1 menu "printk and dmesg options"
4 bool "Show timing information on printks"
7 Selecting this option causes time stamps of the printk()
8 messages to be added to the output of the syslog() system
9 call and at the console.
11 The timestamp is always recorded internally, and exported
12 to /dev/kmsg. This flag just specifies if the timestamp should
13 be included, not that the timestamp is recorded.
15 The behavior is also controlled by the kernel command line
16 parameter printk.time=1. See Documentation/kernel-parameters.txt
19 bool "Show process information on printks"
22 Selecting this option causes process to be
23 included in printk output. Or add printk.process=1 at boot-time.
25 config MESSAGE_LOGLEVEL_DEFAULT
26 int "Default message log level (1-7)"
30 Default log level for printk statements with no specified priority.
32 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
33 that are auditing their logs closely may want to set it to a lower
36 config BOOT_PRINTK_DELAY
37 bool "Delay each boot printk message by N milliseconds"
38 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
40 This build option allows you to read kernel boot messages
41 by inserting a short delay after each one. The delay is
42 specified in milliseconds on the kernel command line,
45 It is likely that you would also need to use "lpj=M" to preset
46 the "loops per jiffie" value.
47 See a previous boot log for the "lpj" value to use for your
48 system, and then set "lpj=M" before setting "boot_delay=N".
49 NOTE: Using this option may adversely affect SMP systems.
50 I.e., processors other than the first one may not boot up.
51 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
52 what it believes to be lockup conditions.
55 bool "Enable dynamic printk() support"
61 Compiles debug level messages into the kernel, which would not
62 otherwise be available at runtime. These messages can then be
63 enabled/disabled based on various levels of scope - per source file,
64 function, module, format string, and line number. This mechanism
65 implicitly compiles in all pr_debug() and dev_dbg() calls, which
66 enlarges the kernel text size by about 2%.
68 If a source file is compiled with DEBUG flag set, any
69 pr_debug() calls in it are enabled by default, but can be
70 disabled at runtime as below. Note that DEBUG flag is
71 turned on by many CONFIG_*DEBUG* options.
75 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
76 which is contained in the 'debugfs' filesystem. Thus, the debugfs
77 filesystem must first be mounted before making use of this feature.
78 We refer the control file as: <debugfs>/dynamic_debug/control. This
79 file contains a list of the debug statements that can be enabled. The
80 format for each line of the file is:
82 filename:lineno [module]function flags format
84 filename : source file of the debug statement
85 lineno : line number of the debug statement
86 module : module that contains the debug statement
87 function : function that contains the debug statement
88 flags : '=p' means the line is turned 'on' for printing
89 format : the format used for the debug statement
93 nullarbor:~ # cat <debugfs>/dynamic_debug/control
94 # filename:lineno [module]function flags format
95 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
96 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
97 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
101 // enable the message at line 1603 of file svcsock.c
102 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
103 <debugfs>/dynamic_debug/control
105 // enable all the messages in file svcsock.c
106 nullarbor:~ # echo -n 'file svcsock.c +p' >
107 <debugfs>/dynamic_debug/control
109 // enable all the messages in the NFS server module
110 nullarbor:~ # echo -n 'module nfsd +p' >
111 <debugfs>/dynamic_debug/control
113 // enable all 12 messages in the function svc_process()
114 nullarbor:~ # echo -n 'func svc_process +p' >
115 <debugfs>/dynamic_debug/control
117 // disable all 12 messages in the function svc_process()
118 nullarbor:~ # echo -n 'func svc_process -p' >
119 <debugfs>/dynamic_debug/control
121 See Documentation/dynamic-debug-howto.txt for additional information.
123 endmenu # "printk and dmesg options"
125 menu "Compile-time checks and compiler options"
128 bool "Compile the kernel with debug info"
129 depends on DEBUG_KERNEL && !COMPILE_TEST
131 If you say Y here the resulting kernel image will include
132 debugging info resulting in a larger kernel image.
133 This adds debug symbols to the kernel and modules (gcc -g), and
134 is needed if you intend to use kernel crashdump or binary object
135 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
136 Say Y here only if you plan to debug the kernel.
140 config DEBUG_INFO_REDUCED
141 bool "Reduce debugging information"
142 depends on DEBUG_INFO
144 If you say Y here gcc is instructed to generate less debugging
145 information for structure types. This means that tools that
146 need full debugging information (like kgdb or systemtap) won't
147 be happy. But if you merely need debugging information to
148 resolve line numbers there is no loss. Advantage is that
149 build directory object sizes shrink dramatically over a full
150 DEBUG_INFO build and compile times are reduced too.
151 Only works with newer gcc versions.
153 config DEBUG_INFO_SPLIT
154 bool "Produce split debuginfo in .dwo files"
155 depends on DEBUG_INFO && !FRV
157 Generate debug info into separate .dwo files. This significantly
158 reduces the build directory size for builds with DEBUG_INFO,
159 because it stores the information only once on disk in .dwo
160 files instead of multiple times in object files and executables.
161 In addition the debug information is also compressed.
163 Requires recent gcc (4.7+) and recent gdb/binutils.
164 Any tool that packages or reads debug information would need
165 to know about the .dwo files and include them.
166 Incompatible with older versions of ccache.
168 config DEBUG_INFO_DWARF4
169 bool "Generate dwarf4 debuginfo"
170 depends on DEBUG_INFO
172 Generate dwarf4 debug info. This requires recent versions
173 of gcc and gdb. It makes the debug information larger.
174 But it significantly improves the success of resolving
175 variables in gdb on optimized code.
178 bool "Provide GDB scripts for kernel debugging"
179 depends on DEBUG_INFO
181 This creates the required links to GDB helper scripts in the
182 build directory. If you load vmlinux into gdb, the helper
183 scripts will be automatically imported by gdb as well, and
184 additional functions are available to analyze a Linux kernel
185 instance. See Documentation/gdb-kernel-debugging.txt for further
188 config ENABLE_WARN_DEPRECATED
189 bool "Enable __deprecated logic"
192 Enable the __deprecated logic in the kernel build.
193 Disable this to suppress the "warning: 'foo' is deprecated
194 (declared at kernel/power/somefile.c:1234)" messages.
196 config ENABLE_MUST_CHECK
197 bool "Enable __must_check logic"
200 Enable the __must_check logic in the kernel build. Disable this to
201 suppress the "warning: ignoring return value of 'foo', declared with
202 attribute warn_unused_result" messages.
205 int "Warn for stack frames larger than (needs gcc 4.4)"
208 default 1024 if !64BIT
209 default 2048 if 64BIT
211 Tell gcc to warn at build time for stack frames larger than this.
212 Setting this too low will cause a lot of warnings.
213 Setting it to 0 disables the warning.
216 config STRIP_ASM_SYMS
217 bool "Strip assembler-generated symbols during link"
220 Strip internal assembler-generated symbols during a link (symbols
221 that look like '.Lxxx') so they don't pollute the output of
222 get_wchan() and suchlike.
225 bool "Generate readable assembler code"
226 depends on DEBUG_KERNEL
228 Disable some compiler optimizations that tend to generate human unreadable
229 assembler output. This may make the kernel slightly slower, but it helps
230 to keep kernel developers who have to stare a lot at assembler listings
233 config UNUSED_SYMBOLS
234 bool "Enable unused/obsolete exported symbols"
237 Unused but exported symbols make the kernel needlessly bigger. For
238 that reason most of these unused exports will soon be removed. This
239 option is provided temporarily to provide a transition period in case
240 some external kernel module needs one of these symbols anyway. If you
241 encounter such a case in your module, consider if you are actually
242 using the right API. (rationale: since nobody in the kernel is using
243 this in a module, there is a pretty good chance it's actually the
244 wrong interface to use). If you really need the symbol, please send a
245 mail to the linux kernel mailing list mentioning the symbol and why
246 you really need it, and what the merge plan to the mainline kernel for
250 bool "Track page owner"
251 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT
254 select PAGE_EXTENSION
256 This keeps track of what call chain is the owner of a page, may
257 help to find bare alloc_page(s) leaks. Even if you include this
258 feature on your build, it is disabled in default. You should pass
259 "page_owner=on" to boot parameter in order to enable it. Eats
260 a fair amount of memory if enabled. See tools/vm/page_owner_sort.c
261 for user-space helper.
266 bool "Debug Filesystem"
268 debugfs is a virtual file system that kernel developers use to put
269 debugging files into. Enable this option to be able to read and
270 write to these files.
272 For detailed documentation on the debugfs API, see
273 Documentation/DocBook/filesystems.
278 bool "Run 'make headers_check' when building vmlinux"
281 This option will extract the user-visible kernel headers whenever
282 building the kernel, and will run basic sanity checks on them to
283 ensure that exported files do not attempt to include files which
284 were not exported, etc.
286 If you're making modifications to header files which are
287 relevant for userspace, say 'Y', and check the headers
288 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
289 your build tree), to make sure they're suitable.
291 config DEBUG_SECTION_MISMATCH
292 bool "Enable full Section mismatch analysis"
294 The section mismatch analysis checks if there are illegal
295 references from one section to another section.
296 During linktime or runtime, some sections are dropped;
297 any use of code/data previously in these sections would
298 most likely result in an oops.
299 In the code, functions and variables are annotated with
300 __init,, etc. (see the full list in include/linux/init.h),
301 which results in the code/data being placed in specific sections.
302 The section mismatch analysis is always performed after a full
303 kernel build, and enabling this option causes the following
304 additional steps to occur:
305 - Add the option -fno-inline-functions-called-once to gcc commands.
306 When inlining a function annotated with __init in a non-init
307 function, we would lose the section information and thus
308 the analysis would not catch the illegal reference.
309 This option tells gcc to inline less (but it does result in
311 - Run the section mismatch analysis for each module/built-in.o file.
312 When we run the section mismatch analysis on vmlinux.o, we
313 lose valueble information about where the mismatch was
315 Running the analysis for each module/built-in.o file
316 tells where the mismatch happens much closer to the
317 source. The drawback is that the same mismatch is
318 reported at least twice.
319 - Enable verbose reporting from modpost in order to help resolve
320 the section mismatches that are reported.
322 config SECTION_MISMATCH_WARN_ONLY
323 bool "Make section mismatch errors non-fatal"
326 If you say N here, the build process will fail if there are any
327 section mismatch, instead of just throwing warnings.
332 # Select this config option from the architecture Kconfig, if it
333 # is preferred to always offer frame pointers as a config
334 # option on the architecture (regardless of KERNEL_DEBUG):
336 config ARCH_WANT_FRAME_POINTERS
341 bool "Compile the kernel with frame pointers"
342 depends on DEBUG_KERNEL && \
343 (CRIS || M68K || FRV || UML || \
344 AVR32 || SUPERH || BLACKFIN || MN10300 || METAG) || \
345 ARCH_WANT_FRAME_POINTERS
346 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
348 If you say Y here the resulting kernel image will be slightly
349 larger and slower, but it gives very useful debugging information
350 in case of kernel bugs. (precise oopses/stacktraces/warnings)
352 config DEBUG_FORCE_WEAK_PER_CPU
353 bool "Force weak per-cpu definitions"
354 depends on DEBUG_KERNEL
356 s390 and alpha require percpu variables in modules to be
357 defined weak to work around addressing range issue which
358 puts the following two restrictions on percpu variable
361 1. percpu symbols must be unique whether static or not
362 2. percpu variables can't be defined inside a function
364 To ensure that generic code follows the above rules, this
365 option forces all percpu variables to be defined as weak.
367 endmenu # "Compiler options"
370 bool "Magic SysRq key"
373 If you say Y here, you will have some control over the system even
374 if the system crashes for example during kernel debugging (e.g., you
375 will be able to flush the buffer cache to disk, reboot the system
376 immediately or dump some status information). This is accomplished
377 by pressing various keys while holding SysRq (Alt+PrintScreen). It
378 also works on a serial console (on PC hardware at least), if you
379 send a BREAK and then within 5 seconds a command keypress. The
380 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
381 unless you really know what this hack does.
383 config MAGIC_SYSRQ_DEFAULT_ENABLE
384 hex "Enable magic SysRq key functions by default"
385 depends on MAGIC_SYSRQ
388 Specifies which SysRq key functions are enabled by default.
389 This may be set to 1 or 0 to enable or disable them all, or
390 to a bitmask as described in Documentation/sysrq.txt.
393 bool "Kernel debugging"
395 Say Y here if you are developing drivers or trying to debug and
396 identify kernel problems.
398 menu "Memory Debugging"
400 source mm/Kconfig.debug
403 bool "Debug object operations"
404 depends on DEBUG_KERNEL
406 If you say Y here, additional code will be inserted into the
407 kernel to track the life time of various objects and validate
408 the operations on those objects.
410 config DEBUG_OBJECTS_SELFTEST
411 bool "Debug objects selftest"
412 depends on DEBUG_OBJECTS
414 This enables the selftest of the object debug code.
416 config DEBUG_OBJECTS_FREE
417 bool "Debug objects in freed memory"
418 depends on DEBUG_OBJECTS
420 This enables checks whether a k/v free operation frees an area
421 which contains an object which has not been deactivated
422 properly. This can make kmalloc/kfree-intensive workloads
425 config DEBUG_OBJECTS_TIMERS
426 bool "Debug timer objects"
427 depends on DEBUG_OBJECTS
429 If you say Y here, additional code will be inserted into the
430 timer routines to track the life time of timer objects and
431 validate the timer operations.
433 config DEBUG_OBJECTS_WORK
434 bool "Debug work objects"
435 depends on DEBUG_OBJECTS
437 If you say Y here, additional code will be inserted into the
438 work queue routines to track the life time of work objects and
439 validate the work operations.
441 config DEBUG_OBJECTS_RCU_HEAD
442 bool "Debug RCU callbacks objects"
443 depends on DEBUG_OBJECTS
445 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
447 config DEBUG_OBJECTS_PERCPU_COUNTER
448 bool "Debug percpu counter objects"
449 depends on DEBUG_OBJECTS
451 If you say Y here, additional code will be inserted into the
452 percpu counter routines to track the life time of percpu counter
453 objects and validate the percpu counter operations.
455 config DEBUG_OBJECTS_ENABLE_DEFAULT
456 int "debug_objects bootup default value (0-1)"
459 depends on DEBUG_OBJECTS
461 Debug objects boot parameter default value
464 bool "Debug slab memory allocations"
465 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
467 Say Y here to have the kernel do limited verification on memory
468 allocation as well as poisoning memory on free to catch use of freed
469 memory. This can make kmalloc/kfree-intensive workloads much slower.
471 config DEBUG_SLAB_LEAK
472 bool "Memory leak debugging"
473 depends on DEBUG_SLAB
476 bool "SLUB debugging on by default"
477 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
480 Boot with debugging on by default. SLUB boots by default with
481 the runtime debug capabilities switched off. Enabling this is
482 equivalent to specifying the "slub_debug" parameter on boot.
483 There is no support for more fine grained debug control like
484 possible with slub_debug=xxx. SLUB debugging may be switched
485 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
490 bool "Enable SLUB performance statistics"
491 depends on SLUB && SYSFS
493 SLUB statistics are useful to debug SLUBs allocation behavior in
494 order find ways to optimize the allocator. This should never be
495 enabled for production use since keeping statistics slows down
496 the allocator by a few percentage points. The slabinfo command
497 supports the determination of the most active slabs to figure
498 out which slabs are relevant to a particular load.
499 Try running: slabinfo -DA
501 config HAVE_DEBUG_KMEMLEAK
504 config DEBUG_KMEMLEAK
505 bool "Kernel memory leak detector"
506 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
508 select STACKTRACE if STACKTRACE_SUPPORT
512 Say Y here if you want to enable the memory leak
513 detector. The memory allocation/freeing is traced in a way
514 similar to the Boehm's conservative garbage collector, the
515 difference being that the orphan objects are not freed but
516 only shown in /sys/kernel/debug/kmemleak. Enabling this
517 feature will introduce an overhead to memory
518 allocations. See Documentation/kmemleak.txt for more
521 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
522 of finding leaks due to the slab objects poisoning.
524 In order to access the kmemleak file, debugfs needs to be
525 mounted (usually at /sys/kernel/debug).
527 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
528 int "Maximum kmemleak early log entries"
529 depends on DEBUG_KMEMLEAK
533 Kmemleak must track all the memory allocations to avoid
534 reporting false positives. Since memory may be allocated or
535 freed before kmemleak is initialised, an early log buffer is
536 used to store these actions. If kmemleak reports "early log
537 buffer exceeded", please increase this value.
539 config DEBUG_KMEMLEAK_TEST
540 tristate "Simple test for the kernel memory leak detector"
541 depends on DEBUG_KMEMLEAK && m
543 This option enables a module that explicitly leaks memory.
547 config DEBUG_KMEMLEAK_DEFAULT_OFF
548 bool "Default kmemleak to off"
549 depends on DEBUG_KMEMLEAK
551 Say Y here to disable kmemleak by default. It can then be enabled
552 on the command line via kmemleak=on.
554 config DEBUG_STACK_USAGE
555 bool "Stack utilization instrumentation"
556 depends on DEBUG_KERNEL && !IA64 && !PARISC && !METAG
558 Enables the display of the minimum amount of free stack which each
559 task has ever had available in the sysrq-T and sysrq-P debug output.
561 This option will slow down process creation somewhat.
565 depends on DEBUG_KERNEL
567 Enable this to turn on extended checks in the virtual-memory system
568 that may impact performance.
572 config DEBUG_VM_VMACACHE
573 bool "Debug VMA caching"
576 Enable this to turn on VMA caching debug information. Doing so
577 can cause significant overhead, so only enable it in non-production
583 bool "Debug VM red-black trees"
586 Enable VM red-black tree debugging information and extra validations.
591 bool "Debug VM translations"
592 depends on DEBUG_KERNEL && X86
594 Enable some costly sanity checks in virtual to page code. This can
595 catch mistakes with virt_to_page() and friends.
599 config DEBUG_NOMMU_REGIONS
600 bool "Debug the global anon/private NOMMU mapping region tree"
601 depends on DEBUG_KERNEL && !MMU
603 This option causes the global tree of anonymous and private mapping
604 regions to be regularly checked for invalid topology.
606 config DEBUG_MEMORY_INIT
607 bool "Debug memory initialisation" if EXPERT
610 Enable this for additional checks during memory initialisation.
611 The sanity checks verify aspects of the VM such as the memory model
612 and other information provided by the architecture. Verbose
613 information will be printed at KERN_DEBUG loglevel depending
614 on the mminit_loglevel= command-line option.
618 config MEMORY_NOTIFIER_ERROR_INJECT
619 tristate "Memory hotplug notifier error injection module"
620 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
622 This option provides the ability to inject artificial errors to
623 memory hotplug notifier chain callbacks. It is controlled through
624 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
626 If the notifier call chain should be failed with some events
627 notified, write the error code to "actions/<notifier event>/error".
629 Example: Inject memory hotplug offline error (-12 == -ENOMEM)
631 # cd /sys/kernel/debug/notifier-error-inject/memory
632 # echo -12 > actions/MEM_GOING_OFFLINE/error
633 # echo offline > /sys/devices/system/memory/memoryXXX/state
634 bash: echo: write error: Cannot allocate memory
636 To compile this code as a module, choose M here: the module will
637 be called memory-notifier-error-inject.
641 config DEBUG_PER_CPU_MAPS
642 bool "Debug access to per_cpu maps"
643 depends on DEBUG_KERNEL
646 Say Y to verify that the per_cpu map being accessed has
647 been set up. This adds a fair amount of code to kernel memory
648 and decreases performance.
653 bool "Highmem debugging"
654 depends on DEBUG_KERNEL && HIGHMEM
656 This option enables additional error checking for high memory
657 systems. Disable for production systems.
659 config HAVE_DEBUG_STACKOVERFLOW
662 config DEBUG_STACKOVERFLOW
663 bool "Check for stack overflows"
664 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
666 Say Y here if you want to check for overflows of kernel, IRQ
667 and exception stacks (if your architecture uses them). This
668 option will show detailed messages if free stack space drops
669 below a certain limit.
671 These kinds of bugs usually occur when call-chains in the
672 kernel get too deep, especially when interrupts are
675 Use this in cases where you see apparently random memory
676 corruption, especially if it appears in 'struct thread_info'
678 If in doubt, say "N".
680 source "lib/Kconfig.kmemcheck"
682 source "lib/Kconfig.kasan"
684 endmenu # "Memory Debugging"
689 KCOV does not have any arch-specific code, but currently it is enabled
690 only for x86_64. KCOV requires testing on other archs, and most likely
691 disabling of instrumentation for some early boot code.
694 bool "Code coverage for fuzzing"
695 depends on ARCH_HAS_KCOV
698 KCOV exposes kernel code coverage information in a form suitable
699 for coverage-guided fuzzing (randomized testing).
701 If RANDOMIZE_BASE is enabled, PC values will not be stable across
702 different machines and across reboots. If you need stable PC values,
703 disable RANDOMIZE_BASE.
705 For more details, see Documentation/kcov.txt.
708 bool "Debug shared IRQ handlers"
709 depends on DEBUG_KERNEL
711 Enable this to generate a spurious interrupt as soon as a shared
712 interrupt handler is registered, and just before one is deregistered.
713 Drivers ought to be able to handle interrupts coming in at those
714 points; some don't and need to be caught.
716 menu "Debug Lockups and Hangs"
718 config LOCKUP_DETECTOR
719 bool "Detect Hard and Soft Lockups"
720 depends on DEBUG_KERNEL && !S390
722 Say Y here to enable the kernel to act as a watchdog to detect
723 hard and soft lockups.
725 Softlockups are bugs that cause the kernel to loop in kernel
726 mode for more than 20 seconds, without giving other tasks a
727 chance to run. The current stack trace is displayed upon
728 detection and the system will stay locked up.
730 Hardlockups are bugs that cause the CPU to loop in kernel mode
731 for more than 10 seconds, without letting other interrupts have a
732 chance to run. The current stack trace is displayed upon detection
733 and the system will stay locked up.
735 The overhead should be minimal. A periodic hrtimer runs to
736 generate interrupts and kick the watchdog task every 4 seconds.
737 An NMI is generated every 10 seconds or so to check for hardlockups.
738 If NMIs are not available on the platform, every 12 seconds the
739 hrtimer interrupt on one cpu will be used to check for hardlockups
742 The frequency of hrtimer and NMI events and the soft and hard lockup
743 thresholds can be controlled through the sysctl watchdog_thresh.
745 config HARDLOCKUP_DETECTOR_NMI
747 depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
748 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
750 config HARDLOCKUP_DETECTOR_OTHER_CPU
752 depends on LOCKUP_DETECTOR && SMP
753 depends on !HARDLOCKUP_DETECTOR_NMI && !HAVE_NMI_WATCHDOG
755 config HARDLOCKUP_DETECTOR
757 depends on HARDLOCKUP_DETECTOR_NMI || HARDLOCKUP_DETECTOR_OTHER_CPU
759 config BOOTPARAM_HARDLOCKUP_PANIC
760 bool "Panic (Reboot) On Hard Lockups"
761 depends on HARDLOCKUP_DETECTOR
763 Say Y here to enable the kernel to panic on "hard lockups",
764 which are bugs that cause the kernel to loop in kernel
765 mode with interrupts disabled for more than 10 seconds (configurable
766 using the watchdog_thresh sysctl).
770 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
772 depends on HARDLOCKUP_DETECTOR
774 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
775 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
777 config BOOTPARAM_SOFTLOCKUP_PANIC
778 bool "Panic (Reboot) On Soft Lockups"
779 depends on LOCKUP_DETECTOR
781 Say Y here to enable the kernel to panic on "soft lockups",
782 which are bugs that cause the kernel to loop in kernel
783 mode for more than 20 seconds (configurable using the watchdog_thresh
784 sysctl), without giving other tasks a chance to run.
786 The panic can be used in combination with panic_timeout,
787 to cause the system to reboot automatically after a
788 lockup has been detected. This feature is useful for
789 high-availability systems that have uptime guarantees and
790 where a lockup must be resolved ASAP.
794 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
796 depends on LOCKUP_DETECTOR
798 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
799 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
801 config DETECT_HUNG_TASK
802 bool "Detect Hung Tasks"
803 depends on DEBUG_KERNEL
804 default LOCKUP_DETECTOR
806 Say Y here to enable the kernel to detect "hung tasks",
807 which are bugs that cause the task to be stuck in
808 uninterruptible "D" state indefinitiley.
810 When a hung task is detected, the kernel will print the
811 current stack trace (which you should report), but the
812 task will stay in uninterruptible state. If lockdep is
813 enabled then all held locks will also be reported. This
814 feature has negligible overhead.
816 config DEFAULT_HUNG_TASK_TIMEOUT
817 int "Default timeout for hung task detection (in seconds)"
818 depends on DETECT_HUNG_TASK
821 This option controls the default timeout (in seconds) used
822 to determine when a task has become non-responsive and should
825 It can be adjusted at runtime via the kernel.hung_task_timeout_secs
826 sysctl or by writing a value to
827 /proc/sys/kernel/hung_task_timeout_secs.
829 A timeout of 0 disables the check. The default is two minutes.
830 Keeping the default should be fine in most cases.
832 config BOOTPARAM_HUNG_TASK_PANIC
833 bool "Panic (Reboot) On Hung Tasks"
834 depends on DETECT_HUNG_TASK
836 Say Y here to enable the kernel to panic on "hung tasks",
837 which are bugs that cause the kernel to leave a task stuck
838 in uninterruptible "D" state.
840 The panic can be used in combination with panic_timeout,
841 to cause the system to reboot automatically after a
842 hung task has been detected. This feature is useful for
843 high-availability systems that have uptime guarantees and
844 where a hung tasks must be resolved ASAP.
848 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
850 depends on DETECT_HUNG_TASK
852 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
853 default 1 if BOOTPARAM_HUNG_TASK_PANIC
855 endmenu # "Debug lockups and hangs"
860 Say Y here to enable the kernel to panic when it oopses. This
861 has the same effect as setting oops=panic on the kernel command
864 This feature is useful to ensure that the kernel does not do
865 anything erroneous after an oops which could result in data
866 corruption or other issues.
870 config PANIC_ON_OOPS_VALUE
873 default 0 if !PANIC_ON_OOPS
874 default 1 if PANIC_ON_OOPS
880 Set the timeout value (in seconds) until a reboot occurs when the
881 the kernel panics. If n = 0, then we wait forever. A timeout
882 value n > 0 will wait n seconds before rebooting, while a timeout
883 value n < 0 will reboot immediately.
886 bool "Collect scheduler debugging info"
887 depends on DEBUG_KERNEL && PROC_FS
890 If you say Y here, the /proc/sched_debug file will be provided
891 that can help debug the scheduler. The runtime overhead of this
898 config PANIC_ON_RT_THROTTLING
899 bool "Panic on RT throttling"
901 Say Y here to enable the kernel to panic when a realtime
902 runqueue is throttled. This may be useful for detecting
903 and debugging RT throttling issues.
908 bool "Collect scheduler statistics"
909 depends on DEBUG_KERNEL && PROC_FS
912 If you say Y here, additional code will be inserted into the
913 scheduler and related routines to collect statistics about
914 scheduler behavior and provide them in /proc/schedstat. These
915 stats may be useful for both tuning and debugging the scheduler
916 If you aren't debugging the scheduler or trying to tune a specific
917 application, you can say N to avoid the very slight overhead
920 config SCHED_STACK_END_CHECK
921 bool "Detect stack corruption on calls to schedule()"
922 depends on DEBUG_KERNEL
925 This option checks for a stack overrun on calls to schedule().
926 If the stack end location is found to be over written always panic as
927 the content of the corrupted region can no longer be trusted.
928 This is to ensure no erroneous behaviour occurs which could result in
929 data corruption or a sporadic crash at a later stage once the region
930 is examined. The runtime overhead introduced is minimal.
932 config DEBUG_TIMEKEEPING
933 bool "Enable extra timekeeping sanity checking"
935 This option will enable additional timekeeping sanity checks
936 which may be helpful when diagnosing issues where timekeeping
937 problems are suspected.
939 This may include checks in the timekeeping hotpaths, so this
940 option may have a (very small) performance impact to some
946 bool "Collect kernel timers statistics"
947 depends on DEBUG_KERNEL && PROC_FS
949 If you say Y here, additional code will be inserted into the
950 timer routines to collect statistics about kernel timers being
951 reprogrammed. The statistics can be read from /proc/timer_stats.
952 The statistics collection is started by writing 1 to /proc/timer_stats,
953 writing 0 stops it. This feature is useful to collect information
954 about timer usage patterns in kernel and userspace. This feature
955 is lightweight if enabled in the kernel config but not activated
956 (it defaults to deactivated on bootup and will only be activated
957 if some application like powertop activates it explicitly).
960 bool "Debug preemptible kernel"
961 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
964 If you say Y here then the kernel will use a debug variant of the
965 commonly used smp_processor_id() function and will print warnings
966 if kernel code uses it in a preemption-unsafe way. Also, the kernel
967 will detect preemption count underflows.
969 menu "Lock Debugging (spinlocks, mutexes, etc...)"
971 config DEBUG_RT_MUTEXES
972 bool "RT Mutex debugging, deadlock detection"
973 depends on DEBUG_KERNEL && RT_MUTEXES
975 This allows rt mutex semantics violations and rt mutex related
976 deadlocks (lockups) to be detected and reported automatically.
978 config DEBUG_SPINLOCK
979 bool "Spinlock and rw-lock debugging: basic checks"
980 depends on DEBUG_KERNEL
981 select UNINLINE_SPIN_UNLOCK
983 Say Y here and build SMP to catch missing spinlock initialization
984 and certain other kinds of spinlock errors commonly made. This is
985 best used in conjunction with the NMI watchdog so that spinlock
986 deadlocks are also debuggable.
989 bool "Mutex debugging: basic checks"
990 depends on DEBUG_KERNEL
992 This feature allows mutex semantics violations to be detected and
995 config DEBUG_WW_MUTEX_SLOWPATH
996 bool "Wait/wound mutex debugging: Slowpath testing"
997 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
998 select DEBUG_LOCK_ALLOC
999 select DEBUG_SPINLOCK
1000 select DEBUG_MUTEXES
1002 This feature enables slowpath testing for w/w mutex users by
1003 injecting additional -EDEADLK wound/backoff cases. Together with
1004 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this
1005 will test all possible w/w mutex interface abuse with the
1006 exception of simply not acquiring all the required locks.
1007 Note that this feature can introduce significant overhead, so
1008 it really should not be enabled in a production or distro kernel,
1009 even a debug kernel. If you are a driver writer, enable it. If
1010 you are a distro, do not.
1012 config DEBUG_LOCK_ALLOC
1013 bool "Lock debugging: detect incorrect freeing of live locks"
1014 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1015 select DEBUG_SPINLOCK
1016 select DEBUG_MUTEXES
1019 This feature will check whether any held lock (spinlock, rwlock,
1020 mutex or rwsem) is incorrectly freed by the kernel, via any of the
1021 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
1022 vfree(), etc.), whether a live lock is incorrectly reinitialized via
1023 spin_lock_init()/mutex_init()/etc., or whether there is any lock
1024 held during task exit.
1026 config PROVE_LOCKING
1027 bool "Lock debugging: prove locking correctness"
1028 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1030 select DEBUG_SPINLOCK
1031 select DEBUG_MUTEXES
1032 select DEBUG_LOCK_ALLOC
1033 select TRACE_IRQFLAGS
1036 This feature enables the kernel to prove that all locking
1037 that occurs in the kernel runtime is mathematically
1038 correct: that under no circumstance could an arbitrary (and
1039 not yet triggered) combination of observed locking
1040 sequences (on an arbitrary number of CPUs, running an
1041 arbitrary number of tasks and interrupt contexts) cause a
1044 In short, this feature enables the kernel to report locking
1045 related deadlocks before they actually occur.
1047 The proof does not depend on how hard and complex a
1048 deadlock scenario would be to trigger: how many
1049 participant CPUs, tasks and irq-contexts would be needed
1050 for it to trigger. The proof also does not depend on
1051 timing: if a race and a resulting deadlock is possible
1052 theoretically (no matter how unlikely the race scenario
1053 is), it will be proven so and will immediately be
1054 reported by the kernel (once the event is observed that
1055 makes the deadlock theoretically possible).
1057 If a deadlock is impossible (i.e. the locking rules, as
1058 observed by the kernel, are mathematically correct), the
1059 kernel reports nothing.
1061 NOTE: this feature can also be enabled for rwlocks, mutexes
1062 and rwsems - in which case all dependencies between these
1063 different locking variants are observed and mapped too, and
1064 the proof of observed correctness is also maintained for an
1065 arbitrary combination of these separate locking variants.
1067 For more details, see Documentation/locking/lockdep-design.txt.
1071 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1073 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE && !ARC && !SCORE
1078 bool "Lock usage statistics"
1079 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1081 select DEBUG_SPINLOCK
1082 select DEBUG_MUTEXES
1083 select DEBUG_LOCK_ALLOC
1086 This feature enables tracking lock contention points
1088 For more details, see Documentation/locking/lockstat.txt
1090 This also enables lock events required by "perf lock",
1092 If you want to use "perf lock", you also need to turn on
1093 CONFIG_EVENT_TRACING.
1095 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
1096 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
1098 config DEBUG_LOCKDEP
1099 bool "Lock dependency engine debugging"
1100 depends on DEBUG_KERNEL && LOCKDEP
1102 If you say Y here, the lock dependency engine will do
1103 additional runtime checks to debug itself, at the price
1104 of more runtime overhead.
1106 config DEBUG_ATOMIC_SLEEP
1107 bool "Sleep inside atomic section checking"
1108 select PREEMPT_COUNT
1109 depends on DEBUG_KERNEL
1111 If you say Y here, various routines which may sleep will become very
1112 noisy if they are called inside atomic sections: when a spinlock is
1113 held, inside an rcu read side critical section, inside preempt disabled
1114 sections, inside an interrupt, etc...
1116 config DEBUG_LOCKING_API_SELFTESTS
1117 bool "Locking API boot-time self-tests"
1118 depends on DEBUG_KERNEL
1120 Say Y here if you want the kernel to run a short self-test during
1121 bootup. The self-test checks whether common types of locking bugs
1122 are detected by debugging mechanisms or not. (if you disable
1123 lock debugging then those bugs wont be detected of course.)
1124 The following locking APIs are covered: spinlocks, rwlocks,
1127 config LOCK_TORTURE_TEST
1128 tristate "torture tests for locking"
1129 depends on DEBUG_KERNEL
1133 This option provides a kernel module that runs torture tests
1134 on kernel locking primitives. The kernel module may be built
1135 after the fact on the running kernel to be tested, if desired.
1137 Say Y here if you want kernel locking-primitive torture tests
1138 to be built into the kernel.
1139 Say M if you want these torture tests to build as a module.
1140 Say N if you are unsure.
1142 endmenu # lock debugging
1144 config TRACE_IRQFLAGS
1147 Enables hooks to interrupt enabling and disabling for
1148 either tracing or lock debugging.
1151 bool "Stack backtrace support"
1152 depends on STACKTRACE_SUPPORT
1154 This option causes the kernel to create a /proc/pid/stack for
1155 every process, showing its current stack trace.
1156 It is also used by various kernel debugging features that require
1157 stack trace generation.
1159 config DEBUG_KOBJECT
1160 bool "kobject debugging"
1161 depends on DEBUG_KERNEL
1163 If you say Y here, some extra kobject debugging messages will be sent
1166 config DEBUG_KOBJECT_RELEASE
1167 bool "kobject release debugging"
1168 depends on DEBUG_OBJECTS_TIMERS
1170 kobjects are reference counted objects. This means that their
1171 last reference count put is not predictable, and the kobject can
1172 live on past the point at which a driver decides to drop it's
1173 initial reference to the kobject gained on allocation. An
1174 example of this would be a struct device which has just been
1177 However, some buggy drivers assume that after such an operation,
1178 the memory backing the kobject can be immediately freed. This
1179 goes completely against the principles of a refcounted object.
1181 If you say Y here, the kernel will delay the release of kobjects
1182 on the last reference count to improve the visibility of this
1183 kind of kobject release bug.
1185 config HAVE_DEBUG_BUGVERBOSE
1188 config DEBUG_BUGVERBOSE
1189 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
1190 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
1193 Say Y here to make BUG() panics output the file name and line number
1194 of the BUG call as well as the EIP and oops trace. This aids
1195 debugging but costs about 70-100K of memory.
1198 bool "Debug linked list manipulation"
1199 depends on DEBUG_KERNEL
1201 Enable this to turn on extended checks in the linked-list
1206 config DEBUG_PI_LIST
1207 bool "Debug priority linked list manipulation"
1208 depends on DEBUG_KERNEL
1210 Enable this to turn on extended checks in the priority-ordered
1211 linked-list (plist) walking routines. This checks the entire
1212 list multiple times during each manipulation.
1217 bool "Debug SG table operations"
1218 depends on DEBUG_KERNEL
1220 Enable this to turn on checks on scatter-gather tables. This can
1221 help find problems with drivers that do not properly initialize
1226 config DEBUG_NOTIFIERS
1227 bool "Debug notifier call chains"
1228 depends on DEBUG_KERNEL
1230 Enable this to turn on sanity checking for notifier call chains.
1231 This is most useful for kernel developers to make sure that
1232 modules properly unregister themselves from notifier chains.
1233 This is a relatively cheap check but if you care about maximum
1236 config DEBUG_CREDENTIALS
1237 bool "Debug credential management"
1238 depends on DEBUG_KERNEL
1240 Enable this to turn on some debug checking for credential
1241 management. The additional code keeps track of the number of
1242 pointers from task_structs to any given cred struct, and checks to
1243 see that this number never exceeds the usage count of the cred
1246 Furthermore, if SELinux is enabled, this also checks that the
1247 security pointer in the cred struct is never seen to be invalid.
1251 menu "RCU Debugging"
1254 def_bool PROVE_LOCKING
1256 config PROVE_RCU_REPEATEDLY
1257 bool "RCU debugging: don't disable PROVE_RCU on first splat"
1258 depends on PROVE_RCU
1261 By itself, PROVE_RCU will disable checking upon issuing the
1262 first warning (or "splat"). This feature prevents such
1263 disabling, allowing multiple RCU-lockdep warnings to be printed
1266 Say Y to allow multiple RCU-lockdep warnings per boot.
1268 Say N if you are unsure.
1270 config SPARSE_RCU_POINTER
1271 bool "RCU debugging: sparse-based checks for pointer usage"
1274 This feature enables the __rcu sparse annotation for
1275 RCU-protected pointers. This annotation will cause sparse
1276 to flag any non-RCU used of annotated pointers. This can be
1277 helpful when debugging RCU usage. Please note that this feature
1278 is not intended to enforce code cleanliness; it is instead merely
1281 Say Y to make sparse flag questionable use of RCU-protected pointers
1283 Say N if you are unsure.
1289 config RCU_TORTURE_TEST
1290 tristate "torture tests for RCU"
1291 depends on DEBUG_KERNEL
1297 This option provides a kernel module that runs torture tests
1298 on the RCU infrastructure. The kernel module may be built
1299 after the fact on the running kernel to be tested, if desired.
1301 Say Y here if you want RCU torture tests to be built into
1303 Say M if you want the RCU torture tests to build as a module.
1304 Say N if you are unsure.
1306 config RCU_TORTURE_TEST_RUNNABLE
1307 bool "torture tests for RCU runnable by default"
1308 depends on RCU_TORTURE_TEST = y
1311 This option provides a way to build the RCU torture tests
1312 directly into the kernel without them starting up at boot
1313 time. You can use /proc/sys/kernel/rcutorture_runnable
1314 to manually override this setting. This /proc file is
1315 available only when the RCU torture tests have been built
1318 Say Y here if you want the RCU torture tests to start during
1319 boot (you probably don't).
1320 Say N here if you want the RCU torture tests to start only
1321 after being manually enabled via /proc.
1323 config RCU_TORTURE_TEST_SLOW_PREINIT
1324 bool "Slow down RCU grace-period pre-initialization to expose races"
1325 depends on RCU_TORTURE_TEST
1327 This option delays grace-period pre-initialization (the
1328 propagation of CPU-hotplug changes up the rcu_node combining
1329 tree) for a few jiffies between initializing each pair of
1330 consecutive rcu_node structures. This helps to expose races
1331 involving grace-period pre-initialization, in other words, it
1332 makes your kernel less stable. It can also greatly increase
1333 grace-period latency, especially on systems with large numbers
1334 of CPUs. This is useful when torture-testing RCU, but in
1335 almost no other circumstance.
1337 Say Y here if you want your system to crash and hang more often.
1338 Say N if you want a sane system.
1340 config RCU_TORTURE_TEST_SLOW_PREINIT_DELAY
1341 int "How much to slow down RCU grace-period pre-initialization"
1344 depends on RCU_TORTURE_TEST_SLOW_PREINIT
1346 This option specifies the number of jiffies to wait between
1347 each rcu_node structure pre-initialization step.
1349 config RCU_TORTURE_TEST_SLOW_INIT
1350 bool "Slow down RCU grace-period initialization to expose races"
1351 depends on RCU_TORTURE_TEST
1353 This option delays grace-period initialization for a few
1354 jiffies between initializing each pair of consecutive
1355 rcu_node structures. This helps to expose races involving
1356 grace-period initialization, in other words, it makes your
1357 kernel less stable. It can also greatly increase grace-period
1358 latency, especially on systems with large numbers of CPUs.
1359 This is useful when torture-testing RCU, but in almost no
1362 Say Y here if you want your system to crash and hang more often.
1363 Say N if you want a sane system.
1365 config RCU_TORTURE_TEST_SLOW_INIT_DELAY
1366 int "How much to slow down RCU grace-period initialization"
1369 depends on RCU_TORTURE_TEST_SLOW_INIT
1371 This option specifies the number of jiffies to wait between
1372 each rcu_node structure initialization.
1374 config RCU_TORTURE_TEST_SLOW_CLEANUP
1375 bool "Slow down RCU grace-period cleanup to expose races"
1376 depends on RCU_TORTURE_TEST
1378 This option delays grace-period cleanup for a few jiffies
1379 between cleaning up each pair of consecutive rcu_node
1380 structures. This helps to expose races involving grace-period
1381 cleanup, in other words, it makes your kernel less stable.
1382 It can also greatly increase grace-period latency, especially
1383 on systems with large numbers of CPUs. This is useful when
1384 torture-testing RCU, but in almost no other circumstance.
1386 Say Y here if you want your system to crash and hang more often.
1387 Say N if you want a sane system.
1389 config RCU_TORTURE_TEST_SLOW_CLEANUP_DELAY
1390 int "How much to slow down RCU grace-period cleanup"
1393 depends on RCU_TORTURE_TEST_SLOW_CLEANUP
1395 This option specifies the number of jiffies to wait between
1396 each rcu_node structure cleanup operation.
1398 config RCU_CPU_STALL_TIMEOUT
1399 int "RCU CPU stall timeout in seconds"
1400 depends on RCU_STALL_COMMON
1404 If a given RCU grace period extends more than the specified
1405 number of seconds, a CPU stall warning is printed. If the
1406 RCU grace period persists, additional CPU stall warnings are
1407 printed at more widely spaced intervals.
1410 bool "Enable tracing for RCU"
1411 depends on DEBUG_KERNEL
1414 This option provides tracing in RCU which presents stats
1415 in debugfs for debugging RCU implementation.
1417 Say Y here if you want to enable RCU tracing
1418 Say N if you are unsure.
1420 config RCU_EQS_DEBUG
1421 bool "Provide debugging asserts for adding NO_HZ support to an arch"
1422 depends on DEBUG_KERNEL
1424 This option provides consistency checks in RCU's handling of
1425 NO_HZ. These checks have proven quite helpful in detecting
1426 bugs in arch-specific NO_HZ code.
1428 Say N here if you need ultimate kernel/user switch latencies
1429 Say Y if you are unsure
1431 endmenu # "RCU Debugging"
1433 config DEBUG_BLOCK_EXT_DEVT
1434 bool "Force extended block device numbers and spread them"
1435 depends on DEBUG_KERNEL
1439 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1440 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1441 YOU ARE DOING. Distros, please enable this and fix whatever
1444 Conventionally, block device numbers are allocated from
1445 predetermined contiguous area. However, extended block area
1446 may introduce non-contiguous block device numbers. This
1447 option forces most block device numbers to be allocated from
1448 the extended space and spreads them to discover kernel or
1449 userland code paths which assume predetermined contiguous
1450 device number allocation.
1452 Note that turning on this debug option shuffles all the
1453 device numbers for all IDE and SCSI devices including libata
1454 ones, so root partition specified using device number
1455 directly (via rdev or root=MAJ:MIN) won't work anymore.
1456 Textual device names (root=/dev/sdXn) will continue to work.
1458 Say N if you are unsure.
1460 config NOTIFIER_ERROR_INJECTION
1461 tristate "Notifier error injection"
1462 depends on DEBUG_KERNEL
1465 This option provides the ability to inject artificial errors to
1466 specified notifier chain callbacks. It is useful to test the error
1467 handling of notifier call chain failures.
1471 config CPU_NOTIFIER_ERROR_INJECT
1472 tristate "CPU notifier error injection module"
1473 depends on HOTPLUG_CPU && NOTIFIER_ERROR_INJECTION
1475 This option provides a kernel module that can be used to test
1476 the error handling of the cpu notifiers by injecting artificial
1477 errors to CPU notifier chain callbacks. It is controlled through
1478 debugfs interface under /sys/kernel/debug/notifier-error-inject/cpu
1480 If the notifier call chain should be failed with some events
1481 notified, write the error code to "actions/<notifier event>/error".
1483 Example: Inject CPU offline error (-1 == -EPERM)
1485 # cd /sys/kernel/debug/notifier-error-inject/cpu
1486 # echo -1 > actions/CPU_DOWN_PREPARE/error
1487 # echo 0 > /sys/devices/system/cpu/cpu1/online
1488 bash: echo: write error: Operation not permitted
1490 To compile this code as a module, choose M here: the module will
1491 be called cpu-notifier-error-inject.
1495 config PM_NOTIFIER_ERROR_INJECT
1496 tristate "PM notifier error injection module"
1497 depends on PM && NOTIFIER_ERROR_INJECTION
1498 default m if PM_DEBUG
1500 This option provides the ability to inject artificial errors to
1501 PM notifier chain callbacks. It is controlled through debugfs
1502 interface /sys/kernel/debug/notifier-error-inject/pm
1504 If the notifier call chain should be failed with some events
1505 notified, write the error code to "actions/<notifier event>/error".
1507 Example: Inject PM suspend error (-12 = -ENOMEM)
1509 # cd /sys/kernel/debug/notifier-error-inject/pm/
1510 # echo -12 > actions/PM_SUSPEND_PREPARE/error
1511 # echo mem > /sys/power/state
1512 bash: echo: write error: Cannot allocate memory
1514 To compile this code as a module, choose M here: the module will
1515 be called pm-notifier-error-inject.
1519 config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1520 tristate "OF reconfig notifier error injection module"
1521 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1523 This option provides the ability to inject artificial errors to
1524 OF reconfig notifier chain callbacks. It is controlled
1525 through debugfs interface under
1526 /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1528 If the notifier call chain should be failed with some events
1529 notified, write the error code to "actions/<notifier event>/error".
1531 To compile this code as a module, choose M here: the module will
1532 be called of-reconfig-notifier-error-inject.
1536 config FAULT_INJECTION
1537 bool "Fault-injection framework"
1538 depends on DEBUG_KERNEL
1540 Provide fault-injection framework.
1541 For more details, see Documentation/fault-injection/.
1544 bool "Fault-injection capability for kmalloc"
1545 depends on FAULT_INJECTION
1546 depends on SLAB || SLUB
1548 Provide fault-injection capability for kmalloc.
1550 config FAIL_PAGE_ALLOC
1551 bool "Fault-injection capabilitiy for alloc_pages()"
1552 depends on FAULT_INJECTION
1554 Provide fault-injection capability for alloc_pages().
1556 config FAIL_MAKE_REQUEST
1557 bool "Fault-injection capability for disk IO"
1558 depends on FAULT_INJECTION && BLOCK
1560 Provide fault-injection capability for disk IO.
1562 config FAIL_IO_TIMEOUT
1563 bool "Fault-injection capability for faking disk interrupts"
1564 depends on FAULT_INJECTION && BLOCK
1566 Provide fault-injection capability on end IO handling. This
1567 will make the block layer "forget" an interrupt as configured,
1568 thus exercising the error handling.
1570 Only works with drivers that use the generic timeout handling,
1571 for others it wont do anything.
1573 config FAIL_MMC_REQUEST
1574 bool "Fault-injection capability for MMC IO"
1576 depends on FAULT_INJECTION && MMC
1578 Provide fault-injection capability for MMC IO.
1579 This will make the mmc core return data errors. This is
1580 useful to test the error handling in the mmc block device
1581 and to test how the mmc host driver handles retries from
1585 bool "Fault-injection capability for futexes"
1587 depends on FAULT_INJECTION && FUTEX
1589 Provide fault-injection capability for futexes.
1591 config FAULT_INJECTION_DEBUG_FS
1592 bool "Debugfs entries for fault-injection capabilities"
1593 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1595 Enable configuration of fault-injection capabilities via debugfs.
1597 config FAULT_INJECTION_STACKTRACE_FILTER
1598 bool "stacktrace filter for fault-injection capabilities"
1599 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1602 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC && !SCORE
1604 Provide stacktrace filter for fault-injection capabilities
1607 bool "Latency measuring infrastructure"
1608 depends on HAVE_LATENCYTOP_SUPPORT
1609 depends on DEBUG_KERNEL
1610 depends on STACKTRACE_SUPPORT
1612 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC
1619 Enable this option if you want to use the LatencyTOP tool
1620 to find out which userspace is blocking on what kernel operations.
1622 config ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1625 config DEBUG_STRICT_USER_COPY_CHECKS
1626 bool "Strict user copy size checks"
1627 depends on ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS
1628 depends on DEBUG_KERNEL && !TRACE_BRANCH_PROFILING
1630 Enabling this option turns a certain set of sanity checks for user
1631 copy operations into compile time failures.
1633 The copy_from_user() etc checks are there to help test if there
1634 are sufficient security checks on the length argument of
1635 the copy operation, by having gcc prove that the argument is
1640 source kernel/trace/Kconfig
1642 menu "Runtime Testing"
1645 tristate "Linux Kernel Dump Test Tool Module"
1650 This module enables testing of the different dumping mechanisms by
1651 inducing system failures at predefined crash points.
1652 If you don't need it: say N
1653 Choose M here to compile this code as a module. The module will be
1656 Documentation on how to use the module can be found in
1657 Documentation/fault-injection/provoke-crashes.txt
1659 config TEST_LIST_SORT
1660 bool "Linked list sorting test"
1661 depends on DEBUG_KERNEL
1663 Enable this to turn on 'list_sort()' function test. This test is
1664 executed only once during system boot, so affects only boot time.
1668 config KPROBES_SANITY_TEST
1669 bool "Kprobes sanity tests"
1670 depends on DEBUG_KERNEL
1674 This option provides for testing basic kprobes functionality on
1675 boot. A sample kprobe, jprobe and kretprobe are inserted and
1676 verified for functionality.
1678 Say N if you are unsure.
1680 config BACKTRACE_SELF_TEST
1681 tristate "Self test for the backtrace code"
1682 depends on DEBUG_KERNEL
1685 This option provides a kernel module that can be used to test
1686 the kernel stack backtrace code. This option is not useful
1687 for distributions or general kernels, but only for kernel
1688 developers working on architecture code.
1690 Note that if you want to also test saved backtraces, you will
1691 have to enable STACKTRACE as well.
1693 Say N if you are unsure.
1696 tristate "Red-Black tree test"
1697 depends on DEBUG_KERNEL
1699 A benchmark measuring the performance of the rbtree library.
1700 Also includes rbtree invariant checks.
1702 config INTERVAL_TREE_TEST
1703 tristate "Interval tree test"
1704 depends on m && DEBUG_KERNEL
1705 select INTERVAL_TREE
1707 A benchmark measuring the performance of the interval tree library
1710 tristate "Per cpu operations test"
1711 depends on m && DEBUG_KERNEL
1713 Enable this option to build test module which validates per-cpu
1718 config ATOMIC64_SELFTEST
1719 bool "Perform an atomic64_t self-test at boot"
1721 Enable this option to test the atomic64_t functions at boot.
1725 config ASYNC_RAID6_TEST
1726 tristate "Self test for hardware accelerated raid6 recovery"
1727 depends on ASYNC_RAID6_RECOV
1730 This is a one-shot self test that permutes through the
1731 recovery of all the possible two disk failure scenarios for a
1732 N-disk array. Recovery is performed with the asynchronous
1733 raid6 recovery routines, and will optionally use an offload
1734 engine if one is available.
1739 tristate "Test functions located in the hexdump module at runtime"
1741 config TEST_STRING_HELPERS
1742 tristate "Test functions located in the string_helpers module at runtime"
1745 tristate "Test kstrto*() family of functions at runtime"
1748 tristate "Test printf() family of functions at runtime"
1750 config TEST_RHASHTABLE
1751 tristate "Perform selftest on resizable hash table"
1754 Enable this option to test the rhashtable functions at boot.
1758 endmenu # runtime tests
1760 config PROVIDE_OHCI1394_DMA_INIT
1761 bool "Remote debugging over FireWire early on boot"
1762 depends on PCI && X86
1764 If you want to debug problems which hang or crash the kernel early
1765 on boot and the crashing machine has a FireWire port, you can use
1766 this feature to remotely access the memory of the crashed machine
1767 over FireWire. This employs remote DMA as part of the OHCI1394
1768 specification which is now the standard for FireWire controllers.
1770 With remote DMA, you can monitor the printk buffer remotely using
1771 firescope and access all memory below 4GB using fireproxy from gdb.
1772 Even controlling a kernel debugger is possible using remote DMA.
1776 If ohci1394_dma=early is used as boot parameter, it will initialize
1777 all OHCI1394 controllers which are found in the PCI config space.
1779 As all changes to the FireWire bus such as enabling and disabling
1780 devices cause a bus reset and thereby disable remote DMA for all
1781 devices, be sure to have the cable plugged and FireWire enabled on
1782 the debugging host before booting the debug target for debugging.
1784 This code (~1k) is freed after boot. By then, the firewire stack
1785 in charge of the OHCI-1394 controllers should be used instead.
1787 See Documentation/debugging-via-ohci1394.txt for more information.
1790 bool "Build targets in Documentation/ tree"
1791 depends on HEADERS_CHECK
1793 This option attempts to build objects from the source files in the
1794 kernel Documentation/ tree.
1796 Say N if you are unsure.
1798 config DMA_API_DEBUG
1799 bool "Enable debugging of DMA-API usage"
1800 depends on HAVE_DMA_API_DEBUG
1802 Enable this option to debug the use of the DMA API by device drivers.
1803 With this option you will be able to detect common bugs in device
1804 drivers like double-freeing of DMA mappings or freeing mappings that
1805 were never allocated.
1807 This also attempts to catch cases where a page owned by DMA is
1808 accessed by the cpu in a way that could cause data corruption. For
1809 example, this enables cow_user_page() to check that the source page is
1812 This option causes a performance degradation. Use only if you want to
1813 debug device drivers and dma interactions.
1818 tristate "Test module loading with 'hello world' module"
1822 This builds the "test_module" module that emits "Hello, world"
1823 on printk when loaded. It is designed to be used for basic
1824 evaluation of the module loading subsystem (for example when
1825 validating module verification). It lacks any extra dependencies,
1826 and will not normally be loaded by the system unless explicitly
1831 config TEST_USER_COPY
1832 tristate "Test user/kernel boundary protections"
1836 This builds the "test_user_copy" module that runs sanity checks
1837 on the copy_to/from_user infrastructure, making sure basic
1838 user/kernel boundary testing is working. If it fails to load,
1839 a regression has been detected in the user/kernel memory boundary
1845 tristate "Test BPF filter functionality"
1849 This builds the "test_bpf" module that runs various test vectors
1850 against the BPF interpreter or BPF JIT compiler depending on the
1851 current setting. This is in particular useful for BPF JIT compiler
1852 development, but also to run regression tests against changes in
1853 the interpreter code. It also enables test stubs for eBPF maps and
1854 verifier used by user space verifier testsuite.
1858 config TEST_FIRMWARE
1859 tristate "Test firmware loading via userspace interface"
1861 depends on FW_LOADER
1863 This builds the "test_firmware" module that creates a userspace
1864 interface for testing firmware loading. This can be used to
1865 control the triggering of firmware loading without needing an
1866 actual firmware-using device. The contents can be rechecked by
1872 tristate "udelay test driver"
1875 This builds the "udelay_test" module that helps to make sure
1876 that udelay() is working properly.
1882 depends on HAVE_MEMBLOCK
1884 This option adds a kernel parameter 'memtest', which allows memtest
1886 memtest=0, mean disabled; -- default
1887 memtest=1, mean do 1 test pattern;
1889 memtest=17, mean do 17 test patterns.
1890 If you are unsure how to answer this question, answer N.
1892 config TEST_STATIC_KEYS
1893 tristate "Test static keys"
1897 Test the static key interfaces.
1901 source "samples/Kconfig"
1903 source "lib/Kconfig.kgdb"