3 bool "Show timing information on printks"
6 Selecting this option causes timing information to be
7 included in printk output. This allows you to measure
8 the interval between kernel operations, including bootup
9 operations. This is useful for identifying long delays
12 config ENABLE_WARN_DEPRECATED
13 bool "Enable __deprecated logic"
16 Enable the __deprecated logic in the kernel build.
17 Disable this to suppress the "warning: 'foo' is deprecated
18 (declared at kernel/power/somefile.c:1234)" messages.
20 config ENABLE_MUST_CHECK
21 bool "Enable __must_check logic"
24 Enable the __must_check logic in the kernel build. Disable this to
25 suppress the "warning: ignoring return value of 'foo', declared with
26 attribute warn_unused_result" messages.
29 bool "Magic SysRq key"
32 If you say Y here, you will have some control over the system even
33 if the system crashes for example during kernel debugging (e.g., you
34 will be able to flush the buffer cache to disk, reboot the system
35 immediately or dump some status information). This is accomplished
36 by pressing various keys while holding SysRq (Alt+PrintScreen). It
37 also works on a serial console (on PC hardware at least), if you
38 send a BREAK and then within 5 seconds a command keypress. The
39 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
40 unless you really know what this hack does.
43 bool "Enable unused/obsolete exported symbols"
46 Unused but exported symbols make the kernel needlessly bigger. For
47 that reason most of these unused exports will soon be removed. This
48 option is provided temporarily to provide a transition period in case
49 some external kernel module needs one of these symbols anyway. If you
50 encounter such a case in your module, consider if you are actually
51 using the right API. (rationale: since nobody in the kernel is using
52 this in a module, there is a pretty good chance it's actually the
53 wrong interface to use). If you really need the symbol, please send a
54 mail to the linux kernel mailing list mentioning the symbol and why
55 you really need it, and what the merge plan to the mainline kernel for
59 bool "Debug Filesystem"
62 debugfs is a virtual file system that kernel developers use to put
63 debugging files into. Enable this option to be able to read and
69 bool "Run 'make headers_check' when building vmlinux"
72 This option will extract the user-visible kernel headers whenever
73 building the kernel, and will run basic sanity checks on them to
74 ensure that exported files do not attempt to include files which
75 were not exported, etc.
77 If you're making modifications to header files which are
78 relevant for userspace, say 'Y', and check the headers
79 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
80 your build tree), to make sure they're suitable.
82 config DEBUG_SECTION_MISMATCH
83 bool "Enable full Section mismatch analysis"
86 The section mismatch analysis checks if there are illegal
87 references from one section to another section.
88 Linux will during link or during runtime drop some sections
89 and any use of code/data previously in these sections will
90 most likely result in an oops.
91 In the code functions and variables are annotated with
92 __init, __devinit etc. (see full list in include/linux/init.h)
93 which results in the code/data being placed in specific sections.
94 The section mismatch analysis is always done after a full
95 kernel build but enabling this option will in addition
97 - Add the option -fno-inline-functions-called-once to gcc
98 When inlining a function annotated __init in a non-init
99 function we would lose the section information and thus
100 the analysis would not catch the illegal reference.
101 This option tells gcc to inline less but will also
102 result in a larger kernel.
103 - Run the section mismatch analysis for each module/built-in.o
104 When we run the section mismatch analysis on vmlinux.o we
105 lose valueble information about where the mismatch was
107 Running the analysis for each module/built-in.o file
108 will tell where the mismatch happens much closer to the
109 source. The drawback is that we will report the same
110 mismatch at least twice.
111 - Enable verbose reporting from modpost to help solving
112 the section mismatches reported.
115 bool "Kernel debugging"
117 Say Y here if you are developing drivers or trying to debug and
118 identify kernel problems.
121 bool "Debug shared IRQ handlers"
122 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
124 Enable this to generate a spurious interrupt as soon as a shared
125 interrupt handler is registered, and just before one is deregistered.
126 Drivers ought to be able to handle interrupts coming in at those
127 points; some don't and need to be caught.
129 config DETECT_SOFTLOCKUP
130 bool "Detect Soft Lockups"
131 depends on DEBUG_KERNEL && !S390
134 Say Y here to enable the kernel to detect "soft lockups",
135 which are bugs that cause the kernel to loop in kernel
136 mode for more than 10 seconds, without giving other tasks a
139 When a soft-lockup is detected, the kernel will print the
140 current stack trace (which you should report), but the
141 system will stay locked up. This feature has negligible
144 (Note that "hard lockups" are separate type of bugs that
145 can be detected via the NMI-watchdog, on platforms that
149 bool "Collect scheduler debugging info"
150 depends on DEBUG_KERNEL && PROC_FS
153 If you say Y here, the /proc/sched_debug file will be provided
154 that can help debug the scheduler. The runtime overhead of this
158 bool "Collect scheduler statistics"
159 depends on DEBUG_KERNEL && PROC_FS
161 If you say Y here, additional code will be inserted into the
162 scheduler and related routines to collect statistics about
163 scheduler behavior and provide them in /proc/schedstat. These
164 stats may be useful for both tuning and debugging the scheduler
165 If you aren't debugging the scheduler or trying to tune a specific
166 application, you can say N to avoid the very slight overhead
170 bool "Collect kernel timers statistics"
171 depends on DEBUG_KERNEL && PROC_FS
173 If you say Y here, additional code will be inserted into the
174 timer routines to collect statistics about kernel timers being
175 reprogrammed. The statistics can be read from /proc/timer_stats.
176 The statistics collection is started by writing 1 to /proc/timer_stats,
177 writing 0 stops it. This feature is useful to collect information
178 about timer usage patterns in kernel and userspace. This feature
179 is lightweight if enabled in the kernel config but not activated
180 (it defaults to deactivated on bootup and will only be activated
181 if some application like powertop activates it explicitly).
184 bool "Debug slab memory allocations"
185 depends on DEBUG_KERNEL && SLAB
187 Say Y here to have the kernel do limited verification on memory
188 allocation as well as poisoning memory on free to catch use of freed
189 memory. This can make kmalloc/kfree-intensive workloads much slower.
191 config DEBUG_SLAB_LEAK
192 bool "Memory leak debugging"
193 depends on DEBUG_SLAB
196 bool "SLUB debugging on by default"
197 depends on SLUB && SLUB_DEBUG
200 Boot with debugging on by default. SLUB boots by default with
201 the runtime debug capabilities switched off. Enabling this is
202 equivalent to specifying the "slub_debug" parameter on boot.
203 There is no support for more fine grained debug control like
204 possible with slub_debug=xxx. SLUB debugging may be switched
205 off in a kernel built with CONFIG_SLUB_DEBUG_ON by specifying
210 bool "Enable SLUB performance statistics"
213 SLUB statistics are useful to debug SLUBs allocation behavior in
214 order find ways to optimize the allocator. This should never be
215 enabled for production use since keeping statistics slows down
216 the allocator by a few percentage points. The slabinfo command
217 supports the determination of the most active slabs to figure
218 out which slabs are relevant to a particular load.
219 Try running: slabinfo -DA
222 bool "Debug preemptible kernel"
223 depends on DEBUG_KERNEL && PREEMPT && (TRACE_IRQFLAGS_SUPPORT || PPC64)
226 If you say Y here then the kernel will use a debug variant of the
227 commonly used smp_processor_id() function and will print warnings
228 if kernel code uses it in a preemption-unsafe way. Also, the kernel
229 will detect preemption count underflows.
231 config DEBUG_RT_MUTEXES
232 bool "RT Mutex debugging, deadlock detection"
233 depends on DEBUG_KERNEL && RT_MUTEXES
235 This allows rt mutex semantics violations and rt mutex related
236 deadlocks (lockups) to be detected and reported automatically.
241 depends on DEBUG_RT_MUTEXES
243 config RT_MUTEX_TESTER
244 bool "Built-in scriptable tester for rt-mutexes"
245 depends on DEBUG_KERNEL && RT_MUTEXES
247 This option enables a rt-mutex tester.
249 config DEBUG_SPINLOCK
250 bool "Spinlock and rw-lock debugging: basic checks"
251 depends on DEBUG_KERNEL
253 Say Y here and build SMP to catch missing spinlock initialization
254 and certain other kinds of spinlock errors commonly made. This is
255 best used in conjunction with the NMI watchdog so that spinlock
256 deadlocks are also debuggable.
259 bool "Mutex debugging: basic checks"
260 depends on DEBUG_KERNEL
262 This feature allows mutex semantics violations to be detected and
265 config DEBUG_SEMAPHORE
266 bool "Semaphore debugging"
267 depends on DEBUG_KERNEL
268 depends on ALPHA || FRV
271 If you say Y here then semaphore processing will issue lots of
272 verbose debugging messages. If you suspect a semaphore problem or a
273 kernel hacker asks for this option then say Y. Otherwise say N.
275 config DEBUG_LOCK_ALLOC
276 bool "Lock debugging: detect incorrect freeing of live locks"
277 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
278 select DEBUG_SPINLOCK
282 This feature will check whether any held lock (spinlock, rwlock,
283 mutex or rwsem) is incorrectly freed by the kernel, via any of the
284 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
285 vfree(), etc.), whether a live lock is incorrectly reinitialized via
286 spin_lock_init()/mutex_init()/etc., or whether there is any lock
287 held during task exit.
290 bool "Lock debugging: prove locking correctness"
291 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
293 select DEBUG_SPINLOCK
295 select DEBUG_LOCK_ALLOC
298 This feature enables the kernel to prove that all locking
299 that occurs in the kernel runtime is mathematically
300 correct: that under no circumstance could an arbitrary (and
301 not yet triggered) combination of observed locking
302 sequences (on an arbitrary number of CPUs, running an
303 arbitrary number of tasks and interrupt contexts) cause a
306 In short, this feature enables the kernel to report locking
307 related deadlocks before they actually occur.
309 The proof does not depend on how hard and complex a
310 deadlock scenario would be to trigger: how many
311 participant CPUs, tasks and irq-contexts would be needed
312 for it to trigger. The proof also does not depend on
313 timing: if a race and a resulting deadlock is possible
314 theoretically (no matter how unlikely the race scenario
315 is), it will be proven so and will immediately be
316 reported by the kernel (once the event is observed that
317 makes the deadlock theoretically possible).
319 If a deadlock is impossible (i.e. the locking rules, as
320 observed by the kernel, are mathematically correct), the
321 kernel reports nothing.
323 NOTE: this feature can also be enabled for rwlocks, mutexes
324 and rwsems - in which case all dependencies between these
325 different locking variants are observed and mapped too, and
326 the proof of observed correctness is also maintained for an
327 arbitrary combination of these separate locking variants.
329 For more details, see Documentation/lockdep-design.txt.
333 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
335 select FRAME_POINTER if !X86 && !MIPS
340 bool "Lock usage statistics"
341 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
343 select DEBUG_SPINLOCK
345 select DEBUG_LOCK_ALLOC
348 This feature enables tracking lock contention points
350 For more details, see Documentation/lockstat.txt
353 bool "Lock dependency engine debugging"
354 depends on DEBUG_KERNEL && LOCKDEP
356 If you say Y here, the lock dependency engine will do
357 additional runtime checks to debug itself, at the price
358 of more runtime overhead.
360 config TRACE_IRQFLAGS
361 depends on DEBUG_KERNEL
364 depends on TRACE_IRQFLAGS_SUPPORT
365 depends on PROVE_LOCKING
367 config DEBUG_SPINLOCK_SLEEP
368 bool "Spinlock debugging: sleep-inside-spinlock checking"
369 depends on DEBUG_KERNEL
371 If you say Y here, various routines which may sleep will become very
372 noisy if they are called with a spinlock held.
374 config DEBUG_LOCKING_API_SELFTESTS
375 bool "Locking API boot-time self-tests"
376 depends on DEBUG_KERNEL
378 Say Y here if you want the kernel to run a short self-test during
379 bootup. The self-test checks whether common types of locking bugs
380 are detected by debugging mechanisms or not. (if you disable
381 lock debugging then those bugs wont be detected of course.)
382 The following locking APIs are covered: spinlocks, rwlocks,
387 depends on DEBUG_KERNEL
388 depends on STACKTRACE_SUPPORT
391 bool "kobject debugging"
392 depends on DEBUG_KERNEL
394 If you say Y here, some extra kobject debugging messages will be sent
398 bool "Highmem debugging"
399 depends on DEBUG_KERNEL && HIGHMEM
401 This options enables addition error checking for high memory systems.
402 Disable for production systems.
404 config DEBUG_BUGVERBOSE
405 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED
407 depends on ARM || AVR32 || M32R || M68K || SPARC32 || SPARC64 || FRV || SUPERH || GENERIC_BUG || BLACKFIN
410 Say Y here to make BUG() panics output the file name and line number
411 of the BUG call as well as the EIP and oops trace. This aids
412 debugging but costs about 70-100K of memory.
415 bool "Compile the kernel with debug info"
416 depends on DEBUG_KERNEL
418 If you say Y here the resulting kernel image will include
419 debugging info resulting in a larger kernel image.
420 This adds debug symbols to the kernel and modules (gcc -g), and
421 is needed if you intend to use kernel crashdump or binary object
422 tools like crash, kgdb, LKCD, gdb, etc on the kernel.
423 Say Y here only if you plan to debug the kernel.
429 depends on DEBUG_KERNEL
431 Enable this to turn on extended checks in the virtual-memory system
432 that may impact performance.
437 bool "Debug linked list manipulation"
438 depends on DEBUG_KERNEL
440 Enable this to turn on extended checks in the linked-list
446 bool "Debug SG table operations"
447 depends on DEBUG_KERNEL
449 Enable this to turn on checks on scatter-gather tables. This can
450 help find problems with drivers that do not properly initialize
456 bool "Compile the kernel with frame pointers"
457 depends on DEBUG_KERNEL && (X86 || CRIS || M68K || M68KNOMMU || FRV || UML || S390 || AVR32 || SUPERH || BLACKFIN)
458 default y if DEBUG_INFO && UML
460 If you say Y here the resulting kernel image will be slightly larger
461 and slower, but it might give very useful debugging information on
462 some architectures or if you use external debuggers.
463 If you don't debug the kernel, you can say N.
465 config FORCED_INLINING
466 bool "Force gcc to inline functions marked 'inline'"
467 depends on DEBUG_KERNEL
470 This option determines if the kernel forces gcc to inline the functions
471 developers have marked 'inline'. Doing so takes away freedom from gcc to
472 do what it thinks is best, which is desirable for the gcc 3.x series of
473 compilers. The gcc 4.x series have a rewritten inlining algorithm and
474 disabling this option will generate a smaller kernel there. Hopefully
475 this algorithm is so good that allowing gcc4 to make the decision can
476 become the default in the future, until then this option is there to
479 config BOOT_PRINTK_DELAY
480 bool "Delay each boot printk message by N milliseconds"
481 depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
483 This build option allows you to read kernel boot messages
484 by inserting a short delay after each one. The delay is
485 specified in milliseconds on the kernel command line,
486 using "boot_delay=N".
488 It is likely that you would also need to use "lpj=M" to preset
489 the "loops per jiffie" value.
490 See a previous boot log for the "lpj" value to use for your
491 system, and then set "lpj=M" before setting "boot_delay=N".
492 NOTE: Using this option may adversely affect SMP systems.
493 I.e., processors other than the first one may not boot up.
494 BOOT_PRINTK_DELAY also may cause DETECT_SOFTLOCKUP to detect
495 what it believes to be lockup conditions.
497 config RCU_TORTURE_TEST
498 tristate "torture tests for RCU"
499 depends on DEBUG_KERNEL
503 This option provides a kernel module that runs torture tests
504 on the RCU infrastructure. The kernel module may be built
505 after the fact on the running kernel to be tested, if desired.
507 Say M if you want the RCU torture tests to build as a module.
508 Say N if you are unsure.
510 config KPROBES_SANITY_TEST
511 bool "Kprobes sanity tests"
512 depends on DEBUG_KERNEL
516 This option provides for testing basic kprobes functionality on
517 boot. A sample kprobe, jprobe and kretprobe are inserted and
518 verified for functionality.
520 Say N if you are unsure.
522 config BACKTRACE_SELF_TEST
523 tristate "Self test for the backtrace code"
524 depends on DEBUG_KERNEL
527 This option provides a kernel module that can be used to test
528 the kernel stack backtrace code. This option is not useful
529 for distributions or general kernels, but only for kernel
530 developers working on architecture code.
532 Say N if you are unsure.
535 tristate "Linux Kernel Dump Test Tool Module"
536 depends on DEBUG_KERNEL
540 This module enables testing of the different dumping mechanisms by
541 inducing system failures at predefined crash points.
542 If you don't need it: say N
543 Choose M here to compile this code as a module. The module will be
546 Documentation on how to use the module can be found in
549 config FAULT_INJECTION
550 bool "Fault-injection framework"
551 depends on DEBUG_KERNEL
553 Provide fault-injection framework.
554 For more details, see Documentation/fault-injection/.
557 bool "Fault-injection capability for kmalloc"
558 depends on FAULT_INJECTION
560 Provide fault-injection capability for kmalloc.
562 config FAIL_PAGE_ALLOC
563 bool "Fault-injection capabilitiy for alloc_pages()"
564 depends on FAULT_INJECTION
566 Provide fault-injection capability for alloc_pages().
568 config FAIL_MAKE_REQUEST
569 bool "Fault-injection capability for disk IO"
570 depends on FAULT_INJECTION
572 Provide fault-injection capability for disk IO.
574 config FAULT_INJECTION_DEBUG_FS
575 bool "Debugfs entries for fault-injection capabilities"
576 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
578 Enable configuration of fault-injection capabilities via debugfs.
580 config FAULT_INJECTION_STACKTRACE_FILTER
581 bool "stacktrace filter for fault-injection capabilities"
582 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
587 Provide stacktrace filter for fault-injection capabilities
590 bool "Latency measuring infrastructure"
591 select FRAME_POINTER if !MIPS
597 depends on HAVE_LATENCYTOP_SUPPORT
599 Enable this option if you want to use the LatencyTOP tool
600 to find out which userspace is blocking on what kernel operations.
602 config PROVIDE_OHCI1394_DMA_INIT
603 bool "Provide code for enabling DMA over FireWire early on boot"
604 depends on PCI && X86
606 If you want to debug problems which hang or crash the kernel early
607 on boot and the crashing machine has a FireWire port, you can use
608 this feature to remotely access the memory of the crashed machine
609 over FireWire. This employs remote DMA as part of the OHCI1394
610 specification which is now the standard for FireWire controllers.
612 With remote DMA, you can monitor the printk buffer remotely using
613 firescope and access all memory below 4GB using fireproxy from gdb.
614 Even controlling a kernel debugger is possible using remote DMA.
618 If ohci1394_dma=early is used as boot parameter, it will initialize
619 all OHCI1394 controllers which are found in the PCI config space.
621 As all changes to the FireWire bus such as enabling and disabling
622 devices cause a bus reset and thereby disable remote DMA for all
623 devices, be sure to have the cable plugged and FireWire enabled on
624 the debugging host before booting the debug target for debugging.
626 This code (~1k) is freed after boot. By then, the firewire stack
627 in charge of the OHCI-1394 controllers should be used instead.
629 See Documentation/debugging-via-ohci1394.txt for more information.
631 source "samples/Kconfig"