1menu "printk and dmesg options" 2 3config PRINTK_TIME 4 bool "Show timing information on printks" 5 depends on PRINTK 6 help 7 Selecting this option causes time stamps of the printk() 8 messages to be added to the output of the syslog() system 9 call and at the console. 10 11 The timestamp is always recorded internally, and exported 12 to /dev/kmsg. This flag just specifies if the timestamp should 13 be included, not that the timestamp is recorded. 14 15 The behavior is also controlled by the kernel command line 16 parameter printk.time=1. See Documentation/kernel-parameters.txt 17 18config MESSAGE_LOGLEVEL_DEFAULT 19 int "Default message log level (1-7)" 20 range 1 7 21 default "4" 22 help 23 Default log level for printk statements with no specified priority. 24 25 This was hard-coded to KERN_WARNING since at least 2.6.10 but folks 26 that are auditing their logs closely may want to set it to a lower 27 priority. 28 29config BOOT_PRINTK_DELAY 30 bool "Delay each boot printk message by N milliseconds" 31 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY 32 help 33 This build option allows you to read kernel boot messages 34 by inserting a short delay after each one. The delay is 35 specified in milliseconds on the kernel command line, 36 using "boot_delay=N". 37 38 It is likely that you would also need to use "lpj=M" to preset 39 the "loops per jiffie" value. 40 See a previous boot log for the "lpj" value to use for your 41 system, and then set "lpj=M" before setting "boot_delay=N". 42 NOTE: Using this option may adversely affect SMP systems. 43 I.e., processors other than the first one may not boot up. 44 BOOT_PRINTK_DELAY also may cause LOCKUP_DETECTOR to detect 45 what it believes to be lockup conditions. 46 47config DYNAMIC_DEBUG 48 bool "Enable dynamic printk() support" 49 default n 50 depends on PRINTK 51 depends on DEBUG_FS 52 help 53 54 Compiles debug level messages into the kernel, which would not 55 otherwise be available at runtime. These messages can then be 56 enabled/disabled based on various levels of scope - per source file, 57 function, module, format string, and line number. This mechanism 58 implicitly compiles in all pr_debug() and dev_dbg() calls, which 59 enlarges the kernel text size by about 2%. 60 61 If a source file is compiled with DEBUG flag set, any 62 pr_debug() calls in it are enabled by default, but can be 63 disabled at runtime as below. Note that DEBUG flag is 64 turned on by many CONFIG_*DEBUG* options. 65 66 Usage: 67 68 Dynamic debugging is controlled via the 'dynamic_debug/control' file, 69 which is contained in the 'debugfs' filesystem. Thus, the debugfs 70 filesystem must first be mounted before making use of this feature. 71 We refer the control file as: <debugfs>/dynamic_debug/control. This 72 file contains a list of the debug statements that can be enabled. The 73 format for each line of the file is: 74 75 filename:lineno [module]function flags format 76 77 filename : source file of the debug statement 78 lineno : line number of the debug statement 79 module : module that contains the debug statement 80 function : function that contains the debug statement 81 flags : '=p' means the line is turned 'on' for printing 82 format : the format used for the debug statement 83 84 From a live system: 85 86 nullarbor:~ # cat <debugfs>/dynamic_debug/control 87 # filename:lineno [module]function flags format 88 fs/aio.c:222 [aio]__put_ioctx =_ "__put_ioctx:\040freeing\040%p\012" 89 fs/aio.c:248 [aio]ioctx_alloc =_ "ENOMEM:\040nr_events\040too\040high\012" 90 fs/aio.c:1770 [aio]sys_io_cancel =_ "calling\040cancel\012" 91 92 Example usage: 93 94 // enable the message at line 1603 of file svcsock.c 95 nullarbor:~ # echo -n 'file svcsock.c line 1603 +p' > 96 <debugfs>/dynamic_debug/control 97 98 // enable all the messages in file svcsock.c 99 nullarbor:~ # echo -n 'file svcsock.c +p' > 100 <debugfs>/dynamic_debug/control 101 102 // enable all the messages in the NFS server module 103 nullarbor:~ # echo -n 'module nfsd +p' > 104 <debugfs>/dynamic_debug/control 105 106 // enable all 12 messages in the function svc_process() 107 nullarbor:~ # echo -n 'func svc_process +p' > 108 <debugfs>/dynamic_debug/control 109 110 // disable all 12 messages in the function svc_process() 111 nullarbor:~ # echo -n 'func svc_process -p' > 112 <debugfs>/dynamic_debug/control 113 114 See Documentation/dynamic-debug-howto.txt for additional information. 115 116endmenu # "printk and dmesg options" 117 118menu "Compile-time checks and compiler options" 119 120config DEBUG_INFO 121 bool "Compile the kernel with debug info" 122 depends on DEBUG_KERNEL && !COMPILE_TEST 123 help 124 If you say Y here the resulting kernel image will include 125 debugging info resulting in a larger kernel image. 126 This adds debug symbols to the kernel and modules (gcc -g), and 127 is needed if you intend to use kernel crashdump or binary object 128 tools like crash, kgdb, LKCD, gdb, etc on the kernel. 129 Say Y here only if you plan to debug the kernel. 130 131 If unsure, say N. 132 133config DEBUG_INFO_REDUCED 134 bool "Reduce debugging information" 135 depends on DEBUG_INFO 136 help 137 If you say Y here gcc is instructed to generate less debugging 138 information for structure types. This means that tools that 139 need full debugging information (like kgdb or systemtap) won't 140 be happy. But if you merely need debugging information to 141 resolve line numbers there is no loss. Advantage is that 142 build directory object sizes shrink dramatically over a full 143 DEBUG_INFO build and compile times are reduced too. 144 Only works with newer gcc versions. 145 146config DEBUG_INFO_SPLIT 147 bool "Produce split debuginfo in .dwo files" 148 depends on DEBUG_INFO 149 help 150 Generate debug info into separate .dwo files. This significantly 151 reduces the build directory size for builds with DEBUG_INFO, 152 because it stores the information only once on disk in .dwo 153 files instead of multiple times in object files and executables. 154 In addition the debug information is also compressed. 155 156 Requires recent gcc (4.7+) and recent gdb/binutils. 157 Any tool that packages or reads debug information would need 158 to know about the .dwo files and include them. 159 Incompatible with older versions of ccache. 160 161config DEBUG_INFO_DWARF4 162 bool "Generate dwarf4 debuginfo" 163 depends on DEBUG_INFO 164 help 165 Generate dwarf4 debug info. This requires recent versions 166 of gcc and gdb. It makes the debug information larger. 167 But it significantly improves the success of resolving 168 variables in gdb on optimized code. 169 170config GDB_SCRIPTS 171 bool "Provide GDB scripts for kernel debugging" 172 depends on DEBUG_INFO 173 help 174 This creates the required links to GDB helper scripts in the 175 build directory. If you load vmlinux into gdb, the helper 176 scripts will be automatically imported by gdb as well, and 177 additional functions are available to analyze a Linux kernel 178 instance. See Documentation/gdb-kernel-debugging.txt for further 179 details. 180 181config ENABLE_WARN_DEPRECATED 182 bool "Enable __deprecated logic" 183 default y 184 help 185 Enable the __deprecated logic in the kernel build. 186 Disable this to suppress the "warning: 'foo' is deprecated 187 (declared at kernel/power/somefile.c:1234)" messages. 188 189config ENABLE_MUST_CHECK 190 bool "Enable __must_check logic" 191 default y 192 help 193 Enable the __must_check logic in the kernel build. Disable this to 194 suppress the "warning: ignoring return value of 'foo', declared with 195 attribute warn_unused_result" messages. 196 197config FRAME_WARN 198 int "Warn for stack frames larger than (needs gcc 4.4)" 199 range 0 8192 200 default 0 if KASAN 201 default 1024 if !64BIT 202 default 2048 if 64BIT 203 help 204 Tell gcc to warn at build time for stack frames larger than this. 205 Setting this too low will cause a lot of warnings. 206 Setting it to 0 disables the warning. 207 Requires gcc 4.4 208 209config STRIP_ASM_SYMS 210 bool "Strip assembler-generated symbols during link" 211 default n 212 help 213 Strip internal assembler-generated symbols during a link (symbols 214 that look like '.Lxxx') so they don't pollute the output of 215 get_wchan() and suchlike. 216 217config READABLE_ASM 218 bool "Generate readable assembler code" 219 depends on DEBUG_KERNEL 220 help 221 Disable some compiler optimizations that tend to generate human unreadable 222 assembler output. This may make the kernel slightly slower, but it helps 223 to keep kernel developers who have to stare a lot at assembler listings 224 sane. 225 226config UNUSED_SYMBOLS 227 bool "Enable unused/obsolete exported symbols" 228 default y if X86 229 help 230 Unused but exported symbols make the kernel needlessly bigger. For 231 that reason most of these unused exports will soon be removed. This 232 option is provided temporarily to provide a transition period in case 233 some external kernel module needs one of these symbols anyway. If you 234 encounter such a case in your module, consider if you are actually 235 using the right API. (rationale: since nobody in the kernel is using 236 this in a module, there is a pretty good chance it's actually the 237 wrong interface to use). If you really need the symbol, please send a 238 mail to the linux kernel mailing list mentioning the symbol and why 239 you really need it, and what the merge plan to the mainline kernel for 240 your module is. 241 242config PAGE_OWNER 243 bool "Track page owner" 244 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT 245 select DEBUG_FS 246 select STACKTRACE 247 select PAGE_EXTENSION 248 help 249 This keeps track of what call chain is the owner of a page, may 250 help to find bare alloc_page(s) leaks. Even if you include this 251 feature on your build, it is disabled in default. You should pass 252 "page_owner=on" to boot parameter in order to enable it. Eats 253 a fair amount of memory if enabled. See tools/vm/page_owner_sort.c 254 for user-space helper. 255 256 If unsure, say N. 257 258config DEBUG_FS 259 bool "Debug Filesystem" 260 help 261 debugfs is a virtual file system that kernel developers use to put 262 debugging files into. Enable this option to be able to read and 263 write to these files. 264 265 For detailed documentation on the debugfs API, see 266 Documentation/DocBook/filesystems. 267 268 If unsure, say N. 269 270config HEADERS_CHECK 271 bool "Run 'make headers_check' when building vmlinux" 272 depends on !UML 273 help 274 This option will extract the user-visible kernel headers whenever 275 building the kernel, and will run basic sanity checks on them to 276 ensure that exported files do not attempt to include files which 277 were not exported, etc. 278 279 If you're making modifications to header files which are 280 relevant for userspace, say 'Y', and check the headers 281 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in 282 your build tree), to make sure they're suitable. 283 284config DEBUG_SECTION_MISMATCH 285 bool "Enable full Section mismatch analysis" 286 help 287 The section mismatch analysis checks if there are illegal 288 references from one section to another section. 289 During linktime or runtime, some sections are dropped; 290 any use of code/data previously in these sections would 291 most likely result in an oops. 292 In the code, functions and variables are annotated with 293 __init,, etc. (see the full list in include/linux/init.h), 294 which results in the code/data being placed in specific sections. 295 The section mismatch analysis is always performed after a full 296 kernel build, and enabling this option causes the following 297 additional steps to occur: 298 - Add the option -fno-inline-functions-called-once to gcc commands. 299 When inlining a function annotated with __init in a non-init 300 function, we would lose the section information and thus 301 the analysis would not catch the illegal reference. 302 This option tells gcc to inline less (but it does result in 303 a larger kernel). 304 - Run the section mismatch analysis for each module/built-in.o file. 305 When we run the section mismatch analysis on vmlinux.o, we 306 lose valueble information about where the mismatch was 307 introduced. 308 Running the analysis for each module/built-in.o file 309 tells where the mismatch happens much closer to the 310 source. The drawback is that the same mismatch is 311 reported at least twice. 312 - Enable verbose reporting from modpost in order to help resolve 313 the section mismatches that are reported. 314 315config SECTION_MISMATCH_WARN_ONLY 316 bool "Make section mismatch errors non-fatal" 317 default y 318 help 319 If you say N here, the build process will fail if there are any 320 section mismatch, instead of just throwing warnings. 321 322 If unsure, say Y. 323 324# 325# Select this config option from the architecture Kconfig, if it 326# is preferred to always offer frame pointers as a config 327# option on the architecture (regardless of KERNEL_DEBUG): 328# 329config ARCH_WANT_FRAME_POINTERS 330 bool 331 help 332 333config FRAME_POINTER 334 bool "Compile the kernel with frame pointers" 335 depends on DEBUG_KERNEL && \ 336 (CRIS || M68K || FRV || UML || \ 337 AVR32 || SUPERH || BLACKFIN || MN10300 || METAG) || \ 338 ARCH_WANT_FRAME_POINTERS 339 default y if (DEBUG_INFO && UML) || ARCH_WANT_FRAME_POINTERS 340 help 341 If you say Y here the resulting kernel image will be slightly 342 larger and slower, but it gives very useful debugging information 343 in case of kernel bugs. (precise oopses/stacktraces/warnings) 344 345config DEBUG_FORCE_WEAK_PER_CPU 346 bool "Force weak per-cpu definitions" 347 depends on DEBUG_KERNEL 348 help 349 s390 and alpha require percpu variables in modules to be 350 defined weak to work around addressing range issue which 351 puts the following two restrictions on percpu variable 352 definitions. 353 354 1. percpu symbols must be unique whether static or not 355 2. percpu variables can't be defined inside a function 356 357 To ensure that generic code follows the above rules, this 358 option forces all percpu variables to be defined as weak. 359 360endmenu # "Compiler options" 361 362config MAGIC_SYSRQ 363 bool "Magic SysRq key" 364 depends on !UML 365 help 366 If you say Y here, you will have some control over the system even 367 if the system crashes for example during kernel debugging (e.g., you 368 will be able to flush the buffer cache to disk, reboot the system 369 immediately or dump some status information). This is accomplished 370 by pressing various keys while holding SysRq (Alt+PrintScreen). It 371 also works on a serial console (on PC hardware at least), if you 372 send a BREAK and then within 5 seconds a command keypress. The 373 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y 374 unless you really know what this hack does. 375 376config MAGIC_SYSRQ_DEFAULT_ENABLE 377 hex "Enable magic SysRq key functions by default" 378 depends on MAGIC_SYSRQ 379 default 0x1 380 help 381 Specifies which SysRq key functions are enabled by default. 382 This may be set to 1 or 0 to enable or disable them all, or 383 to a bitmask as described in Documentation/sysrq.txt. 384 385config DEBUG_KERNEL 386 bool "Kernel debugging" 387 help 388 Say Y here if you are developing drivers or trying to debug and 389 identify kernel problems. 390 391menu "Memory Debugging" 392 393source mm/Kconfig.debug 394 395config DEBUG_OBJECTS 396 bool "Debug object operations" 397 depends on DEBUG_KERNEL 398 help 399 If you say Y here, additional code will be inserted into the 400 kernel to track the life time of various objects and validate 401 the operations on those objects. 402 403config DEBUG_OBJECTS_SELFTEST 404 bool "Debug objects selftest" 405 depends on DEBUG_OBJECTS 406 help 407 This enables the selftest of the object debug code. 408 409config DEBUG_OBJECTS_FREE 410 bool "Debug objects in freed memory" 411 depends on DEBUG_OBJECTS 412 help 413 This enables checks whether a k/v free operation frees an area 414 which contains an object which has not been deactivated 415 properly. This can make kmalloc/kfree-intensive workloads 416 much slower. 417 418config DEBUG_OBJECTS_TIMERS 419 bool "Debug timer objects" 420 depends on DEBUG_OBJECTS 421 help 422 If you say Y here, additional code will be inserted into the 423 timer routines to track the life time of timer objects and 424 validate the timer operations. 425 426config DEBUG_OBJECTS_WORK 427 bool "Debug work objects" 428 depends on DEBUG_OBJECTS 429 help 430 If you say Y here, additional code will be inserted into the 431 work queue routines to track the life time of work objects and 432 validate the work operations. 433 434config DEBUG_OBJECTS_RCU_HEAD 435 bool "Debug RCU callbacks objects" 436 depends on DEBUG_OBJECTS 437 help 438 Enable this to turn on debugging of RCU list heads (call_rcu() usage). 439 440config DEBUG_OBJECTS_PERCPU_COUNTER 441 bool "Debug percpu counter objects" 442 depends on DEBUG_OBJECTS 443 help 444 If you say Y here, additional code will be inserted into the 445 percpu counter routines to track the life time of percpu counter 446 objects and validate the percpu counter operations. 447 448config DEBUG_OBJECTS_ENABLE_DEFAULT 449 int "debug_objects bootup default value (0-1)" 450 range 0 1 451 default "1" 452 depends on DEBUG_OBJECTS 453 help 454 Debug objects boot parameter default value 455 456config DEBUG_SLAB 457 bool "Debug slab memory allocations" 458 depends on DEBUG_KERNEL && SLAB && !KMEMCHECK 459 help 460 Say Y here to have the kernel do limited verification on memory 461 allocation as well as poisoning memory on free to catch use of freed 462 memory. This can make kmalloc/kfree-intensive workloads much slower. 463 464config DEBUG_SLAB_LEAK 465 bool "Memory leak debugging" 466 depends on DEBUG_SLAB 467 468config SLUB_DEBUG_ON 469 bool "SLUB debugging on by default" 470 depends on SLUB && SLUB_DEBUG && !KMEMCHECK 471 default n 472 help 473 Boot with debugging on by default. SLUB boots by default with 474 the runtime debug capabilities switched off. Enabling this is 475 equivalent to specifying the "slub_debug" parameter on boot. 476 There is no support for more fine grained debug control like 477 possible with slub_debug=xxx. SLUB debugging may be switched 478 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying 479 "slub_debug=-". 480 481config SLUB_STATS 482 default n 483 bool "Enable SLUB performance statistics" 484 depends on SLUB && SYSFS 485 help 486 SLUB statistics are useful to debug SLUBs allocation behavior in 487 order find ways to optimize the allocator. This should never be 488 enabled for production use since keeping statistics slows down 489 the allocator by a few percentage points. The slabinfo command 490 supports the determination of the most active slabs to figure 491 out which slabs are relevant to a particular load. 492 Try running: slabinfo -DA 493 494config HAVE_DEBUG_KMEMLEAK 495 bool 496 497config DEBUG_KMEMLEAK 498 bool "Kernel memory leak detector" 499 depends on DEBUG_KERNEL && HAVE_DEBUG_KMEMLEAK 500 select DEBUG_FS 501 select STACKTRACE if STACKTRACE_SUPPORT 502 select KALLSYMS 503 select CRC32 504 help 505 Say Y here if you want to enable the memory leak 506 detector. The memory allocation/freeing is traced in a way 507 similar to the Boehm's conservative garbage collector, the 508 difference being that the orphan objects are not freed but 509 only shown in /sys/kernel/debug/kmemleak. Enabling this 510 feature will introduce an overhead to memory 511 allocations. See Documentation/kmemleak.txt for more 512 details. 513 514 Enabling DEBUG_SLAB or SLUB_DEBUG may increase the chances 515 of finding leaks due to the slab objects poisoning. 516 517 In order to access the kmemleak file, debugfs needs to be 518 mounted (usually at /sys/kernel/debug). 519 520config DEBUG_KMEMLEAK_EARLY_LOG_SIZE 521 int "Maximum kmemleak early log entries" 522 depends on DEBUG_KMEMLEAK 523 range 200 40000 524 default 400 525 help 526 Kmemleak must track all the memory allocations to avoid 527 reporting false positives. Since memory may be allocated or 528 freed before kmemleak is initialised, an early log buffer is 529 used to store these actions. If kmemleak reports "early log 530 buffer exceeded", please increase this value. 531 532config DEBUG_KMEMLEAK_TEST 533 tristate "Simple test for the kernel memory leak detector" 534 depends on DEBUG_KMEMLEAK && m 535 help 536 This option enables a module that explicitly leaks memory. 537 538 If unsure, say N. 539 540config DEBUG_KMEMLEAK_DEFAULT_OFF 541 bool "Default kmemleak to off" 542 depends on DEBUG_KMEMLEAK 543 help 544 Say Y here to disable kmemleak by default. It can then be enabled 545 on the command line via kmemleak=on. 546 547config DEBUG_STACK_USAGE 548 bool "Stack utilization instrumentation" 549 depends on DEBUG_KERNEL && !IA64 && !PARISC && !METAG 550 help 551 Enables the display of the minimum amount of free stack which each 552 task has ever had available in the sysrq-T and sysrq-P debug output. 553 554 This option will slow down process creation somewhat. 555 556config DEBUG_VM 557 bool "Debug VM" 558 depends on DEBUG_KERNEL 559 help 560 Enable this to turn on extended checks in the virtual-memory system 561 that may impact performance. 562 563 If unsure, say N. 564 565config DEBUG_VM_VMACACHE 566 bool "Debug VMA caching" 567 depends on DEBUG_VM 568 help 569 Enable this to turn on VMA caching debug information. Doing so 570 can cause significant overhead, so only enable it in non-production 571 environments. 572 573 If unsure, say N. 574 575config DEBUG_VM_RB 576 bool "Debug VM red-black trees" 577 depends on DEBUG_VM 578 help 579 Enable VM red-black tree debugging information and extra validations. 580 581 If unsure, say N. 582 583config DEBUG_VM_PGFLAGS 584 bool "Debug page-flags operations" 585 depends on DEBUG_VM 586 help 587 Enables extra validation on page flags operations. 588 589 If unsure, say N. 590 591config DEBUG_VIRTUAL 592 bool "Debug VM translations" 593 depends on DEBUG_KERNEL && X86 594 help 595 Enable some costly sanity checks in virtual to page code. This can 596 catch mistakes with virt_to_page() and friends. 597 598 If unsure, say N. 599 600config DEBUG_NOMMU_REGIONS 601 bool "Debug the global anon/private NOMMU mapping region tree" 602 depends on DEBUG_KERNEL && !MMU 603 help 604 This option causes the global tree of anonymous and private mapping 605 regions to be regularly checked for invalid topology. 606 607config DEBUG_MEMORY_INIT 608 bool "Debug memory initialisation" if EXPERT 609 default !EXPERT 610 help 611 Enable this for additional checks during memory initialisation. 612 The sanity checks verify aspects of the VM such as the memory model 613 and other information provided by the architecture. Verbose 614 information will be printed at KERN_DEBUG loglevel depending 615 on the mminit_loglevel= command-line option. 616 617 If unsure, say Y 618 619config MEMORY_NOTIFIER_ERROR_INJECT 620 tristate "Memory hotplug notifier error injection module" 621 depends on MEMORY_HOTPLUG_SPARSE && NOTIFIER_ERROR_INJECTION 622 help 623 This option provides the ability to inject artificial errors to 624 memory hotplug notifier chain callbacks. It is controlled through 625 debugfs interface under /sys/kernel/debug/notifier-error-inject/memory 626 627 If the notifier call chain should be failed with some events 628 notified, write the error code to "actions/<notifier event>/error". 629 630 Example: Inject memory hotplug offline error (-12 == -ENOMEM) 631 632 # cd /sys/kernel/debug/notifier-error-inject/memory 633 # echo -12 > actions/MEM_GOING_OFFLINE/error 634 # echo offline > /sys/devices/system/memory/memoryXXX/state 635 bash: echo: write error: Cannot allocate memory 636 637 To compile this code as a module, choose M here: the module will 638 be called memory-notifier-error-inject. 639 640 If unsure, say N. 641 642config DEBUG_PER_CPU_MAPS 643 bool "Debug access to per_cpu maps" 644 depends on DEBUG_KERNEL 645 depends on SMP 646 help 647 Say Y to verify that the per_cpu map being accessed has 648 been set up. This adds a fair amount of code to kernel memory 649 and decreases performance. 650 651 Say N if unsure. 652 653config DEBUG_HIGHMEM 654 bool "Highmem debugging" 655 depends on DEBUG_KERNEL && HIGHMEM 656 help 657 This option enables additional error checking for high memory 658 systems. Disable for production systems. 659 660config HAVE_DEBUG_STACKOVERFLOW 661 bool 662 663config DEBUG_STACKOVERFLOW 664 bool "Check for stack overflows" 665 depends on DEBUG_KERNEL && HAVE_DEBUG_STACKOVERFLOW 666 ---help--- 667 Say Y here if you want to check for overflows of kernel, IRQ 668 and exception stacks (if your architecture uses them). This 669 option will show detailed messages if free stack space drops 670 below a certain limit. 671 672 These kinds of bugs usually occur when call-chains in the 673 kernel get too deep, especially when interrupts are 674 involved. 675 676 Use this in cases where you see apparently random memory 677 corruption, especially if it appears in 'struct thread_info' 678 679 If in doubt, say "N". 680 681source "lib/Kconfig.kmemcheck" 682 683source "lib/Kconfig.kasan" 684 685endmenu # "Memory Debugging" 686 687config DEBUG_SHIRQ 688 bool "Debug shared IRQ handlers" 689 depends on DEBUG_KERNEL 690 help 691 Enable this to generate a spurious interrupt as soon as a shared 692 interrupt handler is registered, and just before one is deregistered. 693 Drivers ought to be able to handle interrupts coming in at those 694 points; some don't and need to be caught. 695 696menu "Debug Lockups and Hangs" 697 698config LOCKUP_DETECTOR 699 bool "Detect Hard and Soft Lockups" 700 depends on DEBUG_KERNEL && !S390 701 help 702 Say Y here to enable the kernel to act as a watchdog to detect 703 hard and soft lockups. 704 705 Softlockups are bugs that cause the kernel to loop in kernel 706 mode for more than 20 seconds, without giving other tasks a 707 chance to run. The current stack trace is displayed upon 708 detection and the system will stay locked up. 709 710 Hardlockups are bugs that cause the CPU to loop in kernel mode 711 for more than 10 seconds, without letting other interrupts have a 712 chance to run. The current stack trace is displayed upon detection 713 and the system will stay locked up. 714 715 The overhead should be minimal. A periodic hrtimer runs to 716 generate interrupts and kick the watchdog task every 4 seconds. 717 An NMI is generated every 10 seconds or so to check for hardlockups. 718 719 The frequency of hrtimer and NMI events and the soft and hard lockup 720 thresholds can be controlled through the sysctl watchdog_thresh. 721 722config HARDLOCKUP_DETECTOR 723 def_bool y 724 depends on LOCKUP_DETECTOR && !HAVE_NMI_WATCHDOG 725 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI 726 727config BOOTPARAM_HARDLOCKUP_PANIC 728 bool "Panic (Reboot) On Hard Lockups" 729 depends on HARDLOCKUP_DETECTOR 730 help 731 Say Y here to enable the kernel to panic on "hard lockups", 732 which are bugs that cause the kernel to loop in kernel 733 mode with interrupts disabled for more than 10 seconds (configurable 734 using the watchdog_thresh sysctl). 735 736 Say N if unsure. 737 738config BOOTPARAM_HARDLOCKUP_PANIC_VALUE 739 int 740 depends on HARDLOCKUP_DETECTOR 741 range 0 1 742 default 0 if !BOOTPARAM_HARDLOCKUP_PANIC 743 default 1 if BOOTPARAM_HARDLOCKUP_PANIC 744 745config BOOTPARAM_SOFTLOCKUP_PANIC 746 bool "Panic (Reboot) On Soft Lockups" 747 depends on LOCKUP_DETECTOR 748 help 749 Say Y here to enable the kernel to panic on "soft lockups", 750 which are bugs that cause the kernel to loop in kernel 751 mode for more than 20 seconds (configurable using the watchdog_thresh 752 sysctl), without giving other tasks a chance to run. 753 754 The panic can be used in combination with panic_timeout, 755 to cause the system to reboot automatically after a 756 lockup has been detected. This feature is useful for 757 high-availability systems that have uptime guarantees and 758 where a lockup must be resolved ASAP. 759 760 Say N if unsure. 761 762config BOOTPARAM_SOFTLOCKUP_PANIC_VALUE 763 int 764 depends on LOCKUP_DETECTOR 765 range 0 1 766 default 0 if !BOOTPARAM_SOFTLOCKUP_PANIC 767 default 1 if BOOTPARAM_SOFTLOCKUP_PANIC 768 769config DETECT_HUNG_TASK 770 bool "Detect Hung Tasks" 771 depends on DEBUG_KERNEL 772 default LOCKUP_DETECTOR 773 help 774 Say Y here to enable the kernel to detect "hung tasks", 775 which are bugs that cause the task to be stuck in 776 uninterruptible "D" state indefinitiley. 777 778 When a hung task is detected, the kernel will print the 779 current stack trace (which you should report), but the 780 task will stay in uninterruptible state. If lockdep is 781 enabled then all held locks will also be reported. This 782 feature has negligible overhead. 783 784config DEFAULT_HUNG_TASK_TIMEOUT 785 int "Default timeout for hung task detection (in seconds)" 786 depends on DETECT_HUNG_TASK 787 default 120 788 help 789 This option controls the default timeout (in seconds) used 790 to determine when a task has become non-responsive and should 791 be considered hung. 792 793 It can be adjusted at runtime via the kernel.hung_task_timeout_secs 794 sysctl or by writing a value to 795 /proc/sys/kernel/hung_task_timeout_secs. 796 797 A timeout of 0 disables the check. The default is two minutes. 798 Keeping the default should be fine in most cases. 799 800config BOOTPARAM_HUNG_TASK_PANIC 801 bool "Panic (Reboot) On Hung Tasks" 802 depends on DETECT_HUNG_TASK 803 help 804 Say Y here to enable the kernel to panic on "hung tasks", 805 which are bugs that cause the kernel to leave a task stuck 806 in uninterruptible "D" state. 807 808 The panic can be used in combination with panic_timeout, 809 to cause the system to reboot automatically after a 810 hung task has been detected. This feature is useful for 811 high-availability systems that have uptime guarantees and 812 where a hung tasks must be resolved ASAP. 813 814 Say N if unsure. 815 816config BOOTPARAM_HUNG_TASK_PANIC_VALUE 817 int 818 depends on DETECT_HUNG_TASK 819 range 0 1 820 default 0 if !BOOTPARAM_HUNG_TASK_PANIC 821 default 1 if BOOTPARAM_HUNG_TASK_PANIC 822 823config WQ_WATCHDOG 824 bool "Detect Workqueue Stalls" 825 depends on DEBUG_KERNEL 826 help 827 Say Y here to enable stall detection on workqueues. If a 828 worker pool doesn't make forward progress on a pending work 829 item for over a given amount of time, 30s by default, a 830 warning message is printed along with dump of workqueue 831 state. This can be configured through kernel parameter 832 "workqueue.watchdog_thresh" and its sysfs counterpart. 833 834endmenu # "Debug lockups and hangs" 835 836config PANIC_ON_OOPS 837 bool "Panic on Oops" 838 help 839 Say Y here to enable the kernel to panic when it oopses. This 840 has the same effect as setting oops=panic on the kernel command 841 line. 842 843 This feature is useful to ensure that the kernel does not do 844 anything erroneous after an oops which could result in data 845 corruption or other issues. 846 847 Say N if unsure. 848 849config PANIC_ON_OOPS_VALUE 850 int 851 range 0 1 852 default 0 if !PANIC_ON_OOPS 853 default 1 if PANIC_ON_OOPS 854 855config PANIC_TIMEOUT 856 int "panic timeout" 857 default 0 858 help 859 Set the timeout value (in seconds) until a reboot occurs when the 860 the kernel panics. If n = 0, then we wait forever. A timeout 861 value n > 0 will wait n seconds before rebooting, while a timeout 862 value n < 0 will reboot immediately. 863 864config SCHED_DEBUG 865 bool "Collect scheduler debugging info" 866 depends on DEBUG_KERNEL && PROC_FS 867 default y 868 help 869 If you say Y here, the /proc/sched_debug file will be provided 870 that can help debug the scheduler. The runtime overhead of this 871 option is minimal. 872 873config SCHED_INFO 874 bool 875 default n 876 877config SCHEDSTATS 878 bool "Collect scheduler statistics" 879 depends on DEBUG_KERNEL && PROC_FS 880 select SCHED_INFO 881 help 882 If you say Y here, additional code will be inserted into the 883 scheduler and related routines to collect statistics about 884 scheduler behavior and provide them in /proc/schedstat. These 885 stats may be useful for both tuning and debugging the scheduler 886 If you aren't debugging the scheduler or trying to tune a specific 887 application, you can say N to avoid the very slight overhead 888 this adds. 889 890config SCHED_STACK_END_CHECK 891 bool "Detect stack corruption on calls to schedule()" 892 depends on DEBUG_KERNEL 893 default n 894 help 895 This option checks for a stack overrun on calls to schedule(). 896 If the stack end location is found to be over written always panic as 897 the content of the corrupted region can no longer be trusted. 898 This is to ensure no erroneous behaviour occurs which could result in 899 data corruption or a sporadic crash at a later stage once the region 900 is examined. The runtime overhead introduced is minimal. 901 902config DEBUG_TIMEKEEPING 903 bool "Enable extra timekeeping sanity checking" 904 help 905 This option will enable additional timekeeping sanity checks 906 which may be helpful when diagnosing issues where timekeeping 907 problems are suspected. 908 909 This may include checks in the timekeeping hotpaths, so this 910 option may have a (very small) performance impact to some 911 workloads. 912 913 If unsure, say N. 914 915config TIMER_STATS 916 bool "Collect kernel timers statistics" 917 depends on DEBUG_KERNEL && PROC_FS 918 help 919 If you say Y here, additional code will be inserted into the 920 timer routines to collect statistics about kernel timers being 921 reprogrammed. The statistics can be read from /proc/timer_stats. 922 The statistics collection is started by writing 1 to /proc/timer_stats, 923 writing 0 stops it. This feature is useful to collect information 924 about timer usage patterns in kernel and userspace. This feature 925 is lightweight if enabled in the kernel config but not activated 926 (it defaults to deactivated on bootup and will only be activated 927 if some application like powertop activates it explicitly). 928 929config DEBUG_PREEMPT 930 bool "Debug preemptible kernel" 931 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT 932 default y 933 help 934 If you say Y here then the kernel will use a debug variant of the 935 commonly used smp_processor_id() function and will print warnings 936 if kernel code uses it in a preemption-unsafe way. Also, the kernel 937 will detect preemption count underflows. 938 939menu "Lock Debugging (spinlocks, mutexes, etc...)" 940 941config DEBUG_RT_MUTEXES 942 bool "RT Mutex debugging, deadlock detection" 943 depends on DEBUG_KERNEL && RT_MUTEXES 944 help 945 This allows rt mutex semantics violations and rt mutex related 946 deadlocks (lockups) to be detected and reported automatically. 947 948config DEBUG_SPINLOCK 949 bool "Spinlock and rw-lock debugging: basic checks" 950 depends on DEBUG_KERNEL 951 select UNINLINE_SPIN_UNLOCK 952 help 953 Say Y here and build SMP to catch missing spinlock initialization 954 and certain other kinds of spinlock errors commonly made. This is 955 best used in conjunction with the NMI watchdog so that spinlock 956 deadlocks are also debuggable. 957 958config DEBUG_MUTEXES 959 bool "Mutex debugging: basic checks" 960 depends on DEBUG_KERNEL 961 help 962 This feature allows mutex semantics violations to be detected and 963 reported. 964 965config DEBUG_WW_MUTEX_SLOWPATH 966 bool "Wait/wound mutex debugging: Slowpath testing" 967 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 968 select DEBUG_LOCK_ALLOC 969 select DEBUG_SPINLOCK 970 select DEBUG_MUTEXES 971 help 972 This feature enables slowpath testing for w/w mutex users by 973 injecting additional -EDEADLK wound/backoff cases. Together with 974 the full mutex checks enabled with (CONFIG_PROVE_LOCKING) this 975 will test all possible w/w mutex interface abuse with the 976 exception of simply not acquiring all the required locks. 977 Note that this feature can introduce significant overhead, so 978 it really should not be enabled in a production or distro kernel, 979 even a debug kernel. If you are a driver writer, enable it. If 980 you are a distro, do not. 981 982config DEBUG_LOCK_ALLOC 983 bool "Lock debugging: detect incorrect freeing of live locks" 984 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 985 select DEBUG_SPINLOCK 986 select DEBUG_MUTEXES 987 select LOCKDEP 988 help 989 This feature will check whether any held lock (spinlock, rwlock, 990 mutex or rwsem) is incorrectly freed by the kernel, via any of the 991 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(), 992 vfree(), etc.), whether a live lock is incorrectly reinitialized via 993 spin_lock_init()/mutex_init()/etc., or whether there is any lock 994 held during task exit. 995 996config PROVE_LOCKING 997 bool "Lock debugging: prove locking correctness" 998 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 999 select LOCKDEP 1000 select DEBUG_SPINLOCK 1001 select DEBUG_MUTEXES 1002 select DEBUG_LOCK_ALLOC 1003 select TRACE_IRQFLAGS 1004 default n 1005 help 1006 This feature enables the kernel to prove that all locking 1007 that occurs in the kernel runtime is mathematically 1008 correct: that under no circumstance could an arbitrary (and 1009 not yet triggered) combination of observed locking 1010 sequences (on an arbitrary number of CPUs, running an 1011 arbitrary number of tasks and interrupt contexts) cause a 1012 deadlock. 1013 1014 In short, this feature enables the kernel to report locking 1015 related deadlocks before they actually occur. 1016 1017 The proof does not depend on how hard and complex a 1018 deadlock scenario would be to trigger: how many 1019 participant CPUs, tasks and irq-contexts would be needed 1020 for it to trigger. The proof also does not depend on 1021 timing: if a race and a resulting deadlock is possible 1022 theoretically (no matter how unlikely the race scenario 1023 is), it will be proven so and will immediately be 1024 reported by the kernel (once the event is observed that 1025 makes the deadlock theoretically possible). 1026 1027 If a deadlock is impossible (i.e. the locking rules, as 1028 observed by the kernel, are mathematically correct), the 1029 kernel reports nothing. 1030 1031 NOTE: this feature can also be enabled for rwlocks, mutexes 1032 and rwsems - in which case all dependencies between these 1033 different locking variants are observed and mapped too, and 1034 the proof of observed correctness is also maintained for an 1035 arbitrary combination of these separate locking variants. 1036 1037 For more details, see Documentation/locking/lockdep-design.txt. 1038 1039config LOCKDEP 1040 bool 1041 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 1042 select STACKTRACE 1043 select FRAME_POINTER if !MIPS && !PPC && !ARM_UNWIND && !S390 && !MICROBLAZE && !ARC && !SCORE 1044 select KALLSYMS 1045 select KALLSYMS_ALL 1046 1047config LOCK_STAT 1048 bool "Lock usage statistics" 1049 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT 1050 select LOCKDEP 1051 select DEBUG_SPINLOCK 1052 select DEBUG_MUTEXES 1053 select DEBUG_LOCK_ALLOC 1054 default n 1055 help 1056 This feature enables tracking lock contention points 1057 1058 For more details, see Documentation/locking/lockstat.txt 1059 1060 This also enables lock events required by "perf lock", 1061 subcommand of perf. 1062 If you want to use "perf lock", you also need to turn on 1063 CONFIG_EVENT_TRACING. 1064 1065 CONFIG_LOCK_STAT defines "contended" and "acquired" lock events. 1066 (CONFIG_LOCKDEP defines "acquire" and "release" events.) 1067 1068config DEBUG_LOCKDEP 1069 bool "Lock dependency engine debugging" 1070 depends on DEBUG_KERNEL && LOCKDEP 1071 help 1072 If you say Y here, the lock dependency engine will do 1073 additional runtime checks to debug itself, at the price 1074 of more runtime overhead. 1075 1076config DEBUG_ATOMIC_SLEEP 1077 bool "Sleep inside atomic section checking" 1078 select PREEMPT_COUNT 1079 depends on DEBUG_KERNEL 1080 help 1081 If you say Y here, various routines which may sleep will become very 1082 noisy if they are called inside atomic sections: when a spinlock is 1083 held, inside an rcu read side critical section, inside preempt disabled 1084 sections, inside an interrupt, etc... 1085 1086config DEBUG_LOCKING_API_SELFTESTS 1087 bool "Locking API boot-time self-tests" 1088 depends on DEBUG_KERNEL 1089 help 1090 Say Y here if you want the kernel to run a short self-test during 1091 bootup. The self-test checks whether common types of locking bugs 1092 are detected by debugging mechanisms or not. (if you disable 1093 lock debugging then those bugs wont be detected of course.) 1094 The following locking APIs are covered: spinlocks, rwlocks, 1095 mutexes and rwsems. 1096 1097config LOCK_TORTURE_TEST 1098 tristate "torture tests for locking" 1099 depends on DEBUG_KERNEL 1100 select TORTURE_TEST 1101 default n 1102 help 1103 This option provides a kernel module that runs torture tests 1104 on kernel locking primitives. The kernel module may be built 1105 after the fact on the running kernel to be tested, if desired. 1106 1107 Say Y here if you want kernel locking-primitive torture tests 1108 to be built into the kernel. 1109 Say M if you want these torture tests to build as a module. 1110 Say N if you are unsure. 1111 1112endmenu # lock debugging 1113 1114config TRACE_IRQFLAGS 1115 bool 1116 help 1117 Enables hooks to interrupt enabling and disabling for 1118 either tracing or lock debugging. 1119 1120config STACKTRACE 1121 bool "Stack backtrace support" 1122 depends on STACKTRACE_SUPPORT 1123 help 1124 This option causes the kernel to create a /proc/pid/stack for 1125 every process, showing its current stack trace. 1126 It is also used by various kernel debugging features that require 1127 stack trace generation. 1128 1129config DEBUG_KOBJECT 1130 bool "kobject debugging" 1131 depends on DEBUG_KERNEL 1132 help 1133 If you say Y here, some extra kobject debugging messages will be sent 1134 to the syslog. 1135 1136config DEBUG_KOBJECT_RELEASE 1137 bool "kobject release debugging" 1138 depends on DEBUG_OBJECTS_TIMERS 1139 help 1140 kobjects are reference counted objects. This means that their 1141 last reference count put is not predictable, and the kobject can 1142 live on past the point at which a driver decides to drop it's 1143 initial reference to the kobject gained on allocation. An 1144 example of this would be a struct device which has just been 1145 unregistered. 1146 1147 However, some buggy drivers assume that after such an operation, 1148 the memory backing the kobject can be immediately freed. This 1149 goes completely against the principles of a refcounted object. 1150 1151 If you say Y here, the kernel will delay the release of kobjects 1152 on the last reference count to improve the visibility of this 1153 kind of kobject release bug. 1154 1155config HAVE_DEBUG_BUGVERBOSE 1156 bool 1157 1158config DEBUG_BUGVERBOSE 1159 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EXPERT 1160 depends on BUG && (GENERIC_BUG || HAVE_DEBUG_BUGVERBOSE) 1161 default y 1162 help 1163 Say Y here to make BUG() panics output the file name and line number 1164 of the BUG call as well as the EIP and oops trace. This aids 1165 debugging but costs about 70-100K of memory. 1166 1167config DEBUG_LIST 1168 bool "Debug linked list manipulation" 1169 depends on DEBUG_KERNEL 1170 help 1171 Enable this to turn on extended checks in the linked-list 1172 walking routines. 1173 1174 If unsure, say N. 1175 1176config DEBUG_PI_LIST 1177 bool "Debug priority linked list manipulation" 1178 depends on DEBUG_KERNEL 1179 help 1180 Enable this to turn on extended checks in the priority-ordered 1181 linked-list (plist) walking routines. This checks the entire 1182 list multiple times during each manipulation. 1183 1184 If unsure, say N. 1185 1186config DEBUG_SG 1187 bool "Debug SG table operations" 1188 depends on DEBUG_KERNEL 1189 help 1190 Enable this to turn on checks on scatter-gather tables. This can 1191 help find problems with drivers that do not properly initialize 1192 their sg tables. 1193 1194 If unsure, say N. 1195 1196config DEBUG_NOTIFIERS 1197 bool "Debug notifier call chains" 1198 depends on DEBUG_KERNEL 1199 help 1200 Enable this to turn on sanity checking for notifier call chains. 1201 This is most useful for kernel developers to make sure that 1202 modules properly unregister themselves from notifier chains. 1203 This is a relatively cheap check but if you care about maximum 1204 performance, say N. 1205 1206config DEBUG_CREDENTIALS 1207 bool "Debug credential management" 1208 depends on DEBUG_KERNEL 1209 help 1210 Enable this to turn on some debug checking for credential 1211 management. The additional code keeps track of the number of 1212 pointers from task_structs to any given cred struct, and checks to 1213 see that this number never exceeds the usage count of the cred 1214 struct. 1215 1216 Furthermore, if SELinux is enabled, this also checks that the 1217 security pointer in the cred struct is never seen to be invalid. 1218 1219 If unsure, say N. 1220 1221menu "RCU Debugging" 1222 1223config PROVE_RCU 1224 def_bool PROVE_LOCKING 1225 1226config PROVE_RCU_REPEATEDLY 1227 bool "RCU debugging: don't disable PROVE_RCU on first splat" 1228 depends on PROVE_RCU 1229 default n 1230 help 1231 By itself, PROVE_RCU will disable checking upon issuing the 1232 first warning (or "splat"). This feature prevents such 1233 disabling, allowing multiple RCU-lockdep warnings to be printed 1234 on a single reboot. 1235 1236 Say Y to allow multiple RCU-lockdep warnings per boot. 1237 1238 Say N if you are unsure. 1239 1240config SPARSE_RCU_POINTER 1241 bool "RCU debugging: sparse-based checks for pointer usage" 1242 default n 1243 help 1244 This feature enables the __rcu sparse annotation for 1245 RCU-protected pointers. This annotation will cause sparse 1246 to flag any non-RCU used of annotated pointers. This can be 1247 helpful when debugging RCU usage. Please note that this feature 1248 is not intended to enforce code cleanliness; it is instead merely 1249 a debugging aid. 1250 1251 Say Y to make sparse flag questionable use of RCU-protected pointers 1252 1253 Say N if you are unsure. 1254 1255config TORTURE_TEST 1256 tristate 1257 default n 1258 1259config RCU_TORTURE_TEST 1260 tristate "torture tests for RCU" 1261 depends on DEBUG_KERNEL 1262 select TORTURE_TEST 1263 select SRCU 1264 select TASKS_RCU 1265 default n 1266 help 1267 This option provides a kernel module that runs torture tests 1268 on the RCU infrastructure. The kernel module may be built 1269 after the fact on the running kernel to be tested, if desired. 1270 1271 Say Y here if you want RCU torture tests to be built into 1272 the kernel. 1273 Say M if you want the RCU torture tests to build as a module. 1274 Say N if you are unsure. 1275 1276config RCU_TORTURE_TEST_RUNNABLE 1277 bool "torture tests for RCU runnable by default" 1278 depends on RCU_TORTURE_TEST = y 1279 default n 1280 help 1281 This option provides a way to build the RCU torture tests 1282 directly into the kernel without them starting up at boot 1283 time. You can use /proc/sys/kernel/rcutorture_runnable 1284 to manually override this setting. This /proc file is 1285 available only when the RCU torture tests have been built 1286 into the kernel. 1287 1288 Say Y here if you want the RCU torture tests to start during 1289 boot (you probably don't). 1290 Say N here if you want the RCU torture tests to start only 1291 after being manually enabled via /proc. 1292 1293config RCU_TORTURE_TEST_SLOW_PREINIT 1294 bool "Slow down RCU grace-period pre-initialization to expose races" 1295 depends on RCU_TORTURE_TEST 1296 help 1297 This option delays grace-period pre-initialization (the 1298 propagation of CPU-hotplug changes up the rcu_node combining 1299 tree) for a few jiffies between initializing each pair of 1300 consecutive rcu_node structures. This helps to expose races 1301 involving grace-period pre-initialization, in other words, it 1302 makes your kernel less stable. It can also greatly increase 1303 grace-period latency, especially on systems with large numbers 1304 of CPUs. This is useful when torture-testing RCU, but in 1305 almost no other circumstance. 1306 1307 Say Y here if you want your system to crash and hang more often. 1308 Say N if you want a sane system. 1309 1310config RCU_TORTURE_TEST_SLOW_PREINIT_DELAY 1311 int "How much to slow down RCU grace-period pre-initialization" 1312 range 0 5 1313 default 3 1314 depends on RCU_TORTURE_TEST_SLOW_PREINIT 1315 help 1316 This option specifies the number of jiffies to wait between 1317 each rcu_node structure pre-initialization step. 1318 1319config RCU_TORTURE_TEST_SLOW_INIT 1320 bool "Slow down RCU grace-period initialization to expose races" 1321 depends on RCU_TORTURE_TEST 1322 help 1323 This option delays grace-period initialization for a few 1324 jiffies between initializing each pair of consecutive 1325 rcu_node structures. This helps to expose races involving 1326 grace-period initialization, in other words, it makes your 1327 kernel less stable. It can also greatly increase grace-period 1328 latency, especially on systems with large numbers of CPUs. 1329 This is useful when torture-testing RCU, but in almost no 1330 other circumstance. 1331 1332 Say Y here if you want your system to crash and hang more often. 1333 Say N if you want a sane system. 1334 1335config RCU_TORTURE_TEST_SLOW_INIT_DELAY 1336 int "How much to slow down RCU grace-period initialization" 1337 range 0 5 1338 default 3 1339 depends on RCU_TORTURE_TEST_SLOW_INIT 1340 help 1341 This option specifies the number of jiffies to wait between 1342 each rcu_node structure initialization. 1343 1344config RCU_TORTURE_TEST_SLOW_CLEANUP 1345 bool "Slow down RCU grace-period cleanup to expose races" 1346 depends on RCU_TORTURE_TEST 1347 help 1348 This option delays grace-period cleanup for a few jiffies 1349 between cleaning up each pair of consecutive rcu_node 1350 structures. This helps to expose races involving grace-period 1351 cleanup, in other words, it makes your kernel less stable. 1352 It can also greatly increase grace-period latency, especially 1353 on systems with large numbers of CPUs. This is useful when 1354 torture-testing RCU, but in almost no other circumstance. 1355 1356 Say Y here if you want your system to crash and hang more often. 1357 Say N if you want a sane system. 1358 1359config RCU_TORTURE_TEST_SLOW_CLEANUP_DELAY 1360 int "How much to slow down RCU grace-period cleanup" 1361 range 0 5 1362 default 3 1363 depends on RCU_TORTURE_TEST_SLOW_CLEANUP 1364 help 1365 This option specifies the number of jiffies to wait between 1366 each rcu_node structure cleanup operation. 1367 1368config RCU_CPU_STALL_TIMEOUT 1369 int "RCU CPU stall timeout in seconds" 1370 depends on RCU_STALL_COMMON 1371 range 3 300 1372 default 21 1373 help 1374 If a given RCU grace period extends more than the specified 1375 number of seconds, a CPU stall warning is printed. If the 1376 RCU grace period persists, additional CPU stall warnings are 1377 printed at more widely spaced intervals. 1378 1379config RCU_TRACE 1380 bool "Enable tracing for RCU" 1381 depends on DEBUG_KERNEL 1382 select TRACE_CLOCK 1383 help 1384 This option provides tracing in RCU which presents stats 1385 in debugfs for debugging RCU implementation. 1386 1387 Say Y here if you want to enable RCU tracing 1388 Say N if you are unsure. 1389 1390config RCU_EQS_DEBUG 1391 bool "Provide debugging asserts for adding NO_HZ support to an arch" 1392 depends on DEBUG_KERNEL 1393 help 1394 This option provides consistency checks in RCU's handling of 1395 NO_HZ. These checks have proven quite helpful in detecting 1396 bugs in arch-specific NO_HZ code. 1397 1398 Say N here if you need ultimate kernel/user switch latencies 1399 Say Y if you are unsure 1400 1401endmenu # "RCU Debugging" 1402 1403config DEBUG_WQ_FORCE_RR_CPU 1404 bool "Force round-robin CPU selection for unbound work items" 1405 depends on DEBUG_KERNEL 1406 default n 1407 help 1408 Workqueue used to implicitly guarantee that work items queued 1409 without explicit CPU specified are put on the local CPU. This 1410 guarantee is no longer true and while local CPU is still 1411 preferred work items may be put on foreign CPUs. Kernel 1412 parameter "workqueue.debug_force_rr_cpu" is added to force 1413 round-robin CPU selection to flush out usages which depend on the 1414 now broken guarantee. This config option enables the debug 1415 feature by default. When enabled, memory and cache locality will 1416 be impacted. 1417 1418config DEBUG_BLOCK_EXT_DEVT 1419 bool "Force extended block device numbers and spread them" 1420 depends on DEBUG_KERNEL 1421 depends on BLOCK 1422 default n 1423 help 1424 BIG FAT WARNING: ENABLING THIS OPTION MIGHT BREAK BOOTING ON 1425 SOME DISTRIBUTIONS. DO NOT ENABLE THIS UNLESS YOU KNOW WHAT 1426 YOU ARE DOING. Distros, please enable this and fix whatever 1427 is broken. 1428 1429 Conventionally, block device numbers are allocated from 1430 predetermined contiguous area. However, extended block area 1431 may introduce non-contiguous block device numbers. This 1432 option forces most block device numbers to be allocated from 1433 the extended space and spreads them to discover kernel or 1434 userland code paths which assume predetermined contiguous 1435 device number allocation. 1436 1437 Note that turning on this debug option shuffles all the 1438 device numbers for all IDE and SCSI devices including libata 1439 ones, so root partition specified using device number 1440 directly (via rdev or root=MAJ:MIN) won't work anymore. 1441 Textual device names (root=/dev/sdXn) will continue to work. 1442 1443 Say N if you are unsure. 1444 1445config NOTIFIER_ERROR_INJECTION 1446 tristate "Notifier error injection" 1447 depends on DEBUG_KERNEL 1448 select DEBUG_FS 1449 help 1450 This option provides the ability to inject artificial errors to 1451 specified notifier chain callbacks. It is useful to test the error 1452 handling of notifier call chain failures. 1453 1454 Say N if unsure. 1455 1456config CPU_NOTIFIER_ERROR_INJECT 1457 tristate "CPU notifier error injection module" 1458 depends on HOTPLUG_CPU && NOTIFIER_ERROR_INJECTION 1459 help 1460 This option provides a kernel module that can be used to test 1461 the error handling of the cpu notifiers by injecting artificial 1462 errors to CPU notifier chain callbacks. It is controlled through 1463 debugfs interface under /sys/kernel/debug/notifier-error-inject/cpu 1464 1465 If the notifier call chain should be failed with some events 1466 notified, write the error code to "actions/<notifier event>/error". 1467 1468 Example: Inject CPU offline error (-1 == -EPERM) 1469 1470 # cd /sys/kernel/debug/notifier-error-inject/cpu 1471 # echo -1 > actions/CPU_DOWN_PREPARE/error 1472 # echo 0 > /sys/devices/system/cpu/cpu1/online 1473 bash: echo: write error: Operation not permitted 1474 1475 To compile this code as a module, choose M here: the module will 1476 be called cpu-notifier-error-inject. 1477 1478 If unsure, say N. 1479 1480config PM_NOTIFIER_ERROR_INJECT 1481 tristate "PM notifier error injection module" 1482 depends on PM && NOTIFIER_ERROR_INJECTION 1483 default m if PM_DEBUG 1484 help 1485 This option provides the ability to inject artificial errors to 1486 PM notifier chain callbacks. It is controlled through debugfs 1487 interface /sys/kernel/debug/notifier-error-inject/pm 1488 1489 If the notifier call chain should be failed with some events 1490 notified, write the error code to "actions/<notifier event>/error". 1491 1492 Example: Inject PM suspend error (-12 = -ENOMEM) 1493 1494 # cd /sys/kernel/debug/notifier-error-inject/pm/ 1495 # echo -12 > actions/PM_SUSPEND_PREPARE/error 1496 # echo mem > /sys/power/state 1497 bash: echo: write error: Cannot allocate memory 1498 1499 To compile this code as a module, choose M here: the module will 1500 be called pm-notifier-error-inject. 1501 1502 If unsure, say N. 1503 1504config OF_RECONFIG_NOTIFIER_ERROR_INJECT 1505 tristate "OF reconfig notifier error injection module" 1506 depends on OF_DYNAMIC && NOTIFIER_ERROR_INJECTION 1507 help 1508 This option provides the ability to inject artificial errors to 1509 OF reconfig notifier chain callbacks. It is controlled 1510 through debugfs interface under 1511 /sys/kernel/debug/notifier-error-inject/OF-reconfig/ 1512 1513 If the notifier call chain should be failed with some events 1514 notified, write the error code to "actions/<notifier event>/error". 1515 1516 To compile this code as a module, choose M here: the module will 1517 be called of-reconfig-notifier-error-inject. 1518 1519 If unsure, say N. 1520 1521config NETDEV_NOTIFIER_ERROR_INJECT 1522 tristate "Netdev notifier error injection module" 1523 depends on NET && NOTIFIER_ERROR_INJECTION 1524 help 1525 This option provides the ability to inject artificial errors to 1526 netdevice notifier chain callbacks. It is controlled through debugfs 1527 interface /sys/kernel/debug/notifier-error-inject/netdev 1528 1529 If the notifier call chain should be failed with some events 1530 notified, write the error code to "actions/<notifier event>/error". 1531 1532 Example: Inject netdevice mtu change error (-22 = -EINVAL) 1533 1534 # cd /sys/kernel/debug/notifier-error-inject/netdev 1535 # echo -22 > actions/NETDEV_CHANGEMTU/error 1536 # ip link set eth0 mtu 1024 1537 RTNETLINK answers: Invalid argument 1538 1539 To compile this code as a module, choose M here: the module will 1540 be called netdev-notifier-error-inject. 1541 1542 If unsure, say N. 1543 1544config FAULT_INJECTION 1545 bool "Fault-injection framework" 1546 depends on DEBUG_KERNEL 1547 help 1548 Provide fault-injection framework. 1549 For more details, see Documentation/fault-injection/. 1550 1551config FAILSLAB 1552 bool "Fault-injection capability for kmalloc" 1553 depends on FAULT_INJECTION 1554 depends on SLAB || SLUB 1555 help 1556 Provide fault-injection capability for kmalloc. 1557 1558config FAIL_PAGE_ALLOC 1559 bool "Fault-injection capabilitiy for alloc_pages()" 1560 depends on FAULT_INJECTION 1561 help 1562 Provide fault-injection capability for alloc_pages(). 1563 1564config FAIL_MAKE_REQUEST 1565 bool "Fault-injection capability for disk IO" 1566 depends on FAULT_INJECTION && BLOCK 1567 help 1568 Provide fault-injection capability for disk IO. 1569 1570config FAIL_IO_TIMEOUT 1571 bool "Fault-injection capability for faking disk interrupts" 1572 depends on FAULT_INJECTION && BLOCK 1573 help 1574 Provide fault-injection capability on end IO handling. This 1575 will make the block layer "forget" an interrupt as configured, 1576 thus exercising the error handling. 1577 1578 Only works with drivers that use the generic timeout handling, 1579 for others it wont do anything. 1580 1581config FAIL_MMC_REQUEST 1582 bool "Fault-injection capability for MMC IO" 1583 depends on FAULT_INJECTION_DEBUG_FS && MMC 1584 help 1585 Provide fault-injection capability for MMC IO. 1586 This will make the mmc core return data errors. This is 1587 useful to test the error handling in the mmc block device 1588 and to test how the mmc host driver handles retries from 1589 the block device. 1590 1591config FAIL_FUTEX 1592 bool "Fault-injection capability for futexes" 1593 select DEBUG_FS 1594 depends on FAULT_INJECTION && FUTEX 1595 help 1596 Provide fault-injection capability for futexes. 1597 1598config FAULT_INJECTION_DEBUG_FS 1599 bool "Debugfs entries for fault-injection capabilities" 1600 depends on FAULT_INJECTION && SYSFS && DEBUG_FS 1601 help 1602 Enable configuration of fault-injection capabilities via debugfs. 1603 1604config FAULT_INJECTION_STACKTRACE_FILTER 1605 bool "stacktrace filter for fault-injection capabilities" 1606 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT 1607 depends on !X86_64 1608 select STACKTRACE 1609 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC && !SCORE 1610 help 1611 Provide stacktrace filter for fault-injection capabilities 1612 1613config LATENCYTOP 1614 bool "Latency measuring infrastructure" 1615 depends on DEBUG_KERNEL 1616 depends on STACKTRACE_SUPPORT 1617 depends on PROC_FS 1618 select FRAME_POINTER if !MIPS && !PPC && !S390 && !MICROBLAZE && !ARM_UNWIND && !ARC 1619 select KALLSYMS 1620 select KALLSYMS_ALL 1621 select STACKTRACE 1622 select SCHEDSTATS 1623 select SCHED_DEBUG 1624 help 1625 Enable this option if you want to use the LatencyTOP tool 1626 to find out which userspace is blocking on what kernel operations. 1627 1628config ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS 1629 bool 1630 1631config DEBUG_STRICT_USER_COPY_CHECKS 1632 bool "Strict user copy size checks" 1633 depends on ARCH_HAS_DEBUG_STRICT_USER_COPY_CHECKS 1634 depends on DEBUG_KERNEL && !TRACE_BRANCH_PROFILING 1635 help 1636 Enabling this option turns a certain set of sanity checks for user 1637 copy operations into compile time failures. 1638 1639 The copy_from_user() etc checks are there to help test if there 1640 are sufficient security checks on the length argument of 1641 the copy operation, by having gcc prove that the argument is 1642 within bounds. 1643 1644 If unsure, say N. 1645 1646source kernel/trace/Kconfig 1647 1648menu "Runtime Testing" 1649 1650config LKDTM 1651 tristate "Linux Kernel Dump Test Tool Module" 1652 depends on DEBUG_FS 1653 depends on BLOCK 1654 default n 1655 help 1656 This module enables testing of the different dumping mechanisms by 1657 inducing system failures at predefined crash points. 1658 If you don't need it: say N 1659 Choose M here to compile this code as a module. The module will be 1660 called lkdtm. 1661 1662 Documentation on how to use the module can be found in 1663 Documentation/fault-injection/provoke-crashes.txt 1664 1665config TEST_LIST_SORT 1666 bool "Linked list sorting test" 1667 depends on DEBUG_KERNEL 1668 help 1669 Enable this to turn on 'list_sort()' function test. This test is 1670 executed only once during system boot, so affects only boot time. 1671 1672 If unsure, say N. 1673 1674config KPROBES_SANITY_TEST 1675 bool "Kprobes sanity tests" 1676 depends on DEBUG_KERNEL 1677 depends on KPROBES 1678 default n 1679 help 1680 This option provides for testing basic kprobes functionality on 1681 boot. A sample kprobe, jprobe and kretprobe are inserted and 1682 verified for functionality. 1683 1684 Say N if you are unsure. 1685 1686config BACKTRACE_SELF_TEST 1687 tristate "Self test for the backtrace code" 1688 depends on DEBUG_KERNEL 1689 default n 1690 help 1691 This option provides a kernel module that can be used to test 1692 the kernel stack backtrace code. This option is not useful 1693 for distributions or general kernels, but only for kernel 1694 developers working on architecture code. 1695 1696 Note that if you want to also test saved backtraces, you will 1697 have to enable STACKTRACE as well. 1698 1699 Say N if you are unsure. 1700 1701config RBTREE_TEST 1702 tristate "Red-Black tree test" 1703 depends on DEBUG_KERNEL 1704 help 1705 A benchmark measuring the performance of the rbtree library. 1706 Also includes rbtree invariant checks. 1707 1708config INTERVAL_TREE_TEST 1709 tristate "Interval tree test" 1710 depends on m && DEBUG_KERNEL 1711 select INTERVAL_TREE 1712 help 1713 A benchmark measuring the performance of the interval tree library 1714 1715config PERCPU_TEST 1716 tristate "Per cpu operations test" 1717 depends on m && DEBUG_KERNEL 1718 help 1719 Enable this option to build test module which validates per-cpu 1720 operations. 1721 1722 If unsure, say N. 1723 1724config ATOMIC64_SELFTEST 1725 bool "Perform an atomic64_t self-test at boot" 1726 help 1727 Enable this option to test the atomic64_t functions at boot. 1728 1729 If unsure, say N. 1730 1731config ASYNC_RAID6_TEST 1732 tristate "Self test for hardware accelerated raid6 recovery" 1733 depends on ASYNC_RAID6_RECOV 1734 select ASYNC_MEMCPY 1735 ---help--- 1736 This is a one-shot self test that permutes through the 1737 recovery of all the possible two disk failure scenarios for a 1738 N-disk array. Recovery is performed with the asynchronous 1739 raid6 recovery routines, and will optionally use an offload 1740 engine if one is available. 1741 1742 If unsure, say N. 1743 1744config TEST_HEXDUMP 1745 tristate "Test functions located in the hexdump module at runtime" 1746 1747config TEST_STRING_HELPERS 1748 tristate "Test functions located in the string_helpers module at runtime" 1749 1750config TEST_KSTRTOX 1751 tristate "Test kstrto*() family of functions at runtime" 1752 1753config TEST_PRINTF 1754 tristate "Test printf() family of functions at runtime" 1755 1756config TEST_BITMAP 1757 tristate "Test bitmap_*() family of functions at runtime" 1758 default n 1759 help 1760 Enable this option to test the bitmap functions at boot. 1761 1762 If unsure, say N. 1763 1764config TEST_RHASHTABLE 1765 tristate "Perform selftest on resizable hash table" 1766 default n 1767 help 1768 Enable this option to test the rhashtable functions at boot. 1769 1770 If unsure, say N. 1771 1772endmenu # runtime tests 1773 1774config PROVIDE_OHCI1394_DMA_INIT 1775 bool "Remote debugging over FireWire early on boot" 1776 depends on PCI && X86 1777 help 1778 If you want to debug problems which hang or crash the kernel early 1779 on boot and the crashing machine has a FireWire port, you can use 1780 this feature to remotely access the memory of the crashed machine 1781 over FireWire. This employs remote DMA as part of the OHCI1394 1782 specification which is now the standard for FireWire controllers. 1783 1784 With remote DMA, you can monitor the printk buffer remotely using 1785 firescope and access all memory below 4GB using fireproxy from gdb. 1786 Even controlling a kernel debugger is possible using remote DMA. 1787 1788 Usage: 1789 1790 If ohci1394_dma=early is used as boot parameter, it will initialize 1791 all OHCI1394 controllers which are found in the PCI config space. 1792 1793 As all changes to the FireWire bus such as enabling and disabling 1794 devices cause a bus reset and thereby disable remote DMA for all 1795 devices, be sure to have the cable plugged and FireWire enabled on 1796 the debugging host before booting the debug target for debugging. 1797 1798 This code (~1k) is freed after boot. By then, the firewire stack 1799 in charge of the OHCI-1394 controllers should be used instead. 1800 1801 See Documentation/debugging-via-ohci1394.txt for more information. 1802 1803config BUILD_DOCSRC 1804 bool "Build targets in Documentation/ tree" 1805 depends on HEADERS_CHECK 1806 help 1807 This option attempts to build objects from the source files in the 1808 kernel Documentation/ tree. 1809 1810 Say N if you are unsure. 1811 1812config DMA_API_DEBUG 1813 bool "Enable debugging of DMA-API usage" 1814 depends on HAVE_DMA_API_DEBUG 1815 help 1816 Enable this option to debug the use of the DMA API by device drivers. 1817 With this option you will be able to detect common bugs in device 1818 drivers like double-freeing of DMA mappings or freeing mappings that 1819 were never allocated. 1820 1821 This also attempts to catch cases where a page owned by DMA is 1822 accessed by the cpu in a way that could cause data corruption. For 1823 example, this enables cow_user_page() to check that the source page is 1824 not undergoing DMA. 1825 1826 This option causes a performance degradation. Use only if you want to 1827 debug device drivers and dma interactions. 1828 1829 If unsure, say N. 1830 1831config TEST_LKM 1832 tristate "Test module loading with 'hello world' module" 1833 default n 1834 depends on m 1835 help 1836 This builds the "test_module" module that emits "Hello, world" 1837 on printk when loaded. It is designed to be used for basic 1838 evaluation of the module loading subsystem (for example when 1839 validating module verification). It lacks any extra dependencies, 1840 and will not normally be loaded by the system unless explicitly 1841 requested by name. 1842 1843 If unsure, say N. 1844 1845config TEST_USER_COPY 1846 tristate "Test user/kernel boundary protections" 1847 default n 1848 depends on m 1849 help 1850 This builds the "test_user_copy" module that runs sanity checks 1851 on the copy_to/from_user infrastructure, making sure basic 1852 user/kernel boundary testing is working. If it fails to load, 1853 a regression has been detected in the user/kernel memory boundary 1854 protections. 1855 1856 If unsure, say N. 1857 1858config TEST_BPF 1859 tristate "Test BPF filter functionality" 1860 default n 1861 depends on m && NET 1862 help 1863 This builds the "test_bpf" module that runs various test vectors 1864 against the BPF interpreter or BPF JIT compiler depending on the 1865 current setting. This is in particular useful for BPF JIT compiler 1866 development, but also to run regression tests against changes in 1867 the interpreter code. It also enables test stubs for eBPF maps and 1868 verifier used by user space verifier testsuite. 1869 1870 If unsure, say N. 1871 1872config TEST_FIRMWARE 1873 tristate "Test firmware loading via userspace interface" 1874 default n 1875 depends on FW_LOADER 1876 help 1877 This builds the "test_firmware" module that creates a userspace 1878 interface for testing firmware loading. This can be used to 1879 control the triggering of firmware loading without needing an 1880 actual firmware-using device. The contents can be rechecked by 1881 userspace. 1882 1883 If unsure, say N. 1884 1885config TEST_UDELAY 1886 tristate "udelay test driver" 1887 default n 1888 help 1889 This builds the "udelay_test" module that helps to make sure 1890 that udelay() is working properly. 1891 1892 If unsure, say N. 1893 1894config MEMTEST 1895 bool "Memtest" 1896 depends on HAVE_MEMBLOCK 1897 ---help--- 1898 This option adds a kernel parameter 'memtest', which allows memtest 1899 to be set. 1900 memtest=0, mean disabled; -- default 1901 memtest=1, mean do 1 test pattern; 1902 ... 1903 memtest=17, mean do 17 test patterns. 1904 If you are unsure how to answer this question, answer N. 1905 1906config TEST_STATIC_KEYS 1907 tristate "Test static keys" 1908 default n 1909 depends on m 1910 help 1911 Test the static key interfaces. 1912 1913 If unsure, say N. 1914 1915source "samples/Kconfig" 1916 1917source "lib/Kconfig.kgdb" 1918 1919source "lib/Kconfig.ubsan" 1920 1921config ARCH_HAS_DEVMEM_IS_ALLOWED 1922 bool 1923 1924config STRICT_DEVMEM 1925 bool "Filter access to /dev/mem" 1926 depends on MMU 1927 depends on ARCH_HAS_DEVMEM_IS_ALLOWED 1928 default y if TILE || PPC 1929 ---help--- 1930 If this option is disabled, you allow userspace (root) access to all 1931 of memory, including kernel and userspace memory. Accidental 1932 access to this is obviously disastrous, but specific access can 1933 be used by people debugging the kernel. Note that with PAT support 1934 enabled, even in this case there are restrictions on /dev/mem 1935 use due to the cache aliasing requirements. 1936 1937 If this option is switched on, and IO_STRICT_DEVMEM=n, the /dev/mem 1938 file only allows userspace access to PCI space and the BIOS code and 1939 data regions. This is sufficient for dosemu and X and all common 1940 users of /dev/mem. 1941 1942 If in doubt, say Y. 1943 1944config IO_STRICT_DEVMEM 1945 bool "Filter I/O access to /dev/mem" 1946 depends on STRICT_DEVMEM 1947 ---help--- 1948 If this option is disabled, you allow userspace (root) access to all 1949 io-memory regardless of whether a driver is actively using that 1950 range. Accidental access to this is obviously disastrous, but 1951 specific access can be used by people debugging kernel drivers. 1952 1953 If this option is switched on, the /dev/mem file only allows 1954 userspace access to *idle* io-memory ranges (see /proc/iomem) This 1955 may break traditional users of /dev/mem (dosemu, legacy X, etc...) 1956 if the driver using a given range cannot be disabled. 1957 1958 If in doubt, say Y. 1959