3 bool "Show timing information on printks"
6 Selecting this option causes timing information to be
7 included in printk output. This allows you to measure
8 the interval between kernel operations, including bootup
9 operations. This is useful for identifying long delays
12 config DEFAULT_MESSAGE_LOGLEVEL
13 int "Default message log level (1-7)"
17 Default log level for printk statements with no specified priority.
19 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
20 that are auditing their logs closely may want to set it to a lower
23 config ENABLE_WARN_DEPRECATED
24 bool "Enable __deprecated logic"
27 Enable the __deprecated logic in the kernel build.
28 Disable this to suppress the "warning: 'foo' is deprecated
29 (declared at kernel/power/somefile.c:1234)" messages.
31 config ENABLE_MUST_CHECK
32 bool "Enable __must_check logic"
35 Enable the __must_check logic in the kernel build. Disable this to
36 suppress the "warning: ignoring return value of 'foo', declared with
37 attribute warn_unused_result" messages.
40 int "Warn for stack frames larger than (needs gcc 4.4)"
42 default 1024 if !64BIT
45 Tell gcc to warn at build time for stack frames larger than this.
46 Setting this too low will cause a lot of warnings.
47 Setting it to 0 disables the warning.
51 bool "Magic SysRq key"
54 If you say Y here, you will have some control over the system even
55 if the system crashes for example during kernel debugging (e.g., you
56 will be able to flush the buffer cache to disk, reboot the system
57 immediately or dump some status information). This is accomplished
58 by pressing various keys while holding SysRq (Alt+PrintScreen). It
59 also works on a serial console (on PC hardware at least), if you
60 send a BREAK and then within 5 seconds a command keypress. The
61 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
62 unless you really know what this hack does.
65 bool "Strip assembler-generated symbols during link"
68 Strip internal assembler-generated symbols during a link (symbols
69 that look like '.Lxxx') so they don't pollute the output of
70 get_wchan() and suchlike.
73 bool "Enable unused/obsolete exported symbols"
76 Unused but exported symbols make the kernel needlessly bigger. For
77 that reason most of these unused exports will soon be removed. This
78 option is provided temporarily to provide a transition period in case
79 some external kernel module needs one of these symbols anyway. If you
80 encounter such a case in your module, consider if you are actually
81 using the right API. (rationale: since nobody in the kernel is using
82 this in a module, there is a pretty good chance it's actually the
83 wrong interface to use). If you really need the symbol, please send a
84 mail to the linux kernel mailing list mentioning the symbol and why
85 you really need it, and what the merge plan to the mainline kernel for
89 bool "Debug Filesystem"
91 debugfs is a virtual file system that kernel developers use to put
92 debugging files into. Enable this option to be able to read and
95 For detailed documentation on the debugfs API, see
96 Documentation/DocBook/filesystems.
101 bool "Run 'make headers_check' when building vmlinux"
104 This option will extract the user-visible kernel headers whenever
105 building the kernel, and will run basic sanity checks on them to
106 ensure that exported files do not attempt to include files which
107 were not exported, etc.
109 If you're making modifications to header files which are
110 relevant for userspace, say 'Y', and check the headers
111 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
112 your build tree), to make sure they're suitable.
114 config DEBUG_SECTION_MISMATCH
115 bool "Enable full Section mismatch analysis"
117 The section mismatch analysis checks if there are illegal
118 references from one section to another section.
119 Linux will during link or during runtime drop some sections
120 and any use of code/data previously in these sections will
121 most likely result in an oops.
122 In the code functions and variables are annotated with
123 __init, __devinit etc. (see full list in include/linux/init.h)
124 which results in the code/data being placed in specific sections.
125 The section mismatch analysis is always done after a full
126 kernel build but enabling this option will in addition
128 - Add the option -fno-inline-functions-called-once to gcc
129 When inlining a function annotated __init in a non-init
130 function we would lose the section information and thus
131 the analysis would not catch the illegal reference.
132 This option tells gcc to inline less but will also
133 result in a larger kernel.
134 - Run the section mismatch analysis for each module/built-in.o
135 When we run the section mismatch analysis on vmlinux.o we
136 lose valueble information about where the mismatch was
138 Running the analysis for each module/built-in.o file
139 will tell where the mismatch happens much closer to the
140 source. The drawback is that we will report the same
141 mismatch at least twice.
142 - Enable verbose reporting from modpost to help solving
143 the section mismatches reported.
146 bool "Kernel debugging"
148 Say Y here if you are developing drivers or trying to debug and
149 identify kernel problems.
152 bool "Debug shared IRQ handlers"
153 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
155 Enable this to generate a spurious interrupt as soon as a shared
156 interrupt handler is registered, and just before one is deregistered.
157 Drivers ought to be able to handle interrupts coming in at those
158 points; some don't and need to be caught.
160 config LOCKUP_DETECTOR
161 bool "Detect Hard and Soft Lockups"
162 depends on DEBUG_KERNEL && !S390
164 Say Y here to enable the kernel to act as a watchdog to detect
165 hard and soft lockups.
167 Softlockups are bugs that cause the kernel to loop in kernel
168 mode for more than 60 seconds, without giving other tasks a
169 chance to run. The current stack trace is displayed upon
170 detection and the system will stay locked up.
172 Hardlockups are bugs that cause the CPU to loop in kernel mode
173 for more than 60 seconds, without letting other interrupts have a
174 chance to run. The current stack trace is displayed upon detection
175 and the system will stay locked up.
177 The overhead should be minimal. A periodic hrtimer runs to
178 generate interrupts and kick the watchdog task every 10-12 seconds.
179 An NMI is generated every 60 seconds or so to check for hardlockups.
181 config HARDLOCKUP_DETECTOR
182 def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI && \
183 !ARCH_HAS_NMI_WATCHDOG
185 config BOOTPARAM_HARDLOCKUP_PANIC
186 bool "Panic (Reboot) On Hard Lockups"
187 depends on LOCKUP_DETECTOR
189 Say Y here to enable the kernel to panic on "hard lockups",
190 which are bugs that cause the kernel to loop in kernel
191 mode with interrupts disabled for more than 60 seconds.
195 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
197 depends on LOCKUP_DETECTOR
199 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
200 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
202 config BOOTPARAM_SOFTLOCKUP_PANIC
203 bool "Panic (Reboot) On Soft Lockups"
204 depends on LOCKUP_DETECTOR
206 Say Y here to enable the kernel to panic on "soft lockups",
207 which are bugs that cause the kernel to loop in kernel
208 mode for more than 60 seconds, without giving other tasks a
211 The panic can be used in combination with panic_timeout,
212 to cause the system to reboot automatically after a
213 lockup has been detected. This feature is useful for
214 high-availability systems that have uptime guarantees and
215 where a lockup must be resolved ASAP.
219 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
221 depends on LOCKUP_DETECTOR
223 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
224 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
226 config DETECT_HUNG_TASK
227 bool "Detect Hung Tasks"
228 depends on DEBUG_KERNEL
229 default DETECT_SOFTLOCKUP
231 Say Y here to enable the kernel to detect "hung tasks",
232 which are bugs that cause the task to be stuck in
233 uninterruptible "D" state indefinitiley.
235 When a hung task is detected, the kernel will print the
236 current stack trace (which you should report), but the
237 task will stay in uninterruptible state. If lockdep is
238 enabled then all held locks will also be reported. This
239 feature has negligible overhead.
241 config BOOTPARAM_HUNG_TASK_PANIC
242 bool "Panic (Reboot) On Hung Tasks"
243 depends on DETECT_HUNG_TASK
245 Say Y here to enable the kernel to panic on "hung tasks",
246 which are bugs that cause the kernel to leave a task stuck
247 in uninterruptible "D" state.
249 The panic can be used in combination with panic_timeout,
250 to cause the system to reboot automatically after a
251 hung task has been detected. This feature is useful for
252 high-availability systems that have uptime guarantees and
253 where a hung tasks must be resolved ASAP.
257 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
259 depends on DETECT_HUNG_TASK
261 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
262 default 1 if BOOTPARAM_HUNG_TASK_PANIC
265 bool "Collect scheduler debugging info"
266 depends on DEBUG_KERNEL && PROC_FS
269 If you say Y here, the /proc/sched_debug file will be provided
270 that can help debug the scheduler. The runtime overhead of this
274 bool "Collect scheduler statistics"
275 depends on DEBUG_KERNEL && PROC_FS
277 If you say Y here, additional code will be inserted into the
278 scheduler and related routines to collect statistics about
279 scheduler behavior and provide them in /proc/schedstat. These
280 stats may be useful for both tuning and debugging the scheduler
281 If you aren't debugging the scheduler or trying to tune a specific
282 application, you can say N to avoid the very slight overhead
286 bool "Collect kernel timers statistics"
287 depends on DEBUG_KERNEL && PROC_FS
289 If you say Y here, additional code will be inserted into the
290 timer routines to collect statistics about kernel timers being
291 reprogrammed. The statistics can be read from /proc/timer_stats.
292 The statistics collection is started by writing 1 to /proc/timer_stats,
293 writing 0 stops it. This feature is useful to collect information
294 about timer usage patterns in kernel and userspace. This feature
295 is lightweight if enabled in the kernel config but not activated
296 (it defaults to deactivated on bootup and will only be activated
297 if some application like powertop activates it explicitly).
300 bool "Debug object operations"
301 depends on DEBUG_KERNEL
303 If you say Y here, additional code will be inserted into the
304 kernel to track the life time of various objects and validate
305 the operations on those objects.
307 config DEBUG_OBJECTS_SELFTEST
308 bool "Debug objects selftest"
309 depends on DEBUG_OBJECTS
311 This enables the selftest of the object debug code.
313 config DEBUG_OBJECTS_FREE
314 bool "Debug objects in freed memory"
315 depends on DEBUG_OBJECTS
317 This enables checks whether a k/v free operation frees an area
318 which contains an object which has not been deactivated
319 properly. This can make kmalloc/kfree-intensive workloads
322 config DEBUG_OBJECTS_TIMERS
323 bool "Debug timer objects"
324 depends on DEBUG_OBJECTS
326 If you say Y here, additional code will be inserted into the
327 timer routines to track the life time of timer objects and
328 validate the timer operations.
330 config DEBUG_OBJECTS_WORK
331 bool "Debug work objects"
332 depends on DEBUG_OBJECTS
334 If you say Y here, additional code will be inserted into the
335 work queue routines to track the life time of work objects and
336 validate the work operations.
338 config DEBUG_OBJECTS_RCU_HEAD
339 bool "Debug RCU callbacks objects"
340 depends on DEBUG_OBJECTS && PREEMPT
342 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
344 config DEBUG_OBJECTS_PERCPU_COUNTER
345 bool "Debug percpu counter objects"
346 depends on DEBUG_OBJECTS
348 If you say Y here, additional code will be inserted into the
349 percpu counter routines to track the life time of percpu counter
350 objects and validate the percpu counter operations.
352 config DEBUG_OBJECTS_ENABLE_DEFAULT
353 int "debug_objects bootup default value (0-1)"
356 depends on DEBUG_OBJECTS
358 Debug objects boot parameter default value
361 bool "Debug slab memory allocations"
362 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
364 Say Y here to have the kernel do limited verification on memory
365 allocation as well as poisoning memory on free to catch use of freed
366 memory. This can make kmalloc/kfree-intensive workloads much slower.
368 config DEBUG_SLAB_LEAK
369 bool "Memory leak debugging"
370 depends on DEBUG_SLAB
373 bool "SLUB debugging on by default"
374 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
377 Boot with debugging on by default. SLUB boots by default with
378 the runtime debug capabilities switched off. Enabling this is
379 equivalent to specifying the "slub_debug" parameter on boot.
380 There is no support for more fine grained debug control like
381 possible with slub_debug=xxx. SLUB debugging may be switched
382 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
387 bool "Enable SLUB performance statistics"
388 depends on SLUB && SYSFS
390 SLUB statistics are useful to debug SLUBs allocation behavior in
391 order find ways to optimize the allocator. This should never be
392 enabled for production use since keeping statistics slows down
393 the allocator by a few percentage points. The slabinfo command
394 supports the determination of the most active slabs to figure
395 out which slabs are relevant to a particular load.
396 Try running: slabinfo -DA
398 config DEBUG_KMEMLEAK
399 bool "Kernel memory leak detector"
400 depends on DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && \
401 (X86 || ARM || PPC || S390 || SPARC64 || SUPERH || MICROBLAZE || TILE)
403 select DEBUG_FS if SYSFS
404 select STACKTRACE if STACKTRACE_SUPPORT
408 Say Y here if you want to enable the memory leak
409 detector. The memory allocation/freeing is traced in a way
410 similar to the Boehm's conservative garbage collector, the
411 difference being that the orphan objects are not freed but
412 only shown in /sys/kernel/debug/kmemleak. Enabling this
413 feature will introduce an overhead to memory
414 allocations. See Documentation/kmemleak.txt for more
417 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
418 of finding leaks due to the slab objects poisoning.
420 In order to access the kmemleak file, debugfs needs to be
421 mounted (usually at /sys/kernel/debug).
423 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
424 int "Maximum kmemleak early log entries"
425 depends on DEBUG_KMEMLEAK
429 Kmemleak must track all the memory allocations to avoid
430 reporting false positives. Since memory may be allocated or
431 freed before kmemleak is initialised, an early log buffer is
432 used to store these actions. If kmemleak reports "early log
433 buffer exceeded", please increase this value.
435 config DEBUG_KMEMLEAK_TEST
436 tristate "Simple test for the kernel memory leak detector"
437 depends on DEBUG_KMEMLEAK
439 Say Y or M here to build a test for the kernel memory leak
440 detector. This option enables a module that explicitly leaks
445 config DEBUG_KMEMLEAK_DEFAULT_OFF
446 bool "Default kmemleak to off"
447 depends on DEBUG_KMEMLEAK
449 Say Y here to disable kmemleak by default. It can then be enabled
450 on the command line via kmemleak=on.
453 bool "Debug preemptible kernel"
454 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
457 If you say Y here then the kernel will use a debug variant of the
458 commonly used smp_processor_id() function and will print warnings
459 if kernel code uses it in a preemption-unsafe way. Also, the kernel
460 will detect preemption count underflows.
462 config DEBUG_RT_MUTEXES
463 bool "RT Mutex debugging, deadlock detection"
464 depends on DEBUG_KERNEL && RT_MUTEXES
466 This allows rt mutex semantics violations and rt mutex related
467 deadlocks (lockups) to be detected and reported automatically.
472 depends on DEBUG_RT_MUTEXES
474 config RT_MUTEX_TESTER
475 bool "Built-in scriptable tester for rt-mutexes"
476 depends on DEBUG_KERNEL && RT_MUTEXES
478 This option enables a rt-mutex tester.
480 config DEBUG_SPINLOCK
481 bool "Spinlock and rw-lock debugging: basic checks"
482 depends on DEBUG_KERNEL
484 Say Y here and build SMP to catch missing spinlock initialization
485 and certain other kinds of spinlock errors commonly made. This is
486 best used in conjunction with the NMI watchdog so that spinlock
487 deadlocks are also debuggable.
490 bool "Mutex debugging: basic checks"
491 depends on DEBUG_KERNEL
493 This feature allows mutex semantics violations to be detected and
496 config DEBUG_LOCK_ALLOC
497 bool "Lock debugging: detect incorrect freeing of live locks"
498 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
499 select DEBUG_SPINLOCK
503 This feature will check whether any held lock (spinlock, rwlock,
504 mutex or rwsem) is incorrectly freed by the kernel, via any of the
505 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
506 vfree(), etc.), whether a live lock is incorrectly reinitialized via
507 spin_lock_init()/mutex_init()/etc., or whether there is any lock
508 held during task exit.
511 bool "Lock debugging: prove locking correctness"
512 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
514 select DEBUG_SPINLOCK
516 select DEBUG_LOCK_ALLOC
517 select TRACE_IRQFLAGS
520 This feature enables the kernel to prove that all locking
521 that occurs in the kernel runtime is mathematically
522 correct: that under no circumstance could an arbitrary (and
523 not yet triggered) combination of observed locking
524 sequences (on an arbitrary number of CPUs, running an
525 arbitrary number of tasks and interrupt contexts) cause a
528 In short, this feature enables the kernel to report locking
529 related deadlocks before they actually occur.
531 The proof does not depend on how hard and complex a
532 deadlock scenario would be to trigger: how many
533 participant CPUs, tasks and irq-contexts would be needed
534 for it to trigger. The proof also does not depend on
535 timing: if a race and a resulting deadlock is possible
536 theoretically (no matter how unlikely the race scenario
537 is), it will be proven so and will immediately be
538 reported by the kernel (once the event is observed that
539 makes the deadlock theoretically possible).
541 If a deadlock is impossible (i.e. the locking rules, as
542 observed by the kernel, are mathematically correct), the
543 kernel reports nothing.
545 NOTE: this feature can also be enabled for rwlocks, mutexes
546 and rwsems - in which case all dependencies between these
547 different locking variants are observed and mapped too, and
548 the proof of observed correctness is also maintained for an
549 arbitrary combination of these separate locking variants.
551 For more details, see Documentation/lockdep-design.txt.
554 bool "RCU debugging: prove RCU correctness"
555 depends on PROVE_LOCKING
558 This feature enables lockdep extensions that check for correct
559 use of RCU APIs. This is currently under development. Say Y
560 if you want to debug RCU usage or help work on the PROVE_RCU
563 Say N if you are unsure.
565 config PROVE_RCU_REPEATEDLY
566 bool "RCU debugging: don't disable PROVE_RCU on first splat"
570 By itself, PROVE_RCU will disable checking upon issuing the
571 first warning (or "splat"). This feature prevents such
572 disabling, allowing multiple RCU-lockdep warnings to be printed
575 Say Y to allow multiple RCU-lockdep warnings per boot.
577 Say N if you are unsure.
579 config SPARSE_RCU_POINTER
580 bool "RCU debugging: sparse-based checks for pointer usage"
583 This feature enables the __rcu sparse annotation for
584 RCU-protected pointers. This annotation will cause sparse
585 to flag any non-RCU used of annotated pointers. This can be
586 helpful when debugging RCU usage. Please note that this feature
587 is not intended to enforce code cleanliness; it is instead merely
590 Say Y to make sparse flag questionable use of RCU-protected pointers
592 Say N if you are unsure.
596 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
598 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE
603 bool "Lock usage statistics"
604 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
606 select DEBUG_SPINLOCK
608 select DEBUG_LOCK_ALLOC
611 This feature enables tracking lock contention points
613 For more details, see Documentation/lockstat.txt
615 This also enables lock events required by "perf lock",
617 If you want to use "perf lock", you also need to turn on
618 CONFIG_EVENT_TRACING.
620 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
621 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
624 bool "Lock dependency engine debugging"
625 depends on DEBUG_KERNEL && LOCKDEP
627 If you say Y here, the lock dependency engine will do
628 additional runtime checks to debug itself, at the price
629 of more runtime overhead.
631 config TRACE_IRQFLAGS
634 Enables hooks to interrupt enabling and disabling for
635 either tracing or lock debugging.
637 config DEBUG_SPINLOCK_SLEEP
638 bool "Spinlock debugging: sleep-inside-spinlock checking"
639 depends on DEBUG_KERNEL
641 If you say Y here, various routines which may sleep will become very
642 noisy if they are called with a spinlock held.
644 config DEBUG_LOCKING_API_SELFTESTS
645 bool "Locking API boot-time self-tests"
646 depends on DEBUG_KERNEL
648 Say Y here if you want the kernel to run a short self-test during
649 bootup. The self-test checks whether common types of locking bugs
650 are detected by debugging mechanisms or not. (if you disable
651 lock debugging then those bugs wont be detected of course.)
652 The following locking APIs are covered: spinlocks, rwlocks,
657 depends on STACKTRACE_SUPPORT
660 bool "kobject debugging"
661 depends on DEBUG_KERNEL
663 If you say Y here, some extra kobject debugging messages will be sent
667 bool "Highmem debugging"
668 depends on DEBUG_KERNEL && HIGHMEM
670 This options enables addition error checking for high memory systems.
671 Disable for production systems.
673 config DEBUG_BUGVERBOSE
674 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
676 depends on ARM || AVR32 || M32R || M68K || SPARC32 || SPARC64 || \
677 FRV || SUPERH || GENERIC_BUG || BLACKFIN || MN10300
680 Say Y here to make BUG() panics output the file name and line number
681 of the BUG call as well as the EIP and oops trace. This aids
682 debugging but costs about 70-100K of memory.
685 bool "Compile the kernel with debug info"
686 depends on DEBUG_KERNEL
688 If you say Y here the resulting kernel image will include
689 debugging info resulting in a larger kernel image.
690 This adds debug symbols to the kernel and modules (gcc -g), and
691 is needed if you intend to use kernel crashdump or binary object
692 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
693 Say Y here only if you plan to debug the kernel.
697 config DEBUG_INFO_REDUCED
698 bool "Reduce debugging information"
699 depends on DEBUG_INFO
701 If you say Y here gcc is instructed to generate less debugging
702 information for structure types. This means that tools that
703 need full debugging information (like kgdb or systemtap) won't
704 be happy. But if you merely need debugging information to
705 resolve line numbers there is no loss. Advantage is that
706 build directory object sizes shrink dramatically over a full
707 DEBUG_INFO build and compile times are reduced too.
708 Only works with newer gcc versions.
712 depends on DEBUG_KERNEL
714 Enable this to turn on extended checks in the virtual-memory system
715 that may impact performance.
720 bool "Debug VM translations"
721 depends on DEBUG_KERNEL && X86
723 Enable some costly sanity checks in virtual to page code. This can
724 catch mistakes with virt_to_page() and friends.
728 config DEBUG_NOMMU_REGIONS
729 bool "Debug the global anon/private NOMMU mapping region tree"
730 depends on DEBUG_KERNEL && !MMU
732 This option causes the global tree of anonymous and private mapping
733 regions to be regularly checked for invalid topology.
735 config DEBUG_WRITECOUNT
736 bool "Debug filesystem writers count"
737 depends on DEBUG_KERNEL
739 Enable this to catch wrong use of the writers count in struct
740 vfsmount. This will increase the size of each file struct by
745 config DEBUG_MEMORY_INIT
746 bool "Debug memory initialisation" if EXPERT
749 Enable this for additional checks during memory initialisation.
750 The sanity checks verify aspects of the VM such as the memory model
751 and other information provided by the architecture. Verbose
752 information will be printed at KERN_DEBUG loglevel depending
753 on the mminit_loglevel= command-line option.
758 bool "Debug linked list manipulation"
759 depends on DEBUG_KERNEL
761 Enable this to turn on extended checks in the linked-list
766 config TEST_LIST_SORT
767 bool "Linked list sorting test"
768 depends on DEBUG_KERNEL
770 Enable this to turn on 'list_sort()' function test. This test is
771 executed only once during system boot, so affects only boot time.
776 bool "Debug SG table operations"
777 depends on DEBUG_KERNEL
779 Enable this to turn on checks on scatter-gather tables. This can
780 help find problems with drivers that do not properly initialize
785 config DEBUG_NOTIFIERS
786 bool "Debug notifier call chains"
787 depends on DEBUG_KERNEL
789 Enable this to turn on sanity checking for notifier call chains.
790 This is most useful for kernel developers to make sure that
791 modules properly unregister themselves from notifier chains.
792 This is a relatively cheap check but if you care about maximum
795 config DEBUG_CREDENTIALS
796 bool "Debug credential management"
797 depends on DEBUG_KERNEL
799 Enable this to turn on some debug checking for credential
800 management. The additional code keeps track of the number of
801 pointers from task_structs to any given cred struct, and checks to
802 see that this number never exceeds the usage count of the cred
805 Furthermore, if SELinux is enabled, this also checks that the
806 security pointer in the cred struct is never seen to be invalid.
811 # Select this config option from the architecture Kconfig, if it
812 # it is preferred to always offer frame pointers as a config
813 # option on the architecture (regardless of KERNEL_DEBUG):
815 config ARCH_WANT_FRAME_POINTERS
820 bool "Compile the kernel with frame pointers"
821 depends on DEBUG_KERNEL && \
822 (CRIS || M68K || FRV || UML || \
823 AVR32 || SUPERH || BLACKFIN || MN10300) || \
824 ARCH_WANT_FRAME_POINTERS
825 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
827 If you say Y here the resulting kernel image will be slightly
828 larger and slower, but it gives very useful debugging information
829 in case of kernel bugs. (precise oopses/stacktraces/warnings)
831 config BOOT_PRINTK_DELAY
832 bool "Delay each boot printk message by N milliseconds"
833 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
835 This build option allows you to read kernel boot messages
836 by inserting a short delay after each one. The delay is
837 specified in milliseconds on the kernel command line,
838 using "boot_delay=N".
840 It is likely that you would also need to use "lpj=M" to preset
841 the "loops per jiffie" value.
842 See a previous boot log for the "lpj" value to use for your
843 system, and then set "lpj=M" before setting "boot_delay=N".
844 NOTE: Using this option may adversely affect SMP systems.
845 I.e., processors other than the first one may not boot up.
846 BOOT_PRINTK_DELAY also may cause DETECT_SOFTLOCKUP to detect
847 what it believes to be lockup conditions.
849 config RCU_TORTURE_TEST
850 tristate "torture tests for RCU"
851 depends on DEBUG_KERNEL
854 This option provides a kernel module that runs torture tests
855 on the RCU infrastructure. The kernel module may be built
856 after the fact on the running kernel to be tested, if desired.
858 Say Y here if you want RCU torture tests to be built into
860 Say M if you want the RCU torture tests to build as a module.
861 Say N if you are unsure.
863 config RCU_TORTURE_TEST_RUNNABLE
864 bool "torture tests for RCU runnable by default"
865 depends on RCU_TORTURE_TEST = y
868 This option provides a way to build the RCU torture tests
869 directly into the kernel without them starting up at boot
870 time. You can use /proc/sys/kernel/rcutorture_runnable
871 to manually override this setting. This /proc file is
872 available only when the RCU torture tests have been built
875 Say Y here if you want the RCU torture tests to start during
876 boot (you probably don't).
877 Say N here if you want the RCU torture tests to start only
878 after being manually enabled via /proc.
880 config RCU_CPU_STALL_DETECTOR
881 bool "Check for stalled CPUs delaying RCU grace periods"
882 depends on TREE_RCU || TREE_PREEMPT_RCU
885 This option causes RCU to printk information on which
886 CPUs are delaying the current grace period, but only when
887 the grace period extends for excessive time periods.
889 Say N if you want to disable such checks.
891 Say Y if you are unsure.
893 config RCU_CPU_STALL_TIMEOUT
894 int "RCU CPU stall timeout in seconds"
895 depends on RCU_CPU_STALL_DETECTOR
899 If a given RCU grace period extends more than the specified
900 number of seconds, a CPU stall warning is printed. If the
901 RCU grace period persists, additional CPU stall warnings are
902 printed at more widely spaced intervals.
904 config RCU_CPU_STALL_DETECTOR_RUNNABLE
905 bool "RCU CPU stall checking starts automatically at boot"
906 depends on RCU_CPU_STALL_DETECTOR
909 If set, start checking for RCU CPU stalls immediately on
910 boot. Otherwise, RCU CPU stall checking must be manually
913 Say Y if you are unsure.
915 Say N if you wish to suppress RCU CPU stall checking during boot.
917 config RCU_CPU_STALL_VERBOSE
918 bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR"
919 depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU
922 This option causes RCU to printk detailed per-task information
923 for any tasks that are stalling the current RCU grace period.
925 Say N if you are unsure.
927 Say Y if you want to enable such checks.
929 config KPROBES_SANITY_TEST
930 bool "Kprobes sanity tests"
931 depends on DEBUG_KERNEL
935 This option provides for testing basic kprobes functionality on
936 boot. A sample kprobe, jprobe and kretprobe are inserted and
937 verified for functionality.
939 Say N if you are unsure.
941 config BACKTRACE_SELF_TEST
942 tristate "Self test for the backtrace code"
943 depends on DEBUG_KERNEL
946 This option provides a kernel module that can be used to test
947 the kernel stack backtrace code. This option is not useful
948 for distributions or general kernels, but only for kernel
949 developers working on architecture code.
951 Note that if you want to also test saved backtraces, you will
952 have to enable STACKTRACE as well.
954 Say N if you are unsure.
956 config DEBUG_BLOCK_EXT_DEVT
957 bool "Force extended block device numbers and spread them"
958 depends on DEBUG_KERNEL
962 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
963 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
964 YOU ARE DOING. Distros, please enable this and fix whatever
967 Conventionally, block device numbers are allocated from
968 predetermined contiguous area. However, extended block area
969 may introduce non-contiguous block device numbers. This
970 option forces most block device numbers to be allocated from
971 the extended space and spreads them to discover kernel or
972 userland code paths which assume predetermined contiguous
973 device number allocation.
975 Note that turning on this debug option shuffles all the
976 device numbers for all IDE and SCSI devices including libata
977 ones, so root partition specified using device number
978 directly (via rdev or root=MAJ:MIN) won't work anymore.
979 Textual device names (root=/dev/sdXn) will continue to work.
981 Say N if you are unsure.
983 config DEBUG_FORCE_WEAK_PER_CPU
984 bool "Force weak per-cpu definitions"
985 depends on DEBUG_KERNEL
987 s390 and alpha require percpu variables in modules to be
988 defined weak to work around addressing range issue which
989 puts the following two restrictions on percpu variable
992 1. percpu symbols must be unique whether static or not
993 2. percpu variables can't be defined inside a function
995 To ensure that generic code follows the above rules, this
996 option forces all percpu variables to be defined as weak.
999 tristate "Linux Kernel Dump Test Tool Module"
1004 This module enables testing of the different dumping mechanisms by
1005 inducing system failures at predefined crash points.
1006 If you don't need it: say N
1007 Choose M here to compile this code as a module. The module will be
1010 Documentation on how to use the module can be found in
1011 Documentation/fault-injection/provoke-crashes.txt
1013 config CPU_NOTIFIER_ERROR_INJECT
1014 tristate "CPU notifier error injection module"
1015 depends on HOTPLUG_CPU && DEBUG_KERNEL
1017 This option provides a kernel module that can be used to test
1018 the error handling of the cpu notifiers
1020 To compile this code as a module, choose M here: the module will
1021 be called cpu-notifier-error-inject.
1025 config FAULT_INJECTION
1026 bool "Fault-injection framework"
1027 depends on DEBUG_KERNEL
1029 Provide fault-injection framework.
1030 For more details, see Documentation/fault-injection/.
1033 bool "Fault-injection capability for kmalloc"
1034 depends on FAULT_INJECTION
1035 depends on SLAB || SLUB
1037 Provide fault-injection capability for kmalloc.
1039 config FAIL_PAGE_ALLOC
1040 bool "Fault-injection capabilitiy for alloc_pages()"
1041 depends on FAULT_INJECTION
1043 Provide fault-injection capability for alloc_pages().
1045 config FAIL_MAKE_REQUEST
1046 bool "Fault-injection capability for disk IO"
1047 depends on FAULT_INJECTION && BLOCK
1049 Provide fault-injection capability for disk IO.
1051 config FAIL_IO_TIMEOUT
1052 bool "Fault-injection capability for faking disk interrupts"
1053 depends on FAULT_INJECTION && BLOCK
1055 Provide fault-injection capability on end IO handling. This
1056 will make the block layer "forget" an interrupt as configured,
1057 thus exercising the error handling.
1059 Only works with drivers that use the generic timeout handling,
1060 for others it wont do anything.
1062 config FAULT_INJECTION_DEBUG_FS
1063 bool "Debugfs entries for fault-injection capabilities"
1064 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1066 Enable configuration of fault-injection capabilities via debugfs.
1068 config FAULT_INJECTION_STACKTRACE_FILTER
1069 bool "stacktrace filter for fault-injection capabilities"
1070 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1073 select FRAME_POINTER if !PPC && !S390 && !MICROBLAZE
1075 Provide stacktrace filter for fault-injection capabilities
1078 bool "Latency measuring infrastructure"
1079 depends on HAVE_LATENCYTOP_SUPPORT
1080 depends on DEBUG_KERNEL
1081 depends on STACKTRACE_SUPPORT
1083 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE
1090 Enable this option if you want to use the LatencyTOP tool
1091 to find out which userspace is blocking on what kernel operations.
1093 config SYSCTL_SYSCALL_CHECK
1094 bool "Sysctl checks"
1097 sys_sysctl uses binary paths that have been found challenging
1098 to properly maintain and use. This enables checks that help
1099 you to keep things correct.
1101 source mm/Kconfig.debug
1102 source kernel/trace/Kconfig
1104 config PROVIDE_OHCI1394_DMA_INIT
1105 bool "Remote debugging over FireWire early on boot"
1106 depends on PCI && X86
1108 If you want to debug problems which hang or crash the kernel early
1109 on boot and the crashing machine has a FireWire port, you can use
1110 this feature to remotely access the memory of the crashed machine
1111 over FireWire. This employs remote DMA as part of the OHCI1394
1112 specification which is now the standard for FireWire controllers.
1114 With remote DMA, you can monitor the printk buffer remotely using
1115 firescope and access all memory below 4GB using fireproxy from gdb.
1116 Even controlling a kernel debugger is possible using remote DMA.
1120 If ohci1394_dma=early is used as boot parameter, it will initialize
1121 all OHCI1394 controllers which are found in the PCI config space.
1123 As all changes to the FireWire bus such as enabling and disabling
1124 devices cause a bus reset and thereby disable remote DMA for all
1125 devices, be sure to have the cable plugged and FireWire enabled on
1126 the debugging host before booting the debug target for debugging.
1128 This code (~1k) is freed after boot. By then, the firewire stack
1129 in charge of the OHCI-1394 controllers should be used instead.
1131 See Documentation/debugging-via-ohci1394.txt for more information.
1133 config FIREWIRE_OHCI_REMOTE_DMA
1134 bool "Remote debugging over FireWire with firewire-ohci"
1135 depends on FIREWIRE_OHCI
1137 This option lets you use the FireWire bus for remote debugging
1138 with help of the firewire-ohci driver. It enables unfiltered
1139 remote DMA in firewire-ohci.
1140 See Documentation/debugging-via-ohci1394.txt for more information.
1145 bool "Build targets in Documentation/ tree"
1146 depends on HEADERS_CHECK
1148 This option attempts to build objects from the source files in the
1149 kernel Documentation/ tree.
1151 Say N if you are unsure.
1153 config DYNAMIC_DEBUG
1154 bool "Enable dynamic printk() support"
1160 Compiles debug level messages into the kernel, which would not
1161 otherwise be available at runtime. These messages can then be
1162 enabled/disabled based on various levels of scope - per source file,
1163 function, module, format string, and line number. This mechanism
1164 implicitly enables all pr_debug() and dev_dbg() calls. The impact of
1165 this compile option is a larger kernel text size of about 2%.
1169 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
1170 which is contained in the 'debugfs' filesystem. Thus, the debugfs
1171 filesystem must first be mounted before making use of this feature.
1172 We refer the control file as: <debugfs>/dynamic_debug/control. This
1173 file contains a list of the debug statements that can be enabled. The
1174 format for each line of the file is:
1176 filename:lineno [module]function flags format
1178 filename : source file of the debug statement
1179 lineno : line number of the debug statement
1180 module : module that contains the debug statement
1181 function : function that contains the debug statement
1182 flags : 'p' means the line is turned 'on' for printing
1183 format : the format used for the debug statement
1187 nullarbor:~ # cat <debugfs>/dynamic_debug/control
1188 # filename:lineno [module]function flags format
1189 fs/aio.c:222 [aio]__put_ioctx - "__put_ioctx:\040freeing\040%p\012"
1190 fs/aio.c:248 [aio]ioctx_alloc - "ENOMEM:\040nr_events\040too\040high\012"
1191 fs/aio.c:1770 [aio]sys_io_cancel - "calling\040cancel\012"
1195 // enable the message at line 1603 of file svcsock.c
1196 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
1197 <debugfs>/dynamic_debug/control
1199 // enable all the messages in file svcsock.c
1200 nullarbor:~ # echo -n 'file svcsock.c +p' >
1201 <debugfs>/dynamic_debug/control
1203 // enable all the messages in the NFS server module
1204 nullarbor:~ # echo -n 'module nfsd +p' >
1205 <debugfs>/dynamic_debug/control
1207 // enable all 12 messages in the function svc_process()
1208 nullarbor:~ # echo -n 'func svc_process +p' >
1209 <debugfs>/dynamic_debug/control
1211 // disable all 12 messages in the function svc_process()
1212 nullarbor:~ # echo -n 'func svc_process -p' >
1213 <debugfs>/dynamic_debug/control
1215 See Documentation/dynamic-debug-howto.txt for additional information.
1217 config DMA_API_DEBUG
1218 bool "Enable debugging of DMA-API usage"
1219 depends on HAVE_DMA_API_DEBUG
1221 Enable this option to debug the use of the DMA API by device drivers.
1222 With this option you will be able to detect common bugs in device
1223 drivers like double-freeing of DMA mappings or freeing mappings that
1224 were never allocated.
1225 This option causes a performance degredation. Use only if you want
1226 to debug device drivers. If unsure, say N.
1228 config ATOMIC64_SELFTEST
1229 bool "Perform an atomic64_t self-test at boot"
1231 Enable this option to test the atomic64_t functions at boot.
1235 config ASYNC_RAID6_TEST
1236 tristate "Self test for hardware accelerated raid6 recovery"
1237 depends on ASYNC_RAID6_RECOV
1240 This is a one-shot self test that permutes through the
1241 recovery of all the possible two disk failure scenarios for a
1242 N-disk array. Recovery is performed with the asynchronous
1243 raid6 recovery routines, and will optionally use an offload
1244 engine if one is available.
1248 source "samples/Kconfig"
1250 source "lib/Kconfig.kgdb"
1252 source "lib/Kconfig.kmemcheck"
1255 tristate "Test kstrto*() family of functions at runtime"