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"
23 tristate "OProfile system profiling"
25 depends on HAVE_OPROFILE
27 select RING_BUFFER_ALLOW_SWAP
29 OProfile is a profiling system capable of profiling the
30 whole system, include the kernel, kernel modules, libraries,
35 config OPROFILE_EVENT_MULTIPLEX
36 bool "OProfile multiplexing support (EXPERIMENTAL)"
38 depends on OPROFILE && X86
40 The number of hardware counters is limited. The multiplexing
41 feature enables OProfile to gather more events than counters
42 are provided by the hardware. This is realized by switching
43 between events at a user specified time interval.
50 config OPROFILE_NMI_TIMER
52 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
57 depends on HAVE_KPROBES
60 Kprobes allows you to trap at almost any kernel address and
61 execute a callback function. register_kprobe() establishes
62 a probepoint and specifies the callback. Kprobes is useful
63 for kernel debugging, non-intrusive instrumentation and testing.
67 bool "Optimize very unlikely/likely branches"
68 depends on HAVE_ARCH_JUMP_LABEL
70 This option enables a transparent branch optimization that
71 makes certain almost-always-true or almost-always-false branch
72 conditions even cheaper to execute within the kernel.
74 Certain performance-sensitive kernel code, such as trace points,
75 scheduler functionality, networking code and KVM have such
76 branches and include support for this optimization technique.
78 If it is detected that the compiler has support for "asm goto",
79 the kernel will compile such branches with just a nop
80 instruction. When the condition flag is toggled to true, the
81 nop will be converted to a jump instruction to execute the
82 conditional block of instructions.
84 This technique lowers overhead and stress on the branch prediction
85 of the processor and generally makes the kernel faster. The update
86 of the condition is slower, but those are always very rare.
88 ( On 32-bit x86, the necessary options added to the compiler
89 flags may increase the size of the kernel slightly. )
91 config STATIC_KEYS_SELFTEST
92 bool "Static key selftest"
95 Boot time self-test of the branch patching code.
99 depends on KPROBES && HAVE_OPTPROBES
100 select TASKS_RCU if PREEMPT
102 config KPROBES_ON_FTRACE
104 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
105 depends on DYNAMIC_FTRACE_WITH_REGS
107 If function tracer is enabled and the arch supports full
108 passing of pt_regs to function tracing, then kprobes can
109 optimize on top of function tracing.
113 depends on ARCH_SUPPORTS_UPROBES
115 Uprobes is the user-space counterpart to kprobes: they
116 enable instrumentation applications (such as 'perf probe')
117 to establish unintrusive probes in user-space binaries and
118 libraries, by executing handler functions when the probes
119 are hit by user-space applications.
121 ( These probes come in the form of single-byte breakpoints,
122 managed by the kernel and kept transparent to the probed
125 config HAVE_64BIT_ALIGNED_ACCESS
126 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
128 Some architectures require 64 bit accesses to be 64 bit
129 aligned, which also requires structs containing 64 bit values
130 to be 64 bit aligned too. This includes some 32 bit
131 architectures which can do 64 bit accesses, as well as 64 bit
132 architectures without unaligned access.
134 This symbol should be selected by an architecture if 64 bit
135 accesses are required to be 64 bit aligned in this way even
136 though it is not a 64 bit architecture.
138 See Documentation/unaligned-memory-access.txt for more
139 information on the topic of unaligned memory accesses.
141 config HAVE_EFFICIENT_UNALIGNED_ACCESS
144 Some architectures are unable to perform unaligned accesses
145 without the use of get_unaligned/put_unaligned. Others are
146 unable to perform such accesses efficiently (e.g. trap on
147 unaligned access and require fixing it up in the exception
150 This symbol should be selected by an architecture if it can
151 perform unaligned accesses efficiently to allow different
152 code paths to be selected for these cases. Some network
153 drivers, for example, could opt to not fix up alignment
154 problems with received packets if doing so would not help
157 See Documentation/unaligned-memory-access.txt for more
158 information on the topic of unaligned memory accesses.
160 config ARCH_USE_BUILTIN_BSWAP
163 Modern versions of GCC (since 4.4) have builtin functions
164 for handling byte-swapping. Using these, instead of the old
165 inline assembler that the architecture code provides in the
166 __arch_bswapXX() macros, allows the compiler to see what's
167 happening and offers more opportunity for optimisation. In
168 particular, the compiler will be able to combine the byteswap
169 with a nearby load or store and use load-and-swap or
170 store-and-swap instructions if the architecture has them. It
171 should almost *never* result in code which is worse than the
172 hand-coded assembler in <asm/swab.h>. But just in case it
173 does, the use of the builtins is optional.
175 Any architecture with load-and-swap or store-and-swap
176 instructions should set this. And it shouldn't hurt to set it
177 on architectures that don't have such instructions.
181 depends on KPROBES && HAVE_KRETPROBES
183 config USER_RETURN_NOTIFIER
185 depends on HAVE_USER_RETURN_NOTIFIER
187 Provide a kernel-internal notification when a cpu is about to
190 config HAVE_IOREMAP_PROT
196 config HAVE_KRETPROBES
199 config HAVE_OPTPROBES
202 config HAVE_KPROBES_ON_FTRACE
205 config HAVE_FUNCTION_ERROR_INJECTION
212 # An arch should select this if it provides all these things:
214 # task_pt_regs() in asm/processor.h or asm/ptrace.h
215 # arch_has_single_step() if there is hardware single-step support
216 # arch_has_block_step() if there is hardware block-step support
217 # asm/syscall.h supplying asm-generic/syscall.h interface
218 # linux/regset.h user_regset interfaces
219 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
220 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
221 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
222 # signal delivery calls tracehook_signal_handler()
224 config HAVE_ARCH_TRACEHOOK
227 config HAVE_DMA_CONTIGUOUS
230 config GENERIC_SMP_IDLE_THREAD
233 config GENERIC_IDLE_POLL_SETUP
236 config ARCH_HAS_FORTIFY_SOURCE
239 An architecture should select this when it can successfully
240 build and run with CONFIG_FORTIFY_SOURCE.
242 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
243 config ARCH_HAS_SET_MEMORY
246 # Select if arch init_task must go in the __init_task_data section
247 config ARCH_TASK_STRUCT_ON_STACK
250 # Select if arch has its private alloc_task_struct() function
251 config ARCH_TASK_STRUCT_ALLOCATOR
254 config HAVE_ARCH_THREAD_STRUCT_WHITELIST
256 depends on !ARCH_TASK_STRUCT_ALLOCATOR
258 An architecture should select this to provide hardened usercopy
259 knowledge about what region of the thread_struct should be
260 whitelisted for copying to userspace. Normally this is only the
261 FPU registers. Specifically, arch_thread_struct_whitelist()
262 should be implemented. Without this, the entire thread_struct
263 field in task_struct will be left whitelisted.
265 # Select if arch has its private alloc_thread_stack() function
266 config ARCH_THREAD_STACK_ALLOCATOR
269 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
270 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
273 config HAVE_REGS_AND_STACK_ACCESS_API
276 This symbol should be selected by an architecure if it supports
277 the API needed to access registers and stack entries from pt_regs,
278 declared in asm/ptrace.h
279 For example the kprobes-based event tracer needs this API.
283 depends on HAVE_REGS_AND_STACK_ACCESS_API
285 This symbol should be selected by an architecture if it
286 supports an implementation of restartable sequences.
291 The <linux/clk.h> calls support software clock gating and
292 thus are a key power management tool on many systems.
294 config HAVE_HW_BREAKPOINT
296 depends on PERF_EVENTS
298 config HAVE_MIXED_BREAKPOINTS_REGS
300 depends on HAVE_HW_BREAKPOINT
302 Depending on the arch implementation of hardware breakpoints,
303 some of them have separate registers for data and instruction
304 breakpoints addresses, others have mixed registers to store
305 them but define the access type in a control register.
306 Select this option if your arch implements breakpoints under the
309 config HAVE_USER_RETURN_NOTIFIER
312 config HAVE_PERF_EVENTS_NMI
315 System hardware can generate an NMI using the perf event
316 subsystem. Also has support for calculating CPU cycle events
317 to determine how many clock cycles in a given period.
319 config HAVE_HARDLOCKUP_DETECTOR_PERF
321 depends on HAVE_PERF_EVENTS_NMI
323 The arch chooses to use the generic perf-NMI-based hardlockup
324 detector. Must define HAVE_PERF_EVENTS_NMI.
326 config HAVE_NMI_WATCHDOG
330 The arch provides a low level NMI watchdog. It provides
331 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
333 config HAVE_HARDLOCKUP_DETECTOR_ARCH
335 select HAVE_NMI_WATCHDOG
337 The arch chooses to provide its own hardlockup detector, which is
338 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
339 interfaces and parameters provided by hardlockup detector subsystem.
341 config HAVE_PERF_REGS
344 Support selective register dumps for perf events. This includes
345 bit-mapping of each registers and a unique architecture id.
347 config HAVE_PERF_USER_STACK_DUMP
350 Support user stack dumps for perf event samples. This needs
351 access to the user stack pointer which is not unified across
354 config HAVE_ARCH_JUMP_LABEL
357 config HAVE_RCU_TABLE_FREE
360 config ARCH_HAVE_NMI_SAFE_CMPXCHG
363 config HAVE_ALIGNED_STRUCT_PAGE
366 This makes sure that struct pages are double word aligned and that
367 e.g. the SLUB allocator can perform double word atomic operations
368 on a struct page for better performance. However selecting this
369 might increase the size of a struct page by a word.
371 config HAVE_CMPXCHG_LOCAL
374 config HAVE_CMPXCHG_DOUBLE
377 config ARCH_WEAK_RELEASE_ACQUIRE
380 config ARCH_WANT_IPC_PARSE_VERSION
383 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
386 config ARCH_WANT_OLD_COMPAT_IPC
387 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
390 config HAVE_ARCH_SECCOMP_FILTER
393 An arch should select this symbol if it provides all of these things:
395 - syscall_get_arguments()
397 - syscall_set_return_value()
398 - SIGSYS siginfo_t support
399 - secure_computing is called from a ptrace_event()-safe context
400 - secure_computing return value is checked and a return value of -1
401 results in the system call being skipped immediately.
402 - seccomp syscall wired up
404 config SECCOMP_FILTER
406 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
408 Enable tasks to build secure computing environments defined
409 in terms of Berkeley Packet Filter programs which implement
410 task-defined system call filtering polices.
412 See Documentation/userspace-api/seccomp_filter.rst for details.
414 preferred-plugin-hostcc := $(if-success,[ $(gcc-version) -ge 40800 ],$(HOSTCXX),$(HOSTCC))
418 default "$(shell,$(srctree)/scripts/gcc-plugin.sh "$(preferred-plugin-hostcc)" "$(HOSTCXX)" "$(CC)")"
420 Host compiler used to build GCC plugins. This can be $(HOSTCXX),
421 $(HOSTCC), or a null string if GCC plugin is unsupported.
423 config HAVE_GCC_PLUGINS
426 An arch should select this symbol if it supports building with
429 menuconfig GCC_PLUGINS
431 depends on HAVE_GCC_PLUGINS
432 depends on PLUGIN_HOSTCC != ""
434 GCC plugins are loadable modules that provide extra features to the
435 compiler. They are useful for runtime instrumentation and static analysis.
437 See Documentation/gcc-plugins.txt for details.
439 config GCC_PLUGIN_CYC_COMPLEXITY
440 bool "Compute the cyclomatic complexity of a function" if EXPERT
441 depends on GCC_PLUGINS
442 depends on !COMPILE_TEST # too noisy
444 The complexity M of a function's control flow graph is defined as:
448 E = the number of edges
449 N = the number of nodes
450 P = the number of connected components (exit nodes).
452 Enabling this plugin reports the complexity to stderr during the
453 build. It mainly serves as a simple example of how to create a
454 gcc plugin for the kernel.
456 config GCC_PLUGIN_SANCOV
458 depends on GCC_PLUGINS
460 This plugin inserts a __sanitizer_cov_trace_pc() call at the start of
461 basic blocks. It supports all gcc versions with plugin support (from
462 gcc-4.5 on). It is based on the commit "Add fuzzing coverage support"
463 by Dmitry Vyukov <dvyukov@google.com>.
465 config GCC_PLUGIN_LATENT_ENTROPY
466 bool "Generate some entropy during boot and runtime"
467 depends on GCC_PLUGINS
469 By saying Y here the kernel will instrument some kernel code to
470 extract some entropy from both original and artificially created
471 program state. This will help especially embedded systems where
472 there is little 'natural' source of entropy normally. The cost
473 is some slowdown of the boot process (about 0.5%) and fork and
476 Note that entropy extracted this way is not cryptographically
479 This plugin was ported from grsecurity/PaX. More information at:
480 * https://grsecurity.net/
481 * https://pax.grsecurity.net/
483 config GCC_PLUGIN_STRUCTLEAK
484 bool "Force initialization of variables containing userspace addresses"
485 depends on GCC_PLUGINS
486 # Currently STRUCTLEAK inserts initialization out of live scope of
487 # variables from KASAN point of view. This leads to KASAN false
488 # positive reports. Prohibit this combination for now.
489 depends on !KASAN_EXTRA
491 This plugin zero-initializes any structures containing a
492 __user attribute. This can prevent some classes of information
495 This plugin was ported from grsecurity/PaX. More information at:
496 * https://grsecurity.net/
497 * https://pax.grsecurity.net/
499 config GCC_PLUGIN_STRUCTLEAK_BYREF_ALL
500 bool "Force initialize all struct type variables passed by reference"
501 depends on GCC_PLUGIN_STRUCTLEAK
502 depends on !COMPILE_TEST
504 Zero initialize any struct type local variable that may be passed by
505 reference without having been initialized.
507 config GCC_PLUGIN_STRUCTLEAK_VERBOSE
508 bool "Report forcefully initialized variables"
509 depends on GCC_PLUGIN_STRUCTLEAK
510 depends on !COMPILE_TEST # too noisy
512 This option will cause a warning to be printed each time the
513 structleak plugin finds a variable it thinks needs to be
514 initialized. Since not all existing initializers are detected
515 by the plugin, this can produce false positive warnings.
517 config GCC_PLUGIN_RANDSTRUCT
518 bool "Randomize layout of sensitive kernel structures"
519 depends on GCC_PLUGINS
520 select MODVERSIONS if MODULES
522 If you say Y here, the layouts of structures that are entirely
523 function pointers (and have not been manually annotated with
524 __no_randomize_layout), or structures that have been explicitly
525 marked with __randomize_layout, will be randomized at compile-time.
526 This can introduce the requirement of an additional information
527 exposure vulnerability for exploits targeting these structure
530 Enabling this feature will introduce some performance impact,
531 slightly increase memory usage, and prevent the use of forensic
532 tools like Volatility against the system (unless the kernel
533 source tree isn't cleaned after kernel installation).
535 The seed used for compilation is located at
536 scripts/gcc-plgins/randomize_layout_seed.h. It remains after
537 a make clean to allow for external modules to be compiled with
538 the existing seed and will be removed by a make mrproper or
541 Note that the implementation requires gcc 4.7 or newer.
543 This plugin was ported from grsecurity/PaX. More information at:
544 * https://grsecurity.net/
545 * https://pax.grsecurity.net/
547 config GCC_PLUGIN_RANDSTRUCT_PERFORMANCE
548 bool "Use cacheline-aware structure randomization"
549 depends on GCC_PLUGIN_RANDSTRUCT
550 depends on !COMPILE_TEST # do not reduce test coverage
552 If you say Y here, the RANDSTRUCT randomization will make a
553 best effort at restricting randomization to cacheline-sized
554 groups of elements. It will further not randomize bitfields
555 in structures. This reduces the performance hit of RANDSTRUCT
556 at the cost of weakened randomization.
558 config HAVE_STACKPROTECTOR
561 An arch should select this symbol if:
562 - it has implemented a stack canary (e.g. __stack_chk_guard)
564 config CC_HAS_STACKPROTECTOR_NONE
565 def_bool $(cc-option,-fno-stack-protector)
567 config STACKPROTECTOR
568 bool "Stack Protector buffer overflow detection"
569 depends on HAVE_STACKPROTECTOR
570 depends on $(cc-option,-fstack-protector)
573 This option turns on the "stack-protector" GCC feature. This
574 feature puts, at the beginning of functions, a canary value on
575 the stack just before the return address, and validates
576 the value just before actually returning. Stack based buffer
577 overflows (that need to overwrite this return address) now also
578 overwrite the canary, which gets detected and the attack is then
579 neutralized via a kernel panic.
581 Functions will have the stack-protector canary logic added if they
582 have an 8-byte or larger character array on the stack.
584 This feature requires gcc version 4.2 or above, or a distribution
585 gcc with the feature backported ("-fstack-protector").
587 On an x86 "defconfig" build, this feature adds canary checks to
588 about 3% of all kernel functions, which increases kernel code size
591 config STACKPROTECTOR_STRONG
592 bool "Strong Stack Protector"
593 depends on STACKPROTECTOR
594 depends on $(cc-option,-fstack-protector-strong)
597 Functions will have the stack-protector canary logic added in any
598 of the following conditions:
600 - local variable's address used as part of the right hand side of an
601 assignment or function argument
602 - local variable is an array (or union containing an array),
603 regardless of array type or length
604 - uses register local variables
606 This feature requires gcc version 4.9 or above, or a distribution
607 gcc with the feature backported ("-fstack-protector-strong").
609 On an x86 "defconfig" build, this feature adds canary checks to
610 about 20% of all kernel functions, which increases the kernel code
613 config HAVE_ARCH_WITHIN_STACK_FRAMES
616 An architecture should select this if it can walk the kernel stack
617 frames to determine if an object is part of either the arguments
618 or local variables (i.e. that it excludes saved return addresses,
619 and similar) by implementing an inline arch_within_stack_frames(),
620 which is used by CONFIG_HARDENED_USERCOPY.
622 config HAVE_CONTEXT_TRACKING
625 Provide kernel/user boundaries probes necessary for subsystems
626 that need it, such as userspace RCU extended quiescent state.
627 Syscalls need to be wrapped inside user_exit()-user_enter() through
628 the slow path using TIF_NOHZ flag. Exceptions handlers must be
629 wrapped as well. Irqs are already protected inside
630 rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on
631 irq exit still need to be protected.
633 config HAVE_VIRT_CPU_ACCOUNTING
636 config ARCH_HAS_SCALED_CPUTIME
639 config HAVE_VIRT_CPU_ACCOUNTING_GEN
643 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
644 Before enabling this option, arch code must be audited
645 to ensure there are no races in concurrent read/write of
646 cputime_t. For example, reading/writing 64-bit cputime_t on
647 some 32-bit arches may require multiple accesses, so proper
648 locking is needed to protect against concurrent accesses.
651 config HAVE_IRQ_TIME_ACCOUNTING
654 Archs need to ensure they use a high enough resolution clock to
655 support irq time accounting and then call enable_sched_clock_irqtime().
657 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
660 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
663 config HAVE_ARCH_HUGE_VMAP
666 config HAVE_ARCH_SOFT_DIRTY
669 config HAVE_MOD_ARCH_SPECIFIC
672 The arch uses struct mod_arch_specific to store data. Many arches
673 just need a simple module loader without arch specific data - those
674 should not enable this.
676 config MODULES_USE_ELF_RELA
679 Modules only use ELF RELA relocations. Modules with ELF REL
680 relocations will give an error.
682 config MODULES_USE_ELF_REL
685 Modules only use ELF REL relocations. Modules with ELF RELA
686 relocations will give an error.
688 config HAVE_IRQ_EXIT_ON_IRQ_STACK
691 Architecture doesn't only execute the irq handler on the irq stack
692 but also irq_exit(). This way we can process softirqs on this irq
693 stack instead of switching to a new one when we call __do_softirq()
694 in the end of an hardirq.
695 This spares a stack switch and improves cache usage on softirq
698 config PGTABLE_LEVELS
702 config ARCH_HAS_ELF_RANDOMIZE
705 An architecture supports choosing randomized locations for
706 stack, mmap, brk, and ET_DYN. Defined functions:
708 - arch_randomize_brk()
710 config HAVE_ARCH_MMAP_RND_BITS
713 An arch should select this symbol if it supports setting a variable
714 number of bits for use in establishing the base address for mmap
715 allocations, has MMU enabled and provides values for both:
716 - ARCH_MMAP_RND_BITS_MIN
717 - ARCH_MMAP_RND_BITS_MAX
719 config HAVE_EXIT_THREAD
722 An architecture implements exit_thread.
724 config ARCH_MMAP_RND_BITS_MIN
727 config ARCH_MMAP_RND_BITS_MAX
730 config ARCH_MMAP_RND_BITS_DEFAULT
733 config ARCH_MMAP_RND_BITS
734 int "Number of bits to use for ASLR of mmap base address" if EXPERT
735 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
736 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
737 default ARCH_MMAP_RND_BITS_MIN
738 depends on HAVE_ARCH_MMAP_RND_BITS
740 This value can be used to select the number of bits to use to
741 determine the random offset to the base address of vma regions
742 resulting from mmap allocations. This value will be bounded
743 by the architecture's minimum and maximum supported values.
745 This value can be changed after boot using the
746 /proc/sys/vm/mmap_rnd_bits tunable
748 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
751 An arch should select this symbol if it supports running applications
752 in compatibility mode, supports setting a variable number of bits for
753 use in establishing the base address for mmap allocations, has MMU
754 enabled and provides values for both:
755 - ARCH_MMAP_RND_COMPAT_BITS_MIN
756 - ARCH_MMAP_RND_COMPAT_BITS_MAX
758 config ARCH_MMAP_RND_COMPAT_BITS_MIN
761 config ARCH_MMAP_RND_COMPAT_BITS_MAX
764 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
767 config ARCH_MMAP_RND_COMPAT_BITS
768 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
769 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
770 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
771 default ARCH_MMAP_RND_COMPAT_BITS_MIN
772 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
774 This value can be used to select the number of bits to use to
775 determine the random offset to the base address of vma regions
776 resulting from mmap allocations for compatible applications This
777 value will be bounded by the architecture's minimum and maximum
780 This value can be changed after boot using the
781 /proc/sys/vm/mmap_rnd_compat_bits tunable
783 config HAVE_ARCH_COMPAT_MMAP_BASES
786 This allows 64bit applications to invoke 32-bit mmap() syscall
787 and vice-versa 32-bit applications to call 64-bit mmap().
788 Required for applications doing different bitness syscalls.
790 config HAVE_COPY_THREAD_TLS
793 Architecture provides copy_thread_tls to accept tls argument via
794 normal C parameter passing, rather than extracting the syscall
795 argument from pt_regs.
797 config HAVE_STACK_VALIDATION
800 Architecture supports the 'objtool check' host tool command, which
801 performs compile-time stack metadata validation.
803 config HAVE_RELIABLE_STACKTRACE
806 Architecture has a save_stack_trace_tsk_reliable() function which
807 only returns a stack trace if it can guarantee the trace is reliable.
809 config HAVE_ARCH_HASH
813 If this is set, the architecture provides an <asm/hash.h>
814 file which provides platform-specific implementations of some
815 functions in <linux/hash.h> or fs/namei.c.
823 config CLONE_BACKWARDS
826 Architecture has tls passed as the 4th argument of clone(2),
829 config CLONE_BACKWARDS2
832 Architecture has the first two arguments of clone(2) swapped.
834 config CLONE_BACKWARDS3
837 Architecture has tls passed as the 3rd argument of clone(2),
840 config ODD_RT_SIGACTION
843 Architecture has unusual rt_sigaction(2) arguments
845 config OLD_SIGSUSPEND
848 Architecture has old sigsuspend(2) syscall, of one-argument variety
850 config OLD_SIGSUSPEND3
853 Even weirder antique ABI - three-argument sigsuspend(2)
858 Architecture has old sigaction(2) syscall. Nope, not the same
859 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
860 but fairly different variant of sigaction(2), thanks to OSF/1
863 config COMPAT_OLD_SIGACTION
867 def_bool ARCH_HAS_64BIT_TIME
869 This should be selected by all architectures that need to support
870 new system calls with a 64-bit time_t. This is relevant on all 32-bit
871 architectures, and 64-bit architectures as part of compat syscall
874 config COMPAT_32BIT_TIME
875 def_bool (!64BIT && 64BIT_TIME) || COMPAT
877 This enables 32 bit time_t support in addition to 64 bit time_t support.
878 This is relevant on all 32-bit architectures, and 64-bit architectures
879 as part of compat syscall handling.
881 config ARCH_NO_COHERENT_DMA_MMAP
884 config ARCH_NO_PREEMPT
887 config CPU_NO_EFFICIENT_FFS
890 config HAVE_ARCH_VMAP_STACK
893 An arch should select this symbol if it can support kernel stacks
894 in vmalloc space. This means:
896 - vmalloc space must be large enough to hold many kernel stacks.
897 This may rule out many 32-bit architectures.
899 - Stacks in vmalloc space need to work reliably. For example, if
900 vmap page tables are created on demand, either this mechanism
901 needs to work while the stack points to a virtual address with
902 unpopulated page tables or arch code (switch_to() and switch_mm(),
903 most likely) needs to ensure that the stack's page table entries
904 are populated before running on a possibly unpopulated stack.
906 - If the stack overflows into a guard page, something reasonable
907 should happen. The definition of "reasonable" is flexible, but
908 instantly rebooting without logging anything would be unfriendly.
912 bool "Use a virtually-mapped stack"
913 depends on HAVE_ARCH_VMAP_STACK && !KASAN
915 Enable this if you want the use virtually-mapped kernel stacks
916 with guard pages. This causes kernel stack overflows to be
917 caught immediately rather than causing difficult-to-diagnose
920 This is presently incompatible with KASAN because KASAN expects
921 the stack to map directly to the KASAN shadow map using a formula
922 that is incorrect if the stack is in vmalloc space.
924 config ARCH_OPTIONAL_KERNEL_RWX
927 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
930 config ARCH_HAS_STRICT_KERNEL_RWX
933 config STRICT_KERNEL_RWX
934 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
935 depends on ARCH_HAS_STRICT_KERNEL_RWX
936 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
938 If this is set, kernel text and rodata memory will be made read-only,
939 and non-text memory will be made non-executable. This provides
940 protection against certain security exploits (e.g. executing the heap
943 These features are considered standard security practice these days.
944 You should say Y here in almost all cases.
946 config ARCH_HAS_STRICT_MODULE_RWX
949 config STRICT_MODULE_RWX
950 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
951 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
952 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
954 If this is set, module text and rodata memory will be made read-only,
955 and non-text memory will be made non-executable. This provides
956 protection against certain security exploits (e.g. writing to text)
958 # select if the architecture provides an asm/dma-direct.h header
959 config ARCH_HAS_PHYS_TO_DMA
962 config ARCH_HAS_REFCOUNT
965 An architecture selects this when it has implemented refcount_t
966 using open coded assembly primitives that provide an optimized
967 refcount_t implementation, possibly at the expense of some full
968 refcount state checks of CONFIG_REFCOUNT_FULL=y.
970 The refcount overflow check behavior, however, must be retained.
971 Catching overflows is the primary security concern for protecting
972 against bugs in reference counts.
975 bool "Perform full reference count validation at the expense of speed"
977 Enabling this switches the refcounting infrastructure from a fast
978 unchecked atomic_t implementation to a fully state checked
979 implementation, which can be (slightly) slower but provides protections
980 against various use-after-free conditions that can be used in
981 security flaw exploits.
983 source "kernel/gcov/Kconfig"