3107 lines · plain
1# SPDX-License-Identifier: GPL-2.0-only2menu "Kernel hacking"3 4menu "printk and dmesg options"5 6config PRINTK_TIME7 bool "Show timing information on printks"8 depends on PRINTK9 help10 Selecting this option causes time stamps of the printk()11 messages to be added to the output of the syslog() system12 call and at the console.13 14 The timestamp is always recorded internally, and exported15 to /dev/kmsg. This flag just specifies if the timestamp should16 be included, not that the timestamp is recorded.17 18 The behavior is also controlled by the kernel command line19 parameter printk.time=1. See Documentation/admin-guide/kernel-parameters.rst20 21config PRINTK_CALLER22 bool "Show caller information on printks"23 depends on PRINTK24 help25 Selecting this option causes printk() to add a caller "thread id" (if26 in task context) or a caller "processor id" (if not in task context)27 to every message.28 29 This option is intended for environments where multiple threads30 concurrently call printk() for many times, for it is difficult to31 interpret without knowing where these lines (or sometimes individual32 line which was divided into multiple lines due to race) came from.33 34 Since toggling after boot makes the code racy, currently there is35 no option to enable/disable at the kernel command line parameter or36 sysfs interface.37 38config STACKTRACE_BUILD_ID39 bool "Show build ID information in stacktraces"40 depends on PRINTK41 help42 Selecting this option adds build ID information for symbols in43 stacktraces printed with the printk format '%p[SR]b'.44 45 This option is intended for distros where debuginfo is not easily46 accessible but can be downloaded given the build ID of the vmlinux or47 kernel module where the function is located.48 49config CONSOLE_LOGLEVEL_DEFAULT50 int "Default console loglevel (1-15)"51 range 1 1552 default "7"53 help54 Default loglevel to determine what will be printed on the console.55 56 Setting a default here is equivalent to passing in loglevel=<x> in57 the kernel bootargs. loglevel=<x> continues to override whatever58 value is specified here as well.59 60 Note: This does not affect the log level of un-prefixed printk()61 usage in the kernel. That is controlled by the MESSAGE_LOGLEVEL_DEFAULT62 option.63 64config CONSOLE_LOGLEVEL_QUIET65 int "quiet console loglevel (1-15)"66 range 1 1567 default "4"68 help69 loglevel to use when "quiet" is passed on the kernel commandline.70 71 When "quiet" is passed on the kernel commandline this loglevel72 will be used as the loglevel. IOW passing "quiet" will be the73 equivalent of passing "loglevel=<CONSOLE_LOGLEVEL_QUIET>"74 75config MESSAGE_LOGLEVEL_DEFAULT76 int "Default message log level (1-7)"77 range 1 778 default "4"79 help80 Default log level for printk statements with no specified priority.81 82 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks83 that are auditing their logs closely may want to set it to a lower84 priority.85 86 Note: This does not affect what message level gets printed on the console87 by default. To change that, use loglevel=<x> in the kernel bootargs,88 or pick a different CONSOLE_LOGLEVEL_DEFAULT configuration value.89 90config BOOT_PRINTK_DELAY91 bool "Delay each boot printk message by N milliseconds"92 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY93 help94 This build option allows you to read kernel boot messages95 by inserting a short delay after each one. The delay is96 specified in milliseconds on the kernel command line,97 using "boot_delay=N".98 99 It is likely that you would also need to use "lpj=M" to preset100 the "loops per jiffy" value.101 See a previous boot log for the "lpj" value to use for your102 system, and then set "lpj=M" before setting "boot_delay=N".103 NOTE: Using this option may adversely affect SMP systems.104 I.e., processors other than the first one may not boot up.105 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect106 what it believes to be lockup conditions.107 108config DYNAMIC_DEBUG109 bool "Enable dynamic printk() support"110 default n111 depends on PRINTK112 depends on (DEBUG_FS || PROC_FS)113 select DYNAMIC_DEBUG_CORE114 help115 116 Compiles debug level messages into the kernel, which would not117 otherwise be available at runtime. These messages can then be118 enabled/disabled based on various levels of scope - per source file,119 function, module, format string, and line number. This mechanism120 implicitly compiles in all pr_debug() and dev_dbg() calls, which121 enlarges the kernel text size by about 2%.122 123 If a source file is compiled with DEBUG flag set, any124 pr_debug() calls in it are enabled by default, but can be125 disabled at runtime as below. Note that DEBUG flag is126 turned on by many CONFIG_*DEBUG* options.127 128 Usage:129 130 Dynamic debugging is controlled via the 'dynamic_debug/control' file,131 which is contained in the 'debugfs' filesystem or procfs.132 Thus, the debugfs or procfs filesystem must first be mounted before133 making use of this feature.134 We refer the control file as: <debugfs>/dynamic_debug/control. This135 file contains a list of the debug statements that can be enabled. The136 format for each line of the file is:137 138 filename:lineno [module]function flags format139 140 filename : source file of the debug statement141 lineno : line number of the debug statement142 module : module that contains the debug statement143 function : function that contains the debug statement144 flags : '=p' means the line is turned 'on' for printing145 format : the format used for the debug statement146 147 From a live system:148 149 nullarbor:~ # cat <debugfs>/dynamic_debug/control150 # filename:lineno [module]function flags format151 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012"152 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012"153 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012"154 155 Example usage:156 157 // enable the message at line 1603 of file svcsock.c158 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' >159 <debugfs>/dynamic_debug/control160 161 // enable all the messages in file svcsock.c162 nullarbor:~ # echo -n 'file svcsock.c +p' >163 <debugfs>/dynamic_debug/control164 165 // enable all the messages in the NFS server module166 nullarbor:~ # echo -n 'module nfsd +p' >167 <debugfs>/dynamic_debug/control168 169 // enable all 12 messages in the function svc_process()170 nullarbor:~ # echo -n 'func svc_process +p' >171 <debugfs>/dynamic_debug/control172 173 // disable all 12 messages in the function svc_process()174 nullarbor:~ # echo -n 'func svc_process -p' >175 <debugfs>/dynamic_debug/control176 177 See Documentation/admin-guide/dynamic-debug-howto.rst for additional178 information.179 180config DYNAMIC_DEBUG_CORE181 bool "Enable core function of dynamic debug support"182 depends on PRINTK183 depends on (DEBUG_FS || PROC_FS)184 help185 Enable core functional support of dynamic debug. It is useful186 when you want to tie dynamic debug to your kernel modules with187 DYNAMIC_DEBUG_MODULE defined for each of them, especially for188 the case of embedded system where the kernel image size is189 sensitive for people.190 191config SYMBOLIC_ERRNAME192 bool "Support symbolic error names in printf"193 default y if PRINTK194 help195 If you say Y here, the kernel's printf implementation will196 be able to print symbolic error names such as ENOSPC instead197 of the number 28. It makes the kernel image slightly larger198 (about 3KB), but can make the kernel logs easier to read.199 200config DEBUG_BUGVERBOSE201 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT202 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE)203 default y204 help205 Say Y here to make BUG() panics output the file name and line number206 of the BUG call as well as the EIP and oops trace. This aids207 debugging but costs about 70-100K of memory.208 209endmenu # "printk and dmesg options"210 211config DEBUG_KERNEL212 bool "Kernel debugging"213 help214 Say Y here if you are developing drivers or trying to debug and215 identify kernel problems.216 217config DEBUG_MISC218 bool "Miscellaneous debug code"219 default DEBUG_KERNEL220 depends on DEBUG_KERNEL221 help222 Say Y here if you need to enable miscellaneous debug code that should223 be under a more specific debug option but isn't.224 225menu "Compile-time checks and compiler options"226 227config DEBUG_INFO228 bool229 help230 A kernel debug info option other than "None" has been selected231 in the "Debug information" choice below, indicating that debug232 information will be generated for build targets.233 234# Clang generates .uleb128 with label differences for DWARF v5, a feature that235# older binutils ports do not support when utilizing RISC-V style linker236# relaxation: https://sourceware.org/bugzilla/show_bug.cgi?id=27215237config AS_HAS_NON_CONST_ULEB128238 def_bool $(as-instr,.uleb128 .Lexpr_end4 - .Lexpr_start3\n.Lexpr_start3:\n.Lexpr_end4:)239 240choice241 prompt "Debug information"242 depends on DEBUG_KERNEL243 help244 Selecting something other than "None" results in a kernel image245 that will include debugging info resulting in a larger kernel image.246 This adds debug symbols to the kernel and modules (gcc -g), and247 is needed if you intend to use kernel crashdump or binary object248 tools like crash, kgdb, LKCD, gdb, etc on the kernel.249 250 Choose which version of DWARF debug info to emit. If unsure,251 select "Toolchain default".252 253config DEBUG_INFO_NONE254 bool "Disable debug information"255 help256 Do not build the kernel with debugging information, which will257 result in a faster and smaller build.258 259config DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT260 bool "Rely on the toolchain's implicit default DWARF version"261 select DEBUG_INFO262 depends on !CC_IS_CLANG || AS_IS_LLVM || CLANG_VERSION < 140000 || (AS_IS_GNU && AS_VERSION >= 23502 && AS_HAS_NON_CONST_ULEB128)263 help264 The implicit default version of DWARF debug info produced by a265 toolchain changes over time.266 267 This can break consumers of the debug info that haven't upgraded to268 support newer revisions, and prevent testing newer versions, but269 those should be less common scenarios.270 271config DEBUG_INFO_DWARF4272 bool "Generate DWARF Version 4 debuginfo"273 select DEBUG_INFO274 depends on !CC_IS_CLANG || AS_IS_LLVM || (AS_IS_GNU && AS_VERSION >= 23502)275 help276 Generate DWARF v4 debug info. This requires gcc 4.5+, binutils 2.35.2277 if using clang without clang's integrated assembler, and gdb 7.0+.278 279 If you have consumers of DWARF debug info that are not ready for280 newer revisions of DWARF, you may wish to choose this or have your281 config select this.282 283config DEBUG_INFO_DWARF5284 bool "Generate DWARF Version 5 debuginfo"285 select DEBUG_INFO286 depends on !ARCH_HAS_BROKEN_DWARF5287 depends on !CC_IS_CLANG || AS_IS_LLVM || (AS_IS_GNU && AS_VERSION >= 23502 && AS_HAS_NON_CONST_ULEB128)288 help289 Generate DWARF v5 debug info. Requires binutils 2.35.2, gcc 5.0+ (gcc290 5.0+ accepts the -gdwarf-5 flag but only had partial support for some291 draft features until 7.0), and gdb 8.0+.292 293 Changes to the structure of debug info in Version 5 allow for around294 15-18% savings in resulting image and debug info section sizes as295 compared to DWARF Version 4. DWARF Version 5 standardizes previous296 extensions such as accelerators for symbol indexing and the format297 for fission (.dwo/.dwp) files. Users may not want to select this298 config if they rely on tooling that has not yet been updated to299 support DWARF Version 5.300 301endchoice # "Debug information"302 303if DEBUG_INFO304 305config DEBUG_INFO_REDUCED306 bool "Reduce debugging information"307 help308 If you say Y here gcc is instructed to generate less debugging309 information for structure types. This means that tools that310 need full debugging information (like kgdb or systemtap) won't311 be happy. But if you merely need debugging information to312 resolve line numbers there is no loss. Advantage is that313 build directory object sizes shrink dramatically over a full314 DEBUG_INFO build and compile times are reduced too.315 Only works with newer gcc versions.316 317choice318 prompt "Compressed Debug information"319 help320 Compress the resulting debug info. Results in smaller debug info sections,321 but requires that consumers are able to decompress the results.322 323 If unsure, choose DEBUG_INFO_COMPRESSED_NONE.324 325config DEBUG_INFO_COMPRESSED_NONE326 bool "Don't compress debug information"327 help328 Don't compress debug info sections.329 330config DEBUG_INFO_COMPRESSED_ZLIB331 bool "Compress debugging information with zlib"332 depends on $(cc-option,-gz=zlib)333 depends on $(ld-option,--compress-debug-sections=zlib)334 help335 Compress the debug information using zlib. Requires GCC 5.0+ or Clang336 5.0+, binutils 2.26+, and zlib.337 338 Users of dpkg-deb via scripts/package/builddeb may find an increase in339 size of their debug .deb packages with this config set, due to the340 debug info being compressed with zlib, then the object files being341 recompressed with a different compression scheme. But this is still342 preferable to setting $KDEB_COMPRESS to "none" which would be even343 larger.344 345config DEBUG_INFO_COMPRESSED_ZSTD346 bool "Compress debugging information with zstd"347 depends on $(cc-option,-gz=zstd)348 depends on $(ld-option,--compress-debug-sections=zstd)349 help350 Compress the debug information using zstd. This may provide better351 compression than zlib, for about the same time costs, but requires newer352 toolchain support. Requires GCC 13.0+ or Clang 16.0+, binutils 2.40+, and353 zstd.354 355endchoice # "Compressed Debug information"356 357config DEBUG_INFO_SPLIT358 bool "Produce split debuginfo in .dwo files"359 depends on $(cc-option,-gsplit-dwarf)360 # RISC-V linker relaxation + -gsplit-dwarf has issues with LLVM and GCC361 # prior to 12.x:362 # https://github.com/llvm/llvm-project/issues/56642363 # https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99090364 depends on !RISCV || GCC_VERSION >= 120000365 help366 Generate debug info into separate .dwo files. This significantly367 reduces the build directory size for builds with DEBUG_INFO,368 because it stores the information only once on disk in .dwo369 files instead of multiple times in object files and executables.370 In addition the debug information is also compressed.371 372 Requires recent gcc (4.7+) and recent gdb/binutils.373 Any tool that packages or reads debug information would need374 to know about the .dwo files and include them.375 Incompatible with older versions of ccache.376 377config DEBUG_INFO_BTF378 bool "Generate BTF type information"379 depends on !DEBUG_INFO_SPLIT && !DEBUG_INFO_REDUCED380 depends on !GCC_PLUGIN_RANDSTRUCT || COMPILE_TEST381 depends on BPF_SYSCALL382 depends on PAHOLE_VERSION >= 116383 depends on DEBUG_INFO_DWARF4 || PAHOLE_VERSION >= 121384 # pahole uses elfutils, which does not have support for Hexagon relocations385 depends on !HEXAGON386 help387 Generate deduplicated BTF type information from DWARF debug info.388 Turning this on requires pahole v1.16 or later (v1.21 or later to389 support DWARF 5), which will convert DWARF type info into equivalent390 deduplicated BTF type info.391 392config PAHOLE_HAS_SPLIT_BTF393 def_bool PAHOLE_VERSION >= 119394 395config PAHOLE_HAS_BTF_TAG396 def_bool PAHOLE_VERSION >= 123397 depends on CC_IS_CLANG398 help399 Decide whether pahole emits btf_tag attributes (btf_type_tag and400 btf_decl_tag) or not. Currently only clang compiler implements401 these attributes, so make the config depend on CC_IS_CLANG.402 403config PAHOLE_HAS_LANG_EXCLUDE404 def_bool PAHOLE_VERSION >= 124405 help406 Support for the --lang_exclude flag which makes pahole exclude407 compilation units from the supplied language. Used in Kbuild to408 omit Rust CUs which are not supported in version 1.24 of pahole,409 otherwise it would emit malformed kernel and module binaries when410 using DEBUG_INFO_BTF_MODULES.411 412config DEBUG_INFO_BTF_MODULES413 bool "Generate BTF type information for kernel modules"414 default y415 depends on DEBUG_INFO_BTF && MODULES && PAHOLE_HAS_SPLIT_BTF416 help417 Generate compact split BTF type information for kernel modules.418 419config MODULE_ALLOW_BTF_MISMATCH420 bool "Allow loading modules with non-matching BTF type info"421 depends on DEBUG_INFO_BTF_MODULES422 help423 For modules whose split BTF does not match vmlinux, load without424 BTF rather than refusing to load. The default behavior with425 module BTF enabled is to reject modules with such mismatches;426 this option will still load module BTF where possible but ignore427 it when a mismatch is found.428 429config GDB_SCRIPTS430 bool "Provide GDB scripts for kernel debugging"431 help432 This creates the required links to GDB helper scripts in the433 build directory. If you load vmlinux into gdb, the helper434 scripts will be automatically imported by gdb as well, and435 additional functions are available to analyze a Linux kernel436 instance. See Documentation/dev-tools/gdb-kernel-debugging.rst437 for further details.438 439endif # DEBUG_INFO440 441config FRAME_WARN442 int "Warn for stack frames larger than"443 range 0 8192444 default 0 if KMSAN445 default 2048 if GCC_PLUGIN_LATENT_ENTROPY446 default 2048 if PARISC447 default 1536 if (!64BIT && XTENSA)448 default 1280 if KASAN && !64BIT449 default 1024 if !64BIT450 default 2048 if 64BIT451 help452 Tell the compiler to warn at build time for stack frames larger than this.453 Setting this too low will cause a lot of warnings.454 Setting it to 0 disables the warning.455 456config STRIP_ASM_SYMS457 bool "Strip assembler-generated symbols during link"458 default n459 help460 Strip internal assembler-generated symbols during a link (symbols461 that look like '.Lxxx') so they don't pollute the output of462 get_wchan() and suchlike.463 464config READABLE_ASM465 bool "Generate readable assembler code"466 depends on DEBUG_KERNEL467 depends on CC_IS_GCC468 help469 Disable some compiler optimizations that tend to generate human unreadable470 assembler output. This may make the kernel slightly slower, but it helps471 to keep kernel developers who have to stare a lot at assembler listings472 sane.473 474config HEADERS_INSTALL475 bool "Install uapi headers to usr/include"476 depends on !UML477 help478 This option will install uapi headers (headers exported to user-space)479 into the usr/include directory for use during the kernel build.480 This is unneeded for building the kernel itself, but needed for some481 user-space program samples. It is also needed by some features such482 as uapi header sanity checks.483 484config DEBUG_SECTION_MISMATCH485 bool "Enable full Section mismatch analysis"486 depends on CC_IS_GCC487 help488 The section mismatch analysis checks if there are illegal489 references from one section to another section.490 During linktime or runtime, some sections are dropped;491 any use of code/data previously in these sections would492 most likely result in an oops.493 In the code, functions and variables are annotated with494 __init,, etc. (see the full list in include/linux/init.h),495 which results in the code/data being placed in specific sections.496 The section mismatch analysis is always performed after a full497 kernel build, and enabling this option causes the following498 additional step to occur:499 - Add the option -fno-inline-functions-called-once to gcc commands.500 When inlining a function annotated with __init in a non-init501 function, we would lose the section information and thus502 the analysis would not catch the illegal reference.503 This option tells gcc to inline less (but it does result in504 a larger kernel).505 506config SECTION_MISMATCH_WARN_ONLY507 bool "Make section mismatch errors non-fatal"508 default y509 help510 If you say N here, the build process will fail if there are any511 section mismatch, instead of just throwing warnings.512 513 If unsure, say Y.514 515config DEBUG_FORCE_FUNCTION_ALIGN_64B516 bool "Force all function address 64B aligned"517 depends on EXPERT && (X86_64 || ARM64 || PPC32 || PPC64 || ARC || RISCV || S390)518 select FUNCTION_ALIGNMENT_64B519 help520 There are cases that a commit from one domain changes the function521 address alignment of other domains, and cause magic performance522 bump (regression or improvement). Enable this option will help to523 verify if the bump is caused by function alignment changes, while524 it will slightly increase the kernel size and affect icache usage.525 526 It is mainly for debug and performance tuning use.527 528#529# Select this config option from the architecture Kconfig, if it530# is preferred to always offer frame pointers as a config531# option on the architecture (regardless of KERNEL_DEBUG):532#533config ARCH_WANT_FRAME_POINTERS534 bool535 536config FRAME_POINTER537 bool "Compile the kernel with frame pointers"538 depends on DEBUG_KERNEL && (M68K || UML || SUPERH) || ARCH_WANT_FRAME_POINTERS539 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS540 help541 If you say Y here the resulting kernel image will be slightly542 larger and slower, but it gives very useful debugging information543 in case of kernel bugs. (precise oopses/stacktraces/warnings)544 545config OBJTOOL546 bool547 548config STACK_VALIDATION549 bool "Compile-time stack metadata validation"550 depends on HAVE_STACK_VALIDATION && UNWINDER_FRAME_POINTER551 select OBJTOOL552 default n553 help554 Validate frame pointer rules at compile-time. This helps ensure that555 runtime stack traces are more reliable.556 557 For more information, see558 tools/objtool/Documentation/objtool.txt.559 560config NOINSTR_VALIDATION561 bool562 depends on HAVE_NOINSTR_VALIDATION && DEBUG_ENTRY563 select OBJTOOL564 default y565 566config VMLINUX_MAP567 bool "Generate vmlinux.map file when linking"568 depends on EXPERT569 help570 Selecting this option will pass "-Map=vmlinux.map" to ld571 when linking vmlinux. That file can be useful for verifying572 and debugging magic section games, and for seeing which573 pieces of code get eliminated with574 CONFIG_LD_DEAD_CODE_DATA_ELIMINATION.575 576config BUILTIN_MODULE_RANGES577 bool "Generate address range information for builtin modules"578 depends on !LTO579 depends on VMLINUX_MAP580 help581 When modules are built into the kernel, there will be no module name582 associated with its symbols in /proc/kallsyms. Tracers may want to583 identify symbols by module name and symbol name regardless of whether584 the module is configured as loadable or not.585 586 This option generates modules.builtin.ranges in the build tree with587 offset ranges (per ELF section) for the module(s) they belong to.588 It also records an anchor symbol to determine the load address of the589 section.590 591config DEBUG_FORCE_WEAK_PER_CPU592 bool "Force weak per-cpu definitions"593 depends on DEBUG_KERNEL594 help595 s390 and alpha require percpu variables in modules to be596 defined weak to work around addressing range issue which597 puts the following two restrictions on percpu variable598 definitions.599 600 1. percpu symbols must be unique whether static or not601 2. percpu variables can't be defined inside a function602 603 To ensure that generic code follows the above rules, this604 option forces all percpu variables to be defined as weak.605 606endmenu # "Compiler options"607 608menu "Generic Kernel Debugging Instruments"609 610config MAGIC_SYSRQ611 bool "Magic SysRq key"612 depends on !UML613 help614 If you say Y here, you will have some control over the system even615 if the system crashes for example during kernel debugging (e.g., you616 will be able to flush the buffer cache to disk, reboot the system617 immediately or dump some status information). This is accomplished618 by pressing various keys while holding SysRq (Alt+PrintScreen). It619 also works on a serial console (on PC hardware at least), if you620 send a BREAK and then within 5 seconds a command keypress. The621 keys are documented in <file:Documentation/admin-guide/sysrq.rst>.622 Don't say Y unless you really know what this hack does.623 624config MAGIC_SYSRQ_DEFAULT_ENABLE625 hex "Enable magic SysRq key functions by default"626 depends on MAGIC_SYSRQ627 default 0x1628 help629 Specifies which SysRq key functions are enabled by default.630 This may be set to 1 or 0 to enable or disable them all, or631 to a bitmask as described in Documentation/admin-guide/sysrq.rst.632 633config MAGIC_SYSRQ_SERIAL634 bool "Enable magic SysRq key over serial"635 depends on MAGIC_SYSRQ636 default y637 help638 Many embedded boards have a disconnected TTL level serial which can639 generate some garbage that can lead to spurious false sysrq detects.640 This option allows you to decide whether you want to enable the641 magic SysRq key.642 643config MAGIC_SYSRQ_SERIAL_SEQUENCE644 string "Char sequence that enables magic SysRq over serial"645 depends on MAGIC_SYSRQ_SERIAL646 default ""647 help648 Specifies a sequence of characters that can follow BREAK to enable649 SysRq on a serial console.650 651 If unsure, leave an empty string and the option will not be enabled.652 653config DEBUG_FS654 bool "Debug Filesystem"655 help656 debugfs is a virtual file system that kernel developers use to put657 debugging files into. Enable this option to be able to read and658 write to these files.659 660 For detailed documentation on the debugfs API, see661 Documentation/filesystems/.662 663 If unsure, say N.664 665choice666 prompt "Debugfs default access"667 depends on DEBUG_FS668 default DEBUG_FS_ALLOW_ALL669 help670 This selects the default access restrictions for debugfs.671 It can be overridden with kernel command line option672 debugfs=[on,no-mount,off]. The restrictions apply for API access673 and filesystem registration.674 675config DEBUG_FS_ALLOW_ALL676 bool "Access normal"677 help678 No restrictions apply. Both API and filesystem registration679 is on. This is the normal default operation.680 681config DEBUG_FS_DISALLOW_MOUNT682 bool "Do not register debugfs as filesystem"683 help684 The API is open but filesystem is not loaded. Clients can still do685 their work and read with debug tools that do not need686 debugfs filesystem.687 688config DEBUG_FS_ALLOW_NONE689 bool "No access"690 help691 Access is off. Clients get -PERM when trying to create nodes in692 debugfs tree and debugfs is not registered as a filesystem.693 Client can then back-off or continue without debugfs access.694 695endchoice696 697source "lib/Kconfig.kgdb"698source "lib/Kconfig.ubsan"699source "lib/Kconfig.kcsan"700 701endmenu702 703menu "Networking Debugging"704 705source "net/Kconfig.debug"706 707endmenu # "Networking Debugging"708 709menu "Memory Debugging"710 711source "mm/Kconfig.debug"712 713config DEBUG_OBJECTS714 bool "Debug object operations"715 depends on DEBUG_KERNEL716 help717 If you say Y here, additional code will be inserted into the718 kernel to track the life time of various objects and validate719 the operations on those objects.720 721config DEBUG_OBJECTS_SELFTEST722 bool "Debug objects selftest"723 depends on DEBUG_OBJECTS724 help725 This enables the selftest of the object debug code.726 727config DEBUG_OBJECTS_FREE728 bool "Debug objects in freed memory"729 depends on DEBUG_OBJECTS730 help731 This enables checks whether a k/v free operation frees an area732 which contains an object which has not been deactivated733 properly. This can make kmalloc/kfree-intensive workloads734 much slower.735 736config DEBUG_OBJECTS_TIMERS737 bool "Debug timer objects"738 depends on DEBUG_OBJECTS739 help740 If you say Y here, additional code will be inserted into the741 timer routines to track the life time of timer objects and742 validate the timer operations.743 744config DEBUG_OBJECTS_WORK745 bool "Debug work objects"746 depends on DEBUG_OBJECTS747 help748 If you say Y here, additional code will be inserted into the749 work queue routines to track the life time of work objects and750 validate the work operations.751 752config DEBUG_OBJECTS_RCU_HEAD753 bool "Debug RCU callbacks objects"754 depends on DEBUG_OBJECTS755 help756 Enable this to turn on debugging of RCU list heads (call_rcu() usage).757 758config DEBUG_OBJECTS_PERCPU_COUNTER759 bool "Debug percpu counter objects"760 depends on DEBUG_OBJECTS761 help762 If you say Y here, additional code will be inserted into the763 percpu counter routines to track the life time of percpu counter764 objects and validate the percpu counter operations.765 766config DEBUG_OBJECTS_ENABLE_DEFAULT767 int "debug_objects bootup default value (0-1)"768 range 0 1769 default "1"770 depends on DEBUG_OBJECTS771 help772 Debug objects boot parameter default value773 774config SHRINKER_DEBUG775 bool "Enable shrinker debugging support"776 depends on DEBUG_FS777 help778 Say Y to enable the shrinker debugfs interface which provides779 visibility into the kernel memory shrinkers subsystem.780 Disable it to avoid an extra memory footprint.781 782config DEBUG_STACK_USAGE783 bool "Stack utilization instrumentation"784 depends on DEBUG_KERNEL785 help786 Enables the display of the minimum amount of free stack which each787 task has ever had available in the sysrq-T and sysrq-P debug output.788 Also emits a message to dmesg when a process exits if that process789 used more stack space than previously exiting processes.790 791 This option will slow down process creation somewhat.792 793config SCHED_STACK_END_CHECK794 bool "Detect stack corruption on calls to schedule()"795 depends on DEBUG_KERNEL796 default n797 help798 This option checks for a stack overrun on calls to schedule().799 If the stack end location is found to be over written always panic as800 the content of the corrupted region can no longer be trusted.801 This is to ensure no erroneous behaviour occurs which could result in802 data corruption or a sporadic crash at a later stage once the region803 is examined. The runtime overhead introduced is minimal.804 805config ARCH_HAS_DEBUG_VM_PGTABLE806 bool807 help808 An architecture should select this when it can successfully809 build and run DEBUG_VM_PGTABLE.810 811config DEBUG_VM_IRQSOFF812 def_bool DEBUG_VM && !PREEMPT_RT813 814config DEBUG_VM815 bool "Debug VM"816 depends on DEBUG_KERNEL817 help818 Enable this to turn on extended checks in the virtual-memory system819 that may impact performance.820 821 If unsure, say N.822 823config DEBUG_VM_SHOOT_LAZIES824 bool "Debug MMU_LAZY_TLB_SHOOTDOWN implementation"825 depends on DEBUG_VM826 depends on MMU_LAZY_TLB_SHOOTDOWN827 help828 Enable additional IPIs that ensure lazy tlb mm references are removed829 before the mm is freed.830 831 If unsure, say N.832 833config DEBUG_VM_MAPLE_TREE834 bool "Debug VM maple trees"835 depends on DEBUG_VM836 select DEBUG_MAPLE_TREE837 help838 Enable VM maple tree debugging information and extra validations.839 840 If unsure, say N.841 842config DEBUG_VM_RB843 bool "Debug VM red-black trees"844 depends on DEBUG_VM845 help846 Enable VM red-black tree debugging information and extra validations.847 848 If unsure, say N.849 850config DEBUG_VM_PGFLAGS851 bool "Debug page-flags operations"852 depends on DEBUG_VM853 help854 Enables extra validation on page flags operations.855 856 If unsure, say N.857 858config DEBUG_VM_PGTABLE859 bool "Debug arch page table for semantics compliance"860 depends on MMU861 depends on ARCH_HAS_DEBUG_VM_PGTABLE862 default y if DEBUG_VM863 help864 This option provides a debug method which can be used to test865 architecture page table helper functions on various platforms in866 verifying if they comply with expected generic MM semantics. This867 will help architecture code in making sure that any changes or868 new additions of these helpers still conform to expected869 semantics of the generic MM. Platforms will have to opt in for870 this through ARCH_HAS_DEBUG_VM_PGTABLE.871 872 If unsure, say N.873 874config ARCH_HAS_DEBUG_VIRTUAL875 bool876 877config DEBUG_VIRTUAL878 bool "Debug VM translations"879 depends on DEBUG_KERNEL && ARCH_HAS_DEBUG_VIRTUAL880 help881 Enable some costly sanity checks in virtual to page code. This can882 catch mistakes with virt_to_page() and friends.883 884 If unsure, say N.885 886config DEBUG_NOMMU_REGIONS887 bool "Debug the global anon/private NOMMU mapping region tree"888 depends on DEBUG_KERNEL && !MMU889 help890 This option causes the global tree of anonymous and private mapping891 regions to be regularly checked for invalid topology.892 893config DEBUG_MEMORY_INIT894 bool "Debug memory initialisation" if EXPERT895 default !EXPERT896 help897 Enable this for additional checks during memory initialisation.898 The sanity checks verify aspects of the VM such as the memory model899 and other information provided by the architecture. Verbose900 information will be printed at KERN_DEBUG loglevel depending901 on the mminit_loglevel= command-line option.902 903 If unsure, say Y904 905config MEMORY_NOTIFIER_ERROR_INJECT906 tristate "Memory hotplug notifier error injection module"907 depends on MEMORY_HOTPLUG && NOTIFIER_ERROR_INJECTION908 help909 This option provides the ability to inject artificial errors to910 memory hotplug notifier chain callbacks. It is controlled through911 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory912 913 If the notifier call chain should be failed with some events914 notified, write the error code to "actions/<notifier event>/error".915 916 Example: Inject memory hotplug offline error (-12 == -ENOMEM)917 918 # cd /sys/kernel/debug/notifier-error-inject/memory919 # echo -12 > actions/MEM_GOING_OFFLINE/error920 # echo offline > /sys/devices/system/memory/memoryXXX/state921 bash: echo: write error: Cannot allocate memory922 923 To compile this code as a module, choose M here: the module will924 be called memory-notifier-error-inject.925 926 If unsure, say N.927 928config DEBUG_PER_CPU_MAPS929 bool "Debug access to per_cpu maps"930 depends on DEBUG_KERNEL931 depends on SMP932 help933 Say Y to verify that the per_cpu map being accessed has934 been set up. This adds a fair amount of code to kernel memory935 and decreases performance.936 937 Say N if unsure.938 939config DEBUG_KMAP_LOCAL940 bool "Debug kmap_local temporary mappings"941 depends on DEBUG_KERNEL && KMAP_LOCAL942 help943 This option enables additional error checking for the kmap_local944 infrastructure. Disable for production use.945 946config ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP947 bool948 949config DEBUG_KMAP_LOCAL_FORCE_MAP950 bool "Enforce kmap_local temporary mappings"951 depends on DEBUG_KERNEL && ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP952 select KMAP_LOCAL953 select DEBUG_KMAP_LOCAL954 help955 This option enforces temporary mappings through the kmap_local956 mechanism for non-highmem pages and on non-highmem systems.957 Disable this for production systems!958 959config DEBUG_HIGHMEM960 bool "Highmem debugging"961 depends on DEBUG_KERNEL && HIGHMEM962 select DEBUG_KMAP_LOCAL_FORCE_MAP if ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP963 select DEBUG_KMAP_LOCAL964 help965 This option enables additional error checking for high memory966 systems. Disable for production systems.967 968config HAVE_DEBUG_STACKOVERFLOW969 bool970 971config DEBUG_STACKOVERFLOW972 bool "Check for stack overflows"973 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW974 help975 Say Y here if you want to check for overflows of kernel, IRQ976 and exception stacks (if your architecture uses them). This977 option will show detailed messages if free stack space drops978 below a certain limit.979 980 These kinds of bugs usually occur when call-chains in the981 kernel get too deep, especially when interrupts are982 involved.983 984 Use this in cases where you see apparently random memory985 corruption, especially if it appears in 'struct thread_info'986 987 If in doubt, say "N".988 989config CODE_TAGGING990 bool991 select KALLSYMS992 993config MEM_ALLOC_PROFILING994 bool "Enable memory allocation profiling"995 default n996 depends on PROC_FS997 depends on !DEBUG_FORCE_WEAK_PER_CPU998 select CODE_TAGGING999 select PAGE_EXTENSION1000 select SLAB_OBJ_EXT1001 help1002 Track allocation source code and record total allocation size1003 initiated at that code location. The mechanism can be used to track1004 memory leaks with a low performance and memory impact.1005 1006config MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT1007 bool "Enable memory allocation profiling by default"1008 default y1009 depends on MEM_ALLOC_PROFILING1010 1011config MEM_ALLOC_PROFILING_DEBUG1012 bool "Memory allocation profiler debugging"1013 default n1014 depends on MEM_ALLOC_PROFILING1015 select MEM_ALLOC_PROFILING_ENABLED_BY_DEFAULT1016 help1017 Adds warnings with helpful error messages for memory allocation1018 profiling.1019 1020source "lib/Kconfig.kasan"1021source "lib/Kconfig.kfence"1022source "lib/Kconfig.kmsan"1023 1024endmenu # "Memory Debugging"1025 1026config DEBUG_SHIRQ1027 bool "Debug shared IRQ handlers"1028 depends on DEBUG_KERNEL1029 help1030 Enable this to generate a spurious interrupt just before a shared1031 interrupt handler is deregistered (generating one when registering1032 is currently disabled). Drivers need to handle this correctly. Some1033 don't and need to be caught.1034 1035menu "Debug Oops, Lockups and Hangs"1036 1037config PANIC_ON_OOPS1038 bool "Panic on Oops"1039 help1040 Say Y here to enable the kernel to panic when it oopses. This1041 has the same effect as setting oops=panic on the kernel command1042 line.1043 1044 This feature is useful to ensure that the kernel does not do1045 anything erroneous after an oops which could result in data1046 corruption or other issues.1047 1048 Say N if unsure.1049 1050config PANIC_ON_OOPS_VALUE1051 int1052 range 0 11053 default 0 if !PANIC_ON_OOPS1054 default 1 if PANIC_ON_OOPS1055 1056config PANIC_TIMEOUT1057 int "panic timeout"1058 default 01059 help1060 Set the timeout value (in seconds) until a reboot occurs when1061 the kernel panics. If n = 0, then we wait forever. A timeout1062 value n > 0 will wait n seconds before rebooting, while a timeout1063 value n < 0 will reboot immediately. This setting can be overridden1064 with the kernel command line option panic=, and from userspace via1065 /proc/sys/kernel/panic.1066 1067config LOCKUP_DETECTOR1068 bool1069 1070config SOFTLOCKUP_DETECTOR1071 bool "Detect Soft Lockups"1072 depends on DEBUG_KERNEL && !S3901073 select LOCKUP_DETECTOR1074 help1075 Say Y here to enable the kernel to act as a watchdog to detect1076 soft lockups.1077 1078 Softlockups are bugs that cause the kernel to loop in kernel1079 mode for more than 20 seconds, without giving other tasks a1080 chance to run. The current stack trace is displayed upon1081 detection and the system will stay locked up.1082 1083config SOFTLOCKUP_DETECTOR_INTR_STORM1084 bool "Detect Interrupt Storm in Soft Lockups"1085 depends on SOFTLOCKUP_DETECTOR && IRQ_TIME_ACCOUNTING1086 select GENERIC_IRQ_STAT_SNAPSHOT1087 default y if NR_CPUS <= 1281088 help1089 Say Y here to enable the kernel to detect interrupt storm1090 during "soft lockups".1091 1092 "soft lockups" can be caused by a variety of reasons. If one is1093 caused by an interrupt storm, then the storming interrupts will not1094 be on the callstack. To detect this case, it is necessary to report1095 the CPU stats and the interrupt counts during the "soft lockups".1096 1097config BOOTPARAM_SOFTLOCKUP_PANIC1098 bool "Panic (Reboot) On Soft Lockups"1099 depends on SOFTLOCKUP_DETECTOR1100 help1101 Say Y here to enable the kernel to panic on "soft lockups",1102 which are bugs that cause the kernel to loop in kernel1103 mode for more than 20 seconds (configurable using the watchdog_thresh1104 sysctl), without giving other tasks a chance to run.1105 1106 The panic can be used in combination with panic_timeout,1107 to cause the system to reboot automatically after a1108 lockup has been detected. This feature is useful for1109 high-availability systems that have uptime guarantees and1110 where a lockup must be resolved ASAP.1111 1112 Say N if unsure.1113 1114config HAVE_HARDLOCKUP_DETECTOR_BUDDY1115 bool1116 depends on SMP1117 default y1118 1119#1120# Global switch whether to build a hardlockup detector at all. It is available1121# only when the architecture supports at least one implementation. There are1122# two exceptions. The hardlockup detector is never enabled on:1123#1124# s390: it reported many false positives there1125#1126# sparc64: has a custom implementation which is not using the common1127# hardlockup command line options and sysctl interface.1128#1129config HARDLOCKUP_DETECTOR1130 bool "Detect Hard Lockups"1131 depends on DEBUG_KERNEL && !S390 && !HARDLOCKUP_DETECTOR_SPARC641132 depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY || HAVE_HARDLOCKUP_DETECTOR_ARCH1133 imply HARDLOCKUP_DETECTOR_PERF1134 imply HARDLOCKUP_DETECTOR_BUDDY1135 imply HARDLOCKUP_DETECTOR_ARCH1136 select LOCKUP_DETECTOR1137 1138 help1139 Say Y here to enable the kernel to act as a watchdog to detect1140 hard lockups.1141 1142 Hardlockups are bugs that cause the CPU to loop in kernel mode1143 for more than 10 seconds, without letting other interrupts have a1144 chance to run. The current stack trace is displayed upon detection1145 and the system will stay locked up.1146 1147#1148# Note that arch-specific variants are always preferred.1149#1150config HARDLOCKUP_DETECTOR_PREFER_BUDDY1151 bool "Prefer the buddy CPU hardlockup detector"1152 depends on HARDLOCKUP_DETECTOR1153 depends on HAVE_HARDLOCKUP_DETECTOR_PERF && HAVE_HARDLOCKUP_DETECTOR_BUDDY1154 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH1155 help1156 Say Y here to prefer the buddy hardlockup detector over the perf one.1157 1158 With the buddy detector, each CPU uses its softlockup hrtimer1159 to check that the next CPU is processing hrtimer interrupts by1160 verifying that a counter is increasing.1161 1162 This hardlockup detector is useful on systems that don't have1163 an arch-specific hardlockup detector or if resources needed1164 for the hardlockup detector are better used for other things.1165 1166config HARDLOCKUP_DETECTOR_PERF1167 bool1168 depends on HARDLOCKUP_DETECTOR1169 depends on HAVE_HARDLOCKUP_DETECTOR_PERF && !HARDLOCKUP_DETECTOR_PREFER_BUDDY1170 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH1171 select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER1172 1173config HARDLOCKUP_DETECTOR_BUDDY1174 bool1175 depends on HARDLOCKUP_DETECTOR1176 depends on HAVE_HARDLOCKUP_DETECTOR_BUDDY1177 depends on !HAVE_HARDLOCKUP_DETECTOR_PERF || HARDLOCKUP_DETECTOR_PREFER_BUDDY1178 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH1179 select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER1180 1181config HARDLOCKUP_DETECTOR_ARCH1182 bool1183 depends on HARDLOCKUP_DETECTOR1184 depends on HAVE_HARDLOCKUP_DETECTOR_ARCH1185 help1186 The arch-specific implementation of the hardlockup detector will1187 be used.1188 1189#1190# Both the "perf" and "buddy" hardlockup detectors count hrtimer1191# interrupts. This config enables functions managing this common code.1192#1193config HARDLOCKUP_DETECTOR_COUNTS_HRTIMER1194 bool1195 select SOFTLOCKUP_DETECTOR1196 1197#1198# Enables a timestamp based low pass filter to compensate for perf based1199# hard lockup detection which runs too fast due to turbo modes.1200#1201config HARDLOCKUP_CHECK_TIMESTAMP1202 bool1203 1204config BOOTPARAM_HARDLOCKUP_PANIC1205 bool "Panic (Reboot) On Hard Lockups"1206 depends on HARDLOCKUP_DETECTOR1207 help1208 Say Y here to enable the kernel to panic on "hard lockups",1209 which are bugs that cause the kernel to loop in kernel1210 mode with interrupts disabled for more than 10 seconds (configurable1211 using the watchdog_thresh sysctl).1212 1213 Say N if unsure.1214 1215config DETECT_HUNG_TASK1216 bool "Detect Hung Tasks"1217 depends on DEBUG_KERNEL1218 default SOFTLOCKUP_DETECTOR1219 help1220 Say Y here to enable the kernel to detect "hung tasks",1221 which are bugs that cause the task to be stuck in1222 uninterruptible "D" state indefinitely.1223 1224 When a hung task is detected, the kernel will print the1225 current stack trace (which you should report), but the1226 task will stay in uninterruptible state. If lockdep is1227 enabled then all held locks will also be reported. This1228 feature has negligible overhead.1229 1230config DEFAULT_HUNG_TASK_TIMEOUT1231 int "Default timeout for hung task detection (in seconds)"1232 depends on DETECT_HUNG_TASK1233 default 1201234 help1235 This option controls the default timeout (in seconds) used1236 to determine when a task has become non-responsive and should1237 be considered hung.1238 1239 It can be adjusted at runtime via the kernel.hung_task_timeout_secs1240 sysctl or by writing a value to1241 /proc/sys/kernel/hung_task_timeout_secs.1242 1243 A timeout of 0 disables the check. The default is two minutes.1244 Keeping the default should be fine in most cases.1245 1246config BOOTPARAM_HUNG_TASK_PANIC1247 bool "Panic (Reboot) On Hung Tasks"1248 depends on DETECT_HUNG_TASK1249 help1250 Say Y here to enable the kernel to panic on "hung tasks",1251 which are bugs that cause the kernel to leave a task stuck1252 in uninterruptible "D" state.1253 1254 The panic can be used in combination with panic_timeout,1255 to cause the system to reboot automatically after a1256 hung task has been detected. This feature is useful for1257 high-availability systems that have uptime guarantees and1258 where a hung tasks must be resolved ASAP.1259 1260 Say N if unsure.1261 1262config WQ_WATCHDOG1263 bool "Detect Workqueue Stalls"1264 depends on DEBUG_KERNEL1265 help1266 Say Y here to enable stall detection on workqueues. If a1267 worker pool doesn't make forward progress on a pending work1268 item for over a given amount of time, 30s by default, a1269 warning message is printed along with dump of workqueue1270 state. This can be configured through kernel parameter1271 "workqueue.watchdog_thresh" and its sysfs counterpart.1272 1273config WQ_CPU_INTENSIVE_REPORT1274 bool "Report per-cpu work items which hog CPU for too long"1275 depends on DEBUG_KERNEL1276 help1277 Say Y here to enable reporting of concurrency-managed per-cpu work1278 items that hog CPUs for longer than1279 workqueue.cpu_intensive_thresh_us. Workqueue automatically1280 detects and excludes them from concurrency management to prevent1281 them from stalling other per-cpu work items. Occassional1282 triggering may not necessarily indicate a problem. Repeated1283 triggering likely indicates that the work item should be switched1284 to use an unbound workqueue.1285 1286config TEST_LOCKUP1287 tristate "Test module to generate lockups"1288 depends on m1289 help1290 This builds the "test_lockup" module that helps to make sure1291 that watchdogs and lockup detectors are working properly.1292 1293 Depending on module parameters it could emulate soft or hard1294 lockup, "hung task", or locking arbitrary lock for a long time.1295 Also it could generate series of lockups with cooling-down periods.1296 1297 If unsure, say N.1298 1299endmenu # "Debug lockups and hangs"1300 1301menu "Scheduler Debugging"1302 1303config SCHED_DEBUG1304 bool "Collect scheduler debugging info"1305 depends on DEBUG_KERNEL && DEBUG_FS1306 default y1307 help1308 If you say Y here, the /sys/kernel/debug/sched file will be provided1309 that can help debug the scheduler. The runtime overhead of this1310 option is minimal.1311 1312config SCHED_INFO1313 bool1314 default n1315 1316config SCHEDSTATS1317 bool "Collect scheduler statistics"1318 depends on PROC_FS1319 select SCHED_INFO1320 help1321 If you say Y here, additional code will be inserted into the1322 scheduler and related routines to collect statistics about1323 scheduler behavior and provide them in /proc/schedstat. These1324 stats may be useful for both tuning and debugging the scheduler1325 If you aren't debugging the scheduler or trying to tune a specific1326 application, you can say N to avoid the very slight overhead1327 this adds.1328 1329endmenu1330 1331config DEBUG_TIMEKEEPING1332 bool "Enable extra timekeeping sanity checking"1333 help1334 This option will enable additional timekeeping sanity checks1335 which may be helpful when diagnosing issues where timekeeping1336 problems are suspected.1337 1338 This may include checks in the timekeeping hotpaths, so this1339 option may have a (very small) performance impact to some1340 workloads.1341 1342 If unsure, say N.1343 1344config DEBUG_PREEMPT1345 bool "Debug preemptible kernel"1346 depends on DEBUG_KERNEL && PREEMPTION && TRACE_IRQFLAGS_SUPPORT1347 help1348 If you say Y here then the kernel will use a debug variant of the1349 commonly used smp_processor_id() function and will print warnings1350 if kernel code uses it in a preemption-unsafe way. Also, the kernel1351 will detect preemption count underflows.1352 1353 This option has potential to introduce high runtime overhead,1354 depending on workload as it triggers debugging routines for each1355 this_cpu operation. It should only be used for debugging purposes.1356 1357menu "Lock Debugging (spinlocks, mutexes, etc...)"1358 1359config LOCK_DEBUGGING_SUPPORT1360 bool1361 depends on TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT1362 default y1363 1364config PROVE_LOCKING1365 bool "Lock debugging: prove locking correctness"1366 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT1367 select LOCKDEP1368 select DEBUG_SPINLOCK1369 select DEBUG_MUTEXES if !PREEMPT_RT1370 select DEBUG_RT_MUTEXES if RT_MUTEXES1371 select DEBUG_RWSEMS if !PREEMPT_RT1372 select DEBUG_WW_MUTEX_SLOWPATH1373 select DEBUG_LOCK_ALLOC1374 select PREEMPT_COUNT if !ARCH_NO_PREEMPT1375 select TRACE_IRQFLAGS1376 default n1377 help1378 This feature enables the kernel to prove that all locking1379 that occurs in the kernel runtime is mathematically1380 correct: that under no circumstance could an arbitrary (and1381 not yet triggered) combination of observed locking1382 sequences (on an arbitrary number of CPUs, running an1383 arbitrary number of tasks and interrupt contexts) cause a1384 deadlock.1385 1386 In short, this feature enables the kernel to report locking1387 related deadlocks before they actually occur.1388 1389 The proof does not depend on how hard and complex a1390 deadlock scenario would be to trigger: how many1391 participant CPUs, tasks and irq-contexts would be needed1392 for it to trigger. The proof also does not depend on1393 timing: if a race and a resulting deadlock is possible1394 theoretically (no matter how unlikely the race scenario1395 is), it will be proven so and will immediately be1396 reported by the kernel (once the event is observed that1397 makes the deadlock theoretically possible).1398 1399 If a deadlock is impossible (i.e. the locking rules, as1400 observed by the kernel, are mathematically correct), the1401 kernel reports nothing.1402 1403 NOTE: this feature can also be enabled for rwlocks, mutexes1404 and rwsems - in which case all dependencies between these1405 different locking variants are observed and mapped too, and1406 the proof of observed correctness is also maintained for an1407 arbitrary combination of these separate locking variants.1408 1409 For more details, see Documentation/locking/lockdep-design.rst.1410 1411config PROVE_RAW_LOCK_NESTING1412 bool "Enable raw_spinlock - spinlock nesting checks"1413 depends on PROVE_LOCKING1414 default n1415 help1416 Enable the raw_spinlock vs. spinlock nesting checks which ensure1417 that the lock nesting rules for PREEMPT_RT enabled kernels are1418 not violated.1419 1420 NOTE: There are known nesting problems. So if you enable this1421 option expect lockdep splats until these problems have been fully1422 addressed which is work in progress. This config switch allows to1423 identify and analyze these problems. It will be removed and the1424 check permanently enabled once the main issues have been fixed.1425 1426 If unsure, select N.1427 1428config LOCK_STAT1429 bool "Lock usage statistics"1430 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT1431 select LOCKDEP1432 select DEBUG_SPINLOCK1433 select DEBUG_MUTEXES if !PREEMPT_RT1434 select DEBUG_RT_MUTEXES if RT_MUTEXES1435 select DEBUG_LOCK_ALLOC1436 default n1437 help1438 This feature enables tracking lock contention points1439 1440 For more details, see Documentation/locking/lockstat.rst1441 1442 This also enables lock events required by "perf lock",1443 subcommand of perf.1444 If you want to use "perf lock", you also need to turn on1445 CONFIG_EVENT_TRACING.1446 1447 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events.1448 (CONFIG_LOCKDEP defines "acquire" and "release" events.)1449 1450config DEBUG_RT_MUTEXES1451 bool "RT Mutex debugging, deadlock detection"1452 depends on DEBUG_KERNEL && RT_MUTEXES1453 help1454 This allows rt mutex semantics violations and rt mutex related1455 deadlocks (lockups) to be detected and reported automatically.1456 1457config DEBUG_SPINLOCK1458 bool "Spinlock and rw-lock debugging: basic checks"1459 depends on DEBUG_KERNEL1460 select UNINLINE_SPIN_UNLOCK1461 help1462 Say Y here and build SMP to catch missing spinlock initialization1463 and certain other kinds of spinlock errors commonly made. This is1464 best used in conjunction with the NMI watchdog so that spinlock1465 deadlocks are also debuggable.1466 1467config DEBUG_MUTEXES1468 bool "Mutex debugging: basic checks"1469 depends on DEBUG_KERNEL && !PREEMPT_RT1470 help1471 This feature allows mutex semantics violations to be detected and1472 reported.1473 1474config DEBUG_WW_MUTEX_SLOWPATH1475 bool "Wait/wound mutex debugging: Slowpath testing"1476 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT1477 select DEBUG_LOCK_ALLOC1478 select DEBUG_SPINLOCK1479 select DEBUG_MUTEXES if !PREEMPT_RT1480 select DEBUG_RT_MUTEXES if PREEMPT_RT1481 help1482 This feature enables slowpath testing for w/w mutex users by1483 injecting additional -EDEADLK wound/backoff cases. Together with1484 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this1485 will test all possible w/w mutex interface abuse with the1486 exception of simply not acquiring all the required locks.1487 Note that this feature can introduce significant overhead, so1488 it really should not be enabled in a production or distro kernel,1489 even a debug kernel. If you are a driver writer, enable it. If1490 you are a distro, do not.1491 1492config DEBUG_RWSEMS1493 bool "RW Semaphore debugging: basic checks"1494 depends on DEBUG_KERNEL && !PREEMPT_RT1495 help1496 This debugging feature allows mismatched rw semaphore locks1497 and unlocks to be detected and reported.1498 1499config DEBUG_LOCK_ALLOC1500 bool "Lock debugging: detect incorrect freeing of live locks"1501 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT1502 select DEBUG_SPINLOCK1503 select DEBUG_MUTEXES if !PREEMPT_RT1504 select DEBUG_RT_MUTEXES if RT_MUTEXES1505 select LOCKDEP1506 help1507 This feature will check whether any held lock (spinlock, rwlock,1508 mutex or rwsem) is incorrectly freed by the kernel, via any of the1509 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),1510 vfree(), etc.), whether a live lock is incorrectly reinitialized via1511 spin_lock_init()/mutex_init()/etc., or whether there is any lock1512 held during task exit.1513 1514config LOCKDEP1515 bool1516 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT1517 select STACKTRACE1518 select KALLSYMS1519 select KALLSYMS_ALL1520 1521config LOCKDEP_SMALL1522 bool1523 1524config LOCKDEP_BITS1525 int "Bitsize for MAX_LOCKDEP_ENTRIES"1526 depends on LOCKDEP && !LOCKDEP_SMALL1527 range 10 301528 default 151529 help1530 Try increasing this value if you hit "BUG: MAX_LOCKDEP_ENTRIES too low!" message.1531 1532config LOCKDEP_CHAINS_BITS1533 int "Bitsize for MAX_LOCKDEP_CHAINS"1534 depends on LOCKDEP && !LOCKDEP_SMALL1535 range 10 211536 default 161537 help1538 Try increasing this value if you hit "BUG: MAX_LOCKDEP_CHAINS too low!" message.1539 1540config LOCKDEP_STACK_TRACE_BITS1541 int "Bitsize for MAX_STACK_TRACE_ENTRIES"1542 depends on LOCKDEP && !LOCKDEP_SMALL1543 range 10 301544 default 191545 help1546 Try increasing this value if you hit "BUG: MAX_STACK_TRACE_ENTRIES too low!" message.1547 1548config LOCKDEP_STACK_TRACE_HASH_BITS1549 int "Bitsize for STACK_TRACE_HASH_SIZE"1550 depends on LOCKDEP && !LOCKDEP_SMALL1551 range 10 301552 default 141553 help1554 Try increasing this value if you need large STACK_TRACE_HASH_SIZE.1555 1556config LOCKDEP_CIRCULAR_QUEUE_BITS1557 int "Bitsize for elements in circular_queue struct"1558 depends on LOCKDEP1559 range 10 301560 default 121561 help1562 Try increasing this value if you hit "lockdep bfs error:-1" warning due to __cq_enqueue() failure.1563 1564config DEBUG_LOCKDEP1565 bool "Lock dependency engine debugging"1566 depends on DEBUG_KERNEL && LOCKDEP1567 select DEBUG_IRQFLAGS1568 help1569 If you say Y here, the lock dependency engine will do1570 additional runtime checks to debug itself, at the price1571 of more runtime overhead.1572 1573config DEBUG_ATOMIC_SLEEP1574 bool "Sleep inside atomic section checking"1575 select PREEMPT_COUNT1576 depends on DEBUG_KERNEL1577 depends on !ARCH_NO_PREEMPT1578 help1579 If you say Y here, various routines which may sleep will become very1580 noisy if they are called inside atomic sections: when a spinlock is1581 held, inside an rcu read side critical section, inside preempt disabled1582 sections, inside an interrupt, etc...1583 1584config DEBUG_LOCKING_API_SELFTESTS1585 bool "Locking API boot-time self-tests"1586 depends on DEBUG_KERNEL1587 help1588 Say Y here if you want the kernel to run a short self-test during1589 bootup. The self-test checks whether common types of locking bugs1590 are detected by debugging mechanisms or not. (if you disable1591 lock debugging then those bugs won't be detected of course.)1592 The following locking APIs are covered: spinlocks, rwlocks,1593 mutexes and rwsems.1594 1595config LOCK_TORTURE_TEST1596 tristate "torture tests for locking"1597 depends on DEBUG_KERNEL1598 select TORTURE_TEST1599 help1600 This option provides a kernel module that runs torture tests1601 on kernel locking primitives. The kernel module may be built1602 after the fact on the running kernel to be tested, if desired.1603 1604 Say Y here if you want kernel locking-primitive torture tests1605 to be built into the kernel.1606 Say M if you want these torture tests to build as a module.1607 Say N if you are unsure.1608 1609config WW_MUTEX_SELFTEST1610 tristate "Wait/wound mutex selftests"1611 help1612 This option provides a kernel module that runs tests on the1613 on the struct ww_mutex locking API.1614 1615 It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction1616 with this test harness.1617 1618 Say M if you want these self tests to build as a module.1619 Say N if you are unsure.1620 1621config SCF_TORTURE_TEST1622 tristate "torture tests for smp_call_function*()"1623 depends on DEBUG_KERNEL1624 select TORTURE_TEST1625 help1626 This option provides a kernel module that runs torture tests1627 on the smp_call_function() family of primitives. The kernel1628 module may be built after the fact on the running kernel to1629 be tested, if desired.1630 1631config CSD_LOCK_WAIT_DEBUG1632 bool "Debugging for csd_lock_wait(), called from smp_call_function*()"1633 depends on DEBUG_KERNEL1634 depends on SMP1635 depends on 64BIT1636 default n1637 help1638 This option enables debug prints when CPUs are slow to respond1639 to the smp_call_function*() IPI wrappers. These debug prints1640 include the IPI handler function currently executing (if any)1641 and relevant stack traces.1642 1643config CSD_LOCK_WAIT_DEBUG_DEFAULT1644 bool "Default csd_lock_wait() debugging on at boot time"1645 depends on CSD_LOCK_WAIT_DEBUG1646 depends on 64BIT1647 default n1648 help1649 This option causes the csdlock_debug= kernel boot parameter to1650 default to 1 (basic debugging) instead of 0 (no debugging).1651 1652endmenu # lock debugging1653 1654config TRACE_IRQFLAGS1655 depends on TRACE_IRQFLAGS_SUPPORT1656 bool1657 help1658 Enables hooks to interrupt enabling and disabling for1659 either tracing or lock debugging.1660 1661config TRACE_IRQFLAGS_NMI1662 def_bool y1663 depends on TRACE_IRQFLAGS1664 depends on TRACE_IRQFLAGS_NMI_SUPPORT1665 1666config NMI_CHECK_CPU1667 bool "Debugging for CPUs failing to respond to backtrace requests"1668 depends on DEBUG_KERNEL1669 depends on X861670 default n1671 help1672 Enables debug prints when a CPU fails to respond to a given1673 backtrace NMI. These prints provide some reasons why a CPU1674 might legitimately be failing to respond, for example, if it1675 is offline of if ignore_nmis is set.1676 1677config DEBUG_IRQFLAGS1678 bool "Debug IRQ flag manipulation"1679 help1680 Enables checks for potentially unsafe enabling or disabling of1681 interrupts, such as calling raw_local_irq_restore() when interrupts1682 are enabled.1683 1684config STACKTRACE1685 bool "Stack backtrace support"1686 depends on STACKTRACE_SUPPORT1687 help1688 This option causes the kernel to create a /proc/pid/stack for1689 every process, showing its current stack trace.1690 It is also used by various kernel debugging features that require1691 stack trace generation.1692 1693config WARN_ALL_UNSEEDED_RANDOM1694 bool "Warn for all uses of unseeded randomness"1695 default n1696 help1697 Some parts of the kernel contain bugs relating to their use of1698 cryptographically secure random numbers before it's actually possible1699 to generate those numbers securely. This setting ensures that these1700 flaws don't go unnoticed, by enabling a message, should this ever1701 occur. This will allow people with obscure setups to know when things1702 are going wrong, so that they might contact developers about fixing1703 it.1704 1705 Unfortunately, on some models of some architectures getting1706 a fully seeded CRNG is extremely difficult, and so this can1707 result in dmesg getting spammed for a surprisingly long1708 time. This is really bad from a security perspective, and1709 so architecture maintainers really need to do what they can1710 to get the CRNG seeded sooner after the system is booted.1711 However, since users cannot do anything actionable to1712 address this, by default this option is disabled.1713 1714 Say Y here if you want to receive warnings for all uses of1715 unseeded randomness. This will be of use primarily for1716 those developers interested in improving the security of1717 Linux kernels running on their architecture (or1718 subarchitecture).1719 1720config DEBUG_KOBJECT1721 bool "kobject debugging"1722 depends on DEBUG_KERNEL1723 help1724 If you say Y here, some extra kobject debugging messages will be sent1725 to the syslog.1726 1727config DEBUG_KOBJECT_RELEASE1728 bool "kobject release debugging"1729 depends on DEBUG_OBJECTS_TIMERS1730 help1731 kobjects are reference counted objects. This means that their1732 last reference count put is not predictable, and the kobject can1733 live on past the point at which a driver decides to drop its1734 initial reference to the kobject gained on allocation. An1735 example of this would be a struct device which has just been1736 unregistered.1737 1738 However, some buggy drivers assume that after such an operation,1739 the memory backing the kobject can be immediately freed. This1740 goes completely against the principles of a refcounted object.1741 1742 If you say Y here, the kernel will delay the release of kobjects1743 on the last reference count to improve the visibility of this1744 kind of kobject release bug.1745 1746config HAVE_DEBUG_BUGVERBOSE1747 bool1748 1749menu "Debug kernel data structures"1750 1751config DEBUG_LIST1752 bool "Debug linked list manipulation"1753 depends on DEBUG_KERNEL1754 select LIST_HARDENED1755 help1756 Enable this to turn on extended checks in the linked-list walking1757 routines.1758 1759 This option trades better quality error reports for performance, and1760 is more suitable for kernel debugging. If you care about performance,1761 you should only enable CONFIG_LIST_HARDENED instead.1762 1763 If unsure, say N.1764 1765config DEBUG_PLIST1766 bool "Debug priority linked list manipulation"1767 depends on DEBUG_KERNEL1768 help1769 Enable this to turn on extended checks in the priority-ordered1770 linked-list (plist) walking routines. This checks the entire1771 list multiple times during each manipulation.1772 1773 If unsure, say N.1774 1775config DEBUG_SG1776 bool "Debug SG table operations"1777 depends on DEBUG_KERNEL1778 help1779 Enable this to turn on checks on scatter-gather tables. This can1780 help find problems with drivers that do not properly initialize1781 their sg tables.1782 1783 If unsure, say N.1784 1785config DEBUG_NOTIFIERS1786 bool "Debug notifier call chains"1787 depends on DEBUG_KERNEL1788 help1789 Enable this to turn on sanity checking for notifier call chains.1790 This is most useful for kernel developers to make sure that1791 modules properly unregister themselves from notifier chains.1792 This is a relatively cheap check but if you care about maximum1793 performance, say N.1794 1795config DEBUG_CLOSURES1796 bool "Debug closures (bcache async widgits)"1797 depends on CLOSURES1798 select DEBUG_FS1799 help1800 Keeps all active closures in a linked list and provides a debugfs1801 interface to list them, which makes it possible to see asynchronous1802 operations that get stuck.1803 1804config DEBUG_MAPLE_TREE1805 bool "Debug maple trees"1806 depends on DEBUG_KERNEL1807 help1808 Enable maple tree debugging information and extra validations.1809 1810 If unsure, say N.1811 1812endmenu1813 1814source "kernel/rcu/Kconfig.debug"1815 1816config DEBUG_WQ_FORCE_RR_CPU1817 bool "Force round-robin CPU selection for unbound work items"1818 depends on DEBUG_KERNEL1819 default n1820 help1821 Workqueue used to implicitly guarantee that work items queued1822 without explicit CPU specified are put on the local CPU. This1823 guarantee is no longer true and while local CPU is still1824 preferred work items may be put on foreign CPUs. Kernel1825 parameter "workqueue.debug_force_rr_cpu" is added to force1826 round-robin CPU selection to flush out usages which depend on the1827 now broken guarantee. This config option enables the debug1828 feature by default. When enabled, memory and cache locality will1829 be impacted.1830 1831config CPU_HOTPLUG_STATE_CONTROL1832 bool "Enable CPU hotplug state control"1833 depends on DEBUG_KERNEL1834 depends on HOTPLUG_CPU1835 default n1836 help1837 Allows to write steps between "offline" and "online" to the CPUs1838 sysfs target file so states can be stepped granular. This is a debug1839 option for now as the hotplug machinery cannot be stopped and1840 restarted at arbitrary points yet.1841 1842 Say N if your are unsure.1843 1844config LATENCYTOP1845 bool "Latency measuring infrastructure"1846 depends on DEBUG_KERNEL1847 depends on STACKTRACE_SUPPORT1848 depends on PROC_FS1849 depends on FRAME_POINTER || MIPS || PPC || S390 || MICROBLAZE || ARM || ARC || X861850 select KALLSYMS1851 select KALLSYMS_ALL1852 select STACKTRACE1853 select SCHEDSTATS1854 help1855 Enable this option if you want to use the LatencyTOP tool1856 to find out which userspace is blocking on what kernel operations.1857 1858config DEBUG_CGROUP_REF1859 bool "Disable inlining of cgroup css reference count functions"1860 depends on DEBUG_KERNEL1861 depends on CGROUPS1862 depends on KPROBES1863 default n1864 help1865 Force cgroup css reference count functions to not be inlined so1866 that they can be kprobed for debugging.1867 1868source "kernel/trace/Kconfig"1869 1870config PROVIDE_OHCI1394_DMA_INIT1871 bool "Remote debugging over FireWire early on boot"1872 depends on PCI && X861873 help1874 If you want to debug problems which hang or crash the kernel early1875 on boot and the crashing machine has a FireWire port, you can use1876 this feature to remotely access the memory of the crashed machine1877 over FireWire. This employs remote DMA as part of the OHCI13941878 specification which is now the standard for FireWire controllers.1879 1880 With remote DMA, you can monitor the printk buffer remotely using1881 firescope and access all memory below 4GB using fireproxy from gdb.1882 Even controlling a kernel debugger is possible using remote DMA.1883 1884 Usage:1885 1886 If ohci1394_dma=early is used as boot parameter, it will initialize1887 all OHCI1394 controllers which are found in the PCI config space.1888 1889 As all changes to the FireWire bus such as enabling and disabling1890 devices cause a bus reset and thereby disable remote DMA for all1891 devices, be sure to have the cable plugged and FireWire enabled on1892 the debugging host before booting the debug target for debugging.1893 1894 This code (~1k) is freed after boot. By then, the firewire stack1895 in charge of the OHCI-1394 controllers should be used instead.1896 1897 See Documentation/core-api/debugging-via-ohci1394.rst for more information.1898 1899source "samples/Kconfig"1900 1901config ARCH_HAS_DEVMEM_IS_ALLOWED1902 bool1903 1904config STRICT_DEVMEM1905 bool "Filter access to /dev/mem"1906 depends on MMU && DEVMEM1907 depends on ARCH_HAS_DEVMEM_IS_ALLOWED || GENERIC_LIB_DEVMEM_IS_ALLOWED1908 default y if PPC || X86 || ARM641909 help1910 If this option is disabled, you allow userspace (root) access to all1911 of memory, including kernel and userspace memory. Accidental1912 access to this is obviously disastrous, but specific access can1913 be used by people debugging the kernel. Note that with PAT support1914 enabled, even in this case there are restrictions on /dev/mem1915 use due to the cache aliasing requirements.1916 1917 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem1918 file only allows userspace access to PCI space and the BIOS code and1919 data regions. This is sufficient for dosemu and X and all common1920 users of /dev/mem.1921 1922 If in doubt, say Y.1923 1924config IO_STRICT_DEVMEM1925 bool "Filter I/O access to /dev/mem"1926 depends on STRICT_DEVMEM1927 help1928 If this option is disabled, you allow userspace (root) access to all1929 io-memory regardless of whether a driver is actively using that1930 range. Accidental access to this is obviously disastrous, but1931 specific access can be used by people debugging kernel drivers.1932 1933 If this option is switched on, the /dev/mem file only allows1934 userspace access to *idle* io-memory ranges (see /proc/iomem) This1935 may break traditional users of /dev/mem (dosemu, legacy X, etc...)1936 if the driver using a given range cannot be disabled.1937 1938 If in doubt, say Y.1939 1940menu "$(SRCARCH) Debugging"1941 1942source "arch/$(SRCARCH)/Kconfig.debug"1943 1944endmenu1945 1946menu "Kernel Testing and Coverage"1947 1948source "lib/kunit/Kconfig"1949 1950config NOTIFIER_ERROR_INJECTION1951 tristate "Notifier error injection"1952 depends on DEBUG_KERNEL1953 select DEBUG_FS1954 help1955 This option provides the ability to inject artificial errors to1956 specified notifier chain callbacks. It is useful to test the error1957 handling of notifier call chain failures.1958 1959 Say N if unsure.1960 1961config PM_NOTIFIER_ERROR_INJECT1962 tristate "PM notifier error injection module"1963 depends on PM && NOTIFIER_ERROR_INJECTION1964 default m if PM_DEBUG1965 help1966 This option provides the ability to inject artificial errors to1967 PM notifier chain callbacks. It is controlled through debugfs1968 interface /sys/kernel/debug/notifier-error-inject/pm1969 1970 If the notifier call chain should be failed with some events1971 notified, write the error code to "actions/<notifier event>/error".1972 1973 Example: Inject PM suspend error (-12 = -ENOMEM)1974 1975 # cd /sys/kernel/debug/notifier-error-inject/pm/1976 # echo -12 > actions/PM_SUSPEND_PREPARE/error1977 # echo mem > /sys/power/state1978 bash: echo: write error: Cannot allocate memory1979 1980 To compile this code as a module, choose M here: the module will1981 be called pm-notifier-error-inject.1982 1983 If unsure, say N.1984 1985config OF_RECONFIG_NOTIFIER_ERROR_INJECT1986 tristate "OF reconfig notifier error injection module"1987 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION1988 help1989 This option provides the ability to inject artificial errors to1990 OF reconfig notifier chain callbacks. It is controlled1991 through debugfs interface under1992 /sys/kernel/debug/notifier-error-inject/OF-reconfig/1993 1994 If the notifier call chain should be failed with some events1995 notified, write the error code to "actions/<notifier event>/error".1996 1997 To compile this code as a module, choose M here: the module will1998 be called of-reconfig-notifier-error-inject.1999 2000 If unsure, say N.2001 2002config NETDEV_NOTIFIER_ERROR_INJECT2003 tristate "Netdev notifier error injection module"2004 depends on NET && NOTIFIER_ERROR_INJECTION2005 help2006 This option provides the ability to inject artificial errors to2007 netdevice notifier chain callbacks. It is controlled through debugfs2008 interface /sys/kernel/debug/notifier-error-inject/netdev2009 2010 If the notifier call chain should be failed with some events2011 notified, write the error code to "actions/<notifier event>/error".2012 2013 Example: Inject netdevice mtu change error (-22 = -EINVAL)2014 2015 # cd /sys/kernel/debug/notifier-error-inject/netdev2016 # echo -22 > actions/NETDEV_CHANGEMTU/error2017 # ip link set eth0 mtu 10242018 RTNETLINK answers: Invalid argument2019 2020 To compile this code as a module, choose M here: the module will2021 be called netdev-notifier-error-inject.2022 2023 If unsure, say N.2024 2025config FUNCTION_ERROR_INJECTION2026 bool "Fault-injections of functions"2027 depends on HAVE_FUNCTION_ERROR_INJECTION && KPROBES2028 help2029 Add fault injections into various functions that are annotated with2030 ALLOW_ERROR_INJECTION() in the kernel. BPF may also modify the return2031 value of these functions. This is useful to test error paths of code.2032 2033 If unsure, say N2034 2035config FAULT_INJECTION2036 bool "Fault-injection framework"2037 depends on DEBUG_KERNEL2038 help2039 Provide fault-injection framework.2040 For more details, see Documentation/fault-injection/.2041 2042config FAILSLAB2043 bool "Fault-injection capability for kmalloc"2044 depends on FAULT_INJECTION2045 help2046 Provide fault-injection capability for kmalloc.2047 2048config FAIL_PAGE_ALLOC2049 bool "Fault-injection capability for alloc_pages()"2050 depends on FAULT_INJECTION2051 help2052 Provide fault-injection capability for alloc_pages().2053 2054config FAULT_INJECTION_USERCOPY2055 bool "Fault injection capability for usercopy functions"2056 depends on FAULT_INJECTION2057 help2058 Provides fault-injection capability to inject failures2059 in usercopy functions (copy_from_user(), get_user(), ...).2060 2061config FAIL_MAKE_REQUEST2062 bool "Fault-injection capability for disk IO"2063 depends on FAULT_INJECTION && BLOCK2064 help2065 Provide fault-injection capability for disk IO.2066 2067config FAIL_IO_TIMEOUT2068 bool "Fault-injection capability for faking disk interrupts"2069 depends on FAULT_INJECTION && BLOCK2070 help2071 Provide fault-injection capability on end IO handling. This2072 will make the block layer "forget" an interrupt as configured,2073 thus exercising the error handling.2074 2075 Only works with drivers that use the generic timeout handling,2076 for others it won't do anything.2077 2078config FAIL_FUTEX2079 bool "Fault-injection capability for futexes"2080 select DEBUG_FS2081 depends on FAULT_INJECTION && FUTEX2082 help2083 Provide fault-injection capability for futexes.2084 2085config FAULT_INJECTION_DEBUG_FS2086 bool "Debugfs entries for fault-injection capabilities"2087 depends on FAULT_INJECTION && SYSFS && DEBUG_FS2088 help2089 Enable configuration of fault-injection capabilities via debugfs.2090 2091config FAIL_FUNCTION2092 bool "Fault-injection capability for functions"2093 depends on FAULT_INJECTION_DEBUG_FS && FUNCTION_ERROR_INJECTION2094 help2095 Provide function-based fault-injection capability.2096 This will allow you to override a specific function with a return2097 with given return value. As a result, function caller will see2098 an error value and have to handle it. This is useful to test the2099 error handling in various subsystems.2100 2101config FAIL_MMC_REQUEST2102 bool "Fault-injection capability for MMC IO"2103 depends on FAULT_INJECTION_DEBUG_FS && MMC2104 help2105 Provide fault-injection capability for MMC IO.2106 This will make the mmc core return data errors. This is2107 useful to test the error handling in the mmc block device2108 and to test how the mmc host driver handles retries from2109 the block device.2110 2111config FAIL_SUNRPC2112 bool "Fault-injection capability for SunRPC"2113 depends on FAULT_INJECTION_DEBUG_FS && SUNRPC_DEBUG2114 help2115 Provide fault-injection capability for SunRPC and2116 its consumers.2117 2118config FAULT_INJECTION_CONFIGFS2119 bool "Configfs interface for fault-injection capabilities"2120 depends on FAULT_INJECTION2121 select CONFIGFS_FS2122 help2123 This option allows configfs-based drivers to dynamically configure2124 fault-injection via configfs. Each parameter for driver-specific2125 fault-injection can be made visible as a configfs attribute in a2126 configfs group.2127 2128 2129config FAULT_INJECTION_STACKTRACE_FILTER2130 bool "stacktrace filter for fault-injection capabilities"2131 depends on FAULT_INJECTION2132 depends on (FAULT_INJECTION_DEBUG_FS || FAULT_INJECTION_CONFIGFS) && STACKTRACE_SUPPORT2133 select STACKTRACE2134 depends on FRAME_POINTER || MIPS || PPC || S390 || MICROBLAZE || ARM || ARC || X862135 help2136 Provide stacktrace filter for fault-injection capabilities2137 2138config ARCH_HAS_KCOV2139 bool2140 help2141 An architecture should select this when it can successfully2142 build and run with CONFIG_KCOV. This typically requires2143 disabling instrumentation for some early boot code.2144 2145config CC_HAS_SANCOV_TRACE_PC2146 def_bool $(cc-option,-fsanitize-coverage=trace-pc)2147 2148 2149config KCOV2150 bool "Code coverage for fuzzing"2151 depends on ARCH_HAS_KCOV2152 depends on CC_HAS_SANCOV_TRACE_PC || GCC_PLUGINS2153 depends on !ARCH_WANTS_NO_INSTR || HAVE_NOINSTR_HACK || \2154 GCC_VERSION >= 120000 || CC_IS_CLANG2155 select DEBUG_FS2156 select GCC_PLUGIN_SANCOV if !CC_HAS_SANCOV_TRACE_PC2157 select OBJTOOL if HAVE_NOINSTR_HACK2158 help2159 KCOV exposes kernel code coverage information in a form suitable2160 for coverage-guided fuzzing (randomized testing).2161 2162 For more details, see Documentation/dev-tools/kcov.rst.2163 2164config KCOV_ENABLE_COMPARISONS2165 bool "Enable comparison operands collection by KCOV"2166 depends on KCOV2167 depends on $(cc-option,-fsanitize-coverage=trace-cmp)2168 help2169 KCOV also exposes operands of every comparison in the instrumented2170 code along with operand sizes and PCs of the comparison instructions.2171 These operands can be used by fuzzing engines to improve the quality2172 of fuzzing coverage.2173 2174config KCOV_INSTRUMENT_ALL2175 bool "Instrument all code by default"2176 depends on KCOV2177 default y2178 help2179 If you are doing generic system call fuzzing (like e.g. syzkaller),2180 then you will want to instrument the whole kernel and you should2181 say y here. If you are doing more targeted fuzzing (like e.g.2182 filesystem fuzzing with AFL) then you will want to enable coverage2183 for more specific subsets of files, and should say n here.2184 2185config KCOV_IRQ_AREA_SIZE2186 hex "Size of interrupt coverage collection area in words"2187 depends on KCOV2188 default 0x400002189 help2190 KCOV uses preallocated per-cpu areas to collect coverage from2191 soft interrupts. This specifies the size of those areas in the2192 number of unsigned long words.2193 2194config KCOV_SELFTEST2195 bool "Perform short selftests on boot"2196 depends on KCOV2197 help2198 Run short KCOV coverage collection selftests on boot.2199 On test failure, causes the kernel to panic. Recommended to be2200 enabled, ensuring critical functionality works as intended.2201 2202menuconfig RUNTIME_TESTING_MENU2203 bool "Runtime Testing"2204 default y2205 2206if RUNTIME_TESTING_MENU2207 2208config TEST_DHRY2209 tristate "Dhrystone benchmark test"2210 help2211 Enable this to include the Dhrystone 2.1 benchmark. This test2212 calculates the number of Dhrystones per second, and the number of2213 DMIPS (Dhrystone MIPS) obtained when the Dhrystone score is divided2214 by 1757 (the number of Dhrystones per second obtained on the VAX2215 11/780, nominally a 1 MIPS machine).2216 2217 To run the benchmark, it needs to be enabled explicitly, either from2218 the kernel command line (when built-in), or from userspace (when2219 built-in or modular).2220 2221 Run once during kernel boot:2222 2223 test_dhry.run2224 2225 Set number of iterations from kernel command line:2226 2227 test_dhry.iterations=<n>2228 2229 Set number of iterations from userspace:2230 2231 echo <n> > /sys/module/test_dhry/parameters/iterations2232 2233 Trigger manual run from userspace:2234 2235 echo y > /sys/module/test_dhry/parameters/run2236 2237 If the number of iterations is <= 0, the test will devise a suitable2238 number of iterations (test runs for at least 2s) automatically.2239 This process takes ca. 4s.2240 2241 If unsure, say N.2242 2243config LKDTM2244 tristate "Linux Kernel Dump Test Tool Module"2245 depends on DEBUG_FS2246 help2247 This module enables testing of the different dumping mechanisms by2248 inducing system failures at predefined crash points.2249 If you don't need it: say N2250 Choose M here to compile this code as a module. The module will be2251 called lkdtm.2252 2253 Documentation on how to use the module can be found in2254 Documentation/fault-injection/provoke-crashes.rst2255 2256config CPUMASK_KUNIT_TEST2257 tristate "KUnit test for cpumask" if !KUNIT_ALL_TESTS2258 depends on KUNIT2259 default KUNIT_ALL_TESTS2260 help2261 Enable to turn on cpumask tests, running at boot or module load time.2262 2263 For more information on KUnit and unit tests in general, please refer2264 to the KUnit documentation in Documentation/dev-tools/kunit/.2265 2266 If unsure, say N.2267 2268config TEST_LIST_SORT2269 tristate "Linked list sorting test" if !KUNIT_ALL_TESTS2270 depends on KUNIT2271 default KUNIT_ALL_TESTS2272 help2273 Enable this to turn on 'list_sort()' function test. This test is2274 executed only once during system boot (so affects only boot time),2275 or at module load time.2276 2277 If unsure, say N.2278 2279config TEST_MIN_HEAP2280 tristate "Min heap test"2281 depends on DEBUG_KERNEL || m2282 help2283 Enable this to turn on min heap function tests. This test is2284 executed only once during system boot (so affects only boot time),2285 or at module load time.2286 2287 If unsure, say N.2288 2289config TEST_SORT2290 tristate "Array-based sort test" if !KUNIT_ALL_TESTS2291 depends on KUNIT2292 default KUNIT_ALL_TESTS2293 help2294 This option enables the self-test function of 'sort()' at boot,2295 or at module load time.2296 2297 If unsure, say N.2298 2299config TEST_DIV642300 tristate "64bit/32bit division and modulo test"2301 depends on DEBUG_KERNEL || m2302 help2303 Enable this to turn on 'do_div()' function test. This test is2304 executed only once during system boot (so affects only boot time),2305 or at module load time.2306 2307 If unsure, say N.2308 2309config TEST_MULDIV642310 tristate "mul_u64_u64_div_u64() test"2311 depends on DEBUG_KERNEL || m2312 help2313 Enable this to turn on 'mul_u64_u64_div_u64()' function test.2314 This test is executed only once during system boot (so affects2315 only boot time), or at module load time.2316 2317 If unsure, say N.2318 2319config TEST_IOV_ITER2320 tristate "Test iov_iter operation" if !KUNIT_ALL_TESTS2321 depends on KUNIT2322 depends on MMU2323 default KUNIT_ALL_TESTS2324 help2325 Enable this to turn on testing of the operation of the I/O iterator2326 (iov_iter). This test is executed only once during system boot (so2327 affects only boot time), or at module load time.2328 2329 If unsure, say N.2330 2331config KPROBES_SANITY_TEST2332 tristate "Kprobes sanity tests" if !KUNIT_ALL_TESTS2333 depends on DEBUG_KERNEL2334 depends on KPROBES2335 depends on KUNIT2336 select STACKTRACE if ARCH_CORRECT_STACKTRACE_ON_KRETPROBE2337 default KUNIT_ALL_TESTS2338 help2339 This option provides for testing basic kprobes functionality on2340 boot. Samples of kprobe and kretprobe are inserted and2341 verified for functionality.2342 2343 Say N if you are unsure.2344 2345config FPROBE_SANITY_TEST2346 bool "Self test for fprobe"2347 depends on DEBUG_KERNEL2348 depends on FPROBE2349 depends on KUNIT=y2350 help2351 This option will enable testing the fprobe when the system boot.2352 A series of tests are made to verify that the fprobe is functioning2353 properly.2354 2355 Say N if you are unsure.2356 2357config BACKTRACE_SELF_TEST2358 tristate "Self test for the backtrace code"2359 depends on DEBUG_KERNEL2360 help2361 This option provides a kernel module that can be used to test2362 the kernel stack backtrace code. This option is not useful2363 for distributions or general kernels, but only for kernel2364 developers working on architecture code.2365 2366 Note that if you want to also test saved backtraces, you will2367 have to enable STACKTRACE as well.2368 2369 Say N if you are unsure.2370 2371config TEST_REF_TRACKER2372 tristate "Self test for reference tracker"2373 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT2374 select REF_TRACKER2375 help2376 This option provides a kernel module performing tests2377 using reference tracker infrastructure.2378 2379 Say N if you are unsure.2380 2381config RBTREE_TEST2382 tristate "Red-Black tree test"2383 depends on DEBUG_KERNEL2384 help2385 A benchmark measuring the performance of the rbtree library.2386 Also includes rbtree invariant checks.2387 2388config REED_SOLOMON_TEST2389 tristate "Reed-Solomon library test"2390 depends on DEBUG_KERNEL || m2391 select REED_SOLOMON2392 select REED_SOLOMON_ENC162393 select REED_SOLOMON_DEC162394 help2395 This option enables the self-test function of rslib at boot,2396 or at module load time.2397 2398 If unsure, say N.2399 2400config INTERVAL_TREE_TEST2401 tristate "Interval tree test"2402 depends on DEBUG_KERNEL2403 select INTERVAL_TREE2404 help2405 A benchmark measuring the performance of the interval tree library2406 2407config PERCPU_TEST2408 tristate "Per cpu operations test"2409 depends on m && DEBUG_KERNEL2410 help2411 Enable this option to build test module which validates per-cpu2412 operations.2413 2414 If unsure, say N.2415 2416config ATOMIC64_SELFTEST2417 tristate "Perform an atomic64_t self-test"2418 help2419 Enable this option to test the atomic64_t functions at boot or2420 at module load time.2421 2422 If unsure, say N.2423 2424config ASYNC_RAID6_TEST2425 tristate "Self test for hardware accelerated raid6 recovery"2426 depends on ASYNC_RAID6_RECOV2427 select ASYNC_MEMCPY2428 help2429 This is a one-shot self test that permutes through the2430 recovery of all the possible two disk failure scenarios for a2431 N-disk array. Recovery is performed with the asynchronous2432 raid6 recovery routines, and will optionally use an offload2433 engine if one is available.2434 2435 If unsure, say N.2436 2437config TEST_HEXDUMP2438 tristate "Test functions located in the hexdump module at runtime"2439 2440config STRING_KUNIT_TEST2441 tristate "KUnit test string functions at runtime" if !KUNIT_ALL_TESTS2442 depends on KUNIT2443 default KUNIT_ALL_TESTS2444 2445config STRING_HELPERS_KUNIT_TEST2446 tristate "KUnit test string helpers at runtime" if !KUNIT_ALL_TESTS2447 depends on KUNIT2448 default KUNIT_ALL_TESTS2449 2450config TEST_KSTRTOX2451 tristate "Test kstrto*() family of functions at runtime"2452 2453config TEST_PRINTF2454 tristate "Test printf() family of functions at runtime"2455 2456config TEST_SCANF2457 tristate "Test scanf() family of functions at runtime"2458 2459config TEST_BITMAP2460 tristate "Test bitmap_*() family of functions at runtime"2461 help2462 Enable this option to test the bitmap functions at boot.2463 2464 If unsure, say N.2465 2466config TEST_UUID2467 tristate "Test functions located in the uuid module at runtime"2468 2469config TEST_XARRAY2470 tristate "Test the XArray code at runtime"2471 2472config TEST_MAPLE_TREE2473 tristate "Test the Maple Tree code at runtime or module load"2474 help2475 Enable this option to test the maple tree code functions at boot, or2476 when the module is loaded. Enable "Debug Maple Trees" will enable2477 more verbose output on failures.2478 2479 If unsure, say N.2480 2481config TEST_RHASHTABLE2482 tristate "Perform selftest on resizable hash table"2483 help2484 Enable this option to test the rhashtable functions at boot.2485 2486 If unsure, say N.2487 2488config TEST_IDA2489 tristate "Perform selftest on IDA functions"2490 2491config TEST_PARMAN2492 tristate "Perform selftest on priority array manager"2493 depends on PARMAN2494 help2495 Enable this option to test priority array manager on boot2496 (or module load).2497 2498 If unsure, say N.2499 2500config TEST_IRQ_TIMINGS2501 bool "IRQ timings selftest"2502 depends on IRQ_TIMINGS2503 help2504 Enable this option to test the irq timings code on boot.2505 2506 If unsure, say N.2507 2508config TEST_LKM2509 tristate "Test module loading with 'hello world' module"2510 depends on m2511 help2512 This builds the "test_module" module that emits "Hello, world"2513 on printk when loaded. It is designed to be used for basic2514 evaluation of the module loading subsystem (for example when2515 validating module verification). It lacks any extra dependencies,2516 and will not normally be loaded by the system unless explicitly2517 requested by name.2518 2519 If unsure, say N.2520 2521config TEST_BITOPS2522 tristate "Test module for compilation of bitops operations"2523 help2524 This builds the "test_bitops" module that is much like the2525 TEST_LKM module except that it does a basic exercise of the2526 set/clear_bit macros and get_count_order/long to make sure there are2527 no compiler warnings from C=1 sparse checker or -Wextra2528 compilations. It has no dependencies and doesn't run or load unless2529 explicitly requested by name. for example: modprobe test_bitops.2530 2531 If unsure, say N.2532 2533config TEST_VMALLOC2534 tristate "Test module for stress/performance analysis of vmalloc allocator"2535 default n2536 depends on MMU2537 depends on m2538 help2539 This builds the "test_vmalloc" module that should be used for2540 stress and performance analysis. So, any new change for vmalloc2541 subsystem can be evaluated from performance and stability point2542 of view.2543 2544 If unsure, say N.2545 2546config TEST_BPF2547 tristate "Test BPF filter functionality"2548 depends on m && NET2549 help2550 This builds the "test_bpf" module that runs various test vectors2551 against the BPF interpreter or BPF JIT compiler depending on the2552 current setting. This is in particular useful for BPF JIT compiler2553 development, but also to run regression tests against changes in2554 the interpreter code. It also enables test stubs for eBPF maps and2555 verifier used by user space verifier testsuite.2556 2557 If unsure, say N.2558 2559config TEST_BLACKHOLE_DEV2560 tristate "Test blackhole netdev functionality"2561 depends on m && NET2562 help2563 This builds the "test_blackhole_dev" module that validates the2564 data path through this blackhole netdev.2565 2566 If unsure, say N.2567 2568config FIND_BIT_BENCHMARK2569 tristate "Test find_bit functions"2570 help2571 This builds the "test_find_bit" module that measure find_*_bit()2572 functions performance.2573 2574 If unsure, say N.2575 2576config TEST_FIRMWARE2577 tristate "Test firmware loading via userspace interface"2578 depends on FW_LOADER2579 help2580 This builds the "test_firmware" module that creates a userspace2581 interface for testing firmware loading. This can be used to2582 control the triggering of firmware loading without needing an2583 actual firmware-using device. The contents can be rechecked by2584 userspace.2585 2586 If unsure, say N.2587 2588config TEST_SYSCTL2589 tristate "sysctl test driver"2590 depends on PROC_SYSCTL2591 help2592 This builds the "test_sysctl" module. This driver enables to test the2593 proc sysctl interfaces available to drivers safely without affecting2594 production knobs which might alter system functionality.2595 2596 If unsure, say N.2597 2598config BITFIELD_KUNIT2599 tristate "KUnit test bitfield functions at runtime" if !KUNIT_ALL_TESTS2600 depends on KUNIT2601 default KUNIT_ALL_TESTS2602 help2603 Enable this option to test the bitfield functions at boot.2604 2605 KUnit tests run during boot and output the results to the debug log2606 in TAP format (http://testanything.org/). Only useful for kernel devs2607 running the KUnit test harness, and not intended for inclusion into a2608 production build.2609 2610 For more information on KUnit and unit tests in general please refer2611 to the KUnit documentation in Documentation/dev-tools/kunit/.2612 2613 If unsure, say N.2614 2615config CHECKSUM_KUNIT2616 tristate "KUnit test checksum functions at runtime" if !KUNIT_ALL_TESTS2617 depends on KUNIT2618 default KUNIT_ALL_TESTS2619 help2620 Enable this option to test the checksum functions at boot.2621 2622 KUnit tests run during boot and output the results to the debug log2623 in TAP format (http://testanything.org/). Only useful for kernel devs2624 running the KUnit test harness, and not intended for inclusion into a2625 production build.2626 2627 For more information on KUnit and unit tests in general please refer2628 to the KUnit documentation in Documentation/dev-tools/kunit/.2629 2630 If unsure, say N.2631 2632config HASH_KUNIT_TEST2633 tristate "KUnit Test for integer hash functions" if !KUNIT_ALL_TESTS2634 depends on KUNIT2635 default KUNIT_ALL_TESTS2636 help2637 Enable this option to test the kernel's string (<linux/stringhash.h>), and2638 integer (<linux/hash.h>) hash functions on boot.2639 2640 KUnit tests run during boot and output the results to the debug log2641 in TAP format (https://testanything.org/). Only useful for kernel devs2642 running the KUnit test harness, and not intended for inclusion into a2643 production build.2644 2645 For more information on KUnit and unit tests in general please refer2646 to the KUnit documentation in Documentation/dev-tools/kunit/.2647 2648 This is intended to help people writing architecture-specific2649 optimized versions. If unsure, say N.2650 2651config RESOURCE_KUNIT_TEST2652 tristate "KUnit test for resource API" if !KUNIT_ALL_TESTS2653 depends on KUNIT2654 default KUNIT_ALL_TESTS2655 select GET_FREE_REGION2656 help2657 This builds the resource API unit test.2658 Tests the logic of API provided by resource.c and ioport.h.2659 For more information on KUnit and unit tests in general please refer2660 to the KUnit documentation in Documentation/dev-tools/kunit/.2661 2662 If unsure, say N.2663 2664config SYSCTL_KUNIT_TEST2665 tristate "KUnit test for sysctl" if !KUNIT_ALL_TESTS2666 depends on KUNIT2667 default KUNIT_ALL_TESTS2668 help2669 This builds the proc sysctl unit test, which runs on boot.2670 Tests the API contract and implementation correctness of sysctl.2671 For more information on KUnit and unit tests in general please refer2672 to the KUnit documentation in Documentation/dev-tools/kunit/.2673 2674 If unsure, say N.2675 2676config LIST_KUNIT_TEST2677 tristate "KUnit Test for Kernel Linked-list structures" if !KUNIT_ALL_TESTS2678 depends on KUNIT2679 default KUNIT_ALL_TESTS2680 help2681 This builds the linked list KUnit test suite.2682 It tests that the API and basic functionality of the list_head type2683 and associated macros.2684 2685 KUnit tests run during boot and output the results to the debug log2686 in TAP format (https://testanything.org/). Only useful for kernel devs2687 running the KUnit test harness, and not intended for inclusion into a2688 production build.2689 2690 For more information on KUnit and unit tests in general please refer2691 to the KUnit documentation in Documentation/dev-tools/kunit/.2692 2693 If unsure, say N.2694 2695config HASHTABLE_KUNIT_TEST2696 tristate "KUnit Test for Kernel Hashtable structures" if !KUNIT_ALL_TESTS2697 depends on KUNIT2698 default KUNIT_ALL_TESTS2699 help2700 This builds the hashtable KUnit test suite.2701 It tests the basic functionality of the API defined in2702 include/linux/hashtable.h. For more information on KUnit and2703 unit tests in general please refer to the KUnit documentation2704 in Documentation/dev-tools/kunit/.2705 2706 If unsure, say N.2707 2708config LINEAR_RANGES_TEST2709 tristate "KUnit test for linear_ranges"2710 depends on KUNIT2711 select LINEAR_RANGES2712 help2713 This builds the linear_ranges unit test, which runs on boot.2714 Tests the linear_ranges logic correctness.2715 For more information on KUnit and unit tests in general please refer2716 to the KUnit documentation in Documentation/dev-tools/kunit/.2717 2718 If unsure, say N.2719 2720config CMDLINE_KUNIT_TEST2721 tristate "KUnit test for cmdline API" if !KUNIT_ALL_TESTS2722 depends on KUNIT2723 default KUNIT_ALL_TESTS2724 help2725 This builds the cmdline API unit test.2726 Tests the logic of API provided by cmdline.c.2727 For more information on KUnit and unit tests in general please refer2728 to the KUnit documentation in Documentation/dev-tools/kunit/.2729 2730 If unsure, say N.2731 2732config BITS_TEST2733 tristate "KUnit test for bits.h" if !KUNIT_ALL_TESTS2734 depends on KUNIT2735 default KUNIT_ALL_TESTS2736 help2737 This builds the bits unit test.2738 Tests the logic of macros defined in bits.h.2739 For more information on KUnit and unit tests in general please refer2740 to the KUnit documentation in Documentation/dev-tools/kunit/.2741 2742 If unsure, say N.2743 2744config SLUB_KUNIT_TEST2745 tristate "KUnit test for SLUB cache error detection" if !KUNIT_ALL_TESTS2746 depends on SLUB_DEBUG && KUNIT2747 default KUNIT_ALL_TESTS2748 help2749 This builds SLUB allocator unit test.2750 Tests SLUB cache debugging functionality.2751 For more information on KUnit and unit tests in general please refer2752 to the KUnit documentation in Documentation/dev-tools/kunit/.2753 2754 If unsure, say N.2755 2756config RATIONAL_KUNIT_TEST2757 tristate "KUnit test for rational.c" if !KUNIT_ALL_TESTS2758 depends on KUNIT && RATIONAL2759 default KUNIT_ALL_TESTS2760 help2761 This builds the rational math unit test.2762 For more information on KUnit and unit tests in general please refer2763 to the KUnit documentation in Documentation/dev-tools/kunit/.2764 2765 If unsure, say N.2766 2767config MEMCPY_KUNIT_TEST2768 tristate "Test memcpy(), memmove(), and memset() functions at runtime" if !KUNIT_ALL_TESTS2769 depends on KUNIT2770 default KUNIT_ALL_TESTS2771 help2772 Builds unit tests for memcpy(), memmove(), and memset() functions.2773 For more information on KUnit and unit tests in general please refer2774 to the KUnit documentation in Documentation/dev-tools/kunit/.2775 2776 If unsure, say N.2777 2778config IS_SIGNED_TYPE_KUNIT_TEST2779 tristate "Test is_signed_type() macro" if !KUNIT_ALL_TESTS2780 depends on KUNIT2781 default KUNIT_ALL_TESTS2782 help2783 Builds unit tests for the is_signed_type() macro.2784 2785 For more information on KUnit and unit tests in general please refer2786 to the KUnit documentation in Documentation/dev-tools/kunit/.2787 2788 If unsure, say N.2789 2790config OVERFLOW_KUNIT_TEST2791 tristate "Test check_*_overflow() functions at runtime" if !KUNIT_ALL_TESTS2792 depends on KUNIT2793 default KUNIT_ALL_TESTS2794 help2795 Builds unit tests for the check_*_overflow(), size_*(), allocation, and2796 related functions.2797 2798 For more information on KUnit and unit tests in general please refer2799 to the KUnit documentation in Documentation/dev-tools/kunit/.2800 2801 If unsure, say N.2802 2803config STACKINIT_KUNIT_TEST2804 tristate "Test level of stack variable initialization" if !KUNIT_ALL_TESTS2805 depends on KUNIT2806 default KUNIT_ALL_TESTS2807 help2808 Test if the kernel is zero-initializing stack variables and2809 padding. Coverage is controlled by compiler flags,2810 CONFIG_INIT_STACK_ALL_PATTERN, CONFIG_INIT_STACK_ALL_ZERO,2811 CONFIG_GCC_PLUGIN_STRUCTLEAK, CONFIG_GCC_PLUGIN_STRUCTLEAK_BYREF,2812 or CONFIG_GCC_PLUGIN_STRUCTLEAK_BYREF_ALL.2813 2814config FORTIFY_KUNIT_TEST2815 tristate "Test fortified str*() and mem*() function internals at runtime" if !KUNIT_ALL_TESTS2816 depends on KUNIT2817 default KUNIT_ALL_TESTS2818 help2819 Builds unit tests for checking internals of FORTIFY_SOURCE as used2820 by the str*() and mem*() family of functions. For testing runtime2821 traps of FORTIFY_SOURCE, see LKDTM's "FORTIFY_*" tests.2822 2823config HW_BREAKPOINT_KUNIT_TEST2824 bool "Test hw_breakpoint constraints accounting" if !KUNIT_ALL_TESTS2825 depends on HAVE_HW_BREAKPOINT2826 depends on KUNIT=y2827 default KUNIT_ALL_TESTS2828 help2829 Tests for hw_breakpoint constraints accounting.2830 2831 If unsure, say N.2832 2833config SIPHASH_KUNIT_TEST2834 tristate "Perform selftest on siphash functions" if !KUNIT_ALL_TESTS2835 depends on KUNIT2836 default KUNIT_ALL_TESTS2837 help2838 Enable this option to test the kernel's siphash (<linux/siphash.h>) hash2839 functions on boot (or module load).2840 2841 This is intended to help people writing architecture-specific2842 optimized versions. If unsure, say N.2843 2844config USERCOPY_KUNIT_TEST2845 tristate "KUnit Test for user/kernel boundary protections"2846 depends on KUNIT2847 default KUNIT_ALL_TESTS2848 help2849 This builds the "usercopy_kunit" module that runs sanity checks2850 on the copy_to/from_user infrastructure, making sure basic2851 user/kernel boundary testing is working.2852 2853config TEST_UDELAY2854 tristate "udelay test driver"2855 help2856 This builds the "udelay_test" module that helps to make sure2857 that udelay() is working properly.2858 2859 If unsure, say N.2860 2861config TEST_STATIC_KEYS2862 tristate "Test static keys"2863 depends on m2864 help2865 Test the static key interfaces.2866 2867 If unsure, say N.2868 2869config TEST_DYNAMIC_DEBUG2870 tristate "Test DYNAMIC_DEBUG"2871 depends on DYNAMIC_DEBUG2872 help2873 This module registers a tracer callback to count enabled2874 pr_debugs in a 'do_debugging' function, then alters their2875 enablements, calls the function, and compares counts.2876 2877 If unsure, say N.2878 2879config TEST_KMOD2880 tristate "kmod stress tester"2881 depends on m2882 depends on NETDEVICES && NET_CORE && INET # for TUN2883 depends on BLOCK2884 depends on PAGE_SIZE_LESS_THAN_256KB # for BTRFS2885 select TEST_LKM2886 select XFS_FS2887 select TUN2888 select BTRFS_FS2889 help2890 Test the kernel's module loading mechanism: kmod. kmod implements2891 support to load modules using the Linux kernel's usermode helper.2892 This test provides a series of tests against kmod.2893 2894 Although technically you can either build test_kmod as a module or2895 into the kernel we disallow building it into the kernel since2896 it stress tests request_module() and this will very likely cause2897 some issues by taking over precious threads available from other2898 module load requests, ultimately this could be fatal.2899 2900 To run tests run:2901 2902 tools/testing/selftests/kmod/kmod.sh --help2903 2904 If unsure, say N.2905 2906config TEST_DEBUG_VIRTUAL2907 tristate "Test CONFIG_DEBUG_VIRTUAL feature"2908 depends on DEBUG_VIRTUAL2909 help2910 Test the kernel's ability to detect incorrect calls to2911 virt_to_phys() done against the non-linear part of the2912 kernel's virtual address map.2913 2914 If unsure, say N.2915 2916config TEST_MEMCAT_P2917 tristate "Test memcat_p() helper function"2918 help2919 Test the memcat_p() helper for correctly merging two2920 pointer arrays together.2921 2922 If unsure, say N.2923 2924config TEST_OBJAGG2925 tristate "Perform selftest on object aggreration manager"2926 default n2927 depends on OBJAGG2928 help2929 Enable this option to test object aggregation manager on boot2930 (or module load).2931 2932config TEST_MEMINIT2933 tristate "Test heap/page initialization"2934 help2935 Test if the kernel is zero-initializing heap and page allocations.2936 This can be useful to test init_on_alloc and init_on_free features.2937 2938 If unsure, say N.2939 2940config TEST_HMM2941 tristate "Test HMM (Heterogeneous Memory Management)"2942 depends on TRANSPARENT_HUGEPAGE2943 depends on DEVICE_PRIVATE2944 select HMM_MIRROR2945 select MMU_NOTIFIER2946 help2947 This is a pseudo device driver solely for testing HMM.2948 Say M here if you want to build the HMM test module.2949 Doing so will allow you to run tools/testing/selftest/vm/hmm-tests.2950 2951 If unsure, say N.2952 2953config TEST_FREE_PAGES2954 tristate "Test freeing pages"2955 help2956 Test that a memory leak does not occur due to a race between2957 freeing a block of pages and a speculative page reference.2958 Loading this module is safe if your kernel has the bug fixed.2959 If the bug is not fixed, it will leak gigabytes of memory and2960 probably OOM your system.2961 2962config TEST_FPU2963 tristate "Test floating point operations in kernel space"2964 depends on ARCH_HAS_KERNEL_FPU_SUPPORT && !KCOV_INSTRUMENT_ALL2965 help2966 Enable this option to add /sys/kernel/debug/selftest_helpers/test_fpu2967 which will trigger a sequence of floating point operations. This is used2968 for self-testing floating point control register setting in2969 kernel_fpu_begin().2970 2971 If unsure, say N.2972 2973config TEST_CLOCKSOURCE_WATCHDOG2974 tristate "Test clocksource watchdog in kernel space"2975 depends on CLOCKSOURCE_WATCHDOG2976 help2977 Enable this option to create a kernel module that will trigger2978 a test of the clocksource watchdog. This module may be loaded2979 via modprobe or insmod in which case it will run upon being2980 loaded, or it may be built in, in which case it will run2981 shortly after boot.2982 2983 If unsure, say N.2984 2985config TEST_OBJPOOL2986 tristate "Test module for correctness and stress of objpool"2987 default n2988 depends on m && DEBUG_KERNEL2989 help2990 This builds the "test_objpool" module that should be used for2991 correctness verification and concurrent testings of objects2992 allocation and reclamation.2993 2994 If unsure, say N.2995 2996endif # RUNTIME_TESTING_MENU2997 2998config ARCH_USE_MEMTEST2999 bool3000 help3001 An architecture should select this when it uses early_memtest()3002 during boot process.3003 3004config MEMTEST3005 bool "Memtest"3006 depends on ARCH_USE_MEMTEST3007 help3008 This option adds a kernel parameter 'memtest', which allows memtest3009 to be set and executed.3010 memtest=0, mean disabled; -- default3011 memtest=1, mean do 1 test pattern;3012 ...3013 memtest=17, mean do 17 test patterns.3014 If you are unsure how to answer this question, answer N.3015 3016 3017 3018config HYPERV_TESTING3019 bool "Microsoft Hyper-V driver testing"3020 default n3021 depends on HYPERV && DEBUG_FS3022 help3023 Select this option to enable Hyper-V vmbus testing.3024 3025endmenu # "Kernel Testing and Coverage"3026 3027menu "Rust hacking"3028 3029config RUST_DEBUG_ASSERTIONS3030 bool "Debug assertions"3031 depends on RUST3032 help3033 Enables rustc's `-Cdebug-assertions` codegen option.3034 3035 This flag lets you turn `cfg(debug_assertions)` conditional3036 compilation on or off. This can be used to enable extra debugging3037 code in development but not in production. For example, it controls3038 the behavior of the standard library's `debug_assert!` macro.3039 3040 Note that this will apply to all Rust code, including `core`.3041 3042 If unsure, say N.3043 3044config RUST_OVERFLOW_CHECKS3045 bool "Overflow checks"3046 default y3047 depends on RUST3048 help3049 Enables rustc's `-Coverflow-checks` codegen option.3050 3051 This flag allows you to control the behavior of runtime integer3052 overflow. When overflow-checks are enabled, a Rust panic will occur3053 on overflow.3054 3055 Note that this will apply to all Rust code, including `core`.3056 3057 If unsure, say Y.3058 3059config RUST_BUILD_ASSERT_ALLOW3060 bool "Allow unoptimized build-time assertions"3061 depends on RUST3062 help3063 Controls how `build_error!` and `build_assert!` are handled during the build.3064 3065 If calls to them exist in the binary, it may indicate a violated invariant3066 or that the optimizer failed to verify the invariant during compilation.3067 3068 This should not happen, thus by default the build is aborted. However,3069 as an escape hatch, you can choose Y here to ignore them during build3070 and let the check be carried at runtime (with `panic!` being called if3071 the check fails).3072 3073 If unsure, say N.3074 3075config RUST_KERNEL_DOCTESTS3076 bool "Doctests for the `kernel` crate" if !KUNIT_ALL_TESTS3077 depends on RUST && KUNIT=y3078 default KUNIT_ALL_TESTS3079 help3080 This builds the documentation tests of the `kernel` crate3081 as KUnit tests.3082 3083 For more information on KUnit and unit tests in general,3084 please refer to the KUnit documentation in Documentation/dev-tools/kunit/.3085 3086 If unsure, say N.3087 3088endmenu # "Rust"3089 3090endmenu # Kernel hacking3091 3092config INT_POW_TEST3093 tristate "Integer exponentiation (int_pow) test" if !KUNIT_ALL_TESTS3094 depends on KUNIT3095 default KUNIT_ALL_TESTS3096 help3097 This option enables the KUnit test suite for the int_pow function,3098 which performs integer exponentiation. The test suite is designed to3099 verify that the implementation of int_pow correctly computes the power3100 of a given base raised to a given exponent.3101 3102 Enabling this option will include tests that check various scenarios3103 and edge cases to ensure the accuracy and reliability of the exponentiation3104 function.3105 3106 If unsure, say N3107