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"
36 depends on HAVE_KPROBES
39 Kprobes allows you to trap at almost any kernel address and
40 execute a callback function. register_kprobe() establishes
41 a probepoint and specifies the callback. Kprobes is useful
42 for kernel debugging, non-intrusive instrumentation and testing.
46 bool "Optimize very unlikely/likely branches"
47 depends on HAVE_ARCH_JUMP_LABEL
48 depends on CC_HAS_ASM_GOTO
50 This option enables a transparent branch optimization that
51 makes certain almost-always-true or almost-always-false branch
52 conditions even cheaper to execute within the kernel.
54 Certain performance-sensitive kernel code, such as trace points,
55 scheduler functionality, networking code and KVM have such
56 branches and include support for this optimization technique.
58 If it is detected that the compiler has support for "asm goto",
59 the kernel will compile such branches with just a nop
60 instruction. When the condition flag is toggled to true, the
61 nop will be converted to a jump instruction to execute the
62 conditional block of instructions.
64 This technique lowers overhead and stress on the branch prediction
65 of the processor and generally makes the kernel faster. The update
66 of the condition is slower, but those are always very rare.
68 ( On 32-bit x86, the necessary options added to the compiler
69 flags may increase the size of the kernel slightly. )
71 config STATIC_KEYS_SELFTEST
72 bool "Static key selftest"
75 Boot time self-test of the branch patching code.
77 config STATIC_CALL_SELFTEST
78 bool "Static call selftest"
79 depends on HAVE_STATIC_CALL
81 Boot time self-test of the call patching code.
85 depends on KPROBES && HAVE_OPTPROBES
86 select TASKS_RCU if PREEMPTION
88 config KPROBES_ON_FTRACE
90 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
91 depends on DYNAMIC_FTRACE_WITH_REGS
93 If function tracer is enabled and the arch supports full
94 passing of pt_regs to function tracing, then kprobes can
95 optimize on top of function tracing.
99 depends on ARCH_SUPPORTS_UPROBES
101 Uprobes is the user-space counterpart to kprobes: they
102 enable instrumentation applications (such as 'perf probe')
103 to establish unintrusive probes in user-space binaries and
104 libraries, by executing handler functions when the probes
105 are hit by user-space applications.
107 ( These probes come in the form of single-byte breakpoints,
108 managed by the kernel and kept transparent to the probed
111 config HAVE_64BIT_ALIGNED_ACCESS
112 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
114 Some architectures require 64 bit accesses to be 64 bit
115 aligned, which also requires structs containing 64 bit values
116 to be 64 bit aligned too. This includes some 32 bit
117 architectures which can do 64 bit accesses, as well as 64 bit
118 architectures without unaligned access.
120 This symbol should be selected by an architecture if 64 bit
121 accesses are required to be 64 bit aligned in this way even
122 though it is not a 64 bit architecture.
124 See Documentation/core-api/unaligned-memory-access.rst for
125 more information on the topic of unaligned memory accesses.
127 config HAVE_EFFICIENT_UNALIGNED_ACCESS
130 Some architectures are unable to perform unaligned accesses
131 without the use of get_unaligned/put_unaligned. Others are
132 unable to perform such accesses efficiently (e.g. trap on
133 unaligned access and require fixing it up in the exception
136 This symbol should be selected by an architecture if it can
137 perform unaligned accesses efficiently to allow different
138 code paths to be selected for these cases. Some network
139 drivers, for example, could opt to not fix up alignment
140 problems with received packets if doing so would not help
143 See Documentation/core-api/unaligned-memory-access.rst for more
144 information on the topic of unaligned memory accesses.
146 config ARCH_USE_BUILTIN_BSWAP
149 Modern versions of GCC (since 4.4) have builtin functions
150 for handling byte-swapping. Using these, instead of the old
151 inline assembler that the architecture code provides in the
152 __arch_bswapXX() macros, allows the compiler to see what's
153 happening and offers more opportunity for optimisation. In
154 particular, the compiler will be able to combine the byteswap
155 with a nearby load or store and use load-and-swap or
156 store-and-swap instructions if the architecture has them. It
157 should almost *never* result in code which is worse than the
158 hand-coded assembler in <asm/swab.h>. But just in case it
159 does, the use of the builtins is optional.
161 Any architecture with load-and-swap or store-and-swap
162 instructions should set this. And it shouldn't hurt to set it
163 on architectures that don't have such instructions.
167 depends on KPROBES && HAVE_KRETPROBES
169 config USER_RETURN_NOTIFIER
171 depends on HAVE_USER_RETURN_NOTIFIER
173 Provide a kernel-internal notification when a cpu is about to
176 config HAVE_IOREMAP_PROT
182 config HAVE_KRETPROBES
185 config HAVE_OPTPROBES
188 config HAVE_KPROBES_ON_FTRACE
191 config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE
194 Since kretprobes modifies return address on the stack, the
195 stacktrace may see the kretprobe trampoline address instead
196 of correct one. If the architecture stacktrace code and
197 unwinder can adjust such entries, select this configuration.
199 config HAVE_FUNCTION_ERROR_INJECTION
205 config HAVE_FUNCTION_DESCRIPTORS
208 config TRACE_IRQFLAGS_SUPPORT
212 # An arch should select this if it provides all these things:
214 # task_pt_regs() in asm/processor.h or asm/ptrace.h
215 # arch_has_single_step() if there is hardware single-step support
216 # arch_has_block_step() if there is hardware block-step support
217 # asm/syscall.h supplying asm-generic/syscall.h interface
218 # linux/regset.h user_regset interfaces
219 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
220 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
221 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
222 # signal delivery calls tracehook_signal_handler()
224 config HAVE_ARCH_TRACEHOOK
227 config HAVE_DMA_CONTIGUOUS
230 config GENERIC_SMP_IDLE_THREAD
233 config GENERIC_IDLE_POLL_SETUP
236 config ARCH_HAS_FORTIFY_SOURCE
239 An architecture should select this when it can successfully
240 build and run with CONFIG_FORTIFY_SOURCE.
243 # Select if the arch provides a historic keepinit alias for the retain_initrd
244 # command line option
246 config ARCH_HAS_KEEPINITRD
249 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
250 config ARCH_HAS_SET_MEMORY
253 # Select if arch has all set_direct_map_invalid/default() functions
254 config ARCH_HAS_SET_DIRECT_MAP
258 # Select if the architecture provides the arch_dma_set_uncached symbol to
259 # either provide an uncached segment alias for a DMA allocation, or
260 # to remap the page tables in place.
262 config ARCH_HAS_DMA_SET_UNCACHED
266 # Select if the architectures provides the arch_dma_clear_uncached symbol
267 # to undo an in-place page table remap for uncached access.
269 config ARCH_HAS_DMA_CLEAR_UNCACHED
272 # Select if arch init_task must go in the __init_task_data section
273 config ARCH_TASK_STRUCT_ON_STACK
276 # Select if arch has its private alloc_task_struct() function
277 config ARCH_TASK_STRUCT_ALLOCATOR
280 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
282 depends on !ARCH_TASK_STRUCT_ALLOCATOR
284 An architecture should select this to provide hardened usercopy
285 knowledge about what region of the thread_struct should be
286 whitelisted for copying to userspace. Normally this is only the
287 FPU registers. Specifically, arch_thread_struct_whitelist()
288 should be implemented. Without this, the entire thread_struct
289 field in task_struct will be left whitelisted.
291 # Select if arch has its private alloc_thread_stack() function
292 config ARCH_THREAD_STACK_ALLOCATOR
295 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
296 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
299 config ARCH_WANTS_NO_INSTR
302 An architecture should select this if the noinstr macro is being used on
303 functions to denote that the toolchain should avoid instrumenting such
304 functions and is required for correctness.
306 config ARCH_32BIT_OFF_T
310 All new 32-bit architectures should have 64-bit off_t type on
311 userspace side which corresponds to the loff_t kernel type. This
312 is the requirement for modern ABIs. Some existing architectures
313 still support 32-bit off_t. This option is enabled for all such
314 architectures explicitly.
316 # Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat
317 config ARCH_32BIT_USTAT_F_TINODE
320 config HAVE_ASM_MODVERSIONS
323 This symbol should be selected by an architecture if it provides
324 <asm/asm-prototypes.h> to support the module versioning for symbols
325 exported from assembly code.
327 config HAVE_REGS_AND_STACK_ACCESS_API
330 This symbol should be selected by an architecture if it supports
331 the API needed to access registers and stack entries from pt_regs,
332 declared in asm/ptrace.h
333 For example the kprobes-based event tracer needs this API.
337 depends on HAVE_REGS_AND_STACK_ACCESS_API
339 This symbol should be selected by an architecture if it
340 supports an implementation of restartable sequences.
342 config HAVE_FUNCTION_ARG_ACCESS_API
345 This symbol should be selected by an architecture if it supports
346 the API needed to access function arguments from pt_regs,
347 declared in asm/ptrace.h
349 config HAVE_HW_BREAKPOINT
351 depends on PERF_EVENTS
353 config HAVE_MIXED_BREAKPOINTS_REGS
355 depends on HAVE_HW_BREAKPOINT
357 Depending on the arch implementation of hardware breakpoints,
358 some of them have separate registers for data and instruction
359 breakpoints addresses, others have mixed registers to store
360 them but define the access type in a control register.
361 Select this option if your arch implements breakpoints under the
364 config HAVE_USER_RETURN_NOTIFIER
367 config HAVE_PERF_EVENTS_NMI
370 System hardware can generate an NMI using the perf event
371 subsystem. Also has support for calculating CPU cycle events
372 to determine how many clock cycles in a given period.
374 config HAVE_HARDLOCKUP_DETECTOR_PERF
376 depends on HAVE_PERF_EVENTS_NMI
378 The arch chooses to use the generic perf-NMI-based hardlockup
379 detector. Must define HAVE_PERF_EVENTS_NMI.
381 config HAVE_NMI_WATCHDOG
385 The arch provides a low level NMI watchdog. It provides
386 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
388 config HAVE_HARDLOCKUP_DETECTOR_ARCH
390 select HAVE_NMI_WATCHDOG
392 The arch chooses to provide its own hardlockup detector, which is
393 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
394 interfaces and parameters provided by hardlockup detector subsystem.
396 config HAVE_PERF_REGS
399 Support selective register dumps for perf events. This includes
400 bit-mapping of each registers and a unique architecture id.
402 config HAVE_PERF_USER_STACK_DUMP
405 Support user stack dumps for perf event samples. This needs
406 access to the user stack pointer which is not unified across
409 config HAVE_ARCH_JUMP_LABEL
412 config HAVE_ARCH_JUMP_LABEL_RELATIVE
415 config MMU_GATHER_TABLE_FREE
418 config MMU_GATHER_RCU_TABLE_FREE
420 select MMU_GATHER_TABLE_FREE
422 config MMU_GATHER_PAGE_SIZE
425 config MMU_GATHER_NO_RANGE
428 config MMU_GATHER_NO_GATHER
430 depends on MMU_GATHER_TABLE_FREE
432 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
435 Temporary select until all architectures can be converted to have
436 irqs disabled over activate_mm. Architectures that do IPI based TLB
437 shootdowns should enable this.
439 config ARCH_HAVE_NMI_SAFE_CMPXCHG
442 config HAVE_ALIGNED_STRUCT_PAGE
445 This makes sure that struct pages are double word aligned and that
446 e.g. the SLUB allocator can perform double word atomic operations
447 on a struct page for better performance. However selecting this
448 might increase the size of a struct page by a word.
450 config HAVE_CMPXCHG_LOCAL
453 config HAVE_CMPXCHG_DOUBLE
456 config ARCH_WEAK_RELEASE_ACQUIRE
459 config ARCH_WANT_IPC_PARSE_VERSION
462 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
465 config ARCH_WANT_OLD_COMPAT_IPC
466 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
469 config HAVE_ARCH_SECCOMP
472 An arch should select this symbol to support seccomp mode 1 (the fixed
473 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
474 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
475 - __NR_seccomp_read_32
476 - __NR_seccomp_write_32
477 - __NR_seccomp_exit_32
478 - __NR_seccomp_sigreturn_32
480 config HAVE_ARCH_SECCOMP_FILTER
482 select HAVE_ARCH_SECCOMP
484 An arch should select this symbol if it provides all of these things:
485 - all the requirements for HAVE_ARCH_SECCOMP
487 - syscall_get_arguments()
489 - syscall_set_return_value()
490 - SIGSYS siginfo_t support
491 - secure_computing is called from a ptrace_event()-safe context
492 - secure_computing return value is checked and a return value of -1
493 results in the system call being skipped immediately.
494 - seccomp syscall wired up
495 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
496 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
497 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
500 prompt "Enable seccomp to safely execute untrusted bytecode"
502 depends on HAVE_ARCH_SECCOMP
504 This kernel feature is useful for number crunching applications
505 that may need to handle untrusted bytecode during their
506 execution. By using pipes or other transports made available
507 to the process as file descriptors supporting the read/write
508 syscalls, it's possible to isolate those applications in their
509 own address space using seccomp. Once seccomp is enabled via
510 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
511 disabled and the task is only allowed to execute a few safe
512 syscalls defined by each seccomp mode.
516 config SECCOMP_FILTER
518 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
520 Enable tasks to build secure computing environments defined
521 in terms of Berkeley Packet Filter programs which implement
522 task-defined system call filtering polices.
524 See Documentation/userspace-api/seccomp_filter.rst for details.
526 config SECCOMP_CACHE_DEBUG
527 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
528 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
531 This enables the /proc/pid/seccomp_cache interface to monitor
532 seccomp cache data. The file format is subject to change. Reading
533 the file requires CAP_SYS_ADMIN.
535 This option is for debugging only. Enabling presents the risk that
536 an adversary may be able to infer the seccomp filter logic.
540 config HAVE_ARCH_STACKLEAK
543 An architecture should select this if it has the code which
544 fills the used part of the kernel stack with the STACKLEAK_POISON
545 value before returning from system calls.
547 config HAVE_STACKPROTECTOR
550 An arch should select this symbol if:
551 - it has implemented a stack canary (e.g. __stack_chk_guard)
553 config STACKPROTECTOR
554 bool "Stack Protector buffer overflow detection"
555 depends on HAVE_STACKPROTECTOR
556 depends on $(cc-option,-fstack-protector)
559 This option turns on the "stack-protector" GCC feature. This
560 feature puts, at the beginning of functions, a canary value on
561 the stack just before the return address, and validates
562 the value just before actually returning. Stack based buffer
563 overflows (that need to overwrite this return address) now also
564 overwrite the canary, which gets detected and the attack is then
565 neutralized via a kernel panic.
567 Functions will have the stack-protector canary logic added if they
568 have an 8-byte or larger character array on the stack.
570 This feature requires gcc version 4.2 or above, or a distribution
571 gcc with the feature backported ("-fstack-protector").
573 On an x86 "defconfig" build, this feature adds canary checks to
574 about 3% of all kernel functions, which increases kernel code size
577 config STACKPROTECTOR_STRONG
578 bool "Strong Stack Protector"
579 depends on STACKPROTECTOR
580 depends on $(cc-option,-fstack-protector-strong)
583 Functions will have the stack-protector canary logic added in any
584 of the following conditions:
586 - local variable's address used as part of the right hand side of an
587 assignment or function argument
588 - local variable is an array (or union containing an array),
589 regardless of array type or length
590 - uses register local variables
592 This feature requires gcc version 4.9 or above, or a distribution
593 gcc with the feature backported ("-fstack-protector-strong").
595 On an x86 "defconfig" build, this feature adds canary checks to
596 about 20% of all kernel functions, which increases the kernel code
599 config ARCH_SUPPORTS_SHADOW_CALL_STACK
602 An architecture should select this if it supports the compiler's
603 Shadow Call Stack and implements runtime support for shadow stack
606 config SHADOW_CALL_STACK
607 bool "Shadow Call Stack"
608 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK
609 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
611 This option enables the compiler's Shadow Call Stack, which
612 uses a shadow stack to protect function return addresses from
613 being overwritten by an attacker. More information can be found
614 in the compiler's documentation:
616 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html
617 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options
619 Note that security guarantees in the kernel differ from the
620 ones documented for user space. The kernel must store addresses
621 of shadow stacks in memory, which means an attacker capable of
622 reading and writing arbitrary memory may be able to locate them
623 and hijack control flow by modifying the stacks.
628 Selected if the kernel will be built using the compiler's LTO feature.
634 Selected if the kernel will be built using Clang's LTO feature.
636 config ARCH_SUPPORTS_LTO_CLANG
639 An architecture should select this option if it supports:
640 - compiling with Clang,
641 - compiling inline assembly with Clang's integrated assembler,
642 - and linking with LLD.
644 config ARCH_SUPPORTS_LTO_CLANG_THIN
647 An architecture should select this option if it can support Clang's
652 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM
653 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm)
654 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm)
655 depends on ARCH_SUPPORTS_LTO_CLANG
656 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT
657 depends on !KASAN || KASAN_HW_TAGS
658 depends on !GCOV_KERNEL
660 The compiler and Kconfig options support building with Clang's
664 prompt "Link Time Optimization (LTO)"
667 This option enables Link Time Optimization (LTO), which allows the
668 compiler to optimize binaries globally.
670 If unsure, select LTO_NONE. Note that LTO is very resource-intensive
671 so it's disabled by default.
676 Build the kernel normally, without Link Time Optimization (LTO).
678 config LTO_CLANG_FULL
679 bool "Clang Full LTO (EXPERIMENTAL)"
680 depends on HAS_LTO_CLANG
681 depends on !COMPILE_TEST
684 This option enables Clang's full Link Time Optimization (LTO), which
685 allows the compiler to optimize the kernel globally. If you enable
686 this option, the compiler generates LLVM bitcode instead of ELF
687 object files, and the actual compilation from bitcode happens at
688 the LTO link step, which may take several minutes depending on the
689 kernel configuration. More information can be found from LLVM's
692 https://llvm.org/docs/LinkTimeOptimization.html
694 During link time, this option can use a large amount of RAM, and
695 may take much longer than the ThinLTO option.
697 config LTO_CLANG_THIN
698 bool "Clang ThinLTO (EXPERIMENTAL)"
699 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN
702 This option enables Clang's ThinLTO, which allows for parallel
703 optimization and faster incremental compiles compared to the
704 CONFIG_LTO_CLANG_FULL option. More information can be found
705 from Clang's documentation:
707 https://clang.llvm.org/docs/ThinLTO.html
712 config ARCH_SUPPORTS_CFI_CLANG
715 An architecture should select this option if it can support Clang's
716 Control-Flow Integrity (CFI) checking.
719 bool "Use Clang's Control Flow Integrity (CFI)"
720 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG
722 # - https://bugs.llvm.org/show_bug.cgi?id=46258
723 # - https://bugs.llvm.org/show_bug.cgi?id=47479
724 depends on CLANG_VERSION >= 120000
727 This option enables Clang’s forward-edge Control Flow Integrity
728 (CFI) checking, where the compiler injects a runtime check to each
729 indirect function call to ensure the target is a valid function with
730 the correct static type. This restricts possible call targets and
731 makes it more difficult for an attacker to exploit bugs that allow
732 the modification of stored function pointers. More information can be
733 found from Clang's documentation:
735 https://clang.llvm.org/docs/ControlFlowIntegrity.html
737 config CFI_CLANG_SHADOW
738 bool "Use CFI shadow to speed up cross-module checks"
740 depends on CFI_CLANG && MODULES
742 If you select this option, the kernel builds a fast look-up table of
743 CFI check functions in loaded modules to reduce performance overhead.
747 config CFI_PERMISSIVE
748 bool "Use CFI in permissive mode"
751 When selected, Control Flow Integrity (CFI) violations result in a
752 warning instead of a kernel panic. This option should only be used
753 for finding indirect call type mismatches during development.
757 config HAVE_ARCH_WITHIN_STACK_FRAMES
760 An architecture should select this if it can walk the kernel stack
761 frames to determine if an object is part of either the arguments
762 or local variables (i.e. that it excludes saved return addresses,
763 and similar) by implementing an inline arch_within_stack_frames(),
764 which is used by CONFIG_HARDENED_USERCOPY.
766 config HAVE_CONTEXT_TRACKING
769 Provide kernel/user boundaries probes necessary for subsystems
770 that need it, such as userspace RCU extended quiescent state.
771 Syscalls need to be wrapped inside user_exit()-user_enter(), either
772 optimized behind static key or through the slow path using TIF_NOHZ
773 flag. Exceptions handlers must be wrapped as well. Irqs are already
774 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
775 handling on irq exit still need to be protected.
777 config HAVE_CONTEXT_TRACKING_OFFSTACK
780 Architecture neither relies on exception_enter()/exception_exit()
781 nor on schedule_user(). Also preempt_schedule_notrace() and
782 preempt_schedule_irq() can't be called in a preemptible section
783 while context tracking is CONTEXT_USER. This feature reflects a sane
784 entry implementation where the following requirements are met on
785 critical entry code, ie: before user_exit() or after user_enter():
787 - Critical entry code isn't preemptible (or better yet:
789 - No use of RCU read side critical sections, unless rcu_nmi_enter()
791 - No use of instrumentation, unless instrumentation_begin() got
797 Arch relies on TIF_NOHZ and syscall slow path to implement context
798 tracking calls to user_enter()/user_exit().
800 config HAVE_VIRT_CPU_ACCOUNTING
803 config HAVE_VIRT_CPU_ACCOUNTING_IDLE
806 Architecture has its own way to account idle CPU time and therefore
807 doesn't implement vtime_account_idle().
809 config ARCH_HAS_SCALED_CPUTIME
812 config HAVE_VIRT_CPU_ACCOUNTING_GEN
816 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
817 Before enabling this option, arch code must be audited
818 to ensure there are no races in concurrent read/write of
819 cputime_t. For example, reading/writing 64-bit cputime_t on
820 some 32-bit arches may require multiple accesses, so proper
821 locking is needed to protect against concurrent accesses.
823 config HAVE_IRQ_TIME_ACCOUNTING
826 Archs need to ensure they use a high enough resolution clock to
827 support irq time accounting and then call enable_sched_clock_irqtime().
832 Architectures that select this are able to move page tables at the
833 PUD level. If there are only 3 page table levels, the move effectively
834 happens at the PGD level.
839 Archs that select this are able to move page tables at the PMD level.
841 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
844 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
847 config HAVE_ARCH_HUGE_VMAP
851 # Archs that select this would be capable of PMD-sized vmaps (i.e.,
852 # arch_vmap_pmd_supported() returns true), and they must make no assumptions
853 # that vmalloc memory is mapped with PAGE_SIZE ptes. The VM_NO_HUGE_VMAP flag
854 # can be used to prohibit arch-specific allocations from using hugepages to
855 # help with this (e.g., modules may require it).
857 config HAVE_ARCH_HUGE_VMALLOC
858 depends on HAVE_ARCH_HUGE_VMAP
861 config ARCH_WANT_HUGE_PMD_SHARE
864 config HAVE_ARCH_SOFT_DIRTY
867 config HAVE_MOD_ARCH_SPECIFIC
870 The arch uses struct mod_arch_specific to store data. Many arches
871 just need a simple module loader without arch specific data - those
872 should not enable this.
874 config MODULES_USE_ELF_RELA
877 Modules only use ELF RELA relocations. Modules with ELF REL
878 relocations will give an error.
880 config MODULES_USE_ELF_REL
883 Modules only use ELF REL relocations. Modules with ELF RELA
884 relocations will give an error.
886 config HAVE_IRQ_EXIT_ON_IRQ_STACK
889 Architecture doesn't only execute the irq handler on the irq stack
890 but also irq_exit(). This way we can process softirqs on this irq
891 stack instead of switching to a new one when we call __do_softirq()
892 in the end of an hardirq.
893 This spares a stack switch and improves cache usage on softirq
896 config HAVE_SOFTIRQ_ON_OWN_STACK
899 Architecture provides a function to run __do_softirq() on a
902 config ALTERNATE_USER_ADDRESS_SPACE
905 Architectures set this when the CPU uses separate address
906 spaces for kernel and user space pointers. In this case, the
907 access_ok() check on a __user pointer is skipped.
909 config PGTABLE_LEVELS
913 config ARCH_HAS_ELF_RANDOMIZE
916 An architecture supports choosing randomized locations for
917 stack, mmap, brk, and ET_DYN. Defined functions:
919 - arch_randomize_brk()
921 config HAVE_ARCH_MMAP_RND_BITS
924 An arch should select this symbol if it supports setting a variable
925 number of bits for use in establishing the base address for mmap
926 allocations, has MMU enabled and provides values for both:
927 - ARCH_MMAP_RND_BITS_MIN
928 - ARCH_MMAP_RND_BITS_MAX
930 config HAVE_EXIT_THREAD
933 An architecture implements exit_thread.
935 config ARCH_MMAP_RND_BITS_MIN
938 config ARCH_MMAP_RND_BITS_MAX
941 config ARCH_MMAP_RND_BITS_DEFAULT
944 config ARCH_MMAP_RND_BITS
945 int "Number of bits to use for ASLR of mmap base address" if EXPERT
946 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
947 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
948 default ARCH_MMAP_RND_BITS_MIN
949 depends on HAVE_ARCH_MMAP_RND_BITS
951 This value can be used to select the number of bits to use to
952 determine the random offset to the base address of vma regions
953 resulting from mmap allocations. This value will be bounded
954 by the architecture's minimum and maximum supported values.
956 This value can be changed after boot using the
957 /proc/sys/vm/mmap_rnd_bits tunable
959 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
962 An arch should select this symbol if it supports running applications
963 in compatibility mode, supports setting a variable number of bits for
964 use in establishing the base address for mmap allocations, has MMU
965 enabled and provides values for both:
966 - ARCH_MMAP_RND_COMPAT_BITS_MIN
967 - ARCH_MMAP_RND_COMPAT_BITS_MAX
969 config ARCH_MMAP_RND_COMPAT_BITS_MIN
972 config ARCH_MMAP_RND_COMPAT_BITS_MAX
975 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
978 config ARCH_MMAP_RND_COMPAT_BITS
979 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
980 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
981 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
982 default ARCH_MMAP_RND_COMPAT_BITS_MIN
983 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
985 This value can be used to select the number of bits to use to
986 determine the random offset to the base address of vma regions
987 resulting from mmap allocations for compatible applications This
988 value will be bounded by the architecture's minimum and maximum
991 This value can be changed after boot using the
992 /proc/sys/vm/mmap_rnd_compat_bits tunable
994 config HAVE_ARCH_COMPAT_MMAP_BASES
997 This allows 64bit applications to invoke 32-bit mmap() syscall
998 and vice-versa 32-bit applications to call 64-bit mmap().
999 Required for applications doing different bitness syscalls.
1001 config PAGE_SIZE_LESS_THAN_64KB
1003 depends on !ARM64_64K_PAGES
1004 depends on !IA64_PAGE_SIZE_64KB
1005 depends on !PAGE_SIZE_64KB
1006 depends on !PARISC_PAGE_SIZE_64KB
1007 depends on !PPC_64K_PAGES
1008 depends on PAGE_SIZE_LESS_THAN_256KB
1010 config PAGE_SIZE_LESS_THAN_256KB
1012 depends on !PPC_256K_PAGES
1013 depends on !PAGE_SIZE_256KB
1015 # This allows to use a set of generic functions to determine mmap base
1016 # address by giving priority to top-down scheme only if the process
1017 # is not in legacy mode (compat task, unlimited stack size or
1018 # sysctl_legacy_va_layout).
1019 # Architecture that selects this option can provide its own version of:
1021 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
1024 select ARCH_HAS_ELF_RANDOMIZE
1026 config HAVE_STACK_VALIDATION
1029 Architecture supports the 'objtool check' host tool command, which
1030 performs compile-time stack metadata validation.
1032 config HAVE_RELIABLE_STACKTRACE
1035 Architecture has either save_stack_trace_tsk_reliable() or
1036 arch_stack_walk_reliable() function which only returns a stack trace
1037 if it can guarantee the trace is reliable.
1039 config HAVE_ARCH_HASH
1043 If this is set, the architecture provides an <asm/hash.h>
1044 file which provides platform-specific implementations of some
1045 functions in <linux/hash.h> or fs/namei.c.
1047 config HAVE_ARCH_NVRAM_OPS
1056 config CLONE_BACKWARDS
1059 Architecture has tls passed as the 4th argument of clone(2),
1062 config CLONE_BACKWARDS2
1065 Architecture has the first two arguments of clone(2) swapped.
1067 config CLONE_BACKWARDS3
1070 Architecture has tls passed as the 3rd argument of clone(2),
1073 config ODD_RT_SIGACTION
1076 Architecture has unusual rt_sigaction(2) arguments
1078 config OLD_SIGSUSPEND
1081 Architecture has old sigsuspend(2) syscall, of one-argument variety
1083 config OLD_SIGSUSPEND3
1086 Even weirder antique ABI - three-argument sigsuspend(2)
1088 config OLD_SIGACTION
1091 Architecture has old sigaction(2) syscall. Nope, not the same
1092 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
1093 but fairly different variant of sigaction(2), thanks to OSF/1
1096 config COMPAT_OLD_SIGACTION
1099 config COMPAT_32BIT_TIME
1100 bool "Provide system calls for 32-bit time_t"
1101 default !64BIT || COMPAT
1103 This enables 32 bit time_t support in addition to 64 bit time_t support.
1104 This is relevant on all 32-bit architectures, and 64-bit architectures
1105 as part of compat syscall handling.
1107 config ARCH_NO_PREEMPT
1110 config ARCH_EPHEMERAL_INODES
1113 An arch should select this symbol if it doesn't keep track of inode
1114 instances on its own, but instead relies on something else (e.g. the
1115 host kernel for an UML kernel).
1117 config ARCH_SUPPORTS_RT
1120 config CPU_NO_EFFICIENT_FFS
1123 config HAVE_ARCH_VMAP_STACK
1126 An arch should select this symbol if it can support kernel stacks
1127 in vmalloc space. This means:
1129 - vmalloc space must be large enough to hold many kernel stacks.
1130 This may rule out many 32-bit architectures.
1132 - Stacks in vmalloc space need to work reliably. For example, if
1133 vmap page tables are created on demand, either this mechanism
1134 needs to work while the stack points to a virtual address with
1135 unpopulated page tables or arch code (switch_to() and switch_mm(),
1136 most likely) needs to ensure that the stack's page table entries
1137 are populated before running on a possibly unpopulated stack.
1139 - If the stack overflows into a guard page, something reasonable
1140 should happen. The definition of "reasonable" is flexible, but
1141 instantly rebooting without logging anything would be unfriendly.
1145 bool "Use a virtually-mapped stack"
1146 depends on HAVE_ARCH_VMAP_STACK
1147 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC
1149 Enable this if you want the use virtually-mapped kernel stacks
1150 with guard pages. This causes kernel stack overflows to be
1151 caught immediately rather than causing difficult-to-diagnose
1154 To use this with software KASAN modes, the architecture must support
1155 backing virtual mappings with real shadow memory, and KASAN_VMALLOC
1158 config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1161 An arch should select this symbol if it can support kernel stack
1162 offset randomization with calls to add_random_kstack_offset()
1163 during syscall entry and choose_random_kstack_offset() during
1164 syscall exit. Careful removal of -fstack-protector-strong and
1165 -fstack-protector should also be applied to the entry code and
1166 closely examined, as the artificial stack bump looks like an array
1167 to the compiler, so it will attempt to add canary checks regardless
1168 of the static branch state.
1170 config RANDOMIZE_KSTACK_OFFSET
1171 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT
1173 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1174 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
1176 The kernel stack offset can be randomized (after pt_regs) by
1177 roughly 5 bits of entropy, frustrating memory corruption
1178 attacks that depend on stack address determinism or
1179 cross-syscall address exposures.
1181 The feature is controlled via the "randomize_kstack_offset=on/off"
1182 kernel boot param, and if turned off has zero overhead due to its use
1183 of static branches (see JUMP_LABEL).
1187 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
1188 bool "Default state of kernel stack offset randomization"
1189 depends on RANDOMIZE_KSTACK_OFFSET
1191 Kernel stack offset randomization is controlled by kernel boot param
1192 "randomize_kstack_offset=on/off", and this config chooses the default
1195 config ARCH_OPTIONAL_KERNEL_RWX
1198 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1201 config ARCH_HAS_STRICT_KERNEL_RWX
1204 config STRICT_KERNEL_RWX
1205 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
1206 depends on ARCH_HAS_STRICT_KERNEL_RWX
1207 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1209 If this is set, kernel text and rodata memory will be made read-only,
1210 and non-text memory will be made non-executable. This provides
1211 protection against certain security exploits (e.g. executing the heap
1214 These features are considered standard security practice these days.
1215 You should say Y here in almost all cases.
1217 config ARCH_HAS_STRICT_MODULE_RWX
1220 config STRICT_MODULE_RWX
1221 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1222 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1223 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1225 If this is set, module text and rodata memory will be made read-only,
1226 and non-text memory will be made non-executable. This provides
1227 protection against certain security exploits (e.g. writing to text)
1229 # select if the architecture provides an asm/dma-direct.h header
1230 config ARCH_HAS_PHYS_TO_DMA
1233 config HAVE_ARCH_COMPILER_H
1236 An architecture can select this if it provides an
1237 asm/compiler.h header that should be included after
1238 linux/compiler-*.h in order to override macro definitions that those
1239 headers generally provide.
1241 config HAVE_ARCH_PREL32_RELOCATIONS
1244 May be selected by an architecture if it supports place-relative
1245 32-bit relocations, both in the toolchain and in the module loader,
1246 in which case relative references can be used in special sections
1247 for PCI fixup, initcalls etc which are only half the size on 64 bit
1248 architectures, and don't require runtime relocation on relocatable
1251 config ARCH_USE_MEMREMAP_PROT
1254 config LOCK_EVENT_COUNTS
1255 bool "Locking event counts collection"
1258 Enable light-weight counting of various locking related events
1259 in the system with minimal performance impact. This reduces
1260 the chance of application behavior change because of timing
1261 differences. The counts are reported via debugfs.
1263 # Select if the architecture has support for applying RELR relocations.
1264 config ARCH_HAS_RELR
1268 bool "Use RELR relocation packing"
1269 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1272 Store the kernel's dynamic relocations in the RELR relocation packing
1273 format. Requires a compatible linker (LLD supports this feature), as
1274 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1277 config ARCH_HAS_MEM_ENCRYPT
1280 config ARCH_HAS_CC_PLATFORM
1283 config HAVE_SPARSE_SYSCALL_NR
1286 An architecture should select this if its syscall numbering is sparse
1287 to save space. For example, MIPS architecture has a syscall array with
1288 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1289 related optimizations for a given architecture.
1291 config ARCH_HAS_VDSO_DATA
1294 config HAVE_STATIC_CALL
1297 config HAVE_STATIC_CALL_INLINE
1299 depends on HAVE_STATIC_CALL
1301 config HAVE_PREEMPT_DYNAMIC
1304 config HAVE_PREEMPT_DYNAMIC_CALL
1306 depends on HAVE_STATIC_CALL
1307 select HAVE_PREEMPT_DYNAMIC
1309 An architecture should select this if it can handle the preemption
1310 model being selected at boot time using static calls.
1312 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a
1313 preemption function will be patched directly.
1315 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any
1316 call to a preemption function will go through a trampoline, and the
1317 trampoline will be patched.
1319 It is strongly advised to support inline static call to avoid any
1322 config HAVE_PREEMPT_DYNAMIC_KEY
1324 depends on HAVE_ARCH_JUMP_LABEL && CC_HAS_ASM_GOTO
1325 select HAVE_PREEMPT_DYNAMIC
1327 An architecture should select this if it can handle the preemption
1328 model being selected at boot time using static keys.
1330 Each preemption function will be given an early return based on a
1331 static key. This should have slightly lower overhead than non-inline
1332 static calls, as this effectively inlines each trampoline into the
1333 start of its callee. This may avoid redundant work, and may
1334 integrate better with CFI schemes.
1336 This will have greater overhead than using inline static calls as
1337 the call to the preemption function cannot be entirely elided.
1339 config ARCH_WANT_LD_ORPHAN_WARN
1342 An arch should select this symbol once all linker sections are explicitly
1343 included, size-asserted, or discarded in the linker scripts. This is
1344 important because we never want expected sections to be placed heuristically
1345 by the linker, since the locations of such sections can change between linker
1348 config HAVE_ARCH_PFN_VALID
1351 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
1354 config ARCH_SUPPORTS_PAGE_TABLE_CHECK
1357 config ARCH_SPLIT_ARG64
1360 If a 32-bit architecture requires 64-bit arguments to be split into
1361 pairs of 32-bit arguments, select this option.
1363 config ARCH_HAS_ELFCORE_COMPAT
1366 config ARCH_HAS_PARANOID_L1D_FLUSH
1369 config DYNAMIC_SIGFRAME
1372 # Select, if arch has a named attribute group bound to NUMA device nodes.
1373 config HAVE_ARCH_NODE_DEV_GROUP
1376 source "kernel/gcov/Kconfig"
1378 source "scripts/gcc-plugins/Kconfig"