2 # Architectures that offer an FUNCTION_TRACER implementation should
3 # select HAVE_FUNCTION_TRACER:
6 config USER_STACKTRACE_SUPPORT
12 config HAVE_FTRACE_NMI_ENTER
15 See Documentation/trace/ftrace-design.txt
17 config HAVE_FUNCTION_TRACER
20 See Documentation/trace/ftrace-design.txt
22 config HAVE_FUNCTION_GRAPH_TRACER
25 See Documentation/trace/ftrace-design.txt
27 config HAVE_FUNCTION_GRAPH_FP_TEST
30 See Documentation/trace/ftrace-design.txt
32 config HAVE_FUNCTION_TRACE_MCOUNT_TEST
35 See Documentation/trace/ftrace-design.txt
37 config HAVE_DYNAMIC_FTRACE
40 See Documentation/trace/ftrace-design.txt
42 config HAVE_FTRACE_MCOUNT_RECORD
45 See Documentation/trace/ftrace-design.txt
47 config HAVE_SYSCALL_TRACEPOINTS
50 See Documentation/trace/ftrace-design.txt
55 Arch supports the gcc options -pg with -mfentry
57 config HAVE_C_RECORDMCOUNT
60 C version of recordmcount available?
62 config TRACER_MAX_TRACE
72 config FTRACE_NMI_ENTER
74 depends on HAVE_FTRACE_NMI_ENTER
78 select CONTEXT_SWITCH_TRACER
81 config EVENT_POWER_TRACING_DEPRECATED
82 depends on EVENT_TRACING
83 bool "Deprecated power event trace API, to be removed"
86 Provides old power event types:
87 C-state/idle accounting events:
90 and old cpufreq accounting event:
92 This is for userspace compatibility
93 and will vanish after 5 kernel iterations,
96 config CONTEXT_SWITCH_TRACER
99 config RING_BUFFER_ALLOW_SWAP
102 Allow the use of ring_buffer_swap_cpu.
103 Adds a very slight overhead to tracing when enabled.
105 # All tracer options should select GENERIC_TRACER. For those options that are
106 # enabled by all tracers (context switch and event tracer) they select TRACING.
107 # This allows those options to appear when no other tracer is selected. But the
108 # options do not appear when something else selects it. We need the two options
109 # GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
110 # hiding of the automatic options.
116 select STACKTRACE if STACKTRACE_SUPPORT
124 config GENERIC_TRACER
129 # Minimum requirements an architecture has to meet for us to
130 # be able to offer generic tracing facilities:
132 config TRACING_SUPPORT
134 # PPC32 has no irqflags tracing support, but it can use most of the
135 # tracers anyway, they were tested to build and work. Note that new
136 # exceptions to this list aren't welcomed, better implement the
137 # irqflags tracing for your architecture.
138 depends on TRACE_IRQFLAGS_SUPPORT || PPC32
139 depends on STACKTRACE_SUPPORT
146 default y if DEBUG_KERNEL
148 Enable the kernel tracing infrastructure.
152 config FUNCTION_TRACER
153 bool "Kernel Function Tracer"
154 depends on HAVE_FUNCTION_TRACER
156 select GENERIC_TRACER
157 select CONTEXT_SWITCH_TRACER
159 Enable the kernel to trace every kernel function. This is done
160 by using a compiler feature to insert a small, 5-byte No-Operation
161 instruction at the beginning of every kernel function, which NOP
162 sequence is then dynamically patched into a tracer call when
163 tracing is enabled by the administrator. If it's runtime disabled
164 (the bootup default), then the overhead of the instructions is very
165 small and not measurable even in micro-benchmarks.
167 config FUNCTION_GRAPH_TRACER
168 bool "Kernel Function Graph Tracer"
169 depends on HAVE_FUNCTION_GRAPH_TRACER
170 depends on FUNCTION_TRACER
171 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
174 Enable the kernel to trace a function at both its return
176 Its first purpose is to trace the duration of functions and
177 draw a call graph for each thread with some information like
178 the return value. This is done by setting the current return
179 address on the current task structure into a stack of calls.
182 config IRQSOFF_TRACER
183 bool "Interrupts-off Latency Tracer"
185 depends on TRACE_IRQFLAGS_SUPPORT
186 depends on !ARCH_USES_GETTIMEOFFSET
187 select TRACE_IRQFLAGS
188 select GENERIC_TRACER
189 select TRACER_MAX_TRACE
190 select RING_BUFFER_ALLOW_SWAP
192 This option measures the time spent in irqs-off critical
193 sections, with microsecond accuracy.
195 The default measurement method is a maximum search, which is
196 disabled by default and can be runtime (re-)started
199 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
201 (Note that kernel size and overhead increase with this option
202 enabled. This option and the preempt-off timing option can be
203 used together or separately.)
205 config PREEMPT_TRACER
206 bool "Preemption-off Latency Tracer"
208 depends on !ARCH_USES_GETTIMEOFFSET
210 select GENERIC_TRACER
211 select TRACER_MAX_TRACE
212 select RING_BUFFER_ALLOW_SWAP
214 This option measures the time spent in preemption-off critical
215 sections, with microsecond accuracy.
217 The default measurement method is a maximum search, which is
218 disabled by default and can be runtime (re-)started
221 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
223 (Note that kernel size and overhead increase with this option
224 enabled. This option and the irqs-off timing option can be
225 used together or separately.)
228 bool "Scheduling Latency Tracer"
229 select GENERIC_TRACER
230 select CONTEXT_SWITCH_TRACER
231 select TRACER_MAX_TRACE
233 This tracer tracks the latency of the highest priority task
234 to be scheduled in, starting from the point it has woken up.
236 config ENABLE_DEFAULT_TRACERS
237 bool "Trace process context switches and events"
238 depends on !GENERIC_TRACER
241 This tracer hooks to various trace points in the kernel,
242 allowing the user to pick and choose which trace point they
243 want to trace. It also includes the sched_switch tracer plugin.
245 config FTRACE_SYSCALLS
246 bool "Trace syscalls"
247 depends on HAVE_SYSCALL_TRACEPOINTS
248 select GENERIC_TRACER
251 Basic tracer to catch the syscall entry and exit events.
253 config TRACE_BRANCH_PROFILING
255 select GENERIC_TRACER
258 prompt "Branch Profiling"
259 default BRANCH_PROFILE_NONE
261 The branch profiling is a software profiler. It will add hooks
262 into the C conditionals to test which path a branch takes.
264 The likely/unlikely profiler only looks at the conditions that
265 are annotated with a likely or unlikely macro.
267 The "all branch" profiler will profile every if-statement in the
268 kernel. This profiler will also enable the likely/unlikely
271 Either of the above profilers adds a bit of overhead to the system.
272 If unsure, choose "No branch profiling".
274 config BRANCH_PROFILE_NONE
275 bool "No branch profiling"
277 No branch profiling. Branch profiling adds a bit of overhead.
278 Only enable it if you want to analyse the branching behavior.
279 Otherwise keep it disabled.
281 config PROFILE_ANNOTATED_BRANCHES
282 bool "Trace likely/unlikely profiler"
283 select TRACE_BRANCH_PROFILING
285 This tracer profiles all likely and unlikely macros
286 in the kernel. It will display the results in:
288 /sys/kernel/debug/tracing/trace_stat/branch_annotated
290 Note: this will add a significant overhead; only turn this
291 on if you need to profile the system's use of these macros.
293 config PROFILE_ALL_BRANCHES
294 bool "Profile all if conditionals"
295 select TRACE_BRANCH_PROFILING
297 This tracer profiles all branch conditions. Every if ()
298 taken in the kernel is recorded whether it hit or miss.
299 The results will be displayed in:
301 /sys/kernel/debug/tracing/trace_stat/branch_all
303 This option also enables the likely/unlikely profiler.
305 This configuration, when enabled, will impose a great overhead
306 on the system. This should only be enabled when the system
307 is to be analyzed in much detail.
310 config TRACING_BRANCHES
313 Selected by tracers that will trace the likely and unlikely
314 conditions. This prevents the tracers themselves from being
315 profiled. Profiling the tracing infrastructure can only happen
316 when the likelys and unlikelys are not being traced.
319 bool "Trace likely/unlikely instances"
320 depends on TRACE_BRANCH_PROFILING
321 select TRACING_BRANCHES
323 This traces the events of likely and unlikely condition
324 calls in the kernel. The difference between this and the
325 "Trace likely/unlikely profiler" is that this is not a
326 histogram of the callers, but actually places the calling
327 events into a running trace buffer to see when and where the
328 events happened, as well as their results.
333 bool "Trace max stack"
334 depends on HAVE_FUNCTION_TRACER
335 select FUNCTION_TRACER
339 This special tracer records the maximum stack footprint of the
340 kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
342 This tracer works by hooking into every function call that the
343 kernel executes, and keeping a maximum stack depth value and
344 stack-trace saved. If this is configured with DYNAMIC_FTRACE
345 then it will not have any overhead while the stack tracer
348 To enable the stack tracer on bootup, pass in 'stacktrace'
349 on the kernel command line.
351 The stack tracer can also be enabled or disabled via the
352 sysctl kernel.stack_tracer_enabled
356 config BLK_DEV_IO_TRACE
357 bool "Support for tracing block IO actions"
363 select GENERIC_TRACER
366 Say Y here if you want to be able to trace the block layer actions
367 on a given queue. Tracing allows you to see any traffic happening
368 on a block device queue. For more information (and the userspace
369 support tools needed), fetch the blktrace tools from:
371 git://git.kernel.dk/blktrace.git
373 Tracing also is possible using the ftrace interface, e.g.:
375 echo 1 > /sys/block/sda/sda1/trace/enable
376 echo blk > /sys/kernel/debug/tracing/current_tracer
377 cat /sys/kernel/debug/tracing/trace_pipe
383 depends on HAVE_REGS_AND_STACK_ACCESS_API
384 bool "Enable kprobes-based dynamic events"
389 This allows the user to add tracing events (similar to tracepoints)
390 on the fly via the ftrace interface. See
391 Documentation/trace/kprobetrace.txt for more details.
393 Those events can be inserted wherever kprobes can probe, and record
394 various register and memory values.
396 This option is also required by perf-probe subcommand of perf tools.
397 If you want to use perf tools, this option is strongly recommended.
400 bool "Enable uprobes-based dynamic events"
401 depends on ARCH_SUPPORTS_UPROBES
408 This allows the user to add tracing events on top of userspace
409 dynamic events (similar to tracepoints) on the fly via the trace
410 events interface. Those events can be inserted wherever uprobes
411 can probe, and record various registers.
412 This option is required if you plan to use perf-probe subcommand
413 of perf tools on user space applications.
418 config DYNAMIC_FTRACE
419 bool "enable/disable ftrace tracepoints dynamically"
420 depends on FUNCTION_TRACER
421 depends on HAVE_DYNAMIC_FTRACE
424 This option will modify all the calls to ftrace dynamically
425 (will patch them out of the binary image and replace them
426 with a No-Op instruction) as they are called. A table is
427 created to dynamically enable them again.
429 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
430 otherwise has native performance as long as no tracing is active.
432 The changes to the code are done by a kernel thread that
433 wakes up once a second and checks to see if any ftrace calls
434 were made. If so, it runs stop_machine (stops all CPUS)
435 and modifies the code to jump over the call to ftrace.
437 config FUNCTION_PROFILER
438 bool "Kernel function profiler"
439 depends on FUNCTION_TRACER
442 This option enables the kernel function profiler. A file is created
443 in debugfs called function_profile_enabled which defaults to zero.
444 When a 1 is echoed into this file profiling begins, and when a
445 zero is entered, profiling stops. A "functions" file is created in
446 the trace_stats directory; this file shows the list of functions that
447 have been hit and their counters.
451 config FTRACE_MCOUNT_RECORD
453 depends on DYNAMIC_FTRACE
454 depends on HAVE_FTRACE_MCOUNT_RECORD
456 config FTRACE_SELFTEST
459 config FTRACE_STARTUP_TEST
460 bool "Perform a startup test on ftrace"
461 depends on GENERIC_TRACER
462 select FTRACE_SELFTEST
464 This option performs a series of startup tests on ftrace. On bootup
465 a series of tests are made to verify that the tracer is
466 functioning properly. It will do tests on all the configured
469 config EVENT_TRACE_TEST_SYSCALLS
470 bool "Run selftest on syscall events"
471 depends on FTRACE_STARTUP_TEST
473 This option will also enable testing every syscall event.
474 It only enables the event and disables it and runs various loads
475 with the event enabled. This adds a bit more time for kernel boot
476 up since it runs this on every system call defined.
478 TBD - enable a way to actually call the syscalls as we test their
482 bool "Memory mapped IO tracing"
483 depends on HAVE_MMIOTRACE_SUPPORT && PCI
484 select GENERIC_TRACER
486 Mmiotrace traces Memory Mapped I/O access and is meant for
487 debugging and reverse engineering. It is called from the ioremap
488 implementation and works via page faults. Tracing is disabled by
489 default and can be enabled at run-time.
491 See Documentation/trace/mmiotrace.txt.
492 If you are not helping to develop drivers, say N.
494 config MMIOTRACE_TEST
495 tristate "Test module for mmiotrace"
496 depends on MMIOTRACE && m
498 This is a dumb module for testing mmiotrace. It is very dangerous
499 as it will write garbage to IO memory starting at a given address.
500 However, it should be safe to use on e.g. unused portion of VRAM.
502 Say N, unless you absolutely know what you are doing.
504 config RING_BUFFER_BENCHMARK
505 tristate "Ring buffer benchmark stress tester"
506 depends on RING_BUFFER
508 This option creates a test to stress the ring buffer and benchmark it.
509 It creates its own ring buffer such that it will not interfere with
510 any other users of the ring buffer (such as ftrace). It then creates
511 a producer and consumer that will run for 10 seconds and sleep for
512 10 seconds. Each interval it will print out the number of events
513 it recorded and give a rough estimate of how long each iteration took.
515 It does not disable interrupts or raise its priority, so it may be
516 affected by processes that are running.
522 endif # TRACING_SUPPORT