Loading...
1#
2# Architectures that offer an FUNCTION_TRACER implementation should
3# select HAVE_FUNCTION_TRACER:
4#
5
6config USER_STACKTRACE_SUPPORT
7 bool
8
9config NOP_TRACER
10 bool
11
12config HAVE_FTRACE_NMI_ENTER
13 bool
14 help
15 See Documentation/trace/ftrace-design.txt
16
17config HAVE_FUNCTION_TRACER
18 bool
19 help
20 See Documentation/trace/ftrace-design.txt
21
22config HAVE_FUNCTION_GRAPH_TRACER
23 bool
24 help
25 See Documentation/trace/ftrace-design.txt
26
27config HAVE_DYNAMIC_FTRACE
28 bool
29 help
30 See Documentation/trace/ftrace-design.txt
31
32config HAVE_DYNAMIC_FTRACE_WITH_REGS
33 bool
34
35config HAVE_FTRACE_MCOUNT_RECORD
36 bool
37 help
38 See Documentation/trace/ftrace-design.txt
39
40config HAVE_SYSCALL_TRACEPOINTS
41 bool
42 help
43 See Documentation/trace/ftrace-design.txt
44
45config HAVE_FENTRY
46 bool
47 help
48 Arch supports the gcc options -pg with -mfentry
49
50config HAVE_C_RECORDMCOUNT
51 bool
52 help
53 C version of recordmcount available?
54
55config TRACER_MAX_TRACE
56 bool
57
58config TRACE_CLOCK
59 bool
60
61config RING_BUFFER
62 bool
63 select TRACE_CLOCK
64 select IRQ_WORK
65
66config FTRACE_NMI_ENTER
67 bool
68 depends on HAVE_FTRACE_NMI_ENTER
69 default y
70
71config EVENT_TRACING
72 select CONTEXT_SWITCH_TRACER
73 select GLOB
74 bool
75
76config CONTEXT_SWITCH_TRACER
77 bool
78
79config RING_BUFFER_ALLOW_SWAP
80 bool
81 help
82 Allow the use of ring_buffer_swap_cpu.
83 Adds a very slight overhead to tracing when enabled.
84
85# All tracer options should select GENERIC_TRACER. For those options that are
86# enabled by all tracers (context switch and event tracer) they select TRACING.
87# This allows those options to appear when no other tracer is selected. But the
88# options do not appear when something else selects it. We need the two options
89# GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
90# hiding of the automatic options.
91
92config TRACING
93 bool
94 select DEBUG_FS
95 select RING_BUFFER
96 select STACKTRACE if STACKTRACE_SUPPORT
97 select TRACEPOINTS
98 select NOP_TRACER
99 select BINARY_PRINTF
100 select EVENT_TRACING
101 select TRACE_CLOCK
102
103config GENERIC_TRACER
104 bool
105 select TRACING
106
107#
108# Minimum requirements an architecture has to meet for us to
109# be able to offer generic tracing facilities:
110#
111config TRACING_SUPPORT
112 bool
113 # PPC32 has no irqflags tracing support, but it can use most of the
114 # tracers anyway, they were tested to build and work. Note that new
115 # exceptions to this list aren't welcomed, better implement the
116 # irqflags tracing for your architecture.
117 depends on TRACE_IRQFLAGS_SUPPORT || PPC32
118 depends on STACKTRACE_SUPPORT
119 default y
120
121if TRACING_SUPPORT
122
123menuconfig FTRACE
124 bool "Tracers"
125 default y if DEBUG_KERNEL
126 help
127 Enable the kernel tracing infrastructure.
128
129if FTRACE
130
131config FUNCTION_TRACER
132 bool "Kernel Function Tracer"
133 depends on HAVE_FUNCTION_TRACER
134 select KALLSYMS
135 select GENERIC_TRACER
136 select CONTEXT_SWITCH_TRACER
137 select GLOB
138 help
139 Enable the kernel to trace every kernel function. This is done
140 by using a compiler feature to insert a small, 5-byte No-Operation
141 instruction at the beginning of every kernel function, which NOP
142 sequence is then dynamically patched into a tracer call when
143 tracing is enabled by the administrator. If it's runtime disabled
144 (the bootup default), then the overhead of the instructions is very
145 small and not measurable even in micro-benchmarks.
146
147config FUNCTION_GRAPH_TRACER
148 bool "Kernel Function Graph Tracer"
149 depends on HAVE_FUNCTION_GRAPH_TRACER
150 depends on FUNCTION_TRACER
151 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
152 default y
153 help
154 Enable the kernel to trace a function at both its return
155 and its entry.
156 Its first purpose is to trace the duration of functions and
157 draw a call graph for each thread with some information like
158 the return value. This is done by setting the current return
159 address on the current task structure into a stack of calls.
160
161
162config IRQSOFF_TRACER
163 bool "Interrupts-off Latency Tracer"
164 default n
165 depends on TRACE_IRQFLAGS_SUPPORT
166 depends on !ARCH_USES_GETTIMEOFFSET
167 select TRACE_IRQFLAGS
168 select GENERIC_TRACER
169 select TRACER_MAX_TRACE
170 select RING_BUFFER_ALLOW_SWAP
171 select TRACER_SNAPSHOT
172 select TRACER_SNAPSHOT_PER_CPU_SWAP
173 help
174 This option measures the time spent in irqs-off critical
175 sections, with microsecond accuracy.
176
177 The default measurement method is a maximum search, which is
178 disabled by default and can be runtime (re-)started
179 via:
180
181 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
182
183 (Note that kernel size and overhead increase with this option
184 enabled. This option and the preempt-off timing option can be
185 used together or separately.)
186
187config PREEMPT_TRACER
188 bool "Preemption-off Latency Tracer"
189 default n
190 depends on !ARCH_USES_GETTIMEOFFSET
191 depends on PREEMPT
192 select GENERIC_TRACER
193 select TRACER_MAX_TRACE
194 select RING_BUFFER_ALLOW_SWAP
195 select TRACER_SNAPSHOT
196 select TRACER_SNAPSHOT_PER_CPU_SWAP
197 help
198 This option measures the time spent in preemption-off critical
199 sections, with microsecond accuracy.
200
201 The default measurement method is a maximum search, which is
202 disabled by default and can be runtime (re-)started
203 via:
204
205 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
206
207 (Note that kernel size and overhead increase with this option
208 enabled. This option and the irqs-off timing option can be
209 used together or separately.)
210
211config SCHED_TRACER
212 bool "Scheduling Latency Tracer"
213 select GENERIC_TRACER
214 select CONTEXT_SWITCH_TRACER
215 select TRACER_MAX_TRACE
216 select TRACER_SNAPSHOT
217 help
218 This tracer tracks the latency of the highest priority task
219 to be scheduled in, starting from the point it has woken up.
220
221config HWLAT_TRACER
222 bool "Tracer to detect hardware latencies (like SMIs)"
223 select GENERIC_TRACER
224 help
225 This tracer, when enabled will create one or more kernel threads,
226 depening on what the cpumask file is set to, which each thread
227 spinning in a loop looking for interruptions caused by
228 something other than the kernel. For example, if a
229 System Management Interrupt (SMI) takes a noticeable amount of
230 time, this tracer will detect it. This is useful for testing
231 if a system is reliable for Real Time tasks.
232
233 Some files are created in the tracing directory when this
234 is enabled:
235
236 hwlat_detector/width - time in usecs for how long to spin for
237 hwlat_detector/window - time in usecs between the start of each
238 iteration
239
240 A kernel thread is created that will spin with interrupts disabled
241 for "width" microseconds in every "widow" cycle. It will not spin
242 for "window - width" microseconds, where the system can
243 continue to operate.
244
245 The output will appear in the trace and trace_pipe files.
246
247 When the tracer is not running, it has no affect on the system,
248 but when it is running, it can cause the system to be
249 periodically non responsive. Do not run this tracer on a
250 production system.
251
252 To enable this tracer, echo in "hwlat" into the current_tracer
253 file. Every time a latency is greater than tracing_thresh, it will
254 be recorded into the ring buffer.
255
256config ENABLE_DEFAULT_TRACERS
257 bool "Trace process context switches and events"
258 depends on !GENERIC_TRACER
259 select TRACING
260 help
261 This tracer hooks to various trace points in the kernel,
262 allowing the user to pick and choose which trace point they
263 want to trace. It also includes the sched_switch tracer plugin.
264
265config FTRACE_SYSCALLS
266 bool "Trace syscalls"
267 depends on HAVE_SYSCALL_TRACEPOINTS
268 select GENERIC_TRACER
269 select KALLSYMS
270 help
271 Basic tracer to catch the syscall entry and exit events.
272
273config TRACER_SNAPSHOT
274 bool "Create a snapshot trace buffer"
275 select TRACER_MAX_TRACE
276 help
277 Allow tracing users to take snapshot of the current buffer using the
278 ftrace interface, e.g.:
279
280 echo 1 > /sys/kernel/debug/tracing/snapshot
281 cat snapshot
282
283config TRACER_SNAPSHOT_PER_CPU_SWAP
284 bool "Allow snapshot to swap per CPU"
285 depends on TRACER_SNAPSHOT
286 select RING_BUFFER_ALLOW_SWAP
287 help
288 Allow doing a snapshot of a single CPU buffer instead of a
289 full swap (all buffers). If this is set, then the following is
290 allowed:
291
292 echo 1 > /sys/kernel/debug/tracing/per_cpu/cpu2/snapshot
293
294 After which, only the tracing buffer for CPU 2 was swapped with
295 the main tracing buffer, and the other CPU buffers remain the same.
296
297 When this is enabled, this adds a little more overhead to the
298 trace recording, as it needs to add some checks to synchronize
299 recording with swaps. But this does not affect the performance
300 of the overall system. This is enabled by default when the preempt
301 or irq latency tracers are enabled, as those need to swap as well
302 and already adds the overhead (plus a lot more).
303
304config TRACE_BRANCH_PROFILING
305 bool
306 select GENERIC_TRACER
307
308choice
309 prompt "Branch Profiling"
310 default BRANCH_PROFILE_NONE
311 help
312 The branch profiling is a software profiler. It will add hooks
313 into the C conditionals to test which path a branch takes.
314
315 The likely/unlikely profiler only looks at the conditions that
316 are annotated with a likely or unlikely macro.
317
318 The "all branch" profiler will profile every if-statement in the
319 kernel. This profiler will also enable the likely/unlikely
320 profiler.
321
322 Either of the above profilers adds a bit of overhead to the system.
323 If unsure, choose "No branch profiling".
324
325config BRANCH_PROFILE_NONE
326 bool "No branch profiling"
327 help
328 No branch profiling. Branch profiling adds a bit of overhead.
329 Only enable it if you want to analyse the branching behavior.
330 Otherwise keep it disabled.
331
332config PROFILE_ANNOTATED_BRANCHES
333 bool "Trace likely/unlikely profiler"
334 select TRACE_BRANCH_PROFILING
335 help
336 This tracer profiles all likely and unlikely macros
337 in the kernel. It will display the results in:
338
339 /sys/kernel/debug/tracing/trace_stat/branch_annotated
340
341 Note: this will add a significant overhead; only turn this
342 on if you need to profile the system's use of these macros.
343
344config PROFILE_ALL_BRANCHES
345 bool "Profile all if conditionals"
346 select TRACE_BRANCH_PROFILING
347 help
348 This tracer profiles all branch conditions. Every if ()
349 taken in the kernel is recorded whether it hit or miss.
350 The results will be displayed in:
351
352 /sys/kernel/debug/tracing/trace_stat/branch_all
353
354 This option also enables the likely/unlikely profiler.
355
356 This configuration, when enabled, will impose a great overhead
357 on the system. This should only be enabled when the system
358 is to be analyzed in much detail.
359endchoice
360
361config TRACING_BRANCHES
362 bool
363 help
364 Selected by tracers that will trace the likely and unlikely
365 conditions. This prevents the tracers themselves from being
366 profiled. Profiling the tracing infrastructure can only happen
367 when the likelys and unlikelys are not being traced.
368
369config BRANCH_TRACER
370 bool "Trace likely/unlikely instances"
371 depends on TRACE_BRANCH_PROFILING
372 select TRACING_BRANCHES
373 help
374 This traces the events of likely and unlikely condition
375 calls in the kernel. The difference between this and the
376 "Trace likely/unlikely profiler" is that this is not a
377 histogram of the callers, but actually places the calling
378 events into a running trace buffer to see when and where the
379 events happened, as well as their results.
380
381 Say N if unsure.
382
383config STACK_TRACER
384 bool "Trace max stack"
385 depends on HAVE_FUNCTION_TRACER
386 select FUNCTION_TRACER
387 select STACKTRACE
388 select KALLSYMS
389 help
390 This special tracer records the maximum stack footprint of the
391 kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
392
393 This tracer works by hooking into every function call that the
394 kernel executes, and keeping a maximum stack depth value and
395 stack-trace saved. If this is configured with DYNAMIC_FTRACE
396 then it will not have any overhead while the stack tracer
397 is disabled.
398
399 To enable the stack tracer on bootup, pass in 'stacktrace'
400 on the kernel command line.
401
402 The stack tracer can also be enabled or disabled via the
403 sysctl kernel.stack_tracer_enabled
404
405 Say N if unsure.
406
407config BLK_DEV_IO_TRACE
408 bool "Support for tracing block IO actions"
409 depends on SYSFS
410 depends on BLOCK
411 select RELAY
412 select DEBUG_FS
413 select TRACEPOINTS
414 select GENERIC_TRACER
415 select STACKTRACE
416 help
417 Say Y here if you want to be able to trace the block layer actions
418 on a given queue. Tracing allows you to see any traffic happening
419 on a block device queue. For more information (and the userspace
420 support tools needed), fetch the blktrace tools from:
421
422 git://git.kernel.dk/blktrace.git
423
424 Tracing also is possible using the ftrace interface, e.g.:
425
426 echo 1 > /sys/block/sda/sda1/trace/enable
427 echo blk > /sys/kernel/debug/tracing/current_tracer
428 cat /sys/kernel/debug/tracing/trace_pipe
429
430 If unsure, say N.
431
432config KPROBE_EVENT
433 depends on KPROBES
434 depends on HAVE_REGS_AND_STACK_ACCESS_API
435 bool "Enable kprobes-based dynamic events"
436 select TRACING
437 select PROBE_EVENTS
438 default y
439 help
440 This allows the user to add tracing events (similar to tracepoints)
441 on the fly via the ftrace interface. See
442 Documentation/trace/kprobetrace.txt for more details.
443
444 Those events can be inserted wherever kprobes can probe, and record
445 various register and memory values.
446
447 This option is also required by perf-probe subcommand of perf tools.
448 If you want to use perf tools, this option is strongly recommended.
449
450config UPROBE_EVENT
451 bool "Enable uprobes-based dynamic events"
452 depends on ARCH_SUPPORTS_UPROBES
453 depends on MMU
454 depends on PERF_EVENTS
455 select UPROBES
456 select PROBE_EVENTS
457 select TRACING
458 default n
459 help
460 This allows the user to add tracing events on top of userspace
461 dynamic events (similar to tracepoints) on the fly via the trace
462 events interface. Those events can be inserted wherever uprobes
463 can probe, and record various registers.
464 This option is required if you plan to use perf-probe subcommand
465 of perf tools on user space applications.
466
467config BPF_EVENTS
468 depends on BPF_SYSCALL
469 depends on (KPROBE_EVENT || UPROBE_EVENT) && PERF_EVENTS
470 bool
471 default y
472 help
473 This allows the user to attach BPF programs to kprobe events.
474
475config PROBE_EVENTS
476 def_bool n
477
478config DYNAMIC_FTRACE
479 bool "enable/disable function tracing dynamically"
480 depends on FUNCTION_TRACER
481 depends on HAVE_DYNAMIC_FTRACE
482 default y
483 help
484 This option will modify all the calls to function tracing
485 dynamically (will patch them out of the binary image and
486 replace them with a No-Op instruction) on boot up. During
487 compile time, a table is made of all the locations that ftrace
488 can function trace, and this table is linked into the kernel
489 image. When this is enabled, functions can be individually
490 enabled, and the functions not enabled will not affect
491 performance of the system.
492
493 See the files in /sys/kernel/debug/tracing:
494 available_filter_functions
495 set_ftrace_filter
496 set_ftrace_notrace
497
498 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
499 otherwise has native performance as long as no tracing is active.
500
501config DYNAMIC_FTRACE_WITH_REGS
502 def_bool y
503 depends on DYNAMIC_FTRACE
504 depends on HAVE_DYNAMIC_FTRACE_WITH_REGS
505
506config FUNCTION_PROFILER
507 bool "Kernel function profiler"
508 depends on FUNCTION_TRACER
509 default n
510 help
511 This option enables the kernel function profiler. A file is created
512 in debugfs called function_profile_enabled which defaults to zero.
513 When a 1 is echoed into this file profiling begins, and when a
514 zero is entered, profiling stops. A "functions" file is created in
515 the trace_stats directory; this file shows the list of functions that
516 have been hit and their counters.
517
518 If in doubt, say N.
519
520config FTRACE_MCOUNT_RECORD
521 def_bool y
522 depends on DYNAMIC_FTRACE
523 depends on HAVE_FTRACE_MCOUNT_RECORD
524
525config FTRACE_SELFTEST
526 bool
527
528config FTRACE_STARTUP_TEST
529 bool "Perform a startup test on ftrace"
530 depends on GENERIC_TRACER
531 select FTRACE_SELFTEST
532 help
533 This option performs a series of startup tests on ftrace. On bootup
534 a series of tests are made to verify that the tracer is
535 functioning properly. It will do tests on all the configured
536 tracers of ftrace.
537
538config EVENT_TRACE_TEST_SYSCALLS
539 bool "Run selftest on syscall events"
540 depends on FTRACE_STARTUP_TEST
541 help
542 This option will also enable testing every syscall event.
543 It only enables the event and disables it and runs various loads
544 with the event enabled. This adds a bit more time for kernel boot
545 up since it runs this on every system call defined.
546
547 TBD - enable a way to actually call the syscalls as we test their
548 events
549
550config MMIOTRACE
551 bool "Memory mapped IO tracing"
552 depends on HAVE_MMIOTRACE_SUPPORT && PCI
553 select GENERIC_TRACER
554 help
555 Mmiotrace traces Memory Mapped I/O access and is meant for
556 debugging and reverse engineering. It is called from the ioremap
557 implementation and works via page faults. Tracing is disabled by
558 default and can be enabled at run-time.
559
560 See Documentation/trace/mmiotrace.txt.
561 If you are not helping to develop drivers, say N.
562
563config TRACING_MAP
564 bool
565 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG
566 help
567 tracing_map is a special-purpose lock-free map for tracing,
568 separated out as a stand-alone facility in order to allow it
569 to be shared between multiple tracers. It isn't meant to be
570 generally used outside of that context, and is normally
571 selected by tracers that use it.
572
573config HIST_TRIGGERS
574 bool "Histogram triggers"
575 depends on ARCH_HAVE_NMI_SAFE_CMPXCHG
576 select TRACING_MAP
577 select TRACING
578 default n
579 help
580 Hist triggers allow one or more arbitrary trace event fields
581 to be aggregated into hash tables and dumped to stdout by
582 reading a debugfs/tracefs file. They're useful for
583 gathering quick and dirty (though precise) summaries of
584 event activity as an initial guide for further investigation
585 using more advanced tools.
586
587 See Documentation/trace/events.txt.
588 If in doubt, say N.
589
590config MMIOTRACE_TEST
591 tristate "Test module for mmiotrace"
592 depends on MMIOTRACE && m
593 help
594 This is a dumb module for testing mmiotrace. It is very dangerous
595 as it will write garbage to IO memory starting at a given address.
596 However, it should be safe to use on e.g. unused portion of VRAM.
597
598 Say N, unless you absolutely know what you are doing.
599
600config TRACEPOINT_BENCHMARK
601 bool "Add tracepoint that benchmarks tracepoints"
602 help
603 This option creates the tracepoint "benchmark:benchmark_event".
604 When the tracepoint is enabled, it kicks off a kernel thread that
605 goes into an infinite loop (calling cond_sched() to let other tasks
606 run), and calls the tracepoint. Each iteration will record the time
607 it took to write to the tracepoint and the next iteration that
608 data will be passed to the tracepoint itself. That is, the tracepoint
609 will report the time it took to do the previous tracepoint.
610 The string written to the tracepoint is a static string of 128 bytes
611 to keep the time the same. The initial string is simply a write of
612 "START". The second string records the cold cache time of the first
613 write which is not added to the rest of the calculations.
614
615 As it is a tight loop, it benchmarks as hot cache. That's fine because
616 we care most about hot paths that are probably in cache already.
617
618 An example of the output:
619
620 START
621 first=3672 [COLD CACHED]
622 last=632 first=3672 max=632 min=632 avg=316 std=446 std^2=199712
623 last=278 first=3672 max=632 min=278 avg=303 std=316 std^2=100337
624 last=277 first=3672 max=632 min=277 avg=296 std=258 std^2=67064
625 last=273 first=3672 max=632 min=273 avg=292 std=224 std^2=50411
626 last=273 first=3672 max=632 min=273 avg=288 std=200 std^2=40389
627 last=281 first=3672 max=632 min=273 avg=287 std=183 std^2=33666
628
629
630config RING_BUFFER_BENCHMARK
631 tristate "Ring buffer benchmark stress tester"
632 depends on RING_BUFFER
633 help
634 This option creates a test to stress the ring buffer and benchmark it.
635 It creates its own ring buffer such that it will not interfere with
636 any other users of the ring buffer (such as ftrace). It then creates
637 a producer and consumer that will run for 10 seconds and sleep for
638 10 seconds. Each interval it will print out the number of events
639 it recorded and give a rough estimate of how long each iteration took.
640
641 It does not disable interrupts or raise its priority, so it may be
642 affected by processes that are running.
643
644 If unsure, say N.
645
646config RING_BUFFER_STARTUP_TEST
647 bool "Ring buffer startup self test"
648 depends on RING_BUFFER
649 help
650 Run a simple self test on the ring buffer on boot up. Late in the
651 kernel boot sequence, the test will start that kicks off
652 a thread per cpu. Each thread will write various size events
653 into the ring buffer. Another thread is created to send IPIs
654 to each of the threads, where the IPI handler will also write
655 to the ring buffer, to test/stress the nesting ability.
656 If any anomalies are discovered, a warning will be displayed
657 and all ring buffers will be disabled.
658
659 The test runs for 10 seconds. This will slow your boot time
660 by at least 10 more seconds.
661
662 At the end of the test, statics and more checks are done.
663 It will output the stats of each per cpu buffer. What
664 was written, the sizes, what was read, what was lost, and
665 other similar details.
666
667 If unsure, say N
668
669config TRACE_ENUM_MAP_FILE
670 bool "Show enum mappings for trace events"
671 depends on TRACING
672 help
673 The "print fmt" of the trace events will show the enum names instead
674 of their values. This can cause problems for user space tools that
675 use this string to parse the raw data as user space does not know
676 how to convert the string to its value.
677
678 To fix this, there's a special macro in the kernel that can be used
679 to convert the enum into its value. If this macro is used, then the
680 print fmt strings will have the enums converted to their values.
681
682 If something does not get converted properly, this option can be
683 used to show what enums the kernel tried to convert.
684
685 This option is for debugging the enum conversions. A file is created
686 in the tracing directory called "enum_map" that will show the enum
687 names matched with their values and what trace event system they
688 belong too.
689
690 Normally, the mapping of the strings to values will be freed after
691 boot up or module load. With this option, they will not be freed, as
692 they are needed for the "enum_map" file. Enabling this option will
693 increase the memory footprint of the running kernel.
694
695 If unsure, say N
696
697config TRACING_EVENTS_GPIO
698 bool "Trace gpio events"
699 depends on GPIOLIB
700 default y
701 help
702 Enable tracing events for gpio subsystem
703
704endif # FTRACE
705
706endif # TRACING_SUPPORT
707
1#
2# Architectures that offer an FUNCTION_TRACER implementation should
3# select HAVE_FUNCTION_TRACER:
4#
5
6config USER_STACKTRACE_SUPPORT
7 bool
8
9config NOP_TRACER
10 bool
11
12config HAVE_FTRACE_NMI_ENTER
13 bool
14 help
15 See Documentation/trace/ftrace-design.txt
16
17config HAVE_FUNCTION_TRACER
18 bool
19 help
20 See Documentation/trace/ftrace-design.txt
21
22config HAVE_FUNCTION_GRAPH_TRACER
23 bool
24 help
25 See Documentation/trace/ftrace-design.txt
26
27config HAVE_FUNCTION_GRAPH_FP_TEST
28 bool
29 help
30 See Documentation/trace/ftrace-design.txt
31
32config HAVE_DYNAMIC_FTRACE
33 bool
34 help
35 See Documentation/trace/ftrace-design.txt
36
37config HAVE_DYNAMIC_FTRACE_WITH_REGS
38 bool
39
40config HAVE_FTRACE_MCOUNT_RECORD
41 bool
42 help
43 See Documentation/trace/ftrace-design.txt
44
45config HAVE_SYSCALL_TRACEPOINTS
46 bool
47 help
48 See Documentation/trace/ftrace-design.txt
49
50config HAVE_FENTRY
51 bool
52 help
53 Arch supports the gcc options -pg with -mfentry
54
55config HAVE_C_RECORDMCOUNT
56 bool
57 help
58 C version of recordmcount available?
59
60config TRACER_MAX_TRACE
61 bool
62
63config TRACE_CLOCK
64 bool
65
66config RING_BUFFER
67 bool
68 select TRACE_CLOCK
69 select IRQ_WORK
70
71config FTRACE_NMI_ENTER
72 bool
73 depends on HAVE_FTRACE_NMI_ENTER
74 default y
75
76config EVENT_TRACING
77 select CONTEXT_SWITCH_TRACER
78 bool
79
80config CONTEXT_SWITCH_TRACER
81 bool
82
83config RING_BUFFER_ALLOW_SWAP
84 bool
85 help
86 Allow the use of ring_buffer_swap_cpu.
87 Adds a very slight overhead to tracing when enabled.
88
89# All tracer options should select GENERIC_TRACER. For those options that are
90# enabled by all tracers (context switch and event tracer) they select TRACING.
91# This allows those options to appear when no other tracer is selected. But the
92# options do not appear when something else selects it. We need the two options
93# GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
94# hiding of the automatic options.
95
96config TRACING
97 bool
98 select DEBUG_FS
99 select RING_BUFFER
100 select STACKTRACE if STACKTRACE_SUPPORT
101 select TRACEPOINTS
102 select NOP_TRACER
103 select BINARY_PRINTF
104 select EVENT_TRACING
105 select TRACE_CLOCK
106
107config GENERIC_TRACER
108 bool
109 select TRACING
110
111#
112# Minimum requirements an architecture has to meet for us to
113# be able to offer generic tracing facilities:
114#
115config TRACING_SUPPORT
116 bool
117 # PPC32 has no irqflags tracing support, but it can use most of the
118 # tracers anyway, they were tested to build and work. Note that new
119 # exceptions to this list aren't welcomed, better implement the
120 # irqflags tracing for your architecture.
121 depends on TRACE_IRQFLAGS_SUPPORT || PPC32
122 depends on STACKTRACE_SUPPORT
123 default y
124
125if TRACING_SUPPORT
126
127menuconfig FTRACE
128 bool "Tracers"
129 default y if DEBUG_KERNEL
130 help
131 Enable the kernel tracing infrastructure.
132
133if FTRACE
134
135config FUNCTION_TRACER
136 bool "Kernel Function Tracer"
137 depends on HAVE_FUNCTION_TRACER
138 select KALLSYMS
139 select GENERIC_TRACER
140 select CONTEXT_SWITCH_TRACER
141 help
142 Enable the kernel to trace every kernel function. This is done
143 by using a compiler feature to insert a small, 5-byte No-Operation
144 instruction at the beginning of every kernel function, which NOP
145 sequence is then dynamically patched into a tracer call when
146 tracing is enabled by the administrator. If it's runtime disabled
147 (the bootup default), then the overhead of the instructions is very
148 small and not measurable even in micro-benchmarks.
149
150config FUNCTION_GRAPH_TRACER
151 bool "Kernel Function Graph Tracer"
152 depends on HAVE_FUNCTION_GRAPH_TRACER
153 depends on FUNCTION_TRACER
154 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
155 default y
156 help
157 Enable the kernel to trace a function at both its return
158 and its entry.
159 Its first purpose is to trace the duration of functions and
160 draw a call graph for each thread with some information like
161 the return value. This is done by setting the current return
162 address on the current task structure into a stack of calls.
163
164
165config IRQSOFF_TRACER
166 bool "Interrupts-off Latency Tracer"
167 default n
168 depends on TRACE_IRQFLAGS_SUPPORT
169 depends on !ARCH_USES_GETTIMEOFFSET
170 select TRACE_IRQFLAGS
171 select GENERIC_TRACER
172 select TRACER_MAX_TRACE
173 select RING_BUFFER_ALLOW_SWAP
174 select TRACER_SNAPSHOT
175 select TRACER_SNAPSHOT_PER_CPU_SWAP
176 help
177 This option measures the time spent in irqs-off critical
178 sections, with microsecond accuracy.
179
180 The default measurement method is a maximum search, which is
181 disabled by default and can be runtime (re-)started
182 via:
183
184 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
185
186 (Note that kernel size and overhead increase with this option
187 enabled. This option and the preempt-off timing option can be
188 used together or separately.)
189
190config PREEMPT_TRACER
191 bool "Preemption-off Latency Tracer"
192 default n
193 depends on !ARCH_USES_GETTIMEOFFSET
194 depends on PREEMPT
195 select GENERIC_TRACER
196 select TRACER_MAX_TRACE
197 select RING_BUFFER_ALLOW_SWAP
198 select TRACER_SNAPSHOT
199 select TRACER_SNAPSHOT_PER_CPU_SWAP
200 help
201 This option measures the time spent in preemption-off critical
202 sections, with microsecond accuracy.
203
204 The default measurement method is a maximum search, which is
205 disabled by default and can be runtime (re-)started
206 via:
207
208 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
209
210 (Note that kernel size and overhead increase with this option
211 enabled. This option and the irqs-off timing option can be
212 used together or separately.)
213
214config SCHED_TRACER
215 bool "Scheduling Latency Tracer"
216 select GENERIC_TRACER
217 select CONTEXT_SWITCH_TRACER
218 select TRACER_MAX_TRACE
219 select TRACER_SNAPSHOT
220 help
221 This tracer tracks the latency of the highest priority task
222 to be scheduled in, starting from the point it has woken up.
223
224config ENABLE_DEFAULT_TRACERS
225 bool "Trace process context switches and events"
226 depends on !GENERIC_TRACER
227 select TRACING
228 help
229 This tracer hooks to various trace points in the kernel,
230 allowing the user to pick and choose which trace point they
231 want to trace. It also includes the sched_switch tracer plugin.
232
233config FTRACE_SYSCALLS
234 bool "Trace syscalls"
235 depends on HAVE_SYSCALL_TRACEPOINTS
236 select GENERIC_TRACER
237 select KALLSYMS
238 help
239 Basic tracer to catch the syscall entry and exit events.
240
241config TRACER_SNAPSHOT
242 bool "Create a snapshot trace buffer"
243 select TRACER_MAX_TRACE
244 help
245 Allow tracing users to take snapshot of the current buffer using the
246 ftrace interface, e.g.:
247
248 echo 1 > /sys/kernel/debug/tracing/snapshot
249 cat snapshot
250
251config TRACER_SNAPSHOT_PER_CPU_SWAP
252 bool "Allow snapshot to swap per CPU"
253 depends on TRACER_SNAPSHOT
254 select RING_BUFFER_ALLOW_SWAP
255 help
256 Allow doing a snapshot of a single CPU buffer instead of a
257 full swap (all buffers). If this is set, then the following is
258 allowed:
259
260 echo 1 > /sys/kernel/debug/tracing/per_cpu/cpu2/snapshot
261
262 After which, only the tracing buffer for CPU 2 was swapped with
263 the main tracing buffer, and the other CPU buffers remain the same.
264
265 When this is enabled, this adds a little more overhead to the
266 trace recording, as it needs to add some checks to synchronize
267 recording with swaps. But this does not affect the performance
268 of the overall system. This is enabled by default when the preempt
269 or irq latency tracers are enabled, as those need to swap as well
270 and already adds the overhead (plus a lot more).
271
272config TRACE_BRANCH_PROFILING
273 bool
274 select GENERIC_TRACER
275
276choice
277 prompt "Branch Profiling"
278 default BRANCH_PROFILE_NONE
279 help
280 The branch profiling is a software profiler. It will add hooks
281 into the C conditionals to test which path a branch takes.
282
283 The likely/unlikely profiler only looks at the conditions that
284 are annotated with a likely or unlikely macro.
285
286 The "all branch" profiler will profile every if-statement in the
287 kernel. This profiler will also enable the likely/unlikely
288 profiler.
289
290 Either of the above profilers adds a bit of overhead to the system.
291 If unsure, choose "No branch profiling".
292
293config BRANCH_PROFILE_NONE
294 bool "No branch profiling"
295 help
296 No branch profiling. Branch profiling adds a bit of overhead.
297 Only enable it if you want to analyse the branching behavior.
298 Otherwise keep it disabled.
299
300config PROFILE_ANNOTATED_BRANCHES
301 bool "Trace likely/unlikely profiler"
302 select TRACE_BRANCH_PROFILING
303 help
304 This tracer profiles all likely and unlikely macros
305 in the kernel. It will display the results in:
306
307 /sys/kernel/debug/tracing/trace_stat/branch_annotated
308
309 Note: this will add a significant overhead; only turn this
310 on if you need to profile the system's use of these macros.
311
312config PROFILE_ALL_BRANCHES
313 bool "Profile all if conditionals"
314 select TRACE_BRANCH_PROFILING
315 help
316 This tracer profiles all branch conditions. Every if ()
317 taken in the kernel is recorded whether it hit or miss.
318 The results will be displayed in:
319
320 /sys/kernel/debug/tracing/trace_stat/branch_all
321
322 This option also enables the likely/unlikely profiler.
323
324 This configuration, when enabled, will impose a great overhead
325 on the system. This should only be enabled when the system
326 is to be analyzed in much detail.
327endchoice
328
329config TRACING_BRANCHES
330 bool
331 help
332 Selected by tracers that will trace the likely and unlikely
333 conditions. This prevents the tracers themselves from being
334 profiled. Profiling the tracing infrastructure can only happen
335 when the likelys and unlikelys are not being traced.
336
337config BRANCH_TRACER
338 bool "Trace likely/unlikely instances"
339 depends on TRACE_BRANCH_PROFILING
340 select TRACING_BRANCHES
341 help
342 This traces the events of likely and unlikely condition
343 calls in the kernel. The difference between this and the
344 "Trace likely/unlikely profiler" is that this is not a
345 histogram of the callers, but actually places the calling
346 events into a running trace buffer to see when and where the
347 events happened, as well as their results.
348
349 Say N if unsure.
350
351config STACK_TRACER
352 bool "Trace max stack"
353 depends on HAVE_FUNCTION_TRACER
354 select FUNCTION_TRACER
355 select STACKTRACE
356 select KALLSYMS
357 help
358 This special tracer records the maximum stack footprint of the
359 kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
360
361 This tracer works by hooking into every function call that the
362 kernel executes, and keeping a maximum stack depth value and
363 stack-trace saved. If this is configured with DYNAMIC_FTRACE
364 then it will not have any overhead while the stack tracer
365 is disabled.
366
367 To enable the stack tracer on bootup, pass in 'stacktrace'
368 on the kernel command line.
369
370 The stack tracer can also be enabled or disabled via the
371 sysctl kernel.stack_tracer_enabled
372
373 Say N if unsure.
374
375config BLK_DEV_IO_TRACE
376 bool "Support for tracing block IO actions"
377 depends on SYSFS
378 depends on BLOCK
379 select RELAY
380 select DEBUG_FS
381 select TRACEPOINTS
382 select GENERIC_TRACER
383 select STACKTRACE
384 help
385 Say Y here if you want to be able to trace the block layer actions
386 on a given queue. Tracing allows you to see any traffic happening
387 on a block device queue. For more information (and the userspace
388 support tools needed), fetch the blktrace tools from:
389
390 git://git.kernel.dk/blktrace.git
391
392 Tracing also is possible using the ftrace interface, e.g.:
393
394 echo 1 > /sys/block/sda/sda1/trace/enable
395 echo blk > /sys/kernel/debug/tracing/current_tracer
396 cat /sys/kernel/debug/tracing/trace_pipe
397
398 If unsure, say N.
399
400config KPROBE_EVENT
401 depends on KPROBES
402 depends on HAVE_REGS_AND_STACK_ACCESS_API
403 bool "Enable kprobes-based dynamic events"
404 select TRACING
405 select PROBE_EVENTS
406 default y
407 help
408 This allows the user to add tracing events (similar to tracepoints)
409 on the fly via the ftrace interface. See
410 Documentation/trace/kprobetrace.txt for more details.
411
412 Those events can be inserted wherever kprobes can probe, and record
413 various register and memory values.
414
415 This option is also required by perf-probe subcommand of perf tools.
416 If you want to use perf tools, this option is strongly recommended.
417
418config UPROBE_EVENT
419 bool "Enable uprobes-based dynamic events"
420 depends on ARCH_SUPPORTS_UPROBES
421 depends on MMU
422 depends on PERF_EVENTS
423 select UPROBES
424 select PROBE_EVENTS
425 select TRACING
426 default n
427 help
428 This allows the user to add tracing events on top of userspace
429 dynamic events (similar to tracepoints) on the fly via the trace
430 events interface. Those events can be inserted wherever uprobes
431 can probe, and record various registers.
432 This option is required if you plan to use perf-probe subcommand
433 of perf tools on user space applications.
434
435config BPF_EVENTS
436 depends on BPF_SYSCALL
437 depends on (KPROBE_EVENT || UPROBE_EVENT) && PERF_EVENTS
438 bool
439 default y
440 help
441 This allows the user to attach BPF programs to kprobe events.
442
443config PROBE_EVENTS
444 def_bool n
445
446config DYNAMIC_FTRACE
447 bool "enable/disable function tracing dynamically"
448 depends on FUNCTION_TRACER
449 depends on HAVE_DYNAMIC_FTRACE
450 default y
451 help
452 This option will modify all the calls to function tracing
453 dynamically (will patch them out of the binary image and
454 replace them with a No-Op instruction) on boot up. During
455 compile time, a table is made of all the locations that ftrace
456 can function trace, and this table is linked into the kernel
457 image. When this is enabled, functions can be individually
458 enabled, and the functions not enabled will not affect
459 performance of the system.
460
461 See the files in /sys/kernel/debug/tracing:
462 available_filter_functions
463 set_ftrace_filter
464 set_ftrace_notrace
465
466 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
467 otherwise has native performance as long as no tracing is active.
468
469config DYNAMIC_FTRACE_WITH_REGS
470 def_bool y
471 depends on DYNAMIC_FTRACE
472 depends on HAVE_DYNAMIC_FTRACE_WITH_REGS
473
474config FUNCTION_PROFILER
475 bool "Kernel function profiler"
476 depends on FUNCTION_TRACER
477 default n
478 help
479 This option enables the kernel function profiler. A file is created
480 in debugfs called function_profile_enabled which defaults to zero.
481 When a 1 is echoed into this file profiling begins, and when a
482 zero is entered, profiling stops. A "functions" file is created in
483 the trace_stats directory; this file shows the list of functions that
484 have been hit and their counters.
485
486 If in doubt, say N.
487
488config FTRACE_MCOUNT_RECORD
489 def_bool y
490 depends on DYNAMIC_FTRACE
491 depends on HAVE_FTRACE_MCOUNT_RECORD
492
493config FTRACE_SELFTEST
494 bool
495
496config FTRACE_STARTUP_TEST
497 bool "Perform a startup test on ftrace"
498 depends on GENERIC_TRACER
499 select FTRACE_SELFTEST
500 help
501 This option performs a series of startup tests on ftrace. On bootup
502 a series of tests are made to verify that the tracer is
503 functioning properly. It will do tests on all the configured
504 tracers of ftrace.
505
506config EVENT_TRACE_TEST_SYSCALLS
507 bool "Run selftest on syscall events"
508 depends on FTRACE_STARTUP_TEST
509 help
510 This option will also enable testing every syscall event.
511 It only enables the event and disables it and runs various loads
512 with the event enabled. This adds a bit more time for kernel boot
513 up since it runs this on every system call defined.
514
515 TBD - enable a way to actually call the syscalls as we test their
516 events
517
518config MMIOTRACE
519 bool "Memory mapped IO tracing"
520 depends on HAVE_MMIOTRACE_SUPPORT && PCI
521 select GENERIC_TRACER
522 help
523 Mmiotrace traces Memory Mapped I/O access and is meant for
524 debugging and reverse engineering. It is called from the ioremap
525 implementation and works via page faults. Tracing is disabled by
526 default and can be enabled at run-time.
527
528 See Documentation/trace/mmiotrace.txt.
529 If you are not helping to develop drivers, say N.
530
531config MMIOTRACE_TEST
532 tristate "Test module for mmiotrace"
533 depends on MMIOTRACE && m
534 help
535 This is a dumb module for testing mmiotrace. It is very dangerous
536 as it will write garbage to IO memory starting at a given address.
537 However, it should be safe to use on e.g. unused portion of VRAM.
538
539 Say N, unless you absolutely know what you are doing.
540
541config TRACEPOINT_BENCHMARK
542 bool "Add tracepoint that benchmarks tracepoints"
543 help
544 This option creates the tracepoint "benchmark:benchmark_event".
545 When the tracepoint is enabled, it kicks off a kernel thread that
546 goes into an infinite loop (calling cond_sched() to let other tasks
547 run), and calls the tracepoint. Each iteration will record the time
548 it took to write to the tracepoint and the next iteration that
549 data will be passed to the tracepoint itself. That is, the tracepoint
550 will report the time it took to do the previous tracepoint.
551 The string written to the tracepoint is a static string of 128 bytes
552 to keep the time the same. The initial string is simply a write of
553 "START". The second string records the cold cache time of the first
554 write which is not added to the rest of the calculations.
555
556 As it is a tight loop, it benchmarks as hot cache. That's fine because
557 we care most about hot paths that are probably in cache already.
558
559 An example of the output:
560
561 START
562 first=3672 [COLD CACHED]
563 last=632 first=3672 max=632 min=632 avg=316 std=446 std^2=199712
564 last=278 first=3672 max=632 min=278 avg=303 std=316 std^2=100337
565 last=277 first=3672 max=632 min=277 avg=296 std=258 std^2=67064
566 last=273 first=3672 max=632 min=273 avg=292 std=224 std^2=50411
567 last=273 first=3672 max=632 min=273 avg=288 std=200 std^2=40389
568 last=281 first=3672 max=632 min=273 avg=287 std=183 std^2=33666
569
570
571config RING_BUFFER_BENCHMARK
572 tristate "Ring buffer benchmark stress tester"
573 depends on RING_BUFFER
574 help
575 This option creates a test to stress the ring buffer and benchmark it.
576 It creates its own ring buffer such that it will not interfere with
577 any other users of the ring buffer (such as ftrace). It then creates
578 a producer and consumer that will run for 10 seconds and sleep for
579 10 seconds. Each interval it will print out the number of events
580 it recorded and give a rough estimate of how long each iteration took.
581
582 It does not disable interrupts or raise its priority, so it may be
583 affected by processes that are running.
584
585 If unsure, say N.
586
587config RING_BUFFER_STARTUP_TEST
588 bool "Ring buffer startup self test"
589 depends on RING_BUFFER
590 help
591 Run a simple self test on the ring buffer on boot up. Late in the
592 kernel boot sequence, the test will start that kicks off
593 a thread per cpu. Each thread will write various size events
594 into the ring buffer. Another thread is created to send IPIs
595 to each of the threads, where the IPI handler will also write
596 to the ring buffer, to test/stress the nesting ability.
597 If any anomalies are discovered, a warning will be displayed
598 and all ring buffers will be disabled.
599
600 The test runs for 10 seconds. This will slow your boot time
601 by at least 10 more seconds.
602
603 At the end of the test, statics and more checks are done.
604 It will output the stats of each per cpu buffer. What
605 was written, the sizes, what was read, what was lost, and
606 other similar details.
607
608 If unsure, say N
609
610config TRACE_ENUM_MAP_FILE
611 bool "Show enum mappings for trace events"
612 depends on TRACING
613 help
614 The "print fmt" of the trace events will show the enum names instead
615 of their values. This can cause problems for user space tools that
616 use this string to parse the raw data as user space does not know
617 how to convert the string to its value.
618
619 To fix this, there's a special macro in the kernel that can be used
620 to convert the enum into its value. If this macro is used, then the
621 print fmt strings will have the enums converted to their values.
622
623 If something does not get converted properly, this option can be
624 used to show what enums the kernel tried to convert.
625
626 This option is for debugging the enum conversions. A file is created
627 in the tracing directory called "enum_map" that will show the enum
628 names matched with their values and what trace event system they
629 belong too.
630
631 Normally, the mapping of the strings to values will be freed after
632 boot up or module load. With this option, they will not be freed, as
633 they are needed for the "enum_map" file. Enabling this option will
634 increase the memory footprint of the running kernel.
635
636 If unsure, say N
637
638config TRACING_EVENTS_GPIO
639 bool "Trace gpio events"
640 depends on GPIOLIB
641 default y
642 help
643 Enable tracing events for gpio subsystem
644
645endif # FTRACE
646
647endif # TRACING_SUPPORT
648