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 ptrace_report_syscall_{entry,exit}
221 # TIF_NOTIFY_RESUME calls resume_user_mode_work()
223 config HAVE_ARCH_TRACEHOOK
226 config HAVE_DMA_CONTIGUOUS
229 config GENERIC_SMP_IDLE_THREAD
232 config GENERIC_IDLE_POLL_SETUP
235 config ARCH_HAS_FORTIFY_SOURCE
238 An architecture should select this when it can successfully
239 build and run with CONFIG_FORTIFY_SOURCE.
242 # Select if the arch provides a historic keepinit alias for the retain_initrd
243 # command line option
245 config ARCH_HAS_KEEPINITRD
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 has all set_direct_map_invalid/default() functions
253 config ARCH_HAS_SET_DIRECT_MAP
257 # Select if the architecture provides the arch_dma_set_uncached symbol to
258 # either provide an uncached segment alias for a DMA allocation, or
259 # to remap the page tables in place.
261 config ARCH_HAS_DMA_SET_UNCACHED
265 # Select if the architectures provides the arch_dma_clear_uncached symbol
266 # to undo an in-place page table remap for uncached access.
268 config ARCH_HAS_DMA_CLEAR_UNCACHED
271 # Select if arch init_task must go in the __init_task_data section
272 config ARCH_TASK_STRUCT_ON_STACK
275 # Select if arch has its private alloc_task_struct() function
276 config ARCH_TASK_STRUCT_ALLOCATOR
279 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
281 depends on !ARCH_TASK_STRUCT_ALLOCATOR
283 An architecture should select this to provide hardened usercopy
284 knowledge about what region of the thread_struct should be
285 whitelisted for copying to userspace. Normally this is only the
286 FPU registers. Specifically, arch_thread_struct_whitelist()
287 should be implemented. Without this, the entire thread_struct
288 field in task_struct will be left whitelisted.
290 # Select if arch has its private alloc_thread_stack() function
291 config ARCH_THREAD_STACK_ALLOCATOR
294 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
295 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
298 config ARCH_WANTS_NO_INSTR
301 An architecture should select this if the noinstr macro is being used on
302 functions to denote that the toolchain should avoid instrumenting such
303 functions and is required for correctness.
305 config ARCH_32BIT_OFF_T
309 All new 32-bit architectures should have 64-bit off_t type on
310 userspace side which corresponds to the loff_t kernel type. This
311 is the requirement for modern ABIs. Some existing architectures
312 still support 32-bit off_t. This option is enabled for all such
313 architectures explicitly.
315 # Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat
316 config ARCH_32BIT_USTAT_F_TINODE
319 config HAVE_ASM_MODVERSIONS
322 This symbol should be selected by an architecture if it provides
323 <asm/asm-prototypes.h> to support the module versioning for symbols
324 exported from assembly code.
326 config HAVE_REGS_AND_STACK_ACCESS_API
329 This symbol should be selected by an architecture if it supports
330 the API needed to access registers and stack entries from pt_regs,
331 declared in asm/ptrace.h
332 For example the kprobes-based event tracer needs this API.
336 depends on HAVE_REGS_AND_STACK_ACCESS_API
338 This symbol should be selected by an architecture if it
339 supports an implementation of restartable sequences.
341 config HAVE_FUNCTION_ARG_ACCESS_API
344 This symbol should be selected by an architecture if it supports
345 the API needed to access function arguments from pt_regs,
346 declared in asm/ptrace.h
348 config HAVE_HW_BREAKPOINT
350 depends on PERF_EVENTS
352 config HAVE_MIXED_BREAKPOINTS_REGS
354 depends on HAVE_HW_BREAKPOINT
356 Depending on the arch implementation of hardware breakpoints,
357 some of them have separate registers for data and instruction
358 breakpoints addresses, others have mixed registers to store
359 them but define the access type in a control register.
360 Select this option if your arch implements breakpoints under the
363 config HAVE_USER_RETURN_NOTIFIER
366 config HAVE_PERF_EVENTS_NMI
369 System hardware can generate an NMI using the perf event
370 subsystem. Also has support for calculating CPU cycle events
371 to determine how many clock cycles in a given period.
373 config HAVE_HARDLOCKUP_DETECTOR_PERF
375 depends on HAVE_PERF_EVENTS_NMI
377 The arch chooses to use the generic perf-NMI-based hardlockup
378 detector. Must define HAVE_PERF_EVENTS_NMI.
380 config HAVE_NMI_WATCHDOG
384 The arch provides a low level NMI watchdog. It provides
385 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
387 config HAVE_HARDLOCKUP_DETECTOR_ARCH
389 select HAVE_NMI_WATCHDOG
391 The arch chooses to provide its own hardlockup detector, which is
392 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
393 interfaces and parameters provided by hardlockup detector subsystem.
395 config HAVE_PERF_REGS
398 Support selective register dumps for perf events. This includes
399 bit-mapping of each registers and a unique architecture id.
401 config HAVE_PERF_USER_STACK_DUMP
404 Support user stack dumps for perf event samples. This needs
405 access to the user stack pointer which is not unified across
408 config HAVE_ARCH_JUMP_LABEL
411 config HAVE_ARCH_JUMP_LABEL_RELATIVE
414 config MMU_GATHER_TABLE_FREE
417 config MMU_GATHER_RCU_TABLE_FREE
419 select MMU_GATHER_TABLE_FREE
421 config MMU_GATHER_PAGE_SIZE
424 config MMU_GATHER_NO_RANGE
427 config MMU_GATHER_NO_GATHER
429 depends on MMU_GATHER_TABLE_FREE
431 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
434 Temporary select until all architectures can be converted to have
435 irqs disabled over activate_mm. Architectures that do IPI based TLB
436 shootdowns should enable this.
438 config ARCH_HAVE_NMI_SAFE_CMPXCHG
441 config HAVE_ALIGNED_STRUCT_PAGE
444 This makes sure that struct pages are double word aligned and that
445 e.g. the SLUB allocator can perform double word atomic operations
446 on a struct page for better performance. However selecting this
447 might increase the size of a struct page by a word.
449 config HAVE_CMPXCHG_LOCAL
452 config HAVE_CMPXCHG_DOUBLE
455 config ARCH_WEAK_RELEASE_ACQUIRE
458 config ARCH_WANT_IPC_PARSE_VERSION
461 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
464 config ARCH_WANT_OLD_COMPAT_IPC
465 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
468 config HAVE_ARCH_SECCOMP
471 An arch should select this symbol to support seccomp mode 1 (the fixed
472 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
473 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
474 - __NR_seccomp_read_32
475 - __NR_seccomp_write_32
476 - __NR_seccomp_exit_32
477 - __NR_seccomp_sigreturn_32
479 config HAVE_ARCH_SECCOMP_FILTER
481 select HAVE_ARCH_SECCOMP
483 An arch should select this symbol if it provides all of these things:
484 - all the requirements for HAVE_ARCH_SECCOMP
486 - syscall_get_arguments()
488 - syscall_set_return_value()
489 - SIGSYS siginfo_t support
490 - secure_computing is called from a ptrace_event()-safe context
491 - secure_computing return value is checked and a return value of -1
492 results in the system call being skipped immediately.
493 - seccomp syscall wired up
494 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
495 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
496 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
499 prompt "Enable seccomp to safely execute untrusted bytecode"
501 depends on HAVE_ARCH_SECCOMP
503 This kernel feature is useful for number crunching applications
504 that may need to handle untrusted bytecode during their
505 execution. By using pipes or other transports made available
506 to the process as file descriptors supporting the read/write
507 syscalls, it's possible to isolate those applications in their
508 own address space using seccomp. Once seccomp is enabled via
509 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
510 disabled and the task is only allowed to execute a few safe
511 syscalls defined by each seccomp mode.
515 config SECCOMP_FILTER
517 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
519 Enable tasks to build secure computing environments defined
520 in terms of Berkeley Packet Filter programs which implement
521 task-defined system call filtering polices.
523 See Documentation/userspace-api/seccomp_filter.rst for details.
525 config SECCOMP_CACHE_DEBUG
526 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
527 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
530 This enables the /proc/pid/seccomp_cache interface to monitor
531 seccomp cache data. The file format is subject to change. Reading
532 the file requires CAP_SYS_ADMIN.
534 This option is for debugging only. Enabling presents the risk that
535 an adversary may be able to infer the seccomp filter logic.
539 config HAVE_ARCH_STACKLEAK
542 An architecture should select this if it has the code which
543 fills the used part of the kernel stack with the STACKLEAK_POISON
544 value before returning from system calls.
546 config HAVE_STACKPROTECTOR
549 An arch should select this symbol if:
550 - it has implemented a stack canary (e.g. __stack_chk_guard)
552 config STACKPROTECTOR
553 bool "Stack Protector buffer overflow detection"
554 depends on HAVE_STACKPROTECTOR
555 depends on $(cc-option,-fstack-protector)
558 This option turns on the "stack-protector" GCC feature. This
559 feature puts, at the beginning of functions, a canary value on
560 the stack just before the return address, and validates
561 the value just before actually returning. Stack based buffer
562 overflows (that need to overwrite this return address) now also
563 overwrite the canary, which gets detected and the attack is then
564 neutralized via a kernel panic.
566 Functions will have the stack-protector canary logic added if they
567 have an 8-byte or larger character array on the stack.
569 This feature requires gcc version 4.2 or above, or a distribution
570 gcc with the feature backported ("-fstack-protector").
572 On an x86 "defconfig" build, this feature adds canary checks to
573 about 3% of all kernel functions, which increases kernel code size
576 config STACKPROTECTOR_STRONG
577 bool "Strong Stack Protector"
578 depends on STACKPROTECTOR
579 depends on $(cc-option,-fstack-protector-strong)
582 Functions will have the stack-protector canary logic added in any
583 of the following conditions:
585 - local variable's address used as part of the right hand side of an
586 assignment or function argument
587 - local variable is an array (or union containing an array),
588 regardless of array type or length
589 - uses register local variables
591 This feature requires gcc version 4.9 or above, or a distribution
592 gcc with the feature backported ("-fstack-protector-strong").
594 On an x86 "defconfig" build, this feature adds canary checks to
595 about 20% of all kernel functions, which increases the kernel code
598 config ARCH_SUPPORTS_SHADOW_CALL_STACK
601 An architecture should select this if it supports the compiler's
602 Shadow Call Stack and implements runtime support for shadow stack
605 config SHADOW_CALL_STACK
606 bool "Shadow Call Stack"
607 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK
608 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
610 This option enables the compiler's Shadow Call Stack, which
611 uses a shadow stack to protect function return addresses from
612 being overwritten by an attacker. More information can be found
613 in the compiler's documentation:
615 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html
616 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options
618 Note that security guarantees in the kernel differ from the
619 ones documented for user space. The kernel must store addresses
620 of shadow stacks in memory, which means an attacker capable of
621 reading and writing arbitrary memory may be able to locate them
622 and hijack control flow by modifying the stacks.
627 Selected if the kernel will be built using the compiler's LTO feature.
633 Selected if the kernel will be built using Clang's LTO feature.
635 config ARCH_SUPPORTS_LTO_CLANG
638 An architecture should select this option if it supports:
639 - compiling with Clang,
640 - compiling inline assembly with Clang's integrated assembler,
641 - and linking with LLD.
643 config ARCH_SUPPORTS_LTO_CLANG_THIN
646 An architecture should select this option if it can support Clang's
651 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM
652 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm)
653 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm)
654 depends on ARCH_SUPPORTS_LTO_CLANG
655 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT
656 depends on !KASAN || KASAN_HW_TAGS
657 depends on !GCOV_KERNEL
659 The compiler and Kconfig options support building with Clang's
663 prompt "Link Time Optimization (LTO)"
666 This option enables Link Time Optimization (LTO), which allows the
667 compiler to optimize binaries globally.
669 If unsure, select LTO_NONE. Note that LTO is very resource-intensive
670 so it's disabled by default.
675 Build the kernel normally, without Link Time Optimization (LTO).
677 config LTO_CLANG_FULL
678 bool "Clang Full LTO (EXPERIMENTAL)"
679 depends on HAS_LTO_CLANG
680 depends on !COMPILE_TEST
683 This option enables Clang's full Link Time Optimization (LTO), which
684 allows the compiler to optimize the kernel globally. If you enable
685 this option, the compiler generates LLVM bitcode instead of ELF
686 object files, and the actual compilation from bitcode happens at
687 the LTO link step, which may take several minutes depending on the
688 kernel configuration. More information can be found from LLVM's
691 https://llvm.org/docs/LinkTimeOptimization.html
693 During link time, this option can use a large amount of RAM, and
694 may take much longer than the ThinLTO option.
696 config LTO_CLANG_THIN
697 bool "Clang ThinLTO (EXPERIMENTAL)"
698 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN
701 This option enables Clang's ThinLTO, which allows for parallel
702 optimization and faster incremental compiles compared to the
703 CONFIG_LTO_CLANG_FULL option. More information can be found
704 from Clang's documentation:
706 https://clang.llvm.org/docs/ThinLTO.html
711 config ARCH_SUPPORTS_CFI_CLANG
714 An architecture should select this option if it can support Clang's
715 Control-Flow Integrity (CFI) checking.
718 bool "Use Clang's Control Flow Integrity (CFI)"
719 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG
721 # - https://bugs.llvm.org/show_bug.cgi?id=46258
722 # - https://bugs.llvm.org/show_bug.cgi?id=47479
723 depends on CLANG_VERSION >= 120000
726 This option enables Clang’s forward-edge Control Flow Integrity
727 (CFI) checking, where the compiler injects a runtime check to each
728 indirect function call to ensure the target is a valid function with
729 the correct static type. This restricts possible call targets and
730 makes it more difficult for an attacker to exploit bugs that allow
731 the modification of stored function pointers. More information can be
732 found from Clang's documentation:
734 https://clang.llvm.org/docs/ControlFlowIntegrity.html
736 config CFI_CLANG_SHADOW
737 bool "Use CFI shadow to speed up cross-module checks"
739 depends on CFI_CLANG && MODULES
741 If you select this option, the kernel builds a fast look-up table of
742 CFI check functions in loaded modules to reduce performance overhead.
746 config CFI_PERMISSIVE
747 bool "Use CFI in permissive mode"
750 When selected, Control Flow Integrity (CFI) violations result in a
751 warning instead of a kernel panic. This option should only be used
752 for finding indirect call type mismatches during development.
756 config HAVE_ARCH_WITHIN_STACK_FRAMES
759 An architecture should select this if it can walk the kernel stack
760 frames to determine if an object is part of either the arguments
761 or local variables (i.e. that it excludes saved return addresses,
762 and similar) by implementing an inline arch_within_stack_frames(),
763 which is used by CONFIG_HARDENED_USERCOPY.
765 config HAVE_CONTEXT_TRACKING
768 Provide kernel/user boundaries probes necessary for subsystems
769 that need it, such as userspace RCU extended quiescent state.
770 Syscalls need to be wrapped inside user_exit()-user_enter(), either
771 optimized behind static key or through the slow path using TIF_NOHZ
772 flag. Exceptions handlers must be wrapped as well. Irqs are already
773 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
774 handling on irq exit still need to be protected.
776 config HAVE_CONTEXT_TRACKING_OFFSTACK
779 Architecture neither relies on exception_enter()/exception_exit()
780 nor on schedule_user(). Also preempt_schedule_notrace() and
781 preempt_schedule_irq() can't be called in a preemptible section
782 while context tracking is CONTEXT_USER. This feature reflects a sane
783 entry implementation where the following requirements are met on
784 critical entry code, ie: before user_exit() or after user_enter():
786 - Critical entry code isn't preemptible (or better yet:
788 - No use of RCU read side critical sections, unless rcu_nmi_enter()
790 - No use of instrumentation, unless instrumentation_begin() got
796 Arch relies on TIF_NOHZ and syscall slow path to implement context
797 tracking calls to user_enter()/user_exit().
799 config HAVE_VIRT_CPU_ACCOUNTING
802 config HAVE_VIRT_CPU_ACCOUNTING_IDLE
805 Architecture has its own way to account idle CPU time and therefore
806 doesn't implement vtime_account_idle().
808 config ARCH_HAS_SCALED_CPUTIME
811 config HAVE_VIRT_CPU_ACCOUNTING_GEN
815 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
816 Before enabling this option, arch code must be audited
817 to ensure there are no races in concurrent read/write of
818 cputime_t. For example, reading/writing 64-bit cputime_t on
819 some 32-bit arches may require multiple accesses, so proper
820 locking is needed to protect against concurrent accesses.
822 config HAVE_IRQ_TIME_ACCOUNTING
825 Archs need to ensure they use a high enough resolution clock to
826 support irq time accounting and then call enable_sched_clock_irqtime().
831 Architectures that select this are able to move page tables at the
832 PUD level. If there are only 3 page table levels, the move effectively
833 happens at the PGD level.
838 Archs that select this are able to move page tables at the PMD level.
840 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
843 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
846 config HAVE_ARCH_HUGE_VMAP
850 # Archs that select this would be capable of PMD-sized vmaps (i.e.,
851 # arch_vmap_pmd_supported() returns true), and they must make no assumptions
852 # that vmalloc memory is mapped with PAGE_SIZE ptes. The VM_NO_HUGE_VMAP flag
853 # can be used to prohibit arch-specific allocations from using hugepages to
854 # help with this (e.g., modules may require it).
856 config HAVE_ARCH_HUGE_VMALLOC
857 depends on HAVE_ARCH_HUGE_VMAP
860 config ARCH_WANT_HUGE_PMD_SHARE
863 config HAVE_ARCH_SOFT_DIRTY
866 config HAVE_MOD_ARCH_SPECIFIC
869 The arch uses struct mod_arch_specific to store data. Many arches
870 just need a simple module loader without arch specific data - those
871 should not enable this.
873 config MODULES_USE_ELF_RELA
876 Modules only use ELF RELA relocations. Modules with ELF REL
877 relocations will give an error.
879 config MODULES_USE_ELF_REL
882 Modules only use ELF REL relocations. Modules with ELF RELA
883 relocations will give an error.
885 config HAVE_IRQ_EXIT_ON_IRQ_STACK
888 Architecture doesn't only execute the irq handler on the irq stack
889 but also irq_exit(). This way we can process softirqs on this irq
890 stack instead of switching to a new one when we call __do_softirq()
891 in the end of an hardirq.
892 This spares a stack switch and improves cache usage on softirq
895 config HAVE_SOFTIRQ_ON_OWN_STACK
898 Architecture provides a function to run __do_softirq() on a
901 config ALTERNATE_USER_ADDRESS_SPACE
904 Architectures set this when the CPU uses separate address
905 spaces for kernel and user space pointers. In this case, the
906 access_ok() check on a __user pointer is skipped.
908 config PGTABLE_LEVELS
912 config ARCH_HAS_ELF_RANDOMIZE
915 An architecture supports choosing randomized locations for
916 stack, mmap, brk, and ET_DYN. Defined functions:
918 - arch_randomize_brk()
920 config HAVE_ARCH_MMAP_RND_BITS
923 An arch should select this symbol if it supports setting a variable
924 number of bits for use in establishing the base address for mmap
925 allocations, has MMU enabled and provides values for both:
926 - ARCH_MMAP_RND_BITS_MIN
927 - ARCH_MMAP_RND_BITS_MAX
929 config HAVE_EXIT_THREAD
932 An architecture implements exit_thread.
934 config ARCH_MMAP_RND_BITS_MIN
937 config ARCH_MMAP_RND_BITS_MAX
940 config ARCH_MMAP_RND_BITS_DEFAULT
943 config ARCH_MMAP_RND_BITS
944 int "Number of bits to use for ASLR of mmap base address" if EXPERT
945 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
946 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
947 default ARCH_MMAP_RND_BITS_MIN
948 depends on HAVE_ARCH_MMAP_RND_BITS
950 This value can be used to select the number of bits to use to
951 determine the random offset to the base address of vma regions
952 resulting from mmap allocations. This value will be bounded
953 by the architecture's minimum and maximum supported values.
955 This value can be changed after boot using the
956 /proc/sys/vm/mmap_rnd_bits tunable
958 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
961 An arch should select this symbol if it supports running applications
962 in compatibility mode, supports setting a variable number of bits for
963 use in establishing the base address for mmap allocations, has MMU
964 enabled and provides values for both:
965 - ARCH_MMAP_RND_COMPAT_BITS_MIN
966 - ARCH_MMAP_RND_COMPAT_BITS_MAX
968 config ARCH_MMAP_RND_COMPAT_BITS_MIN
971 config ARCH_MMAP_RND_COMPAT_BITS_MAX
974 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
977 config ARCH_MMAP_RND_COMPAT_BITS
978 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
979 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
980 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
981 default ARCH_MMAP_RND_COMPAT_BITS_MIN
982 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
984 This value can be used to select the number of bits to use to
985 determine the random offset to the base address of vma regions
986 resulting from mmap allocations for compatible applications This
987 value will be bounded by the architecture's minimum and maximum
990 This value can be changed after boot using the
991 /proc/sys/vm/mmap_rnd_compat_bits tunable
993 config HAVE_ARCH_COMPAT_MMAP_BASES
996 This allows 64bit applications to invoke 32-bit mmap() syscall
997 and vice-versa 32-bit applications to call 64-bit mmap().
998 Required for applications doing different bitness syscalls.
1000 config PAGE_SIZE_LESS_THAN_64KB
1002 depends on !ARM64_64K_PAGES
1003 depends on !IA64_PAGE_SIZE_64KB
1004 depends on !PAGE_SIZE_64KB
1005 depends on !PARISC_PAGE_SIZE_64KB
1006 depends on !PPC_64K_PAGES
1007 depends on PAGE_SIZE_LESS_THAN_256KB
1009 config PAGE_SIZE_LESS_THAN_256KB
1011 depends on !PPC_256K_PAGES
1012 depends on !PAGE_SIZE_256KB
1014 # This allows to use a set of generic functions to determine mmap base
1015 # address by giving priority to top-down scheme only if the process
1016 # is not in legacy mode (compat task, unlimited stack size or
1017 # sysctl_legacy_va_layout).
1018 # Architecture that selects this option can provide its own version of:
1020 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
1023 select ARCH_HAS_ELF_RANDOMIZE
1025 config HAVE_STACK_VALIDATION
1028 Architecture supports the 'objtool check' host tool command, which
1029 performs compile-time stack metadata validation.
1031 config HAVE_RELIABLE_STACKTRACE
1034 Architecture has either save_stack_trace_tsk_reliable() or
1035 arch_stack_walk_reliable() function which only returns a stack trace
1036 if it can guarantee the trace is reliable.
1038 config HAVE_ARCH_HASH
1042 If this is set, the architecture provides an <asm/hash.h>
1043 file which provides platform-specific implementations of some
1044 functions in <linux/hash.h> or fs/namei.c.
1046 config HAVE_ARCH_NVRAM_OPS
1055 config CLONE_BACKWARDS
1058 Architecture has tls passed as the 4th argument of clone(2),
1061 config CLONE_BACKWARDS2
1064 Architecture has the first two arguments of clone(2) swapped.
1066 config CLONE_BACKWARDS3
1069 Architecture has tls passed as the 3rd argument of clone(2),
1072 config ODD_RT_SIGACTION
1075 Architecture has unusual rt_sigaction(2) arguments
1077 config OLD_SIGSUSPEND
1080 Architecture has old sigsuspend(2) syscall, of one-argument variety
1082 config OLD_SIGSUSPEND3
1085 Even weirder antique ABI - three-argument sigsuspend(2)
1087 config OLD_SIGACTION
1090 Architecture has old sigaction(2) syscall. Nope, not the same
1091 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
1092 but fairly different variant of sigaction(2), thanks to OSF/1
1095 config COMPAT_OLD_SIGACTION
1098 config COMPAT_32BIT_TIME
1099 bool "Provide system calls for 32-bit time_t"
1100 default !64BIT || COMPAT
1102 This enables 32 bit time_t support in addition to 64 bit time_t support.
1103 This is relevant on all 32-bit architectures, and 64-bit architectures
1104 as part of compat syscall handling.
1106 config ARCH_NO_PREEMPT
1109 config ARCH_EPHEMERAL_INODES
1112 An arch should select this symbol if it doesn't keep track of inode
1113 instances on its own, but instead relies on something else (e.g. the
1114 host kernel for an UML kernel).
1116 config ARCH_SUPPORTS_RT
1119 config CPU_NO_EFFICIENT_FFS
1122 config HAVE_ARCH_VMAP_STACK
1125 An arch should select this symbol if it can support kernel stacks
1126 in vmalloc space. This means:
1128 - vmalloc space must be large enough to hold many kernel stacks.
1129 This may rule out many 32-bit architectures.
1131 - Stacks in vmalloc space need to work reliably. For example, if
1132 vmap page tables are created on demand, either this mechanism
1133 needs to work while the stack points to a virtual address with
1134 unpopulated page tables or arch code (switch_to() and switch_mm(),
1135 most likely) needs to ensure that the stack's page table entries
1136 are populated before running on a possibly unpopulated stack.
1138 - If the stack overflows into a guard page, something reasonable
1139 should happen. The definition of "reasonable" is flexible, but
1140 instantly rebooting without logging anything would be unfriendly.
1144 bool "Use a virtually-mapped stack"
1145 depends on HAVE_ARCH_VMAP_STACK
1146 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC
1148 Enable this if you want the use virtually-mapped kernel stacks
1149 with guard pages. This causes kernel stack overflows to be
1150 caught immediately rather than causing difficult-to-diagnose
1153 To use this with software KASAN modes, the architecture must support
1154 backing virtual mappings with real shadow memory, and KASAN_VMALLOC
1157 config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1160 An arch should select this symbol if it can support kernel stack
1161 offset randomization with calls to add_random_kstack_offset()
1162 during syscall entry and choose_random_kstack_offset() during
1163 syscall exit. Careful removal of -fstack-protector-strong and
1164 -fstack-protector should also be applied to the entry code and
1165 closely examined, as the artificial stack bump looks like an array
1166 to the compiler, so it will attempt to add canary checks regardless
1167 of the static branch state.
1169 config RANDOMIZE_KSTACK_OFFSET
1170 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT
1172 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1173 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
1175 The kernel stack offset can be randomized (after pt_regs) by
1176 roughly 5 bits of entropy, frustrating memory corruption
1177 attacks that depend on stack address determinism or
1178 cross-syscall address exposures.
1180 The feature is controlled via the "randomize_kstack_offset=on/off"
1181 kernel boot param, and if turned off has zero overhead due to its use
1182 of static branches (see JUMP_LABEL).
1186 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
1187 bool "Default state of kernel stack offset randomization"
1188 depends on RANDOMIZE_KSTACK_OFFSET
1190 Kernel stack offset randomization is controlled by kernel boot param
1191 "randomize_kstack_offset=on/off", and this config chooses the default
1194 config ARCH_OPTIONAL_KERNEL_RWX
1197 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1200 config ARCH_HAS_STRICT_KERNEL_RWX
1203 config STRICT_KERNEL_RWX
1204 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
1205 depends on ARCH_HAS_STRICT_KERNEL_RWX
1206 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1208 If this is set, kernel text and rodata memory will be made read-only,
1209 and non-text memory will be made non-executable. This provides
1210 protection against certain security exploits (e.g. executing the heap
1213 These features are considered standard security practice these days.
1214 You should say Y here in almost all cases.
1216 config ARCH_HAS_STRICT_MODULE_RWX
1219 config STRICT_MODULE_RWX
1220 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1221 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1222 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1224 If this is set, module text and rodata memory will be made read-only,
1225 and non-text memory will be made non-executable. This provides
1226 protection against certain security exploits (e.g. writing to text)
1228 # select if the architecture provides an asm/dma-direct.h header
1229 config ARCH_HAS_PHYS_TO_DMA
1232 config HAVE_ARCH_COMPILER_H
1235 An architecture can select this if it provides an
1236 asm/compiler.h header that should be included after
1237 linux/compiler-*.h in order to override macro definitions that those
1238 headers generally provide.
1240 config HAVE_ARCH_PREL32_RELOCATIONS
1243 May be selected by an architecture if it supports place-relative
1244 32-bit relocations, both in the toolchain and in the module loader,
1245 in which case relative references can be used in special sections
1246 for PCI fixup, initcalls etc which are only half the size on 64 bit
1247 architectures, and don't require runtime relocation on relocatable
1250 config ARCH_USE_MEMREMAP_PROT
1253 config LOCK_EVENT_COUNTS
1254 bool "Locking event counts collection"
1257 Enable light-weight counting of various locking related events
1258 in the system with minimal performance impact. This reduces
1259 the chance of application behavior change because of timing
1260 differences. The counts are reported via debugfs.
1262 # Select if the architecture has support for applying RELR relocations.
1263 config ARCH_HAS_RELR
1267 bool "Use RELR relocation packing"
1268 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1271 Store the kernel's dynamic relocations in the RELR relocation packing
1272 format. Requires a compatible linker (LLD supports this feature), as
1273 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1276 config ARCH_HAS_MEM_ENCRYPT
1279 config ARCH_HAS_CC_PLATFORM
1282 config HAVE_SPARSE_SYSCALL_NR
1285 An architecture should select this if its syscall numbering is sparse
1286 to save space. For example, MIPS architecture has a syscall array with
1287 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1288 related optimizations for a given architecture.
1290 config ARCH_HAS_VDSO_DATA
1293 config HAVE_STATIC_CALL
1296 config HAVE_STATIC_CALL_INLINE
1298 depends on HAVE_STATIC_CALL
1300 config HAVE_PREEMPT_DYNAMIC
1303 config HAVE_PREEMPT_DYNAMIC_CALL
1305 depends on HAVE_STATIC_CALL
1306 select HAVE_PREEMPT_DYNAMIC
1308 An architecture should select this if it can handle the preemption
1309 model being selected at boot time using static calls.
1311 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a
1312 preemption function will be patched directly.
1314 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any
1315 call to a preemption function will go through a trampoline, and the
1316 trampoline will be patched.
1318 It is strongly advised to support inline static call to avoid any
1321 config HAVE_PREEMPT_DYNAMIC_KEY
1323 depends on HAVE_ARCH_JUMP_LABEL && CC_HAS_ASM_GOTO
1324 select HAVE_PREEMPT_DYNAMIC
1326 An architecture should select this if it can handle the preemption
1327 model being selected at boot time using static keys.
1329 Each preemption function will be given an early return based on a
1330 static key. This should have slightly lower overhead than non-inline
1331 static calls, as this effectively inlines each trampoline into the
1332 start of its callee. This may avoid redundant work, and may
1333 integrate better with CFI schemes.
1335 This will have greater overhead than using inline static calls as
1336 the call to the preemption function cannot be entirely elided.
1338 config ARCH_WANT_LD_ORPHAN_WARN
1341 An arch should select this symbol once all linker sections are explicitly
1342 included, size-asserted, or discarded in the linker scripts. This is
1343 important because we never want expected sections to be placed heuristically
1344 by the linker, since the locations of such sections can change between linker
1347 config HAVE_ARCH_PFN_VALID
1350 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
1353 config ARCH_SUPPORTS_PAGE_TABLE_CHECK
1356 config ARCH_SPLIT_ARG64
1359 If a 32-bit architecture requires 64-bit arguments to be split into
1360 pairs of 32-bit arguments, select this option.
1362 config ARCH_HAS_ELFCORE_COMPAT
1365 config ARCH_HAS_PARANOID_L1D_FLUSH
1368 config DYNAMIC_SIGFRAME
1371 # Select, if arch has a named attribute group bound to NUMA device nodes.
1372 config HAVE_ARCH_NODE_DEV_GROUP
1375 source "kernel/gcov/Kconfig"
1377 source "scripts/gcc-plugins/Kconfig"