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"
39 depends on HAVE_KPROBES
42 Kprobes allows you to trap at almost any kernel address and
43 execute a callback function. register_kprobe() establishes
44 a probepoint and specifies the callback. Kprobes is useful
45 for kernel debugging, non-intrusive instrumentation and testing.
49 bool "Optimize very unlikely/likely branches"
50 depends on HAVE_ARCH_JUMP_LABEL
51 depends on CC_HAS_ASM_GOTO
53 This option enables a transparent branch optimization that
54 makes certain almost-always-true or almost-always-false branch
55 conditions even cheaper to execute within the kernel.
57 Certain performance-sensitive kernel code, such as trace points,
58 scheduler functionality, networking code and KVM have such
59 branches and include support for this optimization technique.
61 If it is detected that the compiler has support for "asm goto",
62 the kernel will compile such branches with just a nop
63 instruction. When the condition flag is toggled to true, the
64 nop will be converted to a jump instruction to execute the
65 conditional block of instructions.
67 This technique lowers overhead and stress on the branch prediction
68 of the processor and generally makes the kernel faster. The update
69 of the condition is slower, but those are always very rare.
71 ( On 32-bit x86, the necessary options added to the compiler
72 flags may increase the size of the kernel slightly. )
74 config STATIC_KEYS_SELFTEST
75 bool "Static key selftest"
78 Boot time self-test of the branch patching code.
80 config STATIC_CALL_SELFTEST
81 bool "Static call selftest"
82 depends on HAVE_STATIC_CALL
84 Boot time self-test of the call patching code.
88 depends on KPROBES && HAVE_OPTPROBES
89 select TASKS_RCU if PREEMPTION
91 config KPROBES_ON_FTRACE
93 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
94 depends on DYNAMIC_FTRACE_WITH_REGS
96 If function tracer is enabled and the arch supports full
97 passing of pt_regs to function tracing, then kprobes can
98 optimize on top of function tracing.
102 depends on ARCH_SUPPORTS_UPROBES
104 Uprobes is the user-space counterpart to kprobes: they
105 enable instrumentation applications (such as 'perf probe')
106 to establish unintrusive probes in user-space binaries and
107 libraries, by executing handler functions when the probes
108 are hit by user-space applications.
110 ( These probes come in the form of single-byte breakpoints,
111 managed by the kernel and kept transparent to the probed
114 config HAVE_64BIT_ALIGNED_ACCESS
115 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
117 Some architectures require 64 bit accesses to be 64 bit
118 aligned, which also requires structs containing 64 bit values
119 to be 64 bit aligned too. This includes some 32 bit
120 architectures which can do 64 bit accesses, as well as 64 bit
121 architectures without unaligned access.
123 This symbol should be selected by an architecture if 64 bit
124 accesses are required to be 64 bit aligned in this way even
125 though it is not a 64 bit architecture.
127 See Documentation/core-api/unaligned-memory-access.rst for
128 more information on the topic of unaligned memory accesses.
130 config HAVE_EFFICIENT_UNALIGNED_ACCESS
133 Some architectures are unable to perform unaligned accesses
134 without the use of get_unaligned/put_unaligned. Others are
135 unable to perform such accesses efficiently (e.g. trap on
136 unaligned access and require fixing it up in the exception
139 This symbol should be selected by an architecture if it can
140 perform unaligned accesses efficiently to allow different
141 code paths to be selected for these cases. Some network
142 drivers, for example, could opt to not fix up alignment
143 problems with received packets if doing so would not help
146 See Documentation/core-api/unaligned-memory-access.rst for more
147 information on the topic of unaligned memory accesses.
149 config ARCH_USE_BUILTIN_BSWAP
152 Modern versions of GCC (since 4.4) have builtin functions
153 for handling byte-swapping. Using these, instead of the old
154 inline assembler that the architecture code provides in the
155 __arch_bswapXX() macros, allows the compiler to see what's
156 happening and offers more opportunity for optimisation. In
157 particular, the compiler will be able to combine the byteswap
158 with a nearby load or store and use load-and-swap or
159 store-and-swap instructions if the architecture has them. It
160 should almost *never* result in code which is worse than the
161 hand-coded assembler in <asm/swab.h>. But just in case it
162 does, the use of the builtins is optional.
164 Any architecture with load-and-swap or store-and-swap
165 instructions should set this. And it shouldn't hurt to set it
166 on architectures that don't have such instructions.
170 depends on KPROBES && HAVE_KRETPROBES
172 config USER_RETURN_NOTIFIER
174 depends on HAVE_USER_RETURN_NOTIFIER
176 Provide a kernel-internal notification when a cpu is about to
179 config HAVE_IOREMAP_PROT
185 config HAVE_KRETPROBES
188 config HAVE_OPTPROBES
191 config HAVE_KPROBES_ON_FTRACE
194 config HAVE_FUNCTION_ERROR_INJECTION
200 config TRACE_IRQFLAGS_SUPPORT
204 # An arch should select this if it provides all these things:
206 # task_pt_regs() in asm/processor.h or asm/ptrace.h
207 # arch_has_single_step() if there is hardware single-step support
208 # arch_has_block_step() if there is hardware block-step support
209 # asm/syscall.h supplying asm-generic/syscall.h interface
210 # linux/regset.h user_regset interfaces
211 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
212 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
213 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
214 # signal delivery calls tracehook_signal_handler()
216 config HAVE_ARCH_TRACEHOOK
219 config HAVE_DMA_CONTIGUOUS
222 config GENERIC_SMP_IDLE_THREAD
225 config GENERIC_IDLE_POLL_SETUP
228 config ARCH_HAS_FORTIFY_SOURCE
231 An architecture should select this when it can successfully
232 build and run with CONFIG_FORTIFY_SOURCE.
235 # Select if the arch provides a historic keepinit alias for the retain_initrd
236 # command line option
238 config ARCH_HAS_KEEPINITRD
241 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
242 config ARCH_HAS_SET_MEMORY
245 # Select if arch has all set_direct_map_invalid/default() functions
246 config ARCH_HAS_SET_DIRECT_MAP
250 # Select if the architecture provides the arch_dma_set_uncached symbol to
251 # either provide an uncached segment alias for a DMA allocation, or
252 # to remap the page tables in place.
254 config ARCH_HAS_DMA_SET_UNCACHED
258 # Select if the architectures provides the arch_dma_clear_uncached symbol
259 # to undo an in-place page table remap for uncached access.
261 config ARCH_HAS_DMA_CLEAR_UNCACHED
264 # Select if arch init_task must go in the __init_task_data section
265 config ARCH_TASK_STRUCT_ON_STACK
268 # Select if arch has its private alloc_task_struct() function
269 config ARCH_TASK_STRUCT_ALLOCATOR
272 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
274 depends on !ARCH_TASK_STRUCT_ALLOCATOR
276 An architecture should select this to provide hardened usercopy
277 knowledge about what region of the thread_struct should be
278 whitelisted for copying to userspace. Normally this is only the
279 FPU registers. Specifically, arch_thread_struct_whitelist()
280 should be implemented. Without this, the entire thread_struct
281 field in task_struct will be left whitelisted.
283 # Select if arch has its private alloc_thread_stack() function
284 config ARCH_THREAD_STACK_ALLOCATOR
287 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
288 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
291 config ARCH_WANTS_NO_INSTR
294 An architecture should select this if the noinstr macro is being used on
295 functions to denote that the toolchain should avoid instrumenting such
296 functions and is required for correctness.
298 config ARCH_32BIT_OFF_T
302 All new 32-bit architectures should have 64-bit off_t type on
303 userspace side which corresponds to the loff_t kernel type. This
304 is the requirement for modern ABIs. Some existing architectures
305 still support 32-bit off_t. This option is enabled for all such
306 architectures explicitly.
308 # Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat
309 config ARCH_32BIT_USTAT_F_TINODE
312 config HAVE_ASM_MODVERSIONS
315 This symbol should be selected by an architecture if it provides
316 <asm/asm-prototypes.h> to support the module versioning for symbols
317 exported from assembly code.
319 config HAVE_REGS_AND_STACK_ACCESS_API
322 This symbol should be selected by an architecture if it supports
323 the API needed to access registers and stack entries from pt_regs,
324 declared in asm/ptrace.h
325 For example the kprobes-based event tracer needs this API.
329 depends on HAVE_REGS_AND_STACK_ACCESS_API
331 This symbol should be selected by an architecture if it
332 supports an implementation of restartable sequences.
334 config HAVE_FUNCTION_ARG_ACCESS_API
337 This symbol should be selected by an architecture if it supports
338 the API needed to access function arguments from pt_regs,
339 declared in asm/ptrace.h
341 config HAVE_HW_BREAKPOINT
343 depends on PERF_EVENTS
345 config HAVE_MIXED_BREAKPOINTS_REGS
347 depends on HAVE_HW_BREAKPOINT
349 Depending on the arch implementation of hardware breakpoints,
350 some of them have separate registers for data and instruction
351 breakpoints addresses, others have mixed registers to store
352 them but define the access type in a control register.
353 Select this option if your arch implements breakpoints under the
356 config HAVE_USER_RETURN_NOTIFIER
359 config HAVE_PERF_EVENTS_NMI
362 System hardware can generate an NMI using the perf event
363 subsystem. Also has support for calculating CPU cycle events
364 to determine how many clock cycles in a given period.
366 config HAVE_HARDLOCKUP_DETECTOR_PERF
368 depends on HAVE_PERF_EVENTS_NMI
370 The arch chooses to use the generic perf-NMI-based hardlockup
371 detector. Must define HAVE_PERF_EVENTS_NMI.
373 config HAVE_NMI_WATCHDOG
377 The arch provides a low level NMI watchdog. It provides
378 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
380 config HAVE_HARDLOCKUP_DETECTOR_ARCH
382 select HAVE_NMI_WATCHDOG
384 The arch chooses to provide its own hardlockup detector, which is
385 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
386 interfaces and parameters provided by hardlockup detector subsystem.
388 config HAVE_PERF_REGS
391 Support selective register dumps for perf events. This includes
392 bit-mapping of each registers and a unique architecture id.
394 config HAVE_PERF_USER_STACK_DUMP
397 Support user stack dumps for perf event samples. This needs
398 access to the user stack pointer which is not unified across
401 config HAVE_ARCH_JUMP_LABEL
404 config HAVE_ARCH_JUMP_LABEL_RELATIVE
407 config MMU_GATHER_TABLE_FREE
410 config MMU_GATHER_RCU_TABLE_FREE
412 select MMU_GATHER_TABLE_FREE
414 config MMU_GATHER_PAGE_SIZE
417 config MMU_GATHER_NO_RANGE
420 config MMU_GATHER_NO_GATHER
422 depends on MMU_GATHER_TABLE_FREE
424 config ARCH_WANT_IRQS_OFF_ACTIVATE_MM
427 Temporary select until all architectures can be converted to have
428 irqs disabled over activate_mm. Architectures that do IPI based TLB
429 shootdowns should enable this.
431 config ARCH_HAVE_NMI_SAFE_CMPXCHG
434 config HAVE_ALIGNED_STRUCT_PAGE
437 This makes sure that struct pages are double word aligned and that
438 e.g. the SLUB allocator can perform double word atomic operations
439 on a struct page for better performance. However selecting this
440 might increase the size of a struct page by a word.
442 config HAVE_CMPXCHG_LOCAL
445 config HAVE_CMPXCHG_DOUBLE
448 config ARCH_WEAK_RELEASE_ACQUIRE
451 config ARCH_WANT_IPC_PARSE_VERSION
454 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
457 config ARCH_WANT_OLD_COMPAT_IPC
458 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
461 config HAVE_ARCH_SECCOMP
464 An arch should select this symbol to support seccomp mode 1 (the fixed
465 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn,
466 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment:
467 - __NR_seccomp_read_32
468 - __NR_seccomp_write_32
469 - __NR_seccomp_exit_32
470 - __NR_seccomp_sigreturn_32
472 config HAVE_ARCH_SECCOMP_FILTER
474 select HAVE_ARCH_SECCOMP
476 An arch should select this symbol if it provides all of these things:
477 - all the requirements for HAVE_ARCH_SECCOMP
479 - syscall_get_arguments()
481 - syscall_set_return_value()
482 - SIGSYS siginfo_t support
483 - secure_computing is called from a ptrace_event()-safe context
484 - secure_computing return value is checked and a return value of -1
485 results in the system call being skipped immediately.
486 - seccomp syscall wired up
487 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE,
488 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If
489 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too.
492 prompt "Enable seccomp to safely execute untrusted bytecode"
494 depends on HAVE_ARCH_SECCOMP
496 This kernel feature is useful for number crunching applications
497 that may need to handle untrusted bytecode during their
498 execution. By using pipes or other transports made available
499 to the process as file descriptors supporting the read/write
500 syscalls, it's possible to isolate those applications in their
501 own address space using seccomp. Once seccomp is enabled via
502 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be
503 disabled and the task is only allowed to execute a few safe
504 syscalls defined by each seccomp mode.
508 config SECCOMP_FILTER
510 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
512 Enable tasks to build secure computing environments defined
513 in terms of Berkeley Packet Filter programs which implement
514 task-defined system call filtering polices.
516 See Documentation/userspace-api/seccomp_filter.rst for details.
518 config SECCOMP_CACHE_DEBUG
519 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache"
520 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR
523 This enables the /proc/pid/seccomp_cache interface to monitor
524 seccomp cache data. The file format is subject to change. Reading
525 the file requires CAP_SYS_ADMIN.
527 This option is for debugging only. Enabling presents the risk that
528 an adversary may be able to infer the seccomp filter logic.
532 config HAVE_ARCH_STACKLEAK
535 An architecture should select this if it has the code which
536 fills the used part of the kernel stack with the STACKLEAK_POISON
537 value before returning from system calls.
539 config HAVE_STACKPROTECTOR
542 An arch should select this symbol if:
543 - it has implemented a stack canary (e.g. __stack_chk_guard)
545 config STACKPROTECTOR
546 bool "Stack Protector buffer overflow detection"
547 depends on HAVE_STACKPROTECTOR
548 depends on $(cc-option,-fstack-protector)
551 This option turns on the "stack-protector" GCC feature. This
552 feature puts, at the beginning of functions, a canary value on
553 the stack just before the return address, and validates
554 the value just before actually returning. Stack based buffer
555 overflows (that need to overwrite this return address) now also
556 overwrite the canary, which gets detected and the attack is then
557 neutralized via a kernel panic.
559 Functions will have the stack-protector canary logic added if they
560 have an 8-byte or larger character array on the stack.
562 This feature requires gcc version 4.2 or above, or a distribution
563 gcc with the feature backported ("-fstack-protector").
565 On an x86 "defconfig" build, this feature adds canary checks to
566 about 3% of all kernel functions, which increases kernel code size
569 config STACKPROTECTOR_STRONG
570 bool "Strong Stack Protector"
571 depends on STACKPROTECTOR
572 depends on $(cc-option,-fstack-protector-strong)
575 Functions will have the stack-protector canary logic added in any
576 of the following conditions:
578 - local variable's address used as part of the right hand side of an
579 assignment or function argument
580 - local variable is an array (or union containing an array),
581 regardless of array type or length
582 - uses register local variables
584 This feature requires gcc version 4.9 or above, or a distribution
585 gcc with the feature backported ("-fstack-protector-strong").
587 On an x86 "defconfig" build, this feature adds canary checks to
588 about 20% of all kernel functions, which increases the kernel code
591 config ARCH_SUPPORTS_SHADOW_CALL_STACK
594 An architecture should select this if it supports Clang's Shadow
595 Call Stack and implements runtime support for shadow stack
598 config SHADOW_CALL_STACK
599 bool "Clang Shadow Call Stack"
600 depends on CC_IS_CLANG && ARCH_SUPPORTS_SHADOW_CALL_STACK
601 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER
603 This option enables Clang's Shadow Call Stack, which uses a
604 shadow stack to protect function return addresses from being
605 overwritten by an attacker. More information can be found in
606 Clang's documentation:
608 https://clang.llvm.org/docs/ShadowCallStack.html
610 Note that security guarantees in the kernel differ from the
611 ones documented for user space. The kernel must store addresses
612 of shadow stacks in memory, which means an attacker capable of
613 reading and writing arbitrary memory may be able to locate them
614 and hijack control flow by modifying the stacks.
619 Selected if the kernel will be built using the compiler's LTO feature.
625 Selected if the kernel will be built using Clang's LTO feature.
627 config ARCH_SUPPORTS_LTO_CLANG
630 An architecture should select this option if it supports:
631 - compiling with Clang,
632 - compiling inline assembly with Clang's integrated assembler,
633 - and linking with LLD.
635 config ARCH_SUPPORTS_LTO_CLANG_THIN
638 An architecture should select this option if it can support Clang's
643 # Clang >= 11: https://github.com/ClangBuiltLinux/linux/issues/510
644 depends on CC_IS_CLANG && CLANG_VERSION >= 110000 && LD_IS_LLD && AS_IS_LLVM
645 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm)
646 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm)
647 depends on ARCH_SUPPORTS_LTO_CLANG
648 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT
649 depends on !KASAN || KASAN_HW_TAGS
650 depends on !GCOV_KERNEL
652 The compiler and Kconfig options support building with Clang's
656 prompt "Link Time Optimization (LTO)"
659 This option enables Link Time Optimization (LTO), which allows the
660 compiler to optimize binaries globally.
662 If unsure, select LTO_NONE. Note that LTO is very resource-intensive
663 so it's disabled by default.
668 Build the kernel normally, without Link Time Optimization (LTO).
670 config LTO_CLANG_FULL
671 bool "Clang Full LTO (EXPERIMENTAL)"
672 depends on HAS_LTO_CLANG
673 depends on !COMPILE_TEST
676 This option enables Clang's full Link Time Optimization (LTO), which
677 allows the compiler to optimize the kernel globally. If you enable
678 this option, the compiler generates LLVM bitcode instead of ELF
679 object files, and the actual compilation from bitcode happens at
680 the LTO link step, which may take several minutes depending on the
681 kernel configuration. More information can be found from LLVM's
684 https://llvm.org/docs/LinkTimeOptimization.html
686 During link time, this option can use a large amount of RAM, and
687 may take much longer than the ThinLTO option.
689 config LTO_CLANG_THIN
690 bool "Clang ThinLTO (EXPERIMENTAL)"
691 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN
694 This option enables Clang's ThinLTO, which allows for parallel
695 optimization and faster incremental compiles compared to the
696 CONFIG_LTO_CLANG_FULL option. More information can be found
697 from Clang's documentation:
699 https://clang.llvm.org/docs/ThinLTO.html
704 config ARCH_SUPPORTS_CFI_CLANG
707 An architecture should select this option if it can support Clang's
708 Control-Flow Integrity (CFI) checking.
711 bool "Use Clang's Control Flow Integrity (CFI)"
712 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG
714 # - https://bugs.llvm.org/show_bug.cgi?id=46258
715 # - https://bugs.llvm.org/show_bug.cgi?id=47479
716 depends on CLANG_VERSION >= 120000
719 This option enables Clang’s forward-edge Control Flow Integrity
720 (CFI) checking, where the compiler injects a runtime check to each
721 indirect function call to ensure the target is a valid function with
722 the correct static type. This restricts possible call targets and
723 makes it more difficult for an attacker to exploit bugs that allow
724 the modification of stored function pointers. More information can be
725 found from Clang's documentation:
727 https://clang.llvm.org/docs/ControlFlowIntegrity.html
729 config CFI_CLANG_SHADOW
730 bool "Use CFI shadow to speed up cross-module checks"
732 depends on CFI_CLANG && MODULES
734 If you select this option, the kernel builds a fast look-up table of
735 CFI check functions in loaded modules to reduce performance overhead.
739 config CFI_PERMISSIVE
740 bool "Use CFI in permissive mode"
743 When selected, Control Flow Integrity (CFI) violations result in a
744 warning instead of a kernel panic. This option should only be used
745 for finding indirect call type mismatches during development.
749 config HAVE_ARCH_WITHIN_STACK_FRAMES
752 An architecture should select this if it can walk the kernel stack
753 frames to determine if an object is part of either the arguments
754 or local variables (i.e. that it excludes saved return addresses,
755 and similar) by implementing an inline arch_within_stack_frames(),
756 which is used by CONFIG_HARDENED_USERCOPY.
758 config HAVE_CONTEXT_TRACKING
761 Provide kernel/user boundaries probes necessary for subsystems
762 that need it, such as userspace RCU extended quiescent state.
763 Syscalls need to be wrapped inside user_exit()-user_enter(), either
764 optimized behind static key or through the slow path using TIF_NOHZ
765 flag. Exceptions handlers must be wrapped as well. Irqs are already
766 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal
767 handling on irq exit still need to be protected.
769 config HAVE_CONTEXT_TRACKING_OFFSTACK
772 Architecture neither relies on exception_enter()/exception_exit()
773 nor on schedule_user(). Also preempt_schedule_notrace() and
774 preempt_schedule_irq() can't be called in a preemptible section
775 while context tracking is CONTEXT_USER. This feature reflects a sane
776 entry implementation where the following requirements are met on
777 critical entry code, ie: before user_exit() or after user_enter():
779 - Critical entry code isn't preemptible (or better yet:
781 - No use of RCU read side critical sections, unless rcu_nmi_enter()
783 - No use of instrumentation, unless instrumentation_begin() got
789 Arch relies on TIF_NOHZ and syscall slow path to implement context
790 tracking calls to user_enter()/user_exit().
792 config HAVE_VIRT_CPU_ACCOUNTING
795 config HAVE_VIRT_CPU_ACCOUNTING_IDLE
798 Architecture has its own way to account idle CPU time and therefore
799 doesn't implement vtime_account_idle().
801 config ARCH_HAS_SCALED_CPUTIME
804 config HAVE_VIRT_CPU_ACCOUNTING_GEN
808 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
809 Before enabling this option, arch code must be audited
810 to ensure there are no races in concurrent read/write of
811 cputime_t. For example, reading/writing 64-bit cputime_t on
812 some 32-bit arches may require multiple accesses, so proper
813 locking is needed to protect against concurrent accesses.
815 config HAVE_IRQ_TIME_ACCOUNTING
818 Archs need to ensure they use a high enough resolution clock to
819 support irq time accounting and then call enable_sched_clock_irqtime().
824 Architectures that select this are able to move page tables at the
825 PUD level. If there are only 3 page table levels, the move effectively
826 happens at the PGD level.
831 Archs that select this are able to move page tables at the PMD level.
833 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
836 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
839 config HAVE_ARCH_HUGE_VMAP
843 # Archs that select this would be capable of PMD-sized vmaps (i.e.,
844 # arch_vmap_pmd_supported() returns true), and they must make no assumptions
845 # that vmalloc memory is mapped with PAGE_SIZE ptes. The VM_NO_HUGE_VMAP flag
846 # can be used to prohibit arch-specific allocations from using hugepages to
847 # help with this (e.g., modules may require it).
849 config HAVE_ARCH_HUGE_VMALLOC
850 depends on HAVE_ARCH_HUGE_VMAP
853 config ARCH_WANT_HUGE_PMD_SHARE
856 config HAVE_ARCH_SOFT_DIRTY
859 config HAVE_MOD_ARCH_SPECIFIC
862 The arch uses struct mod_arch_specific to store data. Many arches
863 just need a simple module loader without arch specific data - those
864 should not enable this.
866 config MODULES_USE_ELF_RELA
869 Modules only use ELF RELA relocations. Modules with ELF REL
870 relocations will give an error.
872 config MODULES_USE_ELF_REL
875 Modules only use ELF REL relocations. Modules with ELF RELA
876 relocations will give an error.
878 config HAVE_IRQ_EXIT_ON_IRQ_STACK
881 Architecture doesn't only execute the irq handler on the irq stack
882 but also irq_exit(). This way we can process softirqs on this irq
883 stack instead of switching to a new one when we call __do_softirq()
884 in the end of an hardirq.
885 This spares a stack switch and improves cache usage on softirq
888 config HAVE_SOFTIRQ_ON_OWN_STACK
891 Architecture provides a function to run __do_softirq() on a
894 config PGTABLE_LEVELS
898 config ARCH_HAS_ELF_RANDOMIZE
901 An architecture supports choosing randomized locations for
902 stack, mmap, brk, and ET_DYN. Defined functions:
904 - arch_randomize_brk()
906 config HAVE_ARCH_MMAP_RND_BITS
909 An arch should select this symbol if it supports setting a variable
910 number of bits for use in establishing the base address for mmap
911 allocations, has MMU enabled and provides values for both:
912 - ARCH_MMAP_RND_BITS_MIN
913 - ARCH_MMAP_RND_BITS_MAX
915 config HAVE_EXIT_THREAD
918 An architecture implements exit_thread.
920 config ARCH_MMAP_RND_BITS_MIN
923 config ARCH_MMAP_RND_BITS_MAX
926 config ARCH_MMAP_RND_BITS_DEFAULT
929 config ARCH_MMAP_RND_BITS
930 int "Number of bits to use for ASLR of mmap base address" if EXPERT
931 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
932 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
933 default ARCH_MMAP_RND_BITS_MIN
934 depends on HAVE_ARCH_MMAP_RND_BITS
936 This value can be used to select the number of bits to use to
937 determine the random offset to the base address of vma regions
938 resulting from mmap allocations. This value will be bounded
939 by the architecture's minimum and maximum supported values.
941 This value can be changed after boot using the
942 /proc/sys/vm/mmap_rnd_bits tunable
944 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
947 An arch should select this symbol if it supports running applications
948 in compatibility mode, supports setting a variable number of bits for
949 use in establishing the base address for mmap allocations, has MMU
950 enabled and provides values for both:
951 - ARCH_MMAP_RND_COMPAT_BITS_MIN
952 - ARCH_MMAP_RND_COMPAT_BITS_MAX
954 config ARCH_MMAP_RND_COMPAT_BITS_MIN
957 config ARCH_MMAP_RND_COMPAT_BITS_MAX
960 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
963 config ARCH_MMAP_RND_COMPAT_BITS
964 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
965 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
966 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
967 default ARCH_MMAP_RND_COMPAT_BITS_MIN
968 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
970 This value can be used to select the number of bits to use to
971 determine the random offset to the base address of vma regions
972 resulting from mmap allocations for compatible applications This
973 value will be bounded by the architecture's minimum and maximum
976 This value can be changed after boot using the
977 /proc/sys/vm/mmap_rnd_compat_bits tunable
979 config HAVE_ARCH_COMPAT_MMAP_BASES
982 This allows 64bit applications to invoke 32-bit mmap() syscall
983 and vice-versa 32-bit applications to call 64-bit mmap().
984 Required for applications doing different bitness syscalls.
986 # This allows to use a set of generic functions to determine mmap base
987 # address by giving priority to top-down scheme only if the process
988 # is not in legacy mode (compat task, unlimited stack size or
989 # sysctl_legacy_va_layout).
990 # Architecture that selects this option can provide its own version of:
992 config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT
995 select ARCH_HAS_ELF_RANDOMIZE
997 config HAVE_STACK_VALIDATION
1000 Architecture supports the 'objtool check' host tool command, which
1001 performs compile-time stack metadata validation.
1003 config HAVE_RELIABLE_STACKTRACE
1006 Architecture has either save_stack_trace_tsk_reliable() or
1007 arch_stack_walk_reliable() function which only returns a stack trace
1008 if it can guarantee the trace is reliable.
1010 config HAVE_ARCH_HASH
1014 If this is set, the architecture provides an <asm/hash.h>
1015 file which provides platform-specific implementations of some
1016 functions in <linux/hash.h> or fs/namei.c.
1018 config HAVE_ARCH_NVRAM_OPS
1027 config CLONE_BACKWARDS
1030 Architecture has tls passed as the 4th argument of clone(2),
1033 config CLONE_BACKWARDS2
1036 Architecture has the first two arguments of clone(2) swapped.
1038 config CLONE_BACKWARDS3
1041 Architecture has tls passed as the 3rd argument of clone(2),
1044 config ODD_RT_SIGACTION
1047 Architecture has unusual rt_sigaction(2) arguments
1049 config OLD_SIGSUSPEND
1052 Architecture has old sigsuspend(2) syscall, of one-argument variety
1054 config OLD_SIGSUSPEND3
1057 Even weirder antique ABI - three-argument sigsuspend(2)
1059 config OLD_SIGACTION
1062 Architecture has old sigaction(2) syscall. Nope, not the same
1063 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
1064 but fairly different variant of sigaction(2), thanks to OSF/1
1067 config COMPAT_OLD_SIGACTION
1070 config COMPAT_32BIT_TIME
1071 bool "Provide system calls for 32-bit time_t"
1072 default !64BIT || COMPAT
1074 This enables 32 bit time_t support in addition to 64 bit time_t support.
1075 This is relevant on all 32-bit architectures, and 64-bit architectures
1076 as part of compat syscall handling.
1078 config ARCH_NO_PREEMPT
1081 config ARCH_EPHEMERAL_INODES
1084 An arch should select this symbol if it doesn't keep track of inode
1085 instances on its own, but instead relies on something else (e.g. the
1086 host kernel for an UML kernel).
1088 config ARCH_SUPPORTS_RT
1091 config CPU_NO_EFFICIENT_FFS
1094 config HAVE_ARCH_VMAP_STACK
1097 An arch should select this symbol if it can support kernel stacks
1098 in vmalloc space. This means:
1100 - vmalloc space must be large enough to hold many kernel stacks.
1101 This may rule out many 32-bit architectures.
1103 - Stacks in vmalloc space need to work reliably. For example, if
1104 vmap page tables are created on demand, either this mechanism
1105 needs to work while the stack points to a virtual address with
1106 unpopulated page tables or arch code (switch_to() and switch_mm(),
1107 most likely) needs to ensure that the stack's page table entries
1108 are populated before running on a possibly unpopulated stack.
1110 - If the stack overflows into a guard page, something reasonable
1111 should happen. The definition of "reasonable" is flexible, but
1112 instantly rebooting without logging anything would be unfriendly.
1116 bool "Use a virtually-mapped stack"
1117 depends on HAVE_ARCH_VMAP_STACK
1118 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC
1120 Enable this if you want the use virtually-mapped kernel stacks
1121 with guard pages. This causes kernel stack overflows to be
1122 caught immediately rather than causing difficult-to-diagnose
1125 To use this with software KASAN modes, the architecture must support
1126 backing virtual mappings with real shadow memory, and KASAN_VMALLOC
1129 config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1132 An arch should select this symbol if it can support kernel stack
1133 offset randomization with calls to add_random_kstack_offset()
1134 during syscall entry and choose_random_kstack_offset() during
1135 syscall exit. Careful removal of -fstack-protector-strong and
1136 -fstack-protector should also be applied to the entry code and
1137 closely examined, as the artificial stack bump looks like an array
1138 to the compiler, so it will attempt to add canary checks regardless
1139 of the static branch state.
1141 config RANDOMIZE_KSTACK_OFFSET_DEFAULT
1142 bool "Randomize kernel stack offset on syscall entry"
1143 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET
1145 The kernel stack offset can be randomized (after pt_regs) by
1146 roughly 5 bits of entropy, frustrating memory corruption
1147 attacks that depend on stack address determinism or
1148 cross-syscall address exposures. This feature is controlled
1149 by kernel boot param "randomize_kstack_offset=on/off", and this
1150 config chooses the default boot state.
1152 config ARCH_OPTIONAL_KERNEL_RWX
1155 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1158 config ARCH_HAS_STRICT_KERNEL_RWX
1161 config STRICT_KERNEL_RWX
1162 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
1163 depends on ARCH_HAS_STRICT_KERNEL_RWX
1164 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1166 If this is set, kernel text and rodata memory will be made read-only,
1167 and non-text memory will be made non-executable. This provides
1168 protection against certain security exploits (e.g. executing the heap
1171 These features are considered standard security practice these days.
1172 You should say Y here in almost all cases.
1174 config ARCH_HAS_STRICT_MODULE_RWX
1177 config STRICT_MODULE_RWX
1178 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1179 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1180 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1182 If this is set, module text and rodata memory will be made read-only,
1183 and non-text memory will be made non-executable. This provides
1184 protection against certain security exploits (e.g. writing to text)
1186 # select if the architecture provides an asm/dma-direct.h header
1187 config ARCH_HAS_PHYS_TO_DMA
1190 config HAVE_ARCH_COMPILER_H
1193 An architecture can select this if it provides an
1194 asm/compiler.h header that should be included after
1195 linux/compiler-*.h in order to override macro definitions that those
1196 headers generally provide.
1198 config HAVE_ARCH_PREL32_RELOCATIONS
1201 May be selected by an architecture if it supports place-relative
1202 32-bit relocations, both in the toolchain and in the module loader,
1203 in which case relative references can be used in special sections
1204 for PCI fixup, initcalls etc which are only half the size on 64 bit
1205 architectures, and don't require runtime relocation on relocatable
1208 config ARCH_USE_MEMREMAP_PROT
1211 config LOCK_EVENT_COUNTS
1212 bool "Locking event counts collection"
1215 Enable light-weight counting of various locking related events
1216 in the system with minimal performance impact. This reduces
1217 the chance of application behavior change because of timing
1218 differences. The counts are reported via debugfs.
1220 # Select if the architecture has support for applying RELR relocations.
1221 config ARCH_HAS_RELR
1225 bool "Use RELR relocation packing"
1226 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR
1229 Store the kernel's dynamic relocations in the RELR relocation packing
1230 format. Requires a compatible linker (LLD supports this feature), as
1231 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy
1234 config ARCH_HAS_MEM_ENCRYPT
1237 config HAVE_SPARSE_SYSCALL_NR
1240 An architecture should select this if its syscall numbering is sparse
1241 to save space. For example, MIPS architecture has a syscall array with
1242 entries at 4000, 5000 and 6000 locations. This option turns on syscall
1243 related optimizations for a given architecture.
1245 config ARCH_HAS_VDSO_DATA
1248 config HAVE_STATIC_CALL
1251 config HAVE_STATIC_CALL_INLINE
1253 depends on HAVE_STATIC_CALL
1255 config HAVE_PREEMPT_DYNAMIC
1257 depends on HAVE_STATIC_CALL
1258 depends on GENERIC_ENTRY
1260 Select this if the architecture support boot time preempt setting
1261 on top of static calls. It is strongly advised to support inline
1262 static call to avoid any overhead.
1264 config ARCH_WANT_LD_ORPHAN_WARN
1267 An arch should select this symbol once all linker sections are explicitly
1268 included, size-asserted, or discarded in the linker scripts. This is
1269 important because we never want expected sections to be placed heuristically
1270 by the linker, since the locations of such sections can change between linker
1273 config HAVE_ARCH_PFN_VALID
1276 config ARCH_SUPPORTS_DEBUG_PAGEALLOC
1279 config ARCH_SPLIT_ARG64
1282 If a 32-bit architecture requires 64-bit arguments to be split into
1283 pairs of 32-bit arguments, select this option.
1285 config ARCH_HAS_ELFCORE_COMPAT
1288 config ARCH_HAS_PARANOID_L1D_FLUSH
1291 source "kernel/gcov/Kconfig"
1293 source "scripts/gcc-plugins/Kconfig"