1 # SPDX-License-Identifier: GPL-2.0
3 # General architecture dependent options
17 tristate "OProfile system profiling"
19 depends on HAVE_OPROFILE
21 select RING_BUFFER_ALLOW_SWAP
23 OProfile is a profiling system capable of profiling the
24 whole system, include the kernel, kernel modules, libraries,
29 config OPROFILE_EVENT_MULTIPLEX
30 bool "OProfile multiplexing support (EXPERIMENTAL)"
32 depends on OPROFILE && X86
34 The number of hardware counters is limited. The multiplexing
35 feature enables OProfile to gather more events than counters
36 are provided by the hardware. This is realized by switching
37 between events at a user specified time interval.
44 config OPROFILE_NMI_TIMER
46 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
51 depends on HAVE_KPROBES
54 Kprobes allows you to trap at almost any kernel address and
55 execute a callback function. register_kprobe() establishes
56 a probepoint and specifies the callback. Kprobes is useful
57 for kernel debugging, non-intrusive instrumentation and testing.
61 bool "Optimize very unlikely/likely branches"
62 depends on HAVE_ARCH_JUMP_LABEL
64 This option enables a transparent branch optimization that
65 makes certain almost-always-true or almost-always-false branch
66 conditions even cheaper to execute within the kernel.
68 Certain performance-sensitive kernel code, such as trace points,
69 scheduler functionality, networking code and KVM have such
70 branches and include support for this optimization technique.
72 If it is detected that the compiler has support for "asm goto",
73 the kernel will compile such branches with just a nop
74 instruction. When the condition flag is toggled to true, the
75 nop will be converted to a jump instruction to execute the
76 conditional block of instructions.
78 This technique lowers overhead and stress on the branch prediction
79 of the processor and generally makes the kernel faster. The update
80 of the condition is slower, but those are always very rare.
82 ( On 32-bit x86, the necessary options added to the compiler
83 flags may increase the size of the kernel slightly. )
85 config STATIC_KEYS_SELFTEST
86 bool "Static key selftest"
89 Boot time self-test of the branch patching code.
93 depends on KPROBES && HAVE_OPTPROBES
94 select TASKS_RCU if PREEMPT
96 config KPROBES_ON_FTRACE
98 depends on KPROBES && HAVE_KPROBES_ON_FTRACE
99 depends on DYNAMIC_FTRACE_WITH_REGS
101 If function tracer is enabled and the arch supports full
102 passing of pt_regs to function tracing, then kprobes can
103 optimize on top of function tracing.
107 depends on ARCH_SUPPORTS_UPROBES
109 Uprobes is the user-space counterpart to kprobes: they
110 enable instrumentation applications (such as 'perf probe')
111 to establish unintrusive probes in user-space binaries and
112 libraries, by executing handler functions when the probes
113 are hit by user-space applications.
115 ( These probes come in the form of single-byte breakpoints,
116 managed by the kernel and kept transparent to the probed
119 config HAVE_64BIT_ALIGNED_ACCESS
120 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
122 Some architectures require 64 bit accesses to be 64 bit
123 aligned, which also requires structs containing 64 bit values
124 to be 64 bit aligned too. This includes some 32 bit
125 architectures which can do 64 bit accesses, as well as 64 bit
126 architectures without unaligned access.
128 This symbol should be selected by an architecture if 64 bit
129 accesses are required to be 64 bit aligned in this way even
130 though it is not a 64 bit architecture.
132 See Documentation/unaligned-memory-access.txt for more
133 information on the topic of unaligned memory accesses.
135 config HAVE_EFFICIENT_UNALIGNED_ACCESS
138 Some architectures are unable to perform unaligned accesses
139 without the use of get_unaligned/put_unaligned. Others are
140 unable to perform such accesses efficiently (e.g. trap on
141 unaligned access and require fixing it up in the exception
144 This symbol should be selected by an architecture if it can
145 perform unaligned accesses efficiently to allow different
146 code paths to be selected for these cases. Some network
147 drivers, for example, could opt to not fix up alignment
148 problems with received packets if doing so would not help
151 See Documentation/unaligned-memory-access.txt for more
152 information on the topic of unaligned memory accesses.
154 config ARCH_USE_BUILTIN_BSWAP
157 Modern versions of GCC (since 4.4) have builtin functions
158 for handling byte-swapping. Using these, instead of the old
159 inline assembler that the architecture code provides in the
160 __arch_bswapXX() macros, allows the compiler to see what's
161 happening and offers more opportunity for optimisation. In
162 particular, the compiler will be able to combine the byteswap
163 with a nearby load or store and use load-and-swap or
164 store-and-swap instructions if the architecture has them. It
165 should almost *never* result in code which is worse than the
166 hand-coded assembler in <asm/swab.h>. But just in case it
167 does, the use of the builtins is optional.
169 Any architecture with load-and-swap or store-and-swap
170 instructions should set this. And it shouldn't hurt to set it
171 on architectures that don't have such instructions.
175 depends on KPROBES && HAVE_KRETPROBES
177 config USER_RETURN_NOTIFIER
179 depends on HAVE_USER_RETURN_NOTIFIER
181 Provide a kernel-internal notification when a cpu is about to
184 config HAVE_IOREMAP_PROT
190 config HAVE_KRETPROBES
193 config HAVE_OPTPROBES
196 config HAVE_KPROBES_ON_FTRACE
203 # An arch should select this if it provides all these things:
205 # task_pt_regs() in asm/processor.h or asm/ptrace.h
206 # arch_has_single_step() if there is hardware single-step support
207 # arch_has_block_step() if there is hardware block-step support
208 # asm/syscall.h supplying asm-generic/syscall.h interface
209 # linux/regset.h user_regset interfaces
210 # CORE_DUMP_USE_REGSET #define'd in linux/elf.h
211 # TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit}
212 # TIF_NOTIFY_RESUME calls tracehook_notify_resume()
213 # signal delivery calls tracehook_signal_handler()
215 config HAVE_ARCH_TRACEHOOK
218 config HAVE_DMA_CONTIGUOUS
221 config GENERIC_SMP_IDLE_THREAD
224 config GENERIC_IDLE_POLL_SETUP
227 config ARCH_HAS_FORTIFY_SOURCE
230 An architecture should select this when it can successfully
231 build and run with CONFIG_FORTIFY_SOURCE.
233 # Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
234 config ARCH_HAS_SET_MEMORY
237 # Select if arch init_task initializer is different to init/init_task.c
238 config ARCH_INIT_TASK
241 # Select if arch has its private alloc_task_struct() function
242 config ARCH_TASK_STRUCT_ALLOCATOR
245 # Select if arch has its private alloc_thread_stack() function
246 config ARCH_THREAD_STACK_ALLOCATOR
249 # Select if arch wants to size task_struct dynamically via arch_task_struct_size:
250 config ARCH_WANTS_DYNAMIC_TASK_STRUCT
253 config HAVE_REGS_AND_STACK_ACCESS_API
256 This symbol should be selected by an architecure if it supports
257 the API needed to access registers and stack entries from pt_regs,
258 declared in asm/ptrace.h
259 For example the kprobes-based event tracer needs this API.
264 The <linux/clk.h> calls support software clock gating and
265 thus are a key power management tool on many systems.
267 config HAVE_DMA_API_DEBUG
270 config HAVE_HW_BREAKPOINT
272 depends on PERF_EVENTS
274 config HAVE_MIXED_BREAKPOINTS_REGS
276 depends on HAVE_HW_BREAKPOINT
278 Depending on the arch implementation of hardware breakpoints,
279 some of them have separate registers for data and instruction
280 breakpoints addresses, others have mixed registers to store
281 them but define the access type in a control register.
282 Select this option if your arch implements breakpoints under the
285 config HAVE_USER_RETURN_NOTIFIER
288 config HAVE_PERF_EVENTS_NMI
291 System hardware can generate an NMI using the perf event
292 subsystem. Also has support for calculating CPU cycle events
293 to determine how many clock cycles in a given period.
295 config HAVE_HARDLOCKUP_DETECTOR_PERF
297 depends on HAVE_PERF_EVENTS_NMI
299 The arch chooses to use the generic perf-NMI-based hardlockup
300 detector. Must define HAVE_PERF_EVENTS_NMI.
302 config HAVE_NMI_WATCHDOG
306 The arch provides a low level NMI watchdog. It provides
307 asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
309 config HAVE_HARDLOCKUP_DETECTOR_ARCH
311 select HAVE_NMI_WATCHDOG
313 The arch chooses to provide its own hardlockup detector, which is
314 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
315 interfaces and parameters provided by hardlockup detector subsystem.
317 config HAVE_PERF_REGS
320 Support selective register dumps for perf events. This includes
321 bit-mapping of each registers and a unique architecture id.
323 config HAVE_PERF_USER_STACK_DUMP
326 Support user stack dumps for perf event samples. This needs
327 access to the user stack pointer which is not unified across
330 config HAVE_ARCH_JUMP_LABEL
333 config HAVE_RCU_TABLE_FREE
336 config ARCH_HAVE_NMI_SAFE_CMPXCHG
339 config HAVE_ALIGNED_STRUCT_PAGE
342 This makes sure that struct pages are double word aligned and that
343 e.g. the SLUB allocator can perform double word atomic operations
344 on a struct page for better performance. However selecting this
345 might increase the size of a struct page by a word.
347 config HAVE_CMPXCHG_LOCAL
350 config HAVE_CMPXCHG_DOUBLE
353 config ARCH_WEAK_RELEASE_ACQUIRE
356 config ARCH_WANT_IPC_PARSE_VERSION
359 config ARCH_WANT_COMPAT_IPC_PARSE_VERSION
362 config ARCH_WANT_OLD_COMPAT_IPC
363 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
366 config HAVE_ARCH_SECCOMP_FILTER
369 An arch should select this symbol if it provides all of these things:
371 - syscall_get_arguments()
373 - syscall_set_return_value()
374 - SIGSYS siginfo_t support
375 - secure_computing is called from a ptrace_event()-safe context
376 - secure_computing return value is checked and a return value of -1
377 results in the system call being skipped immediately.
378 - seccomp syscall wired up
380 config SECCOMP_FILTER
382 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
384 Enable tasks to build secure computing environments defined
385 in terms of Berkeley Packet Filter programs which implement
386 task-defined system call filtering polices.
388 See Documentation/prctl/seccomp_filter.txt for details.
390 config HAVE_GCC_PLUGINS
393 An arch should select this symbol if it supports building with
396 menuconfig GCC_PLUGINS
398 depends on HAVE_GCC_PLUGINS
399 depends on !COMPILE_TEST
401 GCC plugins are loadable modules that provide extra features to the
402 compiler. They are useful for runtime instrumentation and static analysis.
404 See Documentation/gcc-plugins.txt for details.
406 config GCC_PLUGIN_CYC_COMPLEXITY
407 bool "Compute the cyclomatic complexity of a function" if EXPERT
408 depends on GCC_PLUGINS
409 depends on !COMPILE_TEST
411 The complexity M of a function's control flow graph is defined as:
415 E = the number of edges
416 N = the number of nodes
417 P = the number of connected components (exit nodes).
419 Enabling this plugin reports the complexity to stderr during the
420 build. It mainly serves as a simple example of how to create a
421 gcc plugin for the kernel.
423 config GCC_PLUGIN_SANCOV
425 depends on GCC_PLUGINS
427 This plugin inserts a __sanitizer_cov_trace_pc() call at the start of
428 basic blocks. It supports all gcc versions with plugin support (from
429 gcc-4.5 on). It is based on the commit "Add fuzzing coverage support"
430 by Dmitry Vyukov <dvyukov@google.com>.
432 config GCC_PLUGIN_LATENT_ENTROPY
433 bool "Generate some entropy during boot and runtime"
434 depends on GCC_PLUGINS
436 By saying Y here the kernel will instrument some kernel code to
437 extract some entropy from both original and artificially created
438 program state. This will help especially embedded systems where
439 there is little 'natural' source of entropy normally. The cost
440 is some slowdown of the boot process (about 0.5%) and fork and
443 Note that entropy extracted this way is not cryptographically
446 This plugin was ported from grsecurity/PaX. More information at:
447 * https://grsecurity.net/
448 * https://pax.grsecurity.net/
450 config GCC_PLUGIN_STRUCTLEAK
451 bool "Force initialization of variables containing userspace addresses"
452 depends on GCC_PLUGINS
454 This plugin zero-initializes any structures containing a
455 __user attribute. This can prevent some classes of information
458 This plugin was ported from grsecurity/PaX. More information at:
459 * https://grsecurity.net/
460 * https://pax.grsecurity.net/
462 config GCC_PLUGIN_STRUCTLEAK_BYREF_ALL
463 bool "Force initialize all struct type variables passed by reference"
464 depends on GCC_PLUGIN_STRUCTLEAK
466 Zero initialize any struct type local variable that may be passed by
467 reference without having been initialized.
469 config GCC_PLUGIN_STRUCTLEAK_VERBOSE
470 bool "Report forcefully initialized variables"
471 depends on GCC_PLUGIN_STRUCTLEAK
472 depends on !COMPILE_TEST
474 This option will cause a warning to be printed each time the
475 structleak plugin finds a variable it thinks needs to be
476 initialized. Since not all existing initializers are detected
477 by the plugin, this can produce false positive warnings.
479 config GCC_PLUGIN_RANDSTRUCT
480 bool "Randomize layout of sensitive kernel structures"
481 depends on GCC_PLUGINS
482 select MODVERSIONS if MODULES
484 If you say Y here, the layouts of structures that are entirely
485 function pointers (and have not been manually annotated with
486 __no_randomize_layout), or structures that have been explicitly
487 marked with __randomize_layout, will be randomized at compile-time.
488 This can introduce the requirement of an additional information
489 exposure vulnerability for exploits targeting these structure
492 Enabling this feature will introduce some performance impact,
493 slightly increase memory usage, and prevent the use of forensic
494 tools like Volatility against the system (unless the kernel
495 source tree isn't cleaned after kernel installation).
497 The seed used for compilation is located at
498 scripts/gcc-plgins/randomize_layout_seed.h. It remains after
499 a make clean to allow for external modules to be compiled with
500 the existing seed and will be removed by a make mrproper or
503 Note that the implementation requires gcc 4.7 or newer.
505 This plugin was ported from grsecurity/PaX. More information at:
506 * https://grsecurity.net/
507 * https://pax.grsecurity.net/
509 config GCC_PLUGIN_RANDSTRUCT_PERFORMANCE
510 bool "Use cacheline-aware structure randomization"
511 depends on GCC_PLUGIN_RANDSTRUCT
512 depends on !COMPILE_TEST
514 If you say Y here, the RANDSTRUCT randomization will make a
515 best effort at restricting randomization to cacheline-sized
516 groups of elements. It will further not randomize bitfields
517 in structures. This reduces the performance hit of RANDSTRUCT
518 at the cost of weakened randomization.
520 config HAVE_CC_STACKPROTECTOR
523 An arch should select this symbol if:
524 - its compiler supports the -fstack-protector option
525 - it has implemented a stack canary (e.g. __stack_chk_guard)
527 config CC_STACKPROTECTOR
530 Set when a stack-protector mode is enabled, so that the build
531 can enable kernel-side support for the GCC feature.
534 prompt "Stack Protector buffer overflow detection"
535 depends on HAVE_CC_STACKPROTECTOR
536 default CC_STACKPROTECTOR_NONE
538 This option turns on the "stack-protector" GCC feature. This
539 feature puts, at the beginning of functions, a canary value on
540 the stack just before the return address, and validates
541 the value just before actually returning. Stack based buffer
542 overflows (that need to overwrite this return address) now also
543 overwrite the canary, which gets detected and the attack is then
544 neutralized via a kernel panic.
546 config CC_STACKPROTECTOR_NONE
549 Disable "stack-protector" GCC feature.
551 config CC_STACKPROTECTOR_REGULAR
553 select CC_STACKPROTECTOR
555 Functions will have the stack-protector canary logic added if they
556 have an 8-byte or larger character array on the stack.
558 This feature requires gcc version 4.2 or above, or a distribution
559 gcc with the feature backported ("-fstack-protector").
561 On an x86 "defconfig" build, this feature adds canary checks to
562 about 3% of all kernel functions, which increases kernel code size
565 config CC_STACKPROTECTOR_STRONG
567 select CC_STACKPROTECTOR
569 Functions will have the stack-protector canary logic added in any
570 of the following conditions:
572 - local variable's address used as part of the right hand side of an
573 assignment or function argument
574 - local variable is an array (or union containing an array),
575 regardless of array type or length
576 - uses register local variables
578 This feature requires gcc version 4.9 or above, or a distribution
579 gcc with the feature backported ("-fstack-protector-strong").
581 On an x86 "defconfig" build, this feature adds canary checks to
582 about 20% of all kernel functions, which increases the kernel code
590 Select this if the architecture wants to use thin archives
591 instead of ld -r to create the built-in.o files.
593 config LD_DEAD_CODE_DATA_ELIMINATION
596 Select this if the architecture wants to do dead code and
597 data elimination with the linker by compiling with
598 -ffunction-sections -fdata-sections and linking with
601 This requires that the arch annotates or otherwise protects
602 its external entry points from being discarded. Linker scripts
603 must also merge .text.*, .data.*, and .bss.* correctly into
604 output sections. Care must be taken not to pull in unrelated
605 sections (e.g., '.text.init'). Typically '.' in section names
606 is used to distinguish them from label names / C identifiers.
611 config ARCH_SUPPORTS_LTO_CLANG
614 An architecture should select this option it supports:
615 - compiling with clang,
616 - compiling inline assembly with clang's integrated assembler,
617 - and linking with either lld or GNU gold w/ LLVMgold.
620 prompt "Link-Time Optimization (LTO) (EXPERIMENTAL)"
623 This option turns on Link-Time Optimization (LTO).
629 bool "Use clang Link Time Optimization (LTO) (EXPERIMENTAL)"
630 depends on ARCH_SUPPORTS_LTO_CLANG
631 depends on !FTRACE_MCOUNT_RECORD || HAVE_C_RECORDMCOUNT
634 select LD_DEAD_CODE_DATA_ELIMINATION
636 This option enables clang's Link Time Optimization (LTO), which allows
637 the compiler to optimize the kernel globally at link time. If you
638 enable this option, the compiler generates LLVM IR instead of object
639 files, and the actual compilation from IR occurs at the LTO link step,
640 which may take several minutes.
642 If you select this option, you must compile the kernel with clang >=
643 5.0 (make CC=clang) and GNU gold from binutils >= 2.27, and have the
644 LLVMgold plug-in in LD_LIBRARY_PATH.
651 config CFI_PERMISSIVE
652 bool "Use CFI in permissive mode"
655 When selected, Control Flow Integrity (CFI) violations result in a
656 warning instead of a kernel panic. This option is useful for finding
657 CFI violations in drivers during development.
660 bool "Use clang Control Flow Integrity (CFI) (EXPERIMENTAL)"
665 This option enables clang Control Flow Integrity (CFI), which adds
666 runtime checking for indirect function calls.
668 config CFI_CLANG_SHADOW
669 bool "Use CFI shadow to speed up cross-module checks"
673 If you select this option, the kernel builds a fast look-up table of
674 CFI check functions in loaded modules to reduce overhead.
676 config HAVE_ARCH_WITHIN_STACK_FRAMES
679 An architecture should select this if it can walk the kernel stack
680 frames to determine if an object is part of either the arguments
681 or local variables (i.e. that it excludes saved return addresses,
682 and similar) by implementing an inline arch_within_stack_frames(),
683 which is used by CONFIG_HARDENED_USERCOPY.
685 config HAVE_CONTEXT_TRACKING
688 Provide kernel/user boundaries probes necessary for subsystems
689 that need it, such as userspace RCU extended quiescent state.
690 Syscalls need to be wrapped inside user_exit()-user_enter() through
691 the slow path using TIF_NOHZ flag. Exceptions handlers must be
692 wrapped as well. Irqs are already protected inside
693 rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on
694 irq exit still need to be protected.
696 config HAVE_VIRT_CPU_ACCOUNTING
699 config ARCH_HAS_SCALED_CPUTIME
702 config HAVE_VIRT_CPU_ACCOUNTING_GEN
706 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
707 Before enabling this option, arch code must be audited
708 to ensure there are no races in concurrent read/write of
709 cputime_t. For example, reading/writing 64-bit cputime_t on
710 some 32-bit arches may require multiple accesses, so proper
711 locking is needed to protect against concurrent accesses.
714 config HAVE_IRQ_TIME_ACCOUNTING
717 Archs need to ensure they use a high enough resolution clock to
718 support irq time accounting and then call enable_sched_clock_irqtime().
720 config HAVE_ARCH_TRANSPARENT_HUGEPAGE
723 config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
726 config HAVE_ARCH_HUGE_VMAP
729 config HAVE_ARCH_SOFT_DIRTY
732 config HAVE_MOD_ARCH_SPECIFIC
735 The arch uses struct mod_arch_specific to store data. Many arches
736 just need a simple module loader without arch specific data - those
737 should not enable this.
739 config MODULES_USE_ELF_RELA
742 Modules only use ELF RELA relocations. Modules with ELF REL
743 relocations will give an error.
745 config MODULES_USE_ELF_REL
748 Modules only use ELF REL relocations. Modules with ELF RELA
749 relocations will give an error.
751 config HAVE_UNDERSCORE_SYMBOL_PREFIX
754 Some architectures generate an _ in front of C symbols; things like
755 module loading and assembly files need to know about this.
757 config HAVE_IRQ_EXIT_ON_IRQ_STACK
760 Architecture doesn't only execute the irq handler on the irq stack
761 but also irq_exit(). This way we can process softirqs on this irq
762 stack instead of switching to a new one when we call __do_softirq()
763 in the end of an hardirq.
764 This spares a stack switch and improves cache usage on softirq
767 config PGTABLE_LEVELS
771 config ARCH_HAS_ELF_RANDOMIZE
774 An architecture supports choosing randomized locations for
775 stack, mmap, brk, and ET_DYN. Defined functions:
777 - arch_randomize_brk()
779 config HAVE_ARCH_MMAP_RND_BITS
782 An arch should select this symbol if it supports setting a variable
783 number of bits for use in establishing the base address for mmap
784 allocations, has MMU enabled and provides values for both:
785 - ARCH_MMAP_RND_BITS_MIN
786 - ARCH_MMAP_RND_BITS_MAX
788 config HAVE_EXIT_THREAD
791 An architecture implements exit_thread.
793 config ARCH_MMAP_RND_BITS_MIN
796 config ARCH_MMAP_RND_BITS_MAX
799 config ARCH_MMAP_RND_BITS_DEFAULT
802 config ARCH_MMAP_RND_BITS
803 int "Number of bits to use for ASLR of mmap base address" if EXPERT
804 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
805 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
806 default ARCH_MMAP_RND_BITS_MIN
807 depends on HAVE_ARCH_MMAP_RND_BITS
809 This value can be used to select the number of bits to use to
810 determine the random offset to the base address of vma regions
811 resulting from mmap allocations. This value will be bounded
812 by the architecture's minimum and maximum supported values.
814 This value can be changed after boot using the
815 /proc/sys/vm/mmap_rnd_bits tunable
817 config HAVE_ARCH_MMAP_RND_COMPAT_BITS
820 An arch should select this symbol if it supports running applications
821 in compatibility mode, supports setting a variable number of bits for
822 use in establishing the base address for mmap allocations, has MMU
823 enabled and provides values for both:
824 - ARCH_MMAP_RND_COMPAT_BITS_MIN
825 - ARCH_MMAP_RND_COMPAT_BITS_MAX
827 config ARCH_MMAP_RND_COMPAT_BITS_MIN
830 config ARCH_MMAP_RND_COMPAT_BITS_MAX
833 config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
836 config ARCH_MMAP_RND_COMPAT_BITS
837 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
838 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
839 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
840 default ARCH_MMAP_RND_COMPAT_BITS_MIN
841 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
843 This value can be used to select the number of bits to use to
844 determine the random offset to the base address of vma regions
845 resulting from mmap allocations for compatible applications This
846 value will be bounded by the architecture's minimum and maximum
849 This value can be changed after boot using the
850 /proc/sys/vm/mmap_rnd_compat_bits tunable
852 config HAVE_ARCH_COMPAT_MMAP_BASES
855 This allows 64bit applications to invoke 32-bit mmap() syscall
856 and vice-versa 32-bit applications to call 64-bit mmap().
857 Required for applications doing different bitness syscalls.
859 config HAVE_COPY_THREAD_TLS
862 Architecture provides copy_thread_tls to accept tls argument via
863 normal C parameter passing, rather than extracting the syscall
864 argument from pt_regs.
866 config HAVE_STACK_VALIDATION
869 Architecture supports the 'objtool check' host tool command, which
870 performs compile-time stack metadata validation.
872 config HAVE_RELIABLE_STACKTRACE
875 Architecture has a save_stack_trace_tsk_reliable() function which
876 only returns a stack trace if it can guarantee the trace is reliable.
878 config HAVE_ARCH_HASH
882 If this is set, the architecture provides an <asm/hash.h>
883 file which provides platform-specific implementations of some
884 functions in <linux/hash.h> or fs/namei.c.
892 config CLONE_BACKWARDS
895 Architecture has tls passed as the 4th argument of clone(2),
898 config CLONE_BACKWARDS2
901 Architecture has the first two arguments of clone(2) swapped.
903 config CLONE_BACKWARDS3
906 Architecture has tls passed as the 3rd argument of clone(2),
909 config ODD_RT_SIGACTION
912 Architecture has unusual rt_sigaction(2) arguments
914 config OLD_SIGSUSPEND
917 Architecture has old sigsuspend(2) syscall, of one-argument variety
919 config OLD_SIGSUSPEND3
922 Even weirder antique ABI - three-argument sigsuspend(2)
927 Architecture has old sigaction(2) syscall. Nope, not the same
928 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
929 but fairly different variant of sigaction(2), thanks to OSF/1
932 config COMPAT_OLD_SIGACTION
935 config ARCH_NO_COHERENT_DMA_MMAP
938 config CPU_NO_EFFICIENT_FFS
941 config HAVE_ARCH_VMAP_STACK
944 An arch should select this symbol if it can support kernel stacks
945 in vmalloc space. This means:
947 - vmalloc space must be large enough to hold many kernel stacks.
948 This may rule out many 32-bit architectures.
950 - Stacks in vmalloc space need to work reliably. For example, if
951 vmap page tables are created on demand, either this mechanism
952 needs to work while the stack points to a virtual address with
953 unpopulated page tables or arch code (switch_to() and switch_mm(),
954 most likely) needs to ensure that the stack's page table entries
955 are populated before running on a possibly unpopulated stack.
957 - If the stack overflows into a guard page, something reasonable
958 should happen. The definition of "reasonable" is flexible, but
959 instantly rebooting without logging anything would be unfriendly.
963 bool "Use a virtually-mapped stack"
964 depends on HAVE_ARCH_VMAP_STACK && !KASAN
966 Enable this if you want the use virtually-mapped kernel stacks
967 with guard pages. This causes kernel stack overflows to be
968 caught immediately rather than causing difficult-to-diagnose
971 This is presently incompatible with KASAN because KASAN expects
972 the stack to map directly to the KASAN shadow map using a formula
973 that is incorrect if the stack is in vmalloc space.
975 config ARCH_OPTIONAL_KERNEL_RWX
978 config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
981 config ARCH_HAS_STRICT_KERNEL_RWX
984 config STRICT_KERNEL_RWX
985 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
986 depends on ARCH_HAS_STRICT_KERNEL_RWX
987 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
989 If this is set, kernel text and rodata memory will be made read-only,
990 and non-text memory will be made non-executable. This provides
991 protection against certain security exploits (e.g. executing the heap
994 These features are considered standard security practice these days.
995 You should say Y here in almost all cases.
997 config ARCH_HAS_STRICT_MODULE_RWX
1000 config STRICT_MODULE_RWX
1001 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
1002 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
1003 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
1005 If this is set, module text and rodata memory will be made read-only,
1006 and non-text memory will be made non-executable. This provides
1007 protection against certain security exploits (e.g. writing to text)
1009 config ARCH_HAS_REFCOUNT
1012 An architecture selects this when it has implemented refcount_t
1013 using open coded assembly primitives that provide an optimized
1014 refcount_t implementation, possibly at the expense of some full
1015 refcount state checks of CONFIG_REFCOUNT_FULL=y.
1017 The refcount overflow check behavior, however, must be retained.
1018 Catching overflows is the primary security concern for protecting
1019 against bugs in reference counts.
1021 config REFCOUNT_FULL
1022 bool "Perform full reference count validation at the expense of speed"
1024 Enabling this switches the refcounting infrastructure from a fast
1025 unchecked atomic_t implementation to a fully state checked
1026 implementation, which can be (slightly) slower but provides protections
1027 against various use-after-free conditions that can be used in
1028 security flaw exploits.
1030 source "kernel/gcov/Kconfig"