3 bool "Show timing information on printks"
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.
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.
14 The behavior is also controlled by the kernel command line
15 parameter printk.time=1. See Documentation/kernel-parameters.txt
17 config DEFAULT_MESSAGE_LOGLEVEL
18 int "Default message log level (1-7)"
22 Default log level for printk statements with no specified priority.
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
28 menu "Compile-time checks and compiler options"
31 bool "Compile the kernel with debug info"
32 depends on DEBUG_KERNEL
34 If you say Y here the resulting kernel image will include
35 debugging info resulting in a larger kernel image.
36 This adds debug symbols to the kernel and modules (gcc -g), and
37 is needed if you intend to use kernel crashdump or binary object
38 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
39 Say Y here only if you plan to debug the kernel.
43 config DEBUG_INFO_REDUCED
44 bool "Reduce debugging information"
47 If you say Y here gcc is instructed to generate less debugging
48 information for structure types. This means that tools that
49 need full debugging information (like kgdb or systemtap) won't
50 be happy. But if you merely need debugging information to
51 resolve line numbers there is no loss. Advantage is that
52 build directory object sizes shrink dramatically over a full
53 DEBUG_INFO build and compile times are reduced too.
54 Only works with newer gcc versions.
56 config ENABLE_WARN_DEPRECATED
57 bool "Enable __deprecated logic"
60 Enable the __deprecated logic in the kernel build.
61 Disable this to suppress the "warning: 'foo' is deprecated
62 (declared at kernel/power/somefile.c:1234)" messages.
64 config ENABLE_MUST_CHECK
65 bool "Enable __must_check logic"
68 Enable the __must_check logic in the kernel build. Disable this to
69 suppress the "warning: ignoring return value of 'foo', declared with
70 attribute warn_unused_result" messages.
73 int "Warn for stack frames larger than (needs gcc 4.4)"
75 default 1024 if !64BIT
78 Tell gcc to warn at build time for stack frames larger than this.
79 Setting this too low will cause a lot of warnings.
80 Setting it to 0 disables the warning.
84 bool "Strip assembler-generated symbols during link"
87 Strip internal assembler-generated symbols during a link (symbols
88 that look like '.Lxxx') so they don't pollute the output of
89 get_wchan() and suchlike.
92 bool "Generate readable assembler code"
93 depends on DEBUG_KERNEL
95 Disable some compiler optimizations that tend to generate human unreadable
96 assembler output. This may make the kernel slightly slower, but it helps
97 to keep kernel developers who have to stare a lot at assembler listings
100 config UNUSED_SYMBOLS
101 bool "Enable unused/obsolete exported symbols"
104 Unused but exported symbols make the kernel needlessly bigger. For
105 that reason most of these unused exports will soon be removed. This
106 option is provided temporarily to provide a transition period in case
107 some external kernel module needs one of these symbols anyway. If you
108 encounter such a case in your module, consider if you are actually
109 using the right API. (rationale: since nobody in the kernel is using
110 this in a module, there is a pretty good chance it's actually the
111 wrong interface to use). If you really need the symbol, please send a
112 mail to the linux kernel mailing list mentioning the symbol and why
113 you really need it, and what the merge plan to the mainline kernel for
117 bool "Debug Filesystem"
119 debugfs is a virtual file system that kernel developers use to put
120 debugging files into. Enable this option to be able to read and
121 write to these files.
123 For detailed documentation on the debugfs API, see
124 Documentation/DocBook/filesystems.
129 bool "Run 'make headers_check' when building vmlinux"
132 This option will extract the user-visible kernel headers whenever
133 building the kernel, and will run basic sanity checks on them to
134 ensure that exported files do not attempt to include files which
135 were not exported, etc.
137 If you're making modifications to header files which are
138 relevant for userspace, say 'Y', and check the headers
139 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
140 your build tree), to make sure they're suitable.
142 config DEBUG_SECTION_MISMATCH
143 bool "Enable full Section mismatch analysis"
145 The section mismatch analysis checks if there are illegal
146 references from one section to another section.
147 During linktime or runtime, some sections are dropped;
148 any use of code/data previously in these sections would
149 most likely result in an oops.
150 In the code, functions and variables are annotated with
151 __init, __cpuinit, etc. (see the full list in include/linux/init.h),
152 which results in the code/data being placed in specific sections.
153 The section mismatch analysis is always performed after a full
154 kernel build, and enabling this option causes the following
155 additional steps to occur:
156 - Add the option -fno-inline-functions-called-once to gcc commands.
157 When inlining a function annotated with __init in a non-init
158 function, we would lose the section information and thus
159 the analysis would not catch the illegal reference.
160 This option tells gcc to inline less (but it does result in
162 - Run the section mismatch analysis for each module/built-in.o file.
163 When we run the section mismatch analysis on vmlinux.o, we
164 lose valueble information about where the mismatch was
166 Running the analysis for each module/built-in.o file
167 tells where the mismatch happens much closer to the
168 source. The drawback is that the same mismatch is
169 reported at least twice.
170 - Enable verbose reporting from modpost in order to help resolve
171 the section mismatches that are reported.
174 # Select this config option from the architecture Kconfig, if it
175 # is preferred to always offer frame pointers as a config
176 # option on the architecture (regardless of KERNEL_DEBUG):
178 config ARCH_WANT_FRAME_POINTERS
183 bool "Compile the kernel with frame pointers"
184 depends on DEBUG_KERNEL && \
185 (CRIS || M68K || FRV || UML || \
186 AVR32 || SUPERH || BLACKFIN || MN10300 || METAG) || \
187 ARCH_WANT_FRAME_POINTERS
188 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
190 If you say Y here the resulting kernel image will be slightly
191 larger and slower, but it gives very useful debugging information
192 in case of kernel bugs. (precise oopses/stacktraces/warnings)
194 config DEBUG_FORCE_WEAK_PER_CPU
195 bool "Force weak per-cpu definitions"
196 depends on DEBUG_KERNEL
198 s390 and alpha require percpu variables in modules to be
199 defined weak to work around addressing range issue which
200 puts the following two restrictions on percpu variable
203 1. percpu symbols must be unique whether static or not
204 2. percpu variables can't be defined inside a function
206 To ensure that generic code follows the above rules, this
207 option forces all percpu variables to be defined as weak.
209 endmenu # "Compiler options"
212 bool "Magic SysRq key"
215 If you say Y here, you will have some control over the system even
216 if the system crashes for example during kernel debugging (e.g., you
217 will be able to flush the buffer cache to disk, reboot the system
218 immediately or dump some status information). This is accomplished
219 by pressing various keys while holding SysRq (Alt+PrintScreen). It
220 also works on a serial console (on PC hardware at least), if you
221 send a BREAK and then within 5 seconds a command keypress. The
222 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
223 unless you really know what this hack does.
226 bool "Kernel debugging"
228 Say Y here if you are developing drivers or trying to debug and
229 identify kernel problems.
231 menu "Memory Debugging"
233 source mm/Kconfig.debug
236 bool "Debug object operations"
237 depends on DEBUG_KERNEL
239 If you say Y here, additional code will be inserted into the
240 kernel to track the life time of various objects and validate
241 the operations on those objects.
243 config DEBUG_OBJECTS_SELFTEST
244 bool "Debug objects selftest"
245 depends on DEBUG_OBJECTS
247 This enables the selftest of the object debug code.
249 config DEBUG_OBJECTS_FREE
250 bool "Debug objects in freed memory"
251 depends on DEBUG_OBJECTS
253 This enables checks whether a k/v free operation frees an area
254 which contains an object which has not been deactivated
255 properly. This can make kmalloc/kfree-intensive workloads
258 config DEBUG_OBJECTS_TIMERS
259 bool "Debug timer objects"
260 depends on DEBUG_OBJECTS
262 If you say Y here, additional code will be inserted into the
263 timer routines to track the life time of timer objects and
264 validate the timer operations.
266 config DEBUG_OBJECTS_WORK
267 bool "Debug work objects"
268 depends on DEBUG_OBJECTS
270 If you say Y here, additional code will be inserted into the
271 work queue routines to track the life time of work objects and
272 validate the work operations.
274 config DEBUG_OBJECTS_RCU_HEAD
275 bool "Debug RCU callbacks objects"
276 depends on DEBUG_OBJECTS
278 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
280 config DEBUG_OBJECTS_PERCPU_COUNTER
281 bool "Debug percpu counter objects"
282 depends on DEBUG_OBJECTS
284 If you say Y here, additional code will be inserted into the
285 percpu counter routines to track the life time of percpu counter
286 objects and validate the percpu counter operations.
288 config DEBUG_OBJECTS_ENABLE_DEFAULT
289 int "debug_objects bootup default value (0-1)"
292 depends on DEBUG_OBJECTS
294 Debug objects boot parameter default value
297 bool "Debug slab memory allocations"
298 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
300 Say Y here to have the kernel do limited verification on memory
301 allocation as well as poisoning memory on free to catch use of freed
302 memory. This can make kmalloc/kfree-intensive workloads much slower.
304 config DEBUG_SLAB_LEAK
305 bool "Memory leak debugging"
306 depends on DEBUG_SLAB
309 bool "SLUB debugging on by default"
310 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
313 Boot with debugging on by default. SLUB boots by default with
314 the runtime debug capabilities switched off. Enabling this is
315 equivalent to specifying the "slub_debug" parameter on boot.
316 There is no support for more fine grained debug control like
317 possible with slub_debug=xxx. SLUB debugging may be switched
318 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
323 bool "Enable SLUB performance statistics"
324 depends on SLUB && SYSFS
326 SLUB statistics are useful to debug SLUBs allocation behavior in
327 order find ways to optimize the allocator. This should never be
328 enabled for production use since keeping statistics slows down
329 the allocator by a few percentage points. The slabinfo command
330 supports the determination of the most active slabs to figure
331 out which slabs are relevant to a particular load.
332 Try running: slabinfo -DA
334 config HAVE_DEBUG_KMEMLEAK
337 config DEBUG_KMEMLEAK
338 bool "Kernel memory leak detector"
339 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
341 select STACKTRACE if STACKTRACE_SUPPORT
345 Say Y here if you want to enable the memory leak
346 detector. The memory allocation/freeing is traced in a way
347 similar to the Boehm's conservative garbage collector, the
348 difference being that the orphan objects are not freed but
349 only shown in /sys/kernel/debug/kmemleak. Enabling this
350 feature will introduce an overhead to memory
351 allocations. See Documentation/kmemleak.txt for more
354 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
355 of finding leaks due to the slab objects poisoning.
357 In order to access the kmemleak file, debugfs needs to be
358 mounted (usually at /sys/kernel/debug).
360 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
361 int "Maximum kmemleak early log entries"
362 depends on DEBUG_KMEMLEAK
366 Kmemleak must track all the memory allocations to avoid
367 reporting false positives. Since memory may be allocated or
368 freed before kmemleak is initialised, an early log buffer is
369 used to store these actions. If kmemleak reports "early log
370 buffer exceeded", please increase this value.
372 config DEBUG_KMEMLEAK_TEST
373 tristate "Simple test for the kernel memory leak detector"
374 depends on DEBUG_KMEMLEAK && m
376 This option enables a module that explicitly leaks memory.
380 config DEBUG_KMEMLEAK_DEFAULT_OFF
381 bool "Default kmemleak to off"
382 depends on DEBUG_KMEMLEAK
384 Say Y here to disable kmemleak by default. It can then be enabled
385 on the command line via kmemleak=on.
387 config DEBUG_STACK_USAGE
388 bool "Stack utilization instrumentation"
389 depends on DEBUG_KERNEL && !IA64 && !PARISC && !METAG
391 Enables the display of the minimum amount of free stack which each
392 task has ever had available in the sysrq-T and sysrq-P debug output.
394 This option will slow down process creation somewhat.
398 depends on DEBUG_KERNEL
400 Enable this to turn on extended checks in the virtual-memory system
401 that may impact performance.
406 bool "Debug VM red-black trees"
409 Enable this to turn on more extended checks in the virtual-memory
410 system that may impact performance.
415 bool "Debug VM translations"
416 depends on DEBUG_KERNEL && X86
418 Enable some costly sanity checks in virtual to page code. This can
419 catch mistakes with virt_to_page() and friends.
423 config DEBUG_NOMMU_REGIONS
424 bool "Debug the global anon/private NOMMU mapping region tree"
425 depends on DEBUG_KERNEL && !MMU
427 This option causes the global tree of anonymous and private mapping
428 regions to be regularly checked for invalid topology.
430 config DEBUG_MEMORY_INIT
431 bool "Debug memory initialisation" if EXPERT
434 Enable this for additional checks during memory initialisation.
435 The sanity checks verify aspects of the VM such as the memory model
436 and other information provided by the architecture. Verbose
437 information will be printed at KERN_DEBUG loglevel depending
438 on the mminit_loglevel= command-line option.
442 config MEMORY_NOTIFIER_ERROR_INJECT
443 tristate "Memory hotplug notifier error injection module"
444 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
446 This option provides the ability to inject artificial errors to
447 memory hotplug notifier chain callbacks. It is controlled through
448 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
450 If the notifier call chain should be failed with some events
451 notified, write the error code to "actions/<notifier event>/error".
453 Example: Inject memory hotplug offline error (-12 == -ENOMEM)
455 # cd /sys/kernel/debug/notifier-error-inject/memory
456 # echo -12 > actions/MEM_GOING_OFFLINE/error
457 # echo offline > /sys/devices/system/memory/memoryXXX/state
458 bash: echo: write error: Cannot allocate memory
460 To compile this code as a module, choose M here: the module will
461 be called memory-notifier-error-inject.
465 config DEBUG_PER_CPU_MAPS
466 bool "Debug access to per_cpu maps"
467 depends on DEBUG_KERNEL
470 Say Y to verify that the per_cpu map being accessed has
471 been set up. This adds a fair amount of code to kernel memory
472 and decreases performance.
477 bool "Highmem debugging"
478 depends on DEBUG_KERNEL && HIGHMEM
480 This options enables addition error checking for high memory systems.
481 Disable for production systems.
483 config HAVE_DEBUG_STACKOVERFLOW
486 config DEBUG_STACKOVERFLOW
487 bool "Check for stack overflows"
488 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
490 Say Y here if you want to check for overflows of kernel, IRQ
491 and exception stacks (if your archicture uses them). This
492 option will show detailed messages if free stack space drops
493 below a certain limit.
495 These kinds of bugs usually occur when call-chains in the
496 kernel get too deep, especially when interrupts are
499 Use this in cases where you see apparently random memory
500 corruption, especially if it appears in 'struct thread_info'
502 If in doubt, say "N".
504 source "lib/Kconfig.kmemcheck"
506 endmenu # "Memory Debugging"
509 bool "Debug shared IRQ handlers"
510 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
512 Enable this to generate a spurious interrupt as soon as a shared
513 interrupt handler is registered, and just before one is deregistered.
514 Drivers ought to be able to handle interrupts coming in at those
515 points; some don't and need to be caught.
517 config LOCKUP_DETECTOR
518 bool "Detect Hard and Soft Lockups"
519 depends on DEBUG_KERNEL && !S390
521 Say Y here to enable the kernel to act as a watchdog to detect
522 hard and soft lockups.
524 Softlockups are bugs that cause the kernel to loop in kernel
525 mode for more than 20 seconds, without giving other tasks a
526 chance to run. The current stack trace is displayed upon
527 detection and the system will stay locked up.
529 Hardlockups are bugs that cause the CPU to loop in kernel mode
530 for more than 10 seconds, without letting other interrupts have a
531 chance to run. The current stack trace is displayed upon detection
532 and the system will stay locked up.
534 The overhead should be minimal. A periodic hrtimer runs to
535 generate interrupts and kick the watchdog task every 4 seconds.
536 An NMI is generated every 10 seconds or so to check for hardlockups.
538 The frequency of hrtimer and NMI events and the soft and hard lockup
539 thresholds can be controlled through the sysctl watchdog_thresh.
541 config HARDLOCKUP_DETECTOR
543 depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
544 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
546 config BOOTPARAM_HARDLOCKUP_PANIC
547 bool "Panic (Reboot) On Hard Lockups"
548 depends on HARDLOCKUP_DETECTOR
550 Say Y here to enable the kernel to panic on "hard lockups",
551 which are bugs that cause the kernel to loop in kernel
552 mode with interrupts disabled for more than 10 seconds (configurable
553 using the watchdog_thresh sysctl).
557 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
559 depends on HARDLOCKUP_DETECTOR
561 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
562 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
564 config BOOTPARAM_SOFTLOCKUP_PANIC
565 bool "Panic (Reboot) On Soft Lockups"
566 depends on LOCKUP_DETECTOR
568 Say Y here to enable the kernel to panic on "soft lockups",
569 which are bugs that cause the kernel to loop in kernel
570 mode for more than 20 seconds (configurable using the watchdog_thresh
571 sysctl), without giving other tasks a chance to run.
573 The panic can be used in combination with panic_timeout,
574 to cause the system to reboot automatically after a
575 lockup has been detected. This feature is useful for
576 high-availability systems that have uptime guarantees and
577 where a lockup must be resolved ASAP.
581 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
583 depends on LOCKUP_DETECTOR
585 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
586 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
591 Say Y here to enable the kernel to panic when it oopses. This
592 has the same effect as setting oops=panic on the kernel command
595 This feature is useful to ensure that the kernel does not do
596 anything erroneous after an oops which could result in data
597 corruption or other issues.
601 config PANIC_ON_OOPS_VALUE
604 default 0 if !PANIC_ON_OOPS
605 default 1 if PANIC_ON_OOPS
607 config DETECT_HUNG_TASK
608 bool "Detect Hung Tasks"
609 depends on DEBUG_KERNEL
610 default LOCKUP_DETECTOR
612 Say Y here to enable the kernel to detect "hung tasks",
613 which are bugs that cause the task to be stuck in
614 uninterruptible "D" state indefinitiley.
616 When a hung task is detected, the kernel will print the
617 current stack trace (which you should report), but the
618 task will stay in uninterruptible state. If lockdep is
619 enabled then all held locks will also be reported. This
620 feature has negligible overhead.
622 config DEFAULT_HUNG_TASK_TIMEOUT
623 int "Default timeout for hung task detection (in seconds)"
624 depends on DETECT_HUNG_TASK
627 This option controls the default timeout (in seconds) used
628 to determine when a task has become non-responsive and should
631 It can be adjusted at runtime via the kernel.hung_task_timeout_secs
632 sysctl or by writing a value to
633 /proc/sys/kernel/hung_task_timeout_secs.
635 A timeout of 0 disables the check. The default is two minutes.
636 Keeping the default should be fine in most cases.
638 config BOOTPARAM_HUNG_TASK_PANIC
639 bool "Panic (Reboot) On Hung Tasks"
640 depends on DETECT_HUNG_TASK
642 Say Y here to enable the kernel to panic on "hung tasks",
643 which are bugs that cause the kernel to leave a task stuck
644 in uninterruptible "D" state.
646 The panic can be used in combination with panic_timeout,
647 to cause the system to reboot automatically after a
648 hung task has been detected. This feature is useful for
649 high-availability systems that have uptime guarantees and
650 where a hung tasks must be resolved ASAP.
654 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
656 depends on DETECT_HUNG_TASK
658 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
659 default 1 if BOOTPARAM_HUNG_TASK_PANIC
662 bool "Collect scheduler debugging info"
663 depends on DEBUG_KERNEL && PROC_FS
666 If you say Y here, the /proc/sched_debug file will be provided
667 that can help debug the scheduler. The runtime overhead of this
671 bool "Collect scheduler statistics"
672 depends on DEBUG_KERNEL && PROC_FS
674 If you say Y here, additional code will be inserted into the
675 scheduler and related routines to collect statistics about
676 scheduler behavior and provide them in /proc/schedstat. These
677 stats may be useful for both tuning and debugging the scheduler
678 If you aren't debugging the scheduler or trying to tune a specific
679 application, you can say N to avoid the very slight overhead
683 bool "Collect kernel timers statistics"
684 depends on DEBUG_KERNEL && PROC_FS
686 If you say Y here, additional code will be inserted into the
687 timer routines to collect statistics about kernel timers being
688 reprogrammed. The statistics can be read from /proc/timer_stats.
689 The statistics collection is started by writing 1 to /proc/timer_stats,
690 writing 0 stops it. This feature is useful to collect information
691 about timer usage patterns in kernel and userspace. This feature
692 is lightweight if enabled in the kernel config but not activated
693 (it defaults to deactivated on bootup and will only be activated
694 if some application like powertop activates it explicitly).
697 bool "Debug preemptible kernel"
698 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
701 If you say Y here then the kernel will use a debug variant of the
702 commonly used smp_processor_id() function and will print warnings
703 if kernel code uses it in a preemption-unsafe way. Also, the kernel
704 will detect preemption count underflows.
706 menu "Lock Debugging (spinlocks, mutexes, etc...)"
708 config DEBUG_RT_MUTEXES
709 bool "RT Mutex debugging, deadlock detection"
710 depends on DEBUG_KERNEL && RT_MUTEXES
712 This allows rt mutex semantics violations and rt mutex related
713 deadlocks (lockups) to be detected and reported automatically.
718 depends on DEBUG_RT_MUTEXES
720 config RT_MUTEX_TESTER
721 bool "Built-in scriptable tester for rt-mutexes"
722 depends on DEBUG_KERNEL && RT_MUTEXES
724 This option enables a rt-mutex tester.
726 config DEBUG_SPINLOCK
727 bool "Spinlock and rw-lock debugging: basic checks"
728 depends on DEBUG_KERNEL
729 select UNINLINE_SPIN_UNLOCK
731 Say Y here and build SMP to catch missing spinlock initialization
732 and certain other kinds of spinlock errors commonly made. This is
733 best used in conjunction with the NMI watchdog so that spinlock
734 deadlocks are also debuggable.
737 bool "Mutex debugging: basic checks"
738 depends on DEBUG_KERNEL
740 This feature allows mutex semantics violations to be detected and
743 config DEBUG_LOCK_ALLOC
744 bool "Lock debugging: detect incorrect freeing of live locks"
745 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
746 select DEBUG_SPINLOCK
750 This feature will check whether any held lock (spinlock, rwlock,
751 mutex or rwsem) is incorrectly freed by the kernel, via any of the
752 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
753 vfree(), etc.), whether a live lock is incorrectly reinitialized via
754 spin_lock_init()/mutex_init()/etc., or whether there is any lock
755 held during task exit.
758 bool "Lock debugging: prove locking correctness"
759 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
761 select DEBUG_SPINLOCK
763 select DEBUG_LOCK_ALLOC
764 select TRACE_IRQFLAGS
767 This feature enables the kernel to prove that all locking
768 that occurs in the kernel runtime is mathematically
769 correct: that under no circumstance could an arbitrary (and
770 not yet triggered) combination of observed locking
771 sequences (on an arbitrary number of CPUs, running an
772 arbitrary number of tasks and interrupt contexts) cause a
775 In short, this feature enables the kernel to report locking
776 related deadlocks before they actually occur.
778 The proof does not depend on how hard and complex a
779 deadlock scenario would be to trigger: how many
780 participant CPUs, tasks and irq-contexts would be needed
781 for it to trigger. The proof also does not depend on
782 timing: if a race and a resulting deadlock is possible
783 theoretically (no matter how unlikely the race scenario
784 is), it will be proven so and will immediately be
785 reported by the kernel (once the event is observed that
786 makes the deadlock theoretically possible).
788 If a deadlock is impossible (i.e. the locking rules, as
789 observed by the kernel, are mathematically correct), the
790 kernel reports nothing.
792 NOTE: this feature can also be enabled for rwlocks, mutexes
793 and rwsems - in which case all dependencies between these
794 different locking variants are observed and mapped too, and
795 the proof of observed correctness is also maintained for an
796 arbitrary combination of these separate locking variants.
798 For more details, see Documentation/lockdep-design.txt.
802 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
804 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE
809 bool "Lock usage statistics"
810 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
812 select DEBUG_SPINLOCK
814 select DEBUG_LOCK_ALLOC
817 This feature enables tracking lock contention points
819 For more details, see Documentation/lockstat.txt
821 This also enables lock events required by "perf lock",
823 If you want to use "perf lock", you also need to turn on
824 CONFIG_EVENT_TRACING.
826 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
827 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
830 bool "Lock dependency engine debugging"
831 depends on DEBUG_KERNEL && LOCKDEP
833 If you say Y here, the lock dependency engine will do
834 additional runtime checks to debug itself, at the price
835 of more runtime overhead.
837 config DEBUG_ATOMIC_SLEEP
838 bool "Sleep inside atomic section checking"
840 depends on DEBUG_KERNEL
842 If you say Y here, various routines which may sleep will become very
843 noisy if they are called inside atomic sections: when a spinlock is
844 held, inside an rcu read side critical section, inside preempt disabled
845 sections, inside an interrupt, etc...
847 config DEBUG_LOCKING_API_SELFTESTS
848 bool "Locking API boot-time self-tests"
849 depends on DEBUG_KERNEL
851 Say Y here if you want the kernel to run a short self-test during
852 bootup. The self-test checks whether common types of locking bugs
853 are detected by debugging mechanisms or not. (if you disable
854 lock debugging then those bugs wont be detected of course.)
855 The following locking APIs are covered: spinlocks, rwlocks,
858 endmenu # lock debugging
860 config TRACE_IRQFLAGS
863 Enables hooks to interrupt enabling and disabling for
864 either tracing or lock debugging.
868 depends on STACKTRACE_SUPPORT
871 bool "kobject debugging"
872 depends on DEBUG_KERNEL
874 If you say Y here, some extra kobject debugging messages will be sent
877 config HAVE_DEBUG_BUGVERBOSE
880 config DEBUG_BUGVERBOSE
881 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
882 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
885 Say Y here to make BUG() panics output the file name and line number
886 of the BUG call as well as the EIP and oops trace. This aids
887 debugging but costs about 70-100K of memory.
889 config DEBUG_WRITECOUNT
890 bool "Debug filesystem writers count"
891 depends on DEBUG_KERNEL
893 Enable this to catch wrong use of the writers count in struct
894 vfsmount. This will increase the size of each file struct by
900 bool "Debug linked list manipulation"
901 depends on DEBUG_KERNEL
903 Enable this to turn on extended checks in the linked-list
909 bool "Debug SG table operations"
910 depends on DEBUG_KERNEL
912 Enable this to turn on checks on scatter-gather tables. This can
913 help find problems with drivers that do not properly initialize
918 config DEBUG_NOTIFIERS
919 bool "Debug notifier call chains"
920 depends on DEBUG_KERNEL
922 Enable this to turn on sanity checking for notifier call chains.
923 This is most useful for kernel developers to make sure that
924 modules properly unregister themselves from notifier chains.
925 This is a relatively cheap check but if you care about maximum
928 config DEBUG_CREDENTIALS
929 bool "Debug credential management"
930 depends on DEBUG_KERNEL
932 Enable this to turn on some debug checking for credential
933 management. The additional code keeps track of the number of
934 pointers from task_structs to any given cred struct, and checks to
935 see that this number never exceeds the usage count of the cred
938 Furthermore, if SELinux is enabled, this also checks that the
939 security pointer in the cred struct is never seen to be invalid.
943 config BOOT_PRINTK_DELAY
944 bool "Delay each boot printk message by N milliseconds"
945 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
947 This build option allows you to read kernel boot messages
948 by inserting a short delay after each one. The delay is
949 specified in milliseconds on the kernel command line,
950 using "boot_delay=N".
952 It is likely that you would also need to use "lpj=M" to preset
953 the "loops per jiffie" value.
954 See a previous boot log for the "lpj" value to use for your
955 system, and then set "lpj=M" before setting "boot_delay=N".
956 NOTE: Using this option may adversely affect SMP systems.
957 I.e., processors other than the first one may not boot up.
958 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
959 what it believes to be lockup conditions.
964 bool "RCU debugging: prove RCU correctness"
965 depends on PROVE_LOCKING
968 This feature enables lockdep extensions that check for correct
969 use of RCU APIs. This is currently under development. Say Y
970 if you want to debug RCU usage or help work on the PROVE_RCU
973 Say N if you are unsure.
975 config PROVE_RCU_REPEATEDLY
976 bool "RCU debugging: don't disable PROVE_RCU on first splat"
980 By itself, PROVE_RCU will disable checking upon issuing the
981 first warning (or "splat"). This feature prevents such
982 disabling, allowing multiple RCU-lockdep warnings to be printed
985 Say Y to allow multiple RCU-lockdep warnings per boot.
987 Say N if you are unsure.
989 config PROVE_RCU_DELAY
990 bool "RCU debugging: preemptible RCU race provocation"
991 depends on DEBUG_KERNEL && PREEMPT_RCU
994 There is a class of races that involve an unlikely preemption
995 of __rcu_read_unlock() just after ->rcu_read_lock_nesting has
996 been set to INT_MIN. This feature inserts a delay at that
997 point to increase the probability of these races.
999 Say Y to increase probability of preemption of __rcu_read_unlock().
1001 Say N if you are unsure.
1003 config SPARSE_RCU_POINTER
1004 bool "RCU debugging: sparse-based checks for pointer usage"
1007 This feature enables the __rcu sparse annotation for
1008 RCU-protected pointers. This annotation will cause sparse
1009 to flag any non-RCU used of annotated pointers. This can be
1010 helpful when debugging RCU usage. Please note that this feature
1011 is not intended to enforce code cleanliness; it is instead merely
1014 Say Y to make sparse flag questionable use of RCU-protected pointers
1016 Say N if you are unsure.
1018 config RCU_TORTURE_TEST
1019 tristate "torture tests for RCU"
1020 depends on DEBUG_KERNEL
1023 This option provides a kernel module that runs torture tests
1024 on the RCU infrastructure. The kernel module may be built
1025 after the fact on the running kernel to be tested, if desired.
1027 Say Y here if you want RCU torture tests to be built into
1029 Say M if you want the RCU torture tests to build as a module.
1030 Say N if you are unsure.
1032 config RCU_TORTURE_TEST_RUNNABLE
1033 bool "torture tests for RCU runnable by default"
1034 depends on RCU_TORTURE_TEST = y
1037 This option provides a way to build the RCU torture tests
1038 directly into the kernel without them starting up at boot
1039 time. You can use /proc/sys/kernel/rcutorture_runnable
1040 to manually override this setting. This /proc file is
1041 available only when the RCU torture tests have been built
1044 Say Y here if you want the RCU torture tests to start during
1045 boot (you probably don't).
1046 Say N here if you want the RCU torture tests to start only
1047 after being manually enabled via /proc.
1049 config RCU_CPU_STALL_TIMEOUT
1050 int "RCU CPU stall timeout in seconds"
1051 depends on RCU_STALL_COMMON
1055 If a given RCU grace period extends more than the specified
1056 number of seconds, a CPU stall warning is printed. If the
1057 RCU grace period persists, additional CPU stall warnings are
1058 printed at more widely spaced intervals.
1060 config RCU_CPU_STALL_VERBOSE
1061 bool "Print additional per-task information for RCU_CPU_STALL_DETECTOR"
1062 depends on TREE_PREEMPT_RCU
1065 This option causes RCU to printk detailed per-task information
1066 for any tasks that are stalling the current RCU grace period.
1068 Say N if you are unsure.
1070 Say Y if you want to enable such checks.
1072 config RCU_CPU_STALL_INFO
1073 bool "Print additional diagnostics on RCU CPU stall"
1074 depends on (TREE_RCU || TREE_PREEMPT_RCU) && DEBUG_KERNEL
1077 For each stalled CPU that is aware of the current RCU grace
1078 period, print out additional per-CPU diagnostic information
1079 regarding scheduling-clock ticks, idle state, and,
1080 for RCU_FAST_NO_HZ kernels, idle-entry state.
1082 Say N if you are unsure.
1084 Say Y if you want to enable such diagnostics.
1087 bool "Enable tracing for RCU"
1088 depends on DEBUG_KERNEL
1091 This option provides tracing in RCU which presents stats
1092 in debugfs for debugging RCU implementation.
1094 Say Y here if you want to enable RCU tracing
1095 Say N if you are unsure.
1097 endmenu # "RCU Debugging"
1099 config DEBUG_BLOCK_EXT_DEVT
1100 bool "Force extended block device numbers and spread them"
1101 depends on DEBUG_KERNEL
1105 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1106 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1107 YOU ARE DOING. Distros, please enable this and fix whatever
1110 Conventionally, block device numbers are allocated from
1111 predetermined contiguous area. However, extended block area
1112 may introduce non-contiguous block device numbers. This
1113 option forces most block device numbers to be allocated from
1114 the extended space and spreads them to discover kernel or
1115 userland code paths which assume predetermined contiguous
1116 device number allocation.
1118 Note that turning on this debug option shuffles all the
1119 device numbers for all IDE and SCSI devices including libata
1120 ones, so root partition specified using device number
1121 directly (via rdev or root=MAJ:MIN) won't work anymore.
1122 Textual device names (root=/dev/sdXn) will continue to work.
1124 Say N if you are unsure.
1126 config NOTIFIER_ERROR_INJECTION
1127 tristate "Notifier error injection"
1128 depends on DEBUG_KERNEL
1131 This option provides the ability to inject artificial errors to
1132 specified notifier chain callbacks. It is useful to test the error
1133 handling of notifier call chain failures.
1137 config CPU_NOTIFIER_ERROR_INJECT
1138 tristate "CPU notifier error injection module"
1139 depends on HOTPLUG_CPU && NOTIFIER_ERROR_INJECTION
1141 This option provides a kernel module that can be used to test
1142 the error handling of the cpu notifiers by injecting artificial
1143 errors to CPU notifier chain callbacks. It is controlled through
1144 debugfs interface under /sys/kernel/debug/notifier-error-inject/cpu
1146 If the notifier call chain should be failed with some events
1147 notified, write the error code to "actions/<notifier event>/error".
1149 Example: Inject CPU offline error (-1 == -EPERM)
1151 # cd /sys/kernel/debug/notifier-error-inject/cpu
1152 # echo -1 > actions/CPU_DOWN_PREPARE/error
1153 # echo 0 > /sys/devices/system/cpu/cpu1/online
1154 bash: echo: write error: Operation not permitted
1156 To compile this code as a module, choose M here: the module will
1157 be called cpu-notifier-error-inject.
1161 config PM_NOTIFIER_ERROR_INJECT
1162 tristate "PM notifier error injection module"
1163 depends on PM && NOTIFIER_ERROR_INJECTION
1164 default m if PM_DEBUG
1166 This option provides the ability to inject artificial errors to
1167 PM notifier chain callbacks. It is controlled through debugfs
1168 interface /sys/kernel/debug/notifier-error-inject/pm
1170 If the notifier call chain should be failed with some events
1171 notified, write the error code to "actions/<notifier event>/error".
1173 Example: Inject PM suspend error (-12 = -ENOMEM)
1175 # cd /sys/kernel/debug/notifier-error-inject/pm/
1176 # echo -12 > actions/PM_SUSPEND_PREPARE/error
1177 # echo mem > /sys/power/state
1178 bash: echo: write error: Cannot allocate memory
1180 To compile this code as a module, choose M here: the module will
1181 be called pm-notifier-error-inject.
1185 config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1186 tristate "OF reconfig notifier error injection module"
1187 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1189 This option provides the ability to inject artificial errors to
1190 OF reconfig notifier chain callbacks. It is controlled
1191 through debugfs interface under
1192 /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1194 If the notifier call chain should be failed with some events
1195 notified, write the error code to "actions/<notifier event>/error".
1197 To compile this code as a module, choose M here: the module will
1198 be called of-reconfig-notifier-error-inject.
1202 config FAULT_INJECTION
1203 bool "Fault-injection framework"
1204 depends on DEBUG_KERNEL
1206 Provide fault-injection framework.
1207 For more details, see Documentation/fault-injection/.
1210 bool "Fault-injection capability for kmalloc"
1211 depends on FAULT_INJECTION
1212 depends on SLAB || SLUB
1214 Provide fault-injection capability for kmalloc.
1216 config FAIL_PAGE_ALLOC
1217 bool "Fault-injection capabilitiy for alloc_pages()"
1218 depends on FAULT_INJECTION
1220 Provide fault-injection capability for alloc_pages().
1222 config FAIL_MAKE_REQUEST
1223 bool "Fault-injection capability for disk IO"
1224 depends on FAULT_INJECTION && BLOCK
1226 Provide fault-injection capability for disk IO.
1228 config FAIL_IO_TIMEOUT
1229 bool "Fault-injection capability for faking disk interrupts"
1230 depends on FAULT_INJECTION && BLOCK
1232 Provide fault-injection capability on end IO handling. This
1233 will make the block layer "forget" an interrupt as configured,
1234 thus exercising the error handling.
1236 Only works with drivers that use the generic timeout handling,
1237 for others it wont do anything.
1239 config FAIL_MMC_REQUEST
1240 bool "Fault-injection capability for MMC IO"
1242 depends on FAULT_INJECTION && MMC
1244 Provide fault-injection capability for MMC IO.
1245 This will make the mmc core return data errors. This is
1246 useful to test the error handling in the mmc block device
1247 and to test how the mmc host driver handles retries from
1250 config FAULT_INJECTION_DEBUG_FS
1251 bool "Debugfs entries for fault-injection capabilities"
1252 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1254 Enable configuration of fault-injection capabilities via debugfs.
1256 config FAULT_INJECTION_STACKTRACE_FILTER
1257 bool "stacktrace filter for fault-injection capabilities"
1258 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1261 select FRAME_POINTER if !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND
1263 Provide stacktrace filter for fault-injection capabilities
1266 bool "Latency measuring infrastructure"
1267 depends on HAVE_LATENCYTOP_SUPPORT
1268 depends on DEBUG_KERNEL
1269 depends on STACKTRACE_SUPPORT
1271 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND
1278 Enable this option if you want to use the LatencyTOP tool
1279 to find out which userspace is blocking on what kernel operations.
1281 config ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1284 config DEBUG_STRICT_USER_COPY_CHECKS
1285 bool "Strict user copy size checks"
1286 depends on ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1287 depends on DEBUG_KERNEL && !TRACE_BRANCH_PROFILING
1289 Enabling this option turns a certain set of sanity checks for user
1290 copy operations into compile time failures.
1292 The copy_from_user() etc checks are there to help test if there
1293 are sufficient security checks on the length argument of
1294 the copy operation, by having gcc prove that the argument is
1299 source kernel/trace/Kconfig
1301 menu "Runtime Testing"
1304 tristate "Linux Kernel Dump Test Tool Module"
1309 This module enables testing of the different dumping mechanisms by
1310 inducing system failures at predefined crash points.
1311 If you don't need it: say N
1312 Choose M here to compile this code as a module. The module will be
1315 Documentation on how to use the module can be found in
1316 Documentation/fault-injection/provoke-crashes.txt
1318 config TEST_LIST_SORT
1319 bool "Linked list sorting test"
1320 depends on DEBUG_KERNEL
1322 Enable this to turn on 'list_sort()' function test. This test is
1323 executed only once during system boot, so affects only boot time.
1327 config KPROBES_SANITY_TEST
1328 bool "Kprobes sanity tests"
1329 depends on DEBUG_KERNEL
1333 This option provides for testing basic kprobes functionality on
1334 boot. A sample kprobe, jprobe and kretprobe are inserted and
1335 verified for functionality.
1337 Say N if you are unsure.
1339 config BACKTRACE_SELF_TEST
1340 tristate "Self test for the backtrace code"
1341 depends on DEBUG_KERNEL
1344 This option provides a kernel module that can be used to test
1345 the kernel stack backtrace code. This option is not useful
1346 for distributions or general kernels, but only for kernel
1347 developers working on architecture code.
1349 Note that if you want to also test saved backtraces, you will
1350 have to enable STACKTRACE as well.
1352 Say N if you are unsure.
1355 tristate "Red-Black tree test"
1356 depends on m && DEBUG_KERNEL
1358 A benchmark measuring the performance of the rbtree library.
1359 Also includes rbtree invariant checks.
1361 config INTERVAL_TREE_TEST
1362 tristate "Interval tree test"
1363 depends on m && DEBUG_KERNEL
1365 A benchmark measuring the performance of the interval tree library
1367 config ATOMIC64_SELFTEST
1368 bool "Perform an atomic64_t self-test at boot"
1370 Enable this option to test the atomic64_t functions at boot.
1374 config ASYNC_RAID6_TEST
1375 tristate "Self test for hardware accelerated raid6 recovery"
1376 depends on ASYNC_RAID6_RECOV
1379 This is a one-shot self test that permutes through the
1380 recovery of all the possible two disk failure scenarios for a
1381 N-disk array. Recovery is performed with the asynchronous
1382 raid6 recovery routines, and will optionally use an offload
1383 engine if one is available.
1387 config TEST_STRING_HELPERS
1388 tristate "Test functions located in the string_helpers module at runtime"
1391 tristate "Test kstrto*() family of functions at runtime"
1393 endmenu # runtime tests
1395 config PROVIDE_OHCI1394_DMA_INIT
1396 bool "Remote debugging over FireWire early on boot"
1397 depends on PCI && X86
1399 If you want to debug problems which hang or crash the kernel early
1400 on boot and the crashing machine has a FireWire port, you can use
1401 this feature to remotely access the memory of the crashed machine
1402 over FireWire. This employs remote DMA as part of the OHCI1394
1403 specification which is now the standard for FireWire controllers.
1405 With remote DMA, you can monitor the printk buffer remotely using
1406 firescope and access all memory below 4GB using fireproxy from gdb.
1407 Even controlling a kernel debugger is possible using remote DMA.
1411 If ohci1394_dma=early is used as boot parameter, it will initialize
1412 all OHCI1394 controllers which are found in the PCI config space.
1414 As all changes to the FireWire bus such as enabling and disabling
1415 devices cause a bus reset and thereby disable remote DMA for all
1416 devices, be sure to have the cable plugged and FireWire enabled on
1417 the debugging host before booting the debug target for debugging.
1419 This code (~1k) is freed after boot. By then, the firewire stack
1420 in charge of the OHCI-1394 controllers should be used instead.
1422 See Documentation/debugging-via-ohci1394.txt for more information.
1424 config FIREWIRE_OHCI_REMOTE_DMA
1425 bool "Remote debugging over FireWire with firewire-ohci"
1426 depends on FIREWIRE_OHCI
1428 This option lets you use the FireWire bus for remote debugging
1429 with help of the firewire-ohci driver. It enables unfiltered
1430 remote DMA in firewire-ohci.
1431 See Documentation/debugging-via-ohci1394.txt for more information.
1436 bool "Build targets in Documentation/ tree"
1437 depends on HEADERS_CHECK
1439 This option attempts to build objects from the source files in the
1440 kernel Documentation/ tree.
1442 Say N if you are unsure.
1444 config DYNAMIC_DEBUG
1445 bool "Enable dynamic printk() support"
1451 Compiles debug level messages into the kernel, which would not
1452 otherwise be available at runtime. These messages can then be
1453 enabled/disabled based on various levels of scope - per source file,
1454 function, module, format string, and line number. This mechanism
1455 implicitly compiles in all pr_debug() and dev_dbg() calls, which
1456 enlarges the kernel text size by about 2%.
1458 If a source file is compiled with DEBUG flag set, any
1459 pr_debug() calls in it are enabled by default, but can be
1460 disabled at runtime as below. Note that DEBUG flag is
1461 turned on by many CONFIG_*DEBUG* options.
1465 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
1466 which is contained in the 'debugfs' filesystem. Thus, the debugfs
1467 filesystem must first be mounted before making use of this feature.
1468 We refer the control file as: <debugfs>/dynamic_debug/control. This
1469 file contains a list of the debug statements that can be enabled. The
1470 format for each line of the file is:
1472 filename:lineno [module]function flags format
1474 filename : source file of the debug statement
1475 lineno : line number of the debug statement
1476 module : module that contains the debug statement
1477 function : function that contains the debug statement
1478 flags : '=p' means the line is turned 'on' for printing
1479 format : the format used for the debug statement
1483 nullarbor:~ # cat <debugfs>/dynamic_debug/control
1484 # filename:lineno [module]function flags format
1485 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
1486 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
1487 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
1491 // enable the message at line 1603 of file svcsock.c
1492 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
1493 <debugfs>/dynamic_debug/control
1495 // enable all the messages in file svcsock.c
1496 nullarbor:~ # echo -n 'file svcsock.c +p' >
1497 <debugfs>/dynamic_debug/control
1499 // enable all the messages in the NFS server module
1500 nullarbor:~ # echo -n 'module nfsd +p' >
1501 <debugfs>/dynamic_debug/control
1503 // enable all 12 messages in the function svc_process()
1504 nullarbor:~ # echo -n 'func svc_process +p' >
1505 <debugfs>/dynamic_debug/control
1507 // disable all 12 messages in the function svc_process()
1508 nullarbor:~ # echo -n 'func svc_process -p' >
1509 <debugfs>/dynamic_debug/control
1511 See Documentation/dynamic-debug-howto.txt for additional information.
1513 config DMA_API_DEBUG
1514 bool "Enable debugging of DMA-API usage"
1515 depends on HAVE_DMA_API_DEBUG
1517 Enable this option to debug the use of the DMA API by device drivers.
1518 With this option you will be able to detect common bugs in device
1519 drivers like double-freeing of DMA mappings or freeing mappings that
1520 were never allocated.
1521 This option causes a performance degredation. Use only if you want
1522 to debug device drivers. If unsure, say N.
1524 source "samples/Kconfig"
1526 source "lib/Kconfig.kgdb"