| blob | d5a19413d4f8a3945cdb53cd64c2fd887773730d |
1 # SPDX-License-Identifier: GPL-2.0-only
2 #
3 # Architectures that offer an FUNCTION_TRACER implementation should
4 # select HAVE_FUNCTION_TRACER:
5 #
7 config USER_STACKTRACE_SUPPORT
8 bool
10 config NOP_TRACER
11 bool
13 config HAVE_FUNCTION_TRACER
14 bool
15 help
16 See Documentation/trace/ftrace-design.rst
18 config HAVE_FUNCTION_GRAPH_TRACER
19 bool
20 help
21 See Documentation/trace/ftrace-design.rst
23 config HAVE_DYNAMIC_FTRACE
24 bool
25 help
26 See Documentation/trace/ftrace-design.rst
28 config HAVE_DYNAMIC_FTRACE_WITH_REGS
29 bool
31 config HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
32 bool
34 config 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().
43 config HAVE_FTRACE_MCOUNT_RECORD
44 bool
45 help
46 See Documentation/trace/ftrace-design.rst
48 config HAVE_SYSCALL_TRACEPOINTS
49 bool
50 help
51 See Documentation/trace/ftrace-design.rst
53 config HAVE_FENTRY
54 bool
55 help
56 Arch supports the gcc options -pg with -mfentry
58 config HAVE_NOP_MCOUNT
59 bool
60 help
61 Arch supports the gcc options -pg with -mrecord-mcount and -nop-mcount
63 config HAVE_C_RECORDMCOUNT
64 bool
65 help
66 C version of recordmcount available?
68 config TRACER_MAX_TRACE
69 bool
71 config TRACE_CLOCK
72 bool
74 config RING_BUFFER
75 bool
76 select TRACE_CLOCK
77 select IRQ_WORK
79 config EVENT_TRACING
80 select CONTEXT_SWITCH_TRACER
81 select GLOB
82 bool
84 config CONTEXT_SWITCH_TRACER
85 bool
87 config RING_BUFFER_ALLOW_SWAP
88 bool
89 help
90 Allow the use of ring_buffer_swap_cpu.
91 Adds a very slight overhead to tracing when enabled.
93 config PREEMPTIRQ_TRACEPOINTS
94 bool
95 depends on TRACE_PREEMPT_TOGGLE || TRACE_IRQFLAGS
96 select TRACING
97 default y
98 help
99 Create preempt/irq toggle tracepoints if needed, so that other parts
100 of the kernel can use them to generate or add hooks to them.
102 # All tracer options should select GENERIC_TRACER. For those options that are
103 # enabled by all tracers (context switch and event tracer) they select TRACING.
104 # This allows those options to appear when no other tracer is selected. But the
105 # options do not appear when something else selects it. We need the two options
106 # GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
107 # hiding of the automatic options.
109 config TRACING
110 bool
111 select RING_BUFFER
112 select STACKTRACE if STACKTRACE_SUPPORT
113 select TRACEPOINTS
114 select NOP_TRACER
115 select BINARY_PRINTF
116 select EVENT_TRACING
117 select TRACE_CLOCK
119 config GENERIC_TRACER
120 bool
121 select TRACING
123 #
124 # Minimum requirements an architecture has to meet for us to
125 # be able to offer generic tracing facilities:
126 #
127 config TRACING_SUPPORT
128 bool
129 depends on TRACE_IRQFLAGS_SUPPORT
130 depends on STACKTRACE_SUPPORT
131 default y
133 if TRACING_SUPPORT
135 menuconfig FTRACE
136 bool "Tracers"
137 default y if DEBUG_KERNEL
138 help
139 Enable the kernel tracing infrastructure.
141 if FTRACE
143 config BOOTTIME_TRACING
144 bool "Boot-time Tracing support"
145 depends on TRACING
146 select BOOT_CONFIG
147 help
148 Enable developer to setup ftrace subsystem via supplemental
149 kernel cmdline at boot time for debugging (tracing) driver
150 initialization and boot process.
152 config FUNCTION_TRACER
153 bool "Kernel Function Tracer"
154 depends on HAVE_FUNCTION_TRACER
155 select KALLSYMS
156 select GENERIC_TRACER
157 select CONTEXT_SWITCH_TRACER
158 select GLOB
159 select TASKS_RCU if PREEMPTION
160 select TASKS_RUDE_RCU
161 help
162 Enable the kernel to trace every kernel function. This is done
163 by using a compiler feature to insert a small, 5-byte No-Operation
164 instruction at the beginning of every kernel function, which NOP
165 sequence is then dynamically patched into a tracer call when
166 tracing is enabled by the administrator. If it's runtime disabled
167 (the bootup default), then the overhead of the instructions is very
168 small and not measurable even in micro-benchmarks.
170 config FUNCTION_GRAPH_TRACER
171 bool "Kernel Function Graph Tracer"
172 depends on HAVE_FUNCTION_GRAPH_TRACER
173 depends on FUNCTION_TRACER
174 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
175 default y
176 help
177 Enable the kernel to trace a function at both its return
178 and its entry.
179 Its first purpose is to trace the duration of functions and
180 draw a call graph for each thread with some information like
181 the return value. This is done by setting the current return
182 address on the current task structure into a stack of calls.
184 config DYNAMIC_FTRACE
185 bool "enable/disable function tracing dynamically"
186 depends on FUNCTION_TRACER
187 depends on HAVE_DYNAMIC_FTRACE
188 default y
189 help
190 This option will modify all the calls to function tracing
191 dynamically (will patch them out of the binary image and
192 replace them with a No-Op instruction) on boot up. During
193 compile time, a table is made of all the locations that ftrace
194 can function trace, and this table is linked into the kernel
195 image. When this is enabled, functions can be individually
196 enabled, and the functions not enabled will not affect
197 performance of the system.
199 See the files in /sys/kernel/debug/tracing:
200 available_filter_functions
201 set_ftrace_filter
202 set_ftrace_notrace
204 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
205 otherwise has native performance as long as no tracing is active.
207 config DYNAMIC_FTRACE_WITH_REGS
208 def_bool y
209 depends on DYNAMIC_FTRACE
210 depends on HAVE_DYNAMIC_FTRACE_WITH_REGS
212 config DYNAMIC_FTRACE_WITH_DIRECT_CALLS
213 def_bool y
214 depends on DYNAMIC_FTRACE_WITH_REGS
215 depends on HAVE_DYNAMIC_FTRACE_WITH_DIRECT_CALLS
217 config FUNCTION_PROFILER
218 bool "Kernel function profiler"
219 depends on FUNCTION_TRACER
220 default n
221 help
222 This option enables the kernel function profiler. A file is created
223 in debugfs called function_profile_enabled which defaults to zero.
224 When a 1 is echoed into this file profiling begins, and when a
225 zero is entered, profiling stops. A "functions" file is created in
226 the trace_stat directory; this file shows the list of functions that
227 have been hit and their counters.
229 If in doubt, say N.
231 config STACK_TRACER
232 bool "Trace max stack"
233 depends on HAVE_FUNCTION_TRACER
234 select FUNCTION_TRACER
235 select STACKTRACE
236 select KALLSYMS
237 help
238 This special tracer records the maximum stack footprint of the
239 kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
241 This tracer works by hooking into every function call that the
242 kernel executes, and keeping a maximum stack depth value and
243 stack-trace saved. If this is configured with DYNAMIC_FTRACE
244 then it will not have any overhead while the stack tracer
245 is disabled.
247 To enable the stack tracer on bootup, pass in 'stacktrace'
248 on the kernel command line.
250 The stack tracer can also be enabled or disabled via the
251 sysctl kernel.stack_tracer_enabled
253 Say N if unsure.
255 config TRACE_PREEMPT_TOGGLE
256 bool
257 help
258 Enables hooks which will be called when preemption is first disabled,
259 and last enabled.
261 config IRQSOFF_TRACER
262 bool "Interrupts-off Latency Tracer"
263 default n
264 depends on TRACE_IRQFLAGS_SUPPORT
265 select TRACE_IRQFLAGS
266 select GENERIC_TRACER
267 select TRACER_MAX_TRACE
268 select RING_BUFFER_ALLOW_SWAP
269 select TRACER_SNAPSHOT
270 select TRACER_SNAPSHOT_PER_CPU_SWAP
271 help
272 This option measures the time spent in irqs-off critical
273 sections, with microsecond accuracy.
275 The default measurement method is a maximum search, which is
276 disabled by default and can be runtime (re-)started
277 via:
279 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
281 (Note that kernel size and overhead increase with this option
282 enabled. This option and the preempt-off timing option can be
283 used together or separately.)
285 config PREEMPT_TRACER
286 bool "Preemption-off Latency Tracer"
287 default n
288 depends on PREEMPTION
289 select GENERIC_TRACER
290 select TRACER_MAX_TRACE
291 select RING_BUFFER_ALLOW_SWAP
292 select TRACER_SNAPSHOT
293 select TRACER_SNAPSHOT_PER_CPU_SWAP
294 select TRACE_PREEMPT_TOGGLE
295 help
296 This option measures the time spent in preemption-off critical
297 sections, with microsecond accuracy.
299 The default measurement method is a maximum search, which is
300 disabled by default and can be runtime (re-)started
301 via:
303 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
305 (Note that kernel size and overhead increase with this option
306 enabled. This option and the irqs-off timing option can be
307 used together or separately.)
309 config SCHED_TRACER
310 bool "Scheduling Latency Tracer"
311 select GENERIC_TRACER
312 select CONTEXT_SWITCH_TRACER
313 select TRACER_MAX_TRACE
314 select TRACER_SNAPSHOT
315 help
316 This tracer tracks the latency of the highest priority task
317 to be scheduled in, starting from the point it has woken up.
319 config HWLAT_TRACER
320 bool "Tracer to detect hardware latencies (like SMIs)"
321 select GENERIC_TRACER
322 help
323 This tracer, when enabled will create one or more kernel threads,
324 depending on what the cpumask file is set to, which each thread
325 spinning in a loop looking for interruptions caused by
326 something other than the kernel. For example, if a
327 System Management Interrupt (SMI) takes a noticeable amount of
328 time, this tracer will detect it. This is useful for testing
329 if a system is reliable for Real Time tasks.
331 Some files are created in the tracing directory when this
332 is enabled:
334 hwlat_detector/width - time in usecs for how long to spin for
335 hwlat_detector/window - time in usecs between the start of each
336 iteration
338 A kernel thread is created that will spin with interrupts disabled
339 for "width" microseconds in every "window" cycle. It will not spin
340 for "window - width" microseconds, where the system can
341 continue to operate.
343 The output will appear in the trace and trace_pipe files.
345 When the tracer is not running, it has no affect on the system,
346 but when it is running, it can cause the system to be
347 periodically non responsive. Do not run this tracer on a
348 production system.
350 To enable this tracer, echo in "hwlat" into the current_tracer
351 file. Every time a latency is greater than tracing_thresh, it will
352 be recorded into the ring buffer.
354 config MMIOTRACE
355 bool "Memory mapped IO tracing"
356 depends on HAVE_MMIOTRACE_SUPPORT && PCI
357 select GENERIC_TRACER
358 help
359 Mmiotrace traces Memory Mapped I/O access and is meant for
360 debugging and reverse engineering. It is called from the ioremap
361 implementation and works via page faults. Tracing is disabled by
362 default and can be enabled at run-time.
364 See Documentation/trace/mmiotrace.rst.
365 If you are not helping to develop drivers, say N.
367 config ENABLE_DEFAULT_TRACERS
368 bool "Trace process context switches and events"
369 depends on !GENERIC_TRACER
370 select TRACING
371 help
372 This tracer hooks to various trace points in the kernel,
373 allowing the user to pick and choose which trace point they
374 want to trace. It also includes the sched_switch tracer plugin.
376 config FTRACE_SYSCALLS
377 bool "Trace syscalls"
378 depends on HAVE_SYSCALL_TRACEPOINTS
379 select GENERIC_TRACER
380 select KALLSYMS
381 help
382 Basic tracer to catch the syscall entry and exit events.
384 config TRACER_SNAPSHOT
385 bool "Create a snapshot trace buffer"
386 select TRACER_MAX_TRACE
387 help
388 Allow tracing users to take snapshot of the current buffer using the
389 ftrace interface, e.g.:
391 echo 1 > /sys/kernel/debug/tracing/snapshot
392 cat snapshot
394 config TRACER_SNAPSHOT_PER_CPU_SWAP
395 bool "Allow snapshot to swap per CPU"
396 depends on TRACER_SNAPSHOT
397 select RING_BUFFER_ALLOW_SWAP
398 help
399 Allow doing a snapshot of a single CPU buffer instead of a
400 full swap (all buffers). If this is set, then the following is
401 allowed:
403 echo 1 > /sys/kernel/debug/tracing/per_cpu/cpu2/snapshot
405 After which, only the tracing buffer for CPU 2 was swapped with
406 the main tracing buffer, and the other CPU buffers remain the same.
408 When this is enabled, this adds a little more overhead to the
409 trace recording, as it needs to add some checks to synchronize
410 recording with swaps. But this does not affect the performance
411 of the overall system. This is enabled by default when the preempt
412 or irq latency tracers are enabled, as those need to swap as well
413 and already adds the overhead (plus a lot more).
415 config TRACE_BRANCH_PROFILING
416 bool
417 select GENERIC_TRACER
419 choice
420 prompt "Branch Profiling"
421 default BRANCH_PROFILE_NONE
422 help
423 The branch profiling is a software profiler. It will add hooks
424 into the C conditionals to test which path a branch takes.
426 The likely/unlikely profiler only looks at the conditions that
427 are annotated with a likely or unlikely macro.
429 The "all branch" profiler will profile every if-statement in the
430 kernel. This profiler will also enable the likely/unlikely
431 profiler.
433 Either of the above profilers adds a bit of overhead to the system.
434 If unsure, choose "No branch profiling".
436 config BRANCH_PROFILE_NONE
437 bool "No branch profiling"
438 help
439 No branch profiling. Branch profiling adds a bit of overhead.
440 Only enable it if you want to analyse the branching behavior.
441 Otherwise keep it disabled.
443 config PROFILE_ANNOTATED_BRANCHES
444 bool "Trace likely/unlikely profiler"
445 select TRACE_BRANCH_PROFILING
446 help
447 This tracer profiles all likely and unlikely macros
448 in the kernel. It will display the results in:
450 /sys/kernel/debug/tracing/trace_stat/branch_annotated
452 Note: this will add a significant overhead; only turn this
453 on if you need to profile the system's use of these macros.
455 config PROFILE_ALL_BRANCHES
456 bool "Profile all if conditionals" if !FORTIFY_SOURCE
457 select TRACE_BRANCH_PROFILING
458 help
459 This tracer profiles all branch conditions. Every if ()
460 taken in the kernel is recorded whether it hit or miss.
461 The results will be displayed in:
463 /sys/kernel/debug/tracing/trace_stat/branch_all
465 This option also enables the likely/unlikely profiler.
467 This configuration, when enabled, will impose a great overhead
468 on the system. This should only be enabled when the system
469 is to be analyzed in much detail.
470 endchoice
472 config TRACING_BRANCHES
473 bool
474 help
475 Selected by tracers that will trace the likely and unlikely
476 conditions. This prevents the tracers themselves from being
477 profiled. Profiling the tracing infrastructure can only happen
478 when the likelys and unlikelys are not being traced.
480 config BRANCH_TRACER
481 bool "Trace likely/unlikely instances"
482 depends on TRACE_BRANCH_PROFILING
483 select TRACING_BRANCHES
484 help
485 This traces the events of likely and unlikely condition
486 calls in the kernel. The difference between this and the
487 "Trace likely/unlikely profiler" is that this is not a
488 histogram of the callers, but actually places the calling
489 events into a running trace buffer to see when and where the
490 events happened, as well as their results.
492 Say N if unsure.
494 config BLK_DEV_IO_TRACE
495 bool "Support for tracing block IO actions"
496 depends on SYSFS
497 depends on BLOCK
498 select RELAY
499 select DEBUG_FS
500 select TRACEPOINTS
501 select GENERIC_TRACER
502 select STACKTRACE
503 help
504 Say Y here if you want to be able to trace the block layer actions
505 on a given queue. Tracing allows you to see any traffic happening
506 on a block device queue. For more information (and the userspace
507 support tools needed), fetch the blktrace tools from:
509 git://git.kernel.dk/blktrace.git
511 Tracing also is possible using the ftrace interface, e.g.:
513 echo 1 > /sys/block/sda/sda1/trace/enable
514 echo blk > /sys/kernel/debug/tracing/current_tracer
515 cat /sys/kernel/debug/tracing/trace_pipe
517 If unsure, say N.
519 config KPROBE_EVENTS
520 depends on KPROBES
521 depends on HAVE_REGS_AND_STACK_ACCESS_API
522 bool "Enable kprobes-based dynamic events"
523 select TRACING
524 select PROBE_EVENTS
525 select DYNAMIC_EVENTS
526 default y
527 help
528 This allows the user to add tracing events (similar to tracepoints)
529 on the fly via the ftrace interface. See
530 Documentation/trace/kprobetrace.rst for more details.
532 Those events can be inserted wherever kprobes can probe, and record
533 various register and memory values.
535 This option is also required by perf-probe subcommand of perf tools.
536 If you want to use perf tools, this option is strongly recommended.
538 config KPROBE_EVENTS_ON_NOTRACE
539 bool "Do NOT protect notrace function from kprobe events"
540 depends on KPROBE_EVENTS
541 depends on KPROBES_ON_FTRACE
542 default n
543 help
544 This is only for the developers who want to debug ftrace itself
545 using kprobe events.
547 If kprobes can use ftrace instead of breakpoint, ftrace related
548 functions are protected from kprobe-events to prevent an infinit
549 recursion or any unexpected execution path which leads to a kernel
550 crash.
552 This option disables such protection and allows you to put kprobe
553 events on ftrace functions for debugging ftrace by itself.
554 Note that this might let you shoot yourself in the foot.
556 If unsure, say N.
558 config UPROBE_EVENTS
559 bool "Enable uprobes-based dynamic events"
560 depends on ARCH_SUPPORTS_UPROBES
561 depends on MMU
562 depends on PERF_EVENTS
563 select UPROBES
564 select PROBE_EVENTS
565 select DYNAMIC_EVENTS
566 select TRACING
567 default y
568 help
569 This allows the user to add tracing events on top of userspace
570 dynamic events (similar to tracepoints) on the fly via the trace
571 events interface. Those events can be inserted wherever uprobes
572 can probe, and record various registers.
573 This option is required if you plan to use perf-probe subcommand
574 of perf tools on user space applications.
576 config BPF_EVENTS
577 depends on BPF_SYSCALL
578 depends on (KPROBE_EVENTS || UPROBE_EVENTS) && PERF_EVENTS
579 bool
580 default y
581 help
582 This allows the user to attach BPF programs to kprobe, uprobe, and
583 tracepoint events.
585 config DYNAMIC_EVENTS
586 def_bool n
588 config PROBE_EVENTS
589 def_bool n
591 config BPF_KPROBE_OVERRIDE
592 bool "Enable BPF programs to override a kprobed function"
593 depends on BPF_EVENTS
594 depends on FUNCTION_ERROR_INJECTION
595 default n
596 help
597 Allows BPF to override the execution of a probed function and
598 set a different return value. This is used for error injection.
600 config FTRACE_MCOUNT_RECORD
601 def_bool y
602 depends on DYNAMIC_FTRACE
603 depends on HAVE_FTRACE_MCOUNT_RECORD
605 config TRACING_MAP
606 bool
607 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG
608 help
609 tracing_map is a special-purpose lock-free map for tracing,
610 separated out as a stand-alone facility in order to allow it
611 to be shared between multiple tracers. It isn't meant to be
612 generally used outside of that context, and is normally
613 selected by tracers that use it.
615 config SYNTH_EVENTS
616 bool "Synthetic trace events"
617 select TRACING
618 select DYNAMIC_EVENTS
619 default n
620 help
621 Synthetic events are user-defined trace events that can be
622 used to combine data from other trace events or in fact any
623 data source. Synthetic events can be generated indirectly
624 via the trace() action of histogram triggers or directly
625 by way of an in-kernel API.
627 See Documentation/trace/events.rst or
628 Documentation/trace/histogram.rst for details and examples.
630 If in doubt, say N.
632 config HIST_TRIGGERS
633 bool "Histogram triggers"
634 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG
635 select TRACING_MAP
636 select TRACING
637 select DYNAMIC_EVENTS
638 select SYNTH_EVENTS
639 default n
640 help
641 Hist triggers allow one or more arbitrary trace event fields
642 to be aggregated into hash tables and dumped to stdout by
643 reading a debugfs/tracefs file. They're useful for
644 gathering quick and dirty (though precise) summaries of
645 event activity as an initial guide for further investigation
646 using more advanced tools.
648 Inter-event tracing of quantities such as latencies is also
649 supported using hist triggers under this option.
651 See Documentation/trace/histogram.rst.
652 If in doubt, say N.
654 config TRACE_EVENT_INJECT
655 bool "Trace event injection"
656 depends on TRACING
657 help
658 Allow user-space to inject a specific trace event into the ring
659 buffer. This is mainly used for testing purpose.
661 If unsure, say N.
663 config TRACEPOINT_BENCHMARK
664 bool "Add tracepoint that benchmarks tracepoints"
665 help
666 This option creates the tracepoint "benchmark:benchmark_event".
667 When the tracepoint is enabled, it kicks off a kernel thread that
668 goes into an infinite loop (calling cond_sched() to let other tasks
669 run), and calls the tracepoint. Each iteration will record the time
670 it took to write to the tracepoint and the next iteration that
671 data will be passed to the tracepoint itself. That is, the tracepoint
672 will report the time it took to do the previous tracepoint.
673 The string written to the tracepoint is a static string of 128 bytes
674 to keep the time the same. The initial string is simply a write of
675 "START". The second string records the cold cache time of the first
676 write which is not added to the rest of the calculations.
678 As it is a tight loop, it benchmarks as hot cache. That's fine because
679 we care most about hot paths that are probably in cache already.
681 An example of the output:
683 START
684 first=3672 [COLD CACHED]
685 last=632 first=3672 max=632 min=632 avg=316 std=446 std^2=199712
686 last=278 first=3672 max=632 min=278 avg=303 std=316 std^2=100337
687 last=277 first=3672 max=632 min=277 avg=296 std=258 std^2=67064
688 last=273 first=3672 max=632 min=273 avg=292 std=224 std^2=50411
689 last=273 first=3672 max=632 min=273 avg=288 std=200 std^2=40389
690 last=281 first=3672 max=632 min=273 avg=287 std=183 std^2=33666
693 config RING_BUFFER_BENCHMARK
694 tristate "Ring buffer benchmark stress tester"
695 depends on RING_BUFFER
696 help
697 This option creates a test to stress the ring buffer and benchmark it.
698 It creates its own ring buffer such that it will not interfere with
699 any other users of the ring buffer (such as ftrace). It then creates
700 a producer and consumer that will run for 10 seconds and sleep for
701 10 seconds. Each interval it will print out the number of events
702 it recorded and give a rough estimate of how long each iteration took.
704 It does not disable interrupts or raise its priority, so it may be
705 affected by processes that are running.
707 If unsure, say N.
709 config TRACE_EVAL_MAP_FILE
710 bool "Show eval mappings for trace events"
711 depends on TRACING
712 help
713 The "print fmt" of the trace events will show the enum/sizeof names
714 instead of their values. This can cause problems for user space tools
715 that use this string to parse the raw data as user space does not know
716 how to convert the string to its value.
718 To fix this, there's a special macro in the kernel that can be used
719 to convert an enum/sizeof into its value. If this macro is used, then
720 the print fmt strings will be converted to their values.
722 If something does not get converted properly, this option can be
723 used to show what enums/sizeof the kernel tried to convert.
725 This option is for debugging the conversions. A file is created
726 in the tracing directory called "eval_map" that will show the
727 names matched with their values and what trace event system they
728 belong too.
730 Normally, the mapping of the strings to values will be freed after
731 boot up or module load. With this option, they will not be freed, as
732 they are needed for the "eval_map" file. Enabling this option will
733 increase the memory footprint of the running kernel.
735 If unsure, say N.
737 config FTRACE_RECORD_RECURSION
738 bool "Record functions that recurse in function tracing"
739 depends on FUNCTION_TRACER
740 help
741 All callbacks that attach to the function tracing have some sort
742 of protection against recursion. Even though the protection exists,
743 it adds overhead. This option will create a file in the tracefs
744 file system called "recursed_functions" that will list the functions
745 that triggered a recursion.
747 This will add more overhead to cases that have recursion.
749 If unsure, say N
751 config FTRACE_RECORD_RECURSION_SIZE
752 int "Max number of recursed functions to record"
753 default 128
754 depends on FTRACE_RECORD_RECURSION
755 help
756 This defines the limit of number of functions that can be
757 listed in the "recursed_functions" file, that lists all
758 the functions that caused a recursion to happen.
759 This file can be reset, but the limit can not change in
760 size at runtime.
762 config RING_BUFFER_RECORD_RECURSION
763 bool "Record functions that recurse in the ring buffer"
764 depends on FTRACE_RECORD_RECURSION
765 # default y, because it is coupled with FTRACE_RECORD_RECURSION
766 default y
767 help
768 The ring buffer has its own internal recursion. Although when
769 recursion happens it wont cause harm because of the protection,
770 but it does cause an unwanted overhead. Enabling this option will
771 place where recursion was detected into the ftrace "recursed_functions"
772 file.
774 This will add more overhead to cases that have recursion.
776 config GCOV_PROFILE_FTRACE
777 bool "Enable GCOV profiling on ftrace subsystem"
778 depends on GCOV_KERNEL
779 help
780 Enable GCOV profiling on ftrace subsystem for checking
781 which functions/lines are tested.
783 If unsure, say N.
785 Note that on a kernel compiled with this config, ftrace will
786 run significantly slower.
788 config FTRACE_SELFTEST
789 bool
791 config FTRACE_STARTUP_TEST
792 bool "Perform a startup test on ftrace"
793 depends on GENERIC_TRACER
794 select FTRACE_SELFTEST
795 help
796 This option performs a series of startup tests on ftrace. On bootup
797 a series of tests are made to verify that the tracer is
798 functioning properly. It will do tests on all the configured
799 tracers of ftrace.
801 config EVENT_TRACE_STARTUP_TEST
802 bool "Run selftest on trace events"
803 depends on FTRACE_STARTUP_TEST
804 default y
805 help
806 This option performs a test on all trace events in the system.
807 It basically just enables each event and runs some code that
808 will trigger events (not necessarily the event it enables)
809 This may take some time run as there are a lot of events.
811 config EVENT_TRACE_TEST_SYSCALLS
812 bool "Run selftest on syscall events"
813 depends on EVENT_TRACE_STARTUP_TEST
814 help
815 This option will also enable testing every syscall event.
816 It only enables the event and disables it and runs various loads
817 with the event enabled. This adds a bit more time for kernel boot
818 up since it runs this on every system call defined.
820 TBD - enable a way to actually call the syscalls as we test their
821 events
823 config RING_BUFFER_STARTUP_TEST
824 bool "Ring buffer startup self test"
825 depends on RING_BUFFER
826 help
827 Run a simple self test on the ring buffer on boot up. Late in the
828 kernel boot sequence, the test will start that kicks off
829 a thread per cpu. Each thread will write various size events
830 into the ring buffer. Another thread is created to send IPIs
831 to each of the threads, where the IPI handler will also write
832 to the ring buffer, to test/stress the nesting ability.
833 If any anomalies are discovered, a warning will be displayed
834 and all ring buffers will be disabled.
836 The test runs for 10 seconds. This will slow your boot time
837 by at least 10 more seconds.
839 At the end of the test, statics and more checks are done.
840 It will output the stats of each per cpu buffer. What
841 was written, the sizes, what was read, what was lost, and
842 other similar details.
844 If unsure, say N
846 config RING_BUFFER_VALIDATE_TIME_DELTAS
847 bool "Verify ring buffer time stamp deltas"
848 depends on RING_BUFFER
849 help
850 This will audit the time stamps on the ring buffer sub
851 buffer to make sure that all the time deltas for the
852 events on a sub buffer matches the current time stamp.
853 This audit is performed for every event that is not
854 interrupted, or interrupting another event. A check
855 is also made when traversing sub buffers to make sure
856 that all the deltas on the previous sub buffer do not
857 add up to be greater than the current time stamp.
859 NOTE: This adds significant overhead to recording of events,
860 and should only be used to test the logic of the ring buffer.
861 Do not use it on production systems.
863 Only say Y if you understand what this does, and you
864 still want it enabled. Otherwise say N
866 config MMIOTRACE_TEST
867 tristate "Test module for mmiotrace"
868 depends on MMIOTRACE && m
869 help
870 This is a dumb module for testing mmiotrace. It is very dangerous
871 as it will write garbage to IO memory starting at a given address.
872 However, it should be safe to use on e.g. unused portion of VRAM.
874 Say N, unless you absolutely know what you are doing.
876 config PREEMPTIRQ_DELAY_TEST
877 tristate "Test module to create a preempt / IRQ disable delay thread to test latency tracers"
878 depends on m
879 help
880 Select this option to build a test module that can help test latency
881 tracers by executing a preempt or irq disable section with a user
882 configurable delay. The module busy waits for the duration of the
883 critical section.
885 For example, the following invocation generates a burst of three
886 irq-disabled critical sections for 500us:
887 modprobe preemptirq_delay_test test_mode=irq delay=500 burst_size=3
889 If unsure, say N
891 config SYNTH_EVENT_GEN_TEST
892 tristate "Test module for in-kernel synthetic event generation"
893 depends on SYNTH_EVENTS
894 help
895 This option creates a test module to check the base
896 functionality of in-kernel synthetic event definition and
897 generation.
899 To test, insert the module, and then check the trace buffer
900 for the generated sample events.
902 If unsure, say N.
904 config KPROBE_EVENT_GEN_TEST
905 tristate "Test module for in-kernel kprobe event generation"
906 depends on KPROBE_EVENTS
907 help
908 This option creates a test module to check the base
909 functionality of in-kernel kprobe event definition.
911 To test, insert the module, and then check the trace buffer
912 for the generated kprobe events.
914 If unsure, say N.
916 config HIST_TRIGGERS_DEBUG
917 bool "Hist trigger debug support"
918 depends on HIST_TRIGGERS
919 help
920 Add "hist_debug" file for each event, which when read will
921 dump out a bunch of internal details about the hist triggers
922 defined on that event.
924 The hist_debug file serves a couple of purposes:
926 - Helps developers verify that nothing is broken.
928 - Provides educational information to support the details
929 of the hist trigger internals as described by
930 Documentation/trace/histogram-design.rst.
932 The hist_debug output only covers the data structures
933 related to the histogram definitions themselves and doesn't
934 display the internals of map buckets or variable values of
935 running histograms.
937 If unsure, say N.
939 endif # FTRACE
941 endif # TRACING_SUPPORT