2 # General architecture dependent options
6 tristate "OProfile system profiling"
8 depends on HAVE_OPROFILE
10 select RING_BUFFER_ALLOW_SWAP
12 OProfile is a profiling system capable of profiling the
13 whole system, include the kernel, kernel modules, libraries,
18 config OPROFILE_EVENT_MULTIPLEX
19 bool "OProfile multiplexing support (EXPERIMENTAL)"
21 depends on OPROFILE && X86
23 The number of hardware counters is limited. The multiplexing
24 feature enables OProfile to gather more events than counters
25 are provided by the hardware. This is realized by switching
26 between events at an user specified time interval.
33 config OPROFILE_NMI_TIMER
35 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
40 depends on HAVE_KPROBES
43 Kprobes allows you to trap at almost any kernel address and
44 execute a callback function. register_kprobe() establishes
45 a probepoint and specifies the callback. Kprobes is useful
46 for kernel debugging, non-intrusive instrumentation and testing.
50 bool "Optimize very unlikely/likely branches"
51 depends on HAVE_ARCH_JUMP_LABEL
53 This option enables a transparent branch optimization that
54 makes certain almost-always-true or almost-always-false branch
55 conditions even cheaper to execute within the kernel.
57 Certain performance-sensitive kernel code, such as trace points,
58 scheduler functionality, networking code and KVM have such
59 branches and include support for this optimization technique.
61 If it is detected that the compiler has support for "asm goto",
62 the kernel will compile such branches with just a nop
63 instruction. When the condition flag is toggled to true, the
64 nop will be converted to a jump instruction to execute the
65 conditional block of instructions.
67 This technique lowers overhead and stress on the branch prediction
68 of the processor and generally makes the kernel faster. The update
69 of the condition is slower, but those are always very rare.
71 ( On 32-bit x86, the necessary options added to the compiler
72 flags may increase the size of the kernel slightly. )
74 config STATIC_KEYS_SELFTEST
75 bool "Static key selftest"
78 Boot time self-test of the branch patching code.
82 depends on KPROBES && HAVE_OPTPROBES
85 config KPROBES_ON_FTRACE
87 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
88 depends on DYNAMIC_FTRACE_WITH_REGS
90 If function tracer is enabled and the arch supports full
91 passing of pt_regs to function tracing, then kprobes can
92 optimize on top of function tracing.
98 Uprobes is the user-space counterpart to kprobes: they
99 enable instrumentation applications (such as 'perf probe')
100 to establish unintrusive probes in user-space binaries and
101 libraries, by executing handler functions when the probes
102 are hit by user-space applications.
104 ( These probes come in the form of single-byte breakpoints,
105 managed by the kernel and kept transparent to the probed
108 config HAVE_64BIT_ALIGNED_ACCESS
109 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
111 Some architectures require 64 bit accesses to be 64 bit
112 aligned, which also requires structs containing 64 bit values
113 to be 64 bit aligned too. This includes some 32 bit
114 architectures which can do 64 bit accesses, as well as 64 bit
115 architectures without unaligned access.
117 This symbol should be selected by an architecture if 64 bit
118 accesses are required to be 64 bit aligned in this way even
119 though it is not a 64 bit architecture.
121 See Documentation/unaligned-memory-access.txt for more
122 information on the topic of unaligned memory accesses.
124 config HAVE_EFFICIENT_UNALIGNED_ACCESS
127 Some architectures are unable to perform unaligned accesses
128 without the use of get_unaligned/put_unaligned. Others are
129 unable to perform such accesses efficiently (e.g. trap on
130 unaligned access and require fixing it up in the exception
133 This symbol should be selected by an architecture if it can
134 perform unaligned accesses efficiently to allow different
135 code paths to be selected for these cases. Some network
136 drivers, for example, could opt to not fix up alignment
137 problems with received packets if doing so would not help
140 See Documentation/unaligned-memory-access.txt for more
141 information on the topic of unaligned memory accesses.
143 config ARCH_USE_BUILTIN_BSWAP
146 Modern versions of GCC (since 4.4) have builtin functions
147 for handling byte-swapping. Using these, instead of the old
148 inline assembler that the architecture code provides in the
149 __arch_bswapXX() macros, allows the compiler to see what's
150 happening and offers more opportunity for optimisation. In
151 particular, the compiler will be able to combine the byteswap
152 with a nearby load or store and use load-and-swap or
153 store-and-swap instructions if the architecture has them. It
154 should almost *never* result in code which is worse than the
155 hand-coded assembler in <asm/swab.h>. But just in case it
156 does, the use of the builtins is optional.
158 Any architecture with load-and-swap or store-and-swap
159 instructions should set this. And it shouldn't hurt to set it
160 on architectures that don't have such instructions.
164 depends on KPROBES && HAVE_KRETPROBES
166 config USER_RETURN_NOTIFIER
168 depends on HAVE_USER_RETURN_NOTIFIER
170 Provide a kernel-internal notification when a cpu is about to
173 config HAVE_IOREMAP_PROT
179 config HAVE_KRETPROBES
182 config HAVE_OPTPROBES
185 config HAVE_KPROBES_ON_FTRACE
188 config HAVE_NMI_WATCHDOG
191 # An arch should select this if it provides all these things:
193 # task_pt_regs() in asm/processor.h or asm/ptrace.h
194 # arch_has_single_step() if there is hardware single-step support
195 # arch_has_block_step() if there is hardware block-step support
196 # asm/syscall.h supplying asm-generic/syscall.h interface
197 # linux/regset.h user_regset interfaces
198 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
199 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
200 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
201 # signal delivery calls tracehook_signal_handler()
203 config HAVE_ARCH_TRACEHOOK
206 config HAVE_DMA_ATTRS
209 config HAVE_DMA_CONTIGUOUS
212 config GENERIC_SMP_IDLE_THREAD
215 config GENERIC_IDLE_POLL_SETUP
218 # Select if arch init_task initializer is different to init/init_task.c
219 config ARCH_INIT_TASK
222 # Select if arch has its private alloc_task_struct() function
223 config ARCH_TASK_STRUCT_ALLOCATOR
226 # Select if arch has its private alloc_thread_info() function
227 config ARCH_THREAD_INFO_ALLOCATOR
230 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
231 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
234 config HAVE_REGS_AND_STACK_ACCESS_API
237 This symbol should be selected by an architecure if it supports
238 the API needed to access registers and stack entries from pt_regs,
239 declared in asm/ptrace.h
240 For example the kprobes-based event tracer needs this API.
245 The <linux/clk.h> calls support software clock gating and
246 thus are a key power management tool on many systems.
248 config HAVE_DMA_API_DEBUG
251 config HAVE_HW_BREAKPOINT
253 depends on PERF_EVENTS
255 config HAVE_MIXED_BREAKPOINTS_REGS
257 depends on HAVE_HW_BREAKPOINT
259 Depending on the arch implementation of hardware breakpoints,
260 some of them have separate registers for data and instruction
261 breakpoints addresses, others have mixed registers to store
262 them but define the access type in a control register.
263 Select this option if your arch implements breakpoints under the
266 config HAVE_USER_RETURN_NOTIFIER
269 config HAVE_PERF_EVENTS_NMI
272 System hardware can generate an NMI using the perf event
273 subsystem. Also has support for calculating CPU cycle events
274 to determine how many clock cycles in a given period.
276 config HAVE_PERF_REGS
279 Support selective register dumps for perf events. This includes
280 bit-mapping of each registers and a unique architecture id.
282 config HAVE_PERF_USER_STACK_DUMP
285 Support user stack dumps for perf event samples. This needs
286 access to the user stack pointer which is not unified across
289 config HAVE_ARCH_JUMP_LABEL
292 config HAVE_RCU_TABLE_FREE
295 config ARCH_HAVE_NMI_SAFE_CMPXCHG
298 config HAVE_ALIGNED_STRUCT_PAGE
301 This makes sure that struct pages are double word aligned and that
302 e.g. the SLUB allocator can perform double word atomic operations
303 on a struct page for better performance. However selecting this
304 might increase the size of a struct page by a word.
306 config HAVE_CMPXCHG_LOCAL
309 config HAVE_CMPXCHG_DOUBLE
312 config ARCH_WANT_IPC_PARSE_VERSION
315 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
318 config ARCH_WANT_OLD_COMPAT_IPC
319 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
322 config HAVE_ARCH_SECCOMP_FILTER
325 An arch should select this symbol if it provides all of these things:
327 - syscall_get_arguments()
329 - syscall_set_return_value()
330 - SIGSYS siginfo_t support
331 - secure_computing is called from a ptrace_event()-safe context
332 - secure_computing return value is checked and a return value of -1
333 results in the system call being skipped immediately.
334 - seccomp syscall wired up
336 For best performance, an arch should use seccomp_phase1 and
337 seccomp_phase2 directly. It should call seccomp_phase1 for all
338 syscalls if TIF_SECCOMP is set, but seccomp_phase1 does not
339 need to be called from a ptrace-safe context. It must then
340 call seccomp_phase2 if seccomp_phase1 returns anything other
341 than SECCOMP_PHASE1_OK or SECCOMP_PHASE1_SKIP.
343 As an additional optimization, an arch may provide seccomp_data
344 directly to seccomp_phase1; this avoids multiple calls
345 to the syscall_xyz helpers for every syscall.
347 config SECCOMP_FILTER
349 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
351 Enable tasks to build secure computing environments defined
352 in terms of Berkeley Packet Filter programs which implement
353 task-defined system call filtering polices.
355 See Documentation/prctl/seccomp_filter.txt for details.
357 config HAVE_CC_STACKPROTECTOR
360 An arch should select this symbol if:
361 - its compiler supports the -fstack-protector option
362 - it has implemented a stack canary (e.g. __stack_chk_guard)
364 config CC_STACKPROTECTOR
367 Set when a stack-protector mode is enabled, so that the build
368 can enable kernel-side support for the GCC feature.
371 prompt "Stack Protector buffer overflow detection"
372 depends on HAVE_CC_STACKPROTECTOR
373 default CC_STACKPROTECTOR_NONE
375 This option turns on the "stack-protector" GCC feature. This
376 feature puts, at the beginning of functions, a canary value on
377 the stack just before the return address, and validates
378 the value just before actually returning. Stack based buffer
379 overflows (that need to overwrite this return address) now also
380 overwrite the canary, which gets detected and the attack is then
381 neutralized via a kernel panic.
383 config CC_STACKPROTECTOR_NONE
386 Disable "stack-protector" GCC feature.
388 config CC_STACKPROTECTOR_REGULAR
390 select CC_STACKPROTECTOR
392 Functions will have the stack-protector canary logic added if they
393 have an 8-byte or larger character array on the stack.
395 This feature requires gcc version 4.2 or above, or a distribution
396 gcc with the feature backported ("-fstack-protector").
398 On an x86 "defconfig" build, this feature adds canary checks to
399 about 3% of all kernel functions, which increases kernel code size
402 config CC_STACKPROTECTOR_STRONG
404 select CC_STACKPROTECTOR
406 Functions will have the stack-protector canary logic added in any
407 of the following conditions:
409 - local variable's address used as part of the right hand side of an
410 assignment or function argument
411 - local variable is an array (or union containing an array),
412 regardless of array type or length
413 - uses register local variables
415 This feature requires gcc version 4.9 or above, or a distribution
416 gcc with the feature backported ("-fstack-protector-strong").
418 On an x86 "defconfig" build, this feature adds canary checks to
419 about 20% of all kernel functions, which increases the kernel code
424 config HAVE_CONTEXT_TRACKING
427 Provide kernel/user boundaries probes necessary for subsystems
428 that need it, such as userspace RCU extended quiescent state.
429 Syscalls need to be wrapped inside user_exit()-user_enter() through
430 the slow path using TIF_NOHZ flag. Exceptions handlers must be
431 wrapped as well. Irqs are already protected inside
432 rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on
433 irq exit still need to be protected.
435 config HAVE_VIRT_CPU_ACCOUNTING
438 config HAVE_VIRT_CPU_ACCOUNTING_GEN
442 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
443 Before enabling this option, arch code must be audited
444 to ensure there are no races in concurrent read/write of
445 cputime_t. For example, reading/writing 64-bit cputime_t on
446 some 32-bit arches may require multiple accesses, so proper
447 locking is needed to protect against concurrent accesses.
450 config HAVE_IRQ_TIME_ACCOUNTING
453 Archs need to ensure they use a high enough resolution clock to
454 support irq time accounting and then call enable_sched_clock_irqtime().
456 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
459 config HAVE_ARCH_HUGE_VMAP
462 config HAVE_ARCH_SOFT_DIRTY
465 config HAVE_MOD_ARCH_SPECIFIC
468 The arch uses struct mod_arch_specific to store data. Many arches
469 just need a simple module loader without arch specific data - those
470 should not enable this.
472 config MODULES_USE_ELF_RELA
475 Modules only use ELF RELA relocations. Modules with ELF REL
476 relocations will give an error.
478 config MODULES_USE_ELF_REL
481 Modules only use ELF REL relocations. Modules with ELF RELA
482 relocations will give an error.
484 config HAVE_UNDERSCORE_SYMBOL_PREFIX
487 Some architectures generate an _ in front of C symbols; things like
488 module loading and assembly files need to know about this.
490 config HAVE_IRQ_EXIT_ON_IRQ_STACK
493 Architecture doesn't only execute the irq handler on the irq stack
494 but also irq_exit(). This way we can process softirqs on this irq
495 stack instead of switching to a new one when we call __do_softirq()
496 in the end of an hardirq.
497 This spares a stack switch and improves cache usage on softirq
500 config PGTABLE_LEVELS
504 config ARCH_HAS_ELF_RANDOMIZE
507 An architecture supports choosing randomized locations for
508 stack, mmap, brk, and ET_DYN. Defined functions:
510 - arch_randomize_brk()
512 config HAVE_COPY_THREAD_TLS
515 Architecture provides copy_thread_tls to accept tls argument via
516 normal C parameter passing, rather than extracting the syscall
517 argument from pt_regs.
522 config CLONE_BACKWARDS
525 Architecture has tls passed as the 4th argument of clone(2),
528 config CLONE_BACKWARDS2
531 Architecture has the first two arguments of clone(2) swapped.
533 config CLONE_BACKWARDS3
536 Architecture has tls passed as the 3rd argument of clone(2),
539 config ODD_RT_SIGACTION
542 Architecture has unusual rt_sigaction(2) arguments
544 config OLD_SIGSUSPEND
547 Architecture has old sigsuspend(2) syscall, of one-argument variety
549 config OLD_SIGSUSPEND3
552 Even weirder antique ABI - three-argument sigsuspend(2)
557 Architecture has old sigaction(2) syscall. Nope, not the same
558 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
559 but fairly different variant of sigaction(2), thanks to OSF/1
562 config COMPAT_OLD_SIGACTION
565 source "kernel/gcov/Kconfig"