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"
27 config ARCH_HAS_SUBPAGE_FAULTS
30 Select if the architecture can check permissions at sub-page
31 granularity (e.g. arm64 MTE). The probe_user_*() functions
43 depends on HAVE_KPROBES
45 select TASKS_RCU if PREEMPTION
47 Kprobes allows you to trap at almost any kernel address and
48 execute a callback function. register_kprobe() establishes
49 a probepoint and specifies the callback. Kprobes is useful
50 for kernel debugging, non-intrusive instrumentation and testing.
54 bool "Optimize very unlikely/likely branches"
55 depends on HAVE_ARCH_JUMP_LABEL
56 depends on CC_HAS_ASM_GOTO
57 select OBJTOOL if HAVE_JUMP_LABEL_HACK
59 This option enables a transparent branch optimization that
60 makes certain almost-always-true or almost-always-false branch
61 conditions even cheaper to execute within the kernel.
63 Certain performance-sensitive kernel code, such as trace points,
64 scheduler functionality, networking code and KVM have such
65 branches and include support for this optimization technique.
67 If it is detected that the compiler has support for "asm goto",
68 the kernel will compile such branches with just a nop
69 instruction. When the condition flag is toggled to true, the
70 nop will be converted to a jump instruction to execute the
71 conditional block of instructions.
73 This technique lowers overhead and stress on the branch prediction
74 of the processor and generally makes the kernel faster. The update
75 of the condition is slower, but those are always very rare.
77 ( On 32-bit x86, the necessary options added to the compiler
78 flags may increase the size of the kernel slightly. )
80 config STATIC_KEYS_SELFTEST
81 bool "Static key selftest"
84 Boot time self-test of the branch patching code.
86 config STATIC_CALL_SELFTEST
87 bool "Static call selftest"
88 depends on HAVE_STATIC_CALL
90 Boot time self-test of the call patching code.
94 depends on KPROBES && HAVE_OPTPROBES
95 select TASKS_RCU if PREEMPTION
97 config KPROBES_ON_FTRACE
99 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
100 depends on DYNAMIC_FTRACE_WITH_REGS
102 If function tracer is enabled and the arch supports full
103 passing of pt_regs to function tracing, then kprobes can
104 optimize on top of function tracing.
108 depends on ARCH_SUPPORTS_UPROBES
110 Uprobes is the user-space counterpart to kprobes: they
111 enable instrumentation applications (such as 'perf probe')
112 to establish unintrusive probes in user-space binaries and
113 libraries, by executing handler functions when the probes
114 are hit by user-space applications.
116 ( These probes come in the form of single-byte breakpoints,
117 managed by the kernel and kept transparent to the probed
120 config HAVE_64BIT_ALIGNED_ACCESS
121 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
123 Some architectures require 64 bit accesses to be 64 bit
124 aligned, which also requires structs containing 64 bit values
125 to be 64 bit aligned too. This includes some 32 bit
126 architectures which can do 64 bit accesses, as well as 64 bit
127 architectures without unaligned access.
129 This symbol should be selected by an architecture if 64 bit
130 accesses are required to be 64 bit aligned in this way even
131 though it is not a 64 bit architecture.
133 See Documentation/core-api/unaligned-memory-access.rst for
134 more information on the topic of unaligned memory accesses.
136 config HAVE_EFFICIENT_UNALIGNED_ACCESS
139 Some architectures are unable to perform unaligned accesses
140 without the use of get_unaligned/put_unaligned. Others are
141 unable to perform such accesses efficiently (e.g. trap on
142 unaligned access and require fixing it up in the exception
145 This symbol should be selected by an architecture if it can
146 perform unaligned accesses efficiently to allow different
147 code paths to be selected for these cases. Some network
148 drivers, for example, could opt to not fix up alignment
149 problems with received packets if doing so would not help
152 See Documentation/core-api/unaligned-memory-access.rst for more
153 information on the topic of unaligned memory accesses.
155 config ARCH_USE_BUILTIN_BSWAP
158 Modern versions of GCC (since 4.4) have builtin functions
159 for handling byte-swapping. Using these, instead of the old
160 inline assembler that the architecture code provides in the
161 __arch_bswapXX() macros, allows the compiler to see what's
162 happening and offers more opportunity for optimisation. In
163 particular, the compiler will be able to combine the byteswap
164 with a nearby load or store and use load-and-swap or
165 store-and-swap instructions if the architecture has them. It
166 should almost *never* result in code which is worse than the
167 hand-coded assembler in <asm/swab.h>. But just in case it
168 does, the use of the builtins is optional.
170 Any architecture with load-and-swap or store-and-swap
171 instructions should set this. And it shouldn't hurt to set it
172 on architectures that don't have such instructions.
176 depends on KPROBES && (HAVE_KRETPROBES || HAVE_RETHOOK)
178 config KRETPROBE_ON_RETHOOK
180 depends on HAVE_RETHOOK
181 depends on KRETPROBES
184 config USER_RETURN_NOTIFIER
186 depends on HAVE_USER_RETURN_NOTIFIER
188 Provide a kernel-internal notification when a cpu is about to
191 config HAVE_IOREMAP_PROT
197 config HAVE_KRETPROBES
200 config HAVE_OPTPROBES
203 config HAVE_KPROBES_ON_FTRACE
206 config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE
209 Since kretprobes modifies return address on the stack, the
210 stacktrace may see the kretprobe trampoline address instead
211 of correct one. If the architecture stacktrace code and
212 unwinder can adjust such entries, select this configuration.
214 config HAVE_FUNCTION_ERROR_INJECTION
220 config HAVE_FUNCTION_DESCRIPTORS
223 config TRACE_IRQFLAGS_SUPPORT
226 config TRACE_IRQFLAGS_NMI_SUPPORT
230 # An arch should select this if it provides all these things:
232 # task_pt_regs() in asm/processor.h or asm/ptrace.h
233 # arch_has_single_step() if there is hardware single-step support
234 # arch_has_block_step() if there is hardware block-step support
235 # asm/syscall.h supplying asm-generic/syscall.h interface
236 # linux/regset.h user_regset interfaces
237 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
238 # TIF_SYSCALL_TRACE calls ptrace_report_syscall_{entry,exit}
239 # TIF_NOTIFY_RESUME calls resume_user_mode_work()
241 config HAVE_ARCH_TRACEHOOK
244 config HAVE_DMA_CONTIGUOUS
247 config GENERIC_SMP_IDLE_THREAD
250 config GENERIC_IDLE_POLL_SETUP
253 config ARCH_HAS_FORTIFY_SOURCE
256 An architecture should select this when it can successfully
257 build and run with CONFIG_FORTIFY_SOURCE.
260 # Select if the arch provides a historic keepinit alias for the retain_initrd
261 # command line option
263 config ARCH_HAS_KEEPINITRD
266 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
267 config ARCH_HAS_SET_MEMORY
270 # Select if arch has all set_direct_map_invalid/default() functions
271 config ARCH_HAS_SET_DIRECT_MAP
275 # Select if the architecture provides the arch_dma_set_uncached symbol to
276 # either provide an uncached segment alias for a DMA allocation, or
277 # to remap the page tables in place.
279 config ARCH_HAS_DMA_SET_UNCACHED
283 # Select if the architectures provides the arch_dma_clear_uncached symbol
284 # to undo an in-place page table remap for uncached access.
286 config ARCH_HAS_DMA_CLEAR_UNCACHED
289 # Select if arch init_task must go in the __init_task_data section
290 config ARCH_TASK_STRUCT_ON_STACK
293 # Select if arch has its private alloc_task_struct() function
294 config ARCH_TASK_STRUCT_ALLOCATOR
297 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
299 depends on !ARCH_TASK_STRUCT_ALLOCATOR
301 An architecture should select this to provide hardened usercopy
302 knowledge about what region of the thread_struct should be
303 whitelisted for copying to userspace. Normally this is only the
304 FPU registers. Specifically, arch_thread_struct_whitelist()
305 should be implemented. Without this, the entire thread_struct
306 field in task_struct will be left whitelisted.
308 # Select if arch has its private alloc_thread_stack() function
309 config ARCH_THREAD_STACK_ALLOCATOR
312 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
313 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
316 config ARCH_WANTS_NO_INSTR
319 An architecture should select this if the noinstr macro is being used on
320 functions to denote that the toolchain should avoid instrumenting such
321 functions and is required for correctness.
323 config ARCH_32BIT_OFF_T
327 All new 32-bit architectures should have 64-bit off_t type on
328 userspace side which corresponds to the loff_t kernel type. This
329 is the requirement for modern ABIs. Some existing architectures
330 still support 32-bit off_t. This option is enabled for all such
331 architectures explicitly.
333 # Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat
334 config ARCH_32BIT_USTAT_F_TINODE
337 config HAVE_ASM_MODVERSIONS
340 This symbol should be selected by an architecture if it provides
341 <asm/asm-prototypes.h> to support the module versioning for symbols
342 exported from assembly code.
344 config HAVE_REGS_AND_STACK_ACCESS_API
347 This symbol should be selected by an architecture if it supports
348 the API needed to access registers and stack entries from pt_regs,
349 declared in asm/ptrace.h
350 For example the kprobes-based event tracer needs this API.
354 depends on HAVE_REGS_AND_STACK_ACCESS_API
356 This symbol should be selected by an architecture if it
357 supports an implementation of restartable sequences.
359 config HAVE_FUNCTION_ARG_ACCESS_API
362 This symbol should be selected by an architecture if it supports
363 the API needed to access function arguments from pt_regs,
364 declared in asm/ptrace.h
366 config HAVE_HW_BREAKPOINT
368 depends on PERF_EVENTS
370 config HAVE_MIXED_BREAKPOINTS_REGS
372 depends on HAVE_HW_BREAKPOINT
374 Depending on the arch implementation of hardware breakpoints,
375 some of them have separate registers for data and instruction
376 breakpoints addresses, others have mixed registers to store
377 them but define the access type in a control register.
378 Select this option if your arch implements breakpoints under the
381 config HAVE_USER_RETURN_NOTIFIER
384 config HAVE_PERF_EVENTS_NMI
387 System hardware can generate an NMI using the perf event
388 subsystem. Also has support for calculating CPU cycle events
389 to determine how many clock cycles in a given period.
391 config HAVE_HARDLOCKUP_DETECTOR_PERF
393 depends on HAVE_PERF_EVENTS_NMI
395 The arch chooses to use the generic perf-NMI-based hardlockup
396 detector. Must define HAVE_PERF_EVENTS_NMI.
398 config HAVE_NMI_WATCHDOG
402 The arch provides a low level NMI watchdog. It provides
403 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
405 config HAVE_HARDLOCKUP_DETECTOR_ARCH
407 select HAVE_NMI_WATCHDOG
409 The arch chooses to provide its own hardlockup detector, which is
410 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
411 interfaces and parameters provided by hardlockup detector subsystem.
413 config HAVE_PERF_REGS
416 Support selective register dumps for perf events. This includes
417 bit-mapping of each registers and a unique architecture id.
419 config HAVE_PERF_USER_STACK_DUMP
422 Support user stack dumps for perf event samples. This needs
423 access to the user stack pointer which is not unified across
426 config HAVE_ARCH_JUMP_LABEL
429 config HAVE_ARCH_JUMP_LABEL_RELATIVE
432 config MMU_GATHER_TABLE_FREE
435 config MMU_GATHER_RCU_TABLE_FREE
437 select MMU_GATHER_TABLE_FREE
439 config MMU_GATHER_PAGE_SIZE
442 config MMU_GATHER_NO_RANGE
445 config MMU_GATHER_NO_GATHER
447 depends on MMU_GATHER_TABLE_FREE
449 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
452 Temporary select until all architectures can be converted to have
453 irqs disabled over activate_mm. Architectures that do IPI based TLB
454 shootdowns should enable this.
456 config ARCH_HAVE_NMI_SAFE_CMPXCHG
459 config HAVE_ALIGNED_STRUCT_PAGE
462 This makes sure that struct pages are double word aligned and that
463 e.g. the SLUB allocator can perform double word atomic operations
464 on a struct page for better performance. However selecting this
465 might increase the size of a struct page by a word.
467 config HAVE_CMPXCHG_LOCAL
470 config HAVE_CMPXCHG_DOUBLE
473 config ARCH_WEAK_RELEASE_ACQUIRE
476 config ARCH_WANT_IPC_PARSE_VERSION
479 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
482 config ARCH_WANT_OLD_COMPAT_IPC
483 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
486 config HAVE_ARCH_SECCOMP
489 An arch should select this symbol to support seccomp mode 1 (the fixed
490 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
491 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
492 - __NR_seccomp_read_32
493 - __NR_seccomp_write_32
494 - __NR_seccomp_exit_32
495 - __NR_seccomp_sigreturn_32
497 config HAVE_ARCH_SECCOMP_FILTER
499 select HAVE_ARCH_SECCOMP
501 An arch should select this symbol if it provides all of these things:
502 - all the requirements for HAVE_ARCH_SECCOMP
504 - syscall_get_arguments()
506 - syscall_set_return_value()
507 - SIGSYS siginfo_t support
508 - secure_computing is called from a ptrace_event()-safe context
509 - secure_computing return value is checked and a return value of -1
510 results in the system call being skipped immediately.
511 - seccomp syscall wired up
512 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
513 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
514 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
517 prompt "Enable seccomp to safely execute untrusted bytecode"
519 depends on HAVE_ARCH_SECCOMP
521 This kernel feature is useful for number crunching applications
522 that may need to handle untrusted bytecode during their
523 execution. By using pipes or other transports made available
524 to the process as file descriptors supporting the read/write
525 syscalls, it's possible to isolate those applications in their
526 own address space using seccomp. Once seccomp is enabled via
527 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
528 disabled and the task is only allowed to execute a few safe
529 syscalls defined by each seccomp mode.
533 config SECCOMP_FILTER
535 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
537 Enable tasks to build secure computing environments defined
538 in terms of Berkeley Packet Filter programs which implement
539 task-defined system call filtering polices.
541 See Documentation/userspace-api/seccomp_filter.rst for details.
543 config SECCOMP_CACHE_DEBUG
544 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
545 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
548 This enables the /proc/pid/seccomp_cache interface to monitor
549 seccomp cache data. The file format is subject to change. Reading
550 the file requires CAP_SYS_ADMIN.
552 This option is for debugging only. Enabling presents the risk that
553 an adversary may be able to infer the seccomp filter logic.
557 config HAVE_ARCH_STACKLEAK
560 An architecture should select this if it has the code which
561 fills the used part of the kernel stack with the STACKLEAK_POISON
562 value before returning from system calls.
564 config HAVE_STACKPROTECTOR
567 An arch should select this symbol if:
568 - it has implemented a stack canary (e.g. __stack_chk_guard)
570 config STACKPROTECTOR
571 bool "Stack Protector buffer overflow detection"
572 depends on HAVE_STACKPROTECTOR
573 depends on $(cc-option,-fstack-protector)
576 This option turns on the "stack-protector" GCC feature. This
577 feature puts, at the beginning of functions, a canary value on
578 the stack just before the return address, and validates
579 the value just before actually returning. Stack based buffer
580 overflows (that need to overwrite this return address) now also
581 overwrite the canary, which gets detected and the attack is then
582 neutralized via a kernel panic.
584 Functions will have the stack-protector canary logic added if they
585 have an 8-byte or larger character array on the stack.
587 This feature requires gcc version 4.2 or above, or a distribution
588 gcc with the feature backported ("-fstack-protector").
590 On an x86 "defconfig" build, this feature adds canary checks to
591 about 3% of all kernel functions, which increases kernel code size
594 config STACKPROTECTOR_STRONG
595 bool "Strong Stack Protector"
596 depends on STACKPROTECTOR
597 depends on $(cc-option,-fstack-protector-strong)
600 Functions will have the stack-protector canary logic added in any
601 of the following conditions:
603 - local variable's address used as part of the right hand side of an
604 assignment or function argument
605 - local variable is an array (or union containing an array),
606 regardless of array type or length
607 - uses register local variables
609 This feature requires gcc version 4.9 or above, or a distribution
610 gcc with the feature backported ("-fstack-protector-strong").
612 On an x86 "defconfig" build, this feature adds canary checks to
613 about 20% of all kernel functions, which increases the kernel code
616 config ARCH_SUPPORTS_SHADOW_CALL_STACK
619 An architecture should select this if it supports the compiler's
620 Shadow Call Stack and implements runtime support for shadow stack
623 config SHADOW_CALL_STACK
624 bool "Shadow Call Stack"
625 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK
626 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
628 This option enables the compiler's Shadow Call Stack, which
629 uses a shadow stack to protect function return addresses from
630 being overwritten by an attacker. More information can be found
631 in the compiler's documentation:
633 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html
634 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options
636 Note that security guarantees in the kernel differ from the
637 ones documented for user space. The kernel must store addresses
638 of shadow stacks in memory, which means an attacker capable of
639 reading and writing arbitrary memory may be able to locate them
640 and hijack control flow by modifying the stacks.
645 Selected if the kernel will be built using the compiler's LTO feature.
651 Selected if the kernel will be built using Clang's LTO feature.
653 config ARCH_SUPPORTS_LTO_CLANG
656 An architecture should select this option if it supports:
657 - compiling with Clang,
658 - compiling inline assembly with Clang's integrated assembler,
659 - and linking with LLD.
661 config ARCH_SUPPORTS_LTO_CLANG_THIN
664 An architecture should select this option if it can support Clang's
669 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM
670 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm)
671 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm)
672 depends on ARCH_SUPPORTS_LTO_CLANG
673 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT
674 depends on !KASAN || KASAN_HW_TAGS
675 depends on !GCOV_KERNEL
677 The compiler and Kconfig options support building with Clang's
681 prompt "Link Time Optimization (LTO)"
684 This option enables Link Time Optimization (LTO), which allows the
685 compiler to optimize binaries globally.
687 If unsure, select LTO_NONE. Note that LTO is very resource-intensive
688 so it's disabled by default.
693 Build the kernel normally, without Link Time Optimization (LTO).
695 config LTO_CLANG_FULL
696 bool "Clang Full LTO (EXPERIMENTAL)"
697 depends on HAS_LTO_CLANG
698 depends on !COMPILE_TEST
701 This option enables Clang's full Link Time Optimization (LTO), which
702 allows the compiler to optimize the kernel globally. If you enable
703 this option, the compiler generates LLVM bitcode instead of ELF
704 object files, and the actual compilation from bitcode happens at
705 the LTO link step, which may take several minutes depending on the
706 kernel configuration. More information can be found from LLVM's
709 https://llvm.org/docs/LinkTimeOptimization.html
711 During link time, this option can use a large amount of RAM, and
712 may take much longer than the ThinLTO option.
714 config LTO_CLANG_THIN
715 bool "Clang ThinLTO (EXPERIMENTAL)"
716 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN
719 This option enables Clang's ThinLTO, which allows for parallel
720 optimization and faster incremental compiles compared to the
721 CONFIG_LTO_CLANG_FULL option. More information can be found
722 from Clang's documentation:
724 https://clang.llvm.org/docs/ThinLTO.html
729 config ARCH_SUPPORTS_CFI_CLANG
732 An architecture should select this option if it can support Clang's
733 Control-Flow Integrity (CFI) checking.
736 bool "Use Clang's Control Flow Integrity (CFI)"
737 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG
738 depends on CLANG_VERSION >= 140000
741 This option enables Clang’s forward-edge Control Flow Integrity
742 (CFI) checking, where the compiler injects a runtime check to each
743 indirect function call to ensure the target is a valid function with
744 the correct static type. This restricts possible call targets and
745 makes it more difficult for an attacker to exploit bugs that allow
746 the modification of stored function pointers. More information can be
747 found from Clang's documentation:
749 https://clang.llvm.org/docs/ControlFlowIntegrity.html
751 config CFI_CLANG_SHADOW
752 bool "Use CFI shadow to speed up cross-module checks"
754 depends on CFI_CLANG && MODULES
756 If you select this option, the kernel builds a fast look-up table of
757 CFI check functions in loaded modules to reduce performance overhead.
761 config CFI_PERMISSIVE
762 bool "Use CFI in permissive mode"
765 When selected, Control Flow Integrity (CFI) violations result in a
766 warning instead of a kernel panic. This option should only be used
767 for finding indirect call type mismatches during development.
771 config HAVE_ARCH_WITHIN_STACK_FRAMES
774 An architecture should select this if it can walk the kernel stack
775 frames to determine if an object is part of either the arguments
776 or local variables (i.e. that it excludes saved return addresses,
777 and similar) by implementing an inline arch_within_stack_frames(),
778 which is used by CONFIG_HARDENED_USERCOPY.
780 config HAVE_CONTEXT_TRACKING
783 Provide kernel/user boundaries probes necessary for subsystems
784 that need it, such as userspace RCU extended quiescent state.
785 Syscalls need to be wrapped inside user_exit()-user_enter(), either
786 optimized behind static key or through the slow path using TIF_NOHZ
787 flag. Exceptions handlers must be wrapped as well. Irqs are already
788 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
789 handling on irq exit still need to be protected.
791 config HAVE_CONTEXT_TRACKING_OFFSTACK
794 Architecture neither relies on exception_enter()/exception_exit()
795 nor on schedule_user(). Also preempt_schedule_notrace() and
796 preempt_schedule_irq() can't be called in a preemptible section
797 while context tracking is CONTEXT_USER. This feature reflects a sane
798 entry implementation where the following requirements are met on
799 critical entry code, ie: before user_exit() or after user_enter():
801 - Critical entry code isn't preemptible (or better yet:
803 - No use of RCU read side critical sections, unless rcu_nmi_enter()
805 - No use of instrumentation, unless instrumentation_begin() got
811 Arch relies on TIF_NOHZ and syscall slow path to implement context
812 tracking calls to user_enter()/user_exit().
814 config HAVE_VIRT_CPU_ACCOUNTING
817 config HAVE_VIRT_CPU_ACCOUNTING_IDLE
820 Architecture has its own way to account idle CPU time and therefore
821 doesn't implement vtime_account_idle().
823 config ARCH_HAS_SCALED_CPUTIME
826 config HAVE_VIRT_CPU_ACCOUNTING_GEN
830 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
831 Before enabling this option, arch code must be audited
832 to ensure there are no races in concurrent read/write of
833 cputime_t. For example, reading/writing 64-bit cputime_t on
834 some 32-bit arches may require multiple accesses, so proper
835 locking is needed to protect against concurrent accesses.
837 config HAVE_IRQ_TIME_ACCOUNTING
840 Archs need to ensure they use a high enough resolution clock to
841 support irq time accounting and then call enable_sched_clock_irqtime().
846 Architectures that select this are able to move page tables at the
847 PUD level. If there are only 3 page table levels, the move effectively
848 happens at the PGD level.
853 Archs that select this are able to move page tables at the PMD level.
855 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
858 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
861 config HAVE_ARCH_HUGE_VMAP
865 # Archs that select this would be capable of PMD-sized vmaps (i.e.,
866 # arch_vmap_pmd_supported() returns true). The VM_ALLOW_HUGE_VMAP flag
867 # must be used to enable allocations to use hugepages.
869 config HAVE_ARCH_HUGE_VMALLOC
870 depends on HAVE_ARCH_HUGE_VMAP
873 config ARCH_WANT_HUGE_PMD_SHARE
876 config HAVE_ARCH_SOFT_DIRTY
879 config HAVE_MOD_ARCH_SPECIFIC
882 The arch uses struct mod_arch_specific to store data. Many arches
883 just need a simple module loader without arch specific data - those
884 should not enable this.
886 config MODULES_USE_ELF_RELA
889 Modules only use ELF RELA relocations. Modules with ELF REL
890 relocations will give an error.
892 config MODULES_USE_ELF_REL
895 Modules only use ELF REL relocations. Modules with ELF RELA
896 relocations will give an error.
898 config ARCH_WANTS_MODULES_DATA_IN_VMALLOC
901 For architectures like powerpc/32 which have constraints on module
902 allocation and need to allocate module data outside of module area.
904 config HAVE_IRQ_EXIT_ON_IRQ_STACK
907 Architecture doesn't only execute the irq handler on the irq stack
908 but also irq_exit(). This way we can process softirqs on this irq
909 stack instead of switching to a new one when we call __do_softirq()
910 in the end of an hardirq.
911 This spares a stack switch and improves cache usage on softirq
914 config HAVE_SOFTIRQ_ON_OWN_STACK
917 Architecture provides a function to run __do_softirq() on a
920 config ALTERNATE_USER_ADDRESS_SPACE
923 Architectures set this when the CPU uses separate address
924 spaces for kernel and user space pointers. In this case, the
925 access_ok() check on a __user pointer is skipped.
927 config PGTABLE_LEVELS
931 config ARCH_HAS_ELF_RANDOMIZE
934 An architecture supports choosing randomized locations for
935 stack, mmap, brk, and ET_DYN. Defined functions:
937 - arch_randomize_brk()
939 config HAVE_ARCH_MMAP_RND_BITS
942 An arch should select this symbol if it supports setting a variable
943 number of bits for use in establishing the base address for mmap
944 allocations, has MMU enabled and provides values for both:
945 - ARCH_MMAP_RND_BITS_MIN
946 - ARCH_MMAP_RND_BITS_MAX
948 config HAVE_EXIT_THREAD
951 An architecture implements exit_thread.
953 config ARCH_MMAP_RND_BITS_MIN
956 config ARCH_MMAP_RND_BITS_MAX
959 config ARCH_MMAP_RND_BITS_DEFAULT
962 config ARCH_MMAP_RND_BITS
963 int "Number of bits to use for ASLR of mmap base address" if EXPERT
964 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
965 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
966 default ARCH_MMAP_RND_BITS_MIN
967 depends on HAVE_ARCH_MMAP_RND_BITS
969 This value can be used to select the number of bits to use to
970 determine the random offset to the base address of vma regions
971 resulting from mmap allocations. This value will be bounded
972 by the architecture's minimum and maximum supported values.
974 This value can be changed after boot using the
975 /proc/sys/vm/mmap_rnd_bits tunable
977 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
980 An arch should select this symbol if it supports running applications
981 in compatibility mode, supports setting a variable number of bits for
982 use in establishing the base address for mmap allocations, has MMU
983 enabled and provides values for both:
984 - ARCH_MMAP_RND_COMPAT_BITS_MIN
985 - ARCH_MMAP_RND_COMPAT_BITS_MAX
987 config ARCH_MMAP_RND_COMPAT_BITS_MIN
990 config ARCH_MMAP_RND_COMPAT_BITS_MAX
993 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
996 config ARCH_MMAP_RND_COMPAT_BITS
997 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
998 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
999 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
1000 default ARCH_MMAP_RND_COMPAT_BITS_MIN
1001 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
1003 This value can be used to select the number of bits to use to
1004 determine the random offset to the base address of vma regions
1005 resulting from mmap allocations for compatible applications This
1006 value will be bounded by the architecture's minimum and maximum
1009 This value can be changed after boot using the
1010 /proc/sys/vm/mmap_rnd_compat_bits tunable
1012 config HAVE_ARCH_COMPAT_MMAP_BASES
1015 This allows 64bit applications to invoke 32-bit mmap() syscall
1016 and vice-versa 32-bit applications to call 64-bit mmap().
1017 Required for applications doing different bitness syscalls.
1019 config PAGE_SIZE_LESS_THAN_64KB
1021 depends on !ARM64_64K_PAGES
1022 depends on !IA64_PAGE_SIZE_64KB
1023 depends on !PAGE_SIZE_64KB
1024 depends on !PARISC_PAGE_SIZE_64KB
1025 depends on PAGE_SIZE_LESS_THAN_256KB
1027 config PAGE_SIZE_LESS_THAN_256KB
1029 depends on !PAGE_SIZE_256KB
1031 # This allows to use a set of generic functions to determine mmap base
1032 # address by giving priority to top-down scheme only if the process
1033 # is not in legacy mode (compat task, unlimited stack size or
1034 # sysctl_legacy_va_layout).
1035 # Architecture that selects this option can provide its own version of:
1037 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
1040 select ARCH_HAS_ELF_RANDOMIZE
1045 config HAVE_JUMP_LABEL_HACK
1048 config HAVE_NOINSTR_HACK
1051 config HAVE_NOINSTR_VALIDATION
1054 config HAVE_UACCESS_VALIDATION
1058 config HAVE_STACK_VALIDATION
1061 Architecture supports objtool compile-time frame pointer rule
1064 config HAVE_RELIABLE_STACKTRACE
1067 Architecture has either save_stack_trace_tsk_reliable() or
1068 arch_stack_walk_reliable() function which only returns a stack trace
1069 if it can guarantee the trace is reliable.
1071 config HAVE_ARCH_HASH
1075 If this is set, the architecture provides an <asm/hash.h>
1076 file which provides platform-specific implementations of some
1077 functions in <linux/hash.h> or fs/namei.c.
1079 config HAVE_ARCH_NVRAM_OPS
1088 config CLONE_BACKWARDS
1091 Architecture has tls passed as the 4th argument of clone(2),
1094 config CLONE_BACKWARDS2
1097 Architecture has the first two arguments of clone(2) swapped.
1099 config CLONE_BACKWARDS3
1102 Architecture has tls passed as the 3rd argument of clone(2),
1105 config ODD_RT_SIGACTION
1108 Architecture has unusual rt_sigaction(2) arguments
1110 config OLD_SIGSUSPEND
1113 Architecture has old sigsuspend(2) syscall, of one-argument variety
1115 config OLD_SIGSUSPEND3
1118 Even weirder antique ABI - three-argument sigsuspend(2)
1120 config OLD_SIGACTION
1123 Architecture has old sigaction(2) syscall. Nope, not the same
1124 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
1125 but fairly different variant of sigaction(2), thanks to OSF/1
1128 config COMPAT_OLD_SIGACTION
1131 config COMPAT_32BIT_TIME
1132 bool "Provide system calls for 32-bit time_t"
1133 default !64BIT || COMPAT
1135 This enables 32 bit time_t support in addition to 64 bit time_t support.
1136 This is relevant on all 32-bit architectures, and 64-bit architectures
1137 as part of compat syscall handling.
1139 config ARCH_NO_PREEMPT
1142 config ARCH_EPHEMERAL_INODES
1145 An arch should select this symbol if it doesn't keep track of inode
1146 instances on its own, but instead relies on something else (e.g. the
1147 host kernel for an UML kernel).
1149 config ARCH_SUPPORTS_RT
1152 config CPU_NO_EFFICIENT_FFS
1155 config HAVE_ARCH_VMAP_STACK
1158 An arch should select this symbol if it can support kernel stacks
1159 in vmalloc space. This means:
1161 - vmalloc space must be large enough to hold many kernel stacks.
1162 This may rule out many 32-bit architectures.
1164 - Stacks in vmalloc space need to work reliably. For example, if
1165 vmap page tables are created on demand, either this mechanism
1166 needs to work while the stack points to a virtual address with
1167 unpopulated page tables or arch code (switch_to() and switch_mm(),
1168 most likely) needs to ensure that the stack's page table entries
1169 are populated before running on a possibly unpopulated stack.
1171 - If the stack overflows into a guard page, something reasonable
1172 should happen. The definition of "reasonable" is flexible, but
1173 instantly rebooting without logging anything would be unfriendly.
1177 bool "Use a virtually-mapped stack"
1178 depends on HAVE_ARCH_VMAP_STACK
1179 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC
1181 Enable this if you want the use virtually-mapped kernel stacks
1182 with guard pages. This causes kernel stack overflows to be
1183 caught immediately rather than causing difficult-to-diagnose
1186 To use this with software KASAN modes, the architecture must support
1187 backing virtual mappings with real shadow memory, and KASAN_VMALLOC
1190 config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1193 An arch should select this symbol if it can support kernel stack
1194 offset randomization with calls to add_random_kstack_offset()
1195 during syscall entry and choose_random_kstack_offset() during
1196 syscall exit. Careful removal of -fstack-protector-strong and
1197 -fstack-protector should also be applied to the entry code and
1198 closely examined, as the artificial stack bump looks like an array
1199 to the compiler, so it will attempt to add canary checks regardless
1200 of the static branch state.
1202 config RANDOMIZE_KSTACK_OFFSET
1203 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT
1205 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1206 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
1208 The kernel stack offset can be randomized (after pt_regs) by
1209 roughly 5 bits of entropy, frustrating memory corruption
1210 attacks that depend on stack address determinism or
1211 cross-syscall address exposures.
1213 The feature is controlled via the "randomize_kstack_offset=on/off"
1214 kernel boot param, and if turned off has zero overhead due to its use
1215 of static branches (see JUMP_LABEL).
1219 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
1220 bool "Default state of kernel stack offset randomization"
1221 depends on RANDOMIZE_KSTACK_OFFSET
1223 Kernel stack offset randomization is controlled by kernel boot param
1224 "randomize_kstack_offset=on/off", and this config chooses the default
1227 config ARCH_OPTIONAL_KERNEL_RWX
1230 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1233 config ARCH_HAS_STRICT_KERNEL_RWX
1236 config STRICT_KERNEL_RWX
1237 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
1238 depends on ARCH_HAS_STRICT_KERNEL_RWX
1239 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1241 If this is set, kernel text and rodata memory will be made read-only,
1242 and non-text memory will be made non-executable. This provides
1243 protection against certain security exploits (e.g. executing the heap
1246 These features are considered standard security practice these days.
1247 You should say Y here in almost all cases.
1249 config ARCH_HAS_STRICT_MODULE_RWX
1252 config STRICT_MODULE_RWX
1253 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1254 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1255 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1257 If this is set, module text and rodata memory will be made read-only,
1258 and non-text memory will be made non-executable. This provides
1259 protection against certain security exploits (e.g. writing to text)
1261 # select if the architecture provides an asm/dma-direct.h header
1262 config ARCH_HAS_PHYS_TO_DMA
1265 config HAVE_ARCH_COMPILER_H
1268 An architecture can select this if it provides an
1269 asm/compiler.h header that should be included after
1270 linux/compiler-*.h in order to override macro definitions that those
1271 headers generally provide.
1273 config HAVE_ARCH_PREL32_RELOCATIONS
1276 May be selected by an architecture if it supports place-relative
1277 32-bit relocations, both in the toolchain and in the module loader,
1278 in which case relative references can be used in special sections
1279 for PCI fixup, initcalls etc which are only half the size on 64 bit
1280 architectures, and don't require runtime relocation on relocatable
1283 config ARCH_USE_MEMREMAP_PROT
1286 config LOCK_EVENT_COUNTS
1287 bool "Locking event counts collection"
1290 Enable light-weight counting of various locking related events
1291 in the system with minimal performance impact. This reduces
1292 the chance of application behavior change because of timing
1293 differences. The counts are reported via debugfs.
1295 # Select if the architecture has support for applying RELR relocations.
1296 config ARCH_HAS_RELR
1300 bool "Use RELR relocation packing"
1301 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1304 Store the kernel's dynamic relocations in the RELR relocation packing
1305 format. Requires a compatible linker (LLD supports this feature), as
1306 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1309 config ARCH_HAS_MEM_ENCRYPT
1312 config ARCH_HAS_CC_PLATFORM
1315 config HAVE_SPARSE_SYSCALL_NR
1318 An architecture should select this if its syscall numbering is sparse
1319 to save space. For example, MIPS architecture has a syscall array with
1320 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1321 related optimizations for a given architecture.
1323 config ARCH_HAS_VDSO_DATA
1326 config HAVE_STATIC_CALL
1329 config HAVE_STATIC_CALL_INLINE
1331 depends on HAVE_STATIC_CALL
1334 config HAVE_PREEMPT_DYNAMIC
1337 config HAVE_PREEMPT_DYNAMIC_CALL
1339 depends on HAVE_STATIC_CALL
1340 select HAVE_PREEMPT_DYNAMIC
1342 An architecture should select this if it can handle the preemption
1343 model being selected at boot time using static calls.
1345 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a
1346 preemption function will be patched directly.
1348 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any
1349 call to a preemption function will go through a trampoline, and the
1350 trampoline will be patched.
1352 It is strongly advised to support inline static call to avoid any
1355 config HAVE_PREEMPT_DYNAMIC_KEY
1357 depends on HAVE_ARCH_JUMP_LABEL && CC_HAS_ASM_GOTO
1358 select HAVE_PREEMPT_DYNAMIC
1360 An architecture should select this if it can handle the preemption
1361 model being selected at boot time using static keys.
1363 Each preemption function will be given an early return based on a
1364 static key. This should have slightly lower overhead than non-inline
1365 static calls, as this effectively inlines each trampoline into the
1366 start of its callee. This may avoid redundant work, and may
1367 integrate better with CFI schemes.
1369 This will have greater overhead than using inline static calls as
1370 the call to the preemption function cannot be entirely elided.
1372 config ARCH_WANT_LD_ORPHAN_WARN
1375 An arch should select this symbol once all linker sections are explicitly
1376 included, size-asserted, or discarded in the linker scripts. This is
1377 important because we never want expected sections to be placed heuristically
1378 by the linker, since the locations of such sections can change between linker
1381 config HAVE_ARCH_PFN_VALID
1384 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
1387 config ARCH_SUPPORTS_PAGE_TABLE_CHECK
1390 config ARCH_SPLIT_ARG64
1393 If a 32-bit architecture requires 64-bit arguments to be split into
1394 pairs of 32-bit arguments, select this option.
1396 config ARCH_HAS_ELFCORE_COMPAT
1399 config ARCH_HAS_PARANOID_L1D_FLUSH
1402 config DYNAMIC_SIGFRAME
1405 # Select, if arch has a named attribute group bound to NUMA device nodes.
1406 config HAVE_ARCH_NODE_DEV_GROUP
1409 source "kernel/gcov/Kconfig"
1411 source "scripts/gcc-plugins/Kconfig"