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 ENABLE_WARN_DEPRECATED
13 bool "Enable __deprecated logic"
16 Enable the __deprecated logic in the kernel build.
17 Disable this to suppress the "warning: 'foo' is deprecated
18 (declared at kernel/power/somefile.c:1234)" messages.
20 config ENABLE_MUST_CHECK
21 bool "Enable __must_check logic"
24 Enable the __must_check logic in the kernel build. Disable this to
25 suppress the "warning: ignoring return value of 'foo', declared with
26 attribute warn_unused_result" messages.
29 int "Warn for stack frames larger than (needs gcc 4.4)"
31 default 1024 if !64BIT
34 Tell gcc to warn at build time for stack frames larger than this.
35 Setting this too low will cause a lot of warnings.
36 Setting it to 0 disables the warning.
40 bool "Magic SysRq key"
43 If you say Y here, you will have some control over the system even
44 if the system crashes for example during kernel debugging (e.g., you
45 will be able to flush the buffer cache to disk, reboot the system
46 immediately or dump some status information). This is accomplished
47 by pressing various keys while holding SysRq (Alt+PrintScreen). It
48 also works on a serial console (on PC hardware at least), if you
49 send a BREAK and then within 5 seconds a command keypress. The
50 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
51 unless you really know what this hack does.
54 bool "Strip assembler-generated symbols during link"
57 Strip internal assembler-generated symbols during a link (symbols
58 that look like '.Lxxx') so they don't pollute the output of
59 get_wchan() and suchlike.
62 bool "Enable unused/obsolete exported symbols"
65 Unused but exported symbols make the kernel needlessly bigger. For
66 that reason most of these unused exports will soon be removed. This
67 option is provided temporarily to provide a transition period in case
68 some external kernel module needs one of these symbols anyway. If you
69 encounter such a case in your module, consider if you are actually
70 using the right API. (rationale: since nobody in the kernel is using
71 this in a module, there is a pretty good chance it's actually the
72 wrong interface to use). If you really need the symbol, please send a
73 mail to the linux kernel mailing list mentioning the symbol and why
74 you really need it, and what the merge plan to the mainline kernel for
78 bool "Debug Filesystem"
80 debugfs is a virtual file system that kernel developers use to put
81 debugging files into. Enable this option to be able to read and
84 For detailed documentation on the debugfs API, see
85 Documentation/DocBook/filesystems.
90 bool "Run 'make headers_check' when building vmlinux"
93 This option will extract the user-visible kernel headers whenever
94 building the kernel, and will run basic sanity checks on them to
95 ensure that exported files do not attempt to include files which
96 were not exported, etc.
98 If you're making modifications to header files which are
99 relevant for userspace, say 'Y', and check the headers
100 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
101 your build tree), to make sure they're suitable.
103 config DEBUG_SECTION_MISMATCH
104 bool "Enable full Section mismatch analysis"
106 The section mismatch analysis checks if there are illegal
107 references from one section to another section.
108 Linux will during link or during runtime drop some sections
109 and any use of code/data previously in these sections will
110 most likely result in an oops.
111 In the code functions and variables are annotated with
112 __init, __devinit etc. (see full list in include/linux/init.h)
113 which results in the code/data being placed in specific sections.
114 The section mismatch analysis is always done after a full
115 kernel build but enabling this option will in addition
117 - Add the option -fno-inline-functions-called-once to gcc
118 When inlining a function annotated __init in a non-init
119 function we would lose the section information and thus
120 the analysis would not catch the illegal reference.
121 This option tells gcc to inline less but will also
122 result in a larger kernel.
123 - Run the section mismatch analysis for each module/built-in.o
124 When we run the section mismatch analysis on vmlinux.o we
125 lose valueble information about where the mismatch was
127 Running the analysis for each module/built-in.o file
128 will tell where the mismatch happens much closer to the
129 source. The drawback is that we will report the same
130 mismatch at least twice.
131 - Enable verbose reporting from modpost to help solving
132 the section mismatches reported.
135 bool "Kernel debugging"
137 Say Y here if you are developing drivers or trying to debug and
138 identify kernel problems.
141 bool "Debug shared IRQ handlers"
142 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
144 Enable this to generate a spurious interrupt as soon as a shared
145 interrupt handler is registered, and just before one is deregistered.
146 Drivers ought to be able to handle interrupts coming in at those
147 points; some don't and need to be caught.
149 config LOCKUP_DETECTOR
150 bool "Detect Hard and Soft Lockups"
151 depends on DEBUG_KERNEL && !S390
153 Say Y here to enable the kernel to act as a watchdog to detect
154 hard and soft lockups.
156 Softlockups are bugs that cause the kernel to loop in kernel
157 mode for more than 60 seconds, without giving other tasks a
158 chance to run. The current stack trace is displayed upon
159 detection and the system will stay locked up.
161 Hardlockups are bugs that cause the CPU to loop in kernel mode
162 for more than 60 seconds, without letting other interrupts have a
163 chance to run. The current stack trace is displayed upon detection
164 and the system will stay locked up.
166 The overhead should be minimal. A periodic hrtimer runs to
167 generate interrupts and kick the watchdog task every 10-12 seconds.
168 An NMI is generated every 60 seconds or so to check for hardlockups.
170 config HARDLOCKUP_DETECTOR
171 def_bool LOCKUP_DETECTOR && PERF_EVENTS && HAVE_PERF_EVENTS_NMI && \
172 !ARCH_HAS_NMI_WATCHDOG
174 config BOOTPARAM_SOFTLOCKUP_PANIC
175 bool "Panic (Reboot) On Soft Lockups"
176 depends on LOCKUP_DETECTOR
178 Say Y here to enable the kernel to panic on "soft lockups",
179 which are bugs that cause the kernel to loop in kernel
180 mode for more than 60 seconds, without giving other tasks a
183 The panic can be used in combination with panic_timeout,
184 to cause the system to reboot automatically after a
185 lockup has been detected. This feature is useful for
186 high-availability systems that have uptime guarantees and
187 where a lockup must be resolved ASAP.
191 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
193 depends on LOCKUP_DETECTOR
195 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
196 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
198 config DETECT_HUNG_TASK
199 bool "Detect Hung Tasks"
200 depends on DEBUG_KERNEL
201 default DETECT_SOFTLOCKUP
203 Say Y here to enable the kernel to detect "hung tasks",
204 which are bugs that cause the task to be stuck in
205 uninterruptible "D" state indefinitiley.
207 When a hung task is detected, the kernel will print the
208 current stack trace (which you should report), but the
209 task will stay in uninterruptible state. If lockdep is
210 enabled then all held locks will also be reported. This
211 feature has negligible overhead.
213 config BOOTPARAM_HUNG_TASK_PANIC
214 bool "Panic (Reboot) On Hung Tasks"
215 depends on DETECT_HUNG_TASK
217 Say Y here to enable the kernel to panic on "hung tasks",
218 which are bugs that cause the kernel to leave a task stuck
219 in uninterruptible "D" state.
221 The panic can be used in combination with panic_timeout,
222 to cause the system to reboot automatically after a
223 hung task has been detected. This feature is useful for
224 high-availability systems that have uptime guarantees and
225 where a hung tasks must be resolved ASAP.
229 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
231 depends on DETECT_HUNG_TASK
233 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
234 default 1 if BOOTPARAM_HUNG_TASK_PANIC
237 bool "Collect scheduler debugging info"
238 depends on DEBUG_KERNEL && PROC_FS
241 If you say Y here, the /proc/sched_debug file will be provided
242 that can help debug the scheduler. The runtime overhead of this
246 bool "Collect scheduler statistics"
247 depends on DEBUG_KERNEL && PROC_FS
249 If you say Y here, additional code will be inserted into the
250 scheduler and related routines to collect statistics about
251 scheduler behavior and provide them in /proc/schedstat. These
252 stats may be useful for both tuning and debugging the scheduler
253 If you aren't debugging the scheduler or trying to tune a specific
254 application, you can say N to avoid the very slight overhead
258 bool "Collect kernel timers statistics"
259 depends on DEBUG_KERNEL && PROC_FS
261 If you say Y here, additional code will be inserted into the
262 timer routines to collect statistics about kernel timers being
263 reprogrammed. The statistics can be read from /proc/timer_stats.
264 The statistics collection is started by writing 1 to /proc/timer_stats,
265 writing 0 stops it. This feature is useful to collect information
266 about timer usage patterns in kernel and userspace. This feature
267 is lightweight if enabled in the kernel config but not activated
268 (it defaults to deactivated on bootup and will only be activated
269 if some application like powertop activates it explicitly).
272 bool "Debug object operations"
273 depends on DEBUG_KERNEL
275 If you say Y here, additional code will be inserted into the
276 kernel to track the life time of various objects and validate
277 the operations on those objects.
279 config DEBUG_OBJECTS_SELFTEST
280 bool "Debug objects selftest"
281 depends on DEBUG_OBJECTS
283 This enables the selftest of the object debug code.
285 config DEBUG_OBJECTS_FREE
286 bool "Debug objects in freed memory"
287 depends on DEBUG_OBJECTS
289 This enables checks whether a k/v free operation frees an area
290 which contains an object which has not been deactivated
291 properly. This can make kmalloc/kfree-intensive workloads
294 config DEBUG_OBJECTS_TIMERS
295 bool "Debug timer objects"
296 depends on DEBUG_OBJECTS
298 If you say Y here, additional code will be inserted into the
299 timer routines to track the life time of timer objects and
300 validate the timer operations.
302 config DEBUG_OBJECTS_WORK
303 bool "Debug work objects"
304 depends on DEBUG_OBJECTS
306 If you say Y here, additional code will be inserted into the
307 work queue routines to track the life time of work objects and
308 validate the work operations.
310 config DEBUG_OBJECTS_RCU_HEAD
311 bool "Debug RCU callbacks objects"
312 depends on DEBUG_OBJECTS && PREEMPT
314 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
316 config DEBUG_OBJECTS_PERCPU_COUNTER
317 bool "Debug percpu counter objects"
318 depends on DEBUG_OBJECTS
320 If you say Y here, additional code will be inserted into the
321 percpu counter routines to track the life time of percpu counter
322 objects and validate the percpu counter operations.
324 config DEBUG_OBJECTS_ENABLE_DEFAULT
325 int "debug_objects bootup default value (0-1)"
328 depends on DEBUG_OBJECTS
330 Debug objects boot parameter default value
333 bool "Debug slab memory allocations"
334 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
336 Say Y here to have the kernel do limited verification on memory
337 allocation as well as poisoning memory on free to catch use of freed
338 memory. This can make kmalloc/kfree-intensive workloads much slower.
340 config DEBUG_SLAB_LEAK
341 bool "Memory leak debugging"
342 depends on DEBUG_SLAB
345 bool "SLUB debugging on by default"
346 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
349 Boot with debugging on by default. SLUB boots by default with
350 the runtime debug capabilities switched off. Enabling this is
351 equivalent to specifying the "slub_debug" parameter on boot.
352 There is no support for more fine grained debug control like
353 possible with slub_debug=xxx. SLUB debugging may be switched
354 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
359 bool "Enable SLUB performance statistics"
360 depends on SLUB && SYSFS
362 SLUB statistics are useful to debug SLUBs allocation behavior in
363 order find ways to optimize the allocator. This should never be
364 enabled for production use since keeping statistics slows down
365 the allocator by a few percentage points. The slabinfo command
366 supports the determination of the most active slabs to figure
367 out which slabs are relevant to a particular load.
368 Try running: slabinfo -DA
370 config DEBUG_KMEMLEAK
371 bool "Kernel memory leak detector"
372 depends on DEBUG_KERNEL && EXPERIMENTAL && !MEMORY_HOTPLUG && \
373 (X86 || ARM || PPC || S390 || SPARC64 || SUPERH || MICROBLAZE || TILE)
375 select DEBUG_FS if SYSFS
376 select STACKTRACE if STACKTRACE_SUPPORT
380 Say Y here if you want to enable the memory leak
381 detector. The memory allocation/freeing is traced in a way
382 similar to the Boehm's conservative garbage collector, the
383 difference being that the orphan objects are not freed but
384 only shown in /sys/kernel/debug/kmemleak. Enabling this
385 feature will introduce an overhead to memory
386 allocations. See Documentation/kmemleak.txt for more
389 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
390 of finding leaks due to the slab objects poisoning.
392 In order to access the kmemleak file, debugfs needs to be
393 mounted (usually at /sys/kernel/debug).
395 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
396 int "Maximum kmemleak early log entries"
397 depends on DEBUG_KMEMLEAK
401 Kmemleak must track all the memory allocations to avoid
402 reporting false positives. Since memory may be allocated or
403 freed before kmemleak is initialised, an early log buffer is
404 used to store these actions. If kmemleak reports "early log
405 buffer exceeded", please increase this value.
407 config DEBUG_KMEMLEAK_TEST
408 tristate "Simple test for the kernel memory leak detector"
409 depends on DEBUG_KMEMLEAK
411 Say Y or M here to build a test for the kernel memory leak
412 detector. This option enables a module that explicitly leaks
417 config DEBUG_KMEMLEAK_DEFAULT_OFF
418 bool "Default kmemleak to off"
419 depends on DEBUG_KMEMLEAK
421 Say Y here to disable kmemleak by default. It can then be enabled
422 on the command line via kmemleak=on.
425 bool "Debug preemptible kernel"
426 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
429 If you say Y here then the kernel will use a debug variant of the
430 commonly used smp_processor_id() function and will print warnings
431 if kernel code uses it in a preemption-unsafe way. Also, the kernel
432 will detect preemption count underflows.
434 config DEBUG_RT_MUTEXES
435 bool "RT Mutex debugging, deadlock detection"
436 depends on DEBUG_KERNEL && RT_MUTEXES
438 This allows rt mutex semantics violations and rt mutex related
439 deadlocks (lockups) to be detected and reported automatically.
444 depends on DEBUG_RT_MUTEXES
446 config RT_MUTEX_TESTER
447 bool "Built-in scriptable tester for rt-mutexes"
448 depends on DEBUG_KERNEL && RT_MUTEXES
450 This option enables a rt-mutex tester.
452 config DEBUG_SPINLOCK
453 bool "Spinlock and rw-lock debugging: basic checks"
454 depends on DEBUG_KERNEL
456 Say Y here and build SMP to catch missing spinlock initialization
457 and certain other kinds of spinlock errors commonly made. This is
458 best used in conjunction with the NMI watchdog so that spinlock
459 deadlocks are also debuggable.
462 bool "Mutex debugging: basic checks"
463 depends on DEBUG_KERNEL
465 This feature allows mutex semantics violations to be detected and
469 bool "Big Kernel Lock" if (SMP || PREEMPT)
472 This is the traditional lock that is used in old code instead
473 of proper locking. All drivers that use the BKL should depend
475 Say Y here unless you are working on removing the BKL.
477 config DEBUG_LOCK_ALLOC
478 bool "Lock debugging: detect incorrect freeing of live locks"
479 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
480 select DEBUG_SPINLOCK
484 This feature will check whether any held lock (spinlock, rwlock,
485 mutex or rwsem) is incorrectly freed by the kernel, via any of the
486 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
487 vfree(), etc.), whether a live lock is incorrectly reinitialized via
488 spin_lock_init()/mutex_init()/etc., or whether there is any lock
489 held during task exit.
492 bool "Lock debugging: prove locking correctness"
493 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
495 select DEBUG_SPINLOCK
497 select DEBUG_LOCK_ALLOC
498 select TRACE_IRQFLAGS
501 This feature enables the kernel to prove that all locking
502 that occurs in the kernel runtime is mathematically
503 correct: that under no circumstance could an arbitrary (and
504 not yet triggered) combination of observed locking
505 sequences (on an arbitrary number of CPUs, running an
506 arbitrary number of tasks and interrupt contexts) cause a
509 In short, this feature enables the kernel to report locking
510 related deadlocks before they actually occur.
512 The proof does not depend on how hard and complex a
513 deadlock scenario would be to trigger: how many
514 participant CPUs, tasks and irq-contexts would be needed
515 for it to trigger. The proof also does not depend on
516 timing: if a race and a resulting deadlock is possible
517 theoretically (no matter how unlikely the race scenario
518 is), it will be proven so and will immediately be
519 reported by the kernel (once the event is observed that
520 makes the deadlock theoretically possible).
522 If a deadlock is impossible (i.e. the locking rules, as
523 observed by the kernel, are mathematically correct), the
524 kernel reports nothing.
526 NOTE: this feature can also be enabled for rwlocks, mutexes
527 and rwsems - in which case all dependencies between these
528 different locking variants are observed and mapped too, and
529 the proof of observed correctness is also maintained for an
530 arbitrary combination of these separate locking variants.
532 For more details, see Documentation/lockdep-design.txt.
535 bool "RCU debugging: prove RCU correctness"
536 depends on PROVE_LOCKING
539 This feature enables lockdep extensions that check for correct
540 use of RCU APIs. This is currently under development. Say Y
541 if you want to debug RCU usage or help work on the PROVE_RCU
544 Say N if you are unsure.
546 config PROVE_RCU_REPEATEDLY
547 bool "RCU debugging: don't disable PROVE_RCU on first splat"
551 By itself, PROVE_RCU will disable checking upon issuing the
552 first warning (or "splat"). This feature prevents such
553 disabling, allowing multiple RCU-lockdep warnings to be printed
556 Say Y to allow multiple RCU-lockdep warnings per boot.
558 Say N if you are unsure.
560 config SPARSE_RCU_POINTER
561 bool "RCU debugging: sparse-based checks for pointer usage"
564 This feature enables the __rcu sparse annotation for
565 RCU-protected pointers. This annotation will cause sparse
566 to flag any non-RCU used of annotated pointers. This can be
567 helpful when debugging RCU usage. Please note that this feature
568 is not intended to enforce code cleanliness; it is instead merely
571 Say Y to make sparse flag questionable use of RCU-protected pointers
573 Say N if you are unsure.
577 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
579 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE
584 bool "Lock usage statistics"
585 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
587 select DEBUG_SPINLOCK
589 select DEBUG_LOCK_ALLOC
592 This feature enables tracking lock contention points
594 For more details, see Documentation/lockstat.txt
596 This also enables lock events required by "perf lock",
598 If you want to use "perf lock", you also need to turn on
599 CONFIG_EVENT_TRACING.
601 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
602 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
605 bool "Lock dependency engine debugging"
606 depends on DEBUG_KERNEL && LOCKDEP
608 If you say Y here, the lock dependency engine will do
609 additional runtime checks to debug itself, at the price
610 of more runtime overhead.
612 config TRACE_IRQFLAGS
615 Enables hooks to interrupt enabling and disabling for
616 either tracing or lock debugging.
618 config DEBUG_SPINLOCK_SLEEP
619 bool "Spinlock debugging: sleep-inside-spinlock checking"
620 depends on DEBUG_KERNEL
622 If you say Y here, various routines which may sleep will become very
623 noisy if they are called with a spinlock held.
625 config DEBUG_LOCKING_API_SELFTESTS
626 bool "Locking API boot-time self-tests"
627 depends on DEBUG_KERNEL
629 Say Y here if you want the kernel to run a short self-test during
630 bootup. The self-test checks whether common types of locking bugs
631 are detected by debugging mechanisms or not. (if you disable
632 lock debugging then those bugs wont be detected of course.)
633 The following locking APIs are covered: spinlocks, rwlocks,
638 depends on STACKTRACE_SUPPORT
641 bool "kobject debugging"
642 depends on DEBUG_KERNEL
644 If you say Y here, some extra kobject debugging messages will be sent
648 bool "Highmem debugging"
649 depends on DEBUG_KERNEL && HIGHMEM
651 This options enables addition error checking for high memory systems.
652 Disable for production systems.
654 config DEBUG_BUGVERBOSE
655 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED
657 depends on ARM || AVR32 || M32R || M68K || SPARC32 || SPARC64 || \
658 FRV || SUPERH || GENERIC_BUG || BLACKFIN || MN10300
661 Say Y here to make BUG() panics output the file name and line number
662 of the BUG call as well as the EIP and oops trace. This aids
663 debugging but costs about 70-100K of memory.
666 bool "Compile the kernel with debug info"
667 depends on DEBUG_KERNEL
669 If you say Y here the resulting kernel image will include
670 debugging info resulting in a larger kernel image.
671 This adds debug symbols to the kernel and modules (gcc -g), and
672 is needed if you intend to use kernel crashdump or binary object
673 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
674 Say Y here only if you plan to debug the kernel.
678 config DEBUG_INFO_REDUCED
679 bool "Reduce debugging information"
680 depends on DEBUG_INFO
682 If you say Y here gcc is instructed to generate less debugging
683 information for structure types. This means that tools that
684 need full debugging information (like kgdb or systemtap) won't
685 be happy. But if you merely need debugging information to
686 resolve line numbers there is no loss. Advantage is that
687 build directory object sizes shrink dramatically over a full
688 DEBUG_INFO build and compile times are reduced too.
689 Only works with newer gcc versions.
693 depends on DEBUG_KERNEL
695 Enable this to turn on extended checks in the virtual-memory system
696 that may impact performance.
701 bool "Debug VM translations"
702 depends on DEBUG_KERNEL && X86
704 Enable some costly sanity checks in virtual to page code. This can
705 catch mistakes with virt_to_page() and friends.
709 config DEBUG_NOMMU_REGIONS
710 bool "Debug the global anon/private NOMMU mapping region tree"
711 depends on DEBUG_KERNEL && !MMU
713 This option causes the global tree of anonymous and private mapping
714 regions to be regularly checked for invalid topology.
716 config DEBUG_WRITECOUNT
717 bool "Debug filesystem writers count"
718 depends on DEBUG_KERNEL
720 Enable this to catch wrong use of the writers count in struct
721 vfsmount. This will increase the size of each file struct by
726 config DEBUG_MEMORY_INIT
727 bool "Debug memory initialisation" if EMBEDDED
730 Enable this for additional checks during memory initialisation.
731 The sanity checks verify aspects of the VM such as the memory model
732 and other information provided by the architecture. Verbose
733 information will be printed at KERN_DEBUG loglevel depending
734 on the mminit_loglevel= command-line option.
739 bool "Debug linked list manipulation"
740 depends on DEBUG_KERNEL
742 Enable this to turn on extended checks in the linked-list
747 config TEST_LIST_SORT
748 bool "Linked list sorting test"
749 depends on DEBUG_KERNEL
751 Enable this to turn on 'list_sort()' function test. This test is
752 executed only once during system boot, so affects only boot time.
757 bool "Debug SG table operations"
758 depends on DEBUG_KERNEL
760 Enable this to turn on checks on scatter-gather tables. This can
761 help find problems with drivers that do not properly initialize
766 config DEBUG_NOTIFIERS
767 bool "Debug notifier call chains"
768 depends on DEBUG_KERNEL
770 Enable this to turn on sanity checking for notifier call chains.
771 This is most useful for kernel developers to make sure that
772 modules properly unregister themselves from notifier chains.
773 This is a relatively cheap check but if you care about maximum
776 config DEBUG_CREDENTIALS
777 bool "Debug credential management"
778 depends on DEBUG_KERNEL
780 Enable this to turn on some debug checking for credential
781 management. The additional code keeps track of the number of
782 pointers from task_structs to any given cred struct, and checks to
783 see that this number never exceeds the usage count of the cred
786 Furthermore, if SELinux is enabled, this also checks that the
787 security pointer in the cred struct is never seen to be invalid.
792 # Select this config option from the architecture Kconfig, if it
793 # it is preferred to always offer frame pointers as a config
794 # option on the architecture (regardless of KERNEL_DEBUG):
796 config ARCH_WANT_FRAME_POINTERS
801 bool "Compile the kernel with frame pointers"
802 depends on DEBUG_KERNEL && \
803 (CRIS || M68K || M68KNOMMU || FRV || UML || \
804 AVR32 || SUPERH || BLACKFIN || MN10300) || \
805 ARCH_WANT_FRAME_POINTERS
806 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
808 If you say Y here the resulting kernel image will be slightly
809 larger and slower, but it gives very useful debugging information
810 in case of kernel bugs. (precise oopses/stacktraces/warnings)
812 config BOOT_PRINTK_DELAY
813 bool "Delay each boot printk message by N milliseconds"
814 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
816 This build option allows you to read kernel boot messages
817 by inserting a short delay after each one. The delay is
818 specified in milliseconds on the kernel command line,
819 using "boot_delay=N".
821 It is likely that you would also need to use "lpj=M" to preset
822 the "loops per jiffie" value.
823 See a previous boot log for the "lpj" value to use for your
824 system, and then set "lpj=M" before setting "boot_delay=N".
825 NOTE: Using this option may adversely affect SMP systems.
826 I.e., processors other than the first one may not boot up.
827 BOOT_PRINTK_DELAY also may cause DETECT_SOFTLOCKUP to detect
828 what it believes to be lockup conditions.
830 config RCU_TORTURE_TEST
831 tristate "torture tests for RCU"
832 depends on DEBUG_KERNEL
835 This option provides a kernel module that runs torture tests
836 on the RCU infrastructure. The kernel module may be built
837 after the fact on the running kernel to be tested, if desired.
839 Say Y here if you want RCU torture tests to be built into
841 Say M if you want the RCU torture tests to build as a module.
842 Say N if you are unsure.
844 config RCU_TORTURE_TEST_RUNNABLE
845 bool "torture tests for RCU runnable by default"
846 depends on RCU_TORTURE_TEST = y
849 This option provides a way to build the RCU torture tests
850 directly into the kernel without them starting up at boot
851 time. You can use /proc/sys/kernel/rcutorture_runnable
852 to manually override this setting. This /proc file is
853 available only when the RCU torture tests have been built
856 Say Y here if you want the RCU torture tests to start during
857 boot (you probably don't).
858 Say N here if you want the RCU torture tests to start only
859 after being manually enabled via /proc.
861 config RCU_CPU_STALL_DETECTOR
862 bool "Check for stalled CPUs delaying RCU grace periods"
863 depends on TREE_RCU || TREE_PREEMPT_RCU
866 This option causes RCU to printk information on which
867 CPUs are delaying the current grace period, but only when
868 the grace period extends for excessive time periods.
870 Say N if you want to disable such checks.
872 Say Y if you are unsure.
874 config RCU_CPU_STALL_TIMEOUT
875 int "RCU CPU stall timeout in seconds"
876 depends on RCU_CPU_STALL_DETECTOR
880 If a given RCU grace period extends more than the specified
881 number of seconds, a CPU stall warning is printed. If the
882 RCU grace period persists, additional CPU stall warnings are
883 printed at more widely spaced intervals.
885 config RCU_CPU_STALL_DETECTOR_RUNNABLE
886 bool "RCU CPU stall checking starts automatically at boot"
887 depends on RCU_CPU_STALL_DETECTOR
890 If set, start checking for RCU CPU stalls immediately on
891 boot. Otherwise, RCU CPU stall checking must be manually
894 Say Y if you are unsure.
896 Say N if you wish to suppress RCU CPU stall checking during boot.
898 config RCU_CPU_STALL_VERBOSE
899 bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR"
900 depends on RCU_CPU_STALL_DETECTOR && TREE_PREEMPT_RCU
903 This option causes RCU to printk detailed per-task information
904 for any tasks that are stalling the current RCU grace period.
906 Say N if you are unsure.
908 Say Y if you want to enable such checks.
910 config KPROBES_SANITY_TEST
911 bool "Kprobes sanity tests"
912 depends on DEBUG_KERNEL
916 This option provides for testing basic kprobes functionality on
917 boot. A sample kprobe, jprobe and kretprobe are inserted and
918 verified for functionality.
920 Say N if you are unsure.
922 config BACKTRACE_SELF_TEST
923 tristate "Self test for the backtrace code"
924 depends on DEBUG_KERNEL
927 This option provides a kernel module that can be used to test
928 the kernel stack backtrace code. This option is not useful
929 for distributions or general kernels, but only for kernel
930 developers working on architecture code.
932 Note that if you want to also test saved backtraces, you will
933 have to enable STACKTRACE as well.
935 Say N if you are unsure.
937 config DEBUG_BLOCK_EXT_DEVT
938 bool "Force extended block device numbers and spread them"
939 depends on DEBUG_KERNEL
943 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
944 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
945 YOU ARE DOING. Distros, please enable this and fix whatever
948 Conventionally, block device numbers are allocated from
949 predetermined contiguous area. However, extended block area
950 may introduce non-contiguous block device numbers. This
951 option forces most block device numbers to be allocated from
952 the extended space and spreads them to discover kernel or
953 userland code paths which assume predetermined contiguous
954 device number allocation.
956 Note that turning on this debug option shuffles all the
957 device numbers for all IDE and SCSI devices including libata
958 ones, so root partition specified using device number
959 directly (via rdev or root=MAJ:MIN) won't work anymore.
960 Textual device names (root=/dev/sdXn) will continue to work.
962 Say N if you are unsure.
964 config DEBUG_FORCE_WEAK_PER_CPU
965 bool "Force weak per-cpu definitions"
966 depends on DEBUG_KERNEL
968 s390 and alpha require percpu variables in modules to be
969 defined weak to work around addressing range issue which
970 puts the following two restrictions on percpu variable
973 1. percpu symbols must be unique whether static or not
974 2. percpu variables can't be defined inside a function
976 To ensure that generic code follows the above rules, this
977 option forces all percpu variables to be defined as weak.
980 tristate "Linux Kernel Dump Test Tool Module"
985 This module enables testing of the different dumping mechanisms by
986 inducing system failures at predefined crash points.
987 If you don't need it: say N
988 Choose M here to compile this code as a module. The module will be
991 Documentation on how to use the module can be found in
992 Documentation/fault-injection/provoke-crashes.txt
994 config CPU_NOTIFIER_ERROR_INJECT
995 tristate "CPU notifier error injection module"
996 depends on HOTPLUG_CPU && DEBUG_KERNEL
998 This option provides a kernel module that can be used to test
999 the error handling of the cpu notifiers
1001 To compile this code as a module, choose M here: the module will
1002 be called cpu-notifier-error-inject.
1006 config FAULT_INJECTION
1007 bool "Fault-injection framework"
1008 depends on DEBUG_KERNEL
1010 Provide fault-injection framework.
1011 For more details, see Documentation/fault-injection/.
1014 bool "Fault-injection capability for kmalloc"
1015 depends on FAULT_INJECTION
1016 depends on SLAB || SLUB
1018 Provide fault-injection capability for kmalloc.
1020 config FAIL_PAGE_ALLOC
1021 bool "Fault-injection capabilitiy for alloc_pages()"
1022 depends on FAULT_INJECTION
1024 Provide fault-injection capability for alloc_pages().
1026 config FAIL_MAKE_REQUEST
1027 bool "Fault-injection capability for disk IO"
1028 depends on FAULT_INJECTION && BLOCK
1030 Provide fault-injection capability for disk IO.
1032 config FAIL_IO_TIMEOUT
1033 bool "Fault-injection capability for faking disk interrupts"
1034 depends on FAULT_INJECTION && BLOCK
1036 Provide fault-injection capability on end IO handling. This
1037 will make the block layer "forget" an interrupt as configured,
1038 thus exercising the error handling.
1040 Only works with drivers that use the generic timeout handling,
1041 for others it wont do anything.
1043 config FAULT_INJECTION_DEBUG_FS
1044 bool "Debugfs entries for fault-injection capabilities"
1045 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1047 Enable configuration of fault-injection capabilities via debugfs.
1049 config FAULT_INJECTION_STACKTRACE_FILTER
1050 bool "stacktrace filter for fault-injection capabilities"
1051 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1054 select FRAME_POINTER if !PPC && !S390 && !MICROBLAZE
1056 Provide stacktrace filter for fault-injection capabilities
1059 bool "Latency measuring infrastructure"
1060 depends on HAVE_LATENCYTOP_SUPPORT
1061 depends on DEBUG_KERNEL
1062 depends on STACKTRACE_SUPPORT
1064 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE
1071 Enable this option if you want to use the LatencyTOP tool
1072 to find out which userspace is blocking on what kernel operations.
1074 config SYSCTL_SYSCALL_CHECK
1075 bool "Sysctl checks"
1078 sys_sysctl uses binary paths that have been found challenging
1079 to properly maintain and use. This enables checks that help
1080 you to keep things correct.
1082 source mm/Kconfig.debug
1083 source kernel/trace/Kconfig
1085 config PROVIDE_OHCI1394_DMA_INIT
1086 bool "Remote debugging over FireWire early on boot"
1087 depends on PCI && X86
1089 If you want to debug problems which hang or crash the kernel early
1090 on boot and the crashing machine has a FireWire port, you can use
1091 this feature to remotely access the memory of the crashed machine
1092 over FireWire. This employs remote DMA as part of the OHCI1394
1093 specification which is now the standard for FireWire controllers.
1095 With remote DMA, you can monitor the printk buffer remotely using
1096 firescope and access all memory below 4GB using fireproxy from gdb.
1097 Even controlling a kernel debugger is possible using remote DMA.
1101 If ohci1394_dma=early is used as boot parameter, it will initialize
1102 all OHCI1394 controllers which are found in the PCI config space.
1104 As all changes to the FireWire bus such as enabling and disabling
1105 devices cause a bus reset and thereby disable remote DMA for all
1106 devices, be sure to have the cable plugged and FireWire enabled on
1107 the debugging host before booting the debug target for debugging.
1109 This code (~1k) is freed after boot. By then, the firewire stack
1110 in charge of the OHCI-1394 controllers should be used instead.
1112 See Documentation/debugging-via-ohci1394.txt for more information.
1114 config FIREWIRE_OHCI_REMOTE_DMA
1115 bool "Remote debugging over FireWire with firewire-ohci"
1116 depends on FIREWIRE_OHCI
1118 This option lets you use the FireWire bus for remote debugging
1119 with help of the firewire-ohci driver. It enables unfiltered
1120 remote DMA in firewire-ohci.
1121 See Documentation/debugging-via-ohci1394.txt for more information.
1126 bool "Build targets in Documentation/ tree"
1127 depends on HEADERS_CHECK
1129 This option attempts to build objects from the source files in the
1130 kernel Documentation/ tree.
1132 Say N if you are unsure.
1134 config DYNAMIC_DEBUG
1135 bool "Enable dynamic printk() support"
1141 Compiles debug level messages into the kernel, which would not
1142 otherwise be available at runtime. These messages can then be
1143 enabled/disabled based on various levels of scope - per source file,
1144 function, module, format string, and line number. This mechanism
1145 implicitly enables all pr_debug() and dev_dbg() calls. The impact of
1146 this compile option is a larger kernel text size of about 2%.
1150 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
1151 which is contained in the 'debugfs' filesystem. Thus, the debugfs
1152 filesystem must first be mounted before making use of this feature.
1153 We refer the control file as: <debugfs>/dynamic_debug/control. This
1154 file contains a list of the debug statements that can be enabled. The
1155 format for each line of the file is:
1157 filename:lineno [module]function flags format
1159 filename : source file of the debug statement
1160 lineno : line number of the debug statement
1161 module : module that contains the debug statement
1162 function : function that contains the debug statement
1163 flags : 'p' means the line is turned 'on' for printing
1164 format : the format used for the debug statement
1168 nullarbor:~ # cat <debugfs>/dynamic_debug/control
1169 # filename:lineno [module]function flags format
1170 fs/aio.c:222 [aio]__put_ioctx - "__put_ioctx:\040freeing\040%p\012"
1171 fs/aio.c:248 [aio]ioctx_alloc - "ENOMEM:\040nr_events\040too\040high\012"
1172 fs/aio.c:1770 [aio]sys_io_cancel - "calling\040cancel\012"
1176 // enable the message at line 1603 of file svcsock.c
1177 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
1178 <debugfs>/dynamic_debug/control
1180 // enable all the messages in file svcsock.c
1181 nullarbor:~ # echo -n 'file svcsock.c +p' >
1182 <debugfs>/dynamic_debug/control
1184 // enable all the messages in the NFS server module
1185 nullarbor:~ # echo -n 'module nfsd +p' >
1186 <debugfs>/dynamic_debug/control
1188 // enable all 12 messages in the function svc_process()
1189 nullarbor:~ # echo -n 'func svc_process +p' >
1190 <debugfs>/dynamic_debug/control
1192 // disable all 12 messages in the function svc_process()
1193 nullarbor:~ # echo -n 'func svc_process -p' >
1194 <debugfs>/dynamic_debug/control
1196 See Documentation/dynamic-debug-howto.txt for additional information.
1198 config DMA_API_DEBUG
1199 bool "Enable debugging of DMA-API usage"
1200 depends on HAVE_DMA_API_DEBUG
1202 Enable this option to debug the use of the DMA API by device drivers.
1203 With this option you will be able to detect common bugs in device
1204 drivers like double-freeing of DMA mappings or freeing mappings that
1205 were never allocated.
1206 This option causes a performance degredation. Use only if you want
1207 to debug device drivers. If unsure, say N.
1209 config ATOMIC64_SELFTEST
1210 bool "Perform an atomic64_t self-test at boot"
1212 Enable this option to test the atomic64_t functions at boot.
1216 config ASYNC_RAID6_TEST
1217 tristate "Self test for hardware accelerated raid6 recovery"
1218 depends on ASYNC_RAID6_RECOV
1221 This is a one-shot self test that permutes through the
1222 recovery of all the possible two disk failure scenarios for a
1223 N-disk array. Recovery is performed with the asynchronous
1224 raid6 recovery routines, and will optionally use an offload
1225 engine if one is available.
1229 source "samples/Kconfig"
1231 source "lib/Kconfig.kgdb"
1233 source "lib/Kconfig.kmemcheck"