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
68 config FTRACE_NMI_ENTER
70 depends on HAVE_FTRACE_NMI_ENTER
74 select CONTEXT_SWITCH_TRACER
77 config EVENT_POWER_TRACING_DEPRECATED
78 depends on EVENT_TRACING
79 bool "Deprecated power event trace API, to be removed"
82 Provides old power event types:
83 C-state/idle accounting events:
86 and old cpufreq accounting event:
88 This is for userspace compatibility
89 and will vanish after 5 kernel iterations,
92 config CONTEXT_SWITCH_TRACER
95 config RING_BUFFER_ALLOW_SWAP
98 Allow the use of ring_buffer_swap_cpu.
99 Adds a very slight overhead to tracing when enabled.
101 # All tracer options should select GENERIC_TRACER. For those options that are
102 # enabled by all tracers (context switch and event tracer) they select TRACING.
103 # This allows those options to appear when no other tracer is selected. But the
104 # options do not appear when something else selects it. We need the two options
105 # GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
106 # hiding of the automatic options.
112 select STACKTRACE if STACKTRACE_SUPPORT
118 config GENERIC_TRACER
123 # Minimum requirements an architecture has to meet for us to
124 # be able to offer generic tracing facilities:
126 config TRACING_SUPPORT
128 # PPC32 has no irqflags tracing support, but it can use most of the
129 # tracers anyway, they were tested to build and work. Note that new
130 # exceptions to this list aren't welcomed, better implement the
131 # irqflags tracing for your architecture.
132 depends on TRACE_IRQFLAGS_SUPPORT || PPC32
133 depends on STACKTRACE_SUPPORT
140 default y if DEBUG_KERNEL
142 Enable the kernel tracing infrastructure.
146 config FUNCTION_TRACER
147 bool "Kernel Function Tracer"
148 depends on HAVE_FUNCTION_TRACER
150 select GENERIC_TRACER
151 select CONTEXT_SWITCH_TRACER
153 Enable the kernel to trace every kernel function. This is done
154 by using a compiler feature to insert a small, 5-byte No-Operation
155 instruction at the beginning of every kernel function, which NOP
156 sequence is then dynamically patched into a tracer call when
157 tracing is enabled by the administrator. If it's runtime disabled
158 (the bootup default), then the overhead of the instructions is very
159 small and not measurable even in micro-benchmarks.
161 config FUNCTION_GRAPH_TRACER
162 bool "Kernel Function Graph Tracer"
163 depends on HAVE_FUNCTION_GRAPH_TRACER
164 depends on FUNCTION_TRACER
165 depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
168 Enable the kernel to trace a function at both its return
170 Its first purpose is to trace the duration of functions and
171 draw a call graph for each thread with some information like
172 the return value. This is done by setting the current return
173 address on the current task structure into a stack of calls.
176 config IRQSOFF_TRACER
177 bool "Interrupts-off Latency Tracer"
179 depends on TRACE_IRQFLAGS_SUPPORT
180 depends on !ARCH_USES_GETTIMEOFFSET
181 select TRACE_IRQFLAGS
182 select GENERIC_TRACER
183 select TRACER_MAX_TRACE
184 select RING_BUFFER_ALLOW_SWAP
186 This option measures the time spent in irqs-off critical
187 sections, with microsecond accuracy.
189 The default measurement method is a maximum search, which is
190 disabled by default and can be runtime (re-)started
193 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
195 (Note that kernel size and overhead increase with this option
196 enabled. This option and the preempt-off timing option can be
197 used together or separately.)
199 config PREEMPT_TRACER
200 bool "Preemption-off Latency Tracer"
202 depends on !ARCH_USES_GETTIMEOFFSET
204 select GENERIC_TRACER
205 select TRACER_MAX_TRACE
206 select RING_BUFFER_ALLOW_SWAP
208 This option measures the time spent in preemption-off critical
209 sections, with microsecond accuracy.
211 The default measurement method is a maximum search, which is
212 disabled by default and can be runtime (re-)started
215 echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
217 (Note that kernel size and overhead increase with this option
218 enabled. This option and the irqs-off timing option can be
219 used together or separately.)
222 bool "Scheduling Latency Tracer"
223 select GENERIC_TRACER
224 select CONTEXT_SWITCH_TRACER
225 select TRACER_MAX_TRACE
227 This tracer tracks the latency of the highest priority task
228 to be scheduled in, starting from the point it has woken up.
230 config ENABLE_DEFAULT_TRACERS
231 bool "Trace process context switches and events"
232 depends on !GENERIC_TRACER
235 This tracer hooks to various trace points in the kernel,
236 allowing the user to pick and choose which trace point they
237 want to trace. It also includes the sched_switch tracer plugin.
239 config FTRACE_SYSCALLS
240 bool "Trace syscalls"
241 depends on HAVE_SYSCALL_TRACEPOINTS
242 select GENERIC_TRACER
245 Basic tracer to catch the syscall entry and exit events.
247 config TRACE_BRANCH_PROFILING
249 select GENERIC_TRACER
252 prompt "Branch Profiling"
253 default BRANCH_PROFILE_NONE
255 The branch profiling is a software profiler. It will add hooks
256 into the C conditionals to test which path a branch takes.
258 The likely/unlikely profiler only looks at the conditions that
259 are annotated with a likely or unlikely macro.
261 The "all branch" profiler will profile every if-statement in the
262 kernel. This profiler will also enable the likely/unlikely
265 Either of the above profilers adds a bit of overhead to the system.
266 If unsure, choose "No branch profiling".
268 config BRANCH_PROFILE_NONE
269 bool "No branch profiling"
271 No branch profiling. Branch profiling adds a bit of overhead.
272 Only enable it if you want to analyse the branching behavior.
273 Otherwise keep it disabled.
275 config PROFILE_ANNOTATED_BRANCHES
276 bool "Trace likely/unlikely profiler"
277 select TRACE_BRANCH_PROFILING
279 This tracer profiles all likely and unlikely macros
280 in the kernel. It will display the results in:
282 /sys/kernel/debug/tracing/trace_stat/branch_annotated
284 Note: this will add a significant overhead; only turn this
285 on if you need to profile the system's use of these macros.
287 config PROFILE_ALL_BRANCHES
288 bool "Profile all if conditionals"
289 select TRACE_BRANCH_PROFILING
291 This tracer profiles all branch conditions. Every if ()
292 taken in the kernel is recorded whether it hit or miss.
293 The results will be displayed in:
295 /sys/kernel/debug/tracing/trace_stat/branch_all
297 This option also enables the likely/unlikely profiler.
299 This configuration, when enabled, will impose a great overhead
300 on the system. This should only be enabled when the system
301 is to be analyzed in much detail.
304 config TRACING_BRANCHES
307 Selected by tracers that will trace the likely and unlikely
308 conditions. This prevents the tracers themselves from being
309 profiled. Profiling the tracing infrastructure can only happen
310 when the likelys and unlikelys are not being traced.
313 bool "Trace likely/unlikely instances"
314 depends on TRACE_BRANCH_PROFILING
315 select TRACING_BRANCHES
317 This traces the events of likely and unlikely condition
318 calls in the kernel. The difference between this and the
319 "Trace likely/unlikely profiler" is that this is not a
320 histogram of the callers, but actually places the calling
321 events into a running trace buffer to see when and where the
322 events happened, as well as their results.
327 bool "Trace max stack"
328 depends on HAVE_FUNCTION_TRACER
329 select FUNCTION_TRACER
333 This special tracer records the maximum stack footprint of the
334 kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
336 This tracer works by hooking into every function call that the
337 kernel executes, and keeping a maximum stack depth value and
338 stack-trace saved. If this is configured with DYNAMIC_FTRACE
339 then it will not have any overhead while the stack tracer
342 To enable the stack tracer on bootup, pass in 'stacktrace'
343 on the kernel command line.
345 The stack tracer can also be enabled or disabled via the
346 sysctl kernel.stack_tracer_enabled
350 config BLK_DEV_IO_TRACE
351 bool "Support for tracing block IO actions"
357 select GENERIC_TRACER
360 Say Y here if you want to be able to trace the block layer actions
361 on a given queue. Tracing allows you to see any traffic happening
362 on a block device queue. For more information (and the userspace
363 support tools needed), fetch the blktrace tools from:
365 git://git.kernel.dk/blktrace.git
367 Tracing also is possible using the ftrace interface, e.g.:
369 echo 1 > /sys/block/sda/sda1/trace/enable
370 echo blk > /sys/kernel/debug/tracing/current_tracer
371 cat /sys/kernel/debug/tracing/trace_pipe
377 depends on HAVE_REGS_AND_STACK_ACCESS_API
378 bool "Enable kprobes-based dynamic events"
383 This allows the user to add tracing events (similar to tracepoints)
384 on the fly via the ftrace interface. See
385 Documentation/trace/kprobetrace.txt for more details.
387 Those events can be inserted wherever kprobes can probe, and record
388 various register and memory values.
390 This option is also required by perf-probe subcommand of perf tools.
391 If you want to use perf tools, this option is strongly recommended.
394 bool "Enable uprobes-based dynamic events"
395 depends on ARCH_SUPPORTS_UPROBES
402 This allows the user to add tracing events on top of userspace
403 dynamic events (similar to tracepoints) on the fly via the trace
404 events interface. Those events can be inserted wherever uprobes
405 can probe, and record various registers.
406 This option is required if you plan to use perf-probe subcommand
407 of perf tools on user space applications.
412 config DYNAMIC_FTRACE
413 bool "enable/disable ftrace tracepoints dynamically"
414 depends on FUNCTION_TRACER
415 depends on HAVE_DYNAMIC_FTRACE
418 This option will modify all the calls to ftrace dynamically
419 (will patch them out of the binary image and replace them
420 with a No-Op instruction) as they are called. A table is
421 created to dynamically enable them again.
423 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
424 otherwise has native performance as long as no tracing is active.
426 The changes to the code are done by a kernel thread that
427 wakes up once a second and checks to see if any ftrace calls
428 were made. If so, it runs stop_machine (stops all CPUS)
429 and modifies the code to jump over the call to ftrace.
431 config FUNCTION_PROFILER
432 bool "Kernel function profiler"
433 depends on FUNCTION_TRACER
436 This option enables the kernel function profiler. A file is created
437 in debugfs called function_profile_enabled which defaults to zero.
438 When a 1 is echoed into this file profiling begins, and when a
439 zero is entered, profiling stops. A "functions" file is created in
440 the trace_stats directory; this file shows the list of functions that
441 have been hit and their counters.
445 config FTRACE_MCOUNT_RECORD
447 depends on DYNAMIC_FTRACE
448 depends on HAVE_FTRACE_MCOUNT_RECORD
450 config FTRACE_SELFTEST
453 config FTRACE_STARTUP_TEST
454 bool "Perform a startup test on ftrace"
455 depends on GENERIC_TRACER
456 select FTRACE_SELFTEST
458 This option performs a series of startup tests on ftrace. On bootup
459 a series of tests are made to verify that the tracer is
460 functioning properly. It will do tests on all the configured
463 config EVENT_TRACE_TEST_SYSCALLS
464 bool "Run selftest on syscall events"
465 depends on FTRACE_STARTUP_TEST
467 This option will also enable testing every syscall event.
468 It only enables the event and disables it and runs various loads
469 with the event enabled. This adds a bit more time for kernel boot
470 up since it runs this on every system call defined.
472 TBD - enable a way to actually call the syscalls as we test their
476 bool "Memory mapped IO tracing"
477 depends on HAVE_MMIOTRACE_SUPPORT && PCI
478 select GENERIC_TRACER
480 Mmiotrace traces Memory Mapped I/O access and is meant for
481 debugging and reverse engineering. It is called from the ioremap
482 implementation and works via page faults. Tracing is disabled by
483 default and can be enabled at run-time.
485 See Documentation/trace/mmiotrace.txt.
486 If you are not helping to develop drivers, say N.
488 config MMIOTRACE_TEST
489 tristate "Test module for mmiotrace"
490 depends on MMIOTRACE && m
492 This is a dumb module for testing mmiotrace. It is very dangerous
493 as it will write garbage to IO memory starting at a given address.
494 However, it should be safe to use on e.g. unused portion of VRAM.
496 Say N, unless you absolutely know what you are doing.
498 config RING_BUFFER_BENCHMARK
499 tristate "Ring buffer benchmark stress tester"
500 depends on RING_BUFFER
502 This option creates a test to stress the ring buffer and benchmark it.
503 It creates its own ring buffer such that it will not interfere with
504 any other users of the ring buffer (such as ftrace). It then creates
505 a producer and consumer that will run for 10 seconds and sleep for
506 10 seconds. Each interval it will print out the number of events
507 it recorded and give a rough estimate of how long each iteration took.
509 It does not disable interrupts or raise its priority, so it may be
510 affected by processes that are running.
516 endif # TRACING_SUPPORT