1menu "printk and dmesg options"
   2
   3config PRINTK_TIME
   4	bool "Show timing information on printks"
   5	depends on PRINTK
   6	help
   7	  Selecting this option causes time stamps of the printk()
   8	  messages to be added to the output of the syslog() system
   9	  call and at the console.
  10
  11	  The timestamp is always recorded internally, and exported
  12	  to /dev/kmsg. This flag just specifies if the timestamp should
  13	  be included, not that the timestamp is recorded.
  14
  15	  The behavior is also controlled by the kernel command line
  16	  parameter printk.time=1. See Documentation/kernel-parameters.txt
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
  17
  18config DEFAULT_MESSAGE_LOGLEVEL
 
 
 
 
  19	int "Default message log level (1-7)"
  20	range 1 7
  21	default "4"
  22	help
  23	  Default log level for printk statements with no specified priority.
  24
  25	  This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
  26	  that are auditing their logs closely may want to set it to a lower
  27	  priority.
  28
 
 
 
 
  29config BOOT_PRINTK_DELAY
  30	bool "Delay each boot printk message by N milliseconds"
  31	depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
  32	help
  33	  This build option allows you to read kernel boot messages
  34	  by inserting a short delay after each one.  The delay is
  35	  specified in milliseconds on the kernel command line,
  36	  using "boot_delay=N".
  37
  38	  It is likely that you would also need to use "lpj=M" to preset
  39	  the "loops per jiffie" value.
  40	  See a previous boot log for the "lpj" value to use for your
  41	  system, and then set "lpj=M" before setting "boot_delay=N".
  42	  NOTE:  Using this option may adversely affect SMP systems.
  43	  I.e., processors other than the first one may not boot up.
  44	  BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
  45	  what it believes to be lockup conditions.
  46
  47config DYNAMIC_DEBUG
  48	bool "Enable dynamic printk() support"
  49	default n
  50	depends on PRINTK
  51	depends on DEBUG_FS
 
  52	help
  53
  54	  Compiles debug level messages into the kernel, which would not
  55	  otherwise be available at runtime. These messages can then be
  56	  enabled/disabled based on various levels of scope - per source file,
  57	  function, module, format string, and line number. This mechanism
  58	  implicitly compiles in all pr_debug() and dev_dbg() calls, which
  59	  enlarges the kernel text size by about 2%.
  60
  61	  If a source file is compiled with DEBUG flag set, any
  62	  pr_debug() calls in it are enabled by default, but can be
  63	  disabled at runtime as below.  Note that DEBUG flag is
  64	  turned on by many CONFIG_*DEBUG* options.
  65
  66	  Usage:
  67
  68	  Dynamic debugging is controlled via the 'dynamic_debug/control' file,
  69	  which is contained in the 'debugfs' filesystem. Thus, the debugfs
  70	  filesystem must first be mounted before making use of this feature.
 
  71	  We refer the control file as: <debugfs>/dynamic_debug/control. This
  72	  file contains a list of the debug statements that can be enabled. The
  73	  format for each line of the file is:
  74
  75		filename:lineno [module]function flags format
  76
  77	  filename : source file of the debug statement
  78	  lineno : line number of the debug statement
  79	  module : module that contains the debug statement
  80	  function : function that contains the debug statement
  81          flags : '=p' means the line is turned 'on' for printing
  82          format : the format used for the debug statement
  83
  84	  From a live system:
  85
  86		nullarbor:~ # cat <debugfs>/dynamic_debug/control
  87		# filename:lineno [module]function flags format
  88		fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
  89		fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
  90		fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
  91
  92	  Example usage:
  93
  94		// enable the message at line 1603 of file svcsock.c
  95		nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
  96						<debugfs>/dynamic_debug/control
  97
  98		// enable all the messages in file svcsock.c
  99		nullarbor:~ # echo -n 'file svcsock.c +p' >
 100						<debugfs>/dynamic_debug/control
 101
 102		// enable all the messages in the NFS server module
 103		nullarbor:~ # echo -n 'module nfsd +p' >
 104						<debugfs>/dynamic_debug/control
 105
 106		// enable all 12 messages in the function svc_process()
 107		nullarbor:~ # echo -n 'func svc_process +p' >
 108						<debugfs>/dynamic_debug/control
 109
 110		// disable all 12 messages in the function svc_process()
 111		nullarbor:~ # echo -n 'func svc_process -p' >
 112						<debugfs>/dynamic_debug/control
 113
 114	  See Documentation/dynamic-debug-howto.txt for additional information.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 115
 116endmenu # "printk and dmesg options"
 117
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 118menu "Compile-time checks and compiler options"
 119
 120config DEBUG_INFO
 121	bool "Compile the kernel with debug info"
 122	depends on DEBUG_KERNEL && !COMPILE_TEST
 123	help
 124          If you say Y here the resulting kernel image will include
 125	  debugging info resulting in a larger kernel image.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 126	  This adds debug symbols to the kernel and modules (gcc -g), and
 127	  is needed if you intend to use kernel crashdump or binary object
 128	  tools like crash, kgdb, LKCD, gdb, etc on the kernel.
 129	  Say Y here only if you plan to debug the kernel.
 130
 131	  If unsure, say N.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 132
 133config DEBUG_INFO_REDUCED
 134	bool "Reduce debugging information"
 135	depends on DEBUG_INFO
 136	help
 137	  If you say Y here gcc is instructed to generate less debugging
 138	  information for structure types. This means that tools that
 139	  need full debugging information (like kgdb or systemtap) won't
 140	  be happy. But if you merely need debugging information to
 141	  resolve line numbers there is no loss. Advantage is that
 142	  build directory object sizes shrink dramatically over a full
 143	  DEBUG_INFO build and compile times are reduced too.
 144	  Only works with newer gcc versions.
 145
 146config ENABLE_WARN_DEPRECATED
 147	bool "Enable __deprecated logic"
 148	default y
 149	help
 150	  Enable the __deprecated logic in the kernel build.
 151	  Disable this to suppress the "warning: 'foo' is deprecated
 152	  (declared at kernel/power/somefile.c:1234)" messages.
 
 153
 154config ENABLE_MUST_CHECK
 155	bool "Enable __must_check logic"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 156	default y
 
 157	help
 158	  Enable the __must_check logic in the kernel build.  Disable this to
 159	  suppress the "warning: ignoring return value of 'foo', declared with
 160	  attribute warn_unused_result" messages.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 161
 162config FRAME_WARN
 163	int "Warn for stack frames larger than (needs gcc 4.4)"
 164	range 0 8192
 
 
 
 
 
 165	default 1024 if !64BIT
 166	default 2048 if 64BIT
 167	help
 168	  Tell gcc to warn at build time for stack frames larger than this.
 169	  Setting this too low will cause a lot of warnings.
 170	  Setting it to 0 disables the warning.
 171	  Requires gcc 4.4
 172
 173config STRIP_ASM_SYMS
 174	bool "Strip assembler-generated symbols during link"
 175	default n
 176	help
 177	  Strip internal assembler-generated symbols during a link (symbols
 178	  that look like '.Lxxx') so they don't pollute the output of
 179	  get_wchan() and suchlike.
 180
 181config READABLE_ASM
 182        bool "Generate readable assembler code"
 183        depends on DEBUG_KERNEL
 184        help
 185          Disable some compiler optimizations that tend to generate human unreadable
 186          assembler output. This may make the kernel slightly slower, but it helps
 187          to keep kernel developers who have to stare a lot at assembler listings
 188          sane.
 189
 190config UNUSED_SYMBOLS
 191	bool "Enable unused/obsolete exported symbols"
 192	default y if X86
 193	help
 194	  Unused but exported symbols make the kernel needlessly bigger.  For
 195	  that reason most of these unused exports will soon be removed.  This
 196	  option is provided temporarily to provide a transition period in case
 197	  some external kernel module needs one of these symbols anyway. If you
 198	  encounter such a case in your module, consider if you are actually
 199	  using the right API.  (rationale: since nobody in the kernel is using
 200	  this in a module, there is a pretty good chance it's actually the
 201	  wrong interface to use).  If you really need the symbol, please send a
 202	  mail to the linux kernel mailing list mentioning the symbol and why
 203	  you really need it, and what the merge plan to the mainline kernel for
 204	  your module is.
 205
 206config DEBUG_FS
 207	bool "Debug Filesystem"
 208	help
 209	  debugfs is a virtual file system that kernel developers use to put
 210	  debugging files into.  Enable this option to be able to read and
 211	  write to these files.
 212
 213	  For detailed documentation on the debugfs API, see
 214	  Documentation/DocBook/filesystems.
 215
 216	  If unsure, say N.
 217
 218config HEADERS_CHECK
 219	bool "Run 'make headers_check' when building vmlinux"
 220	depends on !UML
 221	help
 222	  This option will extract the user-visible kernel headers whenever
 223	  building the kernel, and will run basic sanity checks on them to
 224	  ensure that exported files do not attempt to include files which
 225	  were not exported, etc.
 226
 227	  If you're making modifications to header files which are
 228	  relevant for userspace, say 'Y', and check the headers
 229	  exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
 230	  your build tree), to make sure they're suitable.
 231
 232config DEBUG_SECTION_MISMATCH
 233	bool "Enable full Section mismatch analysis"
 
 234	help
 235	  The section mismatch analysis checks if there are illegal
 236	  references from one section to another section.
 237	  During linktime or runtime, some sections are dropped;
 238	  any use of code/data previously in these sections would
 239	  most likely result in an oops.
 240	  In the code, functions and variables are annotated with
 241	  __init,, etc. (see the full list in include/linux/init.h),
 242	  which results in the code/data being placed in specific sections.
 243	  The section mismatch analysis is always performed after a full
 244	  kernel build, and enabling this option causes the following
 245	  additional steps to occur:
 246	  - Add the option -fno-inline-functions-called-once to gcc commands.
 247	    When inlining a function annotated with __init in a non-init
 248	    function, we would lose the section information and thus
 249	    the analysis would not catch the illegal reference.
 250	    This option tells gcc to inline less (but it does result in
 251	    a larger kernel).
 252	  - Run the section mismatch analysis for each module/built-in.o file.
 253	    When we run the section mismatch analysis on vmlinux.o, we
 254	    lose valueble information about where the mismatch was
 255	    introduced.
 256	    Running the analysis for each module/built-in.o file
 257	    tells where the mismatch happens much closer to the
 258	    source. The drawback is that the same mismatch is
 259	    reported at least twice.
 260	  - Enable verbose reporting from modpost in order to help resolve
 261	    the section mismatches that are reported.
 
 
 
 
 
 
 
 
 
 
 
 
 262
 263#
 264# Select this config option from the architecture Kconfig, if it
 265# is preferred to always offer frame pointers as a config
 266# option on the architecture (regardless of KERNEL_DEBUG):
 267#
 268config ARCH_WANT_FRAME_POINTERS
 269	bool
 270	help
 271
 272config FRAME_POINTER
 273	bool "Compile the kernel with frame pointers"
 274	depends on DEBUG_KERNEL && \
 275		(CRIS || M68K || FRV || UML || \
 276		 AVR32 || SUPERH || BLACKFIN || MN10300 || METAG) || \
 277		ARCH_WANT_FRAME_POINTERS
 278	default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
 279	help
 280	  If you say Y here the resulting kernel image will be slightly
 281	  larger and slower, but it gives very useful debugging information
 282	  in case of kernel bugs. (precise oopses/stacktraces/warnings)
 283
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 284config DEBUG_FORCE_WEAK_PER_CPU
 285	bool "Force weak per-cpu definitions"
 286	depends on DEBUG_KERNEL
 287	help
 288	  s390 and alpha require percpu variables in modules to be
 289	  defined weak to work around addressing range issue which
 290	  puts the following two restrictions on percpu variable
 291	  definitions.
 292
 293	  1. percpu symbols must be unique whether static or not
 294	  2. percpu variables can't be defined inside a function
 295
 296	  To ensure that generic code follows the above rules, this
 297	  option forces all percpu variables to be defined as weak.
 298
 299endmenu # "Compiler options"
 300
 
 
 301config MAGIC_SYSRQ
 302	bool "Magic SysRq key"
 303	depends on !UML
 304	help
 305	  If you say Y here, you will have some control over the system even
 306	  if the system crashes for example during kernel debugging (e.g., you
 307	  will be able to flush the buffer cache to disk, reboot the system
 308	  immediately or dump some status information). This is accomplished
 309	  by pressing various keys while holding SysRq (Alt+PrintScreen). It
 310	  also works on a serial console (on PC hardware at least), if you
 311	  send a BREAK and then within 5 seconds a command keypress. The
 312	  keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
 313	  unless you really know what this hack does.
 314
 315config MAGIC_SYSRQ_DEFAULT_ENABLE
 316	hex "Enable magic SysRq key functions by default"
 317	depends on MAGIC_SYSRQ
 318	default 0x1
 319	help
 320	  Specifies which SysRq key functions are enabled by default.
 321	  This may be set to 1 or 0 to enable or disable them all, or
 322	  to a bitmask as described in Documentation/sysrq.txt.
 323
 324config DEBUG_KERNEL
 325	bool "Kernel debugging"
 
 
 326	help
 327	  Say Y here if you are developing drivers or trying to debug and
 328	  identify kernel problems.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 329
 330menu "Memory Debugging"
 331
 332source mm/Kconfig.debug
 333
 334config DEBUG_OBJECTS
 335	bool "Debug object operations"
 336	depends on DEBUG_KERNEL
 337	help
 338	  If you say Y here, additional code will be inserted into the
 339	  kernel to track the life time of various objects and validate
 340	  the operations on those objects.
 341
 342config DEBUG_OBJECTS_SELFTEST
 343	bool "Debug objects selftest"
 344	depends on DEBUG_OBJECTS
 345	help
 346	  This enables the selftest of the object debug code.
 347
 348config DEBUG_OBJECTS_FREE
 349	bool "Debug objects in freed memory"
 350	depends on DEBUG_OBJECTS
 351	help
 352	  This enables checks whether a k/v free operation frees an area
 353	  which contains an object which has not been deactivated
 354	  properly. This can make kmalloc/kfree-intensive workloads
 355	  much slower.
 356
 357config DEBUG_OBJECTS_TIMERS
 358	bool "Debug timer objects"
 359	depends on DEBUG_OBJECTS
 360	help
 361	  If you say Y here, additional code will be inserted into the
 362	  timer routines to track the life time of timer objects and
 363	  validate the timer operations.
 364
 365config DEBUG_OBJECTS_WORK
 366	bool "Debug work objects"
 367	depends on DEBUG_OBJECTS
 368	help
 369	  If you say Y here, additional code will be inserted into the
 370	  work queue routines to track the life time of work objects and
 371	  validate the work operations.
 372
 373config DEBUG_OBJECTS_RCU_HEAD
 374	bool "Debug RCU callbacks objects"
 375	depends on DEBUG_OBJECTS
 376	help
 377	  Enable this to turn on debugging of RCU list heads (call_rcu() usage).
 378
 379config DEBUG_OBJECTS_PERCPU_COUNTER
 380	bool "Debug percpu counter objects"
 381	depends on DEBUG_OBJECTS
 382	help
 383	  If you say Y here, additional code will be inserted into the
 384	  percpu counter routines to track the life time of percpu counter
 385	  objects and validate the percpu counter operations.
 386
 387config DEBUG_OBJECTS_ENABLE_DEFAULT
 388	int "debug_objects bootup default value (0-1)"
 389        range 0 1
 390        default "1"
 391        depends on DEBUG_OBJECTS
 392        help
 393          Debug objects boot parameter default value
 394
 395config DEBUG_SLAB
 396	bool "Debug slab memory allocations"
 397	depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
 398	help
 399	  Say Y here to have the kernel do limited verification on memory
 400	  allocation as well as poisoning memory on free to catch use of freed
 401	  memory. This can make kmalloc/kfree-intensive workloads much slower.
 402
 403config DEBUG_SLAB_LEAK
 404	bool "Memory leak debugging"
 405	depends on DEBUG_SLAB
 406
 407config SLUB_DEBUG_ON
 408	bool "SLUB debugging on by default"
 409	depends on SLUB && SLUB_DEBUG && !KMEMCHECK
 410	default n
 411	help
 412	  Boot with debugging on by default. SLUB boots by default with
 413	  the runtime debug capabilities switched off. Enabling this is
 414	  equivalent to specifying the "slub_debug" parameter on boot.
 415	  There is no support for more fine grained debug control like
 416	  possible with slub_debug=xxx. SLUB debugging may be switched
 417	  off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
 418	  "slub_debug=-".
 419
 420config SLUB_STATS
 421	default n
 422	bool "Enable SLUB performance statistics"
 423	depends on SLUB && SYSFS
 424	help
 425	  SLUB statistics are useful to debug SLUBs allocation behavior in
 426	  order find ways to optimize the allocator. This should never be
 427	  enabled for production use since keeping statistics slows down
 428	  the allocator by a few percentage points. The slabinfo command
 429	  supports the determination of the most active slabs to figure
 430	  out which slabs are relevant to a particular load.
 431	  Try running: slabinfo -DA
 432
 433config HAVE_DEBUG_KMEMLEAK
 434	bool
 435
 436config DEBUG_KMEMLEAK
 437	bool "Kernel memory leak detector"
 438	depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
 439	select DEBUG_FS
 440	select STACKTRACE if STACKTRACE_SUPPORT
 441	select KALLSYMS
 442	select CRC32
 443	help
 444	  Say Y here if you want to enable the memory leak
 445	  detector. The memory allocation/freeing is traced in a way
 446	  similar to the Boehm's conservative garbage collector, the
 447	  difference being that the orphan objects are not freed but
 448	  only shown in /sys/kernel/debug/kmemleak. Enabling this
 449	  feature will introduce an overhead to memory
 450	  allocations. See Documentation/kmemleak.txt for more
 451	  details.
 452
 453	  Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
 454	  of finding leaks due to the slab objects poisoning.
 455
 456	  In order to access the kmemleak file, debugfs needs to be
 457	  mounted (usually at /sys/kernel/debug).
 458
 459config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
 460	int "Maximum kmemleak early log entries"
 461	depends on DEBUG_KMEMLEAK
 462	range 200 40000
 463	default 400
 464	help
 465	  Kmemleak must track all the memory allocations to avoid
 466	  reporting false positives. Since memory may be allocated or
 467	  freed before kmemleak is initialised, an early log buffer is
 468	  used to store these actions. If kmemleak reports "early log
 469	  buffer exceeded", please increase this value.
 470
 471config DEBUG_KMEMLEAK_TEST
 472	tristate "Simple test for the kernel memory leak detector"
 473	depends on DEBUG_KMEMLEAK && m
 474	help
 475	  This option enables a module that explicitly leaks memory.
 476
 477	  If unsure, say N.
 478
 479config DEBUG_KMEMLEAK_DEFAULT_OFF
 480	bool "Default kmemleak to off"
 481	depends on DEBUG_KMEMLEAK
 482	help
 483	  Say Y here to disable kmemleak by default. It can then be enabled
 484	  on the command line via kmemleak=on.
 
 485
 486config DEBUG_STACK_USAGE
 487	bool "Stack utilization instrumentation"
 488	depends on DEBUG_KERNEL && !IA64 && !PARISC && !METAG
 489	help
 490	  Enables the display of the minimum amount of free stack which each
 491	  task has ever had available in the sysrq-T and sysrq-P debug output.
 
 
 492
 493	  This option will slow down process creation somewhat.
 494
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 495config DEBUG_VM
 496	bool "Debug VM"
 497	depends on DEBUG_KERNEL
 498	help
 499	  Enable this to turn on extended checks in the virtual-memory system
 500          that may impact performance.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 501
 502	  If unsure, say N.
 503
 504config DEBUG_VM_RB
 505	bool "Debug VM red-black trees"
 506	depends on DEBUG_VM
 507	help
 508	  Enable VM red-black tree debugging information and extra validations.
 509
 510	  If unsure, say N.
 511
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 512config DEBUG_VIRTUAL
 513	bool "Debug VM translations"
 514	depends on DEBUG_KERNEL && X86
 515	help
 516	  Enable some costly sanity checks in virtual to page code. This can
 517	  catch mistakes with virt_to_page() and friends.
 518
 519	  If unsure, say N.
 520
 521config DEBUG_NOMMU_REGIONS
 522	bool "Debug the global anon/private NOMMU mapping region tree"
 523	depends on DEBUG_KERNEL && !MMU
 524	help
 525	  This option causes the global tree of anonymous and private mapping
 526	  regions to be regularly checked for invalid topology.
 527
 528config DEBUG_MEMORY_INIT
 529	bool "Debug memory initialisation" if EXPERT
 530	default !EXPERT
 531	help
 532	  Enable this for additional checks during memory initialisation.
 533	  The sanity checks verify aspects of the VM such as the memory model
 534	  and other information provided by the architecture. Verbose
 535	  information will be printed at KERN_DEBUG loglevel depending
 536	  on the mminit_loglevel= command-line option.
 537
 538	  If unsure, say Y
 539
 540config MEMORY_NOTIFIER_ERROR_INJECT
 541	tristate "Memory hotplug notifier error injection module"
 542	depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
 543	help
 544	  This option provides the ability to inject artificial errors to
 545	  memory hotplug notifier chain callbacks.  It is controlled through
 546	  debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
 547
 548	  If the notifier call chain should be failed with some events
 549	  notified, write the error code to "actions/<notifier event>/error".
 550
 551	  Example: Inject memory hotplug offline error (-12 == -ENOMEM)
 552
 553	  # cd /sys/kernel/debug/notifier-error-inject/memory
 554	  # echo -12 > actions/MEM_GOING_OFFLINE/error
 555	  # echo offline > /sys/devices/system/memory/memoryXXX/state
 556	  bash: echo: write error: Cannot allocate memory
 557
 558	  To compile this code as a module, choose M here: the module will
 559	  be called memory-notifier-error-inject.
 560
 561	  If unsure, say N.
 562
 563config DEBUG_PER_CPU_MAPS
 564	bool "Debug access to per_cpu maps"
 565	depends on DEBUG_KERNEL
 566	depends on SMP
 567	help
 568	  Say Y to verify that the per_cpu map being accessed has
 569	  been set up. This adds a fair amount of code to kernel memory
 570	  and decreases performance.
 571
 572	  Say N if unsure.
 573
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 574config DEBUG_HIGHMEM
 575	bool "Highmem debugging"
 576	depends on DEBUG_KERNEL && HIGHMEM
 
 
 577	help
 578	  This options enables addition error checking for high memory systems.
 579	  Disable for production systems.
 580
 581config HAVE_DEBUG_STACKOVERFLOW
 582	bool
 583
 584config DEBUG_STACKOVERFLOW
 585	bool "Check for stack overflows"
 586	depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
 587	---help---
 588	  Say Y here if you want to check for overflows of kernel, IRQ
 589	  and exception stacks (if your archicture uses them). This
 590	  option will show detailed messages if free stack space drops
 591	  below a certain limit.
 592
 593	  These kinds of bugs usually occur when call-chains in the
 594	  kernel get too deep, especially when interrupts are
 595	  involved.
 596
 597	  Use this in cases where you see apparently random memory
 598	  corruption, especially if it appears in 'struct thread_info'
 599
 600	  If in doubt, say "N".
 601
 602source "lib/Kconfig.kmemcheck"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 603
 604endmenu # "Memory Debugging"
 605
 606config DEBUG_SHIRQ
 607	bool "Debug shared IRQ handlers"
 608	depends on DEBUG_KERNEL
 609	help
 610	  Enable this to generate a spurious interrupt as soon as a shared
 611	  interrupt handler is registered, and just before one is deregistered.
 612	  Drivers ought to be able to handle interrupts coming in at those
 613	  points; some don't and need to be caught.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 614
 615menu "Debug Lockups and Hangs"
 
 
 
 
 
 
 
 
 
 616
 617config LOCKUP_DETECTOR
 618	bool "Detect Hard and Soft Lockups"
 
 
 
 619	depends on DEBUG_KERNEL && !S390
 
 620	help
 621	  Say Y here to enable the kernel to act as a watchdog to detect
 622	  hard and soft lockups.
 623
 624	  Softlockups are bugs that cause the kernel to loop in kernel
 625	  mode for more than 20 seconds, without giving other tasks a
 626	  chance to run.  The current stack trace is displayed upon
 627	  detection and the system will stay locked up.
 628
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 629	  Hardlockups are bugs that cause the CPU to loop in kernel mode
 630	  for more than 10 seconds, without letting other interrupts have a
 631	  chance to run.  The current stack trace is displayed upon detection
 632	  and the system will stay locked up.
 633
 634	  The overhead should be minimal.  A periodic hrtimer runs to
 635	  generate interrupts and kick the watchdog task every 4 seconds.
 636	  An NMI is generated every 10 seconds or so to check for hardlockups.
 
 
 
 
 
 
 
 637
 638	  The frequency of hrtimer and NMI events and the soft and hard lockup
 639	  thresholds can be controlled through the sysctl watchdog_thresh.
 
 640
 641config HARDLOCKUP_DETECTOR
 642	def_bool y
 643	depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
 644	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 645
 646config BOOTPARAM_HARDLOCKUP_PANIC
 647	bool "Panic (Reboot) On Hard Lockups"
 648	depends on HARDLOCKUP_DETECTOR
 649	help
 650	  Say Y here to enable the kernel to panic on "hard lockups",
 651	  which are bugs that cause the kernel to loop in kernel
 652	  mode with interrupts disabled for more than 10 seconds (configurable
 653	  using the watchdog_thresh sysctl).
 654
 655	  Say N if unsure.
 656
 657config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
 658	int
 659	depends on HARDLOCKUP_DETECTOR
 660	range 0 1
 661	default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
 662	default 1 if BOOTPARAM_HARDLOCKUP_PANIC
 663
 664config BOOTPARAM_SOFTLOCKUP_PANIC
 665	bool "Panic (Reboot) On Soft Lockups"
 666	depends on LOCKUP_DETECTOR
 667	help
 668	  Say Y here to enable the kernel to panic on "soft lockups",
 669	  which are bugs that cause the kernel to loop in kernel
 670	  mode for more than 20 seconds (configurable using the watchdog_thresh
 671	  sysctl), without giving other tasks a chance to run.
 672
 673	  The panic can be used in combination with panic_timeout,
 674	  to cause the system to reboot automatically after a
 675	  lockup has been detected. This feature is useful for
 676	  high-availability systems that have uptime guarantees and
 677	  where a lockup must be resolved ASAP.
 678
 679	  Say N if unsure.
 680
 681config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
 682	int
 683	depends on LOCKUP_DETECTOR
 684	range 0 1
 685	default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
 686	default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
 687
 688config DETECT_HUNG_TASK
 689	bool "Detect Hung Tasks"
 690	depends on DEBUG_KERNEL
 691	default LOCKUP_DETECTOR
 692	help
 693	  Say Y here to enable the kernel to detect "hung tasks",
 694	  which are bugs that cause the task to be stuck in
 695	  uninterruptible "D" state indefinitiley.
 696
 697	  When a hung task is detected, the kernel will print the
 698	  current stack trace (which you should report), but the
 699	  task will stay in uninterruptible state. If lockdep is
 700	  enabled then all held locks will also be reported. This
 701	  feature has negligible overhead.
 702
 703config DEFAULT_HUNG_TASK_TIMEOUT
 704	int "Default timeout for hung task detection (in seconds)"
 705	depends on DETECT_HUNG_TASK
 706	default 120
 707	help
 708	  This option controls the default timeout (in seconds) used
 709	  to determine when a task has become non-responsive and should
 710	  be considered hung.
 711
 712	  It can be adjusted at runtime via the kernel.hung_task_timeout_secs
 713	  sysctl or by writing a value to
 714	  /proc/sys/kernel/hung_task_timeout_secs.
 715
 716	  A timeout of 0 disables the check.  The default is two minutes.
 717	  Keeping the default should be fine in most cases.
 718
 719config BOOTPARAM_HUNG_TASK_PANIC
 720	bool "Panic (Reboot) On Hung Tasks"
 721	depends on DETECT_HUNG_TASK
 722	help
 723	  Say Y here to enable the kernel to panic on "hung tasks",
 724	  which are bugs that cause the kernel to leave a task stuck
 725	  in uninterruptible "D" state.
 726
 727	  The panic can be used in combination with panic_timeout,
 728	  to cause the system to reboot automatically after a
 729	  hung task has been detected. This feature is useful for
 730	  high-availability systems that have uptime guarantees and
 731	  where a hung tasks must be resolved ASAP.
 732
 733	  Say N if unsure.
 734
 735config BOOTPARAM_HUNG_TASK_PANIC_VALUE
 736	int
 737	depends on DETECT_HUNG_TASK
 738	range 0 1
 739	default 0 if !BOOTPARAM_HUNG_TASK_PANIC
 740	default 1 if BOOTPARAM_HUNG_TASK_PANIC
 741
 742endmenu # "Debug lockups and hangs"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 743
 744config PANIC_ON_OOPS
 745	bool "Panic on Oops"
 
 746	help
 747	  Say Y here to enable the kernel to panic when it oopses. This
 748	  has the same effect as setting oops=panic on the kernel command
 749	  line.
 750
 751	  This feature is useful to ensure that the kernel does not do
 752	  anything erroneous after an oops which could result in data
 753	  corruption or other issues.
 754
 755	  Say N if unsure.
 756
 757config PANIC_ON_OOPS_VALUE
 758	int
 759	range 0 1
 760	default 0 if !PANIC_ON_OOPS
 761	default 1 if PANIC_ON_OOPS
 762
 763config PANIC_TIMEOUT
 764	int "panic timeout"
 765	default 0
 766	help
 767	  Set the timeout value (in seconds) until a reboot occurs when the
 768	  the kernel panics. If n = 0, then we wait forever. A timeout
 769	  value n > 0 will wait n seconds before rebooting, while a timeout
 770	  value n < 0 will reboot immediately.
 771
 772config SCHED_DEBUG
 773	bool "Collect scheduler debugging info"
 774	depends on DEBUG_KERNEL && PROC_FS
 775	default y
 776	help
 777	  If you say Y here, the /proc/sched_debug file will be provided
 778	  that can help debug the scheduler. The runtime overhead of this
 779	  option is minimal.
 780
 
 
 
 
 781config SCHEDSTATS
 782	bool "Collect scheduler statistics"
 783	depends on DEBUG_KERNEL && PROC_FS
 
 784	help
 785	  If you say Y here, additional code will be inserted into the
 786	  scheduler and related routines to collect statistics about
 787	  scheduler behavior and provide them in /proc/schedstat.  These
 788	  stats may be useful for both tuning and debugging the scheduler
 789	  If you aren't debugging the scheduler or trying to tune a specific
 790	  application, you can say N to avoid the very slight overhead
 791	  this adds.
 792
 793config TIMER_STATS
 794	bool "Collect kernel timers statistics"
 795	depends on DEBUG_KERNEL && PROC_FS
 796	help
 797	  If you say Y here, additional code will be inserted into the
 798	  timer routines to collect statistics about kernel timers being
 799	  reprogrammed. The statistics can be read from /proc/timer_stats.
 800	  The statistics collection is started by writing 1 to /proc/timer_stats,
 801	  writing 0 stops it. This feature is useful to collect information
 802	  about timer usage patterns in kernel and userspace. This feature
 803	  is lightweight if enabled in the kernel config but not activated
 804	  (it defaults to deactivated on bootup and will only be activated
 805	  if some application like powertop activates it explicitly).
 806
 807config DEBUG_PREEMPT
 808	bool "Debug preemptible kernel"
 809	depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
 810	default y
 811	help
 812	  If you say Y here then the kernel will use a debug variant of the
 813	  commonly used smp_processor_id() function and will print warnings
 814	  if kernel code uses it in a preemption-unsafe way. Also, the kernel
 815	  will detect preemption count underflows.
 816
 817menu "Lock Debugging (spinlocks, mutexes, etc...)"
 
 
 818
 819config DEBUG_RT_MUTEXES
 820	bool "RT Mutex debugging, deadlock detection"
 821	depends on DEBUG_KERNEL && RT_MUTEXES
 822	help
 823	 This allows rt mutex semantics violations and rt mutex related
 824	 deadlocks (lockups) to be detected and reported automatically.
 825
 826config DEBUG_PI_LIST
 827	bool
 
 828	default y
 829	depends on DEBUG_RT_MUTEXES
 830
 831config RT_MUTEX_TESTER
 832	bool "Built-in scriptable tester for rt-mutexes"
 833	depends on DEBUG_KERNEL && RT_MUTEXES
 834	help
 835	  This option enables a rt-mutex tester.
 836
 837config DEBUG_SPINLOCK
 838	bool "Spinlock and rw-lock debugging: basic checks"
 839	depends on DEBUG_KERNEL
 840	select UNINLINE_SPIN_UNLOCK
 841	help
 842	  Say Y here and build SMP to catch missing spinlock initialization
 843	  and certain other kinds of spinlock errors commonly made.  This is
 844	  best used in conjunction with the NMI watchdog so that spinlock
 845	  deadlocks are also debuggable.
 846
 847config DEBUG_MUTEXES
 848	bool "Mutex debugging: basic checks"
 849	depends on DEBUG_KERNEL
 850	help
 851	 This feature allows mutex semantics violations to be detected and
 852	 reported.
 853
 854config DEBUG_WW_MUTEX_SLOWPATH
 855	bool "Wait/wound mutex debugging: Slowpath testing"
 856	depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
 857	select DEBUG_LOCK_ALLOC
 858	select DEBUG_SPINLOCK
 859	select DEBUG_MUTEXES
 860	help
 861	 This feature enables slowpath testing for w/w mutex users by
 862	 injecting additional -EDEADLK wound/backoff cases. Together with
 863	 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this
 864	 will test all possible w/w mutex interface abuse with the
 865	 exception of simply not acquiring all the required locks.
 866
 867config DEBUG_LOCK_ALLOC
 868	bool "Lock debugging: detect incorrect freeing of live locks"
 869	depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
 870	select DEBUG_SPINLOCK
 871	select DEBUG_MUTEXES
 872	select LOCKDEP
 873	help
 874	 This feature will check whether any held lock (spinlock, rwlock,
 875	 mutex or rwsem) is incorrectly freed by the kernel, via any of the
 876	 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
 877	 vfree(), etc.), whether a live lock is incorrectly reinitialized via
 878	 spin_lock_init()/mutex_init()/etc., or whether there is any lock
 879	 held during task exit.
 880
 881config PROVE_LOCKING
 882	bool "Lock debugging: prove locking correctness"
 883	depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
 884	select LOCKDEP
 885	select DEBUG_SPINLOCK
 886	select DEBUG_MUTEXES
 
 
 
 887	select DEBUG_LOCK_ALLOC
 
 888	select TRACE_IRQFLAGS
 889	default n
 890	help
 891	 This feature enables the kernel to prove that all locking
 892	 that occurs in the kernel runtime is mathematically
 893	 correct: that under no circumstance could an arbitrary (and
 894	 not yet triggered) combination of observed locking
 895	 sequences (on an arbitrary number of CPUs, running an
 896	 arbitrary number of tasks and interrupt contexts) cause a
 897	 deadlock.
 898
 899	 In short, this feature enables the kernel to report locking
 900	 related deadlocks before they actually occur.
 901
 902	 The proof does not depend on how hard and complex a
 903	 deadlock scenario would be to trigger: how many
 904	 participant CPUs, tasks and irq-contexts would be needed
 905	 for it to trigger. The proof also does not depend on
 906	 timing: if a race and a resulting deadlock is possible
 907	 theoretically (no matter how unlikely the race scenario
 908	 is), it will be proven so and will immediately be
 909	 reported by the kernel (once the event is observed that
 910	 makes the deadlock theoretically possible).
 911
 912	 If a deadlock is impossible (i.e. the locking rules, as
 913	 observed by the kernel, are mathematically correct), the
 914	 kernel reports nothing.
 915
 916	 NOTE: this feature can also be enabled for rwlocks, mutexes
 917	 and rwsems - in which case all dependencies between these
 918	 different locking variants are observed and mapped too, and
 919	 the proof of observed correctness is also maintained for an
 920	 arbitrary combination of these separate locking variants.
 921
 922	 For more details, see Documentation/lockdep-design.txt.
 923
 924config LOCKDEP
 925	bool
 926	depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
 927	select STACKTRACE
 928	select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE && !ARC
 929	select KALLSYMS
 930	select KALLSYMS_ALL
 
 931
 932config LOCK_STAT
 933	bool "Lock usage statistics"
 934	depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
 935	select LOCKDEP
 936	select DEBUG_SPINLOCK
 937	select DEBUG_MUTEXES
 
 938	select DEBUG_LOCK_ALLOC
 939	default n
 940	help
 941	 This feature enables tracking lock contention points
 942
 943	 For more details, see Documentation/lockstat.txt
 944
 945	 This also enables lock events required by "perf lock",
 946	 subcommand of perf.
 947	 If you want to use "perf lock", you also need to turn on
 948	 CONFIG_EVENT_TRACING.
 949
 950	 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
 951	 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
 952
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 953config DEBUG_LOCKDEP
 954	bool "Lock dependency engine debugging"
 955	depends on DEBUG_KERNEL && LOCKDEP
 
 956	help
 957	  If you say Y here, the lock dependency engine will do
 958	  additional runtime checks to debug itself, at the price
 959	  of more runtime overhead.
 960
 961config DEBUG_ATOMIC_SLEEP
 962	bool "Sleep inside atomic section checking"
 963	select PREEMPT_COUNT
 964	depends on DEBUG_KERNEL
 
 965	help
 966	  If you say Y here, various routines which may sleep will become very
 967	  noisy if they are called inside atomic sections: when a spinlock is
 968	  held, inside an rcu read side critical section, inside preempt disabled
 969	  sections, inside an interrupt, etc...
 970
 971config DEBUG_LOCKING_API_SELFTESTS
 972	bool "Locking API boot-time self-tests"
 973	depends on DEBUG_KERNEL
 974	help
 975	  Say Y here if you want the kernel to run a short self-test during
 976	  bootup. The self-test checks whether common types of locking bugs
 977	  are detected by debugging mechanisms or not. (if you disable
 978	  lock debugging then those bugs wont be detected of course.)
 979	  The following locking APIs are covered: spinlocks, rwlocks,
 980	  mutexes and rwsems.
 981
 982config LOCK_TORTURE_TEST
 983	tristate "torture tests for locking"
 984	depends on DEBUG_KERNEL
 985	select TORTURE_TEST
 986	default n
 987	help
 988	  This option provides a kernel module that runs torture tests
 989	  on kernel locking primitives.  The kernel module may be built
 990	  after the fact on the running kernel to be tested, if desired.
 991
 992	  Say Y here if you want kernel locking-primitive torture tests
 993	  to be built into the kernel.
 994	  Say M if you want these torture tests to build as a module.
 995	  Say N if you are unsure.
 996
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 997endmenu # lock debugging
 998
 999config TRACE_IRQFLAGS
 
1000	bool
1001	help
1002	  Enables hooks to interrupt enabling and disabling for
1003	  either tracing or lock debugging.
1004
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1005config STACKTRACE
1006	bool
1007	depends on STACKTRACE_SUPPORT
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1008
1009config DEBUG_KOBJECT
1010	bool "kobject debugging"
1011	depends on DEBUG_KERNEL
1012	help
1013	  If you say Y here, some extra kobject debugging messages will be sent
1014	  to the syslog. 
1015
1016config DEBUG_KOBJECT_RELEASE
1017	bool "kobject release debugging"
1018	depends on DEBUG_OBJECTS_TIMERS
1019	help
1020	  kobjects are reference counted objects.  This means that their
1021	  last reference count put is not predictable, and the kobject can
1022	  live on past the point at which a driver decides to drop it's
1023	  initial reference to the kobject gained on allocation.  An
1024	  example of this would be a struct device which has just been
1025	  unregistered.
1026
1027	  However, some buggy drivers assume that after such an operation,
1028	  the memory backing the kobject can be immediately freed.  This
1029	  goes completely against the principles of a refcounted object.
1030
1031	  If you say Y here, the kernel will delay the release of kobjects
1032	  on the last reference count to improve the visibility of this
1033	  kind of kobject release bug.
1034
1035config HAVE_DEBUG_BUGVERBOSE
1036	bool
1037
1038config DEBUG_BUGVERBOSE
1039	bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
1040	depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
1041	default y
1042	help
1043	  Say Y here to make BUG() panics output the file name and line number
1044	  of the BUG call as well as the EIP and oops trace.  This aids
1045	  debugging but costs about 70-100K of memory.
1046
1047config DEBUG_LIST
1048	bool "Debug linked list manipulation"
1049	depends on DEBUG_KERNEL
 
1050	help
1051	  Enable this to turn on extended checks in the linked-list
1052	  walking routines.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1053
1054	  If unsure, say N.
1055
1056config DEBUG_SG
1057	bool "Debug SG table operations"
1058	depends on DEBUG_KERNEL
1059	help
1060	  Enable this to turn on checks on scatter-gather tables. This can
1061	  help find problems with drivers that do not properly initialize
1062	  their sg tables.
1063
1064	  If unsure, say N.
1065
1066config DEBUG_NOTIFIERS
1067	bool "Debug notifier call chains"
1068	depends on DEBUG_KERNEL
1069	help
1070	  Enable this to turn on sanity checking for notifier call chains.
1071	  This is most useful for kernel developers to make sure that
1072	  modules properly unregister themselves from notifier chains.
1073	  This is a relatively cheap check but if you care about maximum
1074	  performance, say N.
1075
1076config DEBUG_CREDENTIALS
1077	bool "Debug credential management"
1078	depends on DEBUG_KERNEL
 
1079	help
1080	  Enable this to turn on some debug checking for credential
1081	  management.  The additional code keeps track of the number of
1082	  pointers from task_structs to any given cred struct, and checks to
1083	  see that this number never exceeds the usage count of the cred
1084	  struct.
1085
1086	  Furthermore, if SELinux is enabled, this also checks that the
1087	  security pointer in the cred struct is never seen to be invalid.
 
 
 
1088
1089	  If unsure, say N.
1090
1091menu "RCU Debugging"
1092
1093config PROVE_RCU
1094	bool "RCU debugging: prove RCU correctness"
1095	depends on PROVE_LOCKING
1096	default n
1097	help
1098	 This feature enables lockdep extensions that check for correct
1099	 use of RCU APIs.  This is currently under development.  Say Y
1100	 if you want to debug RCU usage or help work on the PROVE_RCU
1101	 feature.
1102
1103	 Say N if you are unsure.
1104
1105config PROVE_RCU_REPEATEDLY
1106	bool "RCU debugging: don't disable PROVE_RCU on first splat"
1107	depends on PROVE_RCU
1108	default n
1109	help
1110	 By itself, PROVE_RCU will disable checking upon issuing the
1111	 first warning (or "splat").  This feature prevents such
1112	 disabling, allowing multiple RCU-lockdep warnings to be printed
1113	 on a single reboot.
1114
1115	 Say Y to allow multiple RCU-lockdep warnings per boot.
1116
1117	 Say N if you are unsure.
1118
1119config PROVE_RCU_DELAY
1120	bool "RCU debugging: preemptible RCU race provocation"
1121	depends on DEBUG_KERNEL && PREEMPT_RCU
 
 
1122	default n
1123	help
1124	 There is a class of races that involve an unlikely preemption
1125	 of __rcu_read_unlock() just after ->rcu_read_lock_nesting has
1126	 been set to INT_MIN.  This feature inserts a delay at that
1127	 point to increase the probability of these races.
1128
1129	 Say Y to increase probability of preemption of __rcu_read_unlock().
1130
1131	 Say N if you are unsure.
 
 
 
 
 
 
 
 
 
 
 
 
1132
1133config SPARSE_RCU_POINTER
1134	bool "RCU debugging: sparse-based checks for pointer usage"
 
 
 
1135	default n
1136	help
1137	 This feature enables the __rcu sparse annotation for
1138	 RCU-protected pointers.  This annotation will cause sparse
1139	 to flag any non-RCU used of annotated pointers.  This can be
1140	 helpful when debugging RCU usage.  Please note that this feature
1141	 is not intended to enforce code cleanliness; it is instead merely
1142	 a debugging aid.
1143
1144	 Say Y to make sparse flag questionable use of RCU-protected pointers
1145
1146	 Say N if you are unsure.
 
 
 
 
 
 
 
 
1147
1148config TORTURE_TEST
1149	tristate
1150	default n
1151
1152config RCU_TORTURE_TEST
1153	tristate "torture tests for RCU"
1154	depends on DEBUG_KERNEL
1155	select TORTURE_TEST
1156	default n
1157	help
1158	  This option provides a kernel module that runs torture tests
1159	  on the RCU infrastructure.  The kernel module may be built
1160	  after the fact on the running kernel to be tested, if desired.
1161
1162	  Say Y here if you want RCU torture tests to be built into
1163	  the kernel.
1164	  Say M if you want the RCU torture tests to build as a module.
1165	  Say N if you are unsure.
1166
1167config RCU_TORTURE_TEST_RUNNABLE
1168	bool "torture tests for RCU runnable by default"
1169	depends on RCU_TORTURE_TEST = y
1170	default n
1171	help
1172	  This option provides a way to build the RCU torture tests
1173	  directly into the kernel without them starting up at boot
1174	  time.  You can use /proc/sys/kernel/rcutorture_runnable
1175	  to manually override this setting.  This /proc file is
1176	  available only when the RCU torture tests have been built
1177	  into the kernel.
1178
1179	  Say Y here if you want the RCU torture tests to start during
1180	  boot (you probably don't).
1181	  Say N here if you want the RCU torture tests to start only
1182	  after being manually enabled via /proc.
1183
1184config RCU_CPU_STALL_TIMEOUT
1185	int "RCU CPU stall timeout in seconds"
1186	depends on RCU_STALL_COMMON
1187	range 3 300
1188	default 21
1189	help
1190	  If a given RCU grace period extends more than the specified
1191	  number of seconds, a CPU stall warning is printed.  If the
1192	  RCU grace period persists, additional CPU stall warnings are
1193	  printed at more widely spaced intervals.
1194
1195config RCU_CPU_STALL_VERBOSE
1196	bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR"
1197	depends on TREE_PREEMPT_RCU
1198	default y
1199	help
1200	  This option causes RCU to printk detailed per-task information
1201	  for any tasks that are stalling the current RCU grace period.
1202
1203	  Say N if you are unsure.
 
1204
1205	  Say Y if you want to enable such checks.
1206
1207config RCU_CPU_STALL_INFO
1208	bool "Print additional diagnostics on RCU CPU stall"
1209	depends on (TREE_RCU || TREE_PREEMPT_RCU) && DEBUG_KERNEL
1210	default n
 
 
 
 
 
 
1211	help
1212	  For each stalled CPU that is aware of the current RCU grace
1213	  period, print out additional per-CPU diagnostic information
1214	  regarding scheduling-clock ticks, idle state, and,
1215	  for RCU_FAST_NO_HZ kernels, idle-entry state.
 
 
1216
1217	  Say N if you are unsure.
 
 
 
1218
1219	  Say Y if you want to enable such diagnostics.
1220
1221config RCU_TRACE
1222	bool "Enable tracing for RCU"
1223	depends on DEBUG_KERNEL
1224	select TRACE_CLOCK
1225	help
1226	  This option provides tracing in RCU which presents stats
1227	  in debugfs for debugging RCU implementation.
 
 
1228
1229	  Say Y here if you want to enable RCU tracing
1230	  Say N if you are unsure.
 
 
1231
1232endmenu # "RCU Debugging"
1233
1234config DEBUG_BLOCK_EXT_DEVT
1235        bool "Force extended block device numbers and spread them"
1236	depends on DEBUG_KERNEL
1237	depends on BLOCK
1238	default n
1239	help
1240	  BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1241	  SOME DISTRIBUTIONS.  DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1242	  YOU ARE DOING.  Distros, please enable this and fix whatever
1243	  is broken.
1244
1245	  Conventionally, block device numbers are allocated from
1246	  predetermined contiguous area.  However, extended block area
1247	  may introduce non-contiguous block device numbers.  This
1248	  option forces most block device numbers to be allocated from
1249	  the extended space and spreads them to discover kernel or
1250	  userland code paths which assume predetermined contiguous
1251	  device number allocation.
1252
1253	  Note that turning on this debug option shuffles all the
1254	  device numbers for all IDE and SCSI devices including libata
1255	  ones, so root partition specified using device number
1256	  directly (via rdev or root=MAJ:MIN) won't work anymore.
1257	  Textual device names (root=/dev/sdXn) will continue to work.
1258
1259	  Say N if you are unsure.
 
 
 
 
 
 
1260
1261config NOTIFIER_ERROR_INJECTION
1262	tristate "Notifier error injection"
1263	depends on DEBUG_KERNEL
1264	select DEBUG_FS
1265	help
1266	  This option provides the ability to inject artificial errors to
1267	  specified notifier chain callbacks. It is useful to test the error
1268	  handling of notifier call chain failures.
1269
1270	  Say N if unsure.
1271
1272config CPU_NOTIFIER_ERROR_INJECT
1273	tristate "CPU notifier error injection module"
1274	depends on HOTPLUG_CPU && NOTIFIER_ERROR_INJECTION
1275	help
1276	  This option provides a kernel module that can be used to test
1277	  the error handling of the cpu notifiers by injecting artificial
1278	  errors to CPU notifier chain callbacks.  It is controlled through
1279	  debugfs interface under /sys/kernel/debug/notifier-error-inject/cpu
1280
1281	  If the notifier call chain should be failed with some events
1282	  notified, write the error code to "actions/<notifier event>/error".
1283
1284	  Example: Inject CPU offline error (-1 == -EPERM)
1285
1286	  # cd /sys/kernel/debug/notifier-error-inject/cpu
1287	  # echo -1 > actions/CPU_DOWN_PREPARE/error
1288	  # echo 0 > /sys/devices/system/cpu/cpu1/online
1289	  bash: echo: write error: Operation not permitted
1290
1291	  To compile this code as a module, choose M here: the module will
1292	  be called cpu-notifier-error-inject.
1293
1294	  If unsure, say N.
1295
1296config PM_NOTIFIER_ERROR_INJECT
1297	tristate "PM notifier error injection module"
1298	depends on PM && NOTIFIER_ERROR_INJECTION
1299	default m if PM_DEBUG
1300	help
1301	  This option provides the ability to inject artificial errors to
1302	  PM notifier chain callbacks.  It is controlled through debugfs
1303	  interface /sys/kernel/debug/notifier-error-inject/pm
1304
1305	  If the notifier call chain should be failed with some events
1306	  notified, write the error code to "actions/<notifier event>/error".
1307
1308	  Example: Inject PM suspend error (-12 = -ENOMEM)
1309
1310	  # cd /sys/kernel/debug/notifier-error-inject/pm/
1311	  # echo -12 > actions/PM_SUSPEND_PREPARE/error
1312	  # echo mem > /sys/power/state
1313	  bash: echo: write error: Cannot allocate memory
1314
1315	  To compile this code as a module, choose M here: the module will
1316	  be called pm-notifier-error-inject.
1317
1318	  If unsure, say N.
1319
1320config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1321	tristate "OF reconfig notifier error injection module"
1322	depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1323	help
1324	  This option provides the ability to inject artificial errors to
1325	  OF reconfig notifier chain callbacks.  It is controlled
1326	  through debugfs interface under
1327	  /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1328
1329	  If the notifier call chain should be failed with some events
1330	  notified, write the error code to "actions/<notifier event>/error".
1331
1332	  To compile this code as a module, choose M here: the module will
1333	  be called of-reconfig-notifier-error-inject.
1334
1335	  If unsure, say N.
1336
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1337config FAULT_INJECTION
1338	bool "Fault-injection framework"
1339	depends on DEBUG_KERNEL
1340	help
1341	  Provide fault-injection framework.
1342	  For more details, see Documentation/fault-injection/.
1343
1344config FAILSLAB
1345	bool "Fault-injection capability for kmalloc"
1346	depends on FAULT_INJECTION
1347	depends on SLAB || SLUB
1348	help
1349	  Provide fault-injection capability for kmalloc.
1350
1351config FAIL_PAGE_ALLOC
1352	bool "Fault-injection capabilitiy for alloc_pages()"
1353	depends on FAULT_INJECTION
1354	help
1355	  Provide fault-injection capability for alloc_pages().
1356
 
 
 
 
 
 
 
1357config FAIL_MAKE_REQUEST
1358	bool "Fault-injection capability for disk IO"
1359	depends on FAULT_INJECTION && BLOCK
1360	help
1361	  Provide fault-injection capability for disk IO.
1362
1363config FAIL_IO_TIMEOUT
1364	bool "Fault-injection capability for faking disk interrupts"
1365	depends on FAULT_INJECTION && BLOCK
1366	help
1367	  Provide fault-injection capability on end IO handling. This
1368	  will make the block layer "forget" an interrupt as configured,
1369	  thus exercising the error handling.
1370
1371	  Only works with drivers that use the generic timeout handling,
1372	  for others it wont do anything.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1373
1374config FAIL_MMC_REQUEST
1375	bool "Fault-injection capability for MMC IO"
1376	select DEBUG_FS
1377	depends on FAULT_INJECTION && MMC
1378	help
1379	  Provide fault-injection capability for MMC IO.
1380	  This will make the mmc core return data errors. This is
1381	  useful to test the error handling in the mmc block device
1382	  and to test how the mmc host driver handles retries from
1383	  the block device.
1384
1385config FAULT_INJECTION_DEBUG_FS
1386	bool "Debugfs entries for fault-injection capabilities"
1387	depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1388	help
1389	  Enable configuration of fault-injection capabilities via debugfs.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1390
1391config FAULT_INJECTION_STACKTRACE_FILTER
1392	bool "stacktrace filter for fault-injection capabilities"
1393	depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1394	depends on !X86_64
1395	select STACKTRACE
1396	select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC
1397	help
1398	  Provide stacktrace filter for fault-injection capabilities
1399
1400config LATENCYTOP
1401	bool "Latency measuring infrastructure"
1402	depends on HAVE_LATENCYTOP_SUPPORT
1403	depends on DEBUG_KERNEL
1404	depends on STACKTRACE_SUPPORT
1405	depends on PROC_FS
1406	select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC
1407	select KALLSYMS
1408	select KALLSYMS_ALL
1409	select STACKTRACE
1410	select SCHEDSTATS
1411	select SCHED_DEBUG
1412	help
1413	  Enable this option if you want to use the LatencyTOP tool
1414	  to find out which userspace is blocking on what kernel operations.
 
 
 
 
1415
1416config ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1417	bool
1418
1419config DEBUG_STRICT_USER_COPY_CHECKS
1420	bool "Strict user copy size checks"
1421	depends on ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1422	depends on DEBUG_KERNEL && !TRACE_BRANCH_PROFILING
 
 
 
 
 
1423	help
1424	  Enabling this option turns a certain set of sanity checks for user
1425	  copy operations into compile time failures.
1426
1427	  The copy_from_user() etc checks are there to help test if there
1428	  are sufficient security checks on the length argument of
1429	  the copy operation, by having gcc prove that the argument is
1430	  within bounds.
1431
1432	  If unsure, say N.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1433
1434source kernel/trace/Kconfig
1435
1436menu "Runtime Testing"
 
 
 
 
 
 
 
 
 
 
 
 
1437
1438config LKDTM
1439	tristate "Linux Kernel Dump Test Tool Module"
1440	depends on DEBUG_FS
1441	depends on BLOCK
1442	default n
1443	help
1444	This module enables testing of the different dumping mechanisms by
1445	inducing system failures at predefined crash points.
1446	If you don't need it: say N
1447	Choose M here to compile this code as a module. The module will be
1448	called lkdtm.
1449
1450	Documentation on how to use the module can be found in
1451	Documentation/fault-injection/provoke-crashes.txt
 
 
 
 
 
 
 
 
 
 
 
 
1452
1453config TEST_LIST_SORT
1454	bool "Linked list sorting test"
1455	depends on DEBUG_KERNEL
 
1456	help
1457	  Enable this to turn on 'list_sort()' function test. This test is
1458	  executed only once during system boot, so affects only boot time.
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1459
1460	  If unsure, say N.
1461
1462config KPROBES_SANITY_TEST
1463	bool "Kprobes sanity tests"
1464	depends on DEBUG_KERNEL
1465	depends on KPROBES
1466	default n
 
 
1467	help
1468	  This option provides for testing basic kprobes functionality on
1469	  boot. A sample kprobe, jprobe and kretprobe are inserted and
1470	  verified for functionality.
1471
1472	  Say N if you are unsure.
1473
 
 
 
 
 
 
 
 
 
 
 
 
1474config BACKTRACE_SELF_TEST
1475	tristate "Self test for the backtrace code"
1476	depends on DEBUG_KERNEL
1477	default n
1478	help
1479	  This option provides a kernel module that can be used to test
1480	  the kernel stack backtrace code. This option is not useful
1481	  for distributions or general kernels, but only for kernel
1482	  developers working on architecture code.
1483
1484	  Note that if you want to also test saved backtraces, you will
1485	  have to enable STACKTRACE as well.
1486
1487	  Say N if you are unsure.
1488
 
 
 
 
 
 
 
 
 
 
1489config RBTREE_TEST
1490	tristate "Red-Black tree test"
1491	depends on DEBUG_KERNEL
1492	help
1493	  A benchmark measuring the performance of the rbtree library.
1494	  Also includes rbtree invariant checks.
1495
 
 
 
 
 
 
 
 
 
 
 
 
1496config INTERVAL_TREE_TEST
1497	tristate "Interval tree test"
1498	depends on m && DEBUG_KERNEL
 
1499	help
1500	  A benchmark measuring the performance of the interval tree library
1501
1502config PERCPU_TEST
1503	tristate "Per cpu operations test"
1504	depends on m && DEBUG_KERNEL
1505	help
1506	  Enable this option to build test module which validates per-cpu
1507	  operations.
1508
1509	  If unsure, say N.
1510
1511config ATOMIC64_SELFTEST
1512	bool "Perform an atomic64_t self-test at boot"
1513	help
1514	  Enable this option to test the atomic64_t functions at boot.
 
1515
1516	  If unsure, say N.
1517
1518config ASYNC_RAID6_TEST
1519	tristate "Self test for hardware accelerated raid6 recovery"
1520	depends on ASYNC_RAID6_RECOV
1521	select ASYNC_MEMCPY
1522	---help---
1523	  This is a one-shot self test that permutes through the
1524	  recovery of all the possible two disk failure scenarios for a
1525	  N-disk array.  Recovery is performed with the asynchronous
1526	  raid6 recovery routines, and will optionally use an offload
1527	  engine if one is available.
1528
1529	  If unsure, say N.
1530
1531config TEST_STRING_HELPERS
1532	tristate "Test functions located in the string_helpers module at runtime"
 
 
 
 
 
 
 
 
 
 
1533
1534config TEST_KSTRTOX
1535	tristate "Test kstrto*() family of functions at runtime"
1536
1537endmenu # runtime tests
 
1538
1539config PROVIDE_OHCI1394_DMA_INIT
1540	bool "Remote debugging over FireWire early on boot"
1541	depends on PCI && X86
1542	help
1543	  If you want to debug problems which hang or crash the kernel early
1544	  on boot and the crashing machine has a FireWire port, you can use
1545	  this feature to remotely access the memory of the crashed machine
1546	  over FireWire. This employs remote DMA as part of the OHCI1394
1547	  specification which is now the standard for FireWire controllers.
1548
1549	  With remote DMA, you can monitor the printk buffer remotely using
1550	  firescope and access all memory below 4GB using fireproxy from gdb.
1551	  Even controlling a kernel debugger is possible using remote DMA.
 
1552
1553	  Usage:
1554
1555	  If ohci1394_dma=early is used as boot parameter, it will initialize
1556	  all OHCI1394 controllers which are found in the PCI config space.
1557
1558	  As all changes to the FireWire bus such as enabling and disabling
1559	  devices cause a bus reset and thereby disable remote DMA for all
1560	  devices, be sure to have the cable plugged and FireWire enabled on
1561	  the debugging host before booting the debug target for debugging.
1562
1563	  This code (~1k) is freed after boot. By then, the firewire stack
1564	  in charge of the OHCI-1394 controllers should be used instead.
 
 
 
 
1565
1566	  See Documentation/debugging-via-ohci1394.txt for more information.
1567
1568config BUILD_DOCSRC
1569	bool "Build targets in Documentation/ tree"
1570	depends on HEADERS_CHECK
1571	help
1572	  This option attempts to build objects from the source files in the
1573	  kernel Documentation/ tree.
1574
1575	  Say N if you are unsure.
 
 
 
1576
1577config DMA_API_DEBUG
1578	bool "Enable debugging of DMA-API usage"
1579	depends on HAVE_DMA_API_DEBUG
1580	help
1581	  Enable this option to debug the use of the DMA API by device drivers.
1582	  With this option you will be able to detect common bugs in device
1583	  drivers like double-freeing of DMA mappings or freeing mappings that
1584	  were never allocated.
1585
1586	  This also attempts to catch cases where a page owned by DMA is
1587	  accessed by the cpu in a way that could cause data corruption.  For
1588	  example, this enables cow_user_page() to check that the source page is
1589	  not undergoing DMA.
1590
1591	  This option causes a performance degradation.  Use only if you want to
1592	  debug device drivers and dma interactions.
 
 
 
1593
1594	  If unsure, say N.
1595
1596config TEST_MODULE
1597	tristate "Test module loading with 'hello world' module"
1598	default n
1599	depends on m
1600	help
1601	  This builds the "test_module" module that emits "Hello, world"
1602	  on printk when loaded. It is designed to be used for basic
1603	  evaluation of the module loading subsystem (for example when
1604	  validating module verification). It lacks any extra dependencies,
1605	  and will not normally be loaded by the system unless explicitly
1606	  requested by name.
1607
1608	  If unsure, say N.
1609
1610config TEST_USER_COPY
1611	tristate "Test user/kernel boundary protections"
 
 
 
 
 
 
 
 
 
 
 
 
1612	default n
 
1613	depends on m
1614	help
1615	  This builds the "test_user_copy" module that runs sanity checks
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1616	  on the copy_to/from_user infrastructure, making sure basic
1617	  user/kernel boundary testing is working. If it fails to load,
1618	  a regression has been detected in the user/kernel memory boundary
1619	  protections.
 
 
 
 
 
 
 
 
 
 
 
 
 
1620
1621	  If unsure, say N.
1622
1623source "samples/Kconfig"
 
 
 
 
1624
1625source "lib/Kconfig.kgdb"
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1626