Kconfig 15 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494
  1. #
  2. # Architectures that offer an FUNCTION_TRACER implementation should
  3. # select HAVE_FUNCTION_TRACER:
  4. #
  5. config USER_STACKTRACE_SUPPORT
  6. bool
  7. config NOP_TRACER
  8. bool
  9. config HAVE_FTRACE_NMI_ENTER
  10. bool
  11. config HAVE_FUNCTION_TRACER
  12. bool
  13. config HAVE_FUNCTION_GRAPH_TRACER
  14. bool
  15. config HAVE_FUNCTION_TRACE_MCOUNT_TEST
  16. bool
  17. help
  18. This gets selected when the arch tests the function_trace_stop
  19. variable at the mcount call site. Otherwise, this variable
  20. is tested by the called function.
  21. config HAVE_DYNAMIC_FTRACE
  22. bool
  23. config HAVE_FTRACE_MCOUNT_RECORD
  24. bool
  25. config HAVE_HW_BRANCH_TRACER
  26. bool
  27. config HAVE_FTRACE_SYSCALLS
  28. bool
  29. config TRACER_MAX_TRACE
  30. bool
  31. config RING_BUFFER
  32. bool
  33. config FTRACE_NMI_ENTER
  34. bool
  35. depends on HAVE_FTRACE_NMI_ENTER
  36. default y
  37. config EVENT_TRACING
  38. bool
  39. config TRACING
  40. bool
  41. select DEBUG_FS
  42. select RING_BUFFER
  43. select STACKTRACE if STACKTRACE_SUPPORT
  44. select TRACEPOINTS
  45. select NOP_TRACER
  46. select BINARY_PRINTF
  47. select EVENT_TRACING
  48. #
  49. # Minimum requirements an architecture has to meet for us to
  50. # be able to offer generic tracing facilities:
  51. #
  52. config TRACING_SUPPORT
  53. bool
  54. # PPC32 has no irqflags tracing support, but it can use most of the
  55. # tracers anyway, they were tested to build and work. Note that new
  56. # exceptions to this list aren't welcomed, better implement the
  57. # irqflags tracing for your architecture.
  58. depends on TRACE_IRQFLAGS_SUPPORT || PPC32
  59. depends on STACKTRACE_SUPPORT
  60. default y
  61. if TRACING_SUPPORT
  62. menuconfig FTRACE
  63. bool "Tracers"
  64. default y if DEBUG_KERNEL
  65. help
  66. Enable the kernel tracing infrastructure.
  67. if FTRACE
  68. config FUNCTION_TRACER
  69. bool "Kernel Function Tracer"
  70. depends on HAVE_FUNCTION_TRACER
  71. select FRAME_POINTER
  72. select KALLSYMS
  73. select TRACING
  74. select CONTEXT_SWITCH_TRACER
  75. help
  76. Enable the kernel to trace every kernel function. This is done
  77. by using a compiler feature to insert a small, 5-byte No-Operation
  78. instruction to the beginning of every kernel function, which NOP
  79. sequence is then dynamically patched into a tracer call when
  80. tracing is enabled by the administrator. If it's runtime disabled
  81. (the bootup default), then the overhead of the instructions is very
  82. small and not measurable even in micro-benchmarks.
  83. config FUNCTION_GRAPH_TRACER
  84. bool "Kernel Function Graph Tracer"
  85. depends on HAVE_FUNCTION_GRAPH_TRACER
  86. depends on FUNCTION_TRACER
  87. default y
  88. help
  89. Enable the kernel to trace a function at both its return
  90. and its entry.
  91. Its first purpose is to trace the duration of functions and
  92. draw a call graph for each thread with some information like
  93. the return value. This is done by setting the current return
  94. address on the current task structure into a stack of calls.
  95. config IRQSOFF_TRACER
  96. bool "Interrupts-off Latency Tracer"
  97. default n
  98. depends on TRACE_IRQFLAGS_SUPPORT
  99. depends on GENERIC_TIME
  100. select TRACE_IRQFLAGS
  101. select TRACING
  102. select TRACER_MAX_TRACE
  103. help
  104. This option measures the time spent in irqs-off critical
  105. sections, with microsecond accuracy.
  106. The default measurement method is a maximum search, which is
  107. disabled by default and can be runtime (re-)started
  108. via:
  109. echo 0 > /debugfs/tracing/tracing_max_latency
  110. (Note that kernel size and overhead increases with this option
  111. enabled. This option and the preempt-off timing option can be
  112. used together or separately.)
  113. config PREEMPT_TRACER
  114. bool "Preemption-off Latency Tracer"
  115. default n
  116. depends on GENERIC_TIME
  117. depends on PREEMPT
  118. select TRACING
  119. select TRACER_MAX_TRACE
  120. help
  121. This option measures the time spent in preemption off critical
  122. sections, with microsecond accuracy.
  123. The default measurement method is a maximum search, which is
  124. disabled by default and can be runtime (re-)started
  125. via:
  126. echo 0 > /debugfs/tracing/tracing_max_latency
  127. (Note that kernel size and overhead increases with this option
  128. enabled. This option and the irqs-off timing option can be
  129. used together or separately.)
  130. config SYSPROF_TRACER
  131. bool "Sysprof Tracer"
  132. depends on X86
  133. select TRACING
  134. select CONTEXT_SWITCH_TRACER
  135. help
  136. This tracer provides the trace needed by the 'Sysprof' userspace
  137. tool.
  138. config SCHED_TRACER
  139. bool "Scheduling Latency Tracer"
  140. select TRACING
  141. select CONTEXT_SWITCH_TRACER
  142. select TRACER_MAX_TRACE
  143. help
  144. This tracer tracks the latency of the highest priority task
  145. to be scheduled in, starting from the point it has woken up.
  146. config CONTEXT_SWITCH_TRACER
  147. bool "Trace process context switches"
  148. select TRACING
  149. select MARKERS
  150. help
  151. This tracer gets called from the context switch and records
  152. all switching of tasks.
  153. config ENABLE_EVENT_TRACING
  154. bool "Trace various events in the kernel"
  155. select TRACING
  156. help
  157. This tracer hooks to various trace points in the kernel
  158. allowing the user to pick and choose which trace point they
  159. want to trace.
  160. Note, all tracers enable event tracing. This option is
  161. only a convenience to enable event tracing when no other
  162. tracers are selected.
  163. config FTRACE_SYSCALLS
  164. bool "Trace syscalls"
  165. depends on HAVE_FTRACE_SYSCALLS
  166. select TRACING
  167. select KALLSYMS
  168. help
  169. Basic tracer to catch the syscall entry and exit events.
  170. config BOOT_TRACER
  171. bool "Trace boot initcalls"
  172. select TRACING
  173. select CONTEXT_SWITCH_TRACER
  174. help
  175. This tracer helps developers to optimize boot times: it records
  176. the timings of the initcalls and traces key events and the identity
  177. of tasks that can cause boot delays, such as context-switches.
  178. Its aim is to be parsed by the /scripts/bootgraph.pl tool to
  179. produce pretty graphics about boot inefficiencies, giving a visual
  180. representation of the delays during initcalls - but the raw
  181. /debug/tracing/trace text output is readable too.
  182. You must pass in ftrace=initcall to the kernel command line
  183. to enable this on bootup.
  184. config TRACE_BRANCH_PROFILING
  185. bool
  186. select TRACING
  187. choice
  188. prompt "Branch Profiling"
  189. default BRANCH_PROFILE_NONE
  190. help
  191. The branch profiling is a software profiler. It will add hooks
  192. into the C conditionals to test which path a branch takes.
  193. The likely/unlikely profiler only looks at the conditions that
  194. are annotated with a likely or unlikely macro.
  195. The "all branch" profiler will profile every if statement in the
  196. kernel. This profiler will also enable the likely/unlikely
  197. profiler as well.
  198. Either of the above profilers add a bit of overhead to the system.
  199. If unsure choose "No branch profiling".
  200. config BRANCH_PROFILE_NONE
  201. bool "No branch profiling"
  202. help
  203. No branch profiling. Branch profiling adds a bit of overhead.
  204. Only enable it if you want to analyse the branching behavior.
  205. Otherwise keep it disabled.
  206. config PROFILE_ANNOTATED_BRANCHES
  207. bool "Trace likely/unlikely profiler"
  208. select TRACE_BRANCH_PROFILING
  209. help
  210. This tracer profiles all the the likely and unlikely macros
  211. in the kernel. It will display the results in:
  212. /debugfs/tracing/profile_annotated_branch
  213. Note: this will add a significant overhead, only turn this
  214. on if you need to profile the system's use of these macros.
  215. config PROFILE_ALL_BRANCHES
  216. bool "Profile all if conditionals"
  217. select TRACE_BRANCH_PROFILING
  218. help
  219. This tracer profiles all branch conditions. Every if ()
  220. taken in the kernel is recorded whether it hit or miss.
  221. The results will be displayed in:
  222. /debugfs/tracing/profile_branch
  223. This option also enables the likely/unlikely profiler.
  224. This configuration, when enabled, will impose a great overhead
  225. on the system. This should only be enabled when the system
  226. is to be analyzed
  227. endchoice
  228. config TRACING_BRANCHES
  229. bool
  230. help
  231. Selected by tracers that will trace the likely and unlikely
  232. conditions. This prevents the tracers themselves from being
  233. profiled. Profiling the tracing infrastructure can only happen
  234. when the likelys and unlikelys are not being traced.
  235. config BRANCH_TRACER
  236. bool "Trace likely/unlikely instances"
  237. depends on TRACE_BRANCH_PROFILING
  238. select TRACING_BRANCHES
  239. help
  240. This traces the events of likely and unlikely condition
  241. calls in the kernel. The difference between this and the
  242. "Trace likely/unlikely profiler" is that this is not a
  243. histogram of the callers, but actually places the calling
  244. events into a running trace buffer to see when and where the
  245. events happened, as well as their results.
  246. Say N if unsure.
  247. config POWER_TRACER
  248. bool "Trace power consumption behavior"
  249. depends on X86
  250. select TRACING
  251. help
  252. This tracer helps developers to analyze and optimize the kernels
  253. power management decisions, specifically the C-state and P-state
  254. behavior.
  255. config STACK_TRACER
  256. bool "Trace max stack"
  257. depends on HAVE_FUNCTION_TRACER
  258. select FUNCTION_TRACER
  259. select STACKTRACE
  260. select KALLSYMS
  261. help
  262. This special tracer records the maximum stack footprint of the
  263. kernel and displays it in debugfs/tracing/stack_trace.
  264. This tracer works by hooking into every function call that the
  265. kernel executes, and keeping a maximum stack depth value and
  266. stack-trace saved. If this is configured with DYNAMIC_FTRACE
  267. then it will not have any overhead while the stack tracer
  268. is disabled.
  269. To enable the stack tracer on bootup, pass in 'stacktrace'
  270. on the kernel command line.
  271. The stack tracer can also be enabled or disabled via the
  272. sysctl kernel.stack_tracer_enabled
  273. Say N if unsure.
  274. config HW_BRANCH_TRACER
  275. depends on HAVE_HW_BRANCH_TRACER
  276. bool "Trace hw branches"
  277. select TRACING
  278. help
  279. This tracer records all branches on the system in a circular
  280. buffer giving access to the last N branches for each cpu.
  281. config KMEMTRACE
  282. bool "Trace SLAB allocations"
  283. select TRACING
  284. help
  285. kmemtrace provides tracing for slab allocator functions, such as
  286. kmalloc, kfree, kmem_cache_alloc, kmem_cache_free etc.. Collected
  287. data is then fed to the userspace application in order to analyse
  288. allocation hotspots, internal fragmentation and so on, making it
  289. possible to see how well an allocator performs, as well as debug
  290. and profile kernel code.
  291. This requires an userspace application to use. See
  292. Documentation/trace/kmemtrace.txt for more information.
  293. Saying Y will make the kernel somewhat larger and slower. However,
  294. if you disable kmemtrace at run-time or boot-time, the performance
  295. impact is minimal (depending on the arch the kernel is built for).
  296. If unsure, say N.
  297. config WORKQUEUE_TRACER
  298. bool "Trace workqueues"
  299. select TRACING
  300. help
  301. The workqueue tracer provides some statistical informations
  302. about each cpu workqueue thread such as the number of the
  303. works inserted and executed since their creation. It can help
  304. to evaluate the amount of work each of them have to perform.
  305. For example it can help a developer to decide whether he should
  306. choose a per cpu workqueue instead of a singlethreaded one.
  307. config BLK_DEV_IO_TRACE
  308. bool "Support for tracing block io actions"
  309. depends on SYSFS
  310. depends on BLOCK
  311. select RELAY
  312. select DEBUG_FS
  313. select TRACEPOINTS
  314. select TRACING
  315. select STACKTRACE
  316. help
  317. Say Y here if you want to be able to trace the block layer actions
  318. on a given queue. Tracing allows you to see any traffic happening
  319. on a block device queue. For more information (and the userspace
  320. support tools needed), fetch the blktrace tools from:
  321. git://git.kernel.dk/blktrace.git
  322. Tracing also is possible using the ftrace interface, e.g.:
  323. echo 1 > /sys/block/sda/sda1/trace/enable
  324. echo blk > /sys/kernel/debug/tracing/current_tracer
  325. cat /sys/kernel/debug/tracing/trace_pipe
  326. If unsure, say N.
  327. config DYNAMIC_FTRACE
  328. bool "enable/disable ftrace tracepoints dynamically"
  329. depends on FUNCTION_TRACER
  330. depends on HAVE_DYNAMIC_FTRACE
  331. default y
  332. help
  333. This option will modify all the calls to ftrace dynamically
  334. (will patch them out of the binary image and replaces them
  335. with a No-Op instruction) as they are called. A table is
  336. created to dynamically enable them again.
  337. This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but otherwise
  338. has native performance as long as no tracing is active.
  339. The changes to the code are done by a kernel thread that
  340. wakes up once a second and checks to see if any ftrace calls
  341. were made. If so, it runs stop_machine (stops all CPUS)
  342. and modifies the code to jump over the call to ftrace.
  343. config FUNCTION_PROFILER
  344. bool "Kernel function profiler"
  345. depends on FUNCTION_TRACER
  346. default n
  347. help
  348. This option enables the kernel function profiler. A file is created
  349. in debugfs called function_profile_enabled which defaults to zero.
  350. When a 1 is echoed into this file profiling begins, and when a
  351. zero is entered, profiling stops. A file in the trace_stats
  352. directory called functions, that show the list of functions that
  353. have been hit and their counters.
  354. If in doubt, say N
  355. config FTRACE_MCOUNT_RECORD
  356. def_bool y
  357. depends on DYNAMIC_FTRACE
  358. depends on HAVE_FTRACE_MCOUNT_RECORD
  359. config FTRACE_SELFTEST
  360. bool
  361. config FTRACE_STARTUP_TEST
  362. bool "Perform a startup test on ftrace"
  363. depends on TRACING
  364. select FTRACE_SELFTEST
  365. help
  366. This option performs a series of startup tests on ftrace. On bootup
  367. a series of tests are made to verify that the tracer is
  368. functioning properly. It will do tests on all the configured
  369. tracers of ftrace.
  370. config MMIOTRACE
  371. bool "Memory mapped IO tracing"
  372. depends on HAVE_MMIOTRACE_SUPPORT && PCI
  373. select TRACING
  374. help
  375. Mmiotrace traces Memory Mapped I/O access and is meant for
  376. debugging and reverse engineering. It is called from the ioremap
  377. implementation and works via page faults. Tracing is disabled by
  378. default and can be enabled at run-time.
  379. See Documentation/trace/mmiotrace.txt.
  380. If you are not helping to develop drivers, say N.
  381. config MMIOTRACE_TEST
  382. tristate "Test module for mmiotrace"
  383. depends on MMIOTRACE && m
  384. help
  385. This is a dumb module for testing mmiotrace. It is very dangerous
  386. as it will write garbage to IO memory starting at a given address.
  387. However, it should be safe to use on e.g. unused portion of VRAM.
  388. Say N, unless you absolutely know what you are doing.
  389. config RING_BUFFER_BENCHMARK
  390. tristate "Ring buffer benchmark stress tester"
  391. depends on RING_BUFFER
  392. help
  393. This option creates a test to stress the ring buffer and bench mark it.
  394. It creates its own ring buffer such that it will not interfer with
  395. any other users of the ring buffer (such as ftrace). It then creates
  396. a producer and consumer that will run for 10 seconds and sleep for
  397. 10 seconds. Each interval it will print out the number of events
  398. it recorded and give a rough estimate of how long each iteration took.
  399. It does not disable interrupts or raise its priority, so it may be
  400. affected by processes that are running.
  401. If unsure, say N
  402. endif # FTRACE
  403. endif # TRACING_SUPPORT