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
18 config MESSAGE_LOGLEVEL_DEFAULT
19 int "Default message log level (1-7)"
23 Default log level for printk statements with no specified priority.
25 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
26 that are auditing their logs closely may want to set it to a lower
29 config BOOT_PRINTK_DELAY
30 bool "Delay each boot printk message by N milliseconds"
31 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
33 This build option allows you to read kernel boot messages
34 by inserting a short delay after each one. The delay is
35 specified in milliseconds on the kernel command line,
38 It is likely that you would also need to use "lpj=M" to preset
39 the "loops per jiffie" value.
40 See a previous boot log for the "lpj" value to use for your
41 system, and then set "lpj=M" before setting "boot_delay=N".
42 NOTE: Using this option may adversely affect SMP systems.
43 I.e., processors other than the first one may not boot up.
44 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
45 what it believes to be lockup conditions.
48 bool "Enable dynamic printk() support"
54 Compiles debug level messages into the kernel, which would not
55 otherwise be available at runtime. These messages can then be
56 enabled/disabled based on various levels of scope - per source file,
57 function, module, format string, and line number. This mechanism
58 implicitly compiles in all pr_debug() and dev_dbg() calls, which
59 enlarges the kernel text size by about 2%.
61 If a source file is compiled with DEBUG flag set, any
62 pr_debug() calls in it are enabled by default, but can be
63 disabled at runtime as below. Note that DEBUG flag is
64 turned on by many CONFIG_*DEBUG* options.
68 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
69 which is contained in the 'debugfs' filesystem. Thus, the debugfs
70 filesystem must first be mounted before making use of this feature.
71 We refer the control file as: <debugfs>/dynamic_debug/control. This
72 file contains a list of the debug statements that can be enabled. The
73 format for each line of the file is:
75 filename:lineno [module]function flags format
77 filename : source file of the debug statement
78 lineno : line number of the debug statement
79 module : module that contains the debug statement
80 function : function that contains the debug statement
81 flags : '=p' means the line is turned 'on' for printing
82 format : the format used for the debug statement
86 nullarbor:~ # cat <debugfs>/dynamic_debug/control
87 # filename:lineno [module]function flags format
88 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
89 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
90 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
94 // enable the message at line 1603 of file svcsock.c
95 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
96 <debugfs>/dynamic_debug/control
98 // enable all the messages in file svcsock.c
99 nullarbor:~ # echo -n 'file svcsock.c +p' >
100 <debugfs>/dynamic_debug/control
102 // enable all the messages in the NFS server module
103 nullarbor:~ # echo -n 'module nfsd +p' >
104 <debugfs>/dynamic_debug/control
106 // enable all 12 messages in the function svc_process()
107 nullarbor:~ # echo -n 'func svc_process +p' >
108 <debugfs>/dynamic_debug/control
110 // disable all 12 messages in the function svc_process()
111 nullarbor:~ # echo -n 'func svc_process -p' >
112 <debugfs>/dynamic_debug/control
114 See Documentation/dynamic-debug-howto.txt for additional information.
116 endmenu # "printk and dmesg options"
118 menu "Compile-time checks and compiler options"
121 bool "Compile the kernel with debug info"
122 depends on DEBUG_KERNEL && !COMPILE_TEST
124 If you say Y here the resulting kernel image will include
125 debugging info resulting in a larger kernel image.
126 This adds debug symbols to the kernel and modules (gcc -g), and
127 is needed if you intend to use kernel crashdump or binary object
128 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
129 Say Y here only if you plan to debug the kernel.
133 config DEBUG_INFO_REDUCED
134 bool "Reduce debugging information"
135 depends on DEBUG_INFO
137 If you say Y here gcc is instructed to generate less debugging
138 information for structure types. This means that tools that
139 need full debugging information (like kgdb or systemtap) won't
140 be happy. But if you merely need debugging information to
141 resolve line numbers there is no loss. Advantage is that
142 build directory object sizes shrink dramatically over a full
143 DEBUG_INFO build and compile times are reduced too.
144 Only works with newer gcc versions.
146 config DEBUG_INFO_SPLIT
147 bool "Produce split debuginfo in .dwo files"
148 depends on DEBUG_INFO && !FRV
150 Generate debug info into separate .dwo files. This significantly
151 reduces the build directory size for builds with DEBUG_INFO,
152 because it stores the information only once on disk in .dwo
153 files instead of multiple times in object files and executables.
154 In addition the debug information is also compressed.
156 Requires recent gcc (4.7+) and recent gdb/binutils.
157 Any tool that packages or reads debug information would need
158 to know about the .dwo files and include them.
159 Incompatible with older versions of ccache.
161 config DEBUG_INFO_DWARF4
162 bool "Generate dwarf4 debuginfo"
163 depends on DEBUG_INFO
165 Generate dwarf4 debug info. This requires recent versions
166 of gcc and gdb. It makes the debug information larger.
167 But it significantly improves the success of resolving
168 variables in gdb on optimized code.
171 bool "Provide GDB scripts for kernel debugging"
172 depends on DEBUG_INFO
174 This creates the required links to GDB helper scripts in the
175 build directory. If you load vmlinux into gdb, the helper
176 scripts will be automatically imported by gdb as well, and
177 additional functions are available to analyze a Linux kernel
178 instance. See Documentation/gdb-kernel-debugging.txt for further
181 config ENABLE_WARN_DEPRECATED
182 bool "Enable __deprecated logic"
185 Enable the __deprecated logic in the kernel build.
186 Disable this to suppress the "warning: 'foo' is deprecated
187 (declared at kernel/power/somefile.c:1234)" messages.
189 config ENABLE_MUST_CHECK
190 bool "Enable __must_check logic"
193 Enable the __must_check logic in the kernel build. Disable this to
194 suppress the "warning: ignoring return value of 'foo', declared with
195 attribute warn_unused_result" messages.
198 int "Warn for stack frames larger than (needs gcc 4.4)"
200 default 2048 if GCC_PLUGIN_LATENT_ENTROPY
201 default 1024 if !64BIT
202 default 2048 if 64BIT
204 Tell gcc to warn at build time for stack frames larger than this.
205 Setting this too low will cause a lot of warnings.
206 Setting it to 0 disables the warning.
209 config STRIP_ASM_SYMS
210 bool "Strip assembler-generated symbols during link"
213 Strip internal assembler-generated symbols during a link (symbols
214 that look like '.Lxxx') so they don't pollute the output of
215 get_wchan() and suchlike.
218 bool "Generate readable assembler code"
219 depends on DEBUG_KERNEL
221 Disable some compiler optimizations that tend to generate human unreadable
222 assembler output. This may make the kernel slightly slower, but it helps
223 to keep kernel developers who have to stare a lot at assembler listings
226 config UNUSED_SYMBOLS
227 bool "Enable unused/obsolete exported symbols"
230 Unused but exported symbols make the kernel needlessly bigger. For
231 that reason most of these unused exports will soon be removed. This
232 option is provided temporarily to provide a transition period in case
233 some external kernel module needs one of these symbols anyway. If you
234 encounter such a case in your module, consider if you are actually
235 using the right API. (rationale: since nobody in the kernel is using
236 this in a module, there is a pretty good chance it's actually the
237 wrong interface to use). If you really need the symbol, please send a
238 mail to the linux kernel mailing list mentioning the symbol and why
239 you really need it, and what the merge plan to the mainline kernel for
243 bool "Track page owner"
244 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT
248 select PAGE_EXTENSION
250 This keeps track of what call chain is the owner of a page, may
251 help to find bare alloc_page(s) leaks. Even if you include this
252 feature on your build, it is disabled in default. You should pass
253 "page_owner=on" to boot parameter in order to enable it. Eats
254 a fair amount of memory if enabled. See tools/vm/page_owner_sort.c
255 for user-space helper.
260 bool "Debug Filesystem"
263 debugfs is a virtual file system that kernel developers use to put
264 debugging files into. Enable this option to be able to read and
265 write to these files.
267 For detailed documentation on the debugfs API, see
268 Documentation/DocBook/filesystems.
273 bool "Run 'make headers_check' when building vmlinux"
276 This option will extract the user-visible kernel headers whenever
277 building the kernel, and will run basic sanity checks on them to
278 ensure that exported files do not attempt to include files which
279 were not exported, etc.
281 If you're making modifications to header files which are
282 relevant for userspace, say 'Y', and check the headers
283 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
284 your build tree), to make sure they're suitable.
286 config DEBUG_SECTION_MISMATCH
287 bool "Enable full Section mismatch analysis"
289 The section mismatch analysis checks if there are illegal
290 references from one section to another section.
291 During linktime or runtime, some sections are dropped;
292 any use of code/data previously in these sections would
293 most likely result in an oops.
294 In the code, functions and variables are annotated with
295 __init,, etc. (see the full list in include/linux/init.h),
296 which results in the code/data being placed in specific sections.
297 The section mismatch analysis is always performed after a full
298 kernel build, and enabling this option causes the following
299 additional steps to occur:
300 - Add the option -fno-inline-functions-called-once to gcc commands.
301 When inlining a function annotated with __init in a non-init
302 function, we would lose the section information and thus
303 the analysis would not catch the illegal reference.
304 This option tells gcc to inline less (but it does result in
306 - Run the section mismatch analysis for each module/built-in.o file.
307 When we run the section mismatch analysis on vmlinux.o, we
308 lose valuable information about where the mismatch was
310 Running the analysis for each module/built-in.o file
311 tells where the mismatch happens much closer to the
312 source. The drawback is that the same mismatch is
313 reported at least twice.
314 - Enable verbose reporting from modpost in order to help resolve
315 the section mismatches that are reported.
317 config SECTION_MISMATCH_WARN_ONLY
318 bool "Make section mismatch errors non-fatal"
321 If you say N here, the build process will fail if there are any
322 section mismatch, instead of just throwing warnings.
327 # Select this config option from the architecture Kconfig, if it
328 # is preferred to always offer frame pointers as a config
329 # option on the architecture (regardless of KERNEL_DEBUG):
331 config ARCH_WANT_FRAME_POINTERS
336 bool "Compile the kernel with frame pointers"
337 depends on DEBUG_KERNEL && \
338 (CRIS || M68K || FRV || UML || \
339 AVR32 || SUPERH || BLACKFIN || MN10300 || METAG) || \
340 ARCH_WANT_FRAME_POINTERS
341 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
343 If you say Y here the resulting kernel image will be slightly
344 larger and slower, but it gives very useful debugging information
345 in case of kernel bugs. (precise oopses/stacktraces/warnings)
347 config STACK_VALIDATION
348 bool "Compile-time stack metadata validation"
349 depends on HAVE_STACK_VALIDATION
352 Add compile-time checks to validate stack metadata, including frame
353 pointers (if CONFIG_FRAME_POINTER is enabled). This helps ensure
354 that runtime stack traces are more reliable.
356 For more information, see
357 tools/objtool/Documentation/stack-validation.txt.
359 config DEBUG_FORCE_WEAK_PER_CPU
360 bool "Force weak per-cpu definitions"
361 depends on DEBUG_KERNEL
363 s390 and alpha require percpu variables in modules to be
364 defined weak to work around addressing range issue which
365 puts the following two restrictions on percpu variable
368 1. percpu symbols must be unique whether static or not
369 2. percpu variables can't be defined inside a function
371 To ensure that generic code follows the above rules, this
372 option forces all percpu variables to be defined as weak.
374 endmenu # "Compiler options"
377 bool "Magic SysRq key"
380 If you say Y here, you will have some control over the system even
381 if the system crashes for example during kernel debugging (e.g., you
382 will be able to flush the buffer cache to disk, reboot the system
383 immediately or dump some status information). This is accomplished
384 by pressing various keys while holding SysRq (Alt+PrintScreen). It
385 also works on a serial console (on PC hardware at least), if you
386 send a BREAK and then within 5 seconds a command keypress. The
387 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
388 unless you really know what this hack does.
390 config MAGIC_SYSRQ_DEFAULT_ENABLE
391 hex "Enable magic SysRq key functions by default"
392 depends on MAGIC_SYSRQ
395 Specifies which SysRq key functions are enabled by default.
396 This may be set to 1 or 0 to enable or disable them all, or
397 to a bitmask as described in Documentation/sysrq.txt.
400 bool "Kernel debugging"
402 Say Y here if you are developing drivers or trying to debug and
403 identify kernel problems.
405 menu "Memory Debugging"
407 source mm/Kconfig.debug
410 bool "Debug object operations"
411 depends on DEBUG_KERNEL
413 If you say Y here, additional code will be inserted into the
414 kernel to track the life time of various objects and validate
415 the operations on those objects.
417 config DEBUG_OBJECTS_SELFTEST
418 bool "Debug objects selftest"
419 depends on DEBUG_OBJECTS
421 This enables the selftest of the object debug code.
423 config DEBUG_OBJECTS_FREE
424 bool "Debug objects in freed memory"
425 depends on DEBUG_OBJECTS
427 This enables checks whether a k/v free operation frees an area
428 which contains an object which has not been deactivated
429 properly. This can make kmalloc/kfree-intensive workloads
432 config DEBUG_OBJECTS_TIMERS
433 bool "Debug timer objects"
434 depends on DEBUG_OBJECTS
436 If you say Y here, additional code will be inserted into the
437 timer routines to track the life time of timer objects and
438 validate the timer operations.
440 config DEBUG_OBJECTS_WORK
441 bool "Debug work objects"
442 depends on DEBUG_OBJECTS
444 If you say Y here, additional code will be inserted into the
445 work queue routines to track the life time of work objects and
446 validate the work operations.
448 config DEBUG_OBJECTS_RCU_HEAD
449 bool "Debug RCU callbacks objects"
450 depends on DEBUG_OBJECTS
452 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
454 config DEBUG_OBJECTS_PERCPU_COUNTER
455 bool "Debug percpu counter objects"
456 depends on DEBUG_OBJECTS
458 If you say Y here, additional code will be inserted into the
459 percpu counter routines to track the life time of percpu counter
460 objects and validate the percpu counter operations.
462 config DEBUG_OBJECTS_ENABLE_DEFAULT
463 int "debug_objects bootup default value (0-1)"
466 depends on DEBUG_OBJECTS
468 Debug objects boot parameter default value
471 bool "Debug slab memory allocations"
472 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK
474 Say Y here to have the kernel do limited verification on memory
475 allocation as well as poisoning memory on free to catch use of freed
476 memory. This can make kmalloc/kfree-intensive workloads much slower.
478 config DEBUG_SLAB_LEAK
479 bool "Memory leak debugging"
480 depends on DEBUG_SLAB
483 bool "SLUB debugging on by default"
484 depends on SLUB && SLUB_DEBUG && !KMEMCHECK
487 Boot with debugging on by default. SLUB boots by default with
488 the runtime debug capabilities switched off. Enabling this is
489 equivalent to specifying the "slub_debug" parameter on boot.
490 There is no support for more fine grained debug control like
491 possible with slub_debug=xxx. SLUB debugging may be switched
492 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
497 bool "Enable SLUB performance statistics"
498 depends on SLUB && SYSFS
500 SLUB statistics are useful to debug SLUBs allocation behavior in
501 order find ways to optimize the allocator. This should never be
502 enabled for production use since keeping statistics slows down
503 the allocator by a few percentage points. The slabinfo command
504 supports the determination of the most active slabs to figure
505 out which slabs are relevant to a particular load.
506 Try running: slabinfo -DA
508 config HAVE_DEBUG_KMEMLEAK
511 config DEBUG_KMEMLEAK
512 bool "Kernel memory leak detector"
513 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
515 select STACKTRACE if STACKTRACE_SUPPORT
519 Say Y here if you want to enable the memory leak
520 detector. The memory allocation/freeing is traced in a way
521 similar to the Boehm's conservative garbage collector, the
522 difference being that the orphan objects are not freed but
523 only shown in /sys/kernel/debug/kmemleak. Enabling this
524 feature will introduce an overhead to memory
525 allocations. See Documentation/kmemleak.txt for more
528 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
529 of finding leaks due to the slab objects poisoning.
531 In order to access the kmemleak file, debugfs needs to be
532 mounted (usually at /sys/kernel/debug).
534 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
535 int "Maximum kmemleak early log entries"
536 depends on DEBUG_KMEMLEAK
540 Kmemleak must track all the memory allocations to avoid
541 reporting false positives. Since memory may be allocated or
542 freed before kmemleak is initialised, an early log buffer is
543 used to store these actions. If kmemleak reports "early log
544 buffer exceeded", please increase this value.
546 config DEBUG_KMEMLEAK_TEST
547 tristate "Simple test for the kernel memory leak detector"
548 depends on DEBUG_KMEMLEAK && m
550 This option enables a module that explicitly leaks memory.
554 config DEBUG_KMEMLEAK_DEFAULT_OFF
555 bool "Default kmemleak to off"
556 depends on DEBUG_KMEMLEAK
558 Say Y here to disable kmemleak by default. It can then be enabled
559 on the command line via kmemleak=on.
561 config DEBUG_STACK_USAGE
562 bool "Stack utilization instrumentation"
563 depends on DEBUG_KERNEL && !IA64
565 Enables the display of the minimum amount of free stack which each
566 task has ever had available in the sysrq-T and sysrq-P debug output.
568 This option will slow down process creation somewhat.
572 depends on DEBUG_KERNEL
574 Enable this to turn on extended checks in the virtual-memory system
575 that may impact performance.
579 config DEBUG_VM_VMACACHE
580 bool "Debug VMA caching"
583 Enable this to turn on VMA caching debug information. Doing so
584 can cause significant overhead, so only enable it in non-production
590 bool "Debug VM red-black trees"
593 Enable VM red-black tree debugging information and extra validations.
597 config DEBUG_VM_PGFLAGS
598 bool "Debug page-flags operations"
601 Enables extra validation on page flags operations.
606 bool "Debug VM translations"
607 depends on DEBUG_KERNEL && X86
609 Enable some costly sanity checks in virtual to page code. This can
610 catch mistakes with virt_to_page() and friends.
614 config DEBUG_NOMMU_REGIONS
615 bool "Debug the global anon/private NOMMU mapping region tree"
616 depends on DEBUG_KERNEL && !MMU
618 This option causes the global tree of anonymous and private mapping
619 regions to be regularly checked for invalid topology.
621 config DEBUG_MEMORY_INIT
622 bool "Debug memory initialisation" if EXPERT
625 Enable this for additional checks during memory initialisation.
626 The sanity checks verify aspects of the VM such as the memory model
627 and other information provided by the architecture. Verbose
628 information will be printed at KERN_DEBUG loglevel depending
629 on the mminit_loglevel= command-line option.
633 config MEMORY_NOTIFIER_ERROR_INJECT
634 tristate "Memory hotplug notifier error injection module"
635 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
637 This option provides the ability to inject artificial errors to
638 memory hotplug notifier chain callbacks. It is controlled through
639 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
641 If the notifier call chain should be failed with some events
642 notified, write the error code to "actions/<notifier event>/error".
644 Example: Inject memory hotplug offline error (-12 == -ENOMEM)
646 # cd /sys/kernel/debug/notifier-error-inject/memory
647 # echo -12 > actions/MEM_GOING_OFFLINE/error
648 # echo offline > /sys/devices/system/memory/memoryXXX/state
649 bash: echo: write error: Cannot allocate memory
651 To compile this code as a module, choose M here: the module will
652 be called memory-notifier-error-inject.
656 config DEBUG_PER_CPU_MAPS
657 bool "Debug access to per_cpu maps"
658 depends on DEBUG_KERNEL
661 Say Y to verify that the per_cpu map being accessed has
662 been set up. This adds a fair amount of code to kernel memory
663 and decreases performance.
668 bool "Highmem debugging"
669 depends on DEBUG_KERNEL && HIGHMEM
671 This option enables additional error checking for high memory
672 systems. Disable for production systems.
674 config HAVE_DEBUG_STACKOVERFLOW
677 config DEBUG_STACKOVERFLOW
678 bool "Check for stack overflows"
679 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
681 Say Y here if you want to check for overflows of kernel, IRQ
682 and exception stacks (if your architecture uses them). This
683 option will show detailed messages if free stack space drops
684 below a certain limit.
686 These kinds of bugs usually occur when call-chains in the
687 kernel get too deep, especially when interrupts are
690 Use this in cases where you see apparently random memory
691 corruption, especially if it appears in 'struct thread_info'
693 If in doubt, say "N".
695 source "lib/Kconfig.kmemcheck"
697 source "lib/Kconfig.kasan"
699 endmenu # "Memory Debugging"
704 KCOV does not have any arch-specific code, but currently it is enabled
705 only for x86_64. KCOV requires testing on other archs, and most likely
706 disabling of instrumentation for some early boot code.
709 bool "Code coverage for fuzzing"
710 depends on ARCH_HAS_KCOV
712 select GCC_PLUGINS if !COMPILE_TEST
713 select GCC_PLUGIN_SANCOV if !COMPILE_TEST
715 KCOV exposes kernel code coverage information in a form suitable
716 for coverage-guided fuzzing (randomized testing).
718 If RANDOMIZE_BASE is enabled, PC values will not be stable across
719 different machines and across reboots. If you need stable PC values,
720 disable RANDOMIZE_BASE.
722 For more details, see Documentation/kcov.txt.
724 config KCOV_INSTRUMENT_ALL
725 bool "Instrument all code by default"
729 If you are doing generic system call fuzzing (like e.g. syzkaller),
730 then you will want to instrument the whole kernel and you should
731 say y here. If you are doing more targeted fuzzing (like e.g.
732 filesystem fuzzing with AFL) then you will want to enable coverage
733 for more specific subsets of files, and should say n here.
736 bool "Debug shared IRQ handlers"
737 depends on DEBUG_KERNEL
739 Enable this to generate a spurious interrupt as soon as a shared
740 interrupt handler is registered, and just before one is deregistered.
741 Drivers ought to be able to handle interrupts coming in at those
742 points; some don't and need to be caught.
744 menu "Debug Lockups and Hangs"
746 config LOCKUP_DETECTOR
747 bool "Detect Hard and Soft Lockups"
748 depends on DEBUG_KERNEL && !S390
750 Say Y here to enable the kernel to act as a watchdog to detect
751 hard and soft lockups.
753 Softlockups are bugs that cause the kernel to loop in kernel
754 mode for more than 20 seconds, without giving other tasks a
755 chance to run. The current stack trace is displayed upon
756 detection and the system will stay locked up.
758 Hardlockups are bugs that cause the CPU to loop in kernel mode
759 for more than 10 seconds, without letting other interrupts have a
760 chance to run. The current stack trace is displayed upon detection
761 and the system will stay locked up.
763 The overhead should be minimal. A periodic hrtimer runs to
764 generate interrupts and kick the watchdog task every 4 seconds.
765 An NMI is generated every 10 seconds or so to check for hardlockups.
767 The frequency of hrtimer and NMI events and the soft and hard lockup
768 thresholds can be controlled through the sysctl watchdog_thresh.
770 config HARDLOCKUP_DETECTOR
772 depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG
773 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI
775 config BOOTPARAM_HARDLOCKUP_PANIC
776 bool "Panic (Reboot) On Hard Lockups"
777 depends on HARDLOCKUP_DETECTOR
779 Say Y here to enable the kernel to panic on "hard lockups",
780 which are bugs that cause the kernel to loop in kernel
781 mode with interrupts disabled for more than 10 seconds (configurable
782 using the watchdog_thresh sysctl).
786 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
788 depends on HARDLOCKUP_DETECTOR
790 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
791 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
793 config BOOTPARAM_SOFTLOCKUP_PANIC
794 bool "Panic (Reboot) On Soft Lockups"
795 depends on LOCKUP_DETECTOR
797 Say Y here to enable the kernel to panic on "soft lockups",
798 which are bugs that cause the kernel to loop in kernel
799 mode for more than 20 seconds (configurable using the watchdog_thresh
800 sysctl), without giving other tasks a chance to run.
802 The panic can be used in combination with panic_timeout,
803 to cause the system to reboot automatically after a
804 lockup has been detected. This feature is useful for
805 high-availability systems that have uptime guarantees and
806 where a lockup must be resolved ASAP.
810 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
812 depends on LOCKUP_DETECTOR
814 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
815 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
817 config DETECT_HUNG_TASK
818 bool "Detect Hung Tasks"
819 depends on DEBUG_KERNEL
820 default LOCKUP_DETECTOR
822 Say Y here to enable the kernel to detect "hung tasks",
823 which are bugs that cause the task to be stuck in
824 uninterruptible "D" state indefinitely.
826 When a hung task is detected, the kernel will print the
827 current stack trace (which you should report), but the
828 task will stay in uninterruptible state. If lockdep is
829 enabled then all held locks will also be reported. This
830 feature has negligible overhead.
832 config DEFAULT_HUNG_TASK_TIMEOUT
833 int "Default timeout for hung task detection (in seconds)"
834 depends on DETECT_HUNG_TASK
837 This option controls the default timeout (in seconds) used
838 to determine when a task has become non-responsive and should
841 It can be adjusted at runtime via the kernel.hung_task_timeout_secs
842 sysctl or by writing a value to
843 /proc/sys/kernel/hung_task_timeout_secs.
845 A timeout of 0 disables the check. The default is two minutes.
846 Keeping the default should be fine in most cases.
848 config BOOTPARAM_HUNG_TASK_PANIC
849 bool "Panic (Reboot) On Hung Tasks"
850 depends on DETECT_HUNG_TASK
852 Say Y here to enable the kernel to panic on "hung tasks",
853 which are bugs that cause the kernel to leave a task stuck
854 in uninterruptible "D" state.
856 The panic can be used in combination with panic_timeout,
857 to cause the system to reboot automatically after a
858 hung task has been detected. This feature is useful for
859 high-availability systems that have uptime guarantees and
860 where a hung tasks must be resolved ASAP.
864 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
866 depends on DETECT_HUNG_TASK
868 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
869 default 1 if BOOTPARAM_HUNG_TASK_PANIC
872 bool "Detect Workqueue Stalls"
873 depends on DEBUG_KERNEL
875 Say Y here to enable stall detection on workqueues. If a
876 worker pool doesn't make forward progress on a pending work
877 item for over a given amount of time, 30s by default, a
878 warning message is printed along with dump of workqueue
879 state. This can be configured through kernel parameter
880 "workqueue.watchdog_thresh" and its sysfs counterpart.
882 endmenu # "Debug lockups and hangs"
887 Say Y here to enable the kernel to panic when it oopses. This
888 has the same effect as setting oops=panic on the kernel command
891 This feature is useful to ensure that the kernel does not do
892 anything erroneous after an oops which could result in data
893 corruption or other issues.
897 config PANIC_ON_OOPS_VALUE
900 default 0 if !PANIC_ON_OOPS
901 default 1 if PANIC_ON_OOPS
907 Set the timeout value (in seconds) until a reboot occurs when the
908 the kernel panics. If n = 0, then we wait forever. A timeout
909 value n > 0 will wait n seconds before rebooting, while a timeout
910 value n < 0 will reboot immediately.
913 bool "Collect scheduler debugging info"
914 depends on DEBUG_KERNEL && PROC_FS
917 If you say Y here, the /proc/sched_debug file will be provided
918 that can help debug the scheduler. The runtime overhead of this
926 bool "Collect scheduler statistics"
927 depends on DEBUG_KERNEL && PROC_FS
930 If you say Y here, additional code will be inserted into the
931 scheduler and related routines to collect statistics about
932 scheduler behavior and provide them in /proc/schedstat. These
933 stats may be useful for both tuning and debugging the scheduler
934 If you aren't debugging the scheduler or trying to tune a specific
935 application, you can say N to avoid the very slight overhead
938 config SCHED_STACK_END_CHECK
939 bool "Detect stack corruption on calls to schedule()"
940 depends on DEBUG_KERNEL
943 This option checks for a stack overrun on calls to schedule().
944 If the stack end location is found to be over written always panic as
945 the content of the corrupted region can no longer be trusted.
946 This is to ensure no erroneous behaviour occurs which could result in
947 data corruption or a sporadic crash at a later stage once the region
948 is examined. The runtime overhead introduced is minimal.
950 config DEBUG_TIMEKEEPING
951 bool "Enable extra timekeeping sanity checking"
953 This option will enable additional timekeeping sanity checks
954 which may be helpful when diagnosing issues where timekeeping
955 problems are suspected.
957 This may include checks in the timekeeping hotpaths, so this
958 option may have a (very small) performance impact to some
964 bool "Collect kernel timers statistics"
965 depends on DEBUG_KERNEL && PROC_FS
967 If you say Y here, additional code will be inserted into the
968 timer routines to collect statistics about kernel timers being
969 reprogrammed. The statistics can be read from /proc/timer_stats.
970 The statistics collection is started by writing 1 to /proc/timer_stats,
971 writing 0 stops it. This feature is useful to collect information
972 about timer usage patterns in kernel and userspace. This feature
973 is lightweight if enabled in the kernel config but not activated
974 (it defaults to deactivated on bootup and will only be activated
975 if some application like powertop activates it explicitly).
978 bool "Debug preemptible kernel"
979 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
982 If you say Y here then the kernel will use a debug variant of the
983 commonly used smp_processor_id() function and will print warnings
984 if kernel code uses it in a preemption-unsafe way. Also, the kernel
985 will detect preemption count underflows.
987 menu "Lock Debugging (spinlocks, mutexes, etc...)"
989 config DEBUG_RT_MUTEXES
990 bool "RT Mutex debugging, deadlock detection"
991 depends on DEBUG_KERNEL && RT_MUTEXES
993 This allows rt mutex semantics violations and rt mutex related
994 deadlocks (lockups) to be detected and reported automatically.
996 config DEBUG_SPINLOCK
997 bool "Spinlock and rw-lock debugging: basic checks"
998 depends on DEBUG_KERNEL
999 select UNINLINE_SPIN_UNLOCK
1001 Say Y here and build SMP to catch missing spinlock initialization
1002 and certain other kinds of spinlock errors commonly made. This is
1003 best used in conjunction with the NMI watchdog so that spinlock
1004 deadlocks are also debuggable.
1006 config DEBUG_MUTEXES
1007 bool "Mutex debugging: basic checks"
1008 depends on DEBUG_KERNEL
1010 This feature allows mutex semantics violations to be detected and
1013 config DEBUG_WW_MUTEX_SLOWPATH
1014 bool "Wait/wound mutex debugging: Slowpath testing"
1015 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1016 select DEBUG_LOCK_ALLOC
1017 select DEBUG_SPINLOCK
1018 select DEBUG_MUTEXES
1020 This feature enables slowpath testing for w/w mutex users by
1021 injecting additional -EDEADLK wound/backoff cases. Together with
1022 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this
1023 will test all possible w/w mutex interface abuse with the
1024 exception of simply not acquiring all the required locks.
1025 Note that this feature can introduce significant overhead, so
1026 it really should not be enabled in a production or distro kernel,
1027 even a debug kernel. If you are a driver writer, enable it. If
1028 you are a distro, do not.
1030 config DEBUG_LOCK_ALLOC
1031 bool "Lock debugging: detect incorrect freeing of live locks"
1032 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1033 select DEBUG_SPINLOCK
1034 select DEBUG_MUTEXES
1037 This feature will check whether any held lock (spinlock, rwlock,
1038 mutex or rwsem) is incorrectly freed by the kernel, via any of the
1039 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
1040 vfree(), etc.), whether a live lock is incorrectly reinitialized via
1041 spin_lock_init()/mutex_init()/etc., or whether there is any lock
1042 held during task exit.
1044 config PROVE_LOCKING
1045 bool "Lock debugging: prove locking correctness"
1046 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1048 select DEBUG_SPINLOCK
1049 select DEBUG_MUTEXES
1050 select DEBUG_LOCK_ALLOC
1051 select TRACE_IRQFLAGS
1054 This feature enables the kernel to prove that all locking
1055 that occurs in the kernel runtime is mathematically
1056 correct: that under no circumstance could an arbitrary (and
1057 not yet triggered) combination of observed locking
1058 sequences (on an arbitrary number of CPUs, running an
1059 arbitrary number of tasks and interrupt contexts) cause a
1062 In short, this feature enables the kernel to report locking
1063 related deadlocks before they actually occur.
1065 The proof does not depend on how hard and complex a
1066 deadlock scenario would be to trigger: how many
1067 participant CPUs, tasks and irq-contexts would be needed
1068 for it to trigger. The proof also does not depend on
1069 timing: if a race and a resulting deadlock is possible
1070 theoretically (no matter how unlikely the race scenario
1071 is), it will be proven so and will immediately be
1072 reported by the kernel (once the event is observed that
1073 makes the deadlock theoretically possible).
1075 If a deadlock is impossible (i.e. the locking rules, as
1076 observed by the kernel, are mathematically correct), the
1077 kernel reports nothing.
1079 NOTE: this feature can also be enabled for rwlocks, mutexes
1080 and rwsems - in which case all dependencies between these
1081 different locking variants are observed and mapped too, and
1082 the proof of observed correctness is also maintained for an
1083 arbitrary combination of these separate locking variants.
1085 For more details, see Documentation/locking/lockdep-design.txt.
1087 config PROVE_LOCKING_SMALL
1092 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1094 select FRAME_POINTER if !MIPS && !PPC && !ARM && !S390 && !MICROBLAZE && !ARC && !SCORE
1099 bool "Lock usage statistics"
1100 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1102 select DEBUG_SPINLOCK
1103 select DEBUG_MUTEXES
1104 select DEBUG_LOCK_ALLOC
1107 This feature enables tracking lock contention points
1109 For more details, see Documentation/locking/lockstat.txt
1111 This also enables lock events required by "perf lock",
1113 If you want to use "perf lock", you also need to turn on
1114 CONFIG_EVENT_TRACING.
1116 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
1117 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
1119 config DEBUG_LOCKDEP
1120 bool "Lock dependency engine debugging"
1121 depends on DEBUG_KERNEL && LOCKDEP
1123 If you say Y here, the lock dependency engine will do
1124 additional runtime checks to debug itself, at the price
1125 of more runtime overhead.
1127 config DEBUG_ATOMIC_SLEEP
1128 bool "Sleep inside atomic section checking"
1129 select PREEMPT_COUNT
1130 depends on DEBUG_KERNEL
1132 If you say Y here, various routines which may sleep will become very
1133 noisy if they are called inside atomic sections: when a spinlock is
1134 held, inside an rcu read side critical section, inside preempt disabled
1135 sections, inside an interrupt, etc...
1137 config DEBUG_LOCKING_API_SELFTESTS
1138 bool "Locking API boot-time self-tests"
1139 depends on DEBUG_KERNEL
1141 Say Y here if you want the kernel to run a short self-test during
1142 bootup. The self-test checks whether common types of locking bugs
1143 are detected by debugging mechanisms or not. (if you disable
1144 lock debugging then those bugs wont be detected of course.)
1145 The following locking APIs are covered: spinlocks, rwlocks,
1148 config LOCK_TORTURE_TEST
1149 tristate "torture tests for locking"
1150 depends on DEBUG_KERNEL
1154 This option provides a kernel module that runs torture tests
1155 on kernel locking primitives. The kernel module may be built
1156 after the fact on the running kernel to be tested, if desired.
1158 Say Y here if you want kernel locking-primitive torture tests
1159 to be built into the kernel.
1160 Say M if you want these torture tests to build as a module.
1161 Say N if you are unsure.
1163 endmenu # lock debugging
1165 config TRACE_IRQFLAGS
1168 Enables hooks to interrupt enabling and disabling for
1169 either tracing or lock debugging.
1172 bool "Stack backtrace support"
1173 depends on STACKTRACE_SUPPORT
1175 This option causes the kernel to create a /proc/pid/stack for
1176 every process, showing its current stack trace.
1177 It is also used by various kernel debugging features that require
1178 stack trace generation.
1180 config WARN_ALL_UNSEEDED_RANDOM
1181 bool "Warn for all uses of unseeded randomness"
1184 Some parts of the kernel contain bugs relating to their use of
1185 cryptographically secure random numbers before it's actually possible
1186 to generate those numbers securely. This setting ensures that these
1187 flaws don't go unnoticed, by enabling a message, should this ever
1188 occur. This will allow people with obscure setups to know when things
1189 are going wrong, so that they might contact developers about fixing
1192 Unfortunately, on some models of some architectures getting
1193 a fully seeded CRNG is extremely difficult, and so this can
1194 result in dmesg getting spammed for a surprisingly long
1195 time. This is really bad from a security perspective, and
1196 so architecture maintainers really need to do what they can
1197 to get the CRNG seeded sooner after the system is booted.
1198 However, since users cannot do anything actionable to
1199 address this, by default this option is disabled.
1201 Say Y here if you want to receive warnings for all uses of
1202 unseeded randomness. This will be of use primarily for
1203 those developers interersted in improving the security of
1204 Linux kernels running on their architecture (or
1207 config DEBUG_KOBJECT
1208 bool "kobject debugging"
1209 depends on DEBUG_KERNEL
1211 If you say Y here, some extra kobject debugging messages will be sent
1214 config DEBUG_KOBJECT_RELEASE
1215 bool "kobject release debugging"
1216 depends on DEBUG_OBJECTS_TIMERS
1218 kobjects are reference counted objects. This means that their
1219 last reference count put is not predictable, and the kobject can
1220 live on past the point at which a driver decides to drop it's
1221 initial reference to the kobject gained on allocation. An
1222 example of this would be a struct device which has just been
1225 However, some buggy drivers assume that after such an operation,
1226 the memory backing the kobject can be immediately freed. This
1227 goes completely against the principles of a refcounted object.
1229 If you say Y here, the kernel will delay the release of kobjects
1230 on the last reference count to improve the visibility of this
1231 kind of kobject release bug.
1233 config HAVE_DEBUG_BUGVERBOSE
1236 config DEBUG_BUGVERBOSE
1237 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
1238 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
1241 Say Y here to make BUG() panics output the file name and line number
1242 of the BUG call as well as the EIP and oops trace. This aids
1243 debugging but costs about 70-100K of memory.
1246 bool "Debug linked list manipulation"
1247 depends on DEBUG_KERNEL
1249 Enable this to turn on extended checks in the linked-list
1254 config DEBUG_PI_LIST
1255 bool "Debug priority linked list manipulation"
1256 depends on DEBUG_KERNEL
1258 Enable this to turn on extended checks in the priority-ordered
1259 linked-list (plist) walking routines. This checks the entire
1260 list multiple times during each manipulation.
1265 bool "Debug SG table operations"
1266 depends on DEBUG_KERNEL
1268 Enable this to turn on checks on scatter-gather tables. This can
1269 help find problems with drivers that do not properly initialize
1274 config DEBUG_NOTIFIERS
1275 bool "Debug notifier call chains"
1276 depends on DEBUG_KERNEL
1278 Enable this to turn on sanity checking for notifier call chains.
1279 This is most useful for kernel developers to make sure that
1280 modules properly unregister themselves from notifier chains.
1281 This is a relatively cheap check but if you care about maximum
1284 config DEBUG_CREDENTIALS
1285 bool "Debug credential management"
1286 depends on DEBUG_KERNEL
1288 Enable this to turn on some debug checking for credential
1289 management. The additional code keeps track of the number of
1290 pointers from task_structs to any given cred struct, and checks to
1291 see that this number never exceeds the usage count of the cred
1294 Furthermore, if SELinux is enabled, this also checks that the
1295 security pointer in the cred struct is never seen to be invalid.
1299 menu "RCU Debugging"
1302 def_bool PROVE_LOCKING
1304 config PROVE_RCU_REPEATEDLY
1305 bool "RCU debugging: don't disable PROVE_RCU on first splat"
1306 depends on PROVE_RCU
1309 By itself, PROVE_RCU will disable checking upon issuing the
1310 first warning (or "splat"). This feature prevents such
1311 disabling, allowing multiple RCU-lockdep warnings to be printed
1314 Say Y to allow multiple RCU-lockdep warnings per boot.
1316 Say N if you are unsure.
1318 config SPARSE_RCU_POINTER
1319 bool "RCU debugging: sparse-based checks for pointer usage"
1322 This feature enables the __rcu sparse annotation for
1323 RCU-protected pointers. This annotation will cause sparse
1324 to flag any non-RCU used of annotated pointers. This can be
1325 helpful when debugging RCU usage. Please note that this feature
1326 is not intended to enforce code cleanliness; it is instead merely
1329 Say Y to make sparse flag questionable use of RCU-protected pointers
1331 Say N if you are unsure.
1337 config RCU_PERF_TEST
1338 tristate "performance tests for RCU"
1339 depends on DEBUG_KERNEL
1345 This option provides a kernel module that runs performance
1346 tests on the RCU infrastructure. The kernel module may be built
1347 after the fact on the running kernel to be tested, if desired.
1349 Say Y here if you want RCU performance tests to be built into
1351 Say M if you want the RCU performance tests to build as a module.
1352 Say N if you are unsure.
1354 config RCU_TORTURE_TEST
1355 tristate "torture tests for RCU"
1356 depends on DEBUG_KERNEL
1362 This option provides a kernel module that runs torture tests
1363 on the RCU infrastructure. The kernel module may be built
1364 after the fact on the running kernel to be tested, if desired.
1366 Say Y here if you want RCU torture tests to be built into
1368 Say M if you want the RCU torture tests to build as a module.
1369 Say N if you are unsure.
1371 config RCU_TORTURE_TEST_SLOW_PREINIT
1372 bool "Slow down RCU grace-period pre-initialization to expose races"
1373 depends on RCU_TORTURE_TEST
1375 This option delays grace-period pre-initialization (the
1376 propagation of CPU-hotplug changes up the rcu_node combining
1377 tree) for a few jiffies between initializing each pair of
1378 consecutive rcu_node structures. This helps to expose races
1379 involving grace-period pre-initialization, in other words, it
1380 makes your kernel less stable. It can also greatly increase
1381 grace-period latency, especially on systems with large numbers
1382 of CPUs. This is useful when torture-testing RCU, but in
1383 almost no other circumstance.
1385 Say Y here if you want your system to crash and hang more often.
1386 Say N if you want a sane system.
1388 config RCU_TORTURE_TEST_SLOW_PREINIT_DELAY
1389 int "How much to slow down RCU grace-period pre-initialization"
1392 depends on RCU_TORTURE_TEST_SLOW_PREINIT
1394 This option specifies the number of jiffies to wait between
1395 each rcu_node structure pre-initialization step.
1397 config RCU_TORTURE_TEST_SLOW_INIT
1398 bool "Slow down RCU grace-period initialization to expose races"
1399 depends on RCU_TORTURE_TEST
1401 This option delays grace-period initialization for a few
1402 jiffies between initializing each pair of consecutive
1403 rcu_node structures. This helps to expose races involving
1404 grace-period initialization, in other words, it makes your
1405 kernel less stable. It can also greatly increase grace-period
1406 latency, especially on systems with large numbers of CPUs.
1407 This is useful when torture-testing RCU, but in almost no
1410 Say Y here if you want your system to crash and hang more often.
1411 Say N if you want a sane system.
1413 config RCU_TORTURE_TEST_SLOW_INIT_DELAY
1414 int "How much to slow down RCU grace-period initialization"
1417 depends on RCU_TORTURE_TEST_SLOW_INIT
1419 This option specifies the number of jiffies to wait between
1420 each rcu_node structure initialization.
1422 config RCU_TORTURE_TEST_SLOW_CLEANUP
1423 bool "Slow down RCU grace-period cleanup to expose races"
1424 depends on RCU_TORTURE_TEST
1426 This option delays grace-period cleanup for a few jiffies
1427 between cleaning up each pair of consecutive rcu_node
1428 structures. This helps to expose races involving grace-period
1429 cleanup, in other words, it makes your kernel less stable.
1430 It can also greatly increase grace-period latency, especially
1431 on systems with large numbers of CPUs. This is useful when
1432 torture-testing RCU, but in almost no other circumstance.
1434 Say Y here if you want your system to crash and hang more often.
1435 Say N if you want a sane system.
1437 config RCU_TORTURE_TEST_SLOW_CLEANUP_DELAY
1438 int "How much to slow down RCU grace-period cleanup"
1441 depends on RCU_TORTURE_TEST_SLOW_CLEANUP
1443 This option specifies the number of jiffies to wait between
1444 each rcu_node structure cleanup operation.
1446 config RCU_CPU_STALL_TIMEOUT
1447 int "RCU CPU stall timeout in seconds"
1448 depends on RCU_STALL_COMMON
1452 If a given RCU grace period extends more than the specified
1453 number of seconds, a CPU stall warning is printed. If the
1454 RCU grace period persists, additional CPU stall warnings are
1455 printed at more widely spaced intervals.
1458 bool "Enable tracing for RCU"
1459 depends on DEBUG_KERNEL
1462 This option provides tracing in RCU which presents stats
1463 in debugfs for debugging RCU implementation.
1465 Say Y here if you want to enable RCU tracing
1466 Say N if you are unsure.
1468 config RCU_EQS_DEBUG
1469 bool "Provide debugging asserts for adding NO_HZ support to an arch"
1470 depends on DEBUG_KERNEL
1472 This option provides consistency checks in RCU's handling of
1473 NO_HZ. These checks have proven quite helpful in detecting
1474 bugs in arch-specific NO_HZ code.
1476 Say N here if you need ultimate kernel/user switch latencies
1477 Say Y if you are unsure
1479 endmenu # "RCU Debugging"
1481 config DEBUG_WQ_FORCE_RR_CPU
1482 bool "Force round-robin CPU selection for unbound work items"
1483 depends on DEBUG_KERNEL
1486 Workqueue used to implicitly guarantee that work items queued
1487 without explicit CPU specified are put on the local CPU. This
1488 guarantee is no longer true and while local CPU is still
1489 preferred work items may be put on foreign CPUs. Kernel
1490 parameter "workqueue.debug_force_rr_cpu" is added to force
1491 round-robin CPU selection to flush out usages which depend on the
1492 now broken guarantee. This config option enables the debug
1493 feature by default. When enabled, memory and cache locality will
1496 config DEBUG_BLOCK_EXT_DEVT
1497 bool "Force extended block device numbers and spread them"
1498 depends on DEBUG_KERNEL
1502 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1503 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1504 YOU ARE DOING. Distros, please enable this and fix whatever
1507 Conventionally, block device numbers are allocated from
1508 predetermined contiguous area. However, extended block area
1509 may introduce non-contiguous block device numbers. This
1510 option forces most block device numbers to be allocated from
1511 the extended space and spreads them to discover kernel or
1512 userland code paths which assume predetermined contiguous
1513 device number allocation.
1515 Note that turning on this debug option shuffles all the
1516 device numbers for all IDE and SCSI devices including libata
1517 ones, so root partition specified using device number
1518 directly (via rdev or root=MAJ:MIN) won't work anymore.
1519 Textual device names (root=/dev/sdXn) will continue to work.
1521 Say N if you are unsure.
1523 config CPU_HOTPLUG_STATE_CONTROL
1524 bool "Enable CPU hotplug state control"
1525 depends on DEBUG_KERNEL
1526 depends on HOTPLUG_CPU
1529 Allows to write steps between "offline" and "online" to the CPUs
1530 sysfs target file so states can be stepped granular. This is a debug
1531 option for now as the hotplug machinery cannot be stopped and
1532 restarted at arbitrary points yet.
1534 Say N if your are unsure.
1536 config NOTIFIER_ERROR_INJECTION
1537 tristate "Notifier error injection"
1538 depends on DEBUG_KERNEL
1541 This option provides the ability to inject artificial errors to
1542 specified notifier chain callbacks. It is useful to test the error
1543 handling of notifier call chain failures.
1547 config CPU_NOTIFIER_ERROR_INJECT
1548 tristate "CPU notifier error injection module"
1549 depends on HOTPLUG_CPU && NOTIFIER_ERROR_INJECTION
1551 This option provides a kernel module that can be used to test
1552 the error handling of the cpu notifiers by injecting artificial
1553 errors to CPU notifier chain callbacks. It is controlled through
1554 debugfs interface under /sys/kernel/debug/notifier-error-inject/cpu
1556 If the notifier call chain should be failed with some events
1557 notified, write the error code to "actions/<notifier event>/error".
1559 Example: Inject CPU offline error (-1 == -EPERM)
1561 # cd /sys/kernel/debug/notifier-error-inject/cpu
1562 # echo -1 > actions/CPU_DOWN_PREPARE/error
1563 # echo 0 > /sys/devices/system/cpu/cpu1/online
1564 bash: echo: write error: Operation not permitted
1566 To compile this code as a module, choose M here: the module will
1567 be called cpu-notifier-error-inject.
1571 config PM_NOTIFIER_ERROR_INJECT
1572 tristate "PM notifier error injection module"
1573 depends on PM && NOTIFIER_ERROR_INJECTION
1574 default m if PM_DEBUG
1576 This option provides the ability to inject artificial errors to
1577 PM notifier chain callbacks. It is controlled through debugfs
1578 interface /sys/kernel/debug/notifier-error-inject/pm
1580 If the notifier call chain should be failed with some events
1581 notified, write the error code to "actions/<notifier event>/error".
1583 Example: Inject PM suspend error (-12 = -ENOMEM)
1585 # cd /sys/kernel/debug/notifier-error-inject/pm/
1586 # echo -12 > actions/PM_SUSPEND_PREPARE/error
1587 # echo mem > /sys/power/state
1588 bash: echo: write error: Cannot allocate memory
1590 To compile this code as a module, choose M here: the module will
1591 be called pm-notifier-error-inject.
1595 config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1596 tristate "OF reconfig notifier error injection module"
1597 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1599 This option provides the ability to inject artificial errors to
1600 OF reconfig notifier chain callbacks. It is controlled
1601 through debugfs interface under
1602 /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1604 If the notifier call chain should be failed with some events
1605 notified, write the error code to "actions/<notifier event>/error".
1607 To compile this code as a module, choose M here: the module will
1608 be called of-reconfig-notifier-error-inject.
1612 config NETDEV_NOTIFIER_ERROR_INJECT
1613 tristate "Netdev notifier error injection module"
1614 depends on NET && NOTIFIER_ERROR_INJECTION
1616 This option provides the ability to inject artificial errors to
1617 netdevice notifier chain callbacks. It is controlled through debugfs
1618 interface /sys/kernel/debug/notifier-error-inject/netdev
1620 If the notifier call chain should be failed with some events
1621 notified, write the error code to "actions/<notifier event>/error".
1623 Example: Inject netdevice mtu change error (-22 = -EINVAL)
1625 # cd /sys/kernel/debug/notifier-error-inject/netdev
1626 # echo -22 > actions/NETDEV_CHANGEMTU/error
1627 # ip link set eth0 mtu 1024
1628 RTNETLINK answers: Invalid argument
1630 To compile this code as a module, choose M here: the module will
1631 be called netdev-notifier-error-inject.
1635 config FAULT_INJECTION
1636 bool "Fault-injection framework"
1637 depends on DEBUG_KERNEL
1639 Provide fault-injection framework.
1640 For more details, see Documentation/fault-injection/.
1643 bool "Fault-injection capability for kmalloc"
1644 depends on FAULT_INJECTION
1645 depends on SLAB || SLUB
1647 Provide fault-injection capability for kmalloc.
1649 config FAIL_PAGE_ALLOC
1650 bool "Fault-injection capabilitiy for alloc_pages()"
1651 depends on FAULT_INJECTION
1653 Provide fault-injection capability for alloc_pages().
1655 config FAIL_MAKE_REQUEST
1656 bool "Fault-injection capability for disk IO"
1657 depends on FAULT_INJECTION && BLOCK
1659 Provide fault-injection capability for disk IO.
1661 config FAIL_IO_TIMEOUT
1662 bool "Fault-injection capability for faking disk interrupts"
1663 depends on FAULT_INJECTION && BLOCK
1665 Provide fault-injection capability on end IO handling. This
1666 will make the block layer "forget" an interrupt as configured,
1667 thus exercising the error handling.
1669 Only works with drivers that use the generic timeout handling,
1670 for others it wont do anything.
1672 config FAIL_MMC_REQUEST
1673 bool "Fault-injection capability for MMC IO"
1674 depends on FAULT_INJECTION_DEBUG_FS && MMC
1676 Provide fault-injection capability for MMC IO.
1677 This will make the mmc core return data errors. This is
1678 useful to test the error handling in the mmc block device
1679 and to test how the mmc host driver handles retries from
1683 bool "Fault-injection capability for futexes"
1685 depends on FAULT_INJECTION && FUTEX
1687 Provide fault-injection capability for futexes.
1689 config FAULT_INJECTION_DEBUG_FS
1690 bool "Debugfs entries for fault-injection capabilities"
1691 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1693 Enable configuration of fault-injection capabilities via debugfs.
1695 config FAULT_INJECTION_STACKTRACE_FILTER
1696 bool "stacktrace filter for fault-injection capabilities"
1697 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1700 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM && !ARC && !SCORE
1702 Provide stacktrace filter for fault-injection capabilities
1705 bool "Latency measuring infrastructure"
1706 depends on DEBUG_KERNEL
1707 depends on STACKTRACE_SUPPORT
1709 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM && !ARC
1716 Enable this option if you want to use the LatencyTOP tool
1717 to find out which userspace is blocking on what kernel operations.
1719 source kernel/trace/Kconfig
1721 menu "Runtime Testing"
1724 tristate "Linux Kernel Dump Test Tool Module"
1729 This module enables testing of the different dumping mechanisms by
1730 inducing system failures at predefined crash points.
1731 If you don't need it: say N
1732 Choose M here to compile this code as a module. The module will be
1735 Documentation on how to use the module can be found in
1736 Documentation/fault-injection/provoke-crashes.txt
1738 config TEST_LIST_SORT
1739 bool "Linked list sorting test"
1740 depends on DEBUG_KERNEL
1742 Enable this to turn on 'list_sort()' function test. This test is
1743 executed only once during system boot, so affects only boot time.
1747 config KPROBES_SANITY_TEST
1748 bool "Kprobes sanity tests"
1749 depends on DEBUG_KERNEL
1753 This option provides for testing basic kprobes functionality on
1754 boot. A sample kprobe, jprobe and kretprobe are inserted and
1755 verified for functionality.
1757 Say N if you are unsure.
1759 config BACKTRACE_SELF_TEST
1760 tristate "Self test for the backtrace code"
1761 depends on DEBUG_KERNEL
1764 This option provides a kernel module that can be used to test
1765 the kernel stack backtrace code. This option is not useful
1766 for distributions or general kernels, but only for kernel
1767 developers working on architecture code.
1769 Note that if you want to also test saved backtraces, you will
1770 have to enable STACKTRACE as well.
1772 Say N if you are unsure.
1775 tristate "Red-Black tree test"
1776 depends on DEBUG_KERNEL
1778 A benchmark measuring the performance of the rbtree library.
1779 Also includes rbtree invariant checks.
1781 config INTERVAL_TREE_TEST
1782 tristate "Interval tree test"
1783 depends on m && DEBUG_KERNEL
1784 select INTERVAL_TREE
1786 A benchmark measuring the performance of the interval tree library
1789 tristate "Per cpu operations test"
1790 depends on m && DEBUG_KERNEL
1792 Enable this option to build test module which validates per-cpu
1797 config ATOMIC64_SELFTEST
1798 bool "Perform an atomic64_t self-test at boot"
1800 Enable this option to test the atomic64_t functions at boot.
1804 config ASYNC_RAID6_TEST
1805 tristate "Self test for hardware accelerated raid6 recovery"
1806 depends on ASYNC_RAID6_RECOV
1809 This is a one-shot self test that permutes through the
1810 recovery of all the possible two disk failure scenarios for a
1811 N-disk array. Recovery is performed with the asynchronous
1812 raid6 recovery routines, and will optionally use an offload
1813 engine if one is available.
1818 tristate "Test functions located in the hexdump module at runtime"
1820 config TEST_STRING_HELPERS
1821 tristate "Test functions located in the string_helpers module at runtime"
1824 tristate "Test kstrto*() family of functions at runtime"
1827 tristate "Test printf() family of functions at runtime"
1830 tristate "Test bitmap_*() family of functions at runtime"
1833 Enable this option to test the bitmap functions at boot.
1838 tristate "Test functions located in the uuid module at runtime"
1840 config TEST_RHASHTABLE
1841 tristate "Perform selftest on resizable hash table"
1844 Enable this option to test the rhashtable functions at boot.
1849 tristate "Perform selftest on hash functions"
1852 Enable this option to test the kernel's integer (<linux/hash.h>),
1853 string (<linux/stringhash.h>), and siphash (<linux/siphash.h>)
1854 hash functions on boot (or module load).
1856 This is intended to help people writing architecture-specific
1857 optimized versions. If unsure, say N.
1859 endmenu # runtime tests
1861 config PROVIDE_OHCI1394_DMA_INIT
1862 bool "Remote debugging over FireWire early on boot"
1863 depends on PCI && X86
1865 If you want to debug problems which hang or crash the kernel early
1866 on boot and the crashing machine has a FireWire port, you can use
1867 this feature to remotely access the memory of the crashed machine
1868 over FireWire. This employs remote DMA as part of the OHCI1394
1869 specification which is now the standard for FireWire controllers.
1871 With remote DMA, you can monitor the printk buffer remotely using
1872 firescope and access all memory below 4GB using fireproxy from gdb.
1873 Even controlling a kernel debugger is possible using remote DMA.
1877 If ohci1394_dma=early is used as boot parameter, it will initialize
1878 all OHCI1394 controllers which are found in the PCI config space.
1880 As all changes to the FireWire bus such as enabling and disabling
1881 devices cause a bus reset and thereby disable remote DMA for all
1882 devices, be sure to have the cable plugged and FireWire enabled on
1883 the debugging host before booting the debug target for debugging.
1885 This code (~1k) is freed after boot. By then, the firewire stack
1886 in charge of the OHCI-1394 controllers should be used instead.
1888 See Documentation/debugging-via-ohci1394.txt for more information.
1890 config DMA_API_DEBUG
1891 bool "Enable debugging of DMA-API usage"
1892 depends on HAVE_DMA_API_DEBUG
1894 Enable this option to debug the use of the DMA API by device drivers.
1895 With this option you will be able to detect common bugs in device
1896 drivers like double-freeing of DMA mappings or freeing mappings that
1897 were never allocated.
1899 This also attempts to catch cases where a page owned by DMA is
1900 accessed by the cpu in a way that could cause data corruption. For
1901 example, this enables cow_user_page() to check that the source page is
1904 This option causes a performance degradation. Use only if you want to
1905 debug device drivers and dma interactions.
1910 tristate "Test module loading with 'hello world' module"
1914 This builds the "test_module" module that emits "Hello, world"
1915 on printk when loaded. It is designed to be used for basic
1916 evaluation of the module loading subsystem (for example when
1917 validating module verification). It lacks any extra dependencies,
1918 and will not normally be loaded by the system unless explicitly
1923 config TEST_USER_COPY
1924 tristate "Test user/kernel boundary protections"
1928 This builds the "test_user_copy" module that runs sanity checks
1929 on the copy_to/from_user infrastructure, making sure basic
1930 user/kernel boundary testing is working. If it fails to load,
1931 a regression has been detected in the user/kernel memory boundary
1937 tristate "Test BPF filter functionality"
1941 This builds the "test_bpf" module that runs various test vectors
1942 against the BPF interpreter or BPF JIT compiler depending on the
1943 current setting. This is in particular useful for BPF JIT compiler
1944 development, but also to run regression tests against changes in
1945 the interpreter code. It also enables test stubs for eBPF maps and
1946 verifier used by user space verifier testsuite.
1950 config TEST_FIRMWARE
1951 tristate "Test firmware loading via userspace interface"
1953 depends on FW_LOADER
1955 This builds the "test_firmware" module that creates a userspace
1956 interface for testing firmware loading. This can be used to
1957 control the triggering of firmware loading without needing an
1958 actual firmware-using device. The contents can be rechecked by
1964 tristate "udelay test driver"
1967 This builds the "udelay_test" module that helps to make sure
1968 that udelay() is working properly.
1974 depends on HAVE_MEMBLOCK
1976 This option adds a kernel parameter 'memtest', which allows memtest
1978 memtest=0, mean disabled; -- default
1979 memtest=1, mean do 1 test pattern;
1981 memtest=17, mean do 17 test patterns.
1982 If you are unsure how to answer this question, answer N.
1984 config TEST_STATIC_KEYS
1985 tristate "Test static keys"
1989 Test the static key interfaces.
1993 source "samples/Kconfig"
1995 source "lib/Kconfig.kgdb"
1997 source "lib/Kconfig.ubsan"
1999 config ARCH_HAS_DEVMEM_IS_ALLOWED
2002 config STRICT_DEVMEM
2003 bool "Filter access to /dev/mem"
2005 depends on ARCH_HAS_DEVMEM_IS_ALLOWED
2006 default y if TILE || PPC
2008 If this option is disabled, you allow userspace (root) access to all
2009 of memory, including kernel and userspace memory. Accidental
2010 access to this is obviously disastrous, but specific access can
2011 be used by people debugging the kernel. Note that with PAT support
2012 enabled, even in this case there are restrictions on /dev/mem
2013 use due to the cache aliasing requirements.
2015 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem
2016 file only allows userspace access to PCI space and the BIOS code and
2017 data regions. This is sufficient for dosemu and X and all common
2022 config IO_STRICT_DEVMEM
2023 bool "Filter I/O access to /dev/mem"
2024 depends on STRICT_DEVMEM
2026 If this option is disabled, you allow userspace (root) access to all
2027 io-memory regardless of whether a driver is actively using that
2028 range. Accidental access to this is obviously disastrous, but
2029 specific access can be used by people debugging kernel drivers.
2031 If this option is switched on, the /dev/mem file only allows
2032 userspace access to *idle* io-memory ranges (see /proc/iomem) This
2033 may break traditional users of /dev/mem (dosemu, legacy X, etc...)
2034 if the driver using a given range cannot be disabled.