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_MUST_CHECK
13 bool "Enable __must_check logic"
16 Enable the __must_check logic in the kernel build. Disable this to
17 suppress the "warning: ignoring return value of 'foo', declared with
18 attribute warn_unused_result" messages.
21 bool "Magic SysRq key"
24 If you say Y here, you will have some control over the system even
25 if the system crashes for example during kernel debugging (e.g., you
26 will be able to flush the buffer cache to disk, reboot the system
27 immediately or dump some status information). This is accomplished
28 by pressing various keys while holding SysRq (Alt+PrintScreen). It
29 also works on a serial console (on PC hardware at least), if you
30 send a BREAK and then within 5 seconds a command keypress. The
31 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
32 unless you really know what this hack does.
35 bool "Enable unused/obsolete exported symbols"
38 Unused but exported symbols make the kernel needlessly bigger. For
39 that reason most of these unused exports will soon be removed. This
40 option is provided temporarily to provide a transition period in case
41 some external kernel module needs one of these symbols anyway. If you
42 encounter such a case in your module, consider if you are actually
43 using the right API. (rationale: since nobody in the kernel is using
44 this in a module, there is a pretty good chance it's actually the
45 wrong interface to use). If you really need the symbol, please send a
46 mail to the linux kernel mailing list mentioning the symbol and why
47 you really need it, and what the merge plan to the mainline kernel for
51 bool "Debug Filesystem"
54 debugfs is a virtual file system that kernel developers use to put
55 debugging files into. Enable this option to be able to read and
61 bool "Run 'make headers_check' when building vmlinux"
64 This option will extract the user-visible kernel headers whenever
65 building the kernel, and will run basic sanity checks on them to
66 ensure that exported files do not attempt to include files which
67 were not exported, etc.
69 If you're making modifications to header files which are
70 relevant for userspace, say 'Y', and check the headers
71 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
72 your build tree), to make sure they're suitable.
75 bool "Kernel debugging"
77 Say Y here if you are developing drivers or trying to debug and
78 identify kernel problems.
81 bool "Debug shared IRQ handlers"
82 depends on DEBUG_KERNEL && GENERIC_HARDIRQS
84 Enable this to generate a spurious interrupt as soon as a shared
85 interrupt handler is registered, and just before one is deregistered.
86 Drivers ought to be able to handle interrupts coming in at those
87 points; some don't and need to be caught.
90 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" if DEBUG_KERNEL
92 default 17 if S390 || LOCKDEP
93 default 16 if X86_NUMAQ || IA64
97 Select kernel log buffer size as a power of 2.
98 Defaults and Examples:
99 17 => 128 KB for S/390
100 16 => 64 KB for x86 NUMAQ or IA-64
102 14 => 16 KB for uniprocessor
106 config DETECT_SOFTLOCKUP
107 bool "Detect Soft Lockups"
108 depends on DEBUG_KERNEL && !S390
111 Say Y here to enable the kernel to detect "soft lockups",
112 which are bugs that cause the kernel to loop in kernel
113 mode for more than 10 seconds, without giving other tasks a
116 When a soft-lockup is detected, the kernel will print the
117 current stack trace (which you should report), but the
118 system will stay locked up. This feature has negligible
121 (Note that "hard lockups" are separate type of bugs that
122 can be detected via the NMI-watchdog, on platforms that
126 bool "Collect scheduler statistics"
127 depends on DEBUG_KERNEL && PROC_FS
129 If you say Y here, additional code will be inserted into the
130 scheduler and related routines to collect statistics about
131 scheduler behavior and provide them in /proc/schedstat. These
132 stats may be useful for both tuning and debugging the scheduler
133 If you aren't debugging the scheduler or trying to tune a specific
134 application, you can say N to avoid the very slight overhead
138 bool "Collect kernel timers statistics"
139 depends on DEBUG_KERNEL && PROC_FS
141 If you say Y here, additional code will be inserted into the
142 timer routines to collect statistics about kernel timers being
143 reprogrammed. The statistics can be read from /proc/timer_stats.
144 The statistics collection is started by writing 1 to /proc/timer_stats,
145 writing 0 stops it. This feature is useful to collect information
146 about timer usage patterns in kernel and userspace.
149 bool "Debug slab memory allocations"
150 depends on DEBUG_KERNEL && SLAB
152 Say Y here to have the kernel do limited verification on memory
153 allocation as well as poisoning memory on free to catch use of freed
154 memory. This can make kmalloc/kfree-intensive workloads much slower.
156 config DEBUG_SLAB_LEAK
157 bool "Memory leak debugging"
158 depends on DEBUG_SLAB
161 bool "Debug preemptible kernel"
162 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
165 If you say Y here then the kernel will use a debug variant of the
166 commonly used smp_processor_id() function and will print warnings
167 if kernel code uses it in a preemption-unsafe way. Also, the kernel
168 will detect preemption count underflows.
170 config DEBUG_RT_MUTEXES
171 bool "RT Mutex debugging, deadlock detection"
172 depends on DEBUG_KERNEL && RT_MUTEXES
174 This allows rt mutex semantics violations and rt mutex related
175 deadlocks (lockups) to be detected and reported automatically.
180 depends on DEBUG_RT_MUTEXES
182 config RT_MUTEX_TESTER
183 bool "Built-in scriptable tester for rt-mutexes"
184 depends on DEBUG_KERNEL && RT_MUTEXES
186 This option enables a rt-mutex tester.
188 config DEBUG_SPINLOCK
189 bool "Spinlock and rw-lock debugging: basic checks"
190 depends on DEBUG_KERNEL
192 Say Y here and build SMP to catch missing spinlock initialization
193 and certain other kinds of spinlock errors commonly made. This is
194 best used in conjunction with the NMI watchdog so that spinlock
195 deadlocks are also debuggable.
198 bool "Mutex debugging: basic checks"
199 depends on DEBUG_KERNEL
201 This feature allows mutex semantics violations to be detected and
204 config DEBUG_LOCK_ALLOC
205 bool "Lock debugging: detect incorrect freeing of live locks"
206 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
207 select DEBUG_SPINLOCK
211 This feature will check whether any held lock (spinlock, rwlock,
212 mutex or rwsem) is incorrectly freed by the kernel, via any of the
213 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
214 vfree(), etc.), whether a live lock is incorrectly reinitialized via
215 spin_lock_init()/mutex_init()/etc., or whether there is any lock
216 held during task exit.
219 bool "Lock debugging: prove locking correctness"
220 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
222 select DEBUG_SPINLOCK
224 select DEBUG_LOCK_ALLOC
227 This feature enables the kernel to prove that all locking
228 that occurs in the kernel runtime is mathematically
229 correct: that under no circumstance could an arbitrary (and
230 not yet triggered) combination of observed locking
231 sequences (on an arbitrary number of CPUs, running an
232 arbitrary number of tasks and interrupt contexts) cause a
235 In short, this feature enables the kernel to report locking
236 related deadlocks before they actually occur.
238 The proof does not depend on how hard and complex a
239 deadlock scenario would be to trigger: how many
240 participant CPUs, tasks and irq-contexts would be needed
241 for it to trigger. The proof also does not depend on
242 timing: if a race and a resulting deadlock is possible
243 theoretically (no matter how unlikely the race scenario
244 is), it will be proven so and will immediately be
245 reported by the kernel (once the event is observed that
246 makes the deadlock theoretically possible).
248 If a deadlock is impossible (i.e. the locking rules, as
249 observed by the kernel, are mathematically correct), the
250 kernel reports nothing.
252 NOTE: this feature can also be enabled for rwlocks, mutexes
253 and rwsems - in which case all dependencies between these
254 different locking variants are observed and mapped too, and
255 the proof of observed correctness is also maintained for an
256 arbitrary combination of these separate locking variants.
258 For more details, see Documentation/lockdep-design.txt.
262 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
264 select FRAME_POINTER if !X86
269 bool "Lock dependency engine debugging"
270 depends on DEBUG_KERNEL && LOCKDEP
272 If you say Y here, the lock dependency engine will do
273 additional runtime checks to debug itself, at the price
274 of more runtime overhead.
276 config TRACE_IRQFLAGS
277 depends on DEBUG_KERNEL
280 depends on TRACE_IRQFLAGS_SUPPORT
281 depends on PROVE_LOCKING
283 config DEBUG_SPINLOCK_SLEEP
284 bool "Spinlock debugging: sleep-inside-spinlock checking"
285 depends on DEBUG_KERNEL
287 If you say Y here, various routines which may sleep will become very
288 noisy if they are called with a spinlock held.
290 config DEBUG_LOCKING_API_SELFTESTS
291 bool "Locking API boot-time self-tests"
292 depends on DEBUG_KERNEL
294 Say Y here if you want the kernel to run a short self-test during
295 bootup. The self-test checks whether common types of locking bugs
296 are detected by debugging mechanisms or not. (if you disable
297 lock debugging then those bugs wont be detected of course.)
298 The following locking APIs are covered: spinlocks, rwlocks,
303 depends on DEBUG_KERNEL
304 depends on STACKTRACE_SUPPORT
307 bool "kobject debugging"
308 depends on DEBUG_KERNEL
310 If you say Y here, some extra kobject debugging messages will be sent
314 bool "Highmem debugging"
315 depends on DEBUG_KERNEL && HIGHMEM
317 This options enables addition error checking for high memory systems.
318 Disable for production systems.
320 config DEBUG_BUGVERBOSE
321 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED
323 depends on ARM || ARM26 || AVR32 || M32R || M68K || SPARC32 || SPARC64 || FRV || SUPERH || GENERIC_BUG
326 Say Y here to make BUG() panics output the file name and line number
327 of the BUG call as well as the EIP and oops trace. This aids
328 debugging but costs about 70-100K of memory.
331 bool "Compile the kernel with debug info"
332 depends on DEBUG_KERNEL
334 If you say Y here the resulting kernel image will include
335 debugging info resulting in a larger kernel image.
336 Say Y here only if you plan to debug the kernel.
342 depends on DEBUG_KERNEL
344 Enable this to turn on extended checks in the virtual-memory system
345 that may impact performance.
350 bool "Debug linked list manipulation"
351 depends on DEBUG_KERNEL
353 Enable this to turn on extended checks in the linked-list
359 bool "Compile the kernel with frame pointers"
360 depends on DEBUG_KERNEL && (X86 || CRIS || M68K || M68KNOMMU || FRV || UML || S390 || AVR32 || SUPERH)
361 default y if DEBUG_INFO && UML
363 If you say Y here the resulting kernel image will be slightly larger
364 and slower, but it might give very useful debugging information on
365 some architectures or if you use external debuggers.
366 If you don't debug the kernel, you can say N.
368 config FORCED_INLINING
369 bool "Force gcc to inline functions marked 'inline'"
370 depends on DEBUG_KERNEL
373 This option determines if the kernel forces gcc to inline the functions
374 developers have marked 'inline'. Doing so takes away freedom from gcc to
375 do what it thinks is best, which is desirable for the gcc 3.x series of
376 compilers. The gcc 4.x series have a rewritten inlining algorithm and
377 disabling this option will generate a smaller kernel there. Hopefully
378 this algorithm is so good that allowing gcc4 to make the decision can
379 become the default in the future, until then this option is there to
382 config RCU_TORTURE_TEST
383 tristate "torture tests for RCU"
384 depends on DEBUG_KERNEL
387 This option provides a kernel module that runs torture tests
388 on the RCU infrastructure. The kernel module may be built
389 after the fact on the running kernel to be tested, if desired.
391 Say Y here if you want RCU torture tests to start automatically
392 at boot time (you probably don't).
393 Say M if you want the RCU torture tests to build as a module.
394 Say N if you are unsure.
397 tristate "Linux Kernel Dump Test Tool Module"
398 depends on DEBUG_KERNEL
402 This module enables testing of the different dumping mechanisms by
403 inducing system failures at predefined crash points.
404 If you don't need it: say N
405 Choose M here to compile this code as a module. The module will be
408 Documentation on how to use the module can be found in
411 config FAULT_INJECTION
412 bool "Fault-injection framework"
413 depends on DEBUG_KERNEL
415 Provide fault-injection framework.
416 For more details, see Documentation/fault-injection/.
419 bool "Fault-injection capability for kmalloc"
420 depends on FAULT_INJECTION
422 Provide fault-injection capability for kmalloc.
424 config FAIL_PAGE_ALLOC
425 bool "Fault-injection capabilitiy for alloc_pages()"
426 depends on FAULT_INJECTION
428 Provide fault-injection capability for alloc_pages().
430 config FAIL_MAKE_REQUEST
431 bool "Fault-injection capability for disk IO"
432 depends on FAULT_INJECTION
434 Provide fault-injection capability for disk IO.
436 config FAULT_INJECTION_DEBUG_FS
437 bool "Debugfs entries for fault-injection capabilities"
438 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
440 Enable configuration of fault-injection capabilities via debugfs.
442 config FAULT_INJECTION_STACKTRACE_FILTER
443 bool "stacktrace filter for fault-injection capabilities"
444 depends on FAULT_INJECTION_DEBUG_FS && STACKTRACE_SUPPORT
448 Provide stacktrace filter for fault-injection capabilities