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 config HAVE_FUNCTION_TRACER
18 config HAVE_FUNCTION_GRAPH_TRACER
21 config HAVE_FUNCTION_TRACE_MCOUNT_TEST
24 This gets selected when the arch tests the function_trace_stop
25 variable at the mcount call site. Otherwise, this variable
26 is tested by the called function.
28 config HAVE_DYNAMIC_FTRACE
31 config HAVE_FTRACE_MCOUNT_RECORD
34 config HAVE_HW_BRANCH_TRACER
37 config HAVE_FTRACE_SYSCALLS
40 config TRACER_MAX_TRACE
46 config FTRACE_NMI_ENTER
48 depends on HAVE_FTRACE_NMI_ENTER
55 select STACKTRACE if STACKTRACE_SUPPORT
61 # Minimum requirements an architecture has to meet for us to
62 # be able to offer generic tracing facilities:
64 config TRACING_SUPPORT
66 # PPC32 has no irqflags tracing support, but it can use most of the
67 # tracers anyway, they were tested to build and work. Note that new
68 # exceptions to this list aren't welcomed, better implement the
69 # irqflags tracing for your architecture.
70 depends on TRACE_IRQFLAGS_SUPPORT || PPC32
71 depends on STACKTRACE_SUPPORT
78 config FUNCTION_TRACER
79 bool "Kernel Function Tracer"
80 depends on HAVE_FUNCTION_TRACER
84 select CONTEXT_SWITCH_TRACER
86 Enable the kernel to trace every kernel function. This is done
87 by using a compiler feature to insert a small, 5-byte No-Operation
88 instruction to the beginning of every kernel function, which NOP
89 sequence is then dynamically patched into a tracer call when
90 tracing is enabled by the administrator. If it's runtime disabled
91 (the bootup default), then the overhead of the instructions is very
92 small and not measurable even in micro-benchmarks.
94 config FUNCTION_GRAPH_TRACER
95 bool "Kernel Function Graph Tracer"
96 depends on HAVE_FUNCTION_GRAPH_TRACER
97 depends on FUNCTION_TRACER
100 Enable the kernel to trace a function at both its return
102 Its first purpose is to trace the duration of functions and
103 draw a call graph for each thread with some information like
104 the return value. This is done by setting the current return
105 address on the current task structure into a stack of calls.
108 config IRQSOFF_TRACER
109 bool "Interrupts-off Latency Tracer"
111 depends on TRACE_IRQFLAGS_SUPPORT
112 depends on GENERIC_TIME
113 select TRACE_IRQFLAGS
115 select TRACER_MAX_TRACE
117 This option measures the time spent in irqs-off critical
118 sections, with microsecond accuracy.
120 The default measurement method is a maximum search, which is
121 disabled by default and can be runtime (re-)started
124 echo 0 > /debugfs/tracing/tracing_max_latency
126 (Note that kernel size and overhead increases with this option
127 enabled. This option and the preempt-off timing option can be
128 used together or separately.)
130 config PREEMPT_TRACER
131 bool "Preemption-off Latency Tracer"
133 depends on GENERIC_TIME
136 select TRACER_MAX_TRACE
138 This option measures the time spent in preemption off critical
139 sections, with microsecond accuracy.
141 The default measurement method is a maximum search, which is
142 disabled by default and can be runtime (re-)started
145 echo 0 > /debugfs/tracing/tracing_max_latency
147 (Note that kernel size and overhead increases with this option
148 enabled. This option and the irqs-off timing option can be
149 used together or separately.)
151 config SYSPROF_TRACER
152 bool "Sysprof Tracer"
155 select CONTEXT_SWITCH_TRACER
157 This tracer provides the trace needed by the 'Sysprof' userspace
161 bool "Scheduling Latency Tracer"
163 select CONTEXT_SWITCH_TRACER
164 select TRACER_MAX_TRACE
166 This tracer tracks the latency of the highest priority task
167 to be scheduled in, starting from the point it has woken up.
169 config CONTEXT_SWITCH_TRACER
170 bool "Trace process context switches"
174 This tracer gets called from the context switch and records
175 all switching of tasks.
178 bool "Trace various events in the kernel"
181 This tracer hooks to various trace points in the kernel
182 allowing the user to pick and choose which trace point they
185 config FTRACE_SYSCALLS
186 bool "Trace syscalls"
187 depends on HAVE_FTRACE_SYSCALLS
191 Basic tracer to catch the syscall entry and exit events.
194 bool "Trace boot initcalls"
196 select CONTEXT_SWITCH_TRACER
198 This tracer helps developers to optimize boot times: it records
199 the timings of the initcalls and traces key events and the identity
200 of tasks that can cause boot delays, such as context-switches.
202 Its aim is to be parsed by the /scripts/bootgraph.pl tool to
203 produce pretty graphics about boot inefficiencies, giving a visual
204 representation of the delays during initcalls - but the raw
205 /debug/tracing/trace text output is readable too.
207 You must pass in ftrace=initcall to the kernel command line
208 to enable this on bootup.
210 config TRACE_BRANCH_PROFILING
211 bool "Trace likely/unlikely profiler"
214 This tracer profiles all the the likely and unlikely macros
215 in the kernel. It will display the results in:
217 /debugfs/tracing/profile_annotated_branch
219 Note: this will add a significant overhead, only turn this
220 on if you need to profile the system's use of these macros.
224 config PROFILE_ALL_BRANCHES
225 bool "Profile all if conditionals"
226 depends on TRACE_BRANCH_PROFILING
228 This tracer profiles all branch conditions. Every if ()
229 taken in the kernel is recorded whether it hit or miss.
230 The results will be displayed in:
232 /debugfs/tracing/profile_branch
234 This configuration, when enabled, will impose a great overhead
235 on the system. This should only be enabled when the system
240 config TRACING_BRANCHES
243 Selected by tracers that will trace the likely and unlikely
244 conditions. This prevents the tracers themselves from being
245 profiled. Profiling the tracing infrastructure can only happen
246 when the likelys and unlikelys are not being traced.
249 bool "Trace likely/unlikely instances"
250 depends on TRACE_BRANCH_PROFILING
251 select TRACING_BRANCHES
253 This traces the events of likely and unlikely condition
254 calls in the kernel. The difference between this and the
255 "Trace likely/unlikely profiler" is that this is not a
256 histogram of the callers, but actually places the calling
257 events into a running trace buffer to see when and where the
258 events happened, as well as their results.
263 bool "Trace power consumption behavior"
267 This tracer helps developers to analyze and optimize the kernels
268 power management decisions, specifically the C-state and P-state
273 bool "Trace max stack"
274 depends on HAVE_FUNCTION_TRACER
275 select FUNCTION_TRACER
279 This special tracer records the maximum stack footprint of the
280 kernel and displays it in debugfs/tracing/stack_trace.
282 This tracer works by hooking into every function call that the
283 kernel executes, and keeping a maximum stack depth value and
284 stack-trace saved. If this is configured with DYNAMIC_FTRACE
285 then it will not have any overhead while the stack tracer
288 To enable the stack tracer on bootup, pass in 'stacktrace'
289 on the kernel command line.
291 The stack tracer can also be enabled or disabled via the
292 sysctl kernel.stack_tracer_enabled
296 config HW_BRANCH_TRACER
297 depends on HAVE_HW_BRANCH_TRACER
298 bool "Trace hw branches"
301 This tracer records all branches on the system in a circular
302 buffer giving access to the last N branches for each cpu.
305 bool "Trace SLAB allocations"
308 kmemtrace provides tracing for slab allocator functions, such as
309 kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected
310 data is then fed to the userspace application in order to analyse
311 allocation hotspots, internal fragmentation and so on, making it
312 possible to see how well an allocator performs, as well as debug
313 and profile kernel code.
315 This requires an userspace application to use. See
316 Documentation/vm/kmemtrace.txt for more information.
318 Saying Y will make the kernel somewhat larger and slower. However,
319 if you disable kmemtrace at run-time or boot-time, the performance
320 impact is minimal (depending on the arch the kernel is built for).
324 config WORKQUEUE_TRACER
325 bool "Trace workqueues"
328 The workqueue tracer provides some statistical informations
329 about each cpu workqueue thread such as the number of the
330 works inserted and executed since their creation. It can help
331 to evaluate the amount of work each of them have to perform.
332 For example it can help a developer to decide whether he should
333 choose a per cpu workqueue instead of a singlethreaded one.
335 config BLK_DEV_IO_TRACE
336 bool "Support for tracing block io actions"
345 Say Y here if you want to be able to trace the block layer actions
346 on a given queue. Tracing allows you to see any traffic happening
347 on a block device queue. For more information (and the userspace
348 support tools needed), fetch the blktrace tools from:
350 git://git.kernel.dk/blktrace.git
352 Tracing also is possible using the ftrace interface, e.g.:
354 echo 1 > /sys/block/sda/sda1/trace/enable
355 echo blk > /sys/kernel/debug/tracing/current_tracer
356 cat /sys/kernel/debug/tracing/trace_pipe
360 config DYNAMIC_FTRACE
361 bool "enable/disable ftrace tracepoints dynamically"
362 depends on FUNCTION_TRACER
363 depends on HAVE_DYNAMIC_FTRACE
366 This option will modify all the calls to ftrace dynamically
367 (will patch them out of the binary image and replaces them
368 with a No-Op instruction) as they are called. A table is
369 created to dynamically enable them again.
371 This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but otherwise
372 has native performance as long as no tracing is active.
374 The changes to the code are done by a kernel thread that
375 wakes up once a second and checks to see if any ftrace calls
376 were made. If so, it runs stop_machine (stops all CPUS)
377 and modifies the code to jump over the call to ftrace.
379 config FUNCTION_PROFILER
380 bool "Kernel function profiler"
381 depends on FUNCTION_TRACER
384 This option enables the kernel function profiler. A file is created
385 in debugfs called function_profile_enabled which defaults to zero.
386 When a 1 is echoed into this file profiling begins, and when a
387 zero is entered, profiling stops. A file in the trace_stats
388 directory called functions, that show the list of functions that
389 have been hit and their counters.
393 config FTRACE_MCOUNT_RECORD
395 depends on DYNAMIC_FTRACE
396 depends on HAVE_FTRACE_MCOUNT_RECORD
398 config FTRACE_SELFTEST
401 config FTRACE_STARTUP_TEST
402 bool "Perform a startup test on ftrace"
404 select FTRACE_SELFTEST
406 This option performs a series of startup tests on ftrace. On bootup
407 a series of tests are made to verify that the tracer is
408 functioning properly. It will do tests on all the configured
412 bool "Memory mapped IO tracing"
413 depends on HAVE_MMIOTRACE_SUPPORT && PCI
416 Mmiotrace traces Memory Mapped I/O access and is meant for
417 debugging and reverse engineering. It is called from the ioremap
418 implementation and works via page faults. Tracing is disabled by
419 default and can be enabled at run-time.
421 See Documentation/tracers/mmiotrace.txt.
422 If you are not helping to develop drivers, say N.
424 config MMIOTRACE_TEST
425 tristate "Test module for mmiotrace"
426 depends on MMIOTRACE && m
428 This is a dumb module for testing mmiotrace. It is very dangerous
429 as it will write garbage to IO memory starting at a given address.
430 However, it should be safe to use on e.g. unused portion of VRAM.
432 Say N, unless you absolutely know what you are doing.
436 endif # TRACING_SUPPORT