Kconfig 17 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565
  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. help
  12. See Documentation/trace/ftrace-design.txt
  13. config HAVE_FUNCTION_TRACER
  14. bool
  15. help
  16. See Documentation/trace/ftrace-design.txt
  17. config HAVE_FUNCTION_GRAPH_TRACER
  18. bool
  19. help
  20. See Documentation/trace/ftrace-design.txt
  21. config HAVE_FUNCTION_GRAPH_FP_TEST
  22. bool
  23. help
  24. An arch may pass in a unique value (frame pointer) to both the
  25. entering and exiting of a function. On exit, the value is compared
  26. and if it does not match, then it will panic the kernel.
  27. config HAVE_FUNCTION_TRACE_MCOUNT_TEST
  28. bool
  29. help
  30. See Documentation/trace/ftrace-design.txt
  31. config HAVE_DYNAMIC_FTRACE
  32. bool
  33. help
  34. See Documentation/trace/ftrace-design.txt
  35. config HAVE_FTRACE_MCOUNT_RECORD
  36. bool
  37. help
  38. See Documentation/trace/ftrace-design.txt
  39. config HAVE_HW_BRANCH_TRACER
  40. bool
  41. config HAVE_SYSCALL_TRACEPOINTS
  42. bool
  43. help
  44. See Documentation/trace/ftrace-design.txt
  45. config TRACER_MAX_TRACE
  46. bool
  47. config RING_BUFFER
  48. bool
  49. config FTRACE_NMI_ENTER
  50. bool
  51. depends on HAVE_FTRACE_NMI_ENTER
  52. default y
  53. config EVENT_TRACING
  54. select CONTEXT_SWITCH_TRACER
  55. bool
  56. config CONTEXT_SWITCH_TRACER
  57. bool
  58. config RING_BUFFER_ALLOW_SWAP
  59. bool
  60. help
  61. Allow the use of ring_buffer_swap_cpu.
  62. Adds a very slight overhead to tracing when enabled.
  63. # All tracer options should select GENERIC_TRACER. For those options that are
  64. # enabled by all tracers (context switch and event tracer) they select TRACING.
  65. # This allows those options to appear when no other tracer is selected. But the
  66. # options do not appear when something else selects it. We need the two options
  67. # GENERIC_TRACER and TRACING to avoid circular dependencies to accomplish the
  68. # hiding of the automatic options.
  69. config TRACING
  70. bool
  71. select DEBUG_FS
  72. select RING_BUFFER
  73. select STACKTRACE if STACKTRACE_SUPPORT
  74. select TRACEPOINTS
  75. select NOP_TRACER
  76. select BINARY_PRINTF
  77. select EVENT_TRACING
  78. config GENERIC_TRACER
  79. bool
  80. select TRACING
  81. #
  82. # Minimum requirements an architecture has to meet for us to
  83. # be able to offer generic tracing facilities:
  84. #
  85. config TRACING_SUPPORT
  86. bool
  87. # PPC32 has no irqflags tracing support, but it can use most of the
  88. # tracers anyway, they were tested to build and work. Note that new
  89. # exceptions to this list aren't welcomed, better implement the
  90. # irqflags tracing for your architecture.
  91. depends on TRACE_IRQFLAGS_SUPPORT || PPC32
  92. depends on STACKTRACE_SUPPORT
  93. default y
  94. if TRACING_SUPPORT
  95. menuconfig FTRACE
  96. bool "Tracers"
  97. default y if DEBUG_KERNEL
  98. help
  99. Enable the kernel tracing infrastructure.
  100. if FTRACE
  101. config FUNCTION_TRACER
  102. bool "Kernel Function Tracer"
  103. depends on HAVE_FUNCTION_TRACER
  104. select FRAME_POINTER
  105. select KALLSYMS
  106. select GENERIC_TRACER
  107. select CONTEXT_SWITCH_TRACER
  108. help
  109. Enable the kernel to trace every kernel function. This is done
  110. by using a compiler feature to insert a small, 5-byte No-Operation
  111. instruction at the beginning of every kernel function, which NOP
  112. sequence is then dynamically patched into a tracer call when
  113. tracing is enabled by the administrator. If it's runtime disabled
  114. (the bootup default), then the overhead of the instructions is very
  115. small and not measurable even in micro-benchmarks.
  116. config FUNCTION_GRAPH_TRACER
  117. bool "Kernel Function Graph Tracer"
  118. depends on HAVE_FUNCTION_GRAPH_TRACER
  119. depends on FUNCTION_TRACER
  120. depends on !X86_32 || !CC_OPTIMIZE_FOR_SIZE
  121. default y
  122. help
  123. Enable the kernel to trace a function at both its return
  124. and its entry.
  125. Its first purpose is to trace the duration of functions and
  126. draw a call graph for each thread with some information like
  127. the return value. This is done by setting the current return
  128. address on the current task structure into a stack of calls.
  129. config IRQSOFF_TRACER
  130. bool "Interrupts-off Latency Tracer"
  131. default n
  132. depends on TRACE_IRQFLAGS_SUPPORT
  133. depends on GENERIC_TIME
  134. select TRACE_IRQFLAGS
  135. select GENERIC_TRACER
  136. select TRACER_MAX_TRACE
  137. select RING_BUFFER_ALLOW_SWAP
  138. help
  139. This option measures the time spent in irqs-off critical
  140. 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
  143. via:
  144. echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
  145. (Note that kernel size and overhead increase with this option
  146. enabled. This option and the preempt-off timing option can be
  147. used together or separately.)
  148. config PREEMPT_TRACER
  149. bool "Preemption-off Latency Tracer"
  150. default n
  151. depends on GENERIC_TIME
  152. depends on PREEMPT
  153. select GENERIC_TRACER
  154. select TRACER_MAX_TRACE
  155. select RING_BUFFER_ALLOW_SWAP
  156. help
  157. This option measures the time spent in preemption-off critical
  158. sections, with microsecond accuracy.
  159. The default measurement method is a maximum search, which is
  160. disabled by default and can be runtime (re-)started
  161. via:
  162. echo 0 > /sys/kernel/debug/tracing/tracing_max_latency
  163. (Note that kernel size and overhead increase with this option
  164. enabled. This option and the irqs-off timing option can be
  165. used together or separately.)
  166. config SYSPROF_TRACER
  167. bool "Sysprof Tracer"
  168. depends on X86
  169. select GENERIC_TRACER
  170. select CONTEXT_SWITCH_TRACER
  171. help
  172. This tracer provides the trace needed by the 'Sysprof' userspace
  173. tool.
  174. config SCHED_TRACER
  175. bool "Scheduling Latency Tracer"
  176. select GENERIC_TRACER
  177. select CONTEXT_SWITCH_TRACER
  178. select TRACER_MAX_TRACE
  179. help
  180. This tracer tracks the latency of the highest priority task
  181. to be scheduled in, starting from the point it has woken up.
  182. config ENABLE_DEFAULT_TRACERS
  183. bool "Trace process context switches and events"
  184. depends on !GENERIC_TRACER
  185. select TRACING
  186. help
  187. This tracer hooks to various trace points in the kernel,
  188. allowing the user to pick and choose which trace point they
  189. want to trace. It also includes the sched_switch tracer plugin.
  190. config FTRACE_SYSCALLS
  191. bool "Trace syscalls"
  192. depends on HAVE_SYSCALL_TRACEPOINTS
  193. select GENERIC_TRACER
  194. select KALLSYMS
  195. help
  196. Basic tracer to catch the syscall entry and exit events.
  197. config BOOT_TRACER
  198. bool "Trace boot initcalls"
  199. select GENERIC_TRACER
  200. select CONTEXT_SWITCH_TRACER
  201. help
  202. This tracer helps developers to optimize boot times: it records
  203. the timings of the initcalls and traces key events and the identity
  204. of tasks that can cause boot delays, such as context-switches.
  205. Its aim is to be parsed by the scripts/bootgraph.pl tool to
  206. produce pretty graphics about boot inefficiencies, giving a visual
  207. representation of the delays during initcalls - but the raw
  208. /debug/tracing/trace text output is readable too.
  209. You must pass in initcall_debug and ftrace=initcall to the kernel
  210. command line to enable this on bootup.
  211. config TRACE_BRANCH_PROFILING
  212. bool
  213. select GENERIC_TRACER
  214. choice
  215. prompt "Branch Profiling"
  216. default BRANCH_PROFILE_NONE
  217. help
  218. The branch profiling is a software profiler. It will add hooks
  219. into the C conditionals to test which path a branch takes.
  220. The likely/unlikely profiler only looks at the conditions that
  221. are annotated with a likely or unlikely macro.
  222. The "all branch" profiler will profile every if-statement in the
  223. kernel. This profiler will also enable the likely/unlikely
  224. profiler.
  225. Either of the above profilers adds a bit of overhead to the system.
  226. If unsure, choose "No branch profiling".
  227. config BRANCH_PROFILE_NONE
  228. bool "No branch profiling"
  229. help
  230. No branch profiling. Branch profiling adds a bit of overhead.
  231. Only enable it if you want to analyse the branching behavior.
  232. Otherwise keep it disabled.
  233. config PROFILE_ANNOTATED_BRANCHES
  234. bool "Trace likely/unlikely profiler"
  235. select TRACE_BRANCH_PROFILING
  236. help
  237. This tracer profiles all the the likely and unlikely macros
  238. in the kernel. It will display the results in:
  239. /sys/kernel/debug/tracing/profile_annotated_branch
  240. Note: this will add a significant overhead; only turn this
  241. on if you need to profile the system's use of these macros.
  242. config PROFILE_ALL_BRANCHES
  243. bool "Profile all if conditionals"
  244. select TRACE_BRANCH_PROFILING
  245. help
  246. This tracer profiles all branch conditions. Every if ()
  247. taken in the kernel is recorded whether it hit or miss.
  248. The results will be displayed in:
  249. /sys/kernel/debug/tracing/profile_branch
  250. This option also enables the likely/unlikely profiler.
  251. This configuration, when enabled, will impose a great overhead
  252. on the system. This should only be enabled when the system
  253. is to be analyzed in much detail.
  254. endchoice
  255. config TRACING_BRANCHES
  256. bool
  257. help
  258. Selected by tracers that will trace the likely and unlikely
  259. conditions. This prevents the tracers themselves from being
  260. profiled. Profiling the tracing infrastructure can only happen
  261. when the likelys and unlikelys are not being traced.
  262. config BRANCH_TRACER
  263. bool "Trace likely/unlikely instances"
  264. depends on TRACE_BRANCH_PROFILING
  265. select TRACING_BRANCHES
  266. help
  267. This traces the events of likely and unlikely condition
  268. calls in the kernel. The difference between this and the
  269. "Trace likely/unlikely profiler" is that this is not a
  270. histogram of the callers, but actually places the calling
  271. events into a running trace buffer to see when and where the
  272. events happened, as well as their results.
  273. Say N if unsure.
  274. config KSYM_TRACER
  275. bool "Trace read and write access on kernel memory locations"
  276. depends on HAVE_HW_BREAKPOINT
  277. select TRACING
  278. help
  279. This tracer helps find read and write operations on any given kernel
  280. symbol i.e. /proc/kallsyms.
  281. config PROFILE_KSYM_TRACER
  282. bool "Profile all kernel memory accesses on 'watched' variables"
  283. depends on KSYM_TRACER
  284. help
  285. This tracer profiles kernel accesses on variables watched through the
  286. ksym tracer ftrace plugin. Depending upon the hardware, all read
  287. and write operations on kernel variables can be monitored for
  288. accesses.
  289. The results will be displayed in:
  290. /debugfs/tracing/profile_ksym
  291. Say N if unsure.
  292. config STACK_TRACER
  293. bool "Trace max stack"
  294. depends on HAVE_FUNCTION_TRACER
  295. select FUNCTION_TRACER
  296. select STACKTRACE
  297. select KALLSYMS
  298. help
  299. This special tracer records the maximum stack footprint of the
  300. kernel and displays it in /sys/kernel/debug/tracing/stack_trace.
  301. This tracer works by hooking into every function call that the
  302. kernel executes, and keeping a maximum stack depth value and
  303. stack-trace saved. If this is configured with DYNAMIC_FTRACE
  304. then it will not have any overhead while the stack tracer
  305. is disabled.
  306. To enable the stack tracer on bootup, pass in 'stacktrace'
  307. on the kernel command line.
  308. The stack tracer can also be enabled or disabled via the
  309. sysctl kernel.stack_tracer_enabled
  310. Say N if unsure.
  311. config HW_BRANCH_TRACER
  312. depends on HAVE_HW_BRANCH_TRACER
  313. bool "Trace hw branches"
  314. select GENERIC_TRACER
  315. help
  316. This tracer records all branches on the system in a circular
  317. buffer, giving access to the last N branches for each cpu.
  318. config KMEMTRACE
  319. bool "Trace SLAB allocations"
  320. select GENERIC_TRACER
  321. help
  322. kmemtrace provides tracing for slab allocator functions, such as
  323. kmalloc, kfree, kmem_cache_alloc, kmem_cache_free, etc. Collected
  324. data is then fed to the userspace application in order to analyse
  325. allocation hotspots, internal fragmentation and so on, making it
  326. possible to see how well an allocator performs, as well as debug
  327. and profile kernel code.
  328. This requires an userspace application to use. See
  329. Documentation/trace/kmemtrace.txt for more information.
  330. Saying Y will make the kernel somewhat larger and slower. However,
  331. if you disable kmemtrace at run-time or boot-time, the performance
  332. impact is minimal (depending on the arch the kernel is built for).
  333. If unsure, say N.
  334. config WORKQUEUE_TRACER
  335. bool "Trace workqueues"
  336. select GENERIC_TRACER
  337. help
  338. The workqueue tracer provides some statistical information
  339. about each cpu workqueue thread such as the number of the
  340. works inserted and executed since their creation. It can help
  341. to evaluate the amount of work each of them has to perform.
  342. For example it can help a developer to decide whether he should
  343. choose a per-cpu workqueue instead of a singlethreaded one.
  344. config BLK_DEV_IO_TRACE
  345. bool "Support for tracing block IO actions"
  346. depends on SYSFS
  347. depends on BLOCK
  348. select RELAY
  349. select DEBUG_FS
  350. select TRACEPOINTS
  351. select GENERIC_TRACER
  352. select STACKTRACE
  353. help
  354. Say Y here if you want to be able to trace the block layer actions
  355. on a given queue. Tracing allows you to see any traffic happening
  356. on a block device queue. For more information (and the userspace
  357. support tools needed), fetch the blktrace tools from:
  358. git://git.kernel.dk/blktrace.git
  359. Tracing also is possible using the ftrace interface, e.g.:
  360. echo 1 > /sys/block/sda/sda1/trace/enable
  361. echo blk > /sys/kernel/debug/tracing/current_tracer
  362. cat /sys/kernel/debug/tracing/trace_pipe
  363. If unsure, say N.
  364. config KPROBE_EVENT
  365. depends on KPROBES
  366. depends on X86
  367. bool "Enable kprobes-based dynamic events"
  368. select TRACING
  369. default y
  370. help
  371. This allows the user to add tracing events (similar to tracepoints)
  372. on the fly via the ftrace interface. See
  373. Documentation/trace/kprobetrace.txt for more details.
  374. Those events can be inserted wherever kprobes can probe, and record
  375. various register and memory values.
  376. This option is also required by perf-probe subcommand of perf tools.
  377. If you want to use perf tools, this option is strongly recommended.
  378. config DYNAMIC_FTRACE
  379. bool "enable/disable ftrace tracepoints dynamically"
  380. depends on FUNCTION_TRACER
  381. depends on HAVE_DYNAMIC_FTRACE
  382. default y
  383. help
  384. This option will modify all the calls to ftrace dynamically
  385. (will patch them out of the binary image and replace them
  386. with a No-Op instruction) as they are called. A table is
  387. created to dynamically enable them again.
  388. This way a CONFIG_FUNCTION_TRACER kernel is slightly larger, but
  389. otherwise has native performance as long as no tracing is active.
  390. The changes to the code are done by a kernel thread that
  391. wakes up once a second and checks to see if any ftrace calls
  392. were made. If so, it runs stop_machine (stops all CPUS)
  393. and modifies the code to jump over the call to ftrace.
  394. config FUNCTION_PROFILER
  395. bool "Kernel function profiler"
  396. depends on FUNCTION_TRACER
  397. default n
  398. help
  399. This option enables the kernel function profiler. A file is created
  400. in debugfs called function_profile_enabled which defaults to zero.
  401. When a 1 is echoed into this file profiling begins, and when a
  402. zero is entered, profiling stops. A "functions" file is created in
  403. the trace_stats directory; this file shows the list of functions that
  404. have been hit and their counters.
  405. If in doubt, say N.
  406. config FTRACE_MCOUNT_RECORD
  407. def_bool y
  408. depends on DYNAMIC_FTRACE
  409. depends on HAVE_FTRACE_MCOUNT_RECORD
  410. config FTRACE_SELFTEST
  411. bool
  412. config FTRACE_STARTUP_TEST
  413. bool "Perform a startup test on ftrace"
  414. depends on GENERIC_TRACER
  415. select FTRACE_SELFTEST
  416. help
  417. This option performs a series of startup tests on ftrace. On bootup
  418. a series of tests are made to verify that the tracer is
  419. functioning properly. It will do tests on all the configured
  420. tracers of ftrace.
  421. config EVENT_TRACE_TEST_SYSCALLS
  422. bool "Run selftest on syscall events"
  423. depends on FTRACE_STARTUP_TEST
  424. help
  425. This option will also enable testing every syscall event.
  426. It only enables the event and disables it and runs various loads
  427. with the event enabled. This adds a bit more time for kernel boot
  428. up since it runs this on every system call defined.
  429. TBD - enable a way to actually call the syscalls as we test their
  430. events
  431. config MMIOTRACE
  432. bool "Memory mapped IO tracing"
  433. depends on HAVE_MMIOTRACE_SUPPORT && PCI
  434. select GENERIC_TRACER
  435. help
  436. Mmiotrace traces Memory Mapped I/O access and is meant for
  437. debugging and reverse engineering. It is called from the ioremap
  438. implementation and works via page faults. Tracing is disabled by
  439. default and can be enabled at run-time.
  440. See Documentation/trace/mmiotrace.txt.
  441. If you are not helping to develop drivers, say N.
  442. config MMIOTRACE_TEST
  443. tristate "Test module for mmiotrace"
  444. depends on MMIOTRACE && m
  445. help
  446. This is a dumb module for testing mmiotrace. It is very dangerous
  447. as it will write garbage to IO memory starting at a given address.
  448. However, it should be safe to use on e.g. unused portion of VRAM.
  449. Say N, unless you absolutely know what you are doing.
  450. config RING_BUFFER_BENCHMARK
  451. tristate "Ring buffer benchmark stress tester"
  452. depends on RING_BUFFER
  453. help
  454. This option creates a test to stress the ring buffer and benchmark it.
  455. It creates its own ring buffer such that it will not interfere with
  456. any other users of the ring buffer (such as ftrace). It then creates
  457. a producer and consumer that will run for 10 seconds and sleep for
  458. 10 seconds. Each interval it will print out the number of events
  459. it recorded and give a rough estimate of how long each iteration took.
  460. It does not disable interrupts or raise its priority, so it may be
  461. affected by processes that are running.
  462. If unsure, say N.
  463. endif # FTRACE
  464. endif # TRACING_SUPPORT