1 # SPDX-License-Identifier: GPL-2.0
3 # General architecture dependent options
7 # Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can
8 # override the default values in this file.
10 source "arch/$(SRCARCH)/Kconfig"
12 config ARCH_CONFIGURES_CPU_MITIGATIONS
15 if !ARCH_CONFIGURES_CPU_MITIGATIONS
16 config CPU_MITIGATIONS
21 # Selected by architectures that need custom DMA operations for e.g. legacy
22 # IOMMUs not handled by dma-iommu. Drivers must never select this symbol.
24 config ARCH_HAS_DMA_OPS
26 select DMA_OPS_HELPERS
29 menu "General architecture-dependent options"
31 config ARCH_HAS_SUBPAGE_FAULTS
34 Select if the architecture can check permissions at sub-page
35 granularity (e.g. arm64 MTE). The probe_user_*() functions
41 config SMT_NUM_THREADS_DYNAMIC
44 # Selected by HOTPLUG_CORE_SYNC_DEAD or HOTPLUG_CORE_SYNC_FULL
45 config HOTPLUG_CORE_SYNC
48 # Basic CPU dead synchronization selected by architecture
49 config HOTPLUG_CORE_SYNC_DEAD
51 select HOTPLUG_CORE_SYNC
53 # Full CPU synchronization with alive state selected by architecture
54 config HOTPLUG_CORE_SYNC_FULL
56 select HOTPLUG_CORE_SYNC_DEAD if HOTPLUG_CPU
57 select HOTPLUG_CORE_SYNC
59 config HOTPLUG_SPLIT_STARTUP
61 select HOTPLUG_CORE_SYNC_FULL
63 config HOTPLUG_PARALLEL
65 select HOTPLUG_SPLIT_STARTUP
72 depends on HAVE_KPROBES
77 Kprobes allows you to trap at almost any kernel address and
78 execute a callback function. register_kprobe() establishes
79 a probepoint and specifies the callback. Kprobes is useful
80 for kernel debugging, non-intrusive instrumentation and testing.
84 bool "Optimize very unlikely/likely branches"
85 depends on HAVE_ARCH_JUMP_LABEL
86 select OBJTOOL if HAVE_JUMP_LABEL_HACK
88 This option enables a transparent branch optimization that
89 makes certain almost-always-true or almost-always-false branch
90 conditions even cheaper to execute within the kernel.
92 Certain performance-sensitive kernel code, such as trace points,
93 scheduler functionality, networking code and KVM have such
94 branches and include support for this optimization technique.
96 If it is detected that the compiler has support for "asm goto",
97 the kernel will compile such branches with just a nop
98 instruction. When the condition flag is toggled to true, the
99 nop will be converted to a jump instruction to execute the
100 conditional block of instructions.
102 This technique lowers overhead and stress on the branch prediction
103 of the processor and generally makes the kernel faster. The update
104 of the condition is slower, but those are always very rare.
106 ( On 32-bit x86, the necessary options added to the compiler
107 flags may increase the size of the kernel slightly. )
109 config STATIC_KEYS_SELFTEST
110 bool "Static key selftest"
111 depends on JUMP_LABEL
113 Boot time self-test of the branch patching code.
115 config STATIC_CALL_SELFTEST
116 bool "Static call selftest"
117 depends on HAVE_STATIC_CALL
119 Boot time self-test of the call patching code.
123 depends on KPROBES && HAVE_OPTPROBES
124 select NEED_TASKS_RCU
126 config KPROBES_ON_FTRACE
128 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
129 depends on DYNAMIC_FTRACE_WITH_REGS
131 If function tracer is enabled and the arch supports full
132 passing of pt_regs to function tracing, then kprobes can
133 optimize on top of function tracing.
137 depends on ARCH_SUPPORTS_UPROBES
138 select TASKS_TRACE_RCU
140 Uprobes is the user-space counterpart to kprobes: they
141 enable instrumentation applications (such as 'perf probe')
142 to establish unintrusive probes in user-space binaries and
143 libraries, by executing handler functions when the probes
144 are hit by user-space applications.
146 ( These probes come in the form of single-byte breakpoints,
147 managed by the kernel and kept transparent to the probed
150 config HAVE_64BIT_ALIGNED_ACCESS
151 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
153 Some architectures require 64 bit accesses to be 64 bit
154 aligned, which also requires structs containing 64 bit values
155 to be 64 bit aligned too. This includes some 32 bit
156 architectures which can do 64 bit accesses, as well as 64 bit
157 architectures without unaligned access.
159 This symbol should be selected by an architecture if 64 bit
160 accesses are required to be 64 bit aligned in this way even
161 though it is not a 64 bit architecture.
163 See Documentation/core-api/unaligned-memory-access.rst for
164 more information on the topic of unaligned memory accesses.
166 config HAVE_EFFICIENT_UNALIGNED_ACCESS
169 Some architectures are unable to perform unaligned accesses
170 without the use of get_unaligned/put_unaligned. Others are
171 unable to perform such accesses efficiently (e.g. trap on
172 unaligned access and require fixing it up in the exception
175 This symbol should be selected by an architecture if it can
176 perform unaligned accesses efficiently to allow different
177 code paths to be selected for these cases. Some network
178 drivers, for example, could opt to not fix up alignment
179 problems with received packets if doing so would not help
182 See Documentation/core-api/unaligned-memory-access.rst for more
183 information on the topic of unaligned memory accesses.
185 config ARCH_USE_BUILTIN_BSWAP
188 Modern versions of GCC (since 4.4) have builtin functions
189 for handling byte-swapping. Using these, instead of the old
190 inline assembler that the architecture code provides in the
191 __arch_bswapXX() macros, allows the compiler to see what's
192 happening and offers more opportunity for optimisation. In
193 particular, the compiler will be able to combine the byteswap
194 with a nearby load or store and use load-and-swap or
195 store-and-swap instructions if the architecture has them. It
196 should almost *never* result in code which is worse than the
197 hand-coded assembler in <asm/swab.h>. But just in case it
198 does, the use of the builtins is optional.
200 Any architecture with load-and-swap or store-and-swap
201 instructions should set this. And it shouldn't hurt to set it
202 on architectures that don't have such instructions.
206 depends on KPROBES && (HAVE_KRETPROBES || HAVE_RETHOOK)
208 config KRETPROBE_ON_RETHOOK
210 depends on HAVE_RETHOOK
211 depends on KRETPROBES
214 config USER_RETURN_NOTIFIER
216 depends on HAVE_USER_RETURN_NOTIFIER
218 Provide a kernel-internal notification when a cpu is about to
221 config HAVE_IOREMAP_PROT
227 config HAVE_KRETPROBES
230 config HAVE_OPTPROBES
233 config HAVE_KPROBES_ON_FTRACE
236 config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE
239 Since kretprobes modifies return address on the stack, the
240 stacktrace may see the kretprobe trampoline address instead
241 of correct one. If the architecture stacktrace code and
242 unwinder can adjust such entries, select this configuration.
244 config HAVE_FUNCTION_ERROR_INJECTION
250 config HAVE_FUNCTION_DESCRIPTORS
253 config TRACE_IRQFLAGS_SUPPORT
256 config TRACE_IRQFLAGS_NMI_SUPPORT
260 # An arch should select this if it provides all these things:
262 # task_pt_regs() in asm/processor.h or asm/ptrace.h
263 # arch_has_single_step() if there is hardware single-step support
264 # arch_has_block_step() if there is hardware block-step support
265 # asm/syscall.h supplying asm-generic/syscall.h interface
266 # linux/regset.h user_regset interfaces
267 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
268 # TIF_SYSCALL_TRACE calls ptrace_report_syscall_{entry,exit}
269 # TIF_NOTIFY_RESUME calls resume_user_mode_work()
271 config HAVE_ARCH_TRACEHOOK
274 config HAVE_DMA_CONTIGUOUS
277 config GENERIC_SMP_IDLE_THREAD
280 config GENERIC_IDLE_POLL_SETUP
283 config ARCH_HAS_FORTIFY_SOURCE
286 An architecture should select this when it can successfully
287 build and run with CONFIG_FORTIFY_SOURCE.
290 # Select if the arch provides a historic keepinit alias for the retain_initrd
291 # command line option
293 config ARCH_HAS_KEEPINITRD
296 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
297 config ARCH_HAS_SET_MEMORY
300 # Select if arch has all set_direct_map_invalid/default() functions
301 config ARCH_HAS_SET_DIRECT_MAP
305 # Select if the architecture provides the arch_dma_set_uncached symbol to
306 # either provide an uncached segment alias for a DMA allocation, or
307 # to remap the page tables in place.
309 config ARCH_HAS_DMA_SET_UNCACHED
313 # Select if the architectures provides the arch_dma_clear_uncached symbol
314 # to undo an in-place page table remap for uncached access.
316 config ARCH_HAS_DMA_CLEAR_UNCACHED
319 config ARCH_HAS_CPU_FINALIZE_INIT
322 # The architecture has a per-task state that includes the mm's PASID
323 config ARCH_HAS_CPU_PASID
327 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
330 An architecture should select this to provide hardened usercopy
331 knowledge about what region of the thread_struct should be
332 whitelisted for copying to userspace. Normally this is only the
333 FPU registers. Specifically, arch_thread_struct_whitelist()
334 should be implemented. Without this, the entire thread_struct
335 field in task_struct will be left whitelisted.
337 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
338 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
341 config ARCH_WANTS_NO_INSTR
344 An architecture should select this if the noinstr macro is being used on
345 functions to denote that the toolchain should avoid instrumenting such
346 functions and is required for correctness.
348 config ARCH_32BIT_OFF_T
352 All new 32-bit architectures should have 64-bit off_t type on
353 userspace side which corresponds to the loff_t kernel type. This
354 is the requirement for modern ABIs. Some existing architectures
355 still support 32-bit off_t. This option is enabled for all such
356 architectures explicitly.
358 # Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat
359 config ARCH_32BIT_USTAT_F_TINODE
362 config HAVE_ASM_MODVERSIONS
365 This symbol should be selected by an architecture if it provides
366 <asm/asm-prototypes.h> to support the module versioning for symbols
367 exported from assembly code.
369 config HAVE_REGS_AND_STACK_ACCESS_API
372 This symbol should be selected by an architecture if it supports
373 the API needed to access registers and stack entries from pt_regs,
374 declared in asm/ptrace.h
375 For example the kprobes-based event tracer needs this API.
379 depends on HAVE_REGS_AND_STACK_ACCESS_API
381 This symbol should be selected by an architecture if it
382 supports an implementation of restartable sequences.
387 This symbol should be selected by an architecture if it
390 config HAVE_FUNCTION_ARG_ACCESS_API
393 This symbol should be selected by an architecture if it supports
394 the API needed to access function arguments from pt_regs,
395 declared in asm/ptrace.h
397 config HAVE_HW_BREAKPOINT
399 depends on PERF_EVENTS
401 config HAVE_MIXED_BREAKPOINTS_REGS
403 depends on HAVE_HW_BREAKPOINT
405 Depending on the arch implementation of hardware breakpoints,
406 some of them have separate registers for data and instruction
407 breakpoints addresses, others have mixed registers to store
408 them but define the access type in a control register.
409 Select this option if your arch implements breakpoints under the
412 config HAVE_USER_RETURN_NOTIFIER
415 config HAVE_PERF_EVENTS_NMI
418 System hardware can generate an NMI using the perf event
419 subsystem. Also has support for calculating CPU cycle events
420 to determine how many clock cycles in a given period.
422 config HAVE_HARDLOCKUP_DETECTOR_PERF
424 depends on HAVE_PERF_EVENTS_NMI
426 The arch chooses to use the generic perf-NMI-based hardlockup
427 detector. Must define HAVE_PERF_EVENTS_NMI.
429 config HAVE_HARDLOCKUP_DETECTOR_ARCH
432 The arch provides its own hardlockup detector implementation instead
435 It uses the same command line parameters, and sysctl interface,
436 as the generic hardlockup detectors.
438 config HAVE_PERF_REGS
441 Support selective register dumps for perf events. This includes
442 bit-mapping of each registers and a unique architecture id.
444 config HAVE_PERF_USER_STACK_DUMP
447 Support user stack dumps for perf event samples. This needs
448 access to the user stack pointer which is not unified across
451 config HAVE_ARCH_JUMP_LABEL
454 config HAVE_ARCH_JUMP_LABEL_RELATIVE
457 config MMU_GATHER_TABLE_FREE
460 config MMU_GATHER_RCU_TABLE_FREE
462 select MMU_GATHER_TABLE_FREE
464 config MMU_GATHER_PAGE_SIZE
467 config MMU_GATHER_NO_RANGE
469 select MMU_GATHER_MERGE_VMAS
471 config MMU_GATHER_NO_FLUSH_CACHE
474 config MMU_GATHER_MERGE_VMAS
477 config MMU_GATHER_NO_GATHER
479 depends on MMU_GATHER_TABLE_FREE
481 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
484 Temporary select until all architectures can be converted to have
485 irqs disabled over activate_mm. Architectures that do IPI based TLB
486 shootdowns should enable this.
488 # Use normal mm refcounting for MMU_LAZY_TLB kernel thread references.
489 # MMU_LAZY_TLB_REFCOUNT=n can improve the scalability of context switching
490 # to/from kernel threads when the same mm is running on a lot of CPUs (a large
491 # multi-threaded application), by reducing contention on the mm refcount.
493 # This can be disabled if the architecture ensures no CPUs are using an mm as a
494 # "lazy tlb" beyond its final refcount (i.e., by the time __mmdrop frees the mm
495 # or its kernel page tables). This could be arranged by arch_exit_mmap(), or
496 # final exit(2) TLB flush, for example.
498 # To implement this, an arch *must*:
499 # Ensure the _lazy_tlb variants of mmgrab/mmdrop are used when manipulating
500 # the lazy tlb reference of a kthread's ->active_mm (non-arch code has been
501 # converted already).
502 config MMU_LAZY_TLB_REFCOUNT
504 depends on !MMU_LAZY_TLB_SHOOTDOWN
506 # This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an
507 # mm as a lazy tlb beyond its last reference count, by shooting down these
508 # users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may
509 # be using the mm as a lazy tlb, so that they may switch themselves to using
510 # init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs
511 # may be using mm as a lazy tlb mm.
513 # To implement this, an arch *must*:
514 # - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains
515 # at least all possible CPUs in which the mm is lazy.
516 # - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above).
517 config MMU_LAZY_TLB_SHOOTDOWN
520 config ARCH_HAVE_NMI_SAFE_CMPXCHG
523 config ARCH_HAVE_EXTRA_ELF_NOTES
526 An architecture should select this in order to enable adding an
527 arch-specific ELF note section to core files. It must provide two
528 functions: elf_coredump_extra_notes_size() and
529 elf_coredump_extra_notes_write() which are invoked by the ELF core
532 config ARCH_HAS_NMI_SAFE_THIS_CPU_OPS
535 config HAVE_ALIGNED_STRUCT_PAGE
538 This makes sure that struct pages are double word aligned and that
539 e.g. the SLUB allocator can perform double word atomic operations
540 on a struct page for better performance. However selecting this
541 might increase the size of a struct page by a word.
543 config HAVE_CMPXCHG_LOCAL
546 config HAVE_CMPXCHG_DOUBLE
549 config ARCH_WEAK_RELEASE_ACQUIRE
552 config ARCH_WANT_IPC_PARSE_VERSION
555 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
558 config ARCH_WANT_OLD_COMPAT_IPC
559 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
562 config HAVE_ARCH_SECCOMP
565 An arch should select this symbol to support seccomp mode 1 (the fixed
566 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
567 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
568 - __NR_seccomp_read_32
569 - __NR_seccomp_write_32
570 - __NR_seccomp_exit_32
571 - __NR_seccomp_sigreturn_32
573 config HAVE_ARCH_SECCOMP_FILTER
575 select HAVE_ARCH_SECCOMP
577 An arch should select this symbol if it provides all of these things:
578 - all the requirements for HAVE_ARCH_SECCOMP
580 - syscall_get_arguments()
582 - syscall_set_return_value()
583 - SIGSYS siginfo_t support
584 - secure_computing is called from a ptrace_event()-safe context
585 - secure_computing return value is checked and a return value of -1
586 results in the system call being skipped immediately.
587 - seccomp syscall wired up
588 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
589 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
590 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
593 prompt "Enable seccomp to safely execute untrusted bytecode"
595 depends on HAVE_ARCH_SECCOMP
597 This kernel feature is useful for number crunching applications
598 that may need to handle untrusted bytecode during their
599 execution. By using pipes or other transports made available
600 to the process as file descriptors supporting the read/write
601 syscalls, it's possible to isolate those applications in their
602 own address space using seccomp. Once seccomp is enabled via
603 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
604 disabled and the task is only allowed to execute a few safe
605 syscalls defined by each seccomp mode.
609 config SECCOMP_FILTER
611 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
613 Enable tasks to build secure computing environments defined
614 in terms of Berkeley Packet Filter programs which implement
615 task-defined system call filtering polices.
617 See Documentation/userspace-api/seccomp_filter.rst for details.
619 config SECCOMP_CACHE_DEBUG
620 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
621 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
624 This enables the /proc/pid/seccomp_cache interface to monitor
625 seccomp cache data. The file format is subject to change. Reading
626 the file requires CAP_SYS_ADMIN.
628 This option is for debugging only. Enabling presents the risk that
629 an adversary may be able to infer the seccomp filter logic.
633 config HAVE_ARCH_STACKLEAK
636 An architecture should select this if it has the code which
637 fills the used part of the kernel stack with the STACKLEAK_POISON
638 value before returning from system calls.
640 config HAVE_STACKPROTECTOR
643 An arch should select this symbol if:
644 - it has implemented a stack canary (e.g. __stack_chk_guard)
646 config STACKPROTECTOR
647 bool "Stack Protector buffer overflow detection"
648 depends on HAVE_STACKPROTECTOR
649 depends on $(cc-option,-fstack-protector)
652 This option turns on the "stack-protector" GCC feature. This
653 feature puts, at the beginning of functions, a canary value on
654 the stack just before the return address, and validates
655 the value just before actually returning. Stack based buffer
656 overflows (that need to overwrite this return address) now also
657 overwrite the canary, which gets detected and the attack is then
658 neutralized via a kernel panic.
660 Functions will have the stack-protector canary logic added if they
661 have an 8-byte or larger character array on the stack.
663 This feature requires gcc version 4.2 or above, or a distribution
664 gcc with the feature backported ("-fstack-protector").
666 On an x86 "defconfig" build, this feature adds canary checks to
667 about 3% of all kernel functions, which increases kernel code size
670 config STACKPROTECTOR_STRONG
671 bool "Strong Stack Protector"
672 depends on STACKPROTECTOR
673 depends on $(cc-option,-fstack-protector-strong)
676 Functions will have the stack-protector canary logic added in any
677 of the following conditions:
679 - local variable's address used as part of the right hand side of an
680 assignment or function argument
681 - local variable is an array (or union containing an array),
682 regardless of array type or length
683 - uses register local variables
685 This feature requires gcc version 4.9 or above, or a distribution
686 gcc with the feature backported ("-fstack-protector-strong").
688 On an x86 "defconfig" build, this feature adds canary checks to
689 about 20% of all kernel functions, which increases the kernel code
692 config ARCH_SUPPORTS_SHADOW_CALL_STACK
695 An architecture should select this if it supports the compiler's
696 Shadow Call Stack and implements runtime support for shadow stack
699 config SHADOW_CALL_STACK
700 bool "Shadow Call Stack"
701 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK
702 depends on DYNAMIC_FTRACE_WITH_ARGS || DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
705 This option enables the compiler's Shadow Call Stack, which
706 uses a shadow stack to protect function return addresses from
707 being overwritten by an attacker. More information can be found
708 in the compiler's documentation:
710 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html
711 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options
713 Note that security guarantees in the kernel differ from the
714 ones documented for user space. The kernel must store addresses
715 of shadow stacks in memory, which means an attacker capable of
716 reading and writing arbitrary memory may be able to locate them
717 and hijack control flow by modifying the stacks.
722 Set by the arch code if it relies on code patching to insert the
723 shadow call stack push and pop instructions rather than on the
729 Selected if the kernel will be built using the compiler's LTO feature.
735 Selected if the kernel will be built using Clang's LTO feature.
737 config ARCH_SUPPORTS_LTO_CLANG
740 An architecture should select this option if it supports:
741 - compiling with Clang,
742 - compiling inline assembly with Clang's integrated assembler,
743 - and linking with LLD.
745 config ARCH_SUPPORTS_LTO_CLANG_THIN
748 An architecture should select this option if it can support Clang's
753 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM
754 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm)
755 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm)
756 depends on ARCH_SUPPORTS_LTO_CLANG
757 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT
758 # https://github.com/ClangBuiltLinux/linux/issues/1721
759 depends on (!KASAN || KASAN_HW_TAGS || CLANG_VERSION >= 170000) || !DEBUG_INFO
760 depends on (!KCOV || CLANG_VERSION >= 170000) || !DEBUG_INFO
761 depends on !GCOV_KERNEL
763 The compiler and Kconfig options support building with Clang's
767 prompt "Link Time Optimization (LTO)"
770 This option enables Link Time Optimization (LTO), which allows the
771 compiler to optimize binaries globally.
773 If unsure, select LTO_NONE. Note that LTO is very resource-intensive
774 so it's disabled by default.
779 Build the kernel normally, without Link Time Optimization (LTO).
781 config LTO_CLANG_FULL
782 bool "Clang Full LTO (EXPERIMENTAL)"
783 depends on HAS_LTO_CLANG
784 depends on !COMPILE_TEST
787 This option enables Clang's full Link Time Optimization (LTO), which
788 allows the compiler to optimize the kernel globally. If you enable
789 this option, the compiler generates LLVM bitcode instead of ELF
790 object files, and the actual compilation from bitcode happens at
791 the LTO link step, which may take several minutes depending on the
792 kernel configuration. More information can be found from LLVM's
795 https://llvm.org/docs/LinkTimeOptimization.html
797 During link time, this option can use a large amount of RAM, and
798 may take much longer than the ThinLTO option.
800 config LTO_CLANG_THIN
801 bool "Clang ThinLTO (EXPERIMENTAL)"
802 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN
805 This option enables Clang's ThinLTO, which allows for parallel
806 optimization and faster incremental compiles compared to the
807 CONFIG_LTO_CLANG_FULL option. More information can be found
808 from Clang's documentation:
810 https://clang.llvm.org/docs/ThinLTO.html
815 config ARCH_SUPPORTS_CFI_CLANG
818 An architecture should select this option if it can support Clang's
819 Control-Flow Integrity (CFI) checking.
821 config ARCH_USES_CFI_TRAPS
825 bool "Use Clang's Control Flow Integrity (CFI)"
826 depends on ARCH_SUPPORTS_CFI_CLANG
827 depends on $(cc-option,-fsanitize=kcfi)
829 This option enables Clang's forward-edge Control Flow Integrity
830 (CFI) checking, where the compiler injects a runtime check to each
831 indirect function call to ensure the target is a valid function with
832 the correct static type. This restricts possible call targets and
833 makes it more difficult for an attacker to exploit bugs that allow
834 the modification of stored function pointers. More information can be
835 found from Clang's documentation:
837 https://clang.llvm.org/docs/ControlFlowIntegrity.html
839 config CFI_ICALL_NORMALIZE_INTEGERS
840 bool "Normalize CFI tags for integers"
842 depends on HAVE_CFI_ICALL_NORMALIZE_INTEGERS_CLANG
844 This option normalizes the CFI tags for integer types so that all
845 integer types of the same size and signedness receive the same CFI
848 The option is separate from CONFIG_RUST because it affects the ABI.
849 When working with build systems that care about the ABI, it is
850 convenient to be able to turn on this flag first, before Rust is
853 This option is necessary for using CFI with Rust. If unsure, say N.
855 config HAVE_CFI_ICALL_NORMALIZE_INTEGERS_CLANG
857 depends on $(cc-option,-fsanitize=kcfi -fsanitize-cfi-icall-experimental-normalize-integers)
858 # With GCOV/KASAN we need this fix: https://github.com/llvm/llvm-project/pull/104826
859 depends on CLANG_VERSION >= 190103 || (!GCOV_KERNEL && !KASAN_GENERIC && !KASAN_SW_TAGS)
861 config HAVE_CFI_ICALL_NORMALIZE_INTEGERS_RUSTC
863 depends on HAVE_CFI_ICALL_NORMALIZE_INTEGERS_CLANG
864 depends on RUSTC_VERSION >= 107900
865 # With GCOV/KASAN we need this fix: https://github.com/rust-lang/rust/pull/129373
866 depends on (RUSTC_LLVM_VERSION >= 190103 && RUSTC_VERSION >= 108200) || \
867 (!GCOV_KERNEL && !KASAN_GENERIC && !KASAN_SW_TAGS)
869 config CFI_PERMISSIVE
870 bool "Use CFI in permissive mode"
873 When selected, Control Flow Integrity (CFI) violations result in a
874 warning instead of a kernel panic. This option should only be used
875 for finding indirect call type mismatches during development.
879 config HAVE_ARCH_WITHIN_STACK_FRAMES
882 An architecture should select this if it can walk the kernel stack
883 frames to determine if an object is part of either the arguments
884 or local variables (i.e. that it excludes saved return addresses,
885 and similar) by implementing an inline arch_within_stack_frames(),
886 which is used by CONFIG_HARDENED_USERCOPY.
888 config HAVE_CONTEXT_TRACKING_USER
891 Provide kernel/user boundaries probes necessary for subsystems
892 that need it, such as userspace RCU extended quiescent state.
893 Syscalls need to be wrapped inside user_exit()-user_enter(), either
894 optimized behind static key or through the slow path using TIF_NOHZ
895 flag. Exceptions handlers must be wrapped as well. Irqs are already
896 protected inside ct_irq_enter/ct_irq_exit() but preemption or signal
897 handling on irq exit still need to be protected.
899 config HAVE_CONTEXT_TRACKING_USER_OFFSTACK
902 Architecture neither relies on exception_enter()/exception_exit()
903 nor on schedule_user(). Also preempt_schedule_notrace() and
904 preempt_schedule_irq() can't be called in a preemptible section
905 while context tracking is CT_STATE_USER. This feature reflects a sane
906 entry implementation where the following requirements are met on
907 critical entry code, ie: before user_exit() or after user_enter():
909 - Critical entry code isn't preemptible (or better yet:
911 - No use of RCU read side critical sections, unless ct_nmi_enter()
913 - No use of instrumentation, unless instrumentation_begin() got
919 Arch relies on TIF_NOHZ and syscall slow path to implement context
920 tracking calls to user_enter()/user_exit().
922 config HAVE_VIRT_CPU_ACCOUNTING
925 config HAVE_VIRT_CPU_ACCOUNTING_IDLE
928 Architecture has its own way to account idle CPU time and therefore
929 doesn't implement vtime_account_idle().
931 config ARCH_HAS_SCALED_CPUTIME
934 config HAVE_VIRT_CPU_ACCOUNTING_GEN
938 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
939 Before enabling this option, arch code must be audited
940 to ensure there are no races in concurrent read/write of
941 cputime_t. For example, reading/writing 64-bit cputime_t on
942 some 32-bit arches may require multiple accesses, so proper
943 locking is needed to protect against concurrent accesses.
945 config HAVE_IRQ_TIME_ACCOUNTING
948 Archs need to ensure they use a high enough resolution clock to
949 support irq time accounting and then call enable_sched_clock_irqtime().
954 Architectures that select this are able to move page tables at the
955 PUD level. If there are only 3 page table levels, the move effectively
956 happens at the PGD level.
961 Archs that select this are able to move page tables at the PMD level.
963 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
966 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
969 config HAVE_ARCH_HUGE_VMAP
973 # Archs that select this would be capable of PMD-sized vmaps (i.e.,
974 # arch_vmap_pmd_supported() returns true). The VM_ALLOW_HUGE_VMAP flag
975 # must be used to enable allocations to use hugepages.
977 config HAVE_ARCH_HUGE_VMALLOC
978 depends on HAVE_ARCH_HUGE_VMAP
981 config ARCH_WANT_HUGE_PMD_SHARE
984 # Archs that want to use pmd_mkwrite on kernel memory need it defined even
985 # if there are no userspace memory management features that use it
986 config ARCH_WANT_KERNEL_PMD_MKWRITE
989 config ARCH_WANT_PMD_MKWRITE
990 def_bool TRANSPARENT_HUGEPAGE || ARCH_WANT_KERNEL_PMD_MKWRITE
992 config HAVE_ARCH_SOFT_DIRTY
995 config HAVE_MOD_ARCH_SPECIFIC
998 The arch uses struct mod_arch_specific to store data. Many arches
999 just need a simple module loader without arch specific data - those
1000 should not enable this.
1002 config MODULES_USE_ELF_RELA
1005 Modules only use ELF RELA relocations. Modules with ELF REL
1006 relocations will give an error.
1008 config MODULES_USE_ELF_REL
1011 Modules only use ELF REL relocations. Modules with ELF RELA
1012 relocations will give an error.
1014 config ARCH_WANTS_MODULES_DATA_IN_VMALLOC
1017 For architectures like powerpc/32 which have constraints on module
1018 allocation and need to allocate module data outside of module area.
1020 config ARCH_WANTS_EXECMEM_LATE
1023 For architectures that do not allocate executable memory early on
1024 boot, but rather require its initialization late when there is
1025 enough entropy for module space randomization, for instance
1028 config ARCH_HAS_EXECMEM_ROX
1030 depends on MMU && !HIGHMEM
1032 For architectures that support allocations of executable memory
1033 with read-only execute permissions. Architecture must implement
1034 execmem_fill_trapping_insns() callback to enable this.
1036 config HAVE_IRQ_EXIT_ON_IRQ_STACK
1039 Architecture doesn't only execute the irq handler on the irq stack
1040 but also irq_exit(). This way we can process softirqs on this irq
1041 stack instead of switching to a new one when we call __do_softirq()
1042 in the end of an hardirq.
1043 This spares a stack switch and improves cache usage on softirq
1046 config HAVE_SOFTIRQ_ON_OWN_STACK
1049 Architecture provides a function to run __do_softirq() on a
1052 config SOFTIRQ_ON_OWN_STACK
1053 def_bool HAVE_SOFTIRQ_ON_OWN_STACK && !PREEMPT_RT
1055 config ALTERNATE_USER_ADDRESS_SPACE
1058 Architectures set this when the CPU uses separate address
1059 spaces for kernel and user space pointers. In this case, the
1060 access_ok() check on a __user pointer is skipped.
1062 config PGTABLE_LEVELS
1066 config ARCH_HAS_ELF_RANDOMIZE
1069 An architecture supports choosing randomized locations for
1070 stack, mmap, brk, and ET_DYN. Defined functions:
1072 - arch_randomize_brk()
1074 config HAVE_ARCH_MMAP_RND_BITS
1077 An arch should select this symbol if it supports setting a variable
1078 number of bits for use in establishing the base address for mmap
1079 allocations, has MMU enabled and provides values for both:
1080 - ARCH_MMAP_RND_BITS_MIN
1081 - ARCH_MMAP_RND_BITS_MAX
1083 config HAVE_EXIT_THREAD
1086 An architecture implements exit_thread.
1088 config ARCH_MMAP_RND_BITS_MIN
1091 config ARCH_MMAP_RND_BITS_MAX
1094 config ARCH_MMAP_RND_BITS_DEFAULT
1097 config ARCH_MMAP_RND_BITS
1098 int "Number of bits to use for ASLR of mmap base address" if EXPERT
1099 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
1100 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
1101 default ARCH_MMAP_RND_BITS_MIN
1102 depends on HAVE_ARCH_MMAP_RND_BITS
1104 This value can be used to select the number of bits to use to
1105 determine the random offset to the base address of vma regions
1106 resulting from mmap allocations. This value will be bounded
1107 by the architecture's minimum and maximum supported values.
1109 This value can be changed after boot using the
1110 /proc/sys/vm/mmap_rnd_bits tunable
1112 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
1115 An arch should select this symbol if it supports running applications
1116 in compatibility mode, supports setting a variable number of bits for
1117 use in establishing the base address for mmap allocations, has MMU
1118 enabled and provides values for both:
1119 - ARCH_MMAP_RND_COMPAT_BITS_MIN
1120 - ARCH_MMAP_RND_COMPAT_BITS_MAX
1122 config ARCH_MMAP_RND_COMPAT_BITS_MIN
1125 config ARCH_MMAP_RND_COMPAT_BITS_MAX
1128 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
1131 config ARCH_MMAP_RND_COMPAT_BITS
1132 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
1133 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
1134 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
1135 default ARCH_MMAP_RND_COMPAT_BITS_MIN
1136 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
1138 This value can be used to select the number of bits to use to
1139 determine the random offset to the base address of vma regions
1140 resulting from mmap allocations for compatible applications This
1141 value will be bounded by the architecture's minimum and maximum
1144 This value can be changed after boot using the
1145 /proc/sys/vm/mmap_rnd_compat_bits tunable
1147 config HAVE_ARCH_COMPAT_MMAP_BASES
1150 This allows 64bit applications to invoke 32-bit mmap() syscall
1151 and vice-versa 32-bit applications to call 64-bit mmap().
1152 Required for applications doing different bitness syscalls.
1154 config HAVE_PAGE_SIZE_4KB
1157 config HAVE_PAGE_SIZE_8KB
1160 config HAVE_PAGE_SIZE_16KB
1163 config HAVE_PAGE_SIZE_32KB
1166 config HAVE_PAGE_SIZE_64KB
1169 config HAVE_PAGE_SIZE_256KB
1173 prompt "MMU page size"
1175 config PAGE_SIZE_4KB
1177 depends on HAVE_PAGE_SIZE_4KB
1179 This option select the standard 4KiB Linux page size and the only
1180 available option on many architectures. Using 4KiB page size will
1181 minimize memory consumption and is therefore recommended for low
1183 Some software that is written for x86 systems makes incorrect
1184 assumptions about the page size and only runs on 4KiB pages.
1186 config PAGE_SIZE_8KB
1188 depends on HAVE_PAGE_SIZE_8KB
1190 This option is the only supported page size on a few older
1191 processors, and can be slightly faster than 4KiB pages.
1193 config PAGE_SIZE_16KB
1195 depends on HAVE_PAGE_SIZE_16KB
1197 This option is usually a good compromise between memory
1198 consumption and performance for typical desktop and server
1199 workloads, often saving a level of page table lookups compared
1200 to 4KB pages as well as reducing TLB pressure and overhead of
1201 per-page operations in the kernel at the expense of a larger
1204 config PAGE_SIZE_32KB
1206 depends on HAVE_PAGE_SIZE_32KB
1208 Using 32KiB page size will result in slightly higher performance
1209 kernel at the price of higher memory consumption compared to
1210 16KiB pages. This option is available only on cnMIPS cores.
1211 Note that you will need a suitable Linux distribution to
1214 config PAGE_SIZE_64KB
1216 depends on HAVE_PAGE_SIZE_64KB
1218 Using 64KiB page size will result in slightly higher performance
1219 kernel at the price of much higher memory consumption compared to
1220 4KiB or 16KiB pages.
1221 This is not suitable for general-purpose workloads but the
1222 better performance may be worth the cost for certain types of
1223 supercomputing or database applications that work mostly with
1224 large in-memory data rather than small files.
1226 config PAGE_SIZE_256KB
1228 depends on HAVE_PAGE_SIZE_256KB
1230 256KiB pages have little practical value due to their extreme
1231 memory usage. The kernel will only be able to run applications
1232 that have been compiled with '-zmax-page-size' set to 256KiB
1233 (the default is 64KiB or 4KiB on most architectures).
1237 config PAGE_SIZE_LESS_THAN_64KB
1239 depends on !PAGE_SIZE_64KB
1240 depends on PAGE_SIZE_LESS_THAN_256KB
1242 config PAGE_SIZE_LESS_THAN_256KB
1244 depends on !PAGE_SIZE_256KB
1248 default 12 if PAGE_SIZE_4KB
1249 default 13 if PAGE_SIZE_8KB
1250 default 14 if PAGE_SIZE_16KB
1251 default 15 if PAGE_SIZE_32KB
1252 default 16 if PAGE_SIZE_64KB
1253 default 18 if PAGE_SIZE_256KB
1255 # This allows to use a set of generic functions to determine mmap base
1256 # address by giving priority to top-down scheme only if the process
1257 # is not in legacy mode (compat task, unlimited stack size or
1258 # sysctl_legacy_va_layout).
1259 # Architecture that selects this option can provide its own version of:
1261 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
1264 select ARCH_HAS_ELF_RANDOMIZE
1269 config HAVE_JUMP_LABEL_HACK
1272 config HAVE_NOINSTR_HACK
1275 config HAVE_NOINSTR_VALIDATION
1278 config HAVE_UACCESS_VALIDATION
1282 config HAVE_STACK_VALIDATION
1285 Architecture supports objtool compile-time frame pointer rule
1288 config HAVE_RELIABLE_STACKTRACE
1291 Architecture has either save_stack_trace_tsk_reliable() or
1292 arch_stack_walk_reliable() function which only returns a stack trace
1293 if it can guarantee the trace is reliable.
1295 config HAVE_ARCH_HASH
1299 If this is set, the architecture provides an <asm/hash.h>
1300 file which provides platform-specific implementations of some
1301 functions in <linux/hash.h> or fs/namei.c.
1303 config HAVE_ARCH_NVRAM_OPS
1312 config CLONE_BACKWARDS
1315 Architecture has tls passed as the 4th argument of clone(2),
1318 config CLONE_BACKWARDS2
1321 Architecture has the first two arguments of clone(2) swapped.
1323 config CLONE_BACKWARDS3
1326 Architecture has tls passed as the 3rd argument of clone(2),
1329 config ODD_RT_SIGACTION
1332 Architecture has unusual rt_sigaction(2) arguments
1334 config OLD_SIGSUSPEND
1337 Architecture has old sigsuspend(2) syscall, of one-argument variety
1339 config OLD_SIGSUSPEND3
1342 Even weirder antique ABI - three-argument sigsuspend(2)
1344 config OLD_SIGACTION
1347 Architecture has old sigaction(2) syscall. Nope, not the same
1348 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
1349 but fairly different variant of sigaction(2), thanks to OSF/1
1352 config COMPAT_OLD_SIGACTION
1355 config COMPAT_32BIT_TIME
1356 bool "Provide system calls for 32-bit time_t"
1357 default !64BIT || COMPAT
1359 This enables 32 bit time_t support in addition to 64 bit time_t support.
1360 This is relevant on all 32-bit architectures, and 64-bit architectures
1361 as part of compat syscall handling.
1363 config ARCH_NO_PREEMPT
1366 config ARCH_SUPPORTS_RT
1369 config CPU_NO_EFFICIENT_FFS
1372 config HAVE_ARCH_VMAP_STACK
1375 An arch should select this symbol if it can support kernel stacks
1376 in vmalloc space. This means:
1378 - vmalloc space must be large enough to hold many kernel stacks.
1379 This may rule out many 32-bit architectures.
1381 - Stacks in vmalloc space need to work reliably. For example, if
1382 vmap page tables are created on demand, either this mechanism
1383 needs to work while the stack points to a virtual address with
1384 unpopulated page tables or arch code (switch_to() and switch_mm(),
1385 most likely) needs to ensure that the stack's page table entries
1386 are populated before running on a possibly unpopulated stack.
1388 - If the stack overflows into a guard page, something reasonable
1389 should happen. The definition of "reasonable" is flexible, but
1390 instantly rebooting without logging anything would be unfriendly.
1394 bool "Use a virtually-mapped stack"
1395 depends on HAVE_ARCH_VMAP_STACK
1396 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC
1398 Enable this if you want the use virtually-mapped kernel stacks
1399 with guard pages. This causes kernel stack overflows to be
1400 caught immediately rather than causing difficult-to-diagnose
1403 To use this with software KASAN modes, the architecture must support
1404 backing virtual mappings with real shadow memory, and KASAN_VMALLOC
1407 config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1410 An arch should select this symbol if it can support kernel stack
1411 offset randomization with calls to add_random_kstack_offset()
1412 during syscall entry and choose_random_kstack_offset() during
1413 syscall exit. Careful removal of -fstack-protector-strong and
1414 -fstack-protector should also be applied to the entry code and
1415 closely examined, as the artificial stack bump looks like an array
1416 to the compiler, so it will attempt to add canary checks regardless
1417 of the static branch state.
1419 config RANDOMIZE_KSTACK_OFFSET
1420 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT
1422 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1423 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000
1425 The kernel stack offset can be randomized (after pt_regs) by
1426 roughly 5 bits of entropy, frustrating memory corruption
1427 attacks that depend on stack address determinism or
1428 cross-syscall address exposures.
1430 The feature is controlled via the "randomize_kstack_offset=on/off"
1431 kernel boot param, and if turned off has zero overhead due to its use
1432 of static branches (see JUMP_LABEL).
1436 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
1437 bool "Default state of kernel stack offset randomization"
1438 depends on RANDOMIZE_KSTACK_OFFSET
1440 Kernel stack offset randomization is controlled by kernel boot param
1441 "randomize_kstack_offset=on/off", and this config chooses the default
1444 config ARCH_OPTIONAL_KERNEL_RWX
1447 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1450 config ARCH_HAS_STRICT_KERNEL_RWX
1453 config STRICT_KERNEL_RWX
1454 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
1455 depends on ARCH_HAS_STRICT_KERNEL_RWX
1456 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1458 If this is set, kernel text and rodata memory will be made read-only,
1459 and non-text memory will be made non-executable. This provides
1460 protection against certain security exploits (e.g. executing the heap
1463 These features are considered standard security practice these days.
1464 You should say Y here in almost all cases.
1466 config ARCH_HAS_STRICT_MODULE_RWX
1469 config STRICT_MODULE_RWX
1470 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1471 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1472 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1474 If this is set, module text and rodata memory will be made read-only,
1475 and non-text memory will be made non-executable. This provides
1476 protection against certain security exploits (e.g. writing to text)
1478 # select if the architecture provides an asm/dma-direct.h header
1479 config ARCH_HAS_PHYS_TO_DMA
1482 config HAVE_ARCH_COMPILER_H
1485 An architecture can select this if it provides an
1486 asm/compiler.h header that should be included after
1487 linux/compiler-*.h in order to override macro definitions that those
1488 headers generally provide.
1490 config HAVE_ARCH_PREL32_RELOCATIONS
1493 May be selected by an architecture if it supports place-relative
1494 32-bit relocations, both in the toolchain and in the module loader,
1495 in which case relative references can be used in special sections
1496 for PCI fixup, initcalls etc which are only half the size on 64 bit
1497 architectures, and don't require runtime relocation on relocatable
1500 config ARCH_USE_MEMREMAP_PROT
1503 config LOCK_EVENT_COUNTS
1504 bool "Locking event counts collection"
1507 Enable light-weight counting of various locking related events
1508 in the system with minimal performance impact. This reduces
1509 the chance of application behavior change because of timing
1510 differences. The counts are reported via debugfs.
1512 # Select if the architecture has support for applying RELR relocations.
1513 config ARCH_HAS_RELR
1517 bool "Use RELR relocation packing"
1518 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1521 Store the kernel's dynamic relocations in the RELR relocation packing
1522 format. Requires a compatible linker (LLD supports this feature), as
1523 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1526 config ARCH_HAS_MEM_ENCRYPT
1529 config ARCH_HAS_CC_PLATFORM
1532 config HAVE_SPARSE_SYSCALL_NR
1535 An architecture should select this if its syscall numbering is sparse
1536 to save space. For example, MIPS architecture has a syscall array with
1537 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1538 related optimizations for a given architecture.
1540 config ARCH_HAS_VDSO_TIME_DATA
1543 config HAVE_STATIC_CALL
1546 config HAVE_STATIC_CALL_INLINE
1548 depends on HAVE_STATIC_CALL
1551 config HAVE_PREEMPT_DYNAMIC
1554 config HAVE_PREEMPT_DYNAMIC_CALL
1556 depends on HAVE_STATIC_CALL
1557 select HAVE_PREEMPT_DYNAMIC
1559 An architecture should select this if it can handle the preemption
1560 model being selected at boot time using static calls.
1562 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a
1563 preemption function will be patched directly.
1565 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any
1566 call to a preemption function will go through a trampoline, and the
1567 trampoline will be patched.
1569 It is strongly advised to support inline static call to avoid any
1572 config HAVE_PREEMPT_DYNAMIC_KEY
1574 depends on HAVE_ARCH_JUMP_LABEL
1575 select HAVE_PREEMPT_DYNAMIC
1577 An architecture should select this if it can handle the preemption
1578 model being selected at boot time using static keys.
1580 Each preemption function will be given an early return based on a
1581 static key. This should have slightly lower overhead than non-inline
1582 static calls, as this effectively inlines each trampoline into the
1583 start of its callee. This may avoid redundant work, and may
1584 integrate better with CFI schemes.
1586 This will have greater overhead than using inline static calls as
1587 the call to the preemption function cannot be entirely elided.
1589 config ARCH_WANT_LD_ORPHAN_WARN
1592 An arch should select this symbol once all linker sections are explicitly
1593 included, size-asserted, or discarded in the linker scripts. This is
1594 important because we never want expected sections to be placed heuristically
1595 by the linker, since the locations of such sections can change between linker
1598 config HAVE_ARCH_PFN_VALID
1601 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
1604 config ARCH_SUPPORTS_PAGE_TABLE_CHECK
1607 config ARCH_SPLIT_ARG64
1610 If a 32-bit architecture requires 64-bit arguments to be split into
1611 pairs of 32-bit arguments, select this option.
1613 config ARCH_HAS_ELFCORE_COMPAT
1616 config ARCH_HAS_PARANOID_L1D_FLUSH
1619 config ARCH_HAVE_TRACE_MMIO_ACCESS
1622 config DYNAMIC_SIGFRAME
1625 # Select, if arch has a named attribute group bound to NUMA device nodes.
1626 config HAVE_ARCH_NODE_DEV_GROUP
1629 config ARCH_HAS_HW_PTE_YOUNG
1632 Architectures that select this option are capable of setting the
1633 accessed bit in PTE entries when using them as part of linear address
1634 translations. Architectures that require runtime check should select
1635 this option and override arch_has_hw_pte_young().
1637 config ARCH_HAS_NONLEAF_PMD_YOUNG
1640 Architectures that select this option are capable of setting the
1641 accessed bit in non-leaf PMD entries when using them as part of linear
1642 address translations. Page table walkers that clear the accessed bit
1643 may use this capability to reduce their search space.
1645 config ARCH_HAS_KERNEL_FPU_SUPPORT
1648 Architectures that select this option can run floating-point code in
1649 the kernel, as described in Documentation/core-api/floating-point.rst.
1651 source "kernel/gcov/Kconfig"
1653 source "scripts/gcc-plugins/Kconfig"
1655 config FUNCTION_ALIGNMENT_4B
1658 config FUNCTION_ALIGNMENT_8B
1661 config FUNCTION_ALIGNMENT_16B
1664 config FUNCTION_ALIGNMENT_32B
1667 config FUNCTION_ALIGNMENT_64B
1670 config FUNCTION_ALIGNMENT
1672 default 64 if FUNCTION_ALIGNMENT_64B
1673 default 32 if FUNCTION_ALIGNMENT_32B
1674 default 16 if FUNCTION_ALIGNMENT_16B
1675 default 8 if FUNCTION_ALIGNMENT_8B
1676 default 4 if FUNCTION_ALIGNMENT_4B
1679 config CC_HAS_MIN_FUNCTION_ALIGNMENT
1680 # Detect availability of the GCC option -fmin-function-alignment which
1681 # guarantees minimal alignment for all functions, unlike
1682 # -falign-functions which the compiler ignores for cold functions.
1683 def_bool $(cc-option, -fmin-function-alignment=8)
1685 config CC_HAS_SANE_FUNCTION_ALIGNMENT
1686 # Set if the guaranteed alignment with -fmin-function-alignment is
1687 # available or extra care is required in the kernel. Clang provides
1688 # strict alignment always, even with -falign-functions.
1689 def_bool CC_HAS_MIN_FUNCTION_ALIGNMENT || CC_IS_CLANG
1691 config ARCH_NEED_CMPXCHG_1_EMU
1694 config ARCH_WANTS_PRE_LINK_VMLINUX
1697 An architecture can select this if it provides arch/<arch>/tools/Makefile
1698 with .arch.vmlinux.o target to be linked into vmlinux.