1 config DEBUG_SLEEP_IN_IRQ
2 bool "Warn if we sleep excessively in an irq"
4 Sleeping in an interrupt handler is a bad thing.
5 Enabling this option will emit a backtrace (once per instance).
8 bool "Show timing information on printks"
11 Selecting this option causes timing information to be
12 included in printk output. This allows you to measure
13 the interval between kernel operations, including bootup
14 operations. This is useful for identifying long delays
17 config ENABLE_MUST_CHECK
18 bool "Enable __must_check logic"
21 Enable the __must_check logic in the kernel build. Disable this to
22 suppress the "warning: ignoring return value of 'foo', declared with
23 attribute warn_unused_result" messages.
26 bool "Magic SysRq key"
29 If you say Y here, you will have some control over the system even
30 if the system crashes for example during kernel debugging (e.g., you
31 will be able to flush the buffer cache to disk, reboot the system
32 immediately or dump some status information). This is accomplished
33 by pressing various keys while holding SysRq (Alt+PrintScreen). It
34 also works on a serial console (on PC hardware at least), if you
35 send a BREAK and then within 5 seconds a command keypress. The
36 keys are documented in <file:Documentation/sysrq.txt>. Don't say Y
37 unless you really know what this hack does.
40 bool "Enable unused/obsolete exported symbols"
43 Unused but exported symbols make the kernel needlessly bigger. For
44 that reason most of these unused exports will soon be removed. This
45 option is provided temporarily to provide a transition period in case
46 some external kernel module needs one of these symbols anyway. If you
47 encounter such a case in your module, consider if you are actually
48 using the right API. (rationale: since nobody in the kernel is using
49 this in a module, there is a pretty good chance it's actually the
50 wrong interface to use). If you really need the symbol, please send a
51 mail to the linux kernel mailing list mentioning the symbol and why
52 you really need it, and what the merge plan to the mainline kernel for
56 bool "Debug Filesystem"
59 debugfs is a virtual file system that kernel developers use to put
60 debugging files into. Enable this option to be able to read and
66 bool "Run 'make headers_check' when building vmlinux"
69 This option will extract the user-visible kernel headers whenever
70 building the kernel, and will run basic sanity checks on them to
71 ensure that exported files do not attempt to include files which
72 were not exported, etc.
74 If you're making modifications to header files which are
75 relevant for userspace, say 'Y', and check the headers
76 exported to $(INSTALL_HDR_PATH) (usually 'usr/include' in
77 your build tree), to make sure they're suitable.
80 bool "Kernel debugging"
82 Say Y here if you are developing drivers or trying to debug and
83 identify kernel problems.
86 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" if DEBUG_KERNEL
88 default 17 if S390 || LOCKDEP
89 default 16 if X86_NUMAQ || IA64
93 Select kernel log buffer size as a power of 2.
94 Defaults and Examples:
95 17 => 128 KB for S/390
96 16 => 64 KB for x86 NUMAQ or IA-64
98 14 => 16 KB for uniprocessor
102 config DETECT_SOFTLOCKUP
103 bool "Detect Soft Lockups"
104 depends on DEBUG_KERNEL && !S390
107 Say Y here to enable the kernel to detect "soft lockups",
108 which are bugs that cause the kernel to loop in kernel
109 mode for more than 10 seconds, without giving other tasks a
112 When a soft-lockup is detected, the kernel will print the
113 current stack trace (which you should report), but the
114 system will stay locked up. This feature has negligible
117 (Note that "hard lockups" are separate type of bugs that
118 can be detected via the NMI-watchdog, on platforms that
122 bool "Collect scheduler statistics"
123 depends on DEBUG_KERNEL && PROC_FS
125 If you say Y here, additional code will be inserted into the
126 scheduler and related routines to collect statistics about
127 scheduler behavior and provide them in /proc/schedstat. These
128 stats may be useful for both tuning and debugging the scheduler
129 If you aren't debugging the scheduler or trying to tune a specific
130 application, you can say N to avoid the very slight overhead
134 bool "Debug slab memory allocations"
135 depends on DEBUG_KERNEL && SLAB
137 Say Y here to have the kernel do limited verification on memory
138 allocation as well as poisoning memory on free to catch use of freed
139 memory. This can make kmalloc/kfree-intensive workloads much slower.
141 config DEBUG_SLAB_LEAK
142 bool "Memory leak debugging"
143 depends on DEBUG_SLAB
146 bool "Debug preemptible kernel"
147 depends on DEBUG_KERNEL && PREEMPT && TRACE_IRQFLAGS_SUPPORT
150 If you say Y here then the kernel will use a debug variant of the
151 commonly used smp_processor_id() function and will print warnings
152 if kernel code uses it in a preemption-unsafe way. Also, the kernel
153 will detect preemption count underflows.
155 config DEBUG_RT_MUTEXES
156 bool "RT Mutex debugging, deadlock detection"
157 depends on DEBUG_KERNEL && RT_MUTEXES
159 This allows rt mutex semantics violations and rt mutex related
160 deadlocks (lockups) to be detected and reported automatically.
165 depends on DEBUG_RT_MUTEXES
167 config RT_MUTEX_TESTER
168 bool "Built-in scriptable tester for rt-mutexes"
169 depends on DEBUG_KERNEL && RT_MUTEXES
171 This option enables a rt-mutex tester.
173 config DEBUG_SPINLOCK
174 bool "Spinlock and rw-lock debugging: basic checks"
175 depends on DEBUG_KERNEL
177 Say Y here and build SMP to catch missing spinlock initialization
178 and certain other kinds of spinlock errors commonly made. This is
179 best used in conjunction with the NMI watchdog so that spinlock
180 deadlocks are also debuggable.
183 bool "Mutex debugging: basic checks"
184 depends on DEBUG_KERNEL
186 This feature allows mutex semantics violations to be detected and
190 bool "RW-sem debugging: basic checks"
191 depends on DEBUG_KERNEL
193 This feature allows read-write semaphore semantics violations to
194 be detected and reported.
196 config DEBUG_LOCK_ALLOC
197 bool "Lock debugging: detect incorrect freeing of live locks"
198 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
199 select DEBUG_SPINLOCK
204 This feature will check whether any held lock (spinlock, rwlock,
205 mutex or rwsem) is incorrectly freed by the kernel, via any of the
206 memory-freeing routines (kfree(), kmem_cache_free(), free_pages(),
207 vfree(), etc.), whether a live lock is incorrectly reinitialized via
208 spin_lock_init()/mutex_init()/etc., or whether there is any lock
209 held during task exit.
212 bool "Lock debugging: prove locking correctness"
213 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
215 select DEBUG_SPINLOCK
218 select DEBUG_LOCK_ALLOC
221 This feature enables the kernel to prove that all locking
222 that occurs in the kernel runtime is mathematically
223 correct: that under no circumstance could an arbitrary (and
224 not yet triggered) combination of observed locking
225 sequences (on an arbitrary number of CPUs, running an
226 arbitrary number of tasks and interrupt contexts) cause a
229 In short, this feature enables the kernel to report locking
230 related deadlocks before they actually occur.
232 The proof does not depend on how hard and complex a
233 deadlock scenario would be to trigger: how many
234 participant CPUs, tasks and irq-contexts would be needed
235 for it to trigger. The proof also does not depend on
236 timing: if a race and a resulting deadlock is possible
237 theoretically (no matter how unlikely the race scenario
238 is), it will be proven so and will immediately be
239 reported by the kernel (once the event is observed that
240 makes the deadlock theoretically possible).
242 If a deadlock is impossible (i.e. the locking rules, as
243 observed by the kernel, are mathematically correct), the
244 kernel reports nothing.
246 NOTE: this feature can also be enabled for rwlocks, mutexes
247 and rwsems - in which case all dependencies between these
248 different locking variants are observed and mapped too, and
249 the proof of observed correctness is also maintained for an
250 arbitrary combination of these separate locking variants.
252 For more details, see Documentation/lockdep-design.txt.
256 depends on DEBUG_KERNEL && TRACE_IRQFLAGS_SUPPORT && STACKTRACE_SUPPORT && LOCKDEP_SUPPORT
258 select FRAME_POINTER if !X86
263 bool "Lock dependency engine debugging"
264 depends on DEBUG_KERNEL && LOCKDEP
266 If you say Y here, the lock dependency engine will do
267 additional runtime checks to debug itself, at the price
268 of more runtime overhead.
270 config TRACE_IRQFLAGS
271 depends on DEBUG_KERNEL
274 depends on TRACE_IRQFLAGS_SUPPORT
275 depends on PROVE_LOCKING
277 config DEBUG_SPINLOCK_SLEEP
278 bool "Spinlock debugging: sleep-inside-spinlock checking"
279 depends on DEBUG_KERNEL
281 If you say Y here, various routines which may sleep will become very
282 noisy if they are called with a spinlock held.
284 config DEBUG_LOCKING_API_SELFTESTS
285 bool "Locking API boot-time self-tests"
286 depends on DEBUG_KERNEL
288 Say Y here if you want the kernel to run a short self-test during
289 bootup. The self-test checks whether common types of locking bugs
290 are detected by debugging mechanisms or not. (if you disable
291 lock debugging then those bugs wont be detected of course.)
292 The following locking APIs are covered: spinlocks, rwlocks,
297 depends on DEBUG_KERNEL
298 depends on STACKTRACE_SUPPORT
301 bool "kobject debugging"
302 depends on DEBUG_KERNEL
304 If you say Y here, some extra kobject debugging messages will be sent
308 bool "Highmem debugging"
309 depends on DEBUG_KERNEL && HIGHMEM
311 This options enables addition error checking for high memory systems.
312 Disable for production systems.
314 config DEBUG_BUGVERBOSE
315 bool "Verbose BUG() reporting (adds 70K)" if DEBUG_KERNEL && EMBEDDED
317 depends on ARM || ARM26 || AVR32 || M32R || M68K || SPARC32 || SPARC64 || FRV || SUPERH || GENERIC_BUG
320 Say Y here to make BUG() panics output the file name and line number
321 of the BUG call as well as the EIP and oops trace. This aids
322 debugging but costs about 70-100K of memory.
325 bool "Compile the kernel with debug info"
326 depends on DEBUG_KERNEL
328 If you say Y here the resulting kernel image will include
329 debugging info resulting in a larger kernel image.
330 Say Y here only if you plan to debug the kernel.
336 depends on DEBUG_KERNEL
338 Enable this to turn on extended checks in the virtual-memory system
339 that may impact performance.
344 bool "Debug linked list manipulation"
345 depends on DEBUG_KERNEL
347 Enable this to turn on extended checks in the linked-list
353 bool "Compile the kernel with frame pointers"
354 depends on DEBUG_KERNEL && (X86 || CRIS || M68K || M68KNOMMU || FRV || UML || S390 || AVR32 || SUPERH)
355 default y if DEBUG_INFO && UML
357 If you say Y here the resulting kernel image will be slightly larger
358 and slower, but it might give very useful debugging information on
359 some architectures or if you use external debuggers.
360 If you don't debug the kernel, you can say N.
362 config FORCED_INLINING
363 bool "Force gcc to inline functions marked 'inline'"
364 depends on DEBUG_KERNEL
367 This option determines if the kernel forces gcc to inline the functions
368 developers have marked 'inline'. Doing so takes away freedom from gcc to
369 do what it thinks is best, which is desirable for the gcc 3.x series of
370 compilers. The gcc 4.x series have a rewritten inlining algorithm and
371 disabling this option will generate a smaller kernel there. Hopefully
372 this algorithm is so good that allowing gcc4 to make the decision can
373 become the default in the future, until then this option is there to
377 bool "Delay each boot message by N milliseconds"
378 depends on DEBUG_KERNEL
380 This build option allows you to read kernel boot messages
381 by inserting a short delay after each one. The delay is
382 specified in milliseconds on the kernel command line,
383 using "boot_delay=N".
385 It is likely that you would also need to use "lpj=M" to preset
386 the "loops per jiffie" value.
387 See a previous boot log for the "lpj" value to use for your
388 system, and then set "lpj=M" before setting "boot_delay=N".
389 NOTE: Using this option may adversely affect SMP systems.
390 I.e., processors other than the first one may not boot up.
391 BOOT_DELAY also may cause DETECT_SOFTLOCKUP to detect
392 what it believes to be lockup conditions.
394 config RCU_TORTURE_TEST
395 tristate "torture tests for RCU"
396 depends on DEBUG_KERNEL
399 This option provides a kernel module that runs torture tests
400 on the RCU infrastructure. The kernel module may be built
401 after the fact on the running kernel to be tested, if desired.
403 Say Y here if you want RCU torture tests to start automatically
404 at boot time (you probably don't).
405 Say M if you want the RCU torture tests to build as a module.
406 Say N if you are unsure.
409 tristate "Linux Kernel Dump Test Tool Module"
410 depends on DEBUG_KERNEL
414 This module enables testing of the different dumping mechanisms by
415 inducing system failures at predefined crash points.
416 If you don't need it: say N
417 Choose M here to compile this code as a module. The module will be
420 Documentation on how to use the module can be found in
423 config FAULT_INJECTION
424 bool "Fault-injection framework"
425 depends on DEBUG_KERNEL && STACKTRACE_SUPPORT
429 Provide fault-injection framework.
430 For more details, see Documentation/fault-injection/.
433 bool "Fault-injection capability for kmalloc"
434 depends on FAULT_INJECTION
436 Provide fault-injection capability for kmalloc.
438 config FAIL_PAGE_ALLOC
439 bool "Fault-injection capabilitiy for alloc_pages()"
440 depends on FAULT_INJECTION
442 Provide fault-injection capability for alloc_pages().
444 config FAIL_MAKE_REQUEST
445 bool "Fault-injection capability for disk IO"
446 depends on FAULT_INJECTION
448 Provide fault-injection capability for disk IO.
450 config FAULT_INJECTION_DEBUG_FS
451 bool "Debugfs entries for fault-injection capabilities"
452 depends on FAULT_INJECTION && SYSFS && DEBUG_FS
454 Enable configuration of fault-injection capabilities via debugfs.