1# SPDX-License-Identifier: GPL-2.0-only 2menu "Kernel hacking" 3 4menu "printk and dmesg options" 5 6config PRINTK_TIME 7 bool "Show timing information on printks" 8 depends on PRINTK 9 help 10 Selecting this option causes time stamps of the printk() 11 messages to be added to the output of the syslog() system 12 call and at the console. 13 14 The timestamp is always recorded internally, and exported 15 to /dev/kmsg. This flag just specifies if the timestamp should 16 be included, not that the timestamp is recorded. 17 18 The behavior is also controlled by the kernel command line 19 parameter printk.time=1. See Documentation/admin-guide/kernel-parameters.rst 20 21config PRINTK_CALLER 22 bool "Show caller information on printks" 23 depends on PRINTK 24 help 25 Selecting this option causes printk() to add a caller "thread id" (if 26 in task context) or a caller "processor id" (if not in task context) 27 to every message. 28 29 This option is intended for environments where multiple threads 30 concurrently call printk() for many times, for it is difficult to 31 interpret without knowing where these lines (or sometimes individual 32 line which was divided into multiple lines due to race) came from. 33 34 Since toggling after boot makes the code racy, currently there is 35 no option to enable/disable at the kernel command line parameter or 36 sysfs interface. 37 38config PRINTK_EXECUTION_CTX 39 bool 40 depends on PRINTK 41 help 42 This option extends struct printk_info to include extra execution 43 context in printk, such as task name and CPU number from where the 44 message originated. This is useful for correlating printk messages 45 with specific execution contexts. 46 47 This is automatically enabled when a console driver that supports 48 execution context is selected. 49 50config STACKTRACE_BUILD_ID 51 bool "Show build ID information in stacktraces" 52 depends on PRINTK 53 help 54 Selecting this option adds build ID information for symbols in 55 stacktraces printed with the printk format '%p[SR]b'. 56 57 This option is intended for distros where debuginfo is not easily 58 accessible but can be downloaded given the build ID of the vmlinux or 59 kernel module where the function is located. 60 61config CONSOLE_LOGLEVEL_DEFAULT 62 int "Default console loglevel (1-15)" 63 range 1 15 64 default "7" 65 help 66 Default loglevel to determine what will be printed on the console. 67 68 Setting a default here is equivalent to passing in loglevel=<x> in 69 the kernel bootargs. loglevel=<x> continues to override whatever 70 value is specified here as well. 71 72 Note: This does not affect the log level of un-prefixed printk() 73 usage in the kernel. That is controlled by the MESSAGE_LOGLEVEL_DEFAULT 74 option. 75 76config CONSOLE_LOGLEVEL_QUIET 77 int "quiet console loglevel (1-15)" 78 range 1 15 79 default "4" 80 help 81 loglevel to use when "quiet" is passed on the kernel commandline. 82 83 When "quiet" is passed on the kernel commandline this loglevel 84 will be used as the loglevel. IOW passing "quiet" will be the 85 equivalent of passing "loglevel=<CONSOLE_LOGLEVEL_QUIET>" 86 87config MESSAGE_LOGLEVEL_DEFAULT 88 int "Default message log level (1-7)" 89 range 1 7 90 default "4" 91 help 92 Default log level for printk statements with no specified priority. 93 94 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks 95 that are auditing their logs closely may want to set it to a lower 96 priority. 97 98 Note: This does not affect what message level gets printed on the console 99 by default. To change that, use loglevel=<x> in the kernel bootargs, 100 or pick a different CONSOLE_LOGLEVEL_DEFAULT configuration value. 101 102config BOOT_PRINTK_DELAY 103 bool "Delay each boot printk message by N milliseconds" 104 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY 105 help 106 This build option allows you to read kernel boot messages 107 by inserting a short delay after each one. The delay is 108 specified in milliseconds on the kernel command line, 109 using "boot_delay=N". 110 111 It is likely that you would also need to use "lpj=M" to preset 112 the "loops per jiffy" value. 113 See a previous boot log for the "lpj" value to use for your 114 system, and then set "lpj=M" before setting "boot_delay=N". 115 NOTE: Using this option may adversely affect SMP systems. 116 I.e., processors other than the first one may not boot up. 117 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect 118 what it believes to be lockup conditions. 119 120config DYNAMIC_DEBUG 121 bool "Enable dynamic printk() support" 122 default n 123 depends on PRINTK 124 depends on (DEBUG_FS || PROC_FS) 125 select DYNAMIC_DEBUG_CORE 126 help 127 128 Compiles debug level messages into the kernel, which would not 129 otherwise be available at runtime. These messages can then be 130 enabled/disabled based on various levels of scope - per source file, 131 function, module, format string, and line number. This mechanism 132 implicitly compiles in all pr_debug() and dev_dbg() calls, which 133 enlarges the kernel text size by about 2%. 134 135 If a source file is compiled with DEBUG flag set, any 136 pr_debug() calls in it are enabled by default, but can be 137 disabled at runtime as below. Note that DEBUG flag is 138 turned on by many CONFIG_*DEBUG* options. 139 140 Usage: 141 142 Dynamic debugging is controlled via the 'dynamic_debug/control' file, 143 which is contained in the 'debugfs' filesystem or procfs. 144 Thus, the debugfs or procfs filesystem must first be mounted before 145 making use of this feature. 146 We refer the control file as: <debugfs>/dynamic_debug/control. This 147 file contains a list of the debug statements that can be enabled. The 148 format for each line of the file is: 149 150 filename:lineno [module]function flags format 151 152 filename : source file of the debug statement 153 lineno : line number of the debug statement 154 module : module that contains the debug statement 155 function : function that contains the debug statement 156 flags : '=p' means the line is turned 'on' for printing 157 format : the format used for the debug statement 158 159 From a live system: 160 161 nullarbor:~ # cat <debugfs>/dynamic_debug/control 162 # filename:lineno [module]function flags format 163 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012" 164 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012" 165 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012" 166 167 Example usage: 168 169 // enable the message at line 1603 of file svcsock.c 170 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' > 171 <debugfs>/dynamic_debug/control 172 173 // enable all the messages in file svcsock.c 174 nullarbor:~ # echo -n 'file svcsock.c +p' > 175 <debugfs>/dynamic_debug/control 176 177 // enable all the messages in the NFS server module 178 nullarbor:~ # echo -n 'module nfsd +p' > 179 <debugfs>/dynamic_debug/control 180 181 // enable all 12 messages in the function svc_process() 182 nullarbor:~ # echo -n 'func svc_process +p' > 183 <debugfs>/dynamic_debug/control 184 185 // disable all 12 messages in the function svc_process() 186 nullarbor:~ # echo -n 'func svc_process -p' > 187 <debugfs>/dynamic_debug/control 188 189 See Documentation/admin-guide/dynamic-debug-howto.rst for additional 190 information. 191 192config DYNAMIC_DEBUG_CORE 193 bool "Enable core function of dynamic debug support" 194 depends on PRINTK 195 depends on (DEBUG_FS || PROC_FS) 196 help 197 Enable core functional support of dynamic debug. It is useful 198 when you want to tie dynamic debug to your kernel modules with 199 DYNAMIC_DEBUG_MODULE defined for each of them, especially for 200 the case of embedded system where the kernel image size is 201 sensitive for people. 202 203config SYMBOLIC_ERRNAME 204 bool "Support symbolic error names in printf" 205 default y if PRINTK 206 help 207 If you say Y here, the kernel's printf implementation will 208 be able to print symbolic error names such as ENOSPC instead 209 of the number 28. It makes the kernel image slightly larger 210 (about 3KB), but can make the kernel logs easier to read. 211 212config DEBUG_BUGVERBOSE 213 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT 214 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE) 215 default y 216 help 217 Say Y here to make BUG() panics output the file name and line number 218 of the BUG call as well as the EIP and oops trace. This aids 219 debugging but costs about 70-100K of memory. 220 221config DEBUG_BUGVERBOSE_DETAILED 222 bool "Verbose WARN_ON_ONCE() reporting (adds 100K)" if DEBUG_BUGVERBOSE 223 help 224 Say Y here to make WARN_ON_ONCE() output the condition string of the 225 warning, in addition to the file name and line number. 226 This helps debugging, but costs about 100K of memory. 227 228 Say N if unsure. 229 230 231endmenu # "printk and dmesg options" 232 233config DEBUG_KERNEL 234 bool "Kernel debugging" 235 help 236 Say Y here if you are developing drivers or trying to debug and 237 identify kernel problems. 238 239config DEBUG_MISC 240 bool "Miscellaneous debug code" 241 default DEBUG_KERNEL 242 depends on DEBUG_KERNEL 243 help 244 Say Y here if you need to enable miscellaneous debug code that should 245 be under a more specific debug option but isn't. 246 247menu "Compile-time checks and compiler options" 248 249config DEBUG_INFO 250 bool 251 help 252 A kernel debug info option other than "None" has been selected 253 in the "Debug information" choice below, indicating that debug 254 information will be generated for build targets. 255 256# Clang generates .uleb128 with label differences for DWARF v5, a feature that 257# older binutils ports do not support when utilizing RISC-V style linker 258# relaxation: https://sourceware.org/bugzilla/show_bug.cgi?id=27215 259config AS_HAS_NON_CONST_ULEB128 260 def_bool $(as-instr,.uleb128 .Lexpr_end4 - .Lexpr_start3\n.Lexpr_start3:\n.Lexpr_end4:) 261 262choice 263 prompt "Debug information" 264 depends on DEBUG_KERNEL 265 help 266 Selecting something other than "None" results in a kernel image 267 that will include debugging info resulting in a larger kernel image. 268 This adds debug symbols to the kernel and modules (gcc -g), and 269 is needed if you intend to use kernel crashdump or binary object 270 tools like crash, kgdb, LKCD, gdb, etc on the kernel. 271 272 Choose which version of DWARF debug info to emit. If unsure, 273 select "Toolchain default". 274 275config DEBUG_INFO_NONE 276 bool "Disable debug information" 277 help 278 Do not build the kernel with debugging information, which will 279 result in a faster and smaller build. 280 281config DEBUG_INFO_DWARF_TOOLCHAIN_DEFAULT 282 bool "Rely on the toolchain's implicit default DWARF version" 283 select DEBUG_INFO 284 depends on !CC_IS_CLANG || AS_IS_LLVM || (AS_IS_GNU && AS_VERSION >= 23502 && AS_HAS_NON_CONST_ULEB128) 285 help 286 The implicit default version of DWARF debug info produced by a 287 toolchain changes over time. 288 289 This can break consumers of the debug info that haven't upgraded to 290 support newer revisions, and prevent testing newer versions, but 291 those should be less common scenarios. 292 293config DEBUG_INFO_DWARF4 294 bool "Generate DWARF Version 4 debuginfo" 295 select DEBUG_INFO 296 depends on !CC_IS_CLANG || AS_IS_LLVM || (AS_IS_GNU && AS_VERSION >= 23502) 297 help 298 Generate DWARF v4 debug info. This requires gcc 4.5+, binutils 2.35.2 299 if using clang without clang's integrated assembler, and gdb 7.0+. 300 301 If you have consumers of DWARF debug info that are not ready for 302 newer revisions of DWARF, you may wish to choose this or have your 303 config select this. 304 305config DEBUG_INFO_DWARF5 306 bool "Generate DWARF Version 5 debuginfo" 307 select DEBUG_INFO 308 depends on !ARCH_HAS_BROKEN_DWARF5 309 depends on !CC_IS_CLANG || AS_IS_LLVM || (AS_IS_GNU && AS_VERSION >= 23502 && AS_HAS_NON_CONST_ULEB128) 310 help 311 Generate DWARF v5 debug info. Requires binutils 2.35.2, gcc 5.0+ (gcc 312 5.0+ accepts the -gdwarf-5 flag but only had partial support for some 313 draft features until 7.0), and gdb 8.0+. 314 315 Changes to the structure of debug info in Version 5 allow for around 316 15-18% savings in resulting image and debug info section sizes as 317 compared to DWARF Version 4. DWARF Version 5 standardizes previous 318 extensions such as accelerators for symbol indexing and the format 319 for fission (.dwo/.dwp) files. Users may not want to select this 320 config if they rely on tooling that has not yet been updated to 321 support DWARF Version 5. 322 323endchoice # "Debug information" 324 325if DEBUG_INFO 326 327config DEBUG_INFO_REDUCED 328 bool "Reduce debugging information" 329 help 330 If you say Y here gcc is instructed to generate less debugging 331 information for structure types. This means that tools that 332 need full debugging information (like kgdb or systemtap) won't 333 be happy. But if you merely need debugging information to 334 resolve line numbers there is no loss. Advantage is that 335 build directory object sizes shrink dramatically over a full 336 DEBUG_INFO build and compile times are reduced too. 337 Only works with newer gcc versions. 338 339choice 340 prompt "Compressed Debug information" 341 help 342 Compress the resulting debug info. Results in smaller debug info sections, 343 but requires that consumers are able to decompress the results. 344 345 If unsure, choose DEBUG_INFO_COMPRESSED_NONE. 346 347config DEBUG_INFO_COMPRESSED_NONE 348 bool "Don't compress debug information" 349 help 350 Don't compress debug info sections. 351 352config DEBUG_INFO_COMPRESSED_ZLIB 353 bool "Compress debugging information with zlib" 354 depends on $(cc-option,-gz=zlib) 355 depends on $(ld-option,--compress-debug-sections=zlib) 356 help 357 Compress the debug information using zlib. 358 359 Users of dpkg-deb via debian/rules may find an increase in 360 size of their debug .deb packages with this config set, due to the 361 debug info being compressed with zlib, then the object files being 362 recompressed with a different compression scheme. But this is still 363 preferable to setting KDEB_COMPRESS or DPKG_DEB_COMPRESSOR_TYPE to 364 "none" which would be even larger. 365 366config DEBUG_INFO_COMPRESSED_ZSTD 367 bool "Compress debugging information with zstd" 368 depends on $(cc-option,-gz=zstd) 369 depends on $(ld-option,--compress-debug-sections=zstd) 370 help 371 Compress the debug information using zstd. This may provide better 372 compression than zlib, for about the same time costs, but requires newer 373 toolchain support. Requires GCC 13.0+ or Clang 16.0+, binutils 2.40+, and 374 zstd. 375 376endchoice # "Compressed Debug information" 377 378config DEBUG_INFO_SPLIT 379 bool "Produce split debuginfo in .dwo files" 380 depends on $(cc-option,-gsplit-dwarf) 381 # RISC-V linker relaxation + -gsplit-dwarf has issues with LLVM and GCC 382 # prior to 12.x: 383 # https://github.com/llvm/llvm-project/issues/56642 384 # https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99090 385 depends on !RISCV || GCC_VERSION >= 120000 386 help 387 Generate debug info into separate .dwo files. This significantly 388 reduces the build directory size for builds with DEBUG_INFO, 389 because it stores the information only once on disk in .dwo 390 files instead of multiple times in object files and executables. 391 In addition the debug information is also compressed. 392 393 Requires recent gcc (4.7+) and recent gdb/binutils. 394 Any tool that packages or reads debug information would need 395 to know about the .dwo files and include them. 396 Incompatible with older versions of ccache. 397 398config DEBUG_INFO_BTF 399 bool "Generate BTF type information" 400 depends on !DEBUG_INFO_SPLIT && !DEBUG_INFO_REDUCED 401 depends on !GCC_PLUGIN_RANDSTRUCT || COMPILE_TEST 402 depends on BPF_SYSCALL 403 depends on PAHOLE_VERSION >= 122 404 # pahole uses elfutils, which does not have support for Hexagon relocations 405 depends on !HEXAGON 406 help 407 Generate deduplicated BTF type information from DWARF debug info. 408 Turning this on requires pahole v1.22 or later, which will convert 409 DWARF type info into equivalent deduplicated BTF type info. 410 411config PAHOLE_HAS_BTF_TAG 412 def_bool PAHOLE_VERSION >= 123 413 depends on CC_IS_CLANG 414 help 415 Decide whether pahole emits btf_tag attributes (btf_type_tag and 416 btf_decl_tag) or not. Currently only clang compiler implements 417 these attributes, so make the config depend on CC_IS_CLANG. 418 419config PAHOLE_HAS_LANG_EXCLUDE 420 def_bool PAHOLE_VERSION >= 124 421 help 422 Support for the --lang_exclude flag which makes pahole exclude 423 compilation units from the supplied language. Used in Kbuild to 424 omit Rust CUs which are not supported in version 1.24 of pahole, 425 otherwise it would emit malformed kernel and module binaries when 426 using DEBUG_INFO_BTF_MODULES. 427 428config DEBUG_INFO_BTF_MODULES 429 bool "Generate BTF type information for kernel modules" 430 default y 431 depends on DEBUG_INFO_BTF && MODULES 432 help 433 Generate compact split BTF type information for kernel modules. 434 435config MODULE_ALLOW_BTF_MISMATCH 436 bool "Allow loading modules with non-matching BTF type info" 437 depends on DEBUG_INFO_BTF_MODULES 438 help 439 For modules whose split BTF does not match vmlinux, load without 440 BTF rather than refusing to load. The default behavior with 441 module BTF enabled is to reject modules with such mismatches; 442 this option will still load module BTF where possible but ignore 443 it when a mismatch is found. 444 445config GDB_SCRIPTS 446 bool "Provide GDB scripts for kernel debugging" 447 help 448 This creates the required links to GDB helper scripts in the 449 build directory. If you load vmlinux into gdb, the helper 450 scripts will be automatically imported by gdb as well, and 451 additional functions are available to analyze a Linux kernel 452 instance. See Documentation/process/debugging/gdb-kernel-debugging.rst 453 for further details. 454 455endif # DEBUG_INFO 456 457config FRAME_WARN 458 int "Warn for stack frames larger than" 459 range 0 8192 460 default 0 if KMSAN 461 default 2048 if GCC_PLUGIN_LATENT_ENTROPY 462 default 2048 if PARISC 463 default 1536 if (!64BIT && XTENSA) 464 default 1280 if !64BIT 465 default 2048 if 64BIT 466 help 467 Tell the compiler to warn at build time for stack frames larger than this. 468 Setting this too low will cause a lot of warnings. 469 Setting it to 0 disables the warning. 470 471config STRIP_ASM_SYMS 472 bool "Strip assembler-generated symbols during link" 473 default n 474 help 475 Strip internal assembler-generated symbols during a link (symbols 476 that look like '.Lxxx') so they don't pollute the output of 477 get_wchan() and suchlike. 478 479config READABLE_ASM 480 bool "Generate readable assembler code" 481 depends on DEBUG_KERNEL 482 depends on CC_IS_GCC 483 help 484 Disable some compiler optimizations that tend to generate human unreadable 485 assembler output. This may make the kernel slightly slower, but it helps 486 to keep kernel developers who have to stare a lot at assembler listings 487 sane. 488 489config HEADERS_INSTALL 490 bool "Install uapi headers to usr/include" 491 help 492 This option will install uapi headers (headers exported to user-space) 493 into the usr/include directory for use during the kernel build. 494 This is unneeded for building the kernel itself, but needed for some 495 user-space program samples. It is also needed by some features such 496 as uapi header sanity checks. 497 498config DEBUG_SECTION_MISMATCH 499 bool "Enable full Section mismatch analysis" 500 depends on CC_IS_GCC 501 help 502 The section mismatch analysis checks if there are illegal references 503 from one section to another. During linktime or runtime, some 504 sections are dropped; any use of code/data previously in these 505 sections would most likely result in an oops. 506 507 In the code, functions and variables are annotated with __init, 508 __initdata, and so on (see the full list in include/linux/init.h). 509 This directs the toolchain to place code/data in specific sections. 510 511 The section mismatch analysis is always performed after a full 512 kernel build, and enabling this option causes the option 513 -fno-inline-functions-called-once to be added to gcc commands. 514 515 However, when inlining a function annotated with __init in 516 a non-init function, we would lose the section information and thus 517 the analysis would not catch the illegal reference. This option 518 tells gcc to inline less (but it does result in a larger kernel). 519 520config SECTION_MISMATCH_WARN_ONLY 521 bool "Make section mismatch errors non-fatal" 522 default y 523 help 524 If you say N here, the build process will fail if there are any 525 section mismatch, instead of just throwing warnings. 526 527 If unsure, say Y. 528 529config DEBUG_FORCE_FUNCTION_ALIGN_64B 530 bool "Force all function address 64B aligned" 531 depends on EXPERT && (X86_64 || ARM64 || PPC32 || PPC64 || ARC || RISCV || S390) 532 select FUNCTION_ALIGNMENT_64B 533 help 534 There are cases that a commit from one domain changes the function 535 address alignment of other domains, and cause magic performance 536 bump (regression or improvement). Enable this option will help to 537 verify if the bump is caused by function alignment changes, while 538 it will slightly increase the kernel size and affect icache usage. 539 540 It is mainly for debug and performance tuning use. 541 542# 543# Select this config option from the architecture Kconfig, if it 544# is preferred to always offer frame pointers as a config 545# option on the architecture (regardless of KERNEL_DEBUG): 546# 547config ARCH_WANT_FRAME_POINTERS 548 bool 549 550config FRAME_POINTER 551 bool "Compile the kernel with frame pointers" 552 depends on DEBUG_KERNEL && (M68K || UML || SUPERH) || ARCH_WANT_FRAME_POINTERS 553 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS 554 help 555 If you say Y here the resulting kernel image will be slightly 556 larger and slower, but it gives very useful debugging information 557 in case of kernel bugs. (precise oopses/stacktraces/warnings) 558 559config OBJTOOL 560 bool 561 562config OBJTOOL_WERROR 563 bool "Upgrade objtool warnings to errors" 564 depends on OBJTOOL && !COMPILE_TEST 565 help 566 Fail the build on objtool warnings. 567 568 Objtool warnings can indicate kernel instability, including boot 569 failures. This option is highly recommended. 570 571 If unsure, say Y. 572 573config STACK_VALIDATION 574 bool "Compile-time stack metadata validation" 575 depends on HAVE_STACK_VALIDATION && UNWINDER_FRAME_POINTER 576 select OBJTOOL 577 default n 578 help 579 Validate frame pointer rules at compile-time. This helps ensure that 580 runtime stack traces are more reliable. 581 582 For more information, see 583 tools/objtool/Documentation/objtool.txt. 584 585config NOINSTR_VALIDATION 586 bool 587 depends on HAVE_NOINSTR_VALIDATION && DEBUG_ENTRY 588 select OBJTOOL 589 default y 590 591config VMLINUX_MAP 592 bool "Generate vmlinux.map file when linking" 593 depends on EXPERT 594 help 595 Selecting this option will pass "-Map=vmlinux.map" to ld 596 when linking vmlinux. That file can be useful for verifying 597 and debugging magic section games, and for seeing which 598 pieces of code get eliminated with 599 CONFIG_LD_DEAD_CODE_DATA_ELIMINATION. 600 601config BUILTIN_MODULE_RANGES 602 bool "Generate address range information for builtin modules" 603 depends on !LTO 604 depends on VMLINUX_MAP 605 help 606 When modules are built into the kernel, there will be no module name 607 associated with its symbols in /proc/kallsyms. Tracers may want to 608 identify symbols by module name and symbol name regardless of whether 609 the module is configured as loadable or not. 610 611 This option generates modules.builtin.ranges in the build tree with 612 offset ranges (per ELF section) for the module(s) they belong to. 613 It also records an anchor symbol to determine the load address of the 614 section. 615 616config DEBUG_FORCE_WEAK_PER_CPU 617 bool "Force weak per-cpu definitions" 618 depends on DEBUG_KERNEL 619 help 620 s390 and alpha require percpu variables in modules to be 621 defined weak to work around addressing range issue which 622 puts the following two restrictions on percpu variable 623 definitions. 624 625 1. percpu symbols must be unique whether static or not 626 2. percpu variables can't be defined inside a function 627 628 To ensure that generic code follows the above rules, this 629 option forces all percpu variables to be defined as weak. 630 631config WARN_CONTEXT_ANALYSIS 632 bool "Compiler context-analysis warnings" 633 depends on CC_IS_CLANG && CLANG_VERSION >= 230000 634 # Branch profiling re-defines "if", which messes with the compiler's 635 # ability to analyze __cond_acquires(..), resulting in false positives. 636 depends on !TRACE_BRANCH_PROFILING 637 default y 638 help 639 Context Analysis is a language extension, which enables statically 640 checking that required contexts are active (or inactive) by acquiring 641 and releasing user-definable "context locks". 642 643 Clang's name of the feature is "Thread Safety Analysis". Requires 644 Clang 23 or later. 645 646 Produces warnings by default. Select CONFIG_WERROR if you wish to 647 turn these warnings into errors. 648 649 For more details, see Documentation/dev-tools/context-analysis.rst. 650 651config WARN_CONTEXT_ANALYSIS_ALL 652 bool "Enable context analysis for all source files" 653 depends on WARN_CONTEXT_ANALYSIS 654 depends on EXPERT && !COMPILE_TEST 655 help 656 Enable tree-wide context analysis. This is likely to produce a 657 large number of false positives - enable at your own risk. 658 659 If unsure, say N. 660 661endmenu # "Compiler options" 662 663menu "Generic Kernel Debugging Instruments" 664 665config MAGIC_SYSRQ 666 bool "Magic SysRq key" 667 depends on !UML 668 help 669 If you say Y here, you will have some control over the system even 670 if the system crashes for example during kernel debugging (e.g., you 671 will be able to flush the buffer cache to disk, reboot the system 672 immediately or dump some status information). This is accomplished 673 by pressing various keys while holding SysRq (Alt+PrintScreen). It 674 also works on a serial console (on PC hardware at least), if you 675 send a BREAK and then within 5 seconds a command keypress. The 676 keys are documented in <file:Documentation/admin-guide/sysrq.rst>. 677 Don't say Y unless you really know what this hack does. 678 679config MAGIC_SYSRQ_DEFAULT_ENABLE 680 hex "Enable magic SysRq key functions by default" 681 depends on MAGIC_SYSRQ 682 default 0x1 683 help 684 Specifies which SysRq key functions are enabled by default. 685 This may be set to 1 or 0 to enable or disable them all, or 686 to a bitmask as described in Documentation/admin-guide/sysrq.rst. 687 688config MAGIC_SYSRQ_SERIAL 689 bool "Enable magic SysRq key over serial" 690 depends on MAGIC_SYSRQ 691 default y 692 help 693 Many embedded boards have a disconnected TTL level serial which can 694 generate some garbage that can lead to spurious false sysrq detects. 695 This option allows you to decide whether you want to enable the 696 magic SysRq key. 697 698config MAGIC_SYSRQ_SERIAL_SEQUENCE 699 string "Char sequence that enables magic SysRq over serial" 700 depends on MAGIC_SYSRQ_SERIAL 701 default "" 702 help 703 Specifies a sequence of characters that can follow BREAK to enable 704 SysRq on a serial console. 705 706 If unsure, leave an empty string and the option will not be enabled. 707 708config DEBUG_FS 709 bool "Debug Filesystem" 710 help 711 debugfs is a virtual file system that kernel developers use to put 712 debugging files into. Enable this option to be able to read and 713 write to these files. 714 715 For detailed documentation on the debugfs API, see 716 Documentation/filesystems/. 717 718 If unsure, say N. 719 720choice 721 prompt "Debugfs default access" 722 depends on DEBUG_FS 723 default DEBUG_FS_ALLOW_ALL 724 help 725 This selects the default access restrictions for debugfs. 726 It can be overridden with kernel command line option 727 debugfs=[on,off]. The restrictions apply for API access 728 and filesystem registration. 729 730config DEBUG_FS_ALLOW_ALL 731 bool "Access normal" 732 help 733 No restrictions apply. Both API and filesystem registration 734 is on. This is the normal default operation. 735 736config DEBUG_FS_ALLOW_NONE 737 bool "No access" 738 help 739 Access is off. Clients get -PERM when trying to create nodes in 740 debugfs tree and debugfs is not registered as a filesystem. 741 Client can then back-off or continue without debugfs access. 742 743endchoice 744 745source "lib/Kconfig.kgdb" 746source "lib/Kconfig.ubsan" 747source "lib/Kconfig.kcsan" 748 749endmenu 750 751menu "Networking Debugging" 752 753source "net/Kconfig.debug" 754 755endmenu # "Networking Debugging" 756 757menu "Memory Debugging" 758 759source "mm/Kconfig.debug" 760 761config DEBUG_OBJECTS 762 bool "Debug object operations" 763 depends on PREEMPT_COUNT || !DEFERRED_STRUCT_PAGE_INIT 764 depends on DEBUG_KERNEL 765 help 766 If you say Y here, additional code will be inserted into the 767 kernel to track the life time of various objects and validate 768 the operations on those objects. 769 770config DEBUG_OBJECTS_SELFTEST 771 bool "Debug objects selftest" 772 depends on DEBUG_OBJECTS 773 help 774 This enables the selftest of the object debug code. 775 776config DEBUG_OBJECTS_FREE 777 bool "Debug objects in freed memory" 778 depends on DEBUG_OBJECTS 779 help 780 This enables checks whether a k/v free operation frees an area 781 which contains an object which has not been deactivated 782 properly. This can make kmalloc/kfree-intensive workloads 783 much slower. 784 785config DEBUG_OBJECTS_TIMERS 786 bool "Debug timer objects" 787 depends on DEBUG_OBJECTS 788 help 789 If you say Y here, additional code will be inserted into the 790 timer routines to track the life time of timer objects and 791 validate the timer operations. 792 793config DEBUG_OBJECTS_WORK 794 bool "Debug work objects" 795 depends on DEBUG_OBJECTS 796 help 797 If you say Y here, additional code will be inserted into the 798 work queue routines to track the life time of work objects and 799 validate the work operations. 800 801config DEBUG_OBJECTS_RCU_HEAD 802 bool "Debug RCU callbacks objects" 803 depends on DEBUG_OBJECTS 804 help 805 Enable this to turn on debugging of RCU list heads (call_rcu() usage). 806 807config DEBUG_OBJECTS_PERCPU_COUNTER 808 bool "Debug percpu counter objects" 809 depends on DEBUG_OBJECTS 810 help 811 If you say Y here, additional code will be inserted into the 812 percpu counter routines to track the life time of percpu counter 813 objects and validate the percpu counter operations. 814 815config DEBUG_OBJECTS_ENABLE_DEFAULT 816 int "debug_objects bootup default value (0-1)" 817 range 0 1 818 default "1" 819 depends on DEBUG_OBJECTS 820 help 821 Debug objects boot parameter default value 822 823config SHRINKER_DEBUG 824 bool "Enable shrinker debugging support" 825 depends on DEBUG_FS 826 help 827 Say Y to enable the shrinker debugfs interface which provides 828 visibility into the kernel memory shrinkers subsystem. 829 Disable it to avoid an extra memory footprint. 830 831config DEBUG_STACK_USAGE 832 bool "Stack utilization instrumentation" 833 depends on DEBUG_KERNEL 834 help 835 Enables the display of the minimum amount of free stack which each 836 task has ever had available in the sysrq-T and sysrq-P debug output. 837 Also emits a message to dmesg when a process exits if that process 838 used more stack space than previously exiting processes. 839 840 This option will slow down process creation somewhat. 841 842config SCHED_STACK_END_CHECK 843 bool "Detect stack corruption on calls to schedule()" 844 depends on DEBUG_KERNEL 845 default n 846 help 847 This option checks for a stack overrun on calls to schedule(). 848 If the stack end location is found to be over written always panic as 849 the content of the corrupted region can no longer be trusted. 850 This is to ensure no erroneous behaviour occurs which could result in 851 data corruption or a sporadic crash at a later stage once the region 852 is examined. The runtime overhead introduced is minimal. 853 854config ARCH_HAS_DEBUG_VM_PGTABLE 855 bool 856 help 857 An architecture should select this when it can successfully 858 build and run DEBUG_VM_PGTABLE. 859 860config DEBUG_VFS 861 bool "Debug VFS" 862 depends on DEBUG_KERNEL 863 help 864 Enable this to turn on extended checks in the VFS layer that may impact 865 performance. 866 867 If unsure, say N. 868 869config DEBUG_VM_IRQSOFF 870 def_bool DEBUG_VM && !PREEMPT_RT 871 872config DEBUG_VM 873 bool "Debug VM" 874 depends on DEBUG_KERNEL 875 help 876 Enable this to turn on extended checks in the virtual-memory system 877 that may impact performance. 878 879 If unsure, say N. 880 881config DEBUG_VM_SHOOT_LAZIES 882 bool "Debug MMU_LAZY_TLB_SHOOTDOWN implementation" 883 depends on DEBUG_VM 884 depends on MMU_LAZY_TLB_SHOOTDOWN 885 help 886 Enable additional IPIs that ensure lazy tlb mm references are removed 887 before the mm is freed. 888 889 If unsure, say N. 890 891config DEBUG_VM_MAPLE_TREE 892 bool "Debug VM maple trees" 893 depends on DEBUG_VM 894 select DEBUG_MAPLE_TREE 895 help 896 Enable VM maple tree debugging information and extra validations. 897 898 If unsure, say N. 899 900config DEBUG_VM_RB 901 bool "Debug VM red-black trees" 902 depends on DEBUG_VM 903 help 904 Enable VM red-black tree debugging information and extra validations. 905 906 If unsure, say N. 907 908config DEBUG_VM_PGFLAGS 909 bool "Debug page-flags operations" 910 depends on DEBUG_VM 911 help 912 Enables extra validation on page flags operations. 913 914 If unsure, say N. 915 916config DEBUG_VM_PGTABLE 917 bool "Debug arch page table for semantics compliance" 918 depends on MMU 919 depends on ARCH_HAS_DEBUG_VM_PGTABLE 920 default y if DEBUG_VM 921 help 922 This option provides a debug method which can be used to test 923 architecture page table helper functions on various platforms in 924 verifying if they comply with expected generic MM semantics. This 925 will help architecture code in making sure that any changes or 926 new additions of these helpers still conform to expected 927 semantics of the generic MM. Platforms will have to opt in for 928 this through ARCH_HAS_DEBUG_VM_PGTABLE. 929 930 If unsure, say N. 931 932config ARCH_HAS_DEBUG_VIRTUAL 933 bool 934 935config DEBUG_VIRTUAL 936 bool "Debug VM translations" 937 depends on DEBUG_KERNEL && ARCH_HAS_DEBUG_VIRTUAL 938 help 939 Enable some costly sanity checks in virtual to page code. This can 940 catch mistakes with virt_to_page() and friends. 941 942 If unsure, say N. 943 944config DEBUG_NOMMU_REGIONS 945 bool "Debug the global anon/private NOMMU mapping region tree" 946 depends on DEBUG_KERNEL && !MMU 947 help 948 This option causes the global tree of anonymous and private mapping 949 regions to be regularly checked for invalid topology. 950 951config DEBUG_MEMORY_INIT 952 bool "Debug memory initialisation" if EXPERT 953 default !EXPERT 954 help 955 Enable this for additional checks during memory initialisation. 956 The sanity checks verify aspects of the VM such as the memory model 957 and other information provided by the architecture. Verbose 958 information will be printed at KERN_DEBUG loglevel depending 959 on the mminit_loglevel= command-line option. 960 961 If unsure, say Y 962 963config MEMORY_NOTIFIER_ERROR_INJECT 964 tristate "Memory hotplug notifier error injection module" 965 depends on MEMORY_HOTPLUG && NOTIFIER_ERROR_INJECTION 966 help 967 This option provides the ability to inject artificial errors to 968 memory hotplug notifier chain callbacks. It is controlled through 969 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory 970 971 If the notifier call chain should be failed with some events 972 notified, write the error code to "actions/<notifier event>/error". 973 974 Example: Inject memory hotplug offline error (-12 == -ENOMEM) 975 976 # cd /sys/kernel/debug/notifier-error-inject/memory 977 # echo -12 > actions/MEM_GOING_OFFLINE/error 978 # echo offline > /sys/devices/system/memory/memoryXXX/state 979 bash: echo: write error: Cannot allocate memory 980 981 To compile this code as a module, choose M here: the module will 982 be called memory-notifier-error-inject. 983 984 If unsure, say N. 985 986config DEBUG_PER_CPU_MAPS 987 bool "Debug access to per_cpu maps" 988 depends on DEBUG_KERNEL 989 depends on SMP 990 help 991 Say Y to verify that the per_cpu map being accessed has 992 been set up. This adds a fair amount of code to kernel memory 993 and decreases performance. 994 995 Say N if unsure. 996 997config DEBUG_KMAP_LOCAL 998 bool "Debug kmap_local temporary mappings" 999 depends on DEBUG_KERNEL && KMAP_LOCAL 1000 help 1001 This option enables additional error checking for the kmap_local 1002 infrastructure. Disable for production use. 1003 1004config ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP 1005 bool 1006 1007config DEBUG_KMAP_LOCAL_FORCE_MAP 1008 bool "Enforce kmap_local temporary mappings" 1009 depends on DEBUG_KERNEL && ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP 1010 select KMAP_LOCAL 1011 select DEBUG_KMAP_LOCAL 1012 help 1013 This option enforces temporary mappings through the kmap_local 1014 mechanism for non-highmem pages and on non-highmem systems. 1015 Disable this for production systems! 1016 1017config DEBUG_HIGHMEM 1018 bool "Highmem debugging" 1019 depends on DEBUG_KERNEL && HIGHMEM 1020 select DEBUG_KMAP_LOCAL_FORCE_MAP if ARCH_SUPPORTS_KMAP_LOCAL_FORCE_MAP 1021 select DEBUG_KMAP_LOCAL 1022 help 1023 This option enables additional error checking for high memory 1024 systems. Disable for production systems. 1025 1026config HAVE_DEBUG_STACKOVERFLOW 1027 bool 1028 1029config DEBUG_STACKOVERFLOW 1030 bool "Check for stack overflows" 1031 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW 1032 help 1033 Say Y here if you want to check for overflows of kernel, IRQ 1034 and exception stacks (if your architecture uses them). This 1035 option will show detailed messages if free stack space drops 1036 below a certain limit. 1037 1038 These kinds of bugs usually occur when call-chains in the 1039 kernel get too deep, especially when interrupts are 1040 involved. 1041 1042 Use this in cases where you see apparently random memory 1043 corruption, especially if it appears in 'struct thread_info' 1044 1045 If in doubt, say "N". 1046 1047config CODE_TAGGING 1048 bool 1049 select KALLSYMS 1050 1051source "lib/Kconfig.kasan" 1052source "lib/Kconfig.kfence" 1053source "lib/Kconfig.kmsan" 1054 1055endmenu # "Memory Debugging" 1056 1057config DEBUG_SHIRQ 1058 bool "Debug shared IRQ handlers" 1059 depends on DEBUG_KERNEL 1060 help 1061 Enable this to generate a spurious interrupt just before a shared 1062 interrupt handler is deregistered (generating one when registering 1063 is currently disabled). Drivers need to handle this correctly. Some 1064 don't and need to be caught. 1065 1066menu "Debug Oops, Lockups and Hangs" 1067 1068config PANIC_ON_OOPS 1069 bool "Panic on Oops" 1070 help 1071 Say Y here to enable the kernel to panic when it oopses. This 1072 has the same effect as setting oops=panic on the kernel command 1073 line. 1074 1075 This feature is useful to ensure that the kernel does not do 1076 anything erroneous after an oops which could result in data 1077 corruption or other issues. 1078 1079 Say N if unsure. 1080 1081config PANIC_TIMEOUT 1082 int "panic timeout" 1083 default 0 1084 help 1085 Set the timeout value (in seconds) until a reboot occurs when 1086 the kernel panics. If n = 0, then we wait forever. A timeout 1087 value n > 0 will wait n seconds before rebooting, while a timeout 1088 value n < 0 will reboot immediately. This setting can be overridden 1089 with the kernel command line option panic=, and from userspace via 1090 /proc/sys/kernel/panic. 1091 1092config LOCKUP_DETECTOR 1093 bool 1094 1095config SOFTLOCKUP_DETECTOR 1096 bool "Detect Soft Lockups" 1097 depends on DEBUG_KERNEL && !S390 1098 select LOCKUP_DETECTOR 1099 help 1100 Say Y here to enable the kernel to act as a watchdog to detect 1101 soft lockups. 1102 1103 Softlockups are bugs that cause the kernel to loop in kernel 1104 mode for more than 20 seconds, without giving other tasks a 1105 chance to run. The current stack trace is displayed upon 1106 detection and the system will stay locked up. 1107 1108config SOFTLOCKUP_DETECTOR_INTR_STORM 1109 bool "Detect Interrupt Storm in Soft Lockups" 1110 depends on SOFTLOCKUP_DETECTOR && IRQ_TIME_ACCOUNTING 1111 select GENERIC_IRQ_STAT_SNAPSHOT 1112 default y if NR_CPUS <= 128 1113 help 1114 Say Y here to enable the kernel to detect interrupt storm 1115 during "soft lockups". 1116 1117 "soft lockups" can be caused by a variety of reasons. If one is 1118 caused by an interrupt storm, then the storming interrupts will not 1119 be on the callstack. To detect this case, it is necessary to report 1120 the CPU stats and the interrupt counts during the "soft lockups". 1121 1122config BOOTPARAM_SOFTLOCKUP_PANIC 1123 int "Panic (Reboot) On Soft Lockups" 1124 depends on SOFTLOCKUP_DETECTOR 1125 default 0 1126 help 1127 Set to a non-zero value N to enable the kernel to panic on "soft 1128 lockups", which are bugs that cause the kernel to loop in kernel 1129 mode for more than (N * 20 seconds) (configurable using the 1130 watchdog_thresh sysctl), without giving other tasks a chance to run. 1131 1132 The panic can be used in combination with panic_timeout, 1133 to cause the system to reboot automatically after a 1134 lockup has been detected. This feature is useful for 1135 high-availability systems that have uptime guarantees and 1136 where a lockup must be resolved ASAP. 1137 1138 Say 0 if unsure. 1139 1140config HAVE_HARDLOCKUP_DETECTOR_BUDDY 1141 bool 1142 depends on SMP 1143 default y 1144 1145# 1146# Global switch whether to build a hardlockup detector at all. It is available 1147# only when the architecture supports at least one implementation. There are 1148# two exceptions. The hardlockup detector is never enabled on: 1149# 1150# s390: it reported many false positives there 1151# 1152# sparc64: has a custom implementation which is not using the common 1153# hardlockup command line options and sysctl interface. 1154# 1155config HARDLOCKUP_DETECTOR 1156 bool "Detect Hard Lockups" 1157 depends on DEBUG_KERNEL && !S390 && !HARDLOCKUP_DETECTOR_SPARC64 1158 depends on HAVE_HARDLOCKUP_DETECTOR_PERF || HAVE_HARDLOCKUP_DETECTOR_BUDDY || HAVE_HARDLOCKUP_DETECTOR_ARCH 1159 imply HARDLOCKUP_DETECTOR_PERF 1160 imply HARDLOCKUP_DETECTOR_BUDDY 1161 imply HARDLOCKUP_DETECTOR_ARCH 1162 select LOCKUP_DETECTOR 1163 1164 help 1165 Say Y here to enable the kernel to act as a watchdog to detect 1166 hard lockups. 1167 1168 Hardlockups are bugs that cause the CPU to loop in kernel mode 1169 for more than 10 seconds, without letting other interrupts have a 1170 chance to run. The current stack trace is displayed upon detection 1171 and the system will stay locked up. 1172 1173# 1174# Note that arch-specific variants are always preferred. 1175# 1176config HARDLOCKUP_DETECTOR_PREFER_BUDDY 1177 bool "Prefer the buddy CPU hardlockup detector" 1178 depends on HARDLOCKUP_DETECTOR 1179 depends on HAVE_HARDLOCKUP_DETECTOR_PERF && HAVE_HARDLOCKUP_DETECTOR_BUDDY 1180 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH 1181 help 1182 Say Y here to prefer the buddy hardlockup detector over the perf one. 1183 1184 With the buddy detector, each CPU uses its softlockup hrtimer 1185 to check that the next CPU is processing hrtimer interrupts by 1186 verifying that a counter is increasing. 1187 1188 This hardlockup detector is useful on systems that don't have 1189 an arch-specific hardlockup detector or if resources needed 1190 for the hardlockup detector are better used for other things. 1191 1192config HARDLOCKUP_DETECTOR_PERF 1193 bool 1194 depends on HARDLOCKUP_DETECTOR 1195 depends on HAVE_HARDLOCKUP_DETECTOR_PERF && !HARDLOCKUP_DETECTOR_PREFER_BUDDY 1196 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH 1197 select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER 1198 1199config HARDLOCKUP_DETECTOR_BUDDY 1200 bool 1201 depends on HARDLOCKUP_DETECTOR 1202 depends on HAVE_HARDLOCKUP_DETECTOR_BUDDY 1203 depends on !HAVE_HARDLOCKUP_DETECTOR_PERF || HARDLOCKUP_DETECTOR_PREFER_BUDDY 1204 depends on !HAVE_HARDLOCKUP_DETECTOR_ARCH 1205 select HARDLOCKUP_DETECTOR_COUNTS_HRTIMER 1206 1207config HARDLOCKUP_DETECTOR_ARCH 1208 bool 1209 depends on HARDLOCKUP_DETECTOR 1210 depends on HAVE_HARDLOCKUP_DETECTOR_ARCH 1211 help 1212 The arch-specific implementation of the hardlockup detector will 1213 be used. 1214 1215# 1216# Both the "perf" and "buddy" hardlockup detectors count hrtimer 1217# interrupts. This config enables functions managing this common code. 1218# 1219config HARDLOCKUP_DETECTOR_COUNTS_HRTIMER 1220 bool 1221 select SOFTLOCKUP_DETECTOR 1222 1223# 1224# Enables a timestamp based low pass filter to compensate for perf based 1225# hard lockup detection which runs too fast due to turbo modes. 1226# 1227config HARDLOCKUP_CHECK_TIMESTAMP 1228 bool 1229 1230config BOOTPARAM_HARDLOCKUP_PANIC 1231 bool "Panic (Reboot) On Hard Lockups" 1232 depends on HARDLOCKUP_DETECTOR 1233 help 1234 Say Y here to enable the kernel to panic on "hard lockups", 1235 which are bugs that cause the kernel to loop in kernel 1236 mode with interrupts disabled for more than 10 seconds (configurable 1237 using the watchdog_thresh sysctl). 1238 1239 Say N if unsure. 1240 1241config DETECT_HUNG_TASK 1242 bool "Detect Hung Tasks" 1243 depends on DEBUG_KERNEL 1244 default SOFTLOCKUP_DETECTOR 1245 help 1246 Say Y here to enable the kernel to detect "hung tasks", 1247 which are bugs that cause the task to be stuck in 1248 uninterruptible "D" state indefinitely. 1249 1250 When a hung task is detected, the kernel will print the 1251 current stack trace (which you should report), but the 1252 task will stay in uninterruptible state. If lockdep is 1253 enabled then all held locks will also be reported. This 1254 feature has negligible overhead. 1255 1256config DEFAULT_HUNG_TASK_TIMEOUT 1257 int "Default timeout for hung task detection (in seconds)" 1258 depends on DETECT_HUNG_TASK 1259 default 120 1260 help 1261 This option controls the default timeout (in seconds) used 1262 to determine when a task has become non-responsive and should 1263 be considered hung. 1264 1265 It can be adjusted at runtime via the kernel.hung_task_timeout_secs 1266 sysctl or by writing a value to 1267 /proc/sys/kernel/hung_task_timeout_secs. 1268 1269 A timeout of 0 disables the check. The default is two minutes. 1270 Keeping the default should be fine in most cases. 1271 1272config BOOTPARAM_HUNG_TASK_PANIC 1273 int "Number of hung tasks to trigger kernel panic" 1274 depends on DETECT_HUNG_TASK 1275 default 0 1276 help 1277 When set to a non-zero value, a kernel panic will be triggered 1278 if the number of hung tasks found during a single scan reaches 1279 this value. 1280 1281 The panic can be used in combination with panic_timeout, 1282 to cause the system to reboot automatically after a 1283 hung task has been detected. This feature is useful for 1284 high-availability systems that have uptime guarantees and 1285 where a hung tasks must be resolved ASAP. 1286 1287 Say 0 if unsure. 1288 1289config DETECT_HUNG_TASK_BLOCKER 1290 bool "Dump Hung Tasks Blocker" 1291 depends on DETECT_HUNG_TASK 1292 depends on !PREEMPT_RT 1293 default y 1294 help 1295 Say Y here to show the blocker task's stacktrace who acquires 1296 the mutex lock which "hung tasks" are waiting. 1297 This will add overhead a bit but shows suspicious tasks and 1298 call trace if it comes from waiting a mutex. 1299 1300config WQ_WATCHDOG 1301 bool "Detect Workqueue Stalls" 1302 depends on DEBUG_KERNEL 1303 help 1304 Say Y here to enable stall detection on workqueues. If a 1305 worker pool doesn't make forward progress on a pending work 1306 item for over a given amount of time, 30s by default, a 1307 warning message is printed along with dump of workqueue 1308 state. This can be configured through kernel parameter 1309 "workqueue.watchdog_thresh" and its sysfs counterpart. 1310 1311config BOOTPARAM_WQ_STALL_PANIC 1312 int "Panic on Nth workqueue stall" 1313 default 0 1314 range 0 100 1315 depends on WQ_WATCHDOG 1316 help 1317 Set the number of workqueue stalls to trigger a kernel panic. 1318 A workqueue stall occurs when a worker pool doesn't make forward 1319 progress on a pending work item for over 30 seconds (configurable 1320 using the workqueue.watchdog_thresh parameter). 1321 1322 If n = 0, the kernel will not panic on stall. If n > 0, the kernel 1323 will panic after n stall warnings. 1324 1325 The panic can be used in combination with panic_timeout, 1326 to cause the system to reboot automatically after a 1327 stall has been detected. This feature is useful for 1328 high-availability systems that have uptime guarantees and 1329 where a stall must be resolved ASAP. 1330 1331 This setting can be overridden at runtime via the 1332 workqueue.panic_on_stall kernel parameter. 1333 1334config WQ_CPU_INTENSIVE_REPORT 1335 bool "Report per-cpu work items which hog CPU for too long" 1336 depends on DEBUG_KERNEL 1337 help 1338 Say Y here to enable reporting of concurrency-managed per-cpu work 1339 items that hog CPUs for longer than 1340 workqueue.cpu_intensive_thresh_us. Workqueue automatically 1341 detects and excludes them from concurrency management to prevent 1342 them from stalling other per-cpu work items. Occassional 1343 triggering may not necessarily indicate a problem. Repeated 1344 triggering likely indicates that the work item should be switched 1345 to use an unbound workqueue. 1346 1347config TEST_LOCKUP 1348 tristate "Test module to generate lockups" 1349 depends on m 1350 help 1351 This builds the "test_lockup" module that helps to make sure 1352 that watchdogs and lockup detectors are working properly. 1353 1354 Depending on module parameters it could emulate soft or hard 1355 lockup, "hung task", or locking arbitrary lock for a long time. 1356 Also it could generate series of lockups with cooling-down periods. 1357 1358 If unsure, say N. 1359 1360endmenu # "Debug lockups and hangs" 1361 1362menu "Scheduler Debugging" 1363 1364config SCHED_INFO 1365 bool 1366 default n 1367 1368config SCHEDSTATS 1369 bool "Collect scheduler statistics" 1370 depends on PROC_FS 1371 select SCHED_INFO 1372 help 1373 If you say Y here, additional code will be inserted into the 1374 scheduler and related routines to collect statistics about 1375 scheduler behavior and provide them in /proc/schedstat. These 1376 stats may be useful for both tuning and debugging the scheduler 1377 If you aren't debugging the scheduler or trying to tune a specific 1378 application, you can say N to avoid the very slight overhead 1379 this adds. 1380 1381endmenu 1382 1383config DEBUG_PREEMPT 1384 bool "Debug preemptible kernel" 1385 depends on DEBUG_KERNEL && PREEMPTION && TRACE_IRQFLAGS_SUPPORT 1386 help 1387 If you say Y here then the kernel will use a debug variant of the 1388 commonly used smp_processor_id() function and will print warnings 1389 if kernel code uses it in a preemption-unsafe way. Also, the kernel 1390 will detect preemption count underflows. 1391 1392 This option has potential to introduce high runtime overhead, 1393 depending on workload as it triggers debugging routines for each 1394 this_cpu operation. It should only be used for debugging purposes. 1395 1396config DEBUG_ATOMIC 1397 bool "Debug atomic variables" 1398 depends on DEBUG_KERNEL 1399 help 1400 If you say Y here then the kernel will add a runtime alignment check 1401 to atomic accesses. Useful for architectures that do not have trap on 1402 mis-aligned access. 1403 1404 This option has potentially significant overhead. 1405 1406config DEBUG_ATOMIC_LARGEST_ALIGN 1407 bool "Check alignment only up to __aligned_largest" 1408 depends on DEBUG_ATOMIC 1409 help 1410 If you say Y here then the check for natural alignment of 1411 atomic accesses will be constrained to the compiler's largest 1412 alignment for scalar types. 1413 1414menu "Lock Debugging (spinlocks, mutexes, etc...)" 1415 1416config LOCK_DEBUGGING_SUPPORT 1417 bool 1418 depends on TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 1419 default y 1420 1421config PROVE_LOCKING 1422 bool "Lock debugging: prove locking correctness" 1423 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT 1424 select LOCKDEP 1425 select DEBUG_SPINLOCK 1426 select DEBUG_MUTEXES if !PREEMPT_RT 1427 select DEBUG_RT_MUTEXES if RT_MUTEXES 1428 select DEBUG_RWSEMS if !PREEMPT_RT 1429 select DEBUG_WW_MUTEX_SLOWPATH 1430 select DEBUG_LOCK_ALLOC 1431 select PREEMPT_COUNT if !ARCH_NO_PREEMPT 1432 select TRACE_IRQFLAGS 1433 default n 1434 help 1435 This feature enables the kernel to prove that all locking 1436 that occurs in the kernel runtime is mathematically 1437 correct: that under no circumstance could an arbitrary (and 1438 not yet triggered) combination of observed locking 1439 sequences (on an arbitrary number of CPUs, running an 1440 arbitrary number of tasks and interrupt contexts) cause a 1441 deadlock. 1442 1443 In short, this feature enables the kernel to report locking 1444 related deadlocks before they actually occur. 1445 1446 The proof does not depend on how hard and complex a 1447 deadlock scenario would be to trigger: how many 1448 participant CPUs, tasks and irq-contexts would be needed 1449 for it to trigger. The proof also does not depend on 1450 timing: if a race and a resulting deadlock is possible 1451 theoretically (no matter how unlikely the race scenario 1452 is), it will be proven so and will immediately be 1453 reported by the kernel (once the event is observed that 1454 makes the deadlock theoretically possible). 1455 1456 If a deadlock is impossible (i.e. the locking rules, as 1457 observed by the kernel, are mathematically correct), the 1458 kernel reports nothing. 1459 1460 NOTE: this feature can also be enabled for rwlocks, mutexes 1461 and rwsems - in which case all dependencies between these 1462 different locking variants are observed and mapped too, and 1463 the proof of observed correctness is also maintained for an 1464 arbitrary combination of these separate locking variants. 1465 1466 For more details, see Documentation/locking/lockdep-design.rst. 1467 1468config PROVE_RAW_LOCK_NESTING 1469 bool "Enable raw_spinlock - spinlock nesting checks" if !ARCH_SUPPORTS_RT 1470 depends on PROVE_LOCKING 1471 default y if ARCH_SUPPORTS_RT 1472 help 1473 Enable the raw_spinlock vs. spinlock nesting checks which ensure 1474 that the lock nesting rules for PREEMPT_RT enabled kernels are 1475 not violated. 1476 1477config LOCK_STAT 1478 bool "Lock usage statistics" 1479 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT 1480 select LOCKDEP 1481 select DEBUG_SPINLOCK 1482 select DEBUG_MUTEXES if !PREEMPT_RT 1483 select DEBUG_RT_MUTEXES if RT_MUTEXES 1484 select DEBUG_LOCK_ALLOC 1485 default n 1486 help 1487 This feature enables tracking lock contention points 1488 1489 For more details, see Documentation/locking/lockstat.rst 1490 1491 This also enables lock events required by "perf lock", 1492 subcommand of perf. 1493 If you want to use "perf lock", you also need to turn on 1494 CONFIG_EVENT_TRACING. 1495 1496 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events. 1497 (CONFIG_LOCKDEP defines "acquire" and "release" events.) 1498 1499config DEBUG_RT_MUTEXES 1500 bool "RT Mutex debugging, deadlock detection" 1501 depends on DEBUG_KERNEL && RT_MUTEXES 1502 help 1503 This allows rt mutex semantics violations and rt mutex related 1504 deadlocks (lockups) to be detected and reported automatically. 1505 1506config DEBUG_SPINLOCK 1507 bool "Spinlock and rw-lock debugging: basic checks" 1508 depends on DEBUG_KERNEL 1509 select UNINLINE_SPIN_UNLOCK 1510 help 1511 Say Y here and build SMP to catch missing spinlock initialization 1512 and certain other kinds of spinlock errors commonly made. This is 1513 best used in conjunction with the NMI watchdog so that spinlock 1514 deadlocks are also debuggable. 1515 1516config DEBUG_MUTEXES 1517 bool "Mutex debugging: basic checks" 1518 depends on DEBUG_KERNEL && !PREEMPT_RT 1519 help 1520 This feature allows mutex semantics violations to be detected and 1521 reported. 1522 1523config DEBUG_WW_MUTEX_SLOWPATH 1524 bool "Wait/wound mutex debugging: Slowpath testing" 1525 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT 1526 select DEBUG_LOCK_ALLOC 1527 select DEBUG_SPINLOCK 1528 select DEBUG_MUTEXES if !PREEMPT_RT 1529 select DEBUG_RT_MUTEXES if PREEMPT_RT 1530 help 1531 This feature enables slowpath testing for w/w mutex users by 1532 injecting additional -EDEADLK wound/backoff cases. Together with 1533 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this 1534 will test all possible w/w mutex interface abuse with the 1535 exception of simply not acquiring all the required locks. 1536 Note that this feature can introduce significant overhead, so 1537 it really should not be enabled in a production or distro kernel, 1538 even a debug kernel. If you are a driver writer, enable it. If 1539 you are a distro, do not. 1540 1541config DEBUG_RWSEMS 1542 bool "RW Semaphore debugging: basic checks" 1543 depends on DEBUG_KERNEL && !PREEMPT_RT 1544 help 1545 This debugging feature allows mismatched rw semaphore locks 1546 and unlocks to be detected and reported. 1547 1548config DEBUG_LOCK_ALLOC 1549 bool "Lock debugging: detect incorrect freeing of live locks" 1550 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT 1551 select DEBUG_SPINLOCK 1552 select DEBUG_MUTEXES if !PREEMPT_RT 1553 select DEBUG_RT_MUTEXES if RT_MUTEXES 1554 select LOCKDEP 1555 help 1556 This feature will check whether any held lock (spinlock, rwlock, 1557 mutex or rwsem) is incorrectly freed by the kernel, via any of the 1558 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(), 1559 vfree(), etc.), whether a live lock is incorrectly reinitialized via 1560 spin_lock_init()/mutex_init()/etc., or whether there is any lock 1561 held during task exit. 1562 1563config LOCKDEP 1564 bool 1565 depends on DEBUG_KERNEL && LOCK_DEBUGGING_SUPPORT 1566 select STACKTRACE 1567 select KALLSYMS 1568 select KALLSYMS_ALL 1569 1570config LOCKDEP_SMALL 1571 bool 1572 1573config LOCKDEP_BITS 1574 int "Size for MAX_LOCKDEP_ENTRIES (as Nth power of 2)" 1575 depends on LOCKDEP && !LOCKDEP_SMALL 1576 range 10 24 1577 default 15 1578 help 1579 Try increasing this value if you hit "BUG: MAX_LOCKDEP_ENTRIES too low!" message. 1580 1581config LOCKDEP_CHAINS_BITS 1582 int "Size for MAX_LOCKDEP_CHAINS (as Nth power of 2)" 1583 depends on LOCKDEP && !LOCKDEP_SMALL 1584 range 10 21 1585 default 16 1586 help 1587 Try increasing this value if you hit "BUG: MAX_LOCKDEP_CHAINS too low!" message. 1588 1589config LOCKDEP_STACK_TRACE_BITS 1590 int "Size for MAX_STACK_TRACE_ENTRIES (as Nth power of 2)" 1591 depends on LOCKDEP && !LOCKDEP_SMALL 1592 range 10 26 1593 default 21 if KASAN 1594 default 19 1595 help 1596 Try increasing this value if you hit "BUG: MAX_STACK_TRACE_ENTRIES too low!" message. 1597 1598 KASAN significantly increases stack trace consumption because its 1599 slab tracking interacts with lockdep's dependency validation under 1600 PREEMPT_FULL, creating a feedback loop. The higher default when 1601 KASAN is enabled costs ~12MB extra, which is negligible compared to 1602 KASAN's own shadow memory overhead. 1603 1604config LOCKDEP_STACK_TRACE_HASH_BITS 1605 int "Size for STACK_TRACE_HASH_SIZE (as Nth power of 2)" 1606 depends on LOCKDEP && !LOCKDEP_SMALL 1607 range 10 26 1608 default 16 if KASAN 1609 default 14 1610 help 1611 Try increasing this value if you need large STACK_TRACE_HASH_SIZE. 1612 1613config LOCKDEP_CIRCULAR_QUEUE_BITS 1614 int "Size for elements in circular_queue struct (as Nth power of 2)" 1615 depends on LOCKDEP 1616 range 10 26 1617 default 12 1618 help 1619 Try increasing this value if you hit "lockdep bfs error:-1" warning due to __cq_enqueue() failure. 1620 1621config DEBUG_LOCKDEP 1622 bool "Lock dependency engine debugging" 1623 depends on DEBUG_KERNEL && LOCKDEP 1624 select DEBUG_IRQFLAGS 1625 help 1626 If you say Y here, the lock dependency engine will do 1627 additional runtime checks to debug itself, at the price 1628 of more runtime overhead. 1629 1630config DEBUG_ATOMIC_SLEEP 1631 bool "Sleep inside atomic section checking" 1632 select PREEMPT_COUNT 1633 depends on DEBUG_KERNEL 1634 depends on !ARCH_NO_PREEMPT 1635 help 1636 If you say Y here, various routines which may sleep will become very 1637 noisy if they are called inside atomic sections: when a spinlock is 1638 held, inside an rcu read side critical section, inside preempt disabled 1639 sections, inside an interrupt, etc... 1640 1641config DEBUG_LOCKING_API_SELFTESTS 1642 bool "Locking API boot-time self-tests" 1643 depends on DEBUG_KERNEL 1644 help 1645 Say Y here if you want the kernel to run a short self-test during 1646 bootup. The self-test checks whether common types of locking bugs 1647 are detected by debugging mechanisms or not. (if you disable 1648 lock debugging then those bugs won't be detected of course.) 1649 The following locking APIs are covered: spinlocks, rwlocks, 1650 mutexes and rwsems. 1651 1652config LOCK_TORTURE_TEST 1653 tristate "torture tests for locking" 1654 depends on DEBUG_KERNEL 1655 select TORTURE_TEST 1656 help 1657 This option provides a kernel module that runs torture tests 1658 on kernel locking primitives. The kernel module may be built 1659 after the fact on the running kernel to be tested, if desired. 1660 1661 Say Y here if you want kernel locking-primitive torture tests 1662 to be built into the kernel. 1663 Say M if you want these torture tests to build as a module. 1664 Say N if you are unsure. 1665 1666config WW_MUTEX_SELFTEST 1667 tristate "Wait/wound mutex selftests" 1668 help 1669 This option provides a kernel module that runs tests on the 1670 on the struct ww_mutex locking API. 1671 1672 It is recommended to enable DEBUG_WW_MUTEX_SLOWPATH in conjunction 1673 with this test harness. 1674 1675 Say M if you want these self tests to build as a module. 1676 Say N if you are unsure. 1677 1678config SCF_TORTURE_TEST 1679 tristate "torture tests for smp_call_function*()" 1680 depends on DEBUG_KERNEL 1681 select TORTURE_TEST 1682 help 1683 This option provides a kernel module that runs torture tests 1684 on the smp_call_function() family of primitives. The kernel 1685 module may be built after the fact on the running kernel to 1686 be tested, if desired. 1687 1688config CSD_LOCK_WAIT_DEBUG 1689 bool "Debugging for csd_lock_wait(), called from smp_call_function*()" 1690 depends on DEBUG_KERNEL 1691 depends on SMP 1692 depends on 64BIT 1693 default n 1694 help 1695 This option enables debug prints when CPUs are slow to respond 1696 to the smp_call_function*() IPI wrappers. These debug prints 1697 include the IPI handler function currently executing (if any) 1698 and relevant stack traces. 1699 1700config CSD_LOCK_WAIT_DEBUG_DEFAULT 1701 bool "Default csd_lock_wait() debugging on at boot time" 1702 depends on CSD_LOCK_WAIT_DEBUG 1703 depends on 64BIT 1704 default n 1705 help 1706 This option causes the csdlock_debug= kernel boot parameter to 1707 default to 1 (basic debugging) instead of 0 (no debugging). 1708 1709endmenu # lock debugging 1710 1711config TRACE_IRQFLAGS 1712 depends on TRACE_IRQFLAGS_SUPPORT 1713 bool 1714 help 1715 Enables hooks to interrupt enabling and disabling for 1716 either tracing or lock debugging. 1717 1718config TRACE_IRQFLAGS_NMI 1719 def_bool y 1720 depends on TRACE_IRQFLAGS 1721 depends on TRACE_IRQFLAGS_NMI_SUPPORT 1722 1723config NMI_CHECK_CPU 1724 bool "Debugging for CPUs failing to respond to backtrace requests" 1725 depends on DEBUG_KERNEL 1726 depends on X86 1727 default n 1728 help 1729 Enables debug prints when a CPU fails to respond to a given 1730 backtrace NMI. These prints provide some reasons why a CPU 1731 might legitimately be failing to respond, for example, if it 1732 is offline of if ignore_nmis is set. 1733 1734config DEBUG_IRQFLAGS 1735 bool "Debug IRQ flag manipulation" 1736 help 1737 Enables checks for potentially unsafe enabling or disabling of 1738 interrupts, such as calling raw_local_irq_restore() when interrupts 1739 are enabled. 1740 1741config STACKTRACE 1742 bool "Stack backtrace support" 1743 depends on STACKTRACE_SUPPORT 1744 help 1745 This option causes the kernel to create a /proc/pid/stack for 1746 every process, showing its current stack trace. 1747 It is also used by various kernel debugging features that require 1748 stack trace generation. 1749 1750config DEBUG_KOBJECT 1751 bool "kobject debugging" 1752 depends on DEBUG_KERNEL 1753 help 1754 If you say Y here, some extra kobject debugging messages will be sent 1755 to the syslog. 1756 1757config DEBUG_KOBJECT_RELEASE 1758 bool "kobject release debugging" 1759 depends on DEBUG_OBJECTS_TIMERS 1760 help 1761 kobjects are reference counted objects. This means that their 1762 last reference count put is not predictable, and the kobject can 1763 live on past the point at which a driver decides to drop its 1764 initial reference to the kobject gained on allocation. An 1765 example of this would be a struct device which has just been 1766 unregistered. 1767 1768 However, some buggy drivers assume that after such an operation, 1769 the memory backing the kobject can be immediately freed. This 1770 goes completely against the principles of a refcounted object. 1771 1772 If you say Y here, the kernel will delay the release of kobjects 1773 on the last reference count to improve the visibility of this 1774 kind of kobject release bug. 1775 1776config HAVE_DEBUG_BUGVERBOSE 1777 bool 1778 1779menu "Debug kernel data structures" 1780 1781config DEBUG_LIST 1782 bool "Debug linked list manipulation" 1783 depends on DEBUG_KERNEL 1784 select LIST_HARDENED 1785 help 1786 Enable this to turn on extended checks in the linked-list walking 1787 routines. 1788 1789 This option trades better quality error reports for performance, and 1790 is more suitable for kernel debugging. If you care about performance, 1791 you should only enable CONFIG_LIST_HARDENED instead. 1792 1793 If unsure, say N. 1794 1795config DEBUG_PLIST 1796 bool "Debug priority linked list manipulation" 1797 depends on DEBUG_KERNEL 1798 help 1799 Enable this to turn on extended checks in the priority-ordered 1800 linked-list (plist) walking routines. This checks the entire 1801 list multiple times during each manipulation. 1802 1803 If unsure, say N. 1804 1805config DEBUG_SG 1806 bool "Debug SG table operations" 1807 depends on DEBUG_KERNEL 1808 help 1809 Enable this to turn on checks on scatter-gather tables. This can 1810 help find problems with drivers that do not properly initialize 1811 their sg tables. 1812 1813 If unsure, say N. 1814 1815config DEBUG_NOTIFIERS 1816 bool "Debug notifier call chains" 1817 depends on DEBUG_KERNEL 1818 help 1819 Enable this to turn on sanity checking for notifier call chains. 1820 This is most useful for kernel developers to make sure that 1821 modules properly unregister themselves from notifier chains. 1822 This is a relatively cheap check but if you care about maximum 1823 performance, say N. 1824 1825config DEBUG_CLOSURES 1826 bool "Debug closures (bcache async widgits)" 1827 depends on CLOSURES 1828 select DEBUG_FS 1829 help 1830 Keeps all active closures in a linked list and provides a debugfs 1831 interface to list them, which makes it possible to see asynchronous 1832 operations that get stuck. 1833 1834config DEBUG_MAPLE_TREE 1835 bool "Debug maple trees" 1836 depends on DEBUG_KERNEL 1837 help 1838 Enable maple tree debugging information and extra validations. 1839 1840 If unsure, say N. 1841 1842endmenu 1843 1844source "kernel/rcu/Kconfig.debug" 1845 1846config DEBUG_WQ_FORCE_RR_CPU 1847 bool "Force round-robin CPU selection for unbound work items" 1848 depends on DEBUG_KERNEL 1849 default n 1850 help 1851 Workqueue used to implicitly guarantee that work items queued 1852 without explicit CPU specified are put on the local CPU. This 1853 guarantee is no longer true and while local CPU is still 1854 preferred work items may be put on foreign CPUs. Kernel 1855 parameter "workqueue.debug_force_rr_cpu" is added to force 1856 round-robin CPU selection to flush out usages which depend on the 1857 now broken guarantee. This config option enables the debug 1858 feature by default. When enabled, memory and cache locality will 1859 be impacted. 1860 1861config CPU_HOTPLUG_STATE_CONTROL 1862 bool "Enable CPU hotplug state control" 1863 depends on DEBUG_KERNEL 1864 depends on HOTPLUG_CPU 1865 default n 1866 help 1867 Allows to write steps between "offline" and "online" to the CPUs 1868 sysfs target file so states can be stepped granular. This is a debug 1869 option for now as the hotplug machinery cannot be stopped and 1870 restarted at arbitrary points yet. 1871 1872 Say N if your are unsure. 1873 1874config LATENCYTOP 1875 bool "Latency measuring infrastructure" 1876 depends on DEBUG_KERNEL 1877 depends on STACKTRACE_SUPPORT 1878 depends on PROC_FS 1879 depends on FRAME_POINTER || MIPS || PPC || S390 || MICROBLAZE || ARM || ARC || X86 1880 select KALLSYMS 1881 select KALLSYMS_ALL 1882 select STACKTRACE 1883 select SCHEDSTATS 1884 help 1885 Enable this option if you want to use the LatencyTOP tool 1886 to find out which userspace is blocking on what kernel operations. 1887 1888config DEBUG_CGROUP_REF 1889 bool "Disable inlining of cgroup css reference count functions" 1890 depends on DEBUG_KERNEL 1891 depends on CGROUPS 1892 depends on KPROBES 1893 default n 1894 help 1895 Force cgroup css reference count functions to not be inlined so 1896 that they can be kprobed for debugging. 1897 1898source "kernel/trace/Kconfig" 1899 1900config PROVIDE_OHCI1394_DMA_INIT 1901 bool "Remote debugging over FireWire early on boot" 1902 depends on PCI && X86 1903 help 1904 If you want to debug problems which hang or crash the kernel early 1905 on boot and the crashing machine has a FireWire port, you can use 1906 this feature to remotely access the memory of the crashed machine 1907 over FireWire. This employs remote DMA as part of the OHCI1394 1908 specification which is now the standard for FireWire controllers. 1909 1910 With remote DMA, you can monitor the printk buffer remotely using 1911 firescope and access all memory below 4GB using fireproxy from gdb. 1912 Even controlling a kernel debugger is possible using remote DMA. 1913 1914 Usage: 1915 1916 If ohci1394_dma=early is used as boot parameter, it will initialize 1917 all OHCI1394 controllers which are found in the PCI config space. 1918 1919 As all changes to the FireWire bus such as enabling and disabling 1920 devices cause a bus reset and thereby disable remote DMA for all 1921 devices, be sure to have the cable plugged and FireWire enabled on 1922 the debugging host before booting the debug target for debugging. 1923 1924 This code (~1k) is freed after boot. By then, the firewire stack 1925 in charge of the OHCI-1394 controllers should be used instead. 1926 1927 See Documentation/core-api/debugging-via-ohci1394.rst for more information. 1928 1929source "samples/Kconfig" 1930 1931config ARCH_HAS_DEVMEM_IS_ALLOWED 1932 bool 1933 1934config STRICT_DEVMEM 1935 bool "Filter access to /dev/mem" 1936 depends on MMU && DEVMEM 1937 depends on ARCH_HAS_DEVMEM_IS_ALLOWED || GENERIC_LIB_DEVMEM_IS_ALLOWED 1938 default y if PPC || X86 || ARM64 || S390 1939 help 1940 If this option is disabled, you allow userspace (root) access to all 1941 of memory, including kernel and userspace memory. Accidental 1942 access to this is obviously disastrous, but specific access can 1943 be used by people debugging the kernel. Note that with PAT support 1944 enabled, even in this case there are restrictions on /dev/mem 1945 use due to the cache aliasing requirements. 1946 1947 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem 1948 file only allows userspace access to PCI space and the BIOS code and 1949 data regions. This is sufficient for dosemu and X and all common 1950 users of /dev/mem. 1951 1952 If in doubt, say Y. 1953 1954config IO_STRICT_DEVMEM 1955 bool "Filter I/O access to /dev/mem" 1956 depends on STRICT_DEVMEM 1957 help 1958 If this option is disabled, you allow userspace (root) access to all 1959 io-memory regardless of whether a driver is actively using that 1960 range. Accidental access to this is obviously disastrous, but 1961 specific access can be used by people debugging kernel drivers. 1962 1963 If this option is switched on, the /dev/mem file only allows 1964 userspace access to *idle* io-memory ranges (see /proc/iomem) This 1965 may break traditional users of /dev/mem (dosemu, legacy X, etc...) 1966 if the driver using a given range cannot be disabled. 1967 1968 If in doubt, say Y. 1969 1970menu "$(SRCARCH) Debugging" 1971 1972source "arch/$(SRCARCH)/Kconfig.debug" 1973 1974endmenu 1975 1976menu "Kernel Testing and Coverage" 1977 1978source "lib/kunit/Kconfig" 1979 1980config NOTIFIER_ERROR_INJECTION 1981 tristate "Notifier error injection" 1982 depends on DEBUG_KERNEL 1983 select DEBUG_FS 1984 help 1985 This option provides the ability to inject artificial errors to 1986 specified notifier chain callbacks. It is useful to test the error 1987 handling of notifier call chain failures. 1988 1989 Say N if unsure. 1990 1991config PM_NOTIFIER_ERROR_INJECT 1992 tristate "PM notifier error injection module" 1993 depends on PM && NOTIFIER_ERROR_INJECTION 1994 default m if PM_DEBUG 1995 help 1996 This option provides the ability to inject artificial errors to 1997 PM notifier chain callbacks. It is controlled through debugfs 1998 interface /sys/kernel/debug/notifier-error-inject/pm 1999 2000 If the notifier call chain should be failed with some events 2001 notified, write the error code to "actions/<notifier event>/error". 2002 2003 Example: Inject PM suspend error (-12 = -ENOMEM) 2004 2005 # cd /sys/kernel/debug/notifier-error-inject/pm/ 2006 # echo -12 > actions/PM_SUSPEND_PREPARE/error 2007 # echo mem > /sys/power/state 2008 bash: echo: write error: Cannot allocate memory 2009 2010 To compile this code as a module, choose M here: the module will 2011 be called pm-notifier-error-inject. 2012 2013 If unsure, say N. 2014 2015config OF_RECONFIG_NOTIFIER_ERROR_INJECT 2016 tristate "OF reconfig notifier error injection module" 2017 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION 2018 help 2019 This option provides the ability to inject artificial errors to 2020 OF reconfig notifier chain callbacks. It is controlled 2021 through debugfs interface under 2022 /sys/kernel/debug/notifier-error-inject/OF-reconfig/ 2023 2024 If the notifier call chain should be failed with some events 2025 notified, write the error code to "actions/<notifier event>/error". 2026 2027 To compile this code as a module, choose M here: the module will 2028 be called of-reconfig-notifier-error-inject. 2029 2030 If unsure, say N. 2031 2032config NETDEV_NOTIFIER_ERROR_INJECT 2033 tristate "Netdev notifier error injection module" 2034 depends on NET && NOTIFIER_ERROR_INJECTION 2035 help 2036 This option provides the ability to inject artificial errors to 2037 netdevice notifier chain callbacks. It is controlled through debugfs 2038 interface /sys/kernel/debug/notifier-error-inject/netdev 2039 2040 If the notifier call chain should be failed with some events 2041 notified, write the error code to "actions/<notifier event>/error". 2042 2043 Example: Inject netdevice mtu change error (-22 = -EINVAL) 2044 2045 # cd /sys/kernel/debug/notifier-error-inject/netdev 2046 # echo -22 > actions/NETDEV_CHANGEMTU/error 2047 # ip link set eth0 mtu 1024 2048 RTNETLINK answers: Invalid argument 2049 2050 To compile this code as a module, choose M here: the module will 2051 be called netdev-notifier-error-inject. 2052 2053 If unsure, say N. 2054 2055config FUNCTION_ERROR_INJECTION 2056 bool "Fault-injections of functions" 2057 depends on HAVE_FUNCTION_ERROR_INJECTION && KPROBES 2058 help 2059 Add fault injections into various functions that are annotated with 2060 ALLOW_ERROR_INJECTION() in the kernel. BPF may also modify the return 2061 value of these functions. This is useful to test error paths of code. 2062 2063 If unsure, say N 2064 2065config FAULT_INJECTION 2066 bool "Fault-injection framework" 2067 depends on DEBUG_KERNEL 2068 help 2069 Provide fault-injection framework. 2070 For more details, see Documentation/fault-injection/. 2071 2072config FAILSLAB 2073 bool "Fault-injection capability for kmalloc" 2074 depends on FAULT_INJECTION 2075 help 2076 Provide fault-injection capability for kmalloc. 2077 2078config FAIL_PAGE_ALLOC 2079 bool "Fault-injection capability for alloc_pages()" 2080 depends on FAULT_INJECTION 2081 help 2082 Provide fault-injection capability for alloc_pages(). 2083 2084config FAULT_INJECTION_USERCOPY 2085 bool "Fault injection capability for usercopy functions" 2086 depends on FAULT_INJECTION 2087 help 2088 Provides fault-injection capability to inject failures 2089 in usercopy functions (copy_from_user(), get_user(), ...). 2090 2091config FAIL_MAKE_REQUEST 2092 bool "Fault-injection capability for disk IO" 2093 depends on FAULT_INJECTION && BLOCK 2094 help 2095 Provide fault-injection capability for disk IO. 2096 2097config FAIL_IO_TIMEOUT 2098 bool "Fault-injection capability for faking disk interrupts" 2099 depends on FAULT_INJECTION && BLOCK 2100 help 2101 Provide fault-injection capability on end IO handling. This 2102 will make the block layer "forget" an interrupt as configured, 2103 thus exercising the error handling. 2104 2105 Only works with drivers that use the generic timeout handling, 2106 for others it won't do anything. 2107 2108config FAIL_FUTEX 2109 bool "Fault-injection capability for futexes" 2110 select DEBUG_FS 2111 depends on FAULT_INJECTION && FUTEX 2112 help 2113 Provide fault-injection capability for futexes. 2114 2115config FAULT_INJECTION_DEBUG_FS 2116 bool "Debugfs entries for fault-injection capabilities" 2117 depends on FAULT_INJECTION && SYSFS && DEBUG_FS 2118 help 2119 Enable configuration of fault-injection capabilities via debugfs. 2120 2121config FAIL_FUNCTION 2122 bool "Fault-injection capability for functions" 2123 depends on FAULT_INJECTION_DEBUG_FS && FUNCTION_ERROR_INJECTION 2124 help 2125 Provide function-based fault-injection capability. 2126 This will allow you to override a specific function with a return 2127 with given return value. As a result, function caller will see 2128 an error value and have to handle it. This is useful to test the 2129 error handling in various subsystems. 2130 2131config FAIL_MMC_REQUEST 2132 bool "Fault-injection capability for MMC IO" 2133 depends on FAULT_INJECTION_DEBUG_FS && MMC 2134 help 2135 Provide fault-injection capability for MMC IO. 2136 This will make the mmc core return data errors. This is 2137 useful to test the error handling in the mmc block device 2138 and to test how the mmc host driver handles retries from 2139 the block device. 2140 2141config FAIL_SUNRPC 2142 bool "Fault-injection capability for SunRPC" 2143 depends on FAULT_INJECTION_DEBUG_FS && SUNRPC_DEBUG 2144 help 2145 Provide fault-injection capability for SunRPC and 2146 its consumers. 2147 2148config FAIL_SKB_REALLOC 2149 bool "Fault-injection capability forcing skb to reallocate" 2150 depends on FAULT_INJECTION_DEBUG_FS 2151 help 2152 Provide fault-injection capability that forces the skb to be 2153 reallocated, catching possible invalid pointers to the skb. 2154 2155 For more information, check 2156 Documentation/fault-injection/fault-injection.rst 2157 2158config FAULT_INJECTION_CONFIGFS 2159 bool "Configfs interface for fault-injection capabilities" 2160 depends on FAULT_INJECTION 2161 select CONFIGFS_FS 2162 help 2163 This option allows configfs-based drivers to dynamically configure 2164 fault-injection via configfs. Each parameter for driver-specific 2165 fault-injection can be made visible as a configfs attribute in a 2166 configfs group. 2167 2168 2169config FAULT_INJECTION_STACKTRACE_FILTER 2170 bool "stacktrace filter for fault-injection capabilities" 2171 depends on FAULT_INJECTION 2172 depends on (FAULT_INJECTION_DEBUG_FS || FAULT_INJECTION_CONFIGFS) && STACKTRACE_SUPPORT 2173 select STACKTRACE 2174 depends on FRAME_POINTER || MIPS || PPC || S390 || MICROBLAZE || ARM || ARC || X86 2175 help 2176 Provide stacktrace filter for fault-injection capabilities 2177 2178config ARCH_HAS_KCOV 2179 bool 2180 help 2181 An architecture should select this when it can successfully 2182 build and run with CONFIG_KCOV. This typically requires 2183 disabling instrumentation for some early boot code. 2184 2185config KCOV 2186 bool "Code coverage for fuzzing" 2187 depends on ARCH_HAS_KCOV 2188 depends on !ARCH_WANTS_NO_INSTR || HAVE_NOINSTR_HACK || \ 2189 GCC_VERSION >= 120000 || CC_IS_CLANG 2190 select DEBUG_FS 2191 select OBJTOOL if HAVE_NOINSTR_HACK 2192 help 2193 KCOV exposes kernel code coverage information in a form suitable 2194 for coverage-guided fuzzing (randomized testing). 2195 2196 For more details, see Documentation/dev-tools/kcov.rst. 2197 2198config KCOV_ENABLE_COMPARISONS 2199 bool "Enable comparison operands collection by KCOV" 2200 depends on KCOV 2201 depends on $(cc-option,-fsanitize-coverage=trace-cmp) 2202 help 2203 KCOV also exposes operands of every comparison in the instrumented 2204 code along with operand sizes and PCs of the comparison instructions. 2205 These operands can be used by fuzzing engines to improve the quality 2206 of fuzzing coverage. 2207 2208config KCOV_INSTRUMENT_ALL 2209 bool "Instrument all code by default" 2210 depends on KCOV 2211 default y 2212 help 2213 If you are doing generic system call fuzzing (like e.g. syzkaller), 2214 then you will want to instrument the whole kernel and you should 2215 say y here. If you are doing more targeted fuzzing (like e.g. 2216 filesystem fuzzing with AFL) then you will want to enable coverage 2217 for more specific subsets of files, and should say n here. 2218 2219config KCOV_IRQ_AREA_SIZE 2220 hex "Size of interrupt coverage collection area in words" 2221 depends on KCOV 2222 default 0x40000 2223 help 2224 KCOV uses preallocated per-cpu areas to collect coverage from 2225 soft interrupts. This specifies the size of those areas in the 2226 number of unsigned long words. 2227 2228config KCOV_SELFTEST 2229 bool "Perform short selftests on boot" 2230 depends on KCOV 2231 help 2232 Run short KCOV coverage collection selftests on boot. 2233 On test failure, causes the kernel to panic. Recommended to be 2234 enabled, ensuring critical functionality works as intended. 2235 2236menuconfig RUNTIME_TESTING_MENU 2237 bool "Runtime Testing" 2238 default y 2239 2240if RUNTIME_TESTING_MENU 2241 2242config TEST_DHRY 2243 tristate "Dhrystone benchmark test" 2244 help 2245 Enable this to include the Dhrystone 2.1 benchmark. This test 2246 calculates the number of Dhrystones per second, and the number of 2247 DMIPS (Dhrystone MIPS) obtained when the Dhrystone score is divided 2248 by 1757 (the number of Dhrystones per second obtained on the VAX 2249 11/780, nominally a 1 MIPS machine). 2250 2251 To run the benchmark, it needs to be enabled explicitly, either from 2252 the kernel command line (when built-in), or from userspace (when 2253 built-in or modular). 2254 2255 Run once during kernel boot: 2256 2257 test_dhry.run 2258 2259 Set number of iterations from kernel command line: 2260 2261 test_dhry.iterations=<n> 2262 2263 Set number of iterations from userspace: 2264 2265 echo <n> > /sys/module/test_dhry/parameters/iterations 2266 2267 Trigger manual run from userspace: 2268 2269 echo y > /sys/module/test_dhry/parameters/run 2270 2271 If the number of iterations is <= 0, the test will devise a suitable 2272 number of iterations (test runs for at least 2s) automatically. 2273 This process takes ca. 4s. 2274 2275 If unsure, say N. 2276 2277config LKDTM 2278 tristate "Linux Kernel Dump Test Tool Module" 2279 depends on DEBUG_FS 2280 help 2281 This module enables testing of the different dumping mechanisms by 2282 inducing system failures at predefined crash points. 2283 If you don't need it: say N 2284 Choose M here to compile this code as a module. The module will be 2285 called lkdtm. 2286 2287 Documentation on how to use the module can be found in 2288 Documentation/fault-injection/provoke-crashes.rst 2289 2290config CPUMASK_KUNIT_TEST 2291 tristate "KUnit test for cpumask" if !KUNIT_ALL_TESTS 2292 depends on KUNIT 2293 default KUNIT_ALL_TESTS 2294 help 2295 Enable to turn on cpumask tests, running at boot or module load time. 2296 2297 For more information on KUnit and unit tests in general, please refer 2298 to the KUnit documentation in Documentation/dev-tools/kunit/. 2299 2300 If unsure, say N. 2301 2302config TEST_LIST_SORT 2303 tristate "Linked list sorting test" if !KUNIT_ALL_TESTS 2304 depends on KUNIT 2305 default KUNIT_ALL_TESTS 2306 help 2307 Enable this to turn on 'list_sort()' function test. This test is 2308 executed only once during system boot (so affects only boot time), 2309 or at module load time. 2310 2311 If unsure, say N. 2312 2313config TEST_SORT 2314 tristate "Array-based sort test" if !KUNIT_ALL_TESTS 2315 depends on KUNIT 2316 default KUNIT_ALL_TESTS 2317 help 2318 This option enables the self-test function of 'sort()' at boot, 2319 or at module load time. 2320 2321 If unsure, say N. 2322 2323config TEST_DIV64 2324 tristate "64bit/32bit division and modulo test" 2325 depends on DEBUG_KERNEL || m 2326 help 2327 Enable this to turn on 'do_div()' function test. This test is 2328 executed only once during system boot (so affects only boot time), 2329 or at module load time. 2330 2331 If unsure, say N. 2332 2333config TEST_MULDIV64 2334 tristate "mul_u64_u64_div_u64() test" 2335 depends on DEBUG_KERNEL || m 2336 help 2337 Enable this to turn on 'mul_u64_u64_div_u64()' function test. 2338 This test is executed only once during system boot (so affects 2339 only boot time), or at module load time. 2340 2341 If unsure, say N. 2342 2343config TEST_IOV_ITER 2344 tristate "Test iov_iter operation" if !KUNIT_ALL_TESTS 2345 depends on KUNIT 2346 depends on MMU 2347 default KUNIT_ALL_TESTS 2348 help 2349 Enable this to turn on testing of the operation of the I/O iterator 2350 (iov_iter). This test is executed only once during system boot (so 2351 affects only boot time), or at module load time. 2352 2353 If unsure, say N. 2354 2355config KPROBES_SANITY_TEST 2356 tristate "Kprobes sanity tests" if !KUNIT_ALL_TESTS 2357 depends on DEBUG_KERNEL 2358 depends on KPROBES 2359 depends on KUNIT 2360 select STACKTRACE if ARCH_CORRECT_STACKTRACE_ON_KRETPROBE 2361 default KUNIT_ALL_TESTS 2362 help 2363 This option provides for testing basic kprobes functionality on 2364 boot. Samples of kprobe and kretprobe are inserted and 2365 verified for functionality. 2366 2367 Say N if you are unsure. 2368 2369config FPROBE_SANITY_TEST 2370 bool "Self test for fprobe" 2371 depends on DEBUG_KERNEL 2372 depends on FPROBE 2373 depends on KUNIT=y 2374 help 2375 This option will enable testing the fprobe when the system boot. 2376 A series of tests are made to verify that the fprobe is functioning 2377 properly. 2378 2379 Say N if you are unsure. 2380 2381config BACKTRACE_SELF_TEST 2382 tristate "Self test for the backtrace code" 2383 depends on DEBUG_KERNEL 2384 help 2385 This option provides a kernel module that can be used to test 2386 the kernel stack backtrace code. This option is not useful 2387 for distributions or general kernels, but only for kernel 2388 developers working on architecture code. 2389 2390 Note that if you want to also test saved backtraces, you will 2391 have to enable STACKTRACE as well. 2392 2393 Say N if you are unsure. 2394 2395config TEST_REF_TRACKER 2396 tristate "Self test for reference tracker" 2397 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT 2398 select REF_TRACKER 2399 help 2400 This option provides a kernel module performing tests 2401 using reference tracker infrastructure. 2402 2403 Say N if you are unsure. 2404 2405config RBTREE_TEST 2406 tristate "Red-Black tree test" 2407 depends on DEBUG_KERNEL 2408 help 2409 A benchmark measuring the performance of the rbtree library. 2410 Also includes rbtree invariant checks. 2411 2412config REED_SOLOMON_TEST 2413 tristate "Reed-Solomon library test" 2414 depends on DEBUG_KERNEL || m 2415 select REED_SOLOMON 2416 select REED_SOLOMON_ENC16 2417 select REED_SOLOMON_DEC16 2418 help 2419 This option enables the self-test function of rslib at boot, 2420 or at module load time. 2421 2422 If unsure, say N. 2423 2424config INTERVAL_TREE_TEST 2425 tristate "Interval tree test" 2426 depends on DEBUG_KERNEL 2427 select INTERVAL_TREE 2428 help 2429 A benchmark measuring the performance of the interval tree library 2430 2431config PERCPU_TEST 2432 tristate "Per cpu operations test" 2433 depends on m && DEBUG_KERNEL 2434 help 2435 Enable this option to build test module which validates per-cpu 2436 operations. 2437 2438 If unsure, say N. 2439 2440config ATOMIC64_SELFTEST 2441 tristate "Perform an atomic64_t self-test" 2442 help 2443 Enable this option to test the atomic64_t functions at boot or 2444 at module load time. 2445 2446 If unsure, say N. 2447 2448config ASYNC_RAID6_TEST 2449 tristate "Self test for hardware accelerated raid6 recovery" 2450 depends on ASYNC_RAID6_RECOV 2451 select ASYNC_MEMCPY 2452 help 2453 This is a one-shot self test that permutes through the 2454 recovery of all the possible two disk failure scenarios for a 2455 N-disk array. Recovery is performed with the asynchronous 2456 raid6 recovery routines, and will optionally use an offload 2457 engine if one is available. 2458 2459 If unsure, say N. 2460 2461config TEST_HEXDUMP 2462 tristate "Test functions located in the hexdump module at runtime" 2463 2464config PRINTF_KUNIT_TEST 2465 tristate "KUnit test printf() family of functions at runtime" if !KUNIT_ALL_TESTS 2466 depends on KUNIT 2467 default KUNIT_ALL_TESTS 2468 help 2469 Enable this option to test the printf functions at runtime. 2470 2471 If unsure, say N. 2472 2473config SCANF_KUNIT_TEST 2474 tristate "KUnit test scanf() family of functions at runtime" if !KUNIT_ALL_TESTS 2475 depends on KUNIT 2476 default KUNIT_ALL_TESTS 2477 help 2478 Enable this option to test the scanf functions at runtime. 2479 2480 If unsure, say N. 2481 2482config SEQ_BUF_KUNIT_TEST 2483 tristate "KUnit test for seq_buf" if !KUNIT_ALL_TESTS 2484 depends on KUNIT 2485 default KUNIT_ALL_TESTS 2486 help 2487 This builds unit tests for the seq_buf library. 2488 2489 If unsure, say N. 2490 2491config STRING_KUNIT_TEST 2492 tristate "KUnit test string functions at runtime" if !KUNIT_ALL_TESTS 2493 depends on KUNIT 2494 default KUNIT_ALL_TESTS 2495 2496config STRING_HELPERS_KUNIT_TEST 2497 tristate "KUnit test string helpers at runtime" if !KUNIT_ALL_TESTS 2498 depends on KUNIT 2499 default KUNIT_ALL_TESTS 2500 2501config STRING_KUNIT_BENCH 2502 bool "Benchmark string functions at runtime" 2503 depends on STRING_KUNIT_TEST 2504 help 2505 Enable performance measurement for string functions. 2506 2507 This measures the execution efficiency of string functions 2508 during the KUnit test run. 2509 2510 If unsure, say N. 2511 2512config FFS_KUNIT_TEST 2513 tristate "KUnit test ffs-family functions at runtime" if !KUNIT_ALL_TESTS 2514 depends on KUNIT 2515 default KUNIT_ALL_TESTS 2516 help 2517 This builds KUnit tests for ffs-family bit manipulation functions 2518 including ffs(), __ffs(), fls(), __fls(), fls64(), and __ffs64(). 2519 2520 These tests validate mathematical correctness, edge case handling, 2521 and cross-architecture consistency of bit scanning functions. 2522 2523 For more information on KUnit and unit tests in general, 2524 please refer to Documentation/dev-tools/kunit/. 2525 2526config TEST_KSTRTOX 2527 tristate "Test kstrto*() family of functions at runtime" 2528 2529config TEST_BITMAP 2530 tristate "Test bitmap_*() family of functions at runtime" 2531 help 2532 Enable this option to test the bitmap functions at boot. 2533 2534 If unsure, say N. 2535 2536config TEST_XARRAY 2537 tristate "Test the XArray code at runtime" 2538 2539config TEST_MAPLE_TREE 2540 tristate "Test the Maple Tree code at runtime or module load" 2541 help 2542 Enable this option to test the maple tree code functions at boot, or 2543 when the module is loaded. Enable "Debug Maple Trees" will enable 2544 more verbose output on failures. 2545 2546 If unsure, say N. 2547 2548config TEST_RHASHTABLE 2549 tristate "Perform selftest on resizable hash table" 2550 help 2551 Enable this option to test the rhashtable functions at boot. 2552 2553 If unsure, say N. 2554 2555config TEST_IDA 2556 tristate "Perform selftest on IDA functions" 2557 2558config TEST_MISC_MINOR 2559 bool "miscdevice KUnit test" if !KUNIT_ALL_TESTS 2560 depends on KUNIT=y 2561 default KUNIT_ALL_TESTS 2562 help 2563 Kunit test for miscdevice API, specially its behavior in respect to 2564 static and dynamic minor numbers. 2565 2566 KUnit tests run during boot and output the results to the debug log 2567 in TAP format (https://testanything.org/). Only useful for kernel devs 2568 running the KUnit test harness, and not intended for inclusion into a 2569 production build. 2570 2571 For more information on KUnit and unit tests in general please refer 2572 to the KUnit documentation in Documentation/dev-tools/kunit/. 2573 2574 If unsure, say N. 2575 2576config TEST_PARMAN 2577 tristate "Perform selftest on priority array manager" 2578 depends on PARMAN 2579 help 2580 Enable this option to test priority array manager on boot 2581 (or module load). 2582 2583 If unsure, say N. 2584 2585config TEST_LKM 2586 tristate "Test module loading with 'hello world' module" 2587 depends on m 2588 help 2589 This builds the "test_module" module that emits "Hello, world" 2590 on printk when loaded. It is designed to be used for basic 2591 evaluation of the module loading subsystem (for example when 2592 validating module verification). It lacks any extra dependencies, 2593 and will not normally be loaded by the system unless explicitly 2594 requested by name. 2595 2596 If unsure, say N. 2597 2598config TEST_BITOPS 2599 tristate "Test module for compilation of bitops operations" 2600 help 2601 This builds the "test_bitops" module that is much like the 2602 TEST_LKM module except that it does a basic exercise of the 2603 set/clear_bit macros and get_count_order/long to make sure there are 2604 no compiler warnings from C=1 sparse checker or -Wextra 2605 compilations. It has no dependencies and doesn't run or load unless 2606 explicitly requested by name. for example: modprobe test_bitops. 2607 2608 If unsure, say N. 2609 2610config TEST_VMALLOC 2611 tristate "Test module for stress/performance analysis of vmalloc allocator" 2612 default n 2613 depends on MMU 2614 help 2615 This builds the "test_vmalloc" module that should be used for 2616 stress and performance analysis. So, any new change for vmalloc 2617 subsystem can be evaluated from performance and stability point 2618 of view. 2619 2620 If unsure, say N. 2621 2622config TEST_WORKQUEUE 2623 tristate "Test module for stress/performance analysis of workqueue" 2624 depends on m 2625 default n 2626 help 2627 This builds the "test_workqueue" module for benchmarking 2628 workqueue throughput under contention. Useful for evaluating 2629 affinity scope changes (e.g., cache_shard vs cache). 2630 2631 The test drives sysfs to switch affinity scopes, so it must be 2632 loaded after userspace has mounted sysfs; building it in (=y) 2633 would run module_init before /sys is available. 2634 2635 If unsure, say N. 2636 2637config TEST_BPF 2638 tristate "Test BPF filter functionality" 2639 depends on m && NET 2640 help 2641 This builds the "test_bpf" module that runs various test vectors 2642 against the BPF interpreter or BPF JIT compiler depending on the 2643 current setting. This is in particular useful for BPF JIT compiler 2644 development, but also to run regression tests against changes in 2645 the interpreter code. It also enables test stubs for eBPF maps and 2646 verifier used by user space verifier testsuite. 2647 2648 If unsure, say N. 2649 2650config FIND_BIT_BENCHMARK 2651 tristate "Test find_bit functions" 2652 help 2653 This builds the "test_find_bit" module that measure find_*_bit() 2654 functions performance. 2655 2656 If unsure, say N. 2657 2658config REGION_ALLOC_BENCHMARK 2659 tristate "Benchmark bitmap, IDA and Maple Tree region allocation" 2660 help 2661 This builds a microbenchmark comparing variable-sized region 2662 allocation using bitmaps, IDA and Maple Tree. The benchmark 2663 runs at initialization time. 2664 2665 Usage: 2666 insmod region_alloc_benchmark.ko 2667 insmod region_alloc_benchmark.ko capacities=1024,2048,4096,65536 2668 2669 If unsure, say N. 2670 2671config FIND_BIT_BENCHMARK_RUST 2672 tristate "Test find_bit functions in Rust" 2673 depends on RUST 2674 help 2675 This builds the "find_bit_benchmark_rust" module. It is a micro 2676 benchmark that measures the performance of Rust functions that 2677 correspond to the find_*_bit() operations in C. It follows the 2678 FIND_BIT_BENCHMARK closely but will in general not yield same 2679 numbers due to extra bounds checks and overhead of foreign 2680 function calls. 2681 2682 If unsure, say N. 2683 2684config TEST_FIRMWARE 2685 tristate "Test firmware loading via userspace interface" 2686 depends on FW_LOADER 2687 help 2688 This builds the "test_firmware" module that creates a userspace 2689 interface for testing firmware loading. This can be used to 2690 control the triggering of firmware loading without needing an 2691 actual firmware-using device. The contents can be rechecked by 2692 userspace. 2693 2694 If unsure, say N. 2695 2696config TEST_SYSCTL 2697 tristate "sysctl test driver" 2698 depends on SYSCTL 2699 help 2700 This builds the "test_sysctl" module. This driver enables to test the 2701 proc sysctl interfaces available to drivers safely without affecting 2702 production knobs which might alter system functionality. 2703 2704 If unsure, say N. 2705 2706config BITOPS_KUNIT 2707 tristate "KUnit test for bitops" if !KUNIT_ALL_TESTS 2708 depends on KUNIT 2709 default KUNIT_ALL_TESTS 2710 help 2711 This option enables the KUnit test for the bitops library 2712 which provides functions for bit operations. 2713 2714 Note that this is derived from the original test_bitops module. 2715 For micro-benchmarks and compiler warning checks, enable TEST_BITOPS. 2716 2717 If unsure, say N. 2718 2719config BITFIELD_KUNIT 2720 tristate "KUnit test bitfield functions at runtime" if !KUNIT_ALL_TESTS 2721 depends on KUNIT 2722 default KUNIT_ALL_TESTS 2723 help 2724 Enable this option to test the bitfield functions at boot. 2725 2726 KUnit tests run during boot and output the results to the debug log 2727 in TAP format (http://testanything.org/). Only useful for kernel devs 2728 running the KUnit test harness, and not intended for inclusion into a 2729 production build. 2730 2731 For more information on KUnit and unit tests in general please refer 2732 to the KUnit documentation in Documentation/dev-tools/kunit/. 2733 2734 If unsure, say N. 2735 2736config CHECKSUM_KUNIT 2737 tristate "KUnit test checksum functions at runtime" if !KUNIT_ALL_TESTS 2738 depends on KUNIT 2739 default KUNIT_ALL_TESTS 2740 help 2741 Enable this option to test the checksum functions at boot. 2742 2743 KUnit tests run during boot and output the results to the debug log 2744 in TAP format (http://testanything.org/). Only useful for kernel devs 2745 running the KUnit test harness, and not intended for inclusion into a 2746 production build. 2747 2748 For more information on KUnit and unit tests in general please refer 2749 to the KUnit documentation in Documentation/dev-tools/kunit/. 2750 2751 If unsure, say N. 2752 2753config UTIL_MACROS_KUNIT 2754 tristate "KUnit test util_macros.h functions at runtime" if !KUNIT_ALL_TESTS 2755 depends on KUNIT 2756 default KUNIT_ALL_TESTS 2757 help 2758 Enable this option to test the util_macros.h function at boot. 2759 2760 KUnit tests run during boot and output the results to the debug log 2761 in TAP format (http://testanything.org/). Only useful for kernel devs 2762 running the KUnit test harness, and not intended for inclusion into a 2763 production build. 2764 2765 For more information on KUnit and unit tests in general please refer 2766 to the KUnit documentation in Documentation/dev-tools/kunit/. 2767 2768 If unsure, say N. 2769 2770config HASH_KUNIT_TEST 2771 tristate "KUnit Test for integer hash functions" if !KUNIT_ALL_TESTS 2772 depends on KUNIT 2773 default KUNIT_ALL_TESTS 2774 help 2775 Enable this option to test the kernel's string (<linux/stringhash.h>), and 2776 integer (<linux/hash.h>) hash functions on boot. 2777 2778 KUnit tests run during boot and output the results to the debug log 2779 in TAP format (https://testanything.org/). Only useful for kernel devs 2780 running the KUnit test harness, and not intended for inclusion into a 2781 production build. 2782 2783 For more information on KUnit and unit tests in general please refer 2784 to the KUnit documentation in Documentation/dev-tools/kunit/. 2785 2786 This is intended to help people writing architecture-specific 2787 optimized versions. If unsure, say N. 2788 2789config RESOURCE_KUNIT_TEST 2790 tristate "KUnit test for resource API" if !KUNIT_ALL_TESTS 2791 depends on KUNIT 2792 default KUNIT_ALL_TESTS 2793 select GET_FREE_REGION 2794 help 2795 This builds the resource API unit test. 2796 Tests the logic of API provided by resource.c and ioport.h. 2797 For more information on KUnit and unit tests in general please refer 2798 to the KUnit documentation in Documentation/dev-tools/kunit/. 2799 2800 If unsure, say N. 2801 2802config SYSCTL_KUNIT_TEST 2803 tristate "KUnit test for sysctl" if !KUNIT_ALL_TESTS 2804 depends on KUNIT 2805 default KUNIT_ALL_TESTS 2806 help 2807 This builds the proc sysctl unit test, which runs on boot. 2808 Tests the API contract and implementation correctness of sysctl. 2809 For more information on KUnit and unit tests in general please refer 2810 to the KUnit documentation in Documentation/dev-tools/kunit/. 2811 2812 If unsure, say N. 2813 2814config KFIFO_KUNIT_TEST 2815 tristate "KUnit Test for the generic kernel FIFO implementation" if !KUNIT_ALL_TESTS 2816 depends on KUNIT 2817 default KUNIT_ALL_TESTS 2818 help 2819 This builds the generic FIFO implementation KUnit test suite. 2820 It tests that the API and basic functionality of the kfifo type 2821 and associated macros. 2822 2823 For more information on KUnit and unit tests in general please refer 2824 to the KUnit documentation in Documentation/dev-tools/kunit/. 2825 2826 If unsure, say N. 2827 2828config LIST_KUNIT_TEST 2829 tristate "KUnit Test for Kernel Linked-list structures" if !KUNIT_ALL_TESTS 2830 depends on KUNIT 2831 default KUNIT_ALL_TESTS 2832 help 2833 This builds the linked list KUnit test suite. 2834 It tests that the API and basic functionality of the list_head type 2835 and associated macros. 2836 2837 KUnit tests run during boot and output the results to the debug log 2838 in TAP format (https://testanything.org/). Only useful for kernel devs 2839 running the KUnit test harness, and not intended for inclusion into a 2840 production build. 2841 2842 For more information on KUnit and unit tests in general please refer 2843 to the KUnit documentation in Documentation/dev-tools/kunit/. 2844 2845 If unsure, say N. 2846 2847config LIST_PRIVATE_KUNIT_TEST 2848 tristate "KUnit Test for Kernel Private Linked-list structures" if !KUNIT_ALL_TESTS 2849 depends on KUNIT 2850 default KUNIT_ALL_TESTS 2851 help 2852 This builds the KUnit test for the private linked-list primitives 2853 defined in include/linux/list_private.h. 2854 2855 These primitives allow manipulation of list_head members that are 2856 marked as private and require special accessors (ACCESS_PRIVATE) 2857 to strip qualifiers or handle encapsulation. 2858 2859 If unsure, say N. 2860 2861config HASHTABLE_KUNIT_TEST 2862 tristate "KUnit Test for Kernel Hashtable structures" if !KUNIT_ALL_TESTS 2863 depends on KUNIT 2864 default KUNIT_ALL_TESTS 2865 help 2866 This builds the hashtable KUnit test suite. 2867 It tests the basic functionality of the API defined in 2868 include/linux/hashtable.h. For more information on KUnit and 2869 unit tests in general please refer to the KUnit documentation 2870 in Documentation/dev-tools/kunit/. 2871 2872 If unsure, say N. 2873 2874config LINEAR_RANGES_TEST 2875 tristate "KUnit test for linear_ranges" 2876 depends on KUNIT 2877 select LINEAR_RANGES 2878 help 2879 This builds the linear_ranges unit test, which runs on boot. 2880 Tests the linear_ranges logic correctness. 2881 For more information on KUnit and unit tests in general please refer 2882 to the KUnit documentation in Documentation/dev-tools/kunit/. 2883 2884 If unsure, say N. 2885 2886config CONTEXT_ANALYSIS_TEST 2887 bool "Compiler context-analysis warnings test" 2888 depends on EXPERT 2889 help 2890 This builds the test for compiler-based context analysis. The test 2891 does not add executable code to the kernel, but is meant to test that 2892 common patterns supported by the analysis do not result in false 2893 positive warnings. 2894 2895 When adding support for new context locks, it is strongly recommended 2896 to add supported patterns to this test. 2897 2898 If unsure, say N. 2899 2900config LIVEUPDATE_TEST 2901 bool "Live Update Kernel Test" 2902 default n 2903 depends on LIVEUPDATE 2904 help 2905 Enable a built-in kernel test module for the Live Update 2906 Orchestrator. 2907 2908 This module validates the File-Lifecycle-Bound subsystem by 2909 registering a set of mock FLB objects with any real file handlers 2910 that support live update (such as the memfd handler). 2911 2912 When live update operations are performed, this test module will 2913 output messages to the kernel log (dmesg), confirming that its 2914 registration and various callback functions (preserve, retrieve, 2915 finish, etc.) are being invoked correctly. 2916 2917 This is a debugging and regression testing tool for developers 2918 working on the Live Update subsystem. It should not be enabled in 2919 production kernels. 2920 2921 If unsure, say N 2922 2923config CMDLINE_KUNIT_TEST 2924 tristate "KUnit test for cmdline API" if !KUNIT_ALL_TESTS 2925 depends on KUNIT 2926 default KUNIT_ALL_TESTS 2927 help 2928 This builds the cmdline API unit test. 2929 Tests the logic of API provided by cmdline.c. 2930 For more information on KUnit and unit tests in general please refer 2931 to the KUnit documentation in Documentation/dev-tools/kunit/. 2932 2933 If unsure, say N. 2934 2935config BASE64_KUNIT 2936 tristate "KUnit test for base64 decoding and encoding" if !KUNIT_ALL_TESTS 2937 depends on KUNIT 2938 default KUNIT_ALL_TESTS 2939 help 2940 This builds the base64 unit tests. 2941 2942 The tests cover the encoding and decoding logic of Base64 functions 2943 in the kernel. 2944 In addition to correctness checks, simple performance benchmarks 2945 for both encoding and decoding are also included. 2946 2947 For more information on KUnit and unit tests in general please refer 2948 to the KUnit documentation in Documentation/dev-tools/kunit/. 2949 2950 If unsure, say N. 2951 2952config BITS_TEST 2953 tristate "KUnit test for bit functions and macros" if !KUNIT_ALL_TESTS 2954 depends on KUNIT 2955 default KUNIT_ALL_TESTS 2956 help 2957 This builds the bits unit test. 2958 Tests the logic of macros defined in bits.h. 2959 For more information on KUnit and unit tests in general please refer 2960 to the KUnit documentation in Documentation/dev-tools/kunit/. 2961 2962 If unsure, say N. 2963 2964config SHDI3_KUNIT_TEST 2965 tristate "KUnit test for __ashldi3(), __ashrdi3(), and __lshrdi3()" 2966 depends on KUNIT 2967 depends on ARM || XTENSA || MICROBLAZE || ((RISCV || SPARC) && !64BIT) 2968 help 2969 This builds the unit test for __ashldi3(), __ashrdi3(), and 2970 __lshrdi3() helper functions used to implement 64-bit arithmetic 2971 shift left, arithmetic shift right and logical shift right, 2972 respectively, on a 32-bit CPUs. 2973 2974config SLUB_KUNIT_TEST 2975 tristate "KUnit test for SLUB cache error detection" if !KUNIT_ALL_TESTS 2976 depends on SLUB_DEBUG && KUNIT 2977 default KUNIT_ALL_TESTS 2978 help 2979 This builds SLUB allocator unit test. 2980 Tests SLUB cache debugging functionality. 2981 For more information on KUnit and unit tests in general please refer 2982 to the KUnit documentation in Documentation/dev-tools/kunit/. 2983 2984 If unsure, say N. 2985 2986config RATIONAL_KUNIT_TEST 2987 tristate "KUnit test for rational.c" if !KUNIT_ALL_TESTS 2988 depends on KUNIT && RATIONAL 2989 default KUNIT_ALL_TESTS 2990 help 2991 This builds the rational math unit test. 2992 For more information on KUnit and unit tests in general please refer 2993 to the KUnit documentation in Documentation/dev-tools/kunit/. 2994 2995 If unsure, say N. 2996 2997config MEMCPY_KUNIT_TEST 2998 tristate "Test memcpy(), memmove(), and memset() functions at runtime" if !KUNIT_ALL_TESTS 2999 depends on KUNIT 3000 default KUNIT_ALL_TESTS 3001 help 3002 Builds unit tests for memcpy(), memmove(), and memset() functions. 3003 For more information on KUnit and unit tests in general please refer 3004 to the KUnit documentation in Documentation/dev-tools/kunit/. 3005 3006 If unsure, say N. 3007 3008config MIN_HEAP_KUNIT_TEST 3009 tristate "Min heap test" if !KUNIT_ALL_TESTS 3010 depends on KUNIT 3011 default KUNIT_ALL_TESTS 3012 help 3013 This option enables the KUnit test suite for the min heap library 3014 which provides functions for creating and managing min heaps. 3015 The test suite checks the functionality of the min heap library. 3016 3017 If unsure, say N 3018 3019config IS_SIGNED_TYPE_KUNIT_TEST 3020 tristate "Test is_signed_type() macro" if !KUNIT_ALL_TESTS 3021 depends on KUNIT 3022 default KUNIT_ALL_TESTS 3023 help 3024 Builds unit tests for the is_signed_type() macro. 3025 3026 For more information on KUnit and unit tests in general please refer 3027 to the KUnit documentation in Documentation/dev-tools/kunit/. 3028 3029 If unsure, say N. 3030 3031config OVERFLOW_KUNIT_TEST 3032 tristate "Test check_*_overflow() functions at runtime" if !KUNIT_ALL_TESTS 3033 depends on KUNIT 3034 default KUNIT_ALL_TESTS 3035 help 3036 Builds unit tests for the check_*_overflow(), size_*(), allocation, and 3037 related functions. 3038 3039 For more information on KUnit and unit tests in general please refer 3040 to the KUnit documentation in Documentation/dev-tools/kunit/. 3041 3042 If unsure, say N. 3043 3044config RANDSTRUCT_KUNIT_TEST 3045 tristate "Test randstruct structure layout randomization at runtime" if !KUNIT_ALL_TESTS 3046 depends on KUNIT 3047 default KUNIT_ALL_TESTS 3048 help 3049 Builds unit tests for the checking CONFIG_RANDSTRUCT=y, which 3050 randomizes structure layouts. 3051 3052config STACKINIT_KUNIT_TEST 3053 tristate "Test level of stack variable initialization" if !KUNIT_ALL_TESTS 3054 depends on KUNIT 3055 default KUNIT_ALL_TESTS 3056 help 3057 Test if the kernel is zero-initializing stack variables and 3058 padding. Coverage is controlled by compiler flags, 3059 CONFIG_INIT_STACK_ALL_PATTERN or CONFIG_INIT_STACK_ALL_ZERO. 3060 3061config FORTIFY_KUNIT_TEST 3062 tristate "Test fortified str*() and mem*() function internals at runtime" if !KUNIT_ALL_TESTS 3063 depends on KUNIT 3064 default KUNIT_ALL_TESTS 3065 help 3066 Builds unit tests for checking internals of FORTIFY_SOURCE as used 3067 by the str*() and mem*() family of functions. For testing runtime 3068 traps of FORTIFY_SOURCE, see LKDTM's "FORTIFY_*" tests. 3069 3070config LONGEST_SYM_KUNIT_TEST 3071 tristate "Test the longest symbol possible" if !KUNIT_ALL_TESTS 3072 depends on KUNIT && KPROBES 3073 depends on !CALL_PADDING && !CFI && !GCOV_KERNEL 3074 default KUNIT_ALL_TESTS 3075 help 3076 Tests the longest symbol possible 3077 3078 If unsure, say N. 3079 3080config HW_BREAKPOINT_KUNIT_TEST 3081 bool "Test hw_breakpoint constraints accounting" if !KUNIT_ALL_TESTS 3082 depends on HAVE_HW_BREAKPOINT 3083 depends on KUNIT=y 3084 default KUNIT_ALL_TESTS 3085 help 3086 Tests for hw_breakpoint constraints accounting. 3087 3088 If unsure, say N. 3089 3090source "lib/crypto/tests/Kconfig" 3091 3092config SIPHASH_KUNIT_TEST 3093 tristate "Perform selftest on siphash functions" if !KUNIT_ALL_TESTS 3094 depends on KUNIT 3095 default KUNIT_ALL_TESTS 3096 help 3097 Enable this option to test the kernel's siphash (<linux/siphash.h>) hash 3098 functions on boot (or module load). 3099 3100 This is intended to help people writing architecture-specific 3101 optimized versions. If unsure, say N. 3102 3103config USERCOPY_KUNIT_TEST 3104 tristate "KUnit Test for user/kernel boundary protections" 3105 depends on KUNIT 3106 default KUNIT_ALL_TESTS 3107 help 3108 This builds the "usercopy_kunit" module that runs sanity checks 3109 on the copy_to/from_user infrastructure, making sure basic 3110 user/kernel boundary testing is working. 3111 3112config BLACKHOLE_DEV_KUNIT_TEST 3113 tristate "Test blackhole netdev functionality" if !KUNIT_ALL_TESTS 3114 depends on NET 3115 depends on KUNIT 3116 default KUNIT_ALL_TESTS 3117 help 3118 This builds the "blackhole_dev_kunit" module that validates the 3119 data path through this blackhole netdev. 3120 3121 If unsure, say N. 3122 3123config TEST_UDELAY 3124 tristate "udelay test driver" 3125 help 3126 This builds the "udelay_test" module that helps to make sure 3127 that udelay() is working properly. 3128 3129 If unsure, say N. 3130 3131config TEST_STATIC_KEYS 3132 tristate "Test static keys" 3133 depends on m 3134 help 3135 Test the static key interfaces. 3136 3137 If unsure, say N. 3138 3139config TEST_DYNAMIC_DEBUG 3140 tristate "Test DYNAMIC_DEBUG" 3141 depends on DYNAMIC_DEBUG 3142 help 3143 This module registers a tracer callback to count enabled 3144 pr_debugs in a 'do_debugging' function, then alters their 3145 enablements, calls the function, and compares counts. 3146 3147 If unsure, say N. 3148 3149config TEST_KMOD 3150 tristate "kmod stress tester" 3151 depends on m 3152 select TEST_LKM 3153 help 3154 Test the kernel's module loading mechanism: kmod. kmod implements 3155 support to load modules using the Linux kernel's usermode helper. 3156 This test provides a series of tests against kmod. 3157 3158 Although technically you can either build test_kmod as a module or 3159 into the kernel we disallow building it into the kernel since 3160 it stress tests request_module() and this will very likely cause 3161 some issues by taking over precious threads available from other 3162 module load requests, ultimately this could be fatal. 3163 3164 To run tests run: 3165 3166 tools/testing/selftests/kmod/kmod.sh --help 3167 3168 If unsure, say N. 3169 3170config TEST_RUNTIME 3171 bool 3172 3173config TEST_RUNTIME_MODULE 3174 bool 3175 3176config TEST_KALLSYMS 3177 tristate "module kallsyms find_symbol() test" 3178 depends on m 3179 select TEST_RUNTIME 3180 select TEST_RUNTIME_MODULE 3181 select TEST_KALLSYMS_A 3182 select TEST_KALLSYMS_B 3183 select TEST_KALLSYMS_C 3184 select TEST_KALLSYMS_D 3185 help 3186 This allows us to stress test find_symbol() through the kallsyms 3187 used to place symbols on the kernel ELF kallsyms and modules kallsyms 3188 where we place kernel symbols such as exported symbols. 3189 3190 We have four test modules: 3191 3192 A: has KALLSYSMS_NUMSYMS exported symbols 3193 B: uses one of A's symbols 3194 C: adds KALLSYMS_SCALE_FACTOR * KALLSYSMS_NUMSYMS exported 3195 D: adds 2 * the symbols than C 3196 3197 We stress test find_symbol() through two means: 3198 3199 1) Upon load of B it will trigger simplify_symbols() to look for the 3200 one symbol it uses from the module A with tons of symbols. This is an 3201 indirect way for us to have B call resolve_symbol_wait() upon module 3202 load. This will eventually call find_symbol() which will eventually 3203 try to find the symbols used with find_exported_symbol_in_section(). 3204 find_exported_symbol_in_section() uses bsearch() so a binary search 3205 for each symbol. Binary search will at worst be O(log(n)) so the 3206 larger TEST_MODULE_KALLSYSMS the worse the search. 3207 3208 2) The selftests should load C first, before B. Upon B's load towards 3209 the end right before we call module B's init routine we get 3210 complete_formation() called on the module. That will first check 3211 for duplicate symbols with the call to verify_exported_symbols(). 3212 That is when we'll force iteration on module C's insane symbol list. 3213 Since it has 10 * KALLSYMS_NUMSYMS it means we can first test 3214 just loading B without C. The amount of time it takes to load C Vs 3215 B can give us an idea of the impact growth of the symbol space and 3216 give us projection. Module A only uses one symbol from B so to allow 3217 this scaling in module C to be proportional, if it used more symbols 3218 then the first test would be doing more and increasing just the 3219 search space would be slightly different. The last module, module D 3220 will just increase the search space by twice the number of symbols in 3221 C so to allow for full projects. 3222 3223 tools/testing/selftests/module/find_symbol.sh 3224 3225 The current defaults will incur a build delay of about 7 minutes 3226 on an x86_64 with only 8 cores. Enable this only if you want to 3227 stress test find_symbol() with thousands of symbols. At the same 3228 time this is also useful to test building modules with thousands of 3229 symbols, and if BTF is enabled this also stress tests adding BTF 3230 information for each module. Currently enabling many more symbols 3231 will segfault the build system. 3232 3233 If unsure, say N. 3234 3235if TEST_KALLSYMS 3236 3237config TEST_KALLSYMS_A 3238 tristate 3239 depends on m 3240 3241config TEST_KALLSYMS_B 3242 tristate 3243 depends on m 3244 3245config TEST_KALLSYMS_C 3246 tristate 3247 depends on m 3248 3249config TEST_KALLSYMS_D 3250 tristate 3251 depends on m 3252 3253choice 3254 prompt "Kallsym test range" 3255 default TEST_KALLSYMS_LARGE 3256 help 3257 Selecting something other than "Fast" will enable tests which slow 3258 down the build and may crash your build. 3259 3260config TEST_KALLSYMS_FAST 3261 bool "Fast builds" 3262 help 3263 You won't really be testing kallsysms, so this just helps fast builds 3264 when allmodconfig is used.. 3265 3266config TEST_KALLSYMS_LARGE 3267 bool "Enable testing kallsyms with large exports" 3268 help 3269 This will enable larger number of symbols. This will slow down 3270 your build considerably. 3271 3272config TEST_KALLSYMS_MAX 3273 bool "Known kallsysms limits" 3274 help 3275 This will enable exports to the point we know we'll start crashing 3276 builds. 3277 3278endchoice 3279 3280config TEST_KALLSYMS_NUMSYMS 3281 int "test kallsyms number of symbols" 3282 range 2 10000 3283 default 2 if TEST_KALLSYMS_FAST 3284 default 100 if TEST_KALLSYMS_LARGE 3285 default 10000 if TEST_KALLSYMS_MAX 3286 help 3287 The number of symbols to create on TEST_KALLSYMS_A, only one of which 3288 module TEST_KALLSYMS_B will use. This also will be used 3289 for how many symbols TEST_KALLSYMS_C will have, scaled up by 3290 TEST_KALLSYMS_SCALE_FACTOR. Note that setting this to 10,000 will 3291 trigger a segfault today, don't use anything close to it unless 3292 you are aware that this should not be used for automated build tests. 3293 3294config TEST_KALLSYMS_SCALE_FACTOR 3295 int "test kallsyms scale factor" 3296 default 8 3297 help 3298 How many more unusued symbols will TEST_KALLSYSMS_C have than 3299 TEST_KALLSYMS_A. If 8, then module C will have 8 * syms 3300 than module A. Then TEST_KALLSYMS_D will have double the amount 3301 of symbols than C so to allow projections. 3302 3303endif # TEST_KALLSYMS 3304 3305config TEST_DEBUG_VIRTUAL 3306 tristate "Test CONFIG_DEBUG_VIRTUAL feature" 3307 depends on DEBUG_VIRTUAL 3308 help 3309 Test the kernel's ability to detect incorrect calls to 3310 virt_to_phys() done against the non-linear part of the 3311 kernel's virtual address map. 3312 3313 If unsure, say N. 3314 3315config TEST_MEMCAT_P 3316 tristate "Test memcat_p() helper function" 3317 help 3318 Test the memcat_p() helper for correctly merging two 3319 pointer arrays together. 3320 3321 If unsure, say N. 3322 3323config TEST_OBJAGG 3324 tristate "Perform selftest on object aggreration manager" 3325 default n 3326 depends on OBJAGG 3327 help 3328 Enable this option to test object aggregation manager on boot 3329 (or module load). 3330 3331config TEST_MEMINIT 3332 tristate "Test heap/page initialization" 3333 help 3334 Test if the kernel is zero-initializing heap and page allocations. 3335 This can be useful to test init_on_alloc and init_on_free features. 3336 3337 If unsure, say N. 3338 3339config TEST_HMM 3340 tristate "Test HMM (Heterogeneous Memory Management)" 3341 depends on TRANSPARENT_HUGEPAGE 3342 depends on DEVICE_PRIVATE 3343 select HMM_MIRROR 3344 select MMU_NOTIFIER 3345 help 3346 This is a pseudo device driver solely for testing HMM. 3347 Say M here if you want to build the HMM test module. 3348 Doing so will allow you to run tools/testing/selftest/vm/hmm-tests. 3349 3350 If unsure, say N. 3351 3352config TEST_FREE_PAGES 3353 tristate "Test freeing pages" 3354 help 3355 Test that a memory leak does not occur due to a race between 3356 freeing a block of pages and a speculative page reference. 3357 Loading this module is safe if your kernel has the bug fixed. 3358 If the bug is not fixed, it will leak gigabytes of memory and 3359 probably OOM your system. 3360 3361config TEST_FPU 3362 tristate "Test floating point operations in kernel space" 3363 depends on ARCH_HAS_KERNEL_FPU_SUPPORT && !KCOV_INSTRUMENT_ALL 3364 help 3365 Enable this option to add /sys/kernel/debug/selftest_helpers/test_fpu 3366 which will trigger a sequence of floating point operations. This is used 3367 for self-testing floating point control register setting in 3368 kernel_fpu_begin(). 3369 3370 If unsure, say N. 3371 3372config TEST_CLOCKSOURCE_WATCHDOG 3373 tristate "Test clocksource watchdog in kernel space" 3374 depends on CLOCKSOURCE_WATCHDOG 3375 help 3376 Enable this option to create a kernel module that will trigger 3377 a test of the clocksource watchdog. This module may be loaded 3378 via modprobe or insmod in which case it will run upon being 3379 loaded, or it may be built in, in which case it will run 3380 shortly after boot. 3381 3382 If unsure, say N. 3383 3384config TEST_OBJPOOL 3385 tristate "Test module for correctness and stress of objpool" 3386 default n 3387 depends on m && DEBUG_KERNEL 3388 help 3389 This builds the "test_objpool" module that should be used for 3390 correctness verification and concurrent testings of objects 3391 allocation and reclamation. 3392 3393 If unsure, say N. 3394 3395config TEST_KEXEC_HANDOVER 3396 bool "Test for Kexec HandOver" 3397 default n 3398 depends on KEXEC_HANDOVER 3399 help 3400 This option enables test for Kexec HandOver (KHO). 3401 The test consists of two parts: saving kernel data before kexec and 3402 restoring the data after kexec and verifying that it was properly 3403 handed over. This test module creates and saves data on the boot of 3404 the first kernel and restores and verifies the data on the boot of 3405 kexec'ed kernel. 3406 3407 For detailed documentation about KHO, see Documentation/core-api/kho. 3408 3409 To run the test run: 3410 3411 tools/testing/selftests/kho/vmtest.sh -h 3412 3413 If unsure, say N. 3414 3415config RATELIMIT_KUNIT_TEST 3416 tristate "KUnit Test for correctness and stress of ratelimit" if !KUNIT_ALL_TESTS 3417 depends on KUNIT 3418 default KUNIT_ALL_TESTS 3419 help 3420 This builds the "test_ratelimit" module that should be used 3421 for correctness verification and concurrent testings of rate 3422 limiting. 3423 3424 If unsure, say N. 3425 3426config UUID_KUNIT_TEST 3427 tristate "KUnit test for UUID" if !KUNIT_ALL_TESTS 3428 depends on KUNIT 3429 default KUNIT_ALL_TESTS 3430 help 3431 This option enables the KUnit test suite for the uuid library, 3432 which provides functions for generating and parsing UUID and GUID. 3433 The test suite checks parsing of UUID and GUID strings. 3434 3435 If unsure, say N. 3436 3437config INT_POW_KUNIT_TEST 3438 tristate "Integer exponentiation (int_pow) test" if !KUNIT_ALL_TESTS 3439 depends on KUNIT 3440 default KUNIT_ALL_TESTS 3441 help 3442 This option enables the KUnit test suite for the int_pow function, 3443 which performs integer exponentiation. The test suite is designed to 3444 verify that the implementation of int_pow correctly computes the power 3445 of a given base raised to a given exponent. 3446 3447 Enabling this option will include tests that check various scenarios 3448 and edge cases to ensure the accuracy and reliability of the exponentiation 3449 function. 3450 3451 If unsure, say N 3452 3453config INT_SQRT_KUNIT_TEST 3454 tristate "Integer square root test" if !KUNIT_ALL_TESTS 3455 depends on KUNIT 3456 default KUNIT_ALL_TESTS 3457 help 3458 This option enables the KUnit test suite for the int_sqrt() function, 3459 which performs square root calculation. The test suite checks 3460 various scenarios, including edge cases, to ensure correctness. 3461 3462 Enabling this option will include tests that check various scenarios 3463 and edge cases to ensure the accuracy and reliability of the square root 3464 function. 3465 3466 If unsure, say N 3467 3468config INT_LOG_KUNIT_TEST 3469 tristate "Integer log (int_log) test" if !KUNIT_ALL_TESTS 3470 depends on KUNIT 3471 default KUNIT_ALL_TESTS 3472 help 3473 This option enables the KUnit test suite for the int_log library, which 3474 provides two functions to compute the integer logarithm in base 2 and 3475 base 10, called respectively as intlog2 and intlog10. 3476 3477 If unsure, say N 3478 3479config GCD_KUNIT_TEST 3480 tristate "Greatest common divisor test" if !KUNIT_ALL_TESTS 3481 depends on KUNIT 3482 default KUNIT_ALL_TESTS 3483 help 3484 This option enables the KUnit test suite for the gcd() function, 3485 which computes the greatest common divisor of two numbers. 3486 3487 This test suite verifies the correctness of gcd() across various 3488 scenarios, including edge cases. 3489 3490 If unsure, say N 3491 3492config PRIME_NUMBERS_KUNIT_TEST 3493 tristate "Prime number generator test" if !KUNIT_ALL_TESTS 3494 depends on KUNIT 3495 depends on PRIME_NUMBERS 3496 default KUNIT_ALL_TESTS 3497 help 3498 This option enables the KUnit test suite for the {is,next}_prime_number 3499 functions. 3500 3501 Enabling this option will include tests that compare the prime number 3502 generator functions against a brute force implementation. 3503 3504 If unsure, say N 3505 3506config GLOB_KUNIT_TEST 3507 tristate "Glob matching test" if !KUNIT_ALL_TESTS 3508 depends on GLOB 3509 depends on KUNIT 3510 default KUNIT_ALL_TESTS 3511 help 3512 Enable this option to test the glob functions at runtime. 3513 3514 This test suite verifies the correctness of glob_match() across various 3515 scenarios, including edge cases. 3516 3517 If unsure, say N 3518 3519endif # RUNTIME_TESTING_MENU 3520 3521config ARCH_USE_MEMTEST 3522 bool 3523 help 3524 An architecture should select this when it uses early_memtest() 3525 during boot process. 3526 3527config MEMTEST 3528 bool "Memtest" 3529 depends on ARCH_USE_MEMTEST 3530 help 3531 This option adds a kernel parameter 'memtest', which allows memtest 3532 to be set and executed. 3533 memtest=0, mean disabled; -- default 3534 memtest=1, mean do 1 test pattern; 3535 ... 3536 memtest=17, mean do 17 test patterns. 3537 If you are unsure how to answer this question, answer N. 3538 3539 3540 3541config HYPERV_TESTING 3542 bool "Microsoft Hyper-V driver testing" 3543 default n 3544 depends on HYPERV && DEBUG_FS 3545 help 3546 Select this option to enable Hyper-V vmbus testing. 3547 3548endmenu # "Kernel Testing and Coverage" 3549 3550menu "Rust hacking" 3551 3552config RUST_DEBUG_ASSERTIONS 3553 bool "Debug assertions" 3554 depends on RUST 3555 help 3556 Enables rustc's `-Cdebug-assertions` codegen option. 3557 3558 This flag lets you turn `cfg(debug_assertions)` conditional 3559 compilation on or off. This can be used to enable extra debugging 3560 code in development but not in production. For example, it controls 3561 the behavior of the standard library's `debug_assert!` macro. 3562 3563 Note that this will apply to all Rust code, including `core`. 3564 3565 If unsure, say N. 3566 3567config RUST_OVERFLOW_CHECKS 3568 bool "Overflow checks" 3569 default y 3570 depends on RUST 3571 help 3572 Enables rustc's `-Coverflow-checks` codegen option. 3573 3574 This flag allows you to control the behavior of runtime integer 3575 overflow. When overflow-checks are enabled, a Rust panic will occur 3576 on overflow. 3577 3578 Note that this will apply to all Rust code, including `core`. 3579 3580 If unsure, say Y. 3581 3582config RUST_BUILD_ASSERT_ALLOW 3583 bool "Allow unoptimized build-time assertions" 3584 depends on RUST 3585 help 3586 Controls how `build_error!` and `build_assert!` are handled during the build. 3587 3588 If calls to them exist in the binary, it may indicate a violated invariant 3589 or that the optimizer failed to verify the invariant during compilation. 3590 3591 This should not happen, thus by default the build is aborted. However, 3592 as an escape hatch, you can choose Y here to ignore them during build 3593 and let the check be carried at runtime (with `panic!` being called if 3594 the check fails). 3595 3596 If unsure, say N. 3597 3598config RUST_KERNEL_DOCTESTS 3599 bool "Doctests for the `kernel` crate" if !KUNIT_ALL_TESTS 3600 depends on RUST && KUNIT=y 3601 default KUNIT_ALL_TESTS 3602 help 3603 This builds the documentation tests of the `kernel` crate 3604 as KUnit tests. 3605 3606 For more information on KUnit and unit tests in general, 3607 please refer to the KUnit documentation in Documentation/dev-tools/kunit/. 3608 3609 If unsure, say N. 3610 3611config RUST_INLINE_HELPERS 3612 bool "Inline C helpers into Rust code (EXPERIMENTAL)" 3613 depends on RUST && RUSTC_CLANG_LLVM_COMPATIBLE 3614 depends on EXPERT 3615 depends on ARM64 || X86_64 3616 depends on !UML 3617 help 3618 Inlines C helpers into Rust code using Link Time Optimization. 3619 3620 If this option is enabled, C helper functions declared in 3621 rust/helpers/ are inlined into Rust code, which is helpful for 3622 performance of Rust code. This requires a matching LLVM version for 3623 Clang and rustc. 3624 3625 If you are sure that you're using Clang and rustc with matching LLVM 3626 versions, say Y. Otherwise say N. 3627 3628endmenu # "Rust" 3629 3630endmenu # Kernel hacking 3631