1 # SPDX-License-Identifier: GPL-2.0-only
4 menu "printk and dmesg options"
7 bool "Show timing information on printks"
10 Selecting this option causes time stamps of the printk()
11 messages to be added to the output of the syslog() system
12 call and at the console.
14 The timestamp is always recorded internally, and exported
15 to /dev/kmsg. This flag just specifies if the timestamp should
16 be included, not that the timestamp is recorded.
18 The behavior is also controlled by the kernel command line
19 parameter printk.time=1. See Documentation/admin-guide/kernel-parameters.rst
22 bool "Show caller information on printks"
25 Selecting this option causes printk() to add a caller "thread id" (if
26 in task context) or a caller "processor id" (if not in task context)
29 This option is intended for environments where multiple threads
30 concurrently call printk() for many times, for it is difficult to
31 interpret without knowing where these lines (or sometimes individual
32 line which was divided into multiple lines due to race) came from.
34 Since toggling after boot makes the code racy, currently there is
35 no option to enable/disable at the kernel command line parameter or
38 config CONSOLE_LOGLEVEL_DEFAULT
39 int "Default console loglevel (1-15)"
43 Default loglevel to determine what will be printed on the console.
45 Setting a default here is equivalent to passing in loglevel=<x> in
46 the kernel bootargs. loglevel=<x> continues to override whatever
47 value is specified here as well.
49 Note: This does not affect the log level of un-prefixed printk()
50 usage in the kernel. That is controlled by the MESSAGE_LOGLEVEL_DEFAULT
53 config CONSOLE_LOGLEVEL_QUIET
54 int "quiet console loglevel (1-15)"
58 loglevel to use when "quiet" is passed on the kernel commandline.
60 When "quiet" is passed on the kernel commandline this loglevel
61 will be used as the loglevel. IOW passing "quiet" will be the
62 equivalent of passing "loglevel=<CONSOLE_LOGLEVEL_QUIET>"
64 config MESSAGE_LOGLEVEL_DEFAULT
65 int "Default message log level (1-7)"
69 Default log level for printk statements with no specified priority.
71 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks
72 that are auditing their logs closely may want to set it to a lower
75 Note: This does not affect what message level gets printed on the console
76 by default. To change that, use loglevel=<x> in the kernel bootargs,
77 or pick a different CONSOLE_LOGLEVEL_DEFAULT configuration value.
79 config BOOT_PRINTK_DELAY
80 bool "Delay each boot printk message by N milliseconds"
81 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
83 This build option allows you to read kernel boot messages
84 by inserting a short delay after each one. The delay is
85 specified in milliseconds on the kernel command line,
88 It is likely that you would also need to use "lpj=M" to preset
89 the "loops per jiffie" value.
90 See a previous boot log for the "lpj" value to use for your
91 system, and then set "lpj=M" before setting "boot_delay=N".
92 NOTE: Using this option may adversely affect SMP systems.
93 I.e., processors other than the first one may not boot up.
94 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect
95 what it believes to be lockup conditions.
98 bool "Enable dynamic printk() support"
101 depends on (DEBUG_FS || PROC_FS)
104 Compiles debug level messages into the kernel, which would not
105 otherwise be available at runtime. These messages can then be
106 enabled/disabled based on various levels of scope - per source file,
107 function, module, format string, and line number. This mechanism
108 implicitly compiles in all pr_debug() and dev_dbg() calls, which
109 enlarges the kernel text size by about 2%.
111 If a source file is compiled with DEBUG flag set, any
112 pr_debug() calls in it are enabled by default, but can be
113 disabled at runtime as below. Note that DEBUG flag is
114 turned on by many CONFIG_*DEBUG* options.
118 Dynamic debugging is controlled via the 'dynamic_debug/control' file,
119 which is contained in the 'debugfs' filesystem or procfs.
120 Thus, the debugfs or procfs filesystem must first be mounted before
121 making use of this feature.
122 We refer the control file as: <debugfs>/dynamic_debug/control. This
123 file contains a list of the debug statements that can be enabled. The
124 format for each line of the file is:
126 filename:lineno [module]function flags format
128 filename : source file of the debug statement
129 lineno : line number of the debug statement
130 module : module that contains the debug statement
131 function : function that contains the debug statement
132 flags : '=p' means the line is turned 'on' for printing
133 format : the format used for the debug statement
137 nullarbor:~ # cat <debugfs>/dynamic_debug/control
138 # filename:lineno [module]function flags format
139 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"
140 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"
141 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"
145 // enable the message at line 1603 of file svcsock.c
146 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >
147 <debugfs>/dynamic_debug/control
149 // enable all the messages in file svcsock.c
150 nullarbor:~ # echo -n 'file svcsock.c +p' >
151 <debugfs>/dynamic_debug/control
153 // enable all the messages in the NFS server module
154 nullarbor:~ # echo -n 'module nfsd +p' >
155 <debugfs>/dynamic_debug/control
157 // enable all 12 messages in the function svc_process()
158 nullarbor:~ # echo -n 'func svc_process +p' >
159 <debugfs>/dynamic_debug/control
161 // disable all 12 messages in the function svc_process()
162 nullarbor:~ # echo -n 'func svc_process -p' >
163 <debugfs>/dynamic_debug/control
165 See Documentation/admin-guide/dynamic-debug-howto.rst for additional
168 config SYMBOLIC_ERRNAME
169 bool "Support symbolic error names in printf"
172 If you say Y here, the kernel's printf implementation will
173 be able to print symbolic error names such as ENOSPC instead
174 of the number 28. It makes the kernel image slightly larger
175 (about 3KB), but can make the kernel logs easier to read.
177 config DEBUG_BUGVERBOSE
178 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT
179 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)
182 Say Y here to make BUG() panics output the file name and line number
183 of the BUG call as well as the EIP and oops trace. This aids
184 debugging but costs about 70-100K of memory.
186 endmenu # "printk and dmesg options"
188 menu "Compile-time checks and compiler options"
191 bool "Compile the kernel with debug info"
192 depends on DEBUG_KERNEL && !COMPILE_TEST
194 If you say Y here the resulting kernel image will include
195 debugging info resulting in a larger kernel image.
196 This adds debug symbols to the kernel and modules (gcc -g), and
197 is needed if you intend to use kernel crashdump or binary object
198 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
199 Say Y here only if you plan to debug the kernel.
203 config DEBUG_INFO_REDUCED
204 bool "Reduce debugging information"
205 depends on DEBUG_INFO
207 If you say Y here gcc is instructed to generate less debugging
208 information for structure types. This means that tools that
209 need full debugging information (like kgdb or systemtap) won't
210 be happy. But if you merely need debugging information to
211 resolve line numbers there is no loss. Advantage is that
212 build directory object sizes shrink dramatically over a full
213 DEBUG_INFO build and compile times are reduced too.
214 Only works with newer gcc versions.
216 config DEBUG_INFO_COMPRESSED
217 bool "Compressed debugging information"
218 depends on DEBUG_INFO
219 depends on $(cc-option,-gz=zlib)
220 depends on $(as-option,-Wa$(comma)--compress-debug-sections=zlib)
221 depends on $(ld-option,--compress-debug-sections=zlib)
223 Compress the debug information using zlib. Requires GCC 5.0+ or Clang
224 5.0+, binutils 2.26+, and zlib.
226 Users of dpkg-deb via scripts/package/builddeb may find an increase in
227 size of their debug .deb packages with this config set, due to the
228 debug info being compressed with zlib, then the object files being
229 recompressed with a different compression scheme. But this is still
230 preferable to setting $KDEB_COMPRESS to "none" which would be even
233 config DEBUG_INFO_SPLIT
234 bool "Produce split debuginfo in .dwo files"
235 depends on DEBUG_INFO
236 depends on $(cc-option,-gsplit-dwarf)
238 Generate debug info into separate .dwo files. This significantly
239 reduces the build directory size for builds with DEBUG_INFO,
240 because it stores the information only once on disk in .dwo
241 files instead of multiple times in object files and executables.
242 In addition the debug information is also compressed.
244 Requires recent gcc (4.7+) and recent gdb/binutils.
245 Any tool that packages or reads debug information would need
246 to know about the .dwo files and include them.
247 Incompatible with older versions of ccache.
249 config DEBUG_INFO_DWARF4
250 bool "Generate dwarf4 debuginfo"
251 depends on DEBUG_INFO
252 depends on $(cc-option,-gdwarf-4)
254 Generate dwarf4 debug info. This requires recent versions
255 of gcc and gdb. It makes the debug information larger.
256 But it significantly improves the success of resolving
257 variables in gdb on optimized code.
259 config DEBUG_INFO_BTF
260 bool "Generate BTF typeinfo"
261 depends on DEBUG_INFO
262 depends on !DEBUG_INFO_SPLIT && !DEBUG_INFO_REDUCED
263 depends on !GCC_PLUGIN_RANDSTRUCT || COMPILE_TEST
265 Generate deduplicated BTF type information from DWARF debug info.
266 Turning this on expects presence of pahole tool, which will convert
267 DWARF type info into equivalent deduplicated BTF type info.
270 bool "Provide GDB scripts for kernel debugging"
271 depends on DEBUG_INFO
273 This creates the required links to GDB helper scripts in the
274 build directory. If you load vmlinux into gdb, the helper
275 scripts will be automatically imported by gdb as well, and
276 additional functions are available to analyze a Linux kernel
277 instance. See Documentation/dev-tools/gdb-kernel-debugging.rst
280 config ENABLE_MUST_CHECK
281 bool "Enable __must_check logic"
284 Enable the __must_check logic in the kernel build. Disable this to
285 suppress the "warning: ignoring return value of 'foo', declared with
286 attribute warn_unused_result" messages.
289 int "Warn for stack frames larger than"
291 default 2048 if GCC_PLUGIN_LATENT_ENTROPY
292 default 1280 if (!64BIT && PARISC)
293 default 1024 if (!64BIT && !PARISC)
294 default 2048 if 64BIT
296 Tell gcc to warn at build time for stack frames larger than this.
297 Setting this too low will cause a lot of warnings.
298 Setting it to 0 disables the warning.
300 config STRIP_ASM_SYMS
301 bool "Strip assembler-generated symbols during link"
304 Strip internal assembler-generated symbols during a link (symbols
305 that look like '.Lxxx') so they don't pollute the output of
306 get_wchan() and suchlike.
309 bool "Generate readable assembler code"
310 depends on DEBUG_KERNEL
312 Disable some compiler optimizations that tend to generate human unreadable
313 assembler output. This may make the kernel slightly slower, but it helps
314 to keep kernel developers who have to stare a lot at assembler listings
317 config HEADERS_INSTALL
318 bool "Install uapi headers to usr/include"
321 This option will install uapi headers (headers exported to user-space)
322 into the usr/include directory for use during the kernel build.
323 This is unneeded for building the kernel itself, but needed for some
324 user-space program samples. It is also needed by some features such
325 as uapi header sanity checks.
327 config DEBUG_SECTION_MISMATCH
328 bool "Enable full Section mismatch analysis"
330 The section mismatch analysis checks if there are illegal
331 references from one section to another section.
332 During linktime or runtime, some sections are dropped;
333 any use of code/data previously in these sections would
334 most likely result in an oops.
335 In the code, functions and variables are annotated with
336 __init,, etc. (see the full list in include/linux/init.h),
337 which results in the code/data being placed in specific sections.
338 The section mismatch analysis is always performed after a full
339 kernel build, and enabling this option causes the following
340 additional step to occur:
341 - Add the option -fno-inline-functions-called-once to gcc commands.
342 When inlining a function annotated with __init in a non-init
343 function, we would lose the section information and thus
344 the analysis would not catch the illegal reference.
345 This option tells gcc to inline less (but it does result in
348 config SECTION_MISMATCH_WARN_ONLY
349 bool "Make section mismatch errors non-fatal"
352 If you say N here, the build process will fail if there are any
353 section mismatch, instead of just throwing warnings.
358 # Select this config option from the architecture Kconfig, if it
359 # is preferred to always offer frame pointers as a config
360 # option on the architecture (regardless of KERNEL_DEBUG):
362 config ARCH_WANT_FRAME_POINTERS
366 bool "Compile the kernel with frame pointers"
367 depends on DEBUG_KERNEL && (M68K || UML || SUPERH) || ARCH_WANT_FRAME_POINTERS
368 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS
370 If you say Y here the resulting kernel image will be slightly
371 larger and slower, but it gives very useful debugging information
372 in case of kernel bugs. (precise oopses/stacktraces/warnings)
374 config STACK_VALIDATION
375 bool "Compile-time stack metadata validation"
376 depends on HAVE_STACK_VALIDATION
379 Add compile-time checks to validate stack metadata, including frame
380 pointers (if CONFIG_FRAME_POINTER is enabled). This helps ensure
381 that runtime stack traces are more reliable.
383 This is also a prerequisite for generation of ORC unwind data, which
384 is needed for CONFIG_UNWINDER_ORC.
386 For more information, see
387 tools/objtool/Documentation/stack-validation.txt.
389 config VMLINUX_VALIDATION
391 depends on STACK_VALIDATION && DEBUG_ENTRY && !PARAVIRT
394 config DEBUG_FORCE_WEAK_PER_CPU
395 bool "Force weak per-cpu definitions"
396 depends on DEBUG_KERNEL
398 s390 and alpha require percpu variables in modules to be
399 defined weak to work around addressing range issue which
400 puts the following two restrictions on percpu variable
403 1. percpu symbols must be unique whether static or not
404 2. percpu variables can't be defined inside a function
406 To ensure that generic code follows the above rules, this
407 option forces all percpu variables to be defined as weak.
409 endmenu # "Compiler options"
411 menu "Generic Kernel Debugging Instruments"
414 bool "Magic SysRq key"
417 If you say Y here, you will have some control over the system even
418 if the system crashes for example during kernel debugging (e.g., you
419 will be able to flush the buffer cache to disk, reboot the system
420 immediately or dump some status information). This is accomplished
421 by pressing various keys while holding SysRq (Alt+PrintScreen). It
422 also works on a serial console (on PC hardware at least), if you
423 send a BREAK and then within 5 seconds a command keypress. The
424 keys are documented in <file:Documentation/admin-guide/sysrq.rst>.
425 Don't say Y unless you really know what this hack does.
427 config MAGIC_SYSRQ_DEFAULT_ENABLE
428 hex "Enable magic SysRq key functions by default"
429 depends on MAGIC_SYSRQ
432 Specifies which SysRq key functions are enabled by default.
433 This may be set to 1 or 0 to enable or disable them all, or
434 to a bitmask as described in Documentation/admin-guide/sysrq.rst.
436 config MAGIC_SYSRQ_SERIAL
437 bool "Enable magic SysRq key over serial"
438 depends on MAGIC_SYSRQ
441 Many embedded boards have a disconnected TTL level serial which can
442 generate some garbage that can lead to spurious false sysrq detects.
443 This option allows you to decide whether you want to enable the
446 config MAGIC_SYSRQ_SERIAL_SEQUENCE
447 string "Char sequence that enables magic SysRq over serial"
448 depends on MAGIC_SYSRQ_SERIAL
451 Specifies a sequence of characters that can follow BREAK to enable
452 SysRq on a serial console.
454 If unsure, leave an empty string and the option will not be enabled.
457 bool "Debug Filesystem"
459 debugfs is a virtual file system that kernel developers use to put
460 debugging files into. Enable this option to be able to read and
461 write to these files.
463 For detailed documentation on the debugfs API, see
464 Documentation/filesystems/.
468 source "lib/Kconfig.kgdb"
470 source "lib/Kconfig.ubsan"
475 bool "Kernel debugging"
477 Say Y here if you are developing drivers or trying to debug and
478 identify kernel problems.
481 bool "Miscellaneous debug code"
483 depends on DEBUG_KERNEL
485 Say Y here if you need to enable miscellaneous debug code that should
486 be under a more specific debug option but isn't.
489 menu "Memory Debugging"
491 source "mm/Kconfig.debug"
494 bool "Debug object operations"
495 depends on DEBUG_KERNEL
497 If you say Y here, additional code will be inserted into the
498 kernel to track the life time of various objects and validate
499 the operations on those objects.
501 config DEBUG_OBJECTS_SELFTEST
502 bool "Debug objects selftest"
503 depends on DEBUG_OBJECTS
505 This enables the selftest of the object debug code.
507 config DEBUG_OBJECTS_FREE
508 bool "Debug objects in freed memory"
509 depends on DEBUG_OBJECTS
511 This enables checks whether a k/v free operation frees an area
512 which contains an object which has not been deactivated
513 properly. This can make kmalloc/kfree-intensive workloads
516 config DEBUG_OBJECTS_TIMERS
517 bool "Debug timer objects"
518 depends on DEBUG_OBJECTS
520 If you say Y here, additional code will be inserted into the
521 timer routines to track the life time of timer objects and
522 validate the timer operations.
524 config DEBUG_OBJECTS_WORK
525 bool "Debug work objects"
526 depends on DEBUG_OBJECTS
528 If you say Y here, additional code will be inserted into the
529 work queue routines to track the life time of work objects and
530 validate the work operations.
532 config DEBUG_OBJECTS_RCU_HEAD
533 bool "Debug RCU callbacks objects"
534 depends on DEBUG_OBJECTS
536 Enable this to turn on debugging of RCU list heads (call_rcu() usage).
538 config DEBUG_OBJECTS_PERCPU_COUNTER
539 bool "Debug percpu counter objects"
540 depends on DEBUG_OBJECTS
542 If you say Y here, additional code will be inserted into the
543 percpu counter routines to track the life time of percpu counter
544 objects and validate the percpu counter operations.
546 config DEBUG_OBJECTS_ENABLE_DEFAULT
547 int "debug_objects bootup default value (0-1)"
550 depends on DEBUG_OBJECTS
552 Debug objects boot parameter default value
555 bool "Debug slab memory allocations"
556 depends on DEBUG_KERNEL && SLAB
558 Say Y here to have the kernel do limited verification on memory
559 allocation as well as poisoning memory on free to catch use of freed
560 memory. This can make kmalloc/kfree-intensive workloads much slower.
563 bool "SLUB debugging on by default"
564 depends on SLUB && SLUB_DEBUG
567 Boot with debugging on by default. SLUB boots by default with
568 the runtime debug capabilities switched off. Enabling this is
569 equivalent to specifying the "slub_debug" parameter on boot.
570 There is no support for more fine grained debug control like
571 possible with slub_debug=xxx. SLUB debugging may be switched
572 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
577 bool "Enable SLUB performance statistics"
578 depends on SLUB && SYSFS
580 SLUB statistics are useful to debug SLUBs allocation behavior in
581 order find ways to optimize the allocator. This should never be
582 enabled for production use since keeping statistics slows down
583 the allocator by a few percentage points. The slabinfo command
584 supports the determination of the most active slabs to figure
585 out which slabs are relevant to a particular load.
586 Try running: slabinfo -DA
588 config HAVE_DEBUG_KMEMLEAK
591 config DEBUG_KMEMLEAK
592 bool "Kernel memory leak detector"
593 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK
595 select STACKTRACE if STACKTRACE_SUPPORT
599 Say Y here if you want to enable the memory leak
600 detector. The memory allocation/freeing is traced in a way
601 similar to the Boehm's conservative garbage collector, the
602 difference being that the orphan objects are not freed but
603 only shown in /sys/kernel/debug/kmemleak. Enabling this
604 feature will introduce an overhead to memory
605 allocations. See Documentation/dev-tools/kmemleak.rst for more
608 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances
609 of finding leaks due to the slab objects poisoning.
611 In order to access the kmemleak file, debugfs needs to be
612 mounted (usually at /sys/kernel/debug).
614 config DEBUG_KMEMLEAK_MEM_POOL_SIZE
615 int "Kmemleak memory pool size"
616 depends on DEBUG_KMEMLEAK
620 Kmemleak must track all the memory allocations to avoid
621 reporting false positives. Since memory may be allocated or
622 freed before kmemleak is fully initialised, use a static pool
623 of metadata objects to track such callbacks. After kmemleak is
624 fully initialised, this memory pool acts as an emergency one
625 if slab allocations fail.
627 config DEBUG_KMEMLEAK_TEST
628 tristate "Simple test for the kernel memory leak detector"
629 depends on DEBUG_KMEMLEAK && m
631 This option enables a module that explicitly leaks memory.
635 config DEBUG_KMEMLEAK_DEFAULT_OFF
636 bool "Default kmemleak to off"
637 depends on DEBUG_KMEMLEAK
639 Say Y here to disable kmemleak by default. It can then be enabled
640 on the command line via kmemleak=on.
642 config DEBUG_KMEMLEAK_AUTO_SCAN
643 bool "Enable kmemleak auto scan thread on boot up"
645 depends on DEBUG_KMEMLEAK
647 Depending on the cpu, kmemleak scan may be cpu intensive and can
648 stall user tasks at times. This option enables/disables automatic
649 kmemleak scan at boot up.
651 Say N here to disable kmemleak auto scan thread to stop automatic
652 scanning. Disabling this option disables automatic reporting of
657 config DEBUG_STACK_USAGE
658 bool "Stack utilization instrumentation"
659 depends on DEBUG_KERNEL && !IA64
661 Enables the display of the minimum amount of free stack which each
662 task has ever had available in the sysrq-T and sysrq-P debug output.
664 This option will slow down process creation somewhat.
666 config SCHED_STACK_END_CHECK
667 bool "Detect stack corruption on calls to schedule()"
668 depends on DEBUG_KERNEL
671 This option checks for a stack overrun on calls to schedule().
672 If the stack end location is found to be over written always panic as
673 the content of the corrupted region can no longer be trusted.
674 This is to ensure no erroneous behaviour occurs which could result in
675 data corruption or a sporadic crash at a later stage once the region
676 is examined. The runtime overhead introduced is minimal.
678 config ARCH_HAS_DEBUG_VM_PGTABLE
681 An architecture should select this when it can successfully
682 build and run DEBUG_VM_PGTABLE.
686 depends on DEBUG_KERNEL
688 Enable this to turn on extended checks in the virtual-memory system
689 that may impact performance.
693 config DEBUG_VM_VMACACHE
694 bool "Debug VMA caching"
697 Enable this to turn on VMA caching debug information. Doing so
698 can cause significant overhead, so only enable it in non-production
704 bool "Debug VM red-black trees"
707 Enable VM red-black tree debugging information and extra validations.
711 config DEBUG_VM_PGFLAGS
712 bool "Debug page-flags operations"
715 Enables extra validation on page flags operations.
719 config DEBUG_VM_PGTABLE
720 bool "Debug arch page table for semantics compliance"
722 depends on ARCH_HAS_DEBUG_VM_PGTABLE
723 default y if DEBUG_VM
725 This option provides a debug method which can be used to test
726 architecture page table helper functions on various platforms in
727 verifying if they comply with expected generic MM semantics. This
728 will help architecture code in making sure that any changes or
729 new additions of these helpers still conform to expected
730 semantics of the generic MM. Platforms will have to opt in for
731 this through ARCH_HAS_DEBUG_VM_PGTABLE.
735 config ARCH_HAS_DEBUG_VIRTUAL
739 bool "Debug VM translations"
740 depends on DEBUG_KERNEL && ARCH_HAS_DEBUG_VIRTUAL
742 Enable some costly sanity checks in virtual to page code. This can
743 catch mistakes with virt_to_page() and friends.
747 config DEBUG_NOMMU_REGIONS
748 bool "Debug the global anon/private NOMMU mapping region tree"
749 depends on DEBUG_KERNEL && !MMU
751 This option causes the global tree of anonymous and private mapping
752 regions to be regularly checked for invalid topology.
754 config DEBUG_MEMORY_INIT
755 bool "Debug memory initialisation" if EXPERT
758 Enable this for additional checks during memory initialisation.
759 The sanity checks verify aspects of the VM such as the memory model
760 and other information provided by the architecture. Verbose
761 information will be printed at KERN_DEBUG loglevel depending
762 on the mminit_loglevel= command-line option.
766 config MEMORY_NOTIFIER_ERROR_INJECT
767 tristate "Memory hotplug notifier error injection module"
768 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION
770 This option provides the ability to inject artificial errors to
771 memory hotplug notifier chain callbacks. It is controlled through
772 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory
774 If the notifier call chain should be failed with some events
775 notified, write the error code to "actions/<notifier event>/error".
777 Example: Inject memory hotplug offline error (-12 == -ENOMEM)
779 # cd /sys/kernel/debug/notifier-error-inject/memory
780 # echo -12 > actions/MEM_GOING_OFFLINE/error
781 # echo offline > /sys/devices/system/memory/memoryXXX/state
782 bash: echo: write error: Cannot allocate memory
784 To compile this code as a module, choose M here: the module will
785 be called memory-notifier-error-inject.
789 config DEBUG_PER_CPU_MAPS
790 bool "Debug access to per_cpu maps"
791 depends on DEBUG_KERNEL
794 Say Y to verify that the per_cpu map being accessed has
795 been set up. This adds a fair amount of code to kernel memory
796 and decreases performance.
801 bool "Highmem debugging"
802 depends on DEBUG_KERNEL && HIGHMEM
804 This option enables additional error checking for high memory
805 systems. Disable for production systems.
807 config HAVE_DEBUG_STACKOVERFLOW
810 config DEBUG_STACKOVERFLOW
811 bool "Check for stack overflows"
812 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW
814 Say Y here if you want to check for overflows of kernel, IRQ
815 and exception stacks (if your architecture uses them). This
816 option will show detailed messages if free stack space drops
817 below a certain limit.
819 These kinds of bugs usually occur when call-chains in the
820 kernel get too deep, especially when interrupts are
823 Use this in cases where you see apparently random memory
824 corruption, especially if it appears in 'struct thread_info'
826 If in doubt, say "N".
828 source "lib/Kconfig.kasan"
830 endmenu # "Memory Debugging"
833 bool "Debug shared IRQ handlers"
834 depends on DEBUG_KERNEL
836 Enable this to generate a spurious interrupt as soon as a shared
837 interrupt handler is registered, and just before one is deregistered.
838 Drivers ought to be able to handle interrupts coming in at those
839 points; some don't and need to be caught.
841 menu "Debug Oops, Lockups and Hangs"
846 Say Y here to enable the kernel to panic when it oopses. This
847 has the same effect as setting oops=panic on the kernel command
850 This feature is useful to ensure that the kernel does not do
851 anything erroneous after an oops which could result in data
852 corruption or other issues.
856 config PANIC_ON_OOPS_VALUE
859 default 0 if !PANIC_ON_OOPS
860 default 1 if PANIC_ON_OOPS
866 Set the timeout value (in seconds) until a reboot occurs when the
867 the kernel panics. If n = 0, then we wait forever. A timeout
868 value n > 0 will wait n seconds before rebooting, while a timeout
869 value n < 0 will reboot immediately.
871 config LOCKUP_DETECTOR
874 config SOFTLOCKUP_DETECTOR
875 bool "Detect Soft Lockups"
876 depends on DEBUG_KERNEL && !S390
877 select LOCKUP_DETECTOR
879 Say Y here to enable the kernel to act as a watchdog to detect
882 Softlockups are bugs that cause the kernel to loop in kernel
883 mode for more than 20 seconds, without giving other tasks a
884 chance to run. The current stack trace is displayed upon
885 detection and the system will stay locked up.
887 config BOOTPARAM_SOFTLOCKUP_PANIC
888 bool "Panic (Reboot) On Soft Lockups"
889 depends on SOFTLOCKUP_DETECTOR
891 Say Y here to enable the kernel to panic on "soft lockups",
892 which are bugs that cause the kernel to loop in kernel
893 mode for more than 20 seconds (configurable using the watchdog_thresh
894 sysctl), without giving other tasks a chance to run.
896 The panic can be used in combination with panic_timeout,
897 to cause the system to reboot automatically after a
898 lockup has been detected. This feature is useful for
899 high-availability systems that have uptime guarantees and
900 where a lockup must be resolved ASAP.
904 config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE
906 depends on SOFTLOCKUP_DETECTOR
908 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC
909 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC
911 config HARDLOCKUP_DETECTOR_PERF
913 select SOFTLOCKUP_DETECTOR
916 # Enables a timestamp based low pass filter to compensate for perf based
917 # hard lockup detection which runs too fast due to turbo modes.
919 config HARDLOCKUP_CHECK_TIMESTAMP
923 # arch/ can define HAVE_HARDLOCKUP_DETECTOR_ARCH to provide their own hard
924 # lockup detector rather than the perf based detector.
926 config HARDLOCKUP_DETECTOR
927 bool "Detect Hard Lockups"
928 depends on DEBUG_KERNEL && !S390
929 depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_ARCH
930 select LOCKUP_DETECTOR
931 select HARDLOCKUP_DETECTOR_PERF if HAVE_HARDLOCKUP_DETECTOR_PERF
932 select HARDLOCKUP_DETECTOR_ARCH if HAVE_HARDLOCKUP_DETECTOR_ARCH
934 Say Y here to enable the kernel to act as a watchdog to detect
937 Hardlockups are bugs that cause the CPU to loop in kernel mode
938 for more than 10 seconds, without letting other interrupts have a
939 chance to run. The current stack trace is displayed upon detection
940 and the system will stay locked up.
942 config BOOTPARAM_HARDLOCKUP_PANIC
943 bool "Panic (Reboot) On Hard Lockups"
944 depends on HARDLOCKUP_DETECTOR
946 Say Y here to enable the kernel to panic on "hard lockups",
947 which are bugs that cause the kernel to loop in kernel
948 mode with interrupts disabled for more than 10 seconds (configurable
949 using the watchdog_thresh sysctl).
953 config BOOTPARAM_HARDLOCKUP_PANIC_VALUE
955 depends on HARDLOCKUP_DETECTOR
957 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC
958 default 1 if BOOTPARAM_HARDLOCKUP_PANIC
960 config DETECT_HUNG_TASK
961 bool "Detect Hung Tasks"
962 depends on DEBUG_KERNEL
963 default SOFTLOCKUP_DETECTOR
965 Say Y here to enable the kernel to detect "hung tasks",
966 which are bugs that cause the task to be stuck in
967 uninterruptible "D" state indefinitely.
969 When a hung task is detected, the kernel will print the
970 current stack trace (which you should report), but the
971 task will stay in uninterruptible state. If lockdep is
972 enabled then all held locks will also be reported. This
973 feature has negligible overhead.
975 config DEFAULT_HUNG_TASK_TIMEOUT
976 int "Default timeout for hung task detection (in seconds)"
977 depends on DETECT_HUNG_TASK
980 This option controls the default timeout (in seconds) used
981 to determine when a task has become non-responsive and should
984 It can be adjusted at runtime via the kernel.hung_task_timeout_secs
985 sysctl or by writing a value to
986 /proc/sys/kernel/hung_task_timeout_secs.
988 A timeout of 0 disables the check. The default is two minutes.
989 Keeping the default should be fine in most cases.
991 config BOOTPARAM_HUNG_TASK_PANIC
992 bool "Panic (Reboot) On Hung Tasks"
993 depends on DETECT_HUNG_TASK
995 Say Y here to enable the kernel to panic on "hung tasks",
996 which are bugs that cause the kernel to leave a task stuck
997 in uninterruptible "D" state.
999 The panic can be used in combination with panic_timeout,
1000 to cause the system to reboot automatically after a
1001 hung task has been detected. This feature is useful for
1002 high-availability systems that have uptime guarantees and
1003 where a hung tasks must be resolved ASAP.
1007 config BOOTPARAM_HUNG_TASK_PANIC_VALUE
1009 depends on DETECT_HUNG_TASK
1011 default 0 if !BOOTPARAM_HUNG_TASK_PANIC
1012 default 1 if BOOTPARAM_HUNG_TASK_PANIC
1015 bool "Detect Workqueue Stalls"
1016 depends on DEBUG_KERNEL
1018 Say Y here to enable stall detection on workqueues. If a
1019 worker pool doesn't make forward progress on a pending work
1020 item for over a given amount of time, 30s by default, a
1021 warning message is printed along with dump of workqueue
1022 state. This can be configured through kernel parameter
1023 "workqueue.watchdog_thresh" and its sysfs counterpart.
1026 tristate "Test module to generate lockups"
1028 This builds the "test_lockup" module that helps to make sure
1029 that watchdogs and lockup detectors are working properly.
1031 Depending on module parameters it could emulate soft or hard
1032 lockup, "hung task", or locking arbitrary lock for a long time.
1033 Also it could generate series of lockups with cooling-down periods.
1037 endmenu # "Debug lockups and hangs"
1039 menu "Scheduler Debugging"
1042 bool "Collect scheduler debugging info"
1043 depends on DEBUG_KERNEL && PROC_FS
1046 If you say Y here, the /proc/sched_debug file will be provided
1047 that can help debug the scheduler. The runtime overhead of this
1055 bool "Collect scheduler statistics"
1056 depends on DEBUG_KERNEL && PROC_FS
1059 If you say Y here, additional code will be inserted into the
1060 scheduler and related routines to collect statistics about
1061 scheduler behavior and provide them in /proc/schedstat. These
1062 stats may be useful for both tuning and debugging the scheduler
1063 If you aren't debugging the scheduler or trying to tune a specific
1064 application, you can say N to avoid the very slight overhead
1069 config DEBUG_TIMEKEEPING
1070 bool "Enable extra timekeeping sanity checking"
1072 This option will enable additional timekeeping sanity checks
1073 which may be helpful when diagnosing issues where timekeeping
1074 problems are suspected.
1076 This may include checks in the timekeeping hotpaths, so this
1077 option may have a (very small) performance impact to some
1082 config DEBUG_PREEMPT
1083 bool "Debug preemptible kernel"
1084 depends on DEBUG_KERNEL && PREEMPTION && TRACE_IRQFLAGS_SUPPORT
1087 If you say Y here then the kernel will use a debug variant of the
1088 commonly used smp_processor_id() function and will print warnings
1089 if kernel code uses it in a preemption-unsafe way. Also, the kernel
1090 will detect preemption count underflows.
1092 menu "Lock Debugging (spinlocks, mutexes, etc...)"
1094 config LOCK_DEBUGGING_SUPPORT
1096 depends on TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
1099 config PROVE_LOCKING
1100 bool "Lock debugging: prove locking correctness"
1101 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1103 select DEBUG_SPINLOCK
1104 select DEBUG_MUTEXES
1105 select DEBUG_RT_MUTEXES if RT_MUTEXES
1107 select DEBUG_WW_MUTEX_SLOWPATH
1108 select DEBUG_LOCK_ALLOC
1109 select TRACE_IRQFLAGS
1112 This feature enables the kernel to prove that all locking
1113 that occurs in the kernel runtime is mathematically
1114 correct: that under no circumstance could an arbitrary (and
1115 not yet triggered) combination of observed locking
1116 sequences (on an arbitrary number of CPUs, running an
1117 arbitrary number of tasks and interrupt contexts) cause a
1120 In short, this feature enables the kernel to report locking
1121 related deadlocks before they actually occur.
1123 The proof does not depend on how hard and complex a
1124 deadlock scenario would be to trigger: how many
1125 participant CPUs, tasks and irq-contexts would be needed
1126 for it to trigger. The proof also does not depend on
1127 timing: if a race and a resulting deadlock is possible
1128 theoretically (no matter how unlikely the race scenario
1129 is), it will be proven so and will immediately be
1130 reported by the kernel (once the event is observed that
1131 makes the deadlock theoretically possible).
1133 If a deadlock is impossible (i.e. the locking rules, as
1134 observed by the kernel, are mathematically correct), the
1135 kernel reports nothing.
1137 NOTE: this feature can also be enabled for rwlocks, mutexes
1138 and rwsems - in which case all dependencies between these
1139 different locking variants are observed and mapped too, and
1140 the proof of observed correctness is also maintained for an
1141 arbitrary combination of these separate locking variants.
1143 For more details, see Documentation/locking/lockdep-design.rst.
1145 config PROVE_RAW_LOCK_NESTING
1146 bool "Enable raw_spinlock - spinlock nesting checks"
1147 depends on PROVE_LOCKING
1150 Enable the raw_spinlock vs. spinlock nesting checks which ensure
1151 that the lock nesting rules for PREEMPT_RT enabled kernels are
1154 NOTE: There are known nesting problems. So if you enable this
1155 option expect lockdep splats until these problems have been fully
1156 addressed which is work in progress. This config switch allows to
1157 identify and analyze these problems. It will be removed and the
1158 check permanentely enabled once the main issues have been fixed.
1160 If unsure, select N.
1163 bool "Lock usage statistics"
1164 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1166 select DEBUG_SPINLOCK
1167 select DEBUG_MUTEXES
1168 select DEBUG_RT_MUTEXES if RT_MUTEXES
1169 select DEBUG_LOCK_ALLOC
1172 This feature enables tracking lock contention points
1174 For more details, see Documentation/locking/lockstat.rst
1176 This also enables lock events required by "perf lock",
1178 If you want to use "perf lock", you also need to turn on
1179 CONFIG_EVENT_TRACING.
1181 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.
1182 (CONFIG_LOCKDEP defines "acquire" and "release" events.)
1184 config DEBUG_RT_MUTEXES
1185 bool "RT Mutex debugging, deadlock detection"
1186 depends on DEBUG_KERNEL && RT_MUTEXES
1188 This allows rt mutex semantics violations and rt mutex related
1189 deadlocks (lockups) to be detected and reported automatically.
1191 config DEBUG_SPINLOCK
1192 bool "Spinlock and rw-lock debugging: basic checks"
1193 depends on DEBUG_KERNEL
1194 select UNINLINE_SPIN_UNLOCK
1196 Say Y here and build SMP to catch missing spinlock initialization
1197 and certain other kinds of spinlock errors commonly made. This is
1198 best used in conjunction with the NMI watchdog so that spinlock
1199 deadlocks are also debuggable.
1201 config DEBUG_MUTEXES
1202 bool "Mutex debugging: basic checks"
1203 depends on DEBUG_KERNEL
1205 This feature allows mutex semantics violations to be detected and
1208 config DEBUG_WW_MUTEX_SLOWPATH
1209 bool "Wait/wound mutex debugging: Slowpath testing"
1210 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1211 select DEBUG_LOCK_ALLOC
1212 select DEBUG_SPINLOCK
1213 select DEBUG_MUTEXES
1215 This feature enables slowpath testing for w/w mutex users by
1216 injecting additional -EDEADLK wound/backoff cases. Together with
1217 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this
1218 will test all possible w/w mutex interface abuse with the
1219 exception of simply not acquiring all the required locks.
1220 Note that this feature can introduce significant overhead, so
1221 it really should not be enabled in a production or distro kernel,
1222 even a debug kernel. If you are a driver writer, enable it. If
1223 you are a distro, do not.
1226 bool "RW Semaphore debugging: basic checks"
1227 depends on DEBUG_KERNEL
1229 This debugging feature allows mismatched rw semaphore locks
1230 and unlocks to be detected and reported.
1232 config DEBUG_LOCK_ALLOC
1233 bool "Lock debugging: detect incorrect freeing of live locks"
1234 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1235 select DEBUG_SPINLOCK
1236 select DEBUG_MUTEXES
1237 select DEBUG_RT_MUTEXES if RT_MUTEXES
1240 This feature will check whether any held lock (spinlock, rwlock,
1241 mutex or rwsem) is incorrectly freed by the kernel, via any of the
1242 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
1243 vfree(), etc.), whether a live lock is incorrectly reinitialized via
1244 spin_lock_init()/mutex_init()/etc., or whether there is any lock
1245 held during task exit.
1249 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT
1251 select FRAME_POINTER if !MIPS && !PPC && !ARM && !S390 && !MICROBLAZE && !ARC && !X86
1255 config LOCKDEP_SMALL
1258 config DEBUG_LOCKDEP
1259 bool "Lock dependency engine debugging"
1260 depends on DEBUG_KERNEL && LOCKDEP
1262 If you say Y here, the lock dependency engine will do
1263 additional runtime checks to debug itself, at the price
1264 of more runtime overhead.
1266 config DEBUG_ATOMIC_SLEEP
1267 bool "Sleep inside atomic section checking"
1268 select PREEMPT_COUNT
1269 depends on DEBUG_KERNEL
1270 depends on !ARCH_NO_PREEMPT
1272 If you say Y here, various routines which may sleep will become very
1273 noisy if they are called inside atomic sections: when a spinlock is
1274 held, inside an rcu read side critical section, inside preempt disabled
1275 sections, inside an interrupt, etc...
1277 config DEBUG_LOCKING_API_SELFTESTS
1278 bool "Locking API boot-time self-tests"
1279 depends on DEBUG_KERNEL
1281 Say Y here if you want the kernel to run a short self-test during
1282 bootup. The self-test checks whether common types of locking bugs
1283 are detected by debugging mechanisms or not. (if you disable
1284 lock debugging then those bugs wont be detected of course.)
1285 The following locking APIs are covered: spinlocks, rwlocks,
1288 config LOCK_TORTURE_TEST
1289 tristate "torture tests for locking"
1290 depends on DEBUG_KERNEL
1293 This option provides a kernel module that runs torture tests
1294 on kernel locking primitives. The kernel module may be built
1295 after the fact on the running kernel to be tested, if desired.
1297 Say Y here if you want kernel locking-primitive torture tests
1298 to be built into the kernel.
1299 Say M if you want these torture tests to build as a module.
1300 Say N if you are unsure.
1302 config WW_MUTEX_SELFTEST
1303 tristate "Wait/wound mutex selftests"
1305 This option provides a kernel module that runs tests on the
1306 on the struct ww_mutex locking API.
1308 It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction
1309 with this test harness.
1311 Say M if you want these self tests to build as a module.
1312 Say N if you are unsure.
1314 endmenu # lock debugging
1316 config TRACE_IRQFLAGS
1319 Enables hooks to interrupt enabling and disabling for
1320 either tracing or lock debugging.
1323 bool "Stack backtrace support"
1324 depends on STACKTRACE_SUPPORT
1326 This option causes the kernel to create a /proc/pid/stack for
1327 every process, showing its current stack trace.
1328 It is also used by various kernel debugging features that require
1329 stack trace generation.
1331 config WARN_ALL_UNSEEDED_RANDOM
1332 bool "Warn for all uses of unseeded randomness"
1335 Some parts of the kernel contain bugs relating to their use of
1336 cryptographically secure random numbers before it's actually possible
1337 to generate those numbers securely. This setting ensures that these
1338 flaws don't go unnoticed, by enabling a message, should this ever
1339 occur. This will allow people with obscure setups to know when things
1340 are going wrong, so that they might contact developers about fixing
1343 Unfortunately, on some models of some architectures getting
1344 a fully seeded CRNG is extremely difficult, and so this can
1345 result in dmesg getting spammed for a surprisingly long
1346 time. This is really bad from a security perspective, and
1347 so architecture maintainers really need to do what they can
1348 to get the CRNG seeded sooner after the system is booted.
1349 However, since users cannot do anything actionable to
1350 address this, by default the kernel will issue only a single
1351 warning for the first use of unseeded randomness.
1353 Say Y here if you want to receive warnings for all uses of
1354 unseeded randomness. This will be of use primarily for
1355 those developers interested in improving the security of
1356 Linux kernels running on their architecture (or
1359 config DEBUG_KOBJECT
1360 bool "kobject debugging"
1361 depends on DEBUG_KERNEL
1363 If you say Y here, some extra kobject debugging messages will be sent
1366 config DEBUG_KOBJECT_RELEASE
1367 bool "kobject release debugging"
1368 depends on DEBUG_OBJECTS_TIMERS
1370 kobjects are reference counted objects. This means that their
1371 last reference count put is not predictable, and the kobject can
1372 live on past the point at which a driver decides to drop it's
1373 initial reference to the kobject gained on allocation. An
1374 example of this would be a struct device which has just been
1377 However, some buggy drivers assume that after such an operation,
1378 the memory backing the kobject can be immediately freed. This
1379 goes completely against the principles of a refcounted object.
1381 If you say Y here, the kernel will delay the release of kobjects
1382 on the last reference count to improve the visibility of this
1383 kind of kobject release bug.
1385 config HAVE_DEBUG_BUGVERBOSE
1388 menu "Debug kernel data structures"
1391 bool "Debug linked list manipulation"
1392 depends on DEBUG_KERNEL || BUG_ON_DATA_CORRUPTION
1394 Enable this to turn on extended checks in the linked-list
1400 bool "Debug priority linked list manipulation"
1401 depends on DEBUG_KERNEL
1403 Enable this to turn on extended checks in the priority-ordered
1404 linked-list (plist) walking routines. This checks the entire
1405 list multiple times during each manipulation.
1410 bool "Debug SG table operations"
1411 depends on DEBUG_KERNEL
1413 Enable this to turn on checks on scatter-gather tables. This can
1414 help find problems with drivers that do not properly initialize
1419 config DEBUG_NOTIFIERS
1420 bool "Debug notifier call chains"
1421 depends on DEBUG_KERNEL
1423 Enable this to turn on sanity checking for notifier call chains.
1424 This is most useful for kernel developers to make sure that
1425 modules properly unregister themselves from notifier chains.
1426 This is a relatively cheap check but if you care about maximum
1429 config BUG_ON_DATA_CORRUPTION
1430 bool "Trigger a BUG when data corruption is detected"
1433 Select this option if the kernel should BUG when it encounters
1434 data corruption in kernel memory structures when they get checked
1441 config DEBUG_CREDENTIALS
1442 bool "Debug credential management"
1443 depends on DEBUG_KERNEL
1445 Enable this to turn on some debug checking for credential
1446 management. The additional code keeps track of the number of
1447 pointers from task_structs to any given cred struct, and checks to
1448 see that this number never exceeds the usage count of the cred
1451 Furthermore, if SELinux is enabled, this also checks that the
1452 security pointer in the cred struct is never seen to be invalid.
1456 source "kernel/rcu/Kconfig.debug"
1458 config DEBUG_WQ_FORCE_RR_CPU
1459 bool "Force round-robin CPU selection for unbound work items"
1460 depends on DEBUG_KERNEL
1463 Workqueue used to implicitly guarantee that work items queued
1464 without explicit CPU specified are put on the local CPU. This
1465 guarantee is no longer true and while local CPU is still
1466 preferred work items may be put on foreign CPUs. Kernel
1467 parameter "workqueue.debug_force_rr_cpu" is added to force
1468 round-robin CPU selection to flush out usages which depend on the
1469 now broken guarantee. This config option enables the debug
1470 feature by default. When enabled, memory and cache locality will
1473 config DEBUG_BLOCK_EXT_DEVT
1474 bool "Force extended block device numbers and spread them"
1475 depends on DEBUG_KERNEL
1479 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON
1480 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT
1481 YOU ARE DOING. Distros, please enable this and fix whatever
1484 Conventionally, block device numbers are allocated from
1485 predetermined contiguous area. However, extended block area
1486 may introduce non-contiguous block device numbers. This
1487 option forces most block device numbers to be allocated from
1488 the extended space and spreads them to discover kernel or
1489 userland code paths which assume predetermined contiguous
1490 device number allocation.
1492 Note that turning on this debug option shuffles all the
1493 device numbers for all IDE and SCSI devices including libata
1494 ones, so root partition specified using device number
1495 directly (via rdev or root=MAJ:MIN) won't work anymore.
1496 Textual device names (root=/dev/sdXn) will continue to work.
1498 Say N if you are unsure.
1500 config CPU_HOTPLUG_STATE_CONTROL
1501 bool "Enable CPU hotplug state control"
1502 depends on DEBUG_KERNEL
1503 depends on HOTPLUG_CPU
1506 Allows to write steps between "offline" and "online" to the CPUs
1507 sysfs target file so states can be stepped granular. This is a debug
1508 option for now as the hotplug machinery cannot be stopped and
1509 restarted at arbitrary points yet.
1511 Say N if your are unsure.
1514 bool "Latency measuring infrastructure"
1515 depends on DEBUG_KERNEL
1516 depends on STACKTRACE_SUPPORT
1518 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM && !ARC && !X86
1525 Enable this option if you want to use the LatencyTOP tool
1526 to find out which userspace is blocking on what kernel operations.
1528 source "kernel/trace/Kconfig"
1530 config PROVIDE_OHCI1394_DMA_INIT
1531 bool "Remote debugging over FireWire early on boot"
1532 depends on PCI && X86
1534 If you want to debug problems which hang or crash the kernel early
1535 on boot and the crashing machine has a FireWire port, you can use
1536 this feature to remotely access the memory of the crashed machine
1537 over FireWire. This employs remote DMA as part of the OHCI1394
1538 specification which is now the standard for FireWire controllers.
1540 With remote DMA, you can monitor the printk buffer remotely using
1541 firescope and access all memory below 4GB using fireproxy from gdb.
1542 Even controlling a kernel debugger is possible using remote DMA.
1546 If ohci1394_dma=early is used as boot parameter, it will initialize
1547 all OHCI1394 controllers which are found in the PCI config space.
1549 As all changes to the FireWire bus such as enabling and disabling
1550 devices cause a bus reset and thereby disable remote DMA for all
1551 devices, be sure to have the cable plugged and FireWire enabled on
1552 the debugging host before booting the debug target for debugging.
1554 This code (~1k) is freed after boot. By then, the firewire stack
1555 in charge of the OHCI-1394 controllers should be used instead.
1557 See Documentation/core-api/debugging-via-ohci1394.rst for more information.
1559 source "samples/Kconfig"
1561 config ARCH_HAS_DEVMEM_IS_ALLOWED
1564 config STRICT_DEVMEM
1565 bool "Filter access to /dev/mem"
1566 depends on MMU && DEVMEM
1567 depends on ARCH_HAS_DEVMEM_IS_ALLOWED
1568 default y if PPC || X86 || ARM64
1570 If this option is disabled, you allow userspace (root) access to all
1571 of memory, including kernel and userspace memory. Accidental
1572 access to this is obviously disastrous, but specific access can
1573 be used by people debugging the kernel. Note that with PAT support
1574 enabled, even in this case there are restrictions on /dev/mem
1575 use due to the cache aliasing requirements.
1577 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem
1578 file only allows userspace access to PCI space and the BIOS code and
1579 data regions. This is sufficient for dosemu and X and all common
1584 config IO_STRICT_DEVMEM
1585 bool "Filter I/O access to /dev/mem"
1586 depends on STRICT_DEVMEM
1588 If this option is disabled, you allow userspace (root) access to all
1589 io-memory regardless of whether a driver is actively using that
1590 range. Accidental access to this is obviously disastrous, but
1591 specific access can be used by people debugging kernel drivers.
1593 If this option is switched on, the /dev/mem file only allows
1594 userspace access to *idle* io-memory ranges (see /proc/iomem) This
1595 may break traditional users of /dev/mem (dosemu, legacy X, etc...)
1596 if the driver using a given range cannot be disabled.
1600 menu "$(SRCARCH) Debugging"
1602 source "arch/$(SRCARCH)/Kconfig.debug"
1606 menu "Kernel Testing and Coverage"
1608 source "lib/kunit/Kconfig"
1610 config NOTIFIER_ERROR_INJECTION
1611 tristate "Notifier error injection"
1612 depends on DEBUG_KERNEL
1615 This option provides the ability to inject artificial errors to
1616 specified notifier chain callbacks. It is useful to test the error
1617 handling of notifier call chain failures.
1621 config PM_NOTIFIER_ERROR_INJECT
1622 tristate "PM notifier error injection module"
1623 depends on PM && NOTIFIER_ERROR_INJECTION
1624 default m if PM_DEBUG
1626 This option provides the ability to inject artificial errors to
1627 PM notifier chain callbacks. It is controlled through debugfs
1628 interface /sys/kernel/debug/notifier-error-inject/pm
1630 If the notifier call chain should be failed with some events
1631 notified, write the error code to "actions/<notifier event>/error".
1633 Example: Inject PM suspend error (-12 = -ENOMEM)
1635 # cd /sys/kernel/debug/notifier-error-inject/pm/
1636 # echo -12 > actions/PM_SUSPEND_PREPARE/error
1637 # echo mem > /sys/power/state
1638 bash: echo: write error: Cannot allocate memory
1640 To compile this code as a module, choose M here: the module will
1641 be called pm-notifier-error-inject.
1645 config OF_RECONFIG_NOTIFIER_ERROR_INJECT
1646 tristate "OF reconfig notifier error injection module"
1647 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION
1649 This option provides the ability to inject artificial errors to
1650 OF reconfig notifier chain callbacks. It is controlled
1651 through debugfs interface under
1652 /sys/kernel/debug/notifier-error-inject/OF-reconfig/
1654 If the notifier call chain should be failed with some events
1655 notified, write the error code to "actions/<notifier event>/error".
1657 To compile this code as a module, choose M here: the module will
1658 be called of-reconfig-notifier-error-inject.
1662 config NETDEV_NOTIFIER_ERROR_INJECT
1663 tristate "Netdev notifier error injection module"
1664 depends on NET && NOTIFIER_ERROR_INJECTION
1666 This option provides the ability to inject artificial errors to
1667 netdevice notifier chain callbacks. It is controlled through debugfs
1668 interface /sys/kernel/debug/notifier-error-inject/netdev
1670 If the notifier call chain should be failed with some events
1671 notified, write the error code to "actions/<notifier event>/error".
1673 Example: Inject netdevice mtu change error (-22 = -EINVAL)
1675 # cd /sys/kernel/debug/notifier-error-inject/netdev
1676 # echo -22 > actions/NETDEV_CHANGEMTU/error
1677 # ip link set eth0 mtu 1024
1678 RTNETLINK answers: Invalid argument
1680 To compile this code as a module, choose M here: the module will
1681 be called netdev-notifier-error-inject.
1685 config FUNCTION_ERROR_INJECTION
1687 depends on HAVE_FUNCTION_ERROR_INJECTION && KPROBES
1689 config FAULT_INJECTION
1690 bool "Fault-injection framework"
1691 depends on DEBUG_KERNEL
1693 Provide fault-injection framework.
1694 For more details, see Documentation/fault-injection/.
1697 bool "Fault-injection capability for kmalloc"
1698 depends on FAULT_INJECTION
1699 depends on SLAB || SLUB
1701 Provide fault-injection capability for kmalloc.
1703 config FAIL_PAGE_ALLOC
1704 bool "Fault-injection capability for alloc_pages()"
1705 depends on FAULT_INJECTION
1707 Provide fault-injection capability for alloc_pages().
1709 config FAIL_MAKE_REQUEST
1710 bool "Fault-injection capability for disk IO"
1711 depends on FAULT_INJECTION && BLOCK
1713 Provide fault-injection capability for disk IO.
1715 config FAIL_IO_TIMEOUT
1716 bool "Fault-injection capability for faking disk interrupts"
1717 depends on FAULT_INJECTION && BLOCK
1719 Provide fault-injection capability on end IO handling. This
1720 will make the block layer "forget" an interrupt as configured,
1721 thus exercising the error handling.
1723 Only works with drivers that use the generic timeout handling,
1724 for others it wont do anything.
1727 bool "Fault-injection capability for futexes"
1729 depends on FAULT_INJECTION && FUTEX
1731 Provide fault-injection capability for futexes.
1733 config FAULT_INJECTION_DEBUG_FS
1734 bool "Debugfs entries for fault-injection capabilities"
1735 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
1737 Enable configuration of fault-injection capabilities via debugfs.
1739 config FAIL_FUNCTION
1740 bool "Fault-injection capability for functions"
1741 depends on FAULT_INJECTION_DEBUG_FS && FUNCTION_ERROR_INJECTION
1743 Provide function-based fault-injection capability.
1744 This will allow you to override a specific function with a return
1745 with given return value. As a result, function caller will see
1746 an error value and have to handle it. This is useful to test the
1747 error handling in various subsystems.
1749 config FAIL_MMC_REQUEST
1750 bool "Fault-injection capability for MMC IO"
1751 depends on FAULT_INJECTION_DEBUG_FS && MMC
1753 Provide fault-injection capability for MMC IO.
1754 This will make the mmc core return data errors. This is
1755 useful to test the error handling in the mmc block device
1756 and to test how the mmc host driver handles retries from
1759 config FAULT_INJECTION_STACKTRACE_FILTER
1760 bool "stacktrace filter for fault-injection capabilities"
1761 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
1764 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM && !ARC && !X86
1766 Provide stacktrace filter for fault-injection capabilities
1768 config ARCH_HAS_KCOV
1771 An architecture should select this when it can successfully
1772 build and run with CONFIG_KCOV. This typically requires
1773 disabling instrumentation for some early boot code.
1775 config CC_HAS_SANCOV_TRACE_PC
1776 def_bool $(cc-option,-fsanitize-coverage=trace-pc)
1780 bool "Code coverage for fuzzing"
1781 depends on ARCH_HAS_KCOV
1782 depends on CC_HAS_SANCOV_TRACE_PC || GCC_PLUGINS
1784 select GCC_PLUGIN_SANCOV if !CC_HAS_SANCOV_TRACE_PC
1786 KCOV exposes kernel code coverage information in a form suitable
1787 for coverage-guided fuzzing (randomized testing).
1789 If RANDOMIZE_BASE is enabled, PC values will not be stable across
1790 different machines and across reboots. If you need stable PC values,
1791 disable RANDOMIZE_BASE.
1793 For more details, see Documentation/dev-tools/kcov.rst.
1795 config KCOV_ENABLE_COMPARISONS
1796 bool "Enable comparison operands collection by KCOV"
1798 depends on $(cc-option,-fsanitize-coverage=trace-cmp)
1800 KCOV also exposes operands of every comparison in the instrumented
1801 code along with operand sizes and PCs of the comparison instructions.
1802 These operands can be used by fuzzing engines to improve the quality
1803 of fuzzing coverage.
1805 config KCOV_INSTRUMENT_ALL
1806 bool "Instrument all code by default"
1810 If you are doing generic system call fuzzing (like e.g. syzkaller),
1811 then you will want to instrument the whole kernel and you should
1812 say y here. If you are doing more targeted fuzzing (like e.g.
1813 filesystem fuzzing with AFL) then you will want to enable coverage
1814 for more specific subsets of files, and should say n here.
1816 config KCOV_IRQ_AREA_SIZE
1817 hex "Size of interrupt coverage collection area in words"
1821 KCOV uses preallocated per-cpu areas to collect coverage from
1822 soft interrupts. This specifies the size of those areas in the
1823 number of unsigned long words.
1825 menuconfig RUNTIME_TESTING_MENU
1826 bool "Runtime Testing"
1829 if RUNTIME_TESTING_MENU
1832 tristate "Linux Kernel Dump Test Tool Module"
1835 This module enables testing of the different dumping mechanisms by
1836 inducing system failures at predefined crash points.
1837 If you don't need it: say N
1838 Choose M here to compile this code as a module. The module will be
1841 Documentation on how to use the module can be found in
1842 Documentation/fault-injection/provoke-crashes.rst
1844 config TEST_LIST_SORT
1845 tristate "Linked list sorting test"
1846 depends on DEBUG_KERNEL || m
1848 Enable this to turn on 'list_sort()' function test. This test is
1849 executed only once during system boot (so affects only boot time),
1850 or at module load time.
1854 config TEST_MIN_HEAP
1855 tristate "Min heap test"
1856 depends on DEBUG_KERNEL || m
1858 Enable this to turn on min heap function tests. This test is
1859 executed only once during system boot (so affects only boot time),
1860 or at module load time.
1865 tristate "Array-based sort test"
1866 depends on DEBUG_KERNEL || m
1868 This option enables the self-test function of 'sort()' at boot,
1869 or at module load time.
1873 config KPROBES_SANITY_TEST
1874 bool "Kprobes sanity tests"
1875 depends on DEBUG_KERNEL
1878 This option provides for testing basic kprobes functionality on
1879 boot. Samples of kprobe and kretprobe are inserted and
1880 verified for functionality.
1882 Say N if you are unsure.
1884 config BACKTRACE_SELF_TEST
1885 tristate "Self test for the backtrace code"
1886 depends on DEBUG_KERNEL
1888 This option provides a kernel module that can be used to test
1889 the kernel stack backtrace code. This option is not useful
1890 for distributions or general kernels, but only for kernel
1891 developers working on architecture code.
1893 Note that if you want to also test saved backtraces, you will
1894 have to enable STACKTRACE as well.
1896 Say N if you are unsure.
1899 tristate "Red-Black tree test"
1900 depends on DEBUG_KERNEL
1902 A benchmark measuring the performance of the rbtree library.
1903 Also includes rbtree invariant checks.
1905 config REED_SOLOMON_TEST
1906 tristate "Reed-Solomon library test"
1907 depends on DEBUG_KERNEL || m
1909 select REED_SOLOMON_ENC16
1910 select REED_SOLOMON_DEC16
1912 This option enables the self-test function of rslib at boot,
1913 or at module load time.
1917 config INTERVAL_TREE_TEST
1918 tristate "Interval tree test"
1919 depends on DEBUG_KERNEL
1920 select INTERVAL_TREE
1922 A benchmark measuring the performance of the interval tree library
1925 tristate "Per cpu operations test"
1926 depends on m && DEBUG_KERNEL
1928 Enable this option to build test module which validates per-cpu
1933 config ATOMIC64_SELFTEST
1934 tristate "Perform an atomic64_t self-test"
1936 Enable this option to test the atomic64_t functions at boot or
1937 at module load time.
1941 config ASYNC_RAID6_TEST
1942 tristate "Self test for hardware accelerated raid6 recovery"
1943 depends on ASYNC_RAID6_RECOV
1946 This is a one-shot self test that permutes through the
1947 recovery of all the possible two disk failure scenarios for a
1948 N-disk array. Recovery is performed with the asynchronous
1949 raid6 recovery routines, and will optionally use an offload
1950 engine if one is available.
1955 tristate "Test functions located in the hexdump module at runtime"
1957 config TEST_STRING_HELPERS
1958 tristate "Test functions located in the string_helpers module at runtime"
1961 tristate "Test strscpy*() family of functions at runtime"
1964 tristate "Test kstrto*() family of functions at runtime"
1967 tristate "Test printf() family of functions at runtime"
1970 tristate "Test bitmap_*() family of functions at runtime"
1972 Enable this option to test the bitmap functions at boot.
1976 config TEST_BITFIELD
1977 tristate "Test bitfield functions at runtime"
1979 Enable this option to test the bitfield functions at boot.
1984 tristate "Test functions located in the uuid module at runtime"
1987 tristate "Test the XArray code at runtime"
1989 config TEST_OVERFLOW
1990 tristate "Test check_*_overflow() functions at runtime"
1992 config TEST_RHASHTABLE
1993 tristate "Perform selftest on resizable hash table"
1995 Enable this option to test the rhashtable functions at boot.
2000 tristate "Perform selftest on hash functions"
2002 Enable this option to test the kernel's integer (<linux/hash.h>),
2003 string (<linux/stringhash.h>), and siphash (<linux/siphash.h>)
2004 hash functions on boot (or module load).
2006 This is intended to help people writing architecture-specific
2007 optimized versions. If unsure, say N.
2010 tristate "Perform selftest on IDA functions"
2013 tristate "Perform selftest on priority array manager"
2016 Enable this option to test priority array manager on boot
2021 config TEST_IRQ_TIMINGS
2022 bool "IRQ timings selftest"
2023 depends on IRQ_TIMINGS
2025 Enable this option to test the irq timings code on boot.
2030 tristate "Test module loading with 'hello world' module"
2033 This builds the "test_module" module that emits "Hello, world"
2034 on printk when loaded. It is designed to be used for basic
2035 evaluation of the module loading subsystem (for example when
2036 validating module verification). It lacks any extra dependencies,
2037 and will not normally be loaded by the system unless explicitly
2043 tristate "Test module for compilation of clear_bit/set_bit operations"
2046 This builds the "test_bitops" module that is much like the
2047 TEST_LKM module except that it does a basic exercise of the
2048 clear_bit and set_bit macros to make sure there are no compiler
2049 warnings from C=1 sparse checker or -Wextra compilations. It has
2050 no dependencies and doesn't run or load unless explicitly requested
2051 by name. for example: modprobe test_bitops.
2056 tristate "Test module for stress/performance analysis of vmalloc allocator"
2061 This builds the "test_vmalloc" module that should be used for
2062 stress and performance analysis. So, any new change for vmalloc
2063 subsystem can be evaluated from performance and stability point
2068 config TEST_USER_COPY
2069 tristate "Test user/kernel boundary protections"
2072 This builds the "test_user_copy" module that runs sanity checks
2073 on the copy_to/from_user infrastructure, making sure basic
2074 user/kernel boundary testing is working. If it fails to load,
2075 a regression has been detected in the user/kernel memory boundary
2081 tristate "Test BPF filter functionality"
2084 This builds the "test_bpf" module that runs various test vectors
2085 against the BPF interpreter or BPF JIT compiler depending on the
2086 current setting. This is in particular useful for BPF JIT compiler
2087 development, but also to run regression tests against changes in
2088 the interpreter code. It also enables test stubs for eBPF maps and
2089 verifier used by user space verifier testsuite.
2093 config TEST_BLACKHOLE_DEV
2094 tristate "Test blackhole netdev functionality"
2097 This builds the "test_blackhole_dev" module that validates the
2098 data path through this blackhole netdev.
2102 config FIND_BIT_BENCHMARK
2103 tristate "Test find_bit functions"
2105 This builds the "test_find_bit" module that measure find_*_bit()
2106 functions performance.
2110 config TEST_FIRMWARE
2111 tristate "Test firmware loading via userspace interface"
2112 depends on FW_LOADER
2114 This builds the "test_firmware" module that creates a userspace
2115 interface for testing firmware loading. This can be used to
2116 control the triggering of firmware loading without needing an
2117 actual firmware-using device. The contents can be rechecked by
2123 tristate "sysctl test driver"
2124 depends on PROC_SYSCTL
2126 This builds the "test_sysctl" module. This driver enables to test the
2127 proc sysctl interfaces available to drivers safely without affecting
2128 production knobs which might alter system functionality.
2132 config SYSCTL_KUNIT_TEST
2133 tristate "KUnit test for sysctl"
2136 This builds the proc sysctl unit test, which runs on boot.
2137 Tests the API contract and implementation correctness of sysctl.
2138 For more information on KUnit and unit tests in general please refer
2139 to the KUnit documentation in Documentation/dev-tools/kunit/.
2143 config LIST_KUNIT_TEST
2144 tristate "KUnit Test for Kernel Linked-list structures"
2147 This builds the linked list KUnit test suite.
2148 It tests that the API and basic functionality of the list_head type
2149 and associated macros.
2151 KUnit tests run during boot and output the results to the debug log
2152 in TAP format (http://testanything.org/). Only useful for kernel devs
2153 running the KUnit test harness, and not intended for inclusion into a
2156 For more information on KUnit and unit tests in general please refer
2157 to the KUnit documentation in Documentation/dev-tools/kunit/.
2161 config LINEAR_RANGES_TEST
2162 tristate "KUnit test for linear_ranges"
2164 select LINEAR_RANGES
2166 This builds the linear_ranges unit test, which runs on boot.
2167 Tests the linear_ranges logic correctness.
2168 For more information on KUnit and unit tests in general please refer
2169 to the KUnit documentation in Documentation/dev-tools/kunit/.
2174 tristate "udelay test driver"
2176 This builds the "udelay_test" module that helps to make sure
2177 that udelay() is working properly.
2181 config TEST_STATIC_KEYS
2182 tristate "Test static keys"
2185 Test the static key interfaces.
2190 tristate "kmod stress tester"
2192 depends on NETDEVICES && NET_CORE && INET # for TUN
2199 Test the kernel's module loading mechanism: kmod. kmod implements
2200 support to load modules using the Linux kernel's usermode helper.
2201 This test provides a series of tests against kmod.
2203 Although technically you can either build test_kmod as a module or
2204 into the kernel we disallow building it into the kernel since
2205 it stress tests request_module() and this will very likely cause
2206 some issues by taking over precious threads available from other
2207 module load requests, ultimately this could be fatal.
2211 tools/testing/selftests/kmod/kmod.sh --help
2215 config TEST_DEBUG_VIRTUAL
2216 tristate "Test CONFIG_DEBUG_VIRTUAL feature"
2217 depends on DEBUG_VIRTUAL
2219 Test the kernel's ability to detect incorrect calls to
2220 virt_to_phys() done against the non-linear part of the
2221 kernel's virtual address map.
2225 config TEST_MEMCAT_P
2226 tristate "Test memcat_p() helper function"
2228 Test the memcat_p() helper for correctly merging two
2229 pointer arrays together.
2233 config TEST_LIVEPATCH
2234 tristate "Test livepatching"
2236 depends on DYNAMIC_DEBUG
2237 depends on LIVEPATCH
2240 Test kernel livepatching features for correctness. The tests will
2241 load test modules that will be livepatched in various scenarios.
2243 To run all the livepatching tests:
2245 make -C tools/testing/selftests TARGETS=livepatch run_tests
2247 Alternatively, individual tests may be invoked:
2249 tools/testing/selftests/livepatch/test-callbacks.sh
2250 tools/testing/selftests/livepatch/test-livepatch.sh
2251 tools/testing/selftests/livepatch/test-shadow-vars.sh
2256 tristate "Perform selftest on object aggreration manager"
2260 Enable this option to test object aggregation manager on boot
2264 config TEST_STACKINIT
2265 tristate "Test level of stack variable initialization"
2267 Test if the kernel is zero-initializing stack variables and
2268 padding. Coverage is controlled by compiler flags,
2269 CONFIG_GCC_PLUGIN_STRUCTLEAK, CONFIG_GCC_PLUGIN_STRUCTLEAK_BYREF,
2270 or CONFIG_GCC_PLUGIN_STRUCTLEAK_BYREF_ALL.
2275 tristate "Test heap/page initialization"
2277 Test if the kernel is zero-initializing heap and page allocations.
2278 This can be useful to test init_on_alloc and init_on_free features.
2283 tristate "Test HMM (Heterogeneous Memory Management)"
2284 depends on TRANSPARENT_HUGEPAGE
2285 depends on DEVICE_PRIVATE
2289 This is a pseudo device driver solely for testing HMM.
2290 Say M here if you want to build the HMM test module.
2291 Doing so will allow you to run tools/testing/selftest/vm/hmm-tests.
2295 endif # RUNTIME_TESTING_MENU
2300 This option adds a kernel parameter 'memtest', which allows memtest
2302 memtest=0, mean disabled; -- default
2303 memtest=1, mean do 1 test pattern;
2305 memtest=17, mean do 17 test patterns.
2306 If you are unsure how to answer this question, answer N.
2310 config HYPERV_TESTING
2311 bool "Microsoft Hyper-V driver testing"
2313 depends on HYPERV && DEBUG_FS
2315 Select this option to enable Hyper-V vmbus testing.
2317 endmenu # "Kernel Testing and Coverage"
2319 endmenu # Kernel hacking