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/admin-guide/kernel-parameters.rst
18 config CONSOLE_LOGLEVEL_DEFAULT
19 int "Default console loglevel (1-15)"
23 Default loglevel to determine what will be printed on the console.
25 Setting a default here is equivalent to passing in loglevel=<x> in
26 the kernel bootargs. loglevel=<x> continues to override whatever
27 value is specified here as well.
29 Note: This does not affect the log level of un-prefixed printk()
30 usage in the kernel. That is controlled by the MESSAGE_LOGLEVEL_DEFAULT
33 config CONSOLE_LOGLEVEL_QUIET
34 int "quiet console loglevel (1-15)"
38 loglevel to use when "quiet" is passed on the kernel commandline.
40 When "quiet" is passed on the kernel commandline this loglevel
41 will be used as the loglevel. IOW passing "quiet" will be the
42 equivalent of passing "loglevel=<CONSOLE_LOGLEVEL_QUIET>"
44 config MESSAGE_LOGLEVEL_DEFAULT
45 int "Default message log level (1-7)"
49 Default log level for printk statements with no specified priority.
51 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
52 that are auditing their logs closely may want to set it to a lower
55 Note: This does not affect what message level gets printed on the console
56 by default. To change that, use loglevel=<x> in the kernel bootargs,
57 or pick a different CONSOLE_LOGLEVEL_DEFAULT configuration value.
59 config BOOT_PRINTK_DELAY
60 bool "Delay each boot printk message by N milliseconds"
61 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
63 This build option allows you to read kernel boot messages
64 by inserting a short delay after each one. The delay is
65 specified in milliseconds on the kernel command line,
68 It is likely that you would also need to use "lpj=M" to preset
69 the "loops per jiffie" value.
70 See a previous boot log for the "lpj" value to use for your
71 system, and then set "lpj=M" before setting "boot_delay=N".
72 NOTE: Using this option may adversely affect SMP systems.
73 I.e., processors other than the first one may not boot up.
74 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
75 what it believes to be lockup conditions.
78 bool "Enable dynamic printk() support"
84 Compiles debug level messages into the kernel, which would not
85 otherwise be available at runtime. These messages can then be
86 enabled/disabled based on various levels of scope - per source file,
87 function, module, format string, and line number. This mechanism
88 implicitly compiles in all pr_debug() and dev_dbg() calls, which
89 enlarges the kernel text size by about 2%.
91 If a source file is compiled with DEBUG flag set, any
92 pr_debug() calls in it are enabled by default, but can be
93 disabled at runtime as below. Note that DEBUG flag is
94 turned on by many CONFIG_*DEBUG* options.
98 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
99 which is contained in the 'debugfs' filesystem. Thus, the debugfs
100 filesystem must first be mounted before making use of this feature.
101 We refer the control file as: <debugfs>/dynamic_debug/control. This
102 file contains a list of the debug statements that can be enabled. The
103 format for each line of the file is:
105 filename:lineno [module]function flags format
107 filename : source file of the debug statement
108 lineno : line number of the debug statement
109 module : module that contains the debug statement
110 function : function that contains the debug statement
111 flags : '=p' means the line is turned 'on' for printing
112 format : the format used for the debug statement
116 nullarbor:~ # cat <debugfs>/dynamic_debug/control
117 # filename:lineno [module]function flags format
118 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
119 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
120 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
124 // enable the message at line 1603 of file svcsock.c
125 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
126 <debugfs>/dynamic_debug/control
128 // enable all the messages in file svcsock.c
129 nullarbor:~ # echo -n 'file svcsock.c +p' >
130 <debugfs>/dynamic_debug/control
132 // enable all the messages in the NFS server module
133 nullarbor:~ # echo -n 'module nfsd +p' >
134 <debugfs>/dynamic_debug/control
136 // enable all 12 messages in the function svc_process()
137 nullarbor:~ # echo -n 'func svc_process +p' >
138 <debugfs>/dynamic_debug/control
140 // disable all 12 messages in the function svc_process()
141 nullarbor:~ # echo -n 'func svc_process -p' >
142 <debugfs>/dynamic_debug/control
144 See Documentation/admin-guide/dynamic-debug-howto.rst for additional
147 endmenu # "printk and dmesg options"
149 menu "Compile-time checks and compiler options"
152 bool "Compile the kernel with debug info"
153 depends on DEBUG_KERNEL && !COMPILE_TEST
155 If you say Y here the resulting kernel image will include
156 debugging info resulting in a larger kernel image.
157 This adds debug symbols to the kernel and modules (gcc -g), and
158 is needed if you intend to use kernel crashdump or binary object
159 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
160 Say Y here only if you plan to debug the kernel.
164 config DEBUG_INFO_REDUCED
165 bool "Reduce debugging information"
166 depends on DEBUG_INFO
168 If you say Y here gcc is instructed to generate less debugging
169 information for structure types. This means that tools that
170 need full debugging information (like kgdb or systemtap) won't
171 be happy. But if you merely need debugging information to
172 resolve line numbers there is no loss. Advantage is that
173 build directory object sizes shrink dramatically over a full
174 DEBUG_INFO build and compile times are reduced too.
175 Only works with newer gcc versions.
177 config DEBUG_INFO_SPLIT
178 bool "Produce split debuginfo in .dwo files"
179 depends on DEBUG_INFO
181 Generate debug info into separate .dwo files. This significantly
182 reduces the build directory size for builds with DEBUG_INFO,
183 because it stores the information only once on disk in .dwo
184 files instead of multiple times in object files and executables.
185 In addition the debug information is also compressed.
187 Requires recent gcc (4.7+) and recent gdb/binutils.
188 Any tool that packages or reads debug information would need
189 to know about the .dwo files and include them.
190 Incompatible with older versions of ccache.
192 config DEBUG_INFO_DWARF4
193 bool "Generate dwarf4 debuginfo"
194 depends on DEBUG_INFO
196 Generate dwarf4 debug info. This requires recent versions
197 of gcc and gdb. It makes the debug information larger.
198 But it significantly improves the success of resolving
199 variables in gdb on optimized code.
202 bool "Provide GDB scripts for kernel debugging"
203 depends on DEBUG_INFO
205 This creates the required links to GDB helper scripts in the
206 build directory. If you load vmlinux into gdb, the helper
207 scripts will be automatically imported by gdb as well, and
208 additional functions are available to analyze a Linux kernel
209 instance. See Documentation/dev-tools/gdb-kernel-debugging.rst
212 config ENABLE_WARN_DEPRECATED
213 bool "Enable __deprecated logic"
216 Enable the __deprecated logic in the kernel build.
217 Disable this to suppress the "warning: 'foo' is deprecated
218 (declared at kernel/power/somefile.c:1234)" messages.
220 config ENABLE_MUST_CHECK
221 bool "Enable __must_check logic"
224 Enable the __must_check logic in the kernel build. Disable this to
225 suppress the "warning: ignoring return value of 'foo', declared with
226 attribute warn_unused_result" messages.
229 int "Warn for stack frames larger than (needs gcc 4.4)"
231 default 3072 if KASAN_EXTRA
232 default 2048 if GCC_PLUGIN_LATENT_ENTROPY
233 default 1280 if (!64BIT && PARISC)
234 default 1024 if (!64BIT && !PARISC)
235 default 2048 if 64BIT
237 Tell gcc to warn at build time for stack frames larger than this.
238 Setting this too low will cause a lot of warnings.
239 Setting it to 0 disables the warning.
242 config STRIP_ASM_SYMS
243 bool "Strip assembler-generated symbols during link"
246 Strip internal assembler-generated symbols during a link (symbols
247 that look like '.Lxxx') so they don't pollute the output of
248 get_wchan() and suchlike.
251 bool "Generate readable assembler code"
252 depends on DEBUG_KERNEL
254 Disable some compiler optimizations that tend to generate human unreadable
255 assembler output. This may make the kernel slightly slower, but it helps
256 to keep kernel developers who have to stare a lot at assembler listings
259 config UNUSED_SYMBOLS
260 bool "Enable unused/obsolete exported symbols"
263 Unused but exported symbols make the kernel needlessly bigger. For
264 that reason most of these unused exports will soon be removed. This
265 option is provided temporarily to provide a transition period in case
266 some external kernel module needs one of these symbols anyway. If you
267 encounter such a case in your module, consider if you are actually
268 using the right API. (rationale: since nobody in the kernel is using
269 this in a module, there is a pretty good chance it's actually the
270 wrong interface to use). If you really need the symbol, please send a
271 mail to the linux kernel mailing list mentioning the symbol and why
272 you really need it, and what the merge plan to the mainline kernel for
276 bool "Track page owner"
277 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT
281 select PAGE_EXTENSION
283 This keeps track of what call chain is the owner of a page, may
284 help to find bare alloc_page(s) leaks. Even if you include this
285 feature on your build, it is disabled in default. You should pass
286 "page_owner=on" to boot parameter in order to enable it. Eats
287 a fair amount of memory if enabled. See tools/vm/page_owner_sort.c
288 for user-space helper.
293 bool "Debug Filesystem"
295 debugfs is a virtual file system that kernel developers use to put
296 debugging files into. Enable this option to be able to read and
297 write to these files.
299 For detailed documentation on the debugfs API, see
300 Documentation/filesystems/.
305 bool "Run 'make headers_check' when building vmlinux"
308 This option will extract the user-visible kernel headers whenever
309 building the kernel, and will run basic sanity checks on them to
310 ensure that exported files do not attempt to include files which
311 were not exported, etc.
313 If you're making modifications to header files which are
314 relevant for userspace, say 'Y', and check the headers
315 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
316 your build tree), to make sure they're suitable.
318 config DEBUG_SECTION_MISMATCH
319 bool "Enable full Section mismatch analysis"
321 The section mismatch analysis checks if there are illegal
322 references from one section to another section.
323 During linktime or runtime, some sections are dropped;
324 any use of code/data previously in these sections would
325 most likely result in an oops.
326 In the code, functions and variables are annotated with
327 __init,, etc. (see the full list in include/linux/init.h),
328 which results in the code/data being placed in specific sections.
329 The section mismatch analysis is always performed after a full
330 kernel build, and enabling this option causes the following
331 additional steps to occur:
332 - Add the option -fno-inline-functions-called-once to gcc commands.
333 When inlining a function annotated with __init in a non-init
334 function, we would lose the section information and thus
335 the analysis would not catch the illegal reference.
336 This option tells gcc to inline less (but it does result in
338 - Run the section mismatch analysis for each module/built-in.a file.
339 When we run the section mismatch analysis on vmlinux.o, we
340 lose valuable information about where the mismatch was
342 Running the analysis for each module/built-in.a file
343 tells where the mismatch happens much closer to the
344 source. The drawback is that the same mismatch is
345 reported at least twice.
346 - Enable verbose reporting from modpost in order to help resolve
347 the section mismatches that are reported.
349 config SECTION_MISMATCH_WARN_ONLY
350 bool "Make section mismatch errors non-fatal"
353 If you say N here, the build process will fail if there are any
354 section mismatch, instead of just throwing warnings.
359 # Select this config option from the architecture Kconfig, if it
360 # is preferred to always offer frame pointers as a config
361 # option on the architecture (regardless of KERNEL_DEBUG):
363 config ARCH_WANT_FRAME_POINTERS
367 bool "Compile the kernel with frame pointers"
368 depends on DEBUG_KERNEL && (M68K || UML || SUPERH) || ARCH_WANT_FRAME_POINTERS
369 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
371 If you say Y here the resulting kernel image will be slightly
372 larger and slower, but it gives very useful debugging information
373 in case of kernel bugs. (precise oopses/stacktraces/warnings)
375 config STACK_VALIDATION
376 bool "Compile-time stack metadata validation"
377 depends on HAVE_STACK_VALIDATION
380 Add compile-time checks to validate stack metadata, including frame
381 pointers (if CONFIG_FRAME_POINTER is enabled). This helps ensure
382 that runtime stack traces are more reliable.
384 This is also a prerequisite for generation of ORC unwind data, which
385 is needed for CONFIG_UNWINDER_ORC.
387 For more information, see
388 tools/objtool/Documentation/stack-validation.txt.
390 config DEBUG_FORCE_WEAK_PER_CPU
391 bool "Force weak per-cpu definitions"
392 depends on DEBUG_KERNEL
394 s390 and alpha require percpu variables in modules to be
395 defined weak to work around addressing range issue which
396 puts the following two restrictions on percpu variable
399 1. percpu symbols must be unique whether static or not
400 2. percpu variables can't be defined inside a function
402 To ensure that generic code follows the above rules, this
403 option forces all percpu variables to be defined as weak.
405 endmenu # "Compiler options"
408 bool "Magic SysRq key"
411 If you say Y here, you will have some control over the system even
412 if the system crashes for example during kernel debugging (e.g., you
413 will be able to flush the buffer cache to disk, reboot the system
414 immediately or dump some status information). This is accomplished
415 by pressing various keys while holding SysRq (Alt+PrintScreen). It
416 also works on a serial console (on PC hardware at least), if you
417 send a BREAK and then within 5 seconds a command keypress. The
418 keys are documented in <file:Documentation/admin-guide/sysrq.rst>.
419 Don't say Y unless you really know what this hack does.
421 config MAGIC_SYSRQ_DEFAULT_ENABLE
422 hex "Enable magic SysRq key functions by default"
423 depends on MAGIC_SYSRQ
426 Specifies which SysRq key functions are enabled by default.
427 This may be set to 1 or 0 to enable or disable them all, or
428 to a bitmask as described in Documentation/admin-guide/sysrq.rst.
430 config MAGIC_SYSRQ_SERIAL
431 bool "Enable magic SysRq key over serial"
432 depends on MAGIC_SYSRQ
435 Many embedded boards have a disconnected TTL level serial which can
436 generate some garbage that can lead to spurious false sysrq detects.
437 This option allows you to decide whether you want to enable the
441 bool "Kernel debugging"
443 Say Y here if you are developing drivers or trying to debug and
444 identify kernel problems.
446 menu "Memory Debugging"
448 source mm/Kconfig.debug
451 bool "Debug object operations"
452 depends on DEBUG_KERNEL
454 If you say Y here, additional code will be inserted into the
455 kernel to track the life time of various objects and validate
456 the operations on those objects.
458 config DEBUG_OBJECTS_SELFTEST
459 bool "Debug objects selftest"
460 depends on DEBUG_OBJECTS
462 This enables the selftest of the object debug code.
464 config DEBUG_OBJECTS_FREE
465 bool "Debug objects in freed memory"
466 depends on DEBUG_OBJECTS
468 This enables checks whether a k/v free operation frees an area
469 which contains an object which has not been deactivated
470 properly. This can make kmalloc/kfree-intensive workloads
473 config DEBUG_OBJECTS_TIMERS
474 bool "Debug timer objects"
475 depends on DEBUG_OBJECTS
477 If you say Y here, additional code will be inserted into the
478 timer routines to track the life time of timer objects and
479 validate the timer operations.
481 config DEBUG_OBJECTS_WORK
482 bool "Debug work objects"
483 depends on DEBUG_OBJECTS
485 If you say Y here, additional code will be inserted into the
486 work queue routines to track the life time of work objects and
487 validate the work operations.
489 config DEBUG_OBJECTS_RCU_HEAD
490 bool "Debug RCU callbacks objects"
491 depends on DEBUG_OBJECTS
493 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
495 config DEBUG_OBJECTS_PERCPU_COUNTER
496 bool "Debug percpu counter objects"
497 depends on DEBUG_OBJECTS
499 If you say Y here, additional code will be inserted into the
500 percpu counter routines to track the life time of percpu counter
501 objects and validate the percpu counter operations.
503 config DEBUG_OBJECTS_ENABLE_DEFAULT
504 int "debug_objects bootup default value (0-1)"
507 depends on DEBUG_OBJECTS
509 Debug objects boot parameter default value
512 bool "Debug slab memory allocations"
513 depends on DEBUG_KERNEL && SLAB
515 Say Y here to have the kernel do limited verification on memory
516 allocation as well as poisoning memory on free to catch use of freed
517 memory. This can make kmalloc/kfree-intensive workloads much slower.
519 config DEBUG_SLAB_LEAK
520 bool "Memory leak debugging"
521 depends on DEBUG_SLAB
524 bool "SLUB debugging on by default"
525 depends on SLUB && SLUB_DEBUG
528 Boot with debugging on by default. SLUB boots by default with
529 the runtime debug capabilities switched off. Enabling this is
530 equivalent to specifying the "slub_debug" parameter on boot.
531 There is no support for more fine grained debug control like
532 possible with slub_debug=xxx. SLUB debugging may be switched
533 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
538 bool "Enable SLUB performance statistics"
539 depends on SLUB && SYSFS
541 SLUB statistics are useful to debug SLUBs allocation behavior in
542 order find ways to optimize the allocator. This should never be
543 enabled for production use since keeping statistics slows down
544 the allocator by a few percentage points. The slabinfo command
545 supports the determination of the most active slabs to figure
546 out which slabs are relevant to a particular load.
547 Try running: slabinfo -DA
549 config HAVE_DEBUG_KMEMLEAK
552 config DEBUG_KMEMLEAK
553 bool "Kernel memory leak detector"
554 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
556 select STACKTRACE if STACKTRACE_SUPPORT
560 Say Y here if you want to enable the memory leak
561 detector. The memory allocation/freeing is traced in a way
562 similar to the Boehm's conservative garbage collector, the
563 difference being that the orphan objects are not freed but
564 only shown in /sys/kernel/debug/kmemleak. Enabling this
565 feature will introduce an overhead to memory
566 allocations. See Documentation/dev-tools/kmemleak.rst for more
569 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
570 of finding leaks due to the slab objects poisoning.
572 In order to access the kmemleak file, debugfs needs to be
573 mounted (usually at /sys/kernel/debug).
575 config DEBUG_KMEMLEAK_EARLY_LOG_SIZE
576 int "Maximum kmemleak early log entries"
577 depends on DEBUG_KMEMLEAK
581 Kmemleak must track all the memory allocations to avoid
582 reporting false positives. Since memory may be allocated or
583 freed before kmemleak is initialised, an early log buffer is
584 used to store these actions. If kmemleak reports "early log
585 buffer exceeded", please increase this value.
587 config DEBUG_KMEMLEAK_TEST
588 tristate "Simple test for the kernel memory leak detector"
589 depends on DEBUG_KMEMLEAK && m
591 This option enables a module that explicitly leaks memory.
595 config DEBUG_KMEMLEAK_DEFAULT_OFF
596 bool "Default kmemleak to off"
597 depends on DEBUG_KMEMLEAK
599 Say Y here to disable kmemleak by default. It can then be enabled
600 on the command line via kmemleak=on.
602 config DEBUG_STACK_USAGE
603 bool "Stack utilization instrumentation"
604 depends on DEBUG_KERNEL && !IA64
606 Enables the display of the minimum amount of free stack which each
607 task has ever had available in the sysrq-T and sysrq-P debug output.
609 This option will slow down process creation somewhat.
613 depends on DEBUG_KERNEL
615 Enable this to turn on extended checks in the virtual-memory system
616 that may impact performance.
620 config DEBUG_VM_VMACACHE
621 bool "Debug VMA caching"
624 Enable this to turn on VMA caching debug information. Doing so
625 can cause significant overhead, so only enable it in non-production
631 bool "Debug VM red-black trees"
634 Enable VM red-black tree debugging information and extra validations.
638 config DEBUG_VM_PGFLAGS
639 bool "Debug page-flags operations"
642 Enables extra validation on page flags operations.
646 config ARCH_HAS_DEBUG_VIRTUAL
650 bool "Debug VM translations"
651 depends on DEBUG_KERNEL && ARCH_HAS_DEBUG_VIRTUAL
653 Enable some costly sanity checks in virtual to page code. This can
654 catch mistakes with virt_to_page() and friends.
658 config DEBUG_NOMMU_REGIONS
659 bool "Debug the global anon/private NOMMU mapping region tree"
660 depends on DEBUG_KERNEL && !MMU
662 This option causes the global tree of anonymous and private mapping
663 regions to be regularly checked for invalid topology.
665 config DEBUG_MEMORY_INIT
666 bool "Debug memory initialisation" if EXPERT
669 Enable this for additional checks during memory initialisation.
670 The sanity checks verify aspects of the VM such as the memory model
671 and other information provided by the architecture. Verbose
672 information will be printed at KERN_DEBUG loglevel depending
673 on the mminit_loglevel= command-line option.
677 config MEMORY_NOTIFIER_ERROR_INJECT
678 tristate "Memory hotplug notifier error injection module"
679 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
681 This option provides the ability to inject artificial errors to
682 memory hotplug notifier chain callbacks. It is controlled through
683 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
685 If the notifier call chain should be failed with some events
686 notified, write the error code to "actions/<notifier event>/error".
688 Example: Inject memory hotplug offline error (-12 == -ENOMEM)
690 # cd /sys/kernel/debug/notifier-error-inject/memory
691 # echo -12 > actions/MEM_GOING_OFFLINE/error
692 # echo offline > /sys/devices/system/memory/memoryXXX/state
693 bash: echo: write error: Cannot allocate memory
695 To compile this code as a module, choose M here: the module will
696 be called memory-notifier-error-inject.
700 config DEBUG_PER_CPU_MAPS
701 bool "Debug access to per_cpu maps"
702 depends on DEBUG_KERNEL
705 Say Y to verify that the per_cpu map being accessed has
706 been set up. This adds a fair amount of code to kernel memory
707 and decreases performance.
712 bool "Highmem debugging"
713 depends on DEBUG_KERNEL && HIGHMEM
715 This option enables additional error checking for high memory
716 systems. Disable for production systems.
718 config HAVE_DEBUG_STACKOVERFLOW
721 config DEBUG_STACKOVERFLOW
722 bool "Check for stack overflows"
723 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
725 Say Y here if you want to check for overflows of kernel, IRQ
726 and exception stacks (if your architecture uses them). This
727 option will show detailed messages if free stack space drops
728 below a certain limit.
730 These kinds of bugs usually occur when call-chains in the
731 kernel get too deep, especially when interrupts are
734 Use this in cases where you see apparently random memory
735 corruption, especially if it appears in 'struct thread_info'
737 If in doubt, say "N".
739 source "lib/Kconfig.kasan"
741 endmenu # "Memory Debugging"
746 KCOV does not have any arch-specific code, but currently it is enabled
747 only for x86_64. KCOV requires testing on other archs, and most likely
748 disabling of instrumentation for some early boot code.
750 config CC_HAS_SANCOV_TRACE_PC
751 def_bool $(cc-option,-fsanitize-coverage=trace-pc)
754 bool "Code coverage for fuzzing"
755 depends on ARCH_HAS_KCOV
756 depends on CC_HAS_SANCOV_TRACE_PC || GCC_PLUGINS
758 select GCC_PLUGIN_SANCOV if !CC_HAS_SANCOV_TRACE_PC
760 KCOV exposes kernel code coverage information in a form suitable
761 for coverage-guided fuzzing (randomized testing).
763 If RANDOMIZE_BASE is enabled, PC values will not be stable across
764 different machines and across reboots. If you need stable PC values,
765 disable RANDOMIZE_BASE.
767 For more details, see Documentation/dev-tools/kcov.rst.
769 config KCOV_ENABLE_COMPARISONS
770 bool "Enable comparison operands collection by KCOV"
772 depends on $(cc-option,-fsanitize-coverage=trace-cmp)
774 KCOV also exposes operands of every comparison in the instrumented
775 code along with operand sizes and PCs of the comparison instructions.
776 These operands can be used by fuzzing engines to improve the quality
779 config KCOV_INSTRUMENT_ALL
780 bool "Instrument all code by default"
784 If you are doing generic system call fuzzing (like e.g. syzkaller),
785 then you will want to instrument the whole kernel and you should
786 say y here. If you are doing more targeted fuzzing (like e.g.
787 filesystem fuzzing with AFL) then you will want to enable coverage
788 for more specific subsets of files, and should say n here.
791 bool "Debug shared IRQ handlers"
792 depends on DEBUG_KERNEL
794 Enable this to generate a spurious interrupt as soon as a shared
795 interrupt handler is registered, and just before one is deregistered.
796 Drivers ought to be able to handle interrupts coming in at those
797 points; some don't and need to be caught.
799 menu "Debug Lockups and Hangs"
801 config LOCKUP_DETECTOR
804 config SOFTLOCKUP_DETECTOR
805 bool "Detect Soft Lockups"
806 depends on DEBUG_KERNEL && !S390
807 select LOCKUP_DETECTOR
809 Say Y here to enable the kernel to act as a watchdog to detect
812 Softlockups are bugs that cause the kernel to loop in kernel
813 mode for more than 20 seconds, without giving other tasks a
814 chance to run. The current stack trace is displayed upon
815 detection and the system will stay locked up.
817 config BOOTPARAM_SOFTLOCKUP_PANIC
818 bool "Panic (Reboot) On Soft Lockups"
819 depends on SOFTLOCKUP_DETECTOR
821 Say Y here to enable the kernel to panic on "soft lockups",
822 which are bugs that cause the kernel to loop in kernel
823 mode for more than 20 seconds (configurable using the watchdog_thresh
824 sysctl), without giving other tasks a chance to run.
826 The panic can be used in combination with panic_timeout,
827 to cause the system to reboot automatically after a
828 lockup has been detected. This feature is useful for
829 high-availability systems that have uptime guarantees and
830 where a lockup must be resolved ASAP.
834 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
836 depends on SOFTLOCKUP_DETECTOR
838 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
839 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
841 config HARDLOCKUP_DETECTOR_PERF
843 select SOFTLOCKUP_DETECTOR
846 # Enables a timestamp based low pass filter to compensate for perf based
847 # hard lockup detection which runs too fast due to turbo modes.
849 config HARDLOCKUP_CHECK_TIMESTAMP
853 # arch/ can define HAVE_HARDLOCKUP_DETECTOR_ARCH to provide their own hard
854 # lockup detector rather than the perf based detector.
856 config HARDLOCKUP_DETECTOR
857 bool "Detect Hard Lockups"
858 depends on DEBUG_KERNEL && !S390
859 depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_ARCH
860 select LOCKUP_DETECTOR
861 select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF
862 select HARDLOCKUP_DETECTOR_ARCH if HAVE_HARDLOCKUP_DETECTOR_ARCH
864 Say Y here to enable the kernel to act as a watchdog to detect
867 Hardlockups are bugs that cause the CPU to loop in kernel mode
868 for more than 10 seconds, without letting other interrupts have a
869 chance to run. The current stack trace is displayed upon detection
870 and the system will stay locked up.
872 config BOOTPARAM_HARDLOCKUP_PANIC
873 bool "Panic (Reboot) On Hard Lockups"
874 depends on HARDLOCKUP_DETECTOR
876 Say Y here to enable the kernel to panic on "hard lockups",
877 which are bugs that cause the kernel to loop in kernel
878 mode with interrupts disabled for more than 10 seconds (configurable
879 using the watchdog_thresh sysctl).
883 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
885 depends on HARDLOCKUP_DETECTOR
887 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
888 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
890 config DETECT_HUNG_TASK
891 bool "Detect Hung Tasks"
892 depends on DEBUG_KERNEL
893 default SOFTLOCKUP_DETECTOR
895 Say Y here to enable the kernel to detect "hung tasks",
896 which are bugs that cause the task to be stuck in
897 uninterruptible "D" state indefinitely.
899 When a hung task is detected, the kernel will print the
900 current stack trace (which you should report), but the
901 task will stay in uninterruptible state. If lockdep is
902 enabled then all held locks will also be reported. This
903 feature has negligible overhead.
905 config DEFAULT_HUNG_TASK_TIMEOUT
906 int "Default timeout for hung task detection (in seconds)"
907 depends on DETECT_HUNG_TASK
910 This option controls the default timeout (in seconds) used
911 to determine when a task has become non-responsive and should
914 It can be adjusted at runtime via the kernel.hung_task_timeout_secs
915 sysctl or by writing a value to
916 /proc/sys/kernel/hung_task_timeout_secs.
918 A timeout of 0 disables the check. The default is two minutes.
919 Keeping the default should be fine in most cases.
921 config BOOTPARAM_HUNG_TASK_PANIC
922 bool "Panic (Reboot) On Hung Tasks"
923 depends on DETECT_HUNG_TASK
925 Say Y here to enable the kernel to panic on "hung tasks",
926 which are bugs that cause the kernel to leave a task stuck
927 in uninterruptible "D" state.
929 The panic can be used in combination with panic_timeout,
930 to cause the system to reboot automatically after a
931 hung task has been detected. This feature is useful for
932 high-availability systems that have uptime guarantees and
933 where a hung tasks must be resolved ASAP.
937 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
939 depends on DETECT_HUNG_TASK
941 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
942 default 1 if BOOTPARAM_HUNG_TASK_PANIC
945 bool "Detect Workqueue Stalls"
946 depends on DEBUG_KERNEL
948 Say Y here to enable stall detection on workqueues. If a
949 worker pool doesn't make forward progress on a pending work
950 item for over a given amount of time, 30s by default, a
951 warning message is printed along with dump of workqueue
952 state. This can be configured through kernel parameter
953 "workqueue.watchdog_thresh" and its sysfs counterpart.
955 endmenu # "Debug lockups and hangs"
960 Say Y here to enable the kernel to panic when it oopses. This
961 has the same effect as setting oops=panic on the kernel command
964 This feature is useful to ensure that the kernel does not do
965 anything erroneous after an oops which could result in data
966 corruption or other issues.
970 config PANIC_ON_OOPS_VALUE
973 default 0 if !PANIC_ON_OOPS
974 default 1 if PANIC_ON_OOPS
980 Set the timeout value (in seconds) until a reboot occurs when the
981 the kernel panics. If n = 0, then we wait forever. A timeout
982 value n > 0 will wait n seconds before rebooting, while a timeout
983 value n < 0 will reboot immediately.
986 bool "Collect scheduler debugging info"
987 depends on DEBUG_KERNEL && PROC_FS
990 If you say Y here, the /proc/sched_debug file will be provided
991 that can help debug the scheduler. The runtime overhead of this
999 bool "Collect scheduler statistics"
1000 depends on DEBUG_KERNEL && PROC_FS
1003 If you say Y here, additional code will be inserted into the
1004 scheduler and related routines to collect statistics about
1005 scheduler behavior and provide them in /proc/schedstat. These
1006 stats may be useful for both tuning and debugging the scheduler
1007 If you aren't debugging the scheduler or trying to tune a specific
1008 application, you can say N to avoid the very slight overhead
1011 config SCHED_STACK_END_CHECK
1012 bool "Detect stack corruption on calls to schedule()"
1013 depends on DEBUG_KERNEL
1016 This option checks for a stack overrun on calls to schedule().
1017 If the stack end location is found to be over written always panic as
1018 the content of the corrupted region can no longer be trusted.
1019 This is to ensure no erroneous behaviour occurs which could result in
1020 data corruption or a sporadic crash at a later stage once the region
1021 is examined. The runtime overhead introduced is minimal.
1023 config DEBUG_TIMEKEEPING
1024 bool "Enable extra timekeeping sanity checking"
1026 This option will enable additional timekeeping sanity checks
1027 which may be helpful when diagnosing issues where timekeeping
1028 problems are suspected.
1030 This may include checks in the timekeeping hotpaths, so this
1031 option may have a (very small) performance impact to some
1036 config DEBUG_PREEMPT
1037 bool "Debug preemptible kernel"
1038 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
1041 If you say Y here then the kernel will use a debug variant of the
1042 commonly used smp_processor_id() function and will print warnings
1043 if kernel code uses it in a preemption-unsafe way. Also, the kernel
1044 will detect preemption count underflows.
1046 menu "Lock Debugging (spinlocks, mutexes, etc...)"
1048 config LOCK_DEBUGGING_SUPPORT
1050 depends on TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1053 config PROVE_LOCKING
1054 bool "Lock debugging: prove locking correctness"
1055 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1057 select DEBUG_SPINLOCK
1058 select DEBUG_MUTEXES
1059 select DEBUG_RT_MUTEXES if RT_MUTEXES
1060 select DEBUG_RWSEMS if RWSEM_SPIN_ON_OWNER
1061 select DEBUG_WW_MUTEX_SLOWPATH
1062 select DEBUG_LOCK_ALLOC
1063 select TRACE_IRQFLAGS
1066 This feature enables the kernel to prove that all locking
1067 that occurs in the kernel runtime is mathematically
1068 correct: that under no circumstance could an arbitrary (and
1069 not yet triggered) combination of observed locking
1070 sequences (on an arbitrary number of CPUs, running an
1071 arbitrary number of tasks and interrupt contexts) cause a
1074 In short, this feature enables the kernel to report locking
1075 related deadlocks before they actually occur.
1077 The proof does not depend on how hard and complex a
1078 deadlock scenario would be to trigger: how many
1079 participant CPUs, tasks and irq-contexts would be needed
1080 for it to trigger. The proof also does not depend on
1081 timing: if a race and a resulting deadlock is possible
1082 theoretically (no matter how unlikely the race scenario
1083 is), it will be proven so and will immediately be
1084 reported by the kernel (once the event is observed that
1085 makes the deadlock theoretically possible).
1087 If a deadlock is impossible (i.e. the locking rules, as
1088 observed by the kernel, are mathematically correct), the
1089 kernel reports nothing.
1091 NOTE: this feature can also be enabled for rwlocks, mutexes
1092 and rwsems - in which case all dependencies between these
1093 different locking variants are observed and mapped too, and
1094 the proof of observed correctness is also maintained for an
1095 arbitrary combination of these separate locking variants.
1097 For more details, see Documentation/locking/lockdep-design.txt.
1100 bool "Lock usage statistics"
1101 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1103 select DEBUG_SPINLOCK
1104 select DEBUG_MUTEXES
1105 select DEBUG_RT_MUTEXES if RT_MUTEXES
1106 select DEBUG_LOCK_ALLOC
1109 This feature enables tracking lock contention points
1111 For more details, see Documentation/locking/lockstat.txt
1113 This also enables lock events required by "perf lock",
1115 If you want to use "perf lock", you also need to turn on
1116 CONFIG_EVENT_TRACING.
1118 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
1119 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
1121 config DEBUG_RT_MUTEXES
1122 bool "RT Mutex debugging, deadlock detection"
1123 depends on DEBUG_KERNEL && RT_MUTEXES
1125 This allows rt mutex semantics violations and rt mutex related
1126 deadlocks (lockups) to be detected and reported automatically.
1128 config DEBUG_SPINLOCK
1129 bool "Spinlock and rw-lock debugging: basic checks"
1130 depends on DEBUG_KERNEL
1131 select UNINLINE_SPIN_UNLOCK
1133 Say Y here and build SMP to catch missing spinlock initialization
1134 and certain other kinds of spinlock errors commonly made. This is
1135 best used in conjunction with the NMI watchdog so that spinlock
1136 deadlocks are also debuggable.
1138 config DEBUG_MUTEXES
1139 bool "Mutex debugging: basic checks"
1140 depends on DEBUG_KERNEL
1142 This feature allows mutex semantics violations to be detected and
1145 config DEBUG_WW_MUTEX_SLOWPATH
1146 bool "Wait/wound mutex debugging: Slowpath testing"
1147 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1148 select DEBUG_LOCK_ALLOC
1149 select DEBUG_SPINLOCK
1150 select DEBUG_MUTEXES
1152 This feature enables slowpath testing for w/w mutex users by
1153 injecting additional -EDEADLK wound/backoff cases. Together with
1154 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this
1155 will test all possible w/w mutex interface abuse with the
1156 exception of simply not acquiring all the required locks.
1157 Note that this feature can introduce significant overhead, so
1158 it really should not be enabled in a production or distro kernel,
1159 even a debug kernel. If you are a driver writer, enable it. If
1160 you are a distro, do not.
1163 bool "RW Semaphore debugging: basic checks"
1164 depends on DEBUG_KERNEL && RWSEM_SPIN_ON_OWNER
1166 This debugging feature allows mismatched rw semaphore locks and unlocks
1167 to be detected and reported.
1169 config DEBUG_LOCK_ALLOC
1170 bool "Lock debugging: detect incorrect freeing of live locks"
1171 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1172 select DEBUG_SPINLOCK
1173 select DEBUG_MUTEXES
1174 select DEBUG_RT_MUTEXES if RT_MUTEXES
1177 This feature will check whether any held lock (spinlock, rwlock,
1178 mutex or rwsem) is incorrectly freed by the kernel, via any of the
1179 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
1180 vfree(), etc.), whether a live lock is incorrectly reinitialized via
1181 spin_lock_init()/mutex_init()/etc., or whether there is any lock
1182 held during task exit.
1186 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1188 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE && !ARC && !X86
1192 config LOCKDEP_SMALL
1195 config DEBUG_LOCKDEP
1196 bool "Lock dependency engine debugging"
1197 depends on DEBUG_KERNEL && LOCKDEP
1199 If you say Y here, the lock dependency engine will do
1200 additional runtime checks to debug itself, at the price
1201 of more runtime overhead.
1203 config DEBUG_ATOMIC_SLEEP
1204 bool "Sleep inside atomic section checking"
1205 select PREEMPT_COUNT
1206 depends on DEBUG_KERNEL
1208 If you say Y here, various routines which may sleep will become very
1209 noisy if they are called inside atomic sections: when a spinlock is
1210 held, inside an rcu read side critical section, inside preempt disabled
1211 sections, inside an interrupt, etc...
1213 config DEBUG_LOCKING_API_SELFTESTS
1214 bool "Locking API boot-time self-tests"
1215 depends on DEBUG_KERNEL
1217 Say Y here if you want the kernel to run a short self-test during
1218 bootup. The self-test checks whether common types of locking bugs
1219 are detected by debugging mechanisms or not. (if you disable
1220 lock debugging then those bugs wont be detected of course.)
1221 The following locking APIs are covered: spinlocks, rwlocks,
1224 config LOCK_TORTURE_TEST
1225 tristate "torture tests for locking"
1226 depends on DEBUG_KERNEL
1230 This option provides a kernel module that runs torture tests
1231 on kernel locking primitives. The kernel module may be built
1232 after the fact on the running kernel to be tested, if desired.
1234 Say Y here if you want kernel locking-primitive torture tests
1235 to be built into the kernel.
1236 Say M if you want these torture tests to build as a module.
1237 Say N if you are unsure.
1239 config WW_MUTEX_SELFTEST
1240 tristate "Wait/wound mutex selftests"
1242 This option provides a kernel module that runs tests on the
1243 on the struct ww_mutex locking API.
1245 It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction
1246 with this test harness.
1248 Say M if you want these self tests to build as a module.
1249 Say N if you are unsure.
1251 endmenu # lock debugging
1253 config TRACE_IRQFLAGS
1256 Enables hooks to interrupt enabling and disabling for
1257 either tracing or lock debugging.
1260 bool "Stack backtrace support"
1261 depends on STACKTRACE_SUPPORT
1263 This option causes the kernel to create a /proc/pid/stack for
1264 every process, showing its current stack trace.
1265 It is also used by various kernel debugging features that require
1266 stack trace generation.
1268 config WARN_ALL_UNSEEDED_RANDOM
1269 bool "Warn for all uses of unseeded randomness"
1272 Some parts of the kernel contain bugs relating to their use of
1273 cryptographically secure random numbers before it's actually possible
1274 to generate those numbers securely. This setting ensures that these
1275 flaws don't go unnoticed, by enabling a message, should this ever
1276 occur. This will allow people with obscure setups to know when things
1277 are going wrong, so that they might contact developers about fixing
1280 Unfortunately, on some models of some architectures getting
1281 a fully seeded CRNG is extremely difficult, and so this can
1282 result in dmesg getting spammed for a surprisingly long
1283 time. This is really bad from a security perspective, and
1284 so architecture maintainers really need to do what they can
1285 to get the CRNG seeded sooner after the system is booted.
1286 However, since users can not do anything actionble to
1287 address this, by default the kernel will issue only a single
1288 warning for the first use of unseeded randomness.
1290 Say Y here if you want to receive warnings for all uses of
1291 unseeded randomness. This will be of use primarily for
1292 those developers interersted in improving the security of
1293 Linux kernels running on their architecture (or
1296 config DEBUG_KOBJECT
1297 bool "kobject debugging"
1298 depends on DEBUG_KERNEL
1300 If you say Y here, some extra kobject debugging messages will be sent
1303 config DEBUG_KOBJECT_RELEASE
1304 bool "kobject release debugging"
1305 depends on DEBUG_OBJECTS_TIMERS
1307 kobjects are reference counted objects. This means that their
1308 last reference count put is not predictable, and the kobject can
1309 live on past the point at which a driver decides to drop it's
1310 initial reference to the kobject gained on allocation. An
1311 example of this would be a struct device which has just been
1314 However, some buggy drivers assume that after such an operation,
1315 the memory backing the kobject can be immediately freed. This
1316 goes completely against the principles of a refcounted object.
1318 If you say Y here, the kernel will delay the release of kobjects
1319 on the last reference count to improve the visibility of this
1320 kind of kobject release bug.
1322 config HAVE_DEBUG_BUGVERBOSE
1325 config DEBUG_BUGVERBOSE
1326 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
1327 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
1330 Say Y here to make BUG() panics output the file name and line number
1331 of the BUG call as well as the EIP and oops trace. This aids
1332 debugging but costs about 70-100K of memory.
1335 bool "Debug linked list manipulation"
1336 depends on DEBUG_KERNEL || BUG_ON_DATA_CORRUPTION
1338 Enable this to turn on extended checks in the linked-list
1343 config DEBUG_PI_LIST
1344 bool "Debug priority linked list manipulation"
1345 depends on DEBUG_KERNEL
1347 Enable this to turn on extended checks in the priority-ordered
1348 linked-list (plist) walking routines. This checks the entire
1349 list multiple times during each manipulation.
1354 bool "Debug SG table operations"
1355 depends on DEBUG_KERNEL
1357 Enable this to turn on checks on scatter-gather tables. This can
1358 help find problems with drivers that do not properly initialize
1363 config DEBUG_NOTIFIERS
1364 bool "Debug notifier call chains"
1365 depends on DEBUG_KERNEL
1367 Enable this to turn on sanity checking for notifier call chains.
1368 This is most useful for kernel developers to make sure that
1369 modules properly unregister themselves from notifier chains.
1370 This is a relatively cheap check but if you care about maximum
1373 config DEBUG_CREDENTIALS
1374 bool "Debug credential management"
1375 depends on DEBUG_KERNEL
1377 Enable this to turn on some debug checking for credential
1378 management. The additional code keeps track of the number of
1379 pointers from task_structs to any given cred struct, and checks to
1380 see that this number never exceeds the usage count of the cred
1383 Furthermore, if SELinux is enabled, this also checks that the
1384 security pointer in the cred struct is never seen to be invalid.
1388 source "kernel/rcu/Kconfig.debug"
1390 config DEBUG_WQ_FORCE_RR_CPU
1391 bool "Force round-robin CPU selection for unbound work items"
1392 depends on DEBUG_KERNEL
1395 Workqueue used to implicitly guarantee that work items queued
1396 without explicit CPU specified are put on the local CPU. This
1397 guarantee is no longer true and while local CPU is still
1398 preferred work items may be put on foreign CPUs. Kernel
1399 parameter "workqueue.debug_force_rr_cpu" is added to force
1400 round-robin CPU selection to flush out usages which depend on the
1401 now broken guarantee. This config option enables the debug
1402 feature by default. When enabled, memory and cache locality will
1405 config DEBUG_BLOCK_EXT_DEVT
1406 bool "Force extended block device numbers and spread them"
1407 depends on DEBUG_KERNEL
1411 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1412 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1413 YOU ARE DOING. Distros, please enable this and fix whatever
1416 Conventionally, block device numbers are allocated from
1417 predetermined contiguous area. However, extended block area
1418 may introduce non-contiguous block device numbers. This
1419 option forces most block device numbers to be allocated from
1420 the extended space and spreads them to discover kernel or
1421 userland code paths which assume predetermined contiguous
1422 device number allocation.
1424 Note that turning on this debug option shuffles all the
1425 device numbers for all IDE and SCSI devices including libata
1426 ones, so root partition specified using device number
1427 directly (via rdev or root=MAJ:MIN) won't work anymore.
1428 Textual device names (root=/dev/sdXn) will continue to work.
1430 Say N if you are unsure.
1432 config CPU_HOTPLUG_STATE_CONTROL
1433 bool "Enable CPU hotplug state control"
1434 depends on DEBUG_KERNEL
1435 depends on HOTPLUG_CPU
1438 Allows to write steps between "offline" and "online" to the CPUs
1439 sysfs target file so states can be stepped granular. This is a debug
1440 option for now as the hotplug machinery cannot be stopped and
1441 restarted at arbitrary points yet.
1443 Say N if your are unsure.
1445 config NOTIFIER_ERROR_INJECTION
1446 tristate "Notifier error injection"
1447 depends on DEBUG_KERNEL
1450 This option provides the ability to inject artificial errors to
1451 specified notifier chain callbacks. It is useful to test the error
1452 handling of notifier call chain failures.
1456 config PM_NOTIFIER_ERROR_INJECT
1457 tristate "PM notifier error injection module"
1458 depends on PM && NOTIFIER_ERROR_INJECTION
1459 default m if PM_DEBUG
1461 This option provides the ability to inject artificial errors to
1462 PM notifier chain callbacks. It is controlled through debugfs
1463 interface /sys/kernel/debug/notifier-error-inject/pm
1465 If the notifier call chain should be failed with some events
1466 notified, write the error code to "actions/<notifier event>/error".
1468 Example: Inject PM suspend error (-12 = -ENOMEM)
1470 # cd /sys/kernel/debug/notifier-error-inject/pm/
1471 # echo -12 > actions/PM_SUSPEND_PREPARE/error
1472 # echo mem > /sys/power/state
1473 bash: echo: write error: Cannot allocate memory
1475 To compile this code as a module, choose M here: the module will
1476 be called pm-notifier-error-inject.
1480 config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1481 tristate "OF reconfig notifier error injection module"
1482 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1484 This option provides the ability to inject artificial errors to
1485 OF reconfig notifier chain callbacks. It is controlled
1486 through debugfs interface under
1487 /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1489 If the notifier call chain should be failed with some events
1490 notified, write the error code to "actions/<notifier event>/error".
1492 To compile this code as a module, choose M here: the module will
1493 be called of-reconfig-notifier-error-inject.
1497 config NETDEV_NOTIFIER_ERROR_INJECT
1498 tristate "Netdev notifier error injection module"
1499 depends on NET && NOTIFIER_ERROR_INJECTION
1501 This option provides the ability to inject artificial errors to
1502 netdevice notifier chain callbacks. It is controlled through debugfs
1503 interface /sys/kernel/debug/notifier-error-inject/netdev
1505 If the notifier call chain should be failed with some events
1506 notified, write the error code to "actions/<notifier event>/error".
1508 Example: Inject netdevice mtu change error (-22 = -EINVAL)
1510 # cd /sys/kernel/debug/notifier-error-inject/netdev
1511 # echo -22 > actions/NETDEV_CHANGEMTU/error
1512 # ip link set eth0 mtu 1024
1513 RTNETLINK answers: Invalid argument
1515 To compile this code as a module, choose M here: the module will
1516 be called netdev-notifier-error-inject.
1520 config FUNCTION_ERROR_INJECTION
1522 depends on HAVE_FUNCTION_ERROR_INJECTION && KPROBES
1524 config FAULT_INJECTION
1525 bool "Fault-injection framework"
1526 depends on DEBUG_KERNEL
1528 Provide fault-injection framework.
1529 For more details, see Documentation/fault-injection/.
1532 bool "Fault-injection capability for kmalloc"
1533 depends on FAULT_INJECTION
1534 depends on SLAB || SLUB
1536 Provide fault-injection capability for kmalloc.
1538 config FAIL_PAGE_ALLOC
1539 bool "Fault-injection capabilitiy for alloc_pages()"
1540 depends on FAULT_INJECTION
1542 Provide fault-injection capability for alloc_pages().
1544 config FAIL_MAKE_REQUEST
1545 bool "Fault-injection capability for disk IO"
1546 depends on FAULT_INJECTION && BLOCK
1548 Provide fault-injection capability for disk IO.
1550 config FAIL_IO_TIMEOUT
1551 bool "Fault-injection capability for faking disk interrupts"
1552 depends on FAULT_INJECTION && BLOCK
1554 Provide fault-injection capability on end IO handling. This
1555 will make the block layer "forget" an interrupt as configured,
1556 thus exercising the error handling.
1558 Only works with drivers that use the generic timeout handling,
1559 for others it wont do anything.
1562 bool "Fault-injection capability for futexes"
1564 depends on FAULT_INJECTION && FUTEX
1566 Provide fault-injection capability for futexes.
1568 config FAULT_INJECTION_DEBUG_FS
1569 bool "Debugfs entries for fault-injection capabilities"
1570 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1572 Enable configuration of fault-injection capabilities via debugfs.
1574 config FAIL_FUNCTION
1575 bool "Fault-injection capability for functions"
1576 depends on FAULT_INJECTION_DEBUG_FS && FUNCTION_ERROR_INJECTION
1578 Provide function-based fault-injection capability.
1579 This will allow you to override a specific function with a return
1580 with given return value. As a result, function caller will see
1581 an error value and have to handle it. This is useful to test the
1582 error handling in various subsystems.
1584 config FAIL_MMC_REQUEST
1585 bool "Fault-injection capability for MMC IO"
1586 depends on FAULT_INJECTION_DEBUG_FS && MMC
1588 Provide fault-injection capability for MMC IO.
1589 This will make the mmc core return data errors. This is
1590 useful to test the error handling in the mmc block device
1591 and to test how the mmc host driver handles retries from
1594 config FAULT_INJECTION_STACKTRACE_FILTER
1595 bool "stacktrace filter for fault-injection capabilities"
1596 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1599 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC && !X86
1601 Provide stacktrace filter for fault-injection capabilities
1604 bool "Latency measuring infrastructure"
1605 depends on DEBUG_KERNEL
1606 depends on STACKTRACE_SUPPORT
1608 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC && !X86
1615 Enable this option if you want to use the LatencyTOP tool
1616 to find out which userspace is blocking on what kernel operations.
1618 source kernel/trace/Kconfig
1620 config PROVIDE_OHCI1394_DMA_INIT
1621 bool "Remote debugging over FireWire early on boot"
1622 depends on PCI && X86
1624 If you want to debug problems which hang or crash the kernel early
1625 on boot and the crashing machine has a FireWire port, you can use
1626 this feature to remotely access the memory of the crashed machine
1627 over FireWire. This employs remote DMA as part of the OHCI1394
1628 specification which is now the standard for FireWire controllers.
1630 With remote DMA, you can monitor the printk buffer remotely using
1631 firescope and access all memory below 4GB using fireproxy from gdb.
1632 Even controlling a kernel debugger is possible using remote DMA.
1636 If ohci1394_dma=early is used as boot parameter, it will initialize
1637 all OHCI1394 controllers which are found in the PCI config space.
1639 As all changes to the FireWire bus such as enabling and disabling
1640 devices cause a bus reset and thereby disable remote DMA for all
1641 devices, be sure to have the cable plugged and FireWire enabled on
1642 the debugging host before booting the debug target for debugging.
1644 This code (~1k) is freed after boot. By then, the firewire stack
1645 in charge of the OHCI-1394 controllers should be used instead.
1647 See Documentation/debugging-via-ohci1394.txt for more information.
1649 config DMA_API_DEBUG
1650 bool "Enable debugging of DMA-API usage"
1651 select NEED_DMA_MAP_STATE
1653 Enable this option to debug the use of the DMA API by device drivers.
1654 With this option you will be able to detect common bugs in device
1655 drivers like double-freeing of DMA mappings or freeing mappings that
1656 were never allocated.
1658 This also attempts to catch cases where a page owned by DMA is
1659 accessed by the cpu in a way that could cause data corruption. For
1660 example, this enables cow_user_page() to check that the source page is
1663 This option causes a performance degradation. Use only if you want to
1664 debug device drivers and dma interactions.
1668 config DMA_API_DEBUG_SG
1669 bool "Debug DMA scatter-gather usage"
1671 depends on DMA_API_DEBUG
1673 Perform extra checking that callers of dma_map_sg() have respected the
1674 appropriate segment length/boundary limits for the given device when
1675 preparing DMA scatterlists.
1677 This is particularly likely to have been overlooked in cases where the
1678 dma_map_sg() API is used for general bulk mapping of pages rather than
1679 preparing literal scatter-gather descriptors, where there is a risk of
1680 unexpected behaviour from DMA API implementations if the scatterlist
1681 is technically out-of-spec.
1685 menuconfig RUNTIME_TESTING_MENU
1686 bool "Runtime Testing"
1689 if RUNTIME_TESTING_MENU
1692 tristate "Linux Kernel Dump Test Tool Module"
1697 This module enables testing of the different dumping mechanisms by
1698 inducing system failures at predefined crash points.
1699 If you don't need it: say N
1700 Choose M here to compile this code as a module. The module will be
1703 Documentation on how to use the module can be found in
1704 Documentation/fault-injection/provoke-crashes.txt
1706 config TEST_LIST_SORT
1707 tristate "Linked list sorting test"
1708 depends on DEBUG_KERNEL || m
1710 Enable this to turn on 'list_sort()' function test. This test is
1711 executed only once during system boot (so affects only boot time),
1712 or at module load time.
1717 tristate "Array-based sort test"
1718 depends on DEBUG_KERNEL || m
1720 This option enables the self-test function of 'sort()' at boot,
1721 or at module load time.
1725 config KPROBES_SANITY_TEST
1726 bool "Kprobes sanity tests"
1727 depends on DEBUG_KERNEL
1731 This option provides for testing basic kprobes functionality on
1732 boot. Samples of kprobe and kretprobe are inserted and
1733 verified for functionality.
1735 Say N if you are unsure.
1737 config BACKTRACE_SELF_TEST
1738 tristate "Self test for the backtrace code"
1739 depends on DEBUG_KERNEL
1742 This option provides a kernel module that can be used to test
1743 the kernel stack backtrace code. This option is not useful
1744 for distributions or general kernels, but only for kernel
1745 developers working on architecture code.
1747 Note that if you want to also test saved backtraces, you will
1748 have to enable STACKTRACE as well.
1750 Say N if you are unsure.
1753 tristate "Red-Black tree test"
1754 depends on DEBUG_KERNEL
1756 A benchmark measuring the performance of the rbtree library.
1757 Also includes rbtree invariant checks.
1759 config INTERVAL_TREE_TEST
1760 tristate "Interval tree test"
1761 depends on DEBUG_KERNEL
1762 select INTERVAL_TREE
1764 A benchmark measuring the performance of the interval tree library
1767 tristate "Per cpu operations test"
1768 depends on m && DEBUG_KERNEL
1770 Enable this option to build test module which validates per-cpu
1775 config ATOMIC64_SELFTEST
1776 tristate "Perform an atomic64_t self-test"
1778 Enable this option to test the atomic64_t functions at boot or
1779 at module load time.
1783 config ASYNC_RAID6_TEST
1784 tristate "Self test for hardware accelerated raid6 recovery"
1785 depends on ASYNC_RAID6_RECOV
1788 This is a one-shot self test that permutes through the
1789 recovery of all the possible two disk failure scenarios for a
1790 N-disk array. Recovery is performed with the asynchronous
1791 raid6 recovery routines, and will optionally use an offload
1792 engine if one is available.
1797 tristate "Test functions located in the hexdump module at runtime"
1799 config TEST_STRING_HELPERS
1800 tristate "Test functions located in the string_helpers module at runtime"
1803 tristate "Test kstrto*() family of functions at runtime"
1806 tristate "Test printf() family of functions at runtime"
1809 tristate "Test bitmap_*() family of functions at runtime"
1812 Enable this option to test the bitmap functions at boot.
1817 tristate "Test functions located in the uuid module at runtime"
1819 config TEST_OVERFLOW
1820 tristate "Test check_*_overflow() functions at runtime"
1822 config TEST_RHASHTABLE
1823 tristate "Perform selftest on resizable hash table"
1826 Enable this option to test the rhashtable functions at boot.
1831 tristate "Perform selftest on hash functions"
1834 Enable this option to test the kernel's integer (<linux/hash.h>),
1835 string (<linux/stringhash.h>), and siphash (<linux/siphash.h>)
1836 hash functions on boot (or module load).
1838 This is intended to help people writing architecture-specific
1839 optimized versions. If unsure, say N.
1842 tristate "Perform selftest on priority array manager"
1846 Enable this option to test priority array manager on boot
1852 tristate "Test module loading with 'hello world' module"
1856 This builds the "test_module" module that emits "Hello, world"
1857 on printk when loaded. It is designed to be used for basic
1858 evaluation of the module loading subsystem (for example when
1859 validating module verification). It lacks any extra dependencies,
1860 and will not normally be loaded by the system unless explicitly
1865 config TEST_USER_COPY
1866 tristate "Test user/kernel boundary protections"
1870 This builds the "test_user_copy" module that runs sanity checks
1871 on the copy_to/from_user infrastructure, making sure basic
1872 user/kernel boundary testing is working. If it fails to load,
1873 a regression has been detected in the user/kernel memory boundary
1879 tristate "Test BPF filter functionality"
1883 This builds the "test_bpf" module that runs various test vectors
1884 against the BPF interpreter or BPF JIT compiler depending on the
1885 current setting. This is in particular useful for BPF JIT compiler
1886 development, but also to run regression tests against changes in
1887 the interpreter code. It also enables test stubs for eBPF maps and
1888 verifier used by user space verifier testsuite.
1892 config FIND_BIT_BENCHMARK
1893 tristate "Test find_bit functions"
1896 This builds the "test_find_bit" module that measure find_*_bit()
1897 functions performance.
1901 config TEST_FIRMWARE
1902 tristate "Test firmware loading via userspace interface"
1904 depends on FW_LOADER
1906 This builds the "test_firmware" module that creates a userspace
1907 interface for testing firmware loading. This can be used to
1908 control the triggering of firmware loading without needing an
1909 actual firmware-using device. The contents can be rechecked by
1915 tristate "sysctl test driver"
1917 depends on PROC_SYSCTL
1919 This builds the "test_sysctl" module. This driver enables to test the
1920 proc sysctl interfaces available to drivers safely without affecting
1921 production knobs which might alter system functionality.
1926 tristate "udelay test driver"
1929 This builds the "udelay_test" module that helps to make sure
1930 that udelay() is working properly.
1934 config TEST_STATIC_KEYS
1935 tristate "Test static keys"
1939 Test the static key interfaces.
1944 tristate "kmod stress tester"
1947 depends on BLOCK && (64BIT || LBDAF) # for XFS, BTRFS
1948 depends on NETDEVICES && NET_CORE && INET # for TUN
1954 Test the kernel's module loading mechanism: kmod. kmod implements
1955 support to load modules using the Linux kernel's usermode helper.
1956 This test provides a series of tests against kmod.
1958 Although technically you can either build test_kmod as a module or
1959 into the kernel we disallow building it into the kernel since
1960 it stress tests request_module() and this will very likely cause
1961 some issues by taking over precious threads available from other
1962 module load requests, ultimately this could be fatal.
1966 tools/testing/selftests/kmod/kmod.sh --help
1970 config TEST_DEBUG_VIRTUAL
1971 tristate "Test CONFIG_DEBUG_VIRTUAL feature"
1972 depends on DEBUG_VIRTUAL
1974 Test the kernel's ability to detect incorrect calls to
1975 virt_to_phys() done against the non-linear part of the
1976 kernel's virtual address map.
1980 endif # RUNTIME_TESTING_MENU
1984 depends on HAVE_MEMBLOCK
1986 This option adds a kernel parameter 'memtest', which allows memtest
1988 memtest=0, mean disabled; -- default
1989 memtest=1, mean do 1 test pattern;
1991 memtest=17, mean do 17 test patterns.
1992 If you are unsure how to answer this question, answer N.
1994 config BUG_ON_DATA_CORRUPTION
1995 bool "Trigger a BUG when data corruption is detected"
1998 Select this option if the kernel should BUG when it encounters
1999 data corruption in kernel memory structures when they get checked
2004 source "samples/Kconfig"
2006 source "lib/Kconfig.kgdb"
2008 source "lib/Kconfig.ubsan"
2010 config ARCH_HAS_DEVMEM_IS_ALLOWED
2013 config STRICT_DEVMEM
2014 bool "Filter access to /dev/mem"
2015 depends on MMU && DEVMEM
2016 depends on ARCH_HAS_DEVMEM_IS_ALLOWED
2017 default y if PPC || X86 || ARM64
2019 If this option is disabled, you allow userspace (root) access to all
2020 of memory, including kernel and userspace memory. Accidental
2021 access to this is obviously disastrous, but specific access can
2022 be used by people debugging the kernel. Note that with PAT support
2023 enabled, even in this case there are restrictions on /dev/mem
2024 use due to the cache aliasing requirements.
2026 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem
2027 file only allows userspace access to PCI space and the BIOS code and
2028 data regions. This is sufficient for dosemu and X and all common
2033 config IO_STRICT_DEVMEM
2034 bool "Filter I/O access to /dev/mem"
2035 depends on STRICT_DEVMEM
2037 If this option is disabled, you allow userspace (root) access to all
2038 io-memory regardless of whether a driver is actively using that
2039 range. Accidental access to this is obviously disastrous, but
2040 specific access can be used by people debugging kernel drivers.
2042 If this option is switched on, the /dev/mem file only allows
2043 userspace access to *idle* io-memory ranges (see /proc/iomem) This
2044 may break traditional users of /dev/mem (dosemu, legacy X, etc...)
2045 if the driver using a given range cannot be disabled.