Root/lib/Kconfig.debug

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

Archive Download this file



interactive