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 config ARCH_CONFIGURES_CPU_MITIGATIONS
15 if !ARCH_CONFIGURES_CPU_MITIGATIONS
16 config CPU_MITIGATIONS
20 menu "General architecture-dependent options"
45 tristate "OProfile system profiling"
47 depends on HAVE_OPROFILE
49 select RING_BUFFER_ALLOW_SWAP
51 OProfile is a profiling system capable of profiling the
52 whole system, include the kernel, kernel modules, libraries,
57 config OPROFILE_EVENT_MULTIPLEX
58 bool "OProfile multiplexing support (EXPERIMENTAL)"
60 depends on OPROFILE && X86
62 The number of hardware counters is limited. The multiplexing
63 feature enables OProfile to gather more events than counters
64 are provided by the hardware. This is realized by switching
65 between events at a user specified time interval.
72 config OPROFILE_NMI_TIMER
74 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
79 depends on HAVE_KPROBES
82 Kprobes allows you to trap at almost any kernel address and
83 execute a callback function. register_kprobe() establishes
84 a probepoint and specifies the callback. Kprobes is useful
85 for kernel debugging, non-intrusive instrumentation and testing.
89 bool "Optimize very unlikely/likely branches"
90 depends on HAVE_ARCH_JUMP_LABEL
91 depends on CC_HAS_ASM_GOTO
93 This option enables a transparent branch optimization that
94 makes certain almost-always-true or almost-always-false branch
95 conditions even cheaper to execute within the kernel.
97 Certain performance-sensitive kernel code, such as trace points,
98 scheduler functionality, networking code and KVM have such
99 branches and include support for this optimization technique.
101 If it is detected that the compiler has support for "asm goto",
102 the kernel will compile such branches with just a nop
103 instruction. When the condition flag is toggled to true, the
104 nop will be converted to a jump instruction to execute the
105 conditional block of instructions.
107 This technique lowers overhead and stress on the branch prediction
108 of the processor and generally makes the kernel faster. The update
109 of the condition is slower, but those are always very rare.
111 ( On 32-bit x86, the necessary options added to the compiler
112 flags may increase the size of the kernel slightly. )
114 config STATIC_KEYS_SELFTEST
115 bool "Static key selftest"
116 depends on JUMP_LABEL
118 Boot time self-test of the branch patching code.
120 config STATIC_CALL_SELFTEST
121 bool "Static call selftest"
122 depends on HAVE_STATIC_CALL
124 Boot time self-test of the call patching code.
128 depends on KPROBES && HAVE_OPTPROBES
129 select TASKS_RCU if PREEMPTION
131 config KPROBES_ON_FTRACE
133 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
134 depends on DYNAMIC_FTRACE_WITH_REGS
136 If function tracer is enabled and the arch supports full
137 passing of pt_regs to function tracing, then kprobes can
138 optimize on top of function tracing.
142 depends on ARCH_SUPPORTS_UPROBES
144 Uprobes is the user-space counterpart to kprobes: they
145 enable instrumentation applications (such as 'perf probe')
146 to establish unintrusive probes in user-space binaries and
147 libraries, by executing handler functions when the probes
148 are hit by user-space applications.
150 ( These probes come in the form of single-byte breakpoints,
151 managed by the kernel and kept transparent to the probed
154 config HAVE_64BIT_ALIGNED_ACCESS
155 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
157 Some architectures require 64 bit accesses to be 64 bit
158 aligned, which also requires structs containing 64 bit values
159 to be 64 bit aligned too. This includes some 32 bit
160 architectures which can do 64 bit accesses, as well as 64 bit
161 architectures without unaligned access.
163 This symbol should be selected by an architecture if 64 bit
164 accesses are required to be 64 bit aligned in this way even
165 though it is not a 64 bit architecture.
167 See Documentation/unaligned-memory-access.txt for more
168 information on the topic of unaligned memory accesses.
170 config HAVE_EFFICIENT_UNALIGNED_ACCESS
173 Some architectures are unable to perform unaligned accesses
174 without the use of get_unaligned/put_unaligned. Others are
175 unable to perform such accesses efficiently (e.g. trap on
176 unaligned access and require fixing it up in the exception
179 This symbol should be selected by an architecture if it can
180 perform unaligned accesses efficiently to allow different
181 code paths to be selected for these cases. Some network
182 drivers, for example, could opt to not fix up alignment
183 problems with received packets if doing so would not help
186 See Documentation/core-api/unaligned-memory-access.rst for more
187 information on the topic of unaligned memory accesses.
189 config ARCH_USE_BUILTIN_BSWAP
192 Modern versions of GCC (since 4.4) have builtin functions
193 for handling byte-swapping. Using these, instead of the old
194 inline assembler that the architecture code provides in the
195 __arch_bswapXX() macros, allows the compiler to see what's
196 happening and offers more opportunity for optimisation. In
197 particular, the compiler will be able to combine the byteswap
198 with a nearby load or store and use load-and-swap or
199 store-and-swap instructions if the architecture has them. It
200 should almost *never* result in code which is worse than the
201 hand-coded assembler in <asm/swab.h>. But just in case it
202 does, the use of the builtins is optional.
204 Any architecture with load-and-swap or store-and-swap
205 instructions should set this. And it shouldn't hurt to set it
206 on architectures that don't have such instructions.
210 depends on KPROBES && HAVE_KRETPROBES
212 config USER_RETURN_NOTIFIER
214 depends on HAVE_USER_RETURN_NOTIFIER
216 Provide a kernel-internal notification when a cpu is about to
219 config HAVE_IOREMAP_PROT
225 config HAVE_KRETPROBES
228 config HAVE_OPTPROBES
231 config HAVE_KPROBES_ON_FTRACE
234 config HAVE_FUNCTION_ERROR_INJECTION
241 # An arch should select this if it provides all these things:
243 # task_pt_regs() in asm/processor.h or asm/ptrace.h
244 # arch_has_single_step() if there is hardware single-step support
245 # arch_has_block_step() if there is hardware block-step support
246 # asm/syscall.h supplying asm-generic/syscall.h interface
247 # linux/regset.h user_regset interfaces
248 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
249 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
250 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
251 # signal delivery calls tracehook_signal_handler()
253 config HAVE_ARCH_TRACEHOOK
256 config HAVE_DMA_CONTIGUOUS
259 config GENERIC_SMP_IDLE_THREAD
262 config GENERIC_IDLE_POLL_SETUP
265 config ARCH_HAS_FORTIFY_SOURCE
268 An architecture should select this when it can successfully
269 build and run with CONFIG_FORTIFY_SOURCE.
272 # Select if the arch provides a historic keepinit alias for the retain_initrd
273 # command line option
275 config ARCH_HAS_KEEPINITRD
278 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
279 config ARCH_HAS_SET_MEMORY
282 # Select if arch has all set_direct_map_invalid/default() functions
283 config ARCH_HAS_SET_DIRECT_MAP
287 # Select if the architecture provides the arch_dma_set_uncached symbol to
288 # either provide an uncached segement alias for a DMA allocation, or
289 # to remap the page tables in place.
291 config ARCH_HAS_DMA_SET_UNCACHED
295 # Select if the architectures provides the arch_dma_clear_uncached symbol
296 # to undo an in-place page table remap for uncached access.
298 config ARCH_HAS_DMA_CLEAR_UNCACHED
301 config ARCH_HAS_CPU_FINALIZE_INIT
304 # Select if arch init_task must go in the __init_task_data section
305 config ARCH_TASK_STRUCT_ON_STACK
308 # Select if arch has its private alloc_task_struct() function
309 config ARCH_TASK_STRUCT_ALLOCATOR
312 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
314 depends on !ARCH_TASK_STRUCT_ALLOCATOR
316 An architecture should select this to provide hardened usercopy
317 knowledge about what region of the thread_struct should be
318 whitelisted for copying to userspace. Normally this is only the
319 FPU registers. Specifically, arch_thread_struct_whitelist()
320 should be implemented. Without this, the entire thread_struct
321 field in task_struct will be left whitelisted.
323 # Select if arch has its private alloc_thread_stack() function
324 config ARCH_THREAD_STACK_ALLOCATOR
327 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
328 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
331 config ARCH_32BIT_OFF_T
335 All new 32-bit architectures should have 64-bit off_t type on
336 userspace side which corresponds to the loff_t kernel type. This
337 is the requirement for modern ABIs. Some existing architectures
338 still support 32-bit off_t. This option is enabled for all such
339 architectures explicitly.
341 config HAVE_ASM_MODVERSIONS
344 This symbol should be selected by an architecure if it provides
345 <asm/asm-prototypes.h> to support the module versioning for symbols
346 exported from assembly code.
348 config HAVE_REGS_AND_STACK_ACCESS_API
351 This symbol should be selected by an architecure if it supports
352 the API needed to access registers and stack entries from pt_regs,
353 declared in asm/ptrace.h
354 For example the kprobes-based event tracer needs this API.
358 depends on HAVE_REGS_AND_STACK_ACCESS_API
360 This symbol should be selected by an architecture if it
361 supports an implementation of restartable sequences.
363 config HAVE_FUNCTION_ARG_ACCESS_API
366 This symbol should be selected by an architecure if it supports
367 the API needed to access function arguments from pt_regs,
368 declared in asm/ptrace.h
370 config HAVE_HW_BREAKPOINT
372 depends on PERF_EVENTS
374 config HAVE_MIXED_BREAKPOINTS_REGS
376 depends on HAVE_HW_BREAKPOINT
378 Depending on the arch implementation of hardware breakpoints,
379 some of them have separate registers for data and instruction
380 breakpoints addresses, others have mixed registers to store
381 them but define the access type in a control register.
382 Select this option if your arch implements breakpoints under the
385 config HAVE_USER_RETURN_NOTIFIER
388 config HAVE_PERF_EVENTS_NMI
391 System hardware can generate an NMI using the perf event
392 subsystem. Also has support for calculating CPU cycle events
393 to determine how many clock cycles in a given period.
395 config HAVE_HARDLOCKUP_DETECTOR_PERF
397 depends on HAVE_PERF_EVENTS_NMI
399 The arch chooses to use the generic perf-NMI-based hardlockup
400 detector. Must define HAVE_PERF_EVENTS_NMI.
402 config HAVE_NMI_WATCHDOG
406 The arch provides a low level NMI watchdog. It provides
407 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
409 config HAVE_HARDLOCKUP_DETECTOR_ARCH
411 select HAVE_NMI_WATCHDOG
413 The arch chooses to provide its own hardlockup detector, which is
414 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
415 interfaces and parameters provided by hardlockup detector subsystem.
417 config HAVE_PERF_REGS
420 Support selective register dumps for perf events. This includes
421 bit-mapping of each registers and a unique architecture id.
423 config HAVE_PERF_USER_STACK_DUMP
426 Support user stack dumps for perf event samples. This needs
427 access to the user stack pointer which is not unified across
430 config HAVE_ARCH_JUMP_LABEL
433 config HAVE_ARCH_JUMP_LABEL_RELATIVE
436 config MMU_GATHER_TABLE_FREE
439 config MMU_GATHER_RCU_TABLE_FREE
441 select MMU_GATHER_TABLE_FREE
443 config MMU_GATHER_PAGE_SIZE
446 config MMU_GATHER_NO_RANGE
449 config MMU_GATHER_NO_GATHER
451 depends on MMU_GATHER_TABLE_FREE
453 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
456 Temporary select until all architectures can be converted to have
457 irqs disabled over activate_mm. Architectures that do IPI based TLB
458 shootdowns should enable this.
460 config ARCH_HAVE_NMI_SAFE_CMPXCHG
463 config HAVE_ALIGNED_STRUCT_PAGE
466 This makes sure that struct pages are double word aligned and that
467 e.g. the SLUB allocator can perform double word atomic operations
468 on a struct page for better performance. However selecting this
469 might increase the size of a struct page by a word.
471 config HAVE_CMPXCHG_LOCAL
474 config HAVE_CMPXCHG_DOUBLE
477 config ARCH_WEAK_RELEASE_ACQUIRE
480 config ARCH_WANT_IPC_PARSE_VERSION
483 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
486 config ARCH_WANT_OLD_COMPAT_IPC
487 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
490 config HAVE_ARCH_SECCOMP
493 An arch should select this symbol to support seccomp mode 1 (the fixed
494 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
495 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
496 - __NR_seccomp_read_32
497 - __NR_seccomp_write_32
498 - __NR_seccomp_exit_32
499 - __NR_seccomp_sigreturn_32
501 config HAVE_ARCH_SECCOMP_FILTER
503 select HAVE_ARCH_SECCOMP
505 An arch should select this symbol if it provides all of these things:
506 - all the requirements for HAVE_ARCH_SECCOMP
508 - syscall_get_arguments()
510 - syscall_set_return_value()
511 - SIGSYS siginfo_t support
512 - secure_computing is called from a ptrace_event()-safe context
513 - secure_computing return value is checked and a return value of -1
514 results in the system call being skipped immediately.
515 - seccomp syscall wired up
518 prompt "Enable seccomp to safely execute untrusted bytecode"
520 depends on HAVE_ARCH_SECCOMP
522 This kernel feature is useful for number crunching applications
523 that may need to handle untrusted bytecode during their
524 execution. By using pipes or other transports made available
525 to the process as file descriptors supporting the read/write
526 syscalls, it's possible to isolate those applications in their
527 own address space using seccomp. Once seccomp is enabled via
528 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
529 disabled and the task is only allowed to execute a few safe
530 syscalls defined by each seccomp mode.
534 config SECCOMP_FILTER
536 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
538 Enable tasks to build secure computing environments defined
539 in terms of Berkeley Packet Filter programs which implement
540 task-defined system call filtering polices.
542 See Documentation/userspace-api/seccomp_filter.rst for details.
544 config HAVE_ARCH_STACKLEAK
547 An architecture should select this if it has the code which
548 fills the used part of the kernel stack with the STACKLEAK_POISON
549 value before returning from system calls.
551 config HAVE_STACKPROTECTOR
554 An arch should select this symbol if:
555 - it has implemented a stack canary (e.g. __stack_chk_guard)
557 config STACKPROTECTOR
558 bool "Stack Protector buffer overflow detection"
559 depends on HAVE_STACKPROTECTOR
560 depends on $(cc-option,-fstack-protector)
563 This option turns on the "stack-protector" GCC feature. This
564 feature puts, at the beginning of functions, a canary value on
565 the stack just before the return address, and validates
566 the value just before actually returning. Stack based buffer
567 overflows (that need to overwrite this return address) now also
568 overwrite the canary, which gets detected and the attack is then
569 neutralized via a kernel panic.
571 Functions will have the stack-protector canary logic added if they
572 have an 8-byte or larger character array on the stack.
574 This feature requires gcc version 4.2 or above, or a distribution
575 gcc with the feature backported ("-fstack-protector").
577 On an x86 "defconfig" build, this feature adds canary checks to
578 about 3% of all kernel functions, which increases kernel code size
581 config STACKPROTECTOR_STRONG
582 bool "Strong Stack Protector"
583 depends on STACKPROTECTOR
584 depends on $(cc-option,-fstack-protector-strong)
587 Functions will have the stack-protector canary logic added in any
588 of the following conditions:
590 - local variable's address used as part of the right hand side of an
591 assignment or function argument
592 - local variable is an array (or union containing an array),
593 regardless of array type or length
594 - uses register local variables
596 This feature requires gcc version 4.9 or above, or a distribution
597 gcc with the feature backported ("-fstack-protector-strong").
599 On an x86 "defconfig" build, this feature adds canary checks to
600 about 20% of all kernel functions, which increases the kernel code
603 config ARCH_SUPPORTS_SHADOW_CALL_STACK
606 An architecture should select this if it supports Clang's Shadow
607 Call Stack and implements runtime support for shadow stack
610 config SHADOW_CALL_STACK
611 bool "Clang Shadow Call Stack"
612 depends on CC_IS_CLANG && ARCH_SUPPORTS_SHADOW_CALL_STACK
613 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
615 This option enables Clang's Shadow Call Stack, which uses a
616 shadow stack to protect function return addresses from being
617 overwritten by an attacker. More information can be found in
618 Clang's documentation:
620 https://clang.llvm.org/docs/ShadowCallStack.html
622 Note that security guarantees in the kernel differ from the
623 ones documented for user space. The kernel must store addresses
624 of shadow stacks in memory, which means an attacker capable of
625 reading and writing arbitrary memory may be able to locate them
626 and hijack control flow by modifying the stacks.
628 config HAVE_ARCH_WITHIN_STACK_FRAMES
631 An architecture should select this if it can walk the kernel stack
632 frames to determine if an object is part of either the arguments
633 or local variables (i.e. that it excludes saved return addresses,
634 and similar) by implementing an inline arch_within_stack_frames(),
635 which is used by CONFIG_HARDENED_USERCOPY.
637 config HAVE_CONTEXT_TRACKING
640 Provide kernel/user boundaries probes necessary for subsystems
641 that need it, such as userspace RCU extended quiescent state.
642 Syscalls need to be wrapped inside user_exit()-user_enter(), either
643 optimized behind static key or through the slow path using TIF_NOHZ
644 flag. Exceptions handlers must be wrapped as well. Irqs are already
645 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
646 handling on irq exit still need to be protected.
651 Arch relies on TIF_NOHZ and syscall slow path to implement context
652 tracking calls to user_enter()/user_exit().
654 config HAVE_VIRT_CPU_ACCOUNTING
657 config ARCH_HAS_SCALED_CPUTIME
660 config HAVE_VIRT_CPU_ACCOUNTING_GEN
664 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
665 Before enabling this option, arch code must be audited
666 to ensure there are no races in concurrent read/write of
667 cputime_t. For example, reading/writing 64-bit cputime_t on
668 some 32-bit arches may require multiple accesses, so proper
669 locking is needed to protect against concurrent accesses.
672 config HAVE_IRQ_TIME_ACCOUNTING
675 Archs need to ensure they use a high enough resolution clock to
676 support irq time accounting and then call enable_sched_clock_irqtime().
681 Archs that select this are able to move page tables at the PMD level.
683 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
686 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
689 config HAVE_ARCH_HUGE_VMAP
692 config ARCH_WANT_HUGE_PMD_SHARE
695 config HAVE_ARCH_SOFT_DIRTY
698 config HAVE_MOD_ARCH_SPECIFIC
701 The arch uses struct mod_arch_specific to store data. Many arches
702 just need a simple module loader without arch specific data - those
703 should not enable this.
705 config MODULES_USE_ELF_RELA
708 Modules only use ELF RELA relocations. Modules with ELF REL
709 relocations will give an error.
711 config MODULES_USE_ELF_REL
714 Modules only use ELF REL relocations. Modules with ELF RELA
715 relocations will give an error.
717 config HAVE_IRQ_EXIT_ON_IRQ_STACK
720 Architecture doesn't only execute the irq handler on the irq stack
721 but also irq_exit(). This way we can process softirqs on this irq
722 stack instead of switching to a new one when we call __do_softirq()
723 in the end of an hardirq.
724 This spares a stack switch and improves cache usage on softirq
727 config PGTABLE_LEVELS
731 config ARCH_HAS_ELF_RANDOMIZE
734 An architecture supports choosing randomized locations for
735 stack, mmap, brk, and ET_DYN. Defined functions:
737 - arch_randomize_brk()
739 config HAVE_ARCH_MMAP_RND_BITS
742 An arch should select this symbol if it supports setting a variable
743 number of bits for use in establishing the base address for mmap
744 allocations, has MMU enabled and provides values for both:
745 - ARCH_MMAP_RND_BITS_MIN
746 - ARCH_MMAP_RND_BITS_MAX
748 config HAVE_EXIT_THREAD
751 An architecture implements exit_thread.
753 config ARCH_MMAP_RND_BITS_MIN
756 config ARCH_MMAP_RND_BITS_MAX
759 config ARCH_MMAP_RND_BITS_DEFAULT
762 config ARCH_MMAP_RND_BITS
763 int "Number of bits to use for ASLR of mmap base address" if EXPERT
764 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
765 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
766 default ARCH_MMAP_RND_BITS_MIN
767 depends on HAVE_ARCH_MMAP_RND_BITS
769 This value can be used to select the number of bits to use to
770 determine the random offset to the base address of vma regions
771 resulting from mmap allocations. This value will be bounded
772 by the architecture's minimum and maximum supported values.
774 This value can be changed after boot using the
775 /proc/sys/vm/mmap_rnd_bits tunable
777 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
780 An arch should select this symbol if it supports running applications
781 in compatibility mode, supports setting a variable number of bits for
782 use in establishing the base address for mmap allocations, has MMU
783 enabled and provides values for both:
784 - ARCH_MMAP_RND_COMPAT_BITS_MIN
785 - ARCH_MMAP_RND_COMPAT_BITS_MAX
787 config ARCH_MMAP_RND_COMPAT_BITS_MIN
790 config ARCH_MMAP_RND_COMPAT_BITS_MAX
793 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
796 config ARCH_MMAP_RND_COMPAT_BITS
797 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
798 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
799 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
800 default ARCH_MMAP_RND_COMPAT_BITS_MIN
801 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
803 This value can be used to select the number of bits to use to
804 determine the random offset to the base address of vma regions
805 resulting from mmap allocations for compatible applications This
806 value will be bounded by the architecture's minimum and maximum
809 This value can be changed after boot using the
810 /proc/sys/vm/mmap_rnd_compat_bits tunable
812 config HAVE_ARCH_COMPAT_MMAP_BASES
815 This allows 64bit applications to invoke 32-bit mmap() syscall
816 and vice-versa 32-bit applications to call 64-bit mmap().
817 Required for applications doing different bitness syscalls.
819 # This allows to use a set of generic functions to determine mmap base
820 # address by giving priority to top-down scheme only if the process
821 # is not in legacy mode (compat task, unlimited stack size or
822 # sysctl_legacy_va_layout).
823 # Architecture that selects this option can provide its own version of:
825 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
828 select ARCH_HAS_ELF_RANDOMIZE
830 config HAVE_STACK_VALIDATION
833 Architecture supports the 'objtool check' host tool command, which
834 performs compile-time stack metadata validation.
836 config HAVE_RELIABLE_STACKTRACE
839 Architecture has either save_stack_trace_tsk_reliable() or
840 arch_stack_walk_reliable() function which only returns a stack trace
841 if it can guarantee the trace is reliable.
843 config HAVE_ARCH_HASH
847 If this is set, the architecture provides an <asm/hash.h>
848 file which provides platform-specific implementations of some
849 functions in <linux/hash.h> or fs/namei.c.
851 config HAVE_ARCH_NVRAM_OPS
860 config CLONE_BACKWARDS
863 Architecture has tls passed as the 4th argument of clone(2),
866 config CLONE_BACKWARDS2
869 Architecture has the first two arguments of clone(2) swapped.
871 config CLONE_BACKWARDS3
874 Architecture has tls passed as the 3rd argument of clone(2),
877 config ODD_RT_SIGACTION
880 Architecture has unusual rt_sigaction(2) arguments
882 config OLD_SIGSUSPEND
885 Architecture has old sigsuspend(2) syscall, of one-argument variety
887 config OLD_SIGSUSPEND3
890 Even weirder antique ABI - three-argument sigsuspend(2)
895 Architecture has old sigaction(2) syscall. Nope, not the same
896 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
897 but fairly different variant of sigaction(2), thanks to OSF/1
900 config COMPAT_OLD_SIGACTION
903 config COMPAT_32BIT_TIME
904 bool "Provide system calls for 32-bit time_t"
905 default !64BIT || COMPAT
907 This enables 32 bit time_t support in addition to 64 bit time_t support.
908 This is relevant on all 32-bit architectures, and 64-bit architectures
909 as part of compat syscall handling.
911 config ARCH_NO_PREEMPT
914 config ARCH_SUPPORTS_RT
917 config CPU_NO_EFFICIENT_FFS
920 config HAVE_ARCH_VMAP_STACK
923 An arch should select this symbol if it can support kernel stacks
924 in vmalloc space. This means:
926 - vmalloc space must be large enough to hold many kernel stacks.
927 This may rule out many 32-bit architectures.
929 - Stacks in vmalloc space need to work reliably. For example, if
930 vmap page tables are created on demand, either this mechanism
931 needs to work while the stack points to a virtual address with
932 unpopulated page tables or arch code (switch_to() and switch_mm(),
933 most likely) needs to ensure that the stack's page table entries
934 are populated before running on a possibly unpopulated stack.
936 - If the stack overflows into a guard page, something reasonable
937 should happen. The definition of "reasonable" is flexible, but
938 instantly rebooting without logging anything would be unfriendly.
942 bool "Use a virtually-mapped stack"
943 depends on HAVE_ARCH_VMAP_STACK
944 depends on !KASAN || KASAN_VMALLOC
946 Enable this if you want the use virtually-mapped kernel stacks
947 with guard pages. This causes kernel stack overflows to be
948 caught immediately rather than causing difficult-to-diagnose
951 To use this with KASAN, the architecture must support backing
952 virtual mappings with real shadow memory, and KASAN_VMALLOC must
955 config ARCH_OPTIONAL_KERNEL_RWX
958 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
961 config ARCH_HAS_STRICT_KERNEL_RWX
964 config STRICT_KERNEL_RWX
965 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
966 depends on ARCH_HAS_STRICT_KERNEL_RWX
967 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
969 If this is set, kernel text and rodata memory will be made read-only,
970 and non-text memory will be made non-executable. This provides
971 protection against certain security exploits (e.g. executing the heap
974 These features are considered standard security practice these days.
975 You should say Y here in almost all cases.
977 config ARCH_HAS_STRICT_MODULE_RWX
980 config STRICT_MODULE_RWX
981 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
982 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
983 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
985 If this is set, module text and rodata memory will be made read-only,
986 and non-text memory will be made non-executable. This provides
987 protection against certain security exploits (e.g. writing to text)
989 # select if the architecture provides an asm/dma-direct.h header
990 config ARCH_HAS_PHYS_TO_DMA
993 config HAVE_ARCH_COMPILER_H
996 An architecture can select this if it provides an
997 asm/compiler.h header that should be included after
998 linux/compiler-*.h in order to override macro definitions that those
999 headers generally provide.
1001 config HAVE_ARCH_PREL32_RELOCATIONS
1004 May be selected by an architecture if it supports place-relative
1005 32-bit relocations, both in the toolchain and in the module loader,
1006 in which case relative references can be used in special sections
1007 for PCI fixup, initcalls etc which are only half the size on 64 bit
1008 architectures, and don't require runtime relocation on relocatable
1011 config ARCH_USE_MEMREMAP_PROT
1014 config LOCK_EVENT_COUNTS
1015 bool "Locking event counts collection"
1018 Enable light-weight counting of various locking related events
1019 in the system with minimal performance impact. This reduces
1020 the chance of application behavior change because of timing
1021 differences. The counts are reported via debugfs.
1023 # Select if the architecture has support for applying RELR relocations.
1024 config ARCH_HAS_RELR
1028 bool "Use RELR relocation packing"
1029 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1032 Store the kernel's dynamic relocations in the RELR relocation packing
1033 format. Requires a compatible linker (LLD supports this feature), as
1034 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1037 config ARCH_HAS_MEM_ENCRYPT
1040 config ARCH_HAS_CC_PLATFORM
1043 config HAVE_SPARSE_SYSCALL_NR
1046 An architecture should select this if its syscall numbering is sparse
1047 to save space. For example, MIPS architecture has a syscall array with
1048 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1049 related optimizations for a given architecture.
1051 config ARCH_HAS_VDSO_DATA
1054 config HAVE_STATIC_CALL
1057 config HAVE_STATIC_CALL_INLINE
1059 depends on HAVE_STATIC_CALL
1061 config ARCH_WANT_LD_ORPHAN_WARN
1064 An arch should select this symbol once all linker sections are explicitly
1065 included, size-asserted, or discarded in the linker scripts. This is
1066 important because we never want expected sections to be placed heuristically
1067 by the linker, since the locations of such sections can change between linker
1070 config ARCH_SPLIT_ARG64
1073 If a 32-bit architecture requires 64-bit arguments to be split into
1074 pairs of 32-bit arguments, select this option.
1076 source "kernel/gcov/Kconfig"
1078 source "scripts/gcc-plugins/Kconfig"