 958 959 Only say Y if you understand what this does, and you 960 still want it enabled. Otherwise say N 961 962config MMIOTRACE_TEST 963 tristate "Test module for mmiotrace" 964 depends on MMIOTRACE && m 965 help 966 This is a dumb module for testing mmiotrace. It is very dangerous 967 as it will write garbage to IO memory starting at a given address. 968 However, it should be safe to use on e.g. unused portion of VRAM. 969 970 Say N, unless you absolutely know what you are doing. 971 972config PREEMPTIRQ_DELAY_TEST 973 tristate "Test module to create a preempt / IRQ disable delay thread to test latency tracers" 974 depends on m 975 help 976 Select this option to build a test module that can help test latency 977 tracers by executing a preempt or irq disable section with a user 978 configurable delay. The module busy waits for the duration of the 979 critical section. 980 981 For example, the following invocation generates a burst of three 982 irq-disabled critical sections for 500us: 983 modprobe preemptirq_delay_test test_mode=irq delay=500 burst_size=3 984 985 What's more, if you want to attach the test on the cpu which the latency 986 tracer is running on, specify cpu_affinity=cpu_num at the end of the 987 command. 988 989 If unsure, say N 990 991config SYNTH_EVENT_GEN_TEST 992 tristate "Test module for in-kernel synthetic event generation" 993 depends on SYNTH_EVENTS 994 help 995 This option creates a test module to check the base 996 functionality of in-kernel synthetic event definition and 997 generation. 998 999 To test, insert the module, and then check the trace buffer 1000 for the generated sample events. 1001 1002 If unsure, say N. 1003 1004config KPROBE_EVENT_GEN_TEST 1005 tristate "Test module for in-kernel kprobe event generation" 1006 depends on KPROBE_EVENTS 1007 help 1008 This option creates a test module to check the base 1009 functionality of in-kernel kprobe event definition. 1010 1011 To test, insert the module, and then check the trace buffer 1012 for the generated kprobe events. 1013 1014 If unsure, say N. 1015 1016config HIST_TRIGGERS_DEBUG 1017 bool "Hist trigger debug support" 1018 depends on HIST_TRIGGERS 1019 help 1020 Add "hist_debug" file for each event, which when read will 1021 dump out a bunch of internal details about the hist triggers 1022 defined on that event. 1023 1024 The hist_debug file serves a couple of purposes: 1025 1026 - Helps developers verify that nothing is broken. 1027 1028 - Provides educational information to support the details 1029 of the hist trigger internals as described by 1030 Documentation/trace/histogram-design.rst. 1031 1032 The hist_debug output only covers the data structures 1033 related to the histogram definitions themselves and doesn't 1034 display the internals of map buckets or variable values of 1035 running histograms. 1036 1037 If unsure, say N. 1038 1039endif # FTRACE 1040 1041endif # TRACING_SUPPORT 1042 1043