1 # SPDX-License-Identifier: GPL-2.0
3 # General architecture dependent options
7 # Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can
8 # override the default values in this file.
10 source "arch/$(SRCARCH)/Kconfig"
12 menu "General architecture-dependent options"
28 tristate "OProfile system profiling"
30 depends on HAVE_OPROFILE
32 select RING_BUFFER_ALLOW_SWAP
34 OProfile is a profiling system capable of profiling the
35 whole system, include the kernel, kernel modules, libraries,
40 config OPROFILE_EVENT_MULTIPLEX
41 bool "OProfile multiplexing support (EXPERIMENTAL)"
43 depends on OPROFILE && X86
45 The number of hardware counters is limited. The multiplexing
46 feature enables OProfile to gather more events than counters
47 are provided by the hardware. This is realized by switching
48 between events at a user specified time interval.
55 config OPROFILE_NMI_TIMER
57 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
62 depends on HAVE_KPROBES
65 Kprobes allows you to trap at almost any kernel address and
66 execute a callback function. register_kprobe() establishes
67 a probepoint and specifies the callback. Kprobes is useful
68 for kernel debugging, non-intrusive instrumentation and testing.
72 bool "Optimize very unlikely/likely branches"
73 depends on HAVE_ARCH_JUMP_LABEL
74 depends on CC_HAS_ASM_GOTO
76 This option enables a transparent branch optimization that
77 makes certain almost-always-true or almost-always-false branch
78 conditions even cheaper to execute within the kernel.
80 Certain performance-sensitive kernel code, such as trace points,
81 scheduler functionality, networking code and KVM have such
82 branches and include support for this optimization technique.
84 If it is detected that the compiler has support for "asm goto",
85 the kernel will compile such branches with just a nop
86 instruction. When the condition flag is toggled to true, the
87 nop will be converted to a jump instruction to execute the
88 conditional block of instructions.
90 This technique lowers overhead and stress on the branch prediction
91 of the processor and generally makes the kernel faster. The update
92 of the condition is slower, but those are always very rare.
94 ( On 32-bit x86, the necessary options added to the compiler
95 flags may increase the size of the kernel slightly. )
97 config STATIC_KEYS_SELFTEST
98 bool "Static key selftest"
101 Boot time self-test of the branch patching code.
105 depends on KPROBES && HAVE_OPTPROBES
106 select TASKS_RCU if PREEMPT
108 config KPROBES_ON_FTRACE
110 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
111 depends on DYNAMIC_FTRACE_WITH_REGS
113 If function tracer is enabled and the arch supports full
114 passing of pt_regs to function tracing, then kprobes can
115 optimize on top of function tracing.
119 depends on ARCH_SUPPORTS_UPROBES
121 Uprobes is the user-space counterpart to kprobes: they
122 enable instrumentation applications (such as 'perf probe')
123 to establish unintrusive probes in user-space binaries and
124 libraries, by executing handler functions when the probes
125 are hit by user-space applications.
127 ( These probes come in the form of single-byte breakpoints,
128 managed by the kernel and kept transparent to the probed
131 config HAVE_64BIT_ALIGNED_ACCESS
132 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
134 Some architectures require 64 bit accesses to be 64 bit
135 aligned, which also requires structs containing 64 bit values
136 to be 64 bit aligned too. This includes some 32 bit
137 architectures which can do 64 bit accesses, as well as 64 bit
138 architectures without unaligned access.
140 This symbol should be selected by an architecture if 64 bit
141 accesses are required to be 64 bit aligned in this way even
142 though it is not a 64 bit architecture.
144 See Documentation/unaligned-memory-access.txt for more
145 information on the topic of unaligned memory accesses.
147 config HAVE_EFFICIENT_UNALIGNED_ACCESS
150 Some architectures are unable to perform unaligned accesses
151 without the use of get_unaligned/put_unaligned. Others are
152 unable to perform such accesses efficiently (e.g. trap on
153 unaligned access and require fixing it up in the exception
156 This symbol should be selected by an architecture if it can
157 perform unaligned accesses efficiently to allow different
158 code paths to be selected for these cases. Some network
159 drivers, for example, could opt to not fix up alignment
160 problems with received packets if doing so would not help
163 See Documentation/unaligned-memory-access.txt for more
164 information on the topic of unaligned memory accesses.
166 config ARCH_USE_BUILTIN_BSWAP
169 Modern versions of GCC (since 4.4) have builtin functions
170 for handling byte-swapping. Using these, instead of the old
171 inline assembler that the architecture code provides in the
172 __arch_bswapXX() macros, allows the compiler to see what's
173 happening and offers more opportunity for optimisation. In
174 particular, the compiler will be able to combine the byteswap
175 with a nearby load or store and use load-and-swap or
176 store-and-swap instructions if the architecture has them. It
177 should almost *never* result in code which is worse than the
178 hand-coded assembler in <asm/swab.h>. But just in case it
179 does, the use of the builtins is optional.
181 Any architecture with load-and-swap or store-and-swap
182 instructions should set this. And it shouldn't hurt to set it
183 on architectures that don't have such instructions.
187 depends on KPROBES && HAVE_KRETPROBES
189 config USER_RETURN_NOTIFIER
191 depends on HAVE_USER_RETURN_NOTIFIER
193 Provide a kernel-internal notification when a cpu is about to
196 config HAVE_IOREMAP_PROT
202 config HAVE_KRETPROBES
205 config HAVE_OPTPROBES
208 config HAVE_KPROBES_ON_FTRACE
211 config HAVE_FUNCTION_ERROR_INJECTION
218 # An arch should select this if it provides all these things:
220 # task_pt_regs() in asm/processor.h or asm/ptrace.h
221 # arch_has_single_step() if there is hardware single-step support
222 # arch_has_block_step() if there is hardware block-step support
223 # asm/syscall.h supplying asm-generic/syscall.h interface
224 # linux/regset.h user_regset interfaces
225 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
226 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
227 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
228 # signal delivery calls tracehook_signal_handler()
230 config HAVE_ARCH_TRACEHOOK
233 config HAVE_DMA_CONTIGUOUS
236 config GENERIC_SMP_IDLE_THREAD
239 config GENERIC_IDLE_POLL_SETUP
242 config ARCH_HAS_FORTIFY_SOURCE
245 An architecture should select this when it can successfully
246 build and run with CONFIG_FORTIFY_SOURCE.
248 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
249 config ARCH_HAS_SET_MEMORY
252 # Select if arch init_task must go in the __init_task_data section
253 config ARCH_TASK_STRUCT_ON_STACK
256 # Select if arch has its private alloc_task_struct() function
257 config ARCH_TASK_STRUCT_ALLOCATOR
260 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
262 depends on !ARCH_TASK_STRUCT_ALLOCATOR
264 An architecture should select this to provide hardened usercopy
265 knowledge about what region of the thread_struct should be
266 whitelisted for copying to userspace. Normally this is only the
267 FPU registers. Specifically, arch_thread_struct_whitelist()
268 should be implemented. Without this, the entire thread_struct
269 field in task_struct will be left whitelisted.
271 # Select if arch has its private alloc_thread_stack() function
272 config ARCH_THREAD_STACK_ALLOCATOR
275 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
276 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
279 config HAVE_REGS_AND_STACK_ACCESS_API
282 This symbol should be selected by an architecure if it supports
283 the API needed to access registers and stack entries from pt_regs,
284 declared in asm/ptrace.h
285 For example the kprobes-based event tracer needs this API.
289 depends on HAVE_REGS_AND_STACK_ACCESS_API
291 This symbol should be selected by an architecture if it
292 supports an implementation of restartable sequences.
297 The <linux/clk.h> calls support software clock gating and
298 thus are a key power management tool on many systems.
300 config HAVE_HW_BREAKPOINT
302 depends on PERF_EVENTS
304 config HAVE_MIXED_BREAKPOINTS_REGS
306 depends on HAVE_HW_BREAKPOINT
308 Depending on the arch implementation of hardware breakpoints,
309 some of them have separate registers for data and instruction
310 breakpoints addresses, others have mixed registers to store
311 them but define the access type in a control register.
312 Select this option if your arch implements breakpoints under the
315 config HAVE_USER_RETURN_NOTIFIER
318 config HAVE_PERF_EVENTS_NMI
321 System hardware can generate an NMI using the perf event
322 subsystem. Also has support for calculating CPU cycle events
323 to determine how many clock cycles in a given period.
325 config HAVE_HARDLOCKUP_DETECTOR_PERF
327 depends on HAVE_PERF_EVENTS_NMI
329 The arch chooses to use the generic perf-NMI-based hardlockup
330 detector. Must define HAVE_PERF_EVENTS_NMI.
332 config HAVE_NMI_WATCHDOG
336 The arch provides a low level NMI watchdog. It provides
337 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
339 config HAVE_HARDLOCKUP_DETECTOR_ARCH
341 select HAVE_NMI_WATCHDOG
343 The arch chooses to provide its own hardlockup detector, which is
344 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
345 interfaces and parameters provided by hardlockup detector subsystem.
347 config HAVE_PERF_REGS
350 Support selective register dumps for perf events. This includes
351 bit-mapping of each registers and a unique architecture id.
353 config HAVE_PERF_USER_STACK_DUMP
356 Support user stack dumps for perf event samples. This needs
357 access to the user stack pointer which is not unified across
360 config HAVE_ARCH_JUMP_LABEL
363 config HAVE_RCU_TABLE_FREE
366 config HAVE_RCU_TABLE_INVALIDATE
369 config ARCH_HAVE_NMI_SAFE_CMPXCHG
372 config HAVE_ALIGNED_STRUCT_PAGE
375 This makes sure that struct pages are double word aligned and that
376 e.g. the SLUB allocator can perform double word atomic operations
377 on a struct page for better performance. However selecting this
378 might increase the size of a struct page by a word.
380 config HAVE_CMPXCHG_LOCAL
383 config HAVE_CMPXCHG_DOUBLE
386 config ARCH_WEAK_RELEASE_ACQUIRE
389 config ARCH_WANT_IPC_PARSE_VERSION
392 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
395 config ARCH_WANT_OLD_COMPAT_IPC
396 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
399 config HAVE_ARCH_SECCOMP_FILTER
402 An arch should select this symbol if it provides all of these things:
404 - syscall_get_arguments()
406 - syscall_set_return_value()
407 - SIGSYS siginfo_t support
408 - secure_computing is called from a ptrace_event()-safe context
409 - secure_computing return value is checked and a return value of -1
410 results in the system call being skipped immediately.
411 - seccomp syscall wired up
413 config SECCOMP_FILTER
415 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
417 Enable tasks to build secure computing environments defined
418 in terms of Berkeley Packet Filter programs which implement
419 task-defined system call filtering polices.
421 See Documentation/userspace-api/seccomp_filter.rst for details.
423 config HAVE_STACKPROTECTOR
426 An arch should select this symbol if:
427 - it has implemented a stack canary (e.g. __stack_chk_guard)
429 config CC_HAS_STACKPROTECTOR_NONE
430 def_bool $(cc-option,-fno-stack-protector)
432 config STACKPROTECTOR
433 bool "Stack Protector buffer overflow detection"
434 depends on HAVE_STACKPROTECTOR
435 depends on $(cc-option,-fstack-protector)
438 This option turns on the "stack-protector" GCC feature. This
439 feature puts, at the beginning of functions, a canary value on
440 the stack just before the return address, and validates
441 the value just before actually returning. Stack based buffer
442 overflows (that need to overwrite this return address) now also
443 overwrite the canary, which gets detected and the attack is then
444 neutralized via a kernel panic.
446 Functions will have the stack-protector canary logic added if they
447 have an 8-byte or larger character array on the stack.
449 This feature requires gcc version 4.2 or above, or a distribution
450 gcc with the feature backported ("-fstack-protector").
452 On an x86 "defconfig" build, this feature adds canary checks to
453 about 3% of all kernel functions, which increases kernel code size
456 config STACKPROTECTOR_STRONG
457 bool "Strong Stack Protector"
458 depends on STACKPROTECTOR
459 depends on $(cc-option,-fstack-protector-strong)
462 Functions will have the stack-protector canary logic added in any
463 of the following conditions:
465 - local variable's address used as part of the right hand side of an
466 assignment or function argument
467 - local variable is an array (or union containing an array),
468 regardless of array type or length
469 - uses register local variables
471 This feature requires gcc version 4.9 or above, or a distribution
472 gcc with the feature backported ("-fstack-protector-strong").
474 On an x86 "defconfig" build, this feature adds canary checks to
475 about 20% of all kernel functions, which increases the kernel code
478 config HAVE_ARCH_WITHIN_STACK_FRAMES
481 An architecture should select this if it can walk the kernel stack
482 frames to determine if an object is part of either the arguments
483 or local variables (i.e. that it excludes saved return addresses,
484 and similar) by implementing an inline arch_within_stack_frames(),
485 which is used by CONFIG_HARDENED_USERCOPY.
487 config HAVE_CONTEXT_TRACKING
490 Provide kernel/user boundaries probes necessary for subsystems
491 that need it, such as userspace RCU extended quiescent state.
492 Syscalls need to be wrapped inside user_exit()-user_enter() through
493 the slow path using TIF_NOHZ flag. Exceptions handlers must be
494 wrapped as well. Irqs are already protected inside
495 rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on
496 irq exit still need to be protected.
498 config HAVE_VIRT_CPU_ACCOUNTING
501 config ARCH_HAS_SCALED_CPUTIME
504 config HAVE_VIRT_CPU_ACCOUNTING_GEN
508 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
509 Before enabling this option, arch code must be audited
510 to ensure there are no races in concurrent read/write of
511 cputime_t. For example, reading/writing 64-bit cputime_t on
512 some 32-bit arches may require multiple accesses, so proper
513 locking is needed to protect against concurrent accesses.
516 config HAVE_IRQ_TIME_ACCOUNTING
519 Archs need to ensure they use a high enough resolution clock to
520 support irq time accounting and then call enable_sched_clock_irqtime().
522 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
525 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
528 config HAVE_ARCH_HUGE_VMAP
531 config HAVE_ARCH_SOFT_DIRTY
534 config HAVE_MOD_ARCH_SPECIFIC
537 The arch uses struct mod_arch_specific to store data. Many arches
538 just need a simple module loader without arch specific data - those
539 should not enable this.
541 config MODULES_USE_ELF_RELA
544 Modules only use ELF RELA relocations. Modules with ELF REL
545 relocations will give an error.
547 config MODULES_USE_ELF_REL
550 Modules only use ELF REL relocations. Modules with ELF RELA
551 relocations will give an error.
553 config HAVE_IRQ_EXIT_ON_IRQ_STACK
556 Architecture doesn't only execute the irq handler on the irq stack
557 but also irq_exit(). This way we can process softirqs on this irq
558 stack instead of switching to a new one when we call __do_softirq()
559 in the end of an hardirq.
560 This spares a stack switch and improves cache usage on softirq
563 config PGTABLE_LEVELS
567 config ARCH_HAS_ELF_RANDOMIZE
570 An architecture supports choosing randomized locations for
571 stack, mmap, brk, and ET_DYN. Defined functions:
573 - arch_randomize_brk()
575 config HAVE_ARCH_MMAP_RND_BITS
578 An arch should select this symbol if it supports setting a variable
579 number of bits for use in establishing the base address for mmap
580 allocations, has MMU enabled and provides values for both:
581 - ARCH_MMAP_RND_BITS_MIN
582 - ARCH_MMAP_RND_BITS_MAX
584 config HAVE_EXIT_THREAD
587 An architecture implements exit_thread.
589 config ARCH_MMAP_RND_BITS_MIN
592 config ARCH_MMAP_RND_BITS_MAX
595 config ARCH_MMAP_RND_BITS_DEFAULT
598 config ARCH_MMAP_RND_BITS
599 int "Number of bits to use for ASLR of mmap base address" if EXPERT
600 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
601 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
602 default ARCH_MMAP_RND_BITS_MIN
603 depends on HAVE_ARCH_MMAP_RND_BITS
605 This value can be used to select the number of bits to use to
606 determine the random offset to the base address of vma regions
607 resulting from mmap allocations. This value will be bounded
608 by the architecture's minimum and maximum supported values.
610 This value can be changed after boot using the
611 /proc/sys/vm/mmap_rnd_bits tunable
613 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
616 An arch should select this symbol if it supports running applications
617 in compatibility mode, supports setting a variable number of bits for
618 use in establishing the base address for mmap allocations, has MMU
619 enabled and provides values for both:
620 - ARCH_MMAP_RND_COMPAT_BITS_MIN
621 - ARCH_MMAP_RND_COMPAT_BITS_MAX
623 config ARCH_MMAP_RND_COMPAT_BITS_MIN
626 config ARCH_MMAP_RND_COMPAT_BITS_MAX
629 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
632 config ARCH_MMAP_RND_COMPAT_BITS
633 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
634 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
635 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
636 default ARCH_MMAP_RND_COMPAT_BITS_MIN
637 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
639 This value can be used to select the number of bits to use to
640 determine the random offset to the base address of vma regions
641 resulting from mmap allocations for compatible applications This
642 value will be bounded by the architecture's minimum and maximum
645 This value can be changed after boot using the
646 /proc/sys/vm/mmap_rnd_compat_bits tunable
648 config HAVE_ARCH_COMPAT_MMAP_BASES
651 This allows 64bit applications to invoke 32-bit mmap() syscall
652 and vice-versa 32-bit applications to call 64-bit mmap().
653 Required for applications doing different bitness syscalls.
655 config HAVE_COPY_THREAD_TLS
658 Architecture provides copy_thread_tls to accept tls argument via
659 normal C parameter passing, rather than extracting the syscall
660 argument from pt_regs.
662 config HAVE_STACK_VALIDATION
665 Architecture supports the 'objtool check' host tool command, which
666 performs compile-time stack metadata validation.
668 config HAVE_RELIABLE_STACKTRACE
671 Architecture has a save_stack_trace_tsk_reliable() function which
672 only returns a stack trace if it can guarantee the trace is reliable.
674 config HAVE_ARCH_HASH
678 If this is set, the architecture provides an <asm/hash.h>
679 file which provides platform-specific implementations of some
680 functions in <linux/hash.h> or fs/namei.c.
688 config CLONE_BACKWARDS
691 Architecture has tls passed as the 4th argument of clone(2),
694 config CLONE_BACKWARDS2
697 Architecture has the first two arguments of clone(2) swapped.
699 config CLONE_BACKWARDS3
702 Architecture has tls passed as the 3rd argument of clone(2),
705 config ODD_RT_SIGACTION
708 Architecture has unusual rt_sigaction(2) arguments
710 config OLD_SIGSUSPEND
713 Architecture has old sigsuspend(2) syscall, of one-argument variety
715 config OLD_SIGSUSPEND3
718 Even weirder antique ABI - three-argument sigsuspend(2)
723 Architecture has old sigaction(2) syscall. Nope, not the same
724 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
725 but fairly different variant of sigaction(2), thanks to OSF/1
728 config COMPAT_OLD_SIGACTION
732 def_bool ARCH_HAS_64BIT_TIME
734 This should be selected by all architectures that need to support
735 new system calls with a 64-bit time_t. This is relevant on all 32-bit
736 architectures, and 64-bit architectures as part of compat syscall
739 config COMPAT_32BIT_TIME
740 def_bool (!64BIT && 64BIT_TIME) || COMPAT
742 This enables 32 bit time_t support in addition to 64 bit time_t support.
743 This is relevant on all 32-bit architectures, and 64-bit architectures
744 as part of compat syscall handling.
746 config ARCH_NO_COHERENT_DMA_MMAP
749 config ARCH_NO_PREEMPT
752 config CPU_NO_EFFICIENT_FFS
755 config HAVE_ARCH_VMAP_STACK
758 An arch should select this symbol if it can support kernel stacks
759 in vmalloc space. This means:
761 - vmalloc space must be large enough to hold many kernel stacks.
762 This may rule out many 32-bit architectures.
764 - Stacks in vmalloc space need to work reliably. For example, if
765 vmap page tables are created on demand, either this mechanism
766 needs to work while the stack points to a virtual address with
767 unpopulated page tables or arch code (switch_to() and switch_mm(),
768 most likely) needs to ensure that the stack's page table entries
769 are populated before running on a possibly unpopulated stack.
771 - If the stack overflows into a guard page, something reasonable
772 should happen. The definition of "reasonable" is flexible, but
773 instantly rebooting without logging anything would be unfriendly.
777 bool "Use a virtually-mapped stack"
778 depends on HAVE_ARCH_VMAP_STACK && !KASAN
780 Enable this if you want the use virtually-mapped kernel stacks
781 with guard pages. This causes kernel stack overflows to be
782 caught immediately rather than causing difficult-to-diagnose
785 This is presently incompatible with KASAN because KASAN expects
786 the stack to map directly to the KASAN shadow map using a formula
787 that is incorrect if the stack is in vmalloc space.
789 config ARCH_OPTIONAL_KERNEL_RWX
792 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
795 config ARCH_HAS_STRICT_KERNEL_RWX
798 config STRICT_KERNEL_RWX
799 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
800 depends on ARCH_HAS_STRICT_KERNEL_RWX
801 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
803 If this is set, kernel text and rodata memory will be made read-only,
804 and non-text memory will be made non-executable. This provides
805 protection against certain security exploits (e.g. executing the heap
808 These features are considered standard security practice these days.
809 You should say Y here in almost all cases.
811 config ARCH_HAS_STRICT_MODULE_RWX
814 config STRICT_MODULE_RWX
815 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
816 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
817 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
819 If this is set, module text and rodata memory will be made read-only,
820 and non-text memory will be made non-executable. This provides
821 protection against certain security exploits (e.g. writing to text)
823 # select if the architecture provides an asm/dma-direct.h header
824 config ARCH_HAS_PHYS_TO_DMA
827 config ARCH_HAS_REFCOUNT
830 An architecture selects this when it has implemented refcount_t
831 using open coded assembly primitives that provide an optimized
832 refcount_t implementation, possibly at the expense of some full
833 refcount state checks of CONFIG_REFCOUNT_FULL=y.
835 The refcount overflow check behavior, however, must be retained.
836 Catching overflows is the primary security concern for protecting
837 against bugs in reference counts.
840 bool "Perform full reference count validation at the expense of speed"
842 Enabling this switches the refcounting infrastructure from a fast
843 unchecked atomic_t implementation to a fully state checked
844 implementation, which can be (slightly) slower but provides protections
845 against various use-after-free conditions that can be used in
846 security flaw exploits.
848 config HAVE_ARCH_COMPILER_H
851 An architecture can select this if it provides an
852 asm/compiler.h header that should be included after
853 linux/compiler-*.h in order to override macro definitions that those
854 headers generally provide.
856 config HAVE_ARCH_PREL32_RELOCATIONS
859 May be selected by an architecture if it supports place-relative
860 32-bit relocations, both in the toolchain and in the module loader,
861 in which case relative references can be used in special sections
862 for PCI fixup, initcalls etc which are only half the size on 64 bit
863 architectures, and don't require runtime relocation on relocatable
866 source "kernel/gcov/Kconfig"
868 source "scripts/gcc-plugins/Kconfig"