1# SPDX-License-Identifier: GPL-2.0 2# 3# General architecture dependent options 4# 5 6# 7# Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can 8# override the default values in this file. 9# 10source "arch/$(SRCARCH)/Kconfig" 11 12config ARCH_CONFIGURES_CPU_MITIGATIONS 13 bool 14 15if !ARCH_CONFIGURES_CPU_MITIGATIONS 16config CPU_MITIGATIONS 17 def_bool y 18endif 19 20menu "General architecture-dependent options" 21 22config ARCH_HAS_SUBPAGE_FAULTS 23 bool 24 help 25 Select if the architecture can check permissions at sub-page 26 granularity (e.g. arm64 MTE). The probe_user_*() functions 27 must be implemented. 28 29config HOTPLUG_SMT 30 bool 31 32config SMT_NUM_THREADS_DYNAMIC 33 bool 34 35# Selected by HOTPLUG_CORE_SYNC_DEAD or HOTPLUG_CORE_SYNC_FULL 36config HOTPLUG_CORE_SYNC 37 bool 38 39# Basic CPU dead synchronization selected by architecture 40config HOTPLUG_CORE_SYNC_DEAD 41 bool 42 select HOTPLUG_CORE_SYNC 43 44# Full CPU synchronization with alive state selected by architecture 45config HOTPLUG_CORE_SYNC_FULL 46 bool 47 select HOTPLUG_CORE_SYNC_DEAD if HOTPLUG_CPU 48 select HOTPLUG_CORE_SYNC 49 50config HOTPLUG_SPLIT_STARTUP 51 bool 52 select HOTPLUG_CORE_SYNC_FULL 53 54config HOTPLUG_PARALLEL 55 bool 56 select HOTPLUG_SPLIT_STARTUP 57 58config GENERIC_ENTRY 59 bool 60 61config KPROBES 62 bool "Kprobes" 63 depends on HAVE_KPROBES 64 select KALLSYMS 65 select EXECMEM 66 select NEED_TASKS_RCU 67 help 68 Kprobes allows you to trap at almost any kernel address and 69 execute a callback function. register_kprobe() establishes 70 a probepoint and specifies the callback. Kprobes is useful 71 for kernel debugging, non-intrusive instrumentation and testing. 72 If in doubt, say "N". 73 74config JUMP_LABEL 75 bool "Optimize very unlikely/likely branches" 76 depends on HAVE_ARCH_JUMP_LABEL 77 select OBJTOOL if HAVE_JUMP_LABEL_HACK 78 help 79 This option enables a transparent branch optimization that 80 makes certain almost-always-true or almost-always-false branch 81 conditions even cheaper to execute within the kernel. 82 83 Certain performance-sensitive kernel code, such as trace points, 84 scheduler functionality, networking code and KVM have such 85 branches and include support for this optimization technique. 86 87 If it is detected that the compiler has support for "asm goto", 88 the kernel will compile such branches with just a nop 89 instruction. When the condition flag is toggled to true, the 90 nop will be converted to a jump instruction to execute the 91 conditional block of instructions. 92 93 This technique lowers overhead and stress on the branch prediction 94 of the processor and generally makes the kernel faster. The update 95 of the condition is slower, but those are always very rare. 96 97 ( On 32-bit x86, the necessary options added to the compiler 98 flags may increase the size of the kernel slightly. ) 99 100config STATIC_KEYS_SELFTEST 101 bool "Static key selftest" 102 depends on JUMP_LABEL 103 help 104 Boot time self-test of the branch patching code. 105 106config STATIC_CALL_SELFTEST 107 bool "Static call selftest" 108 depends on HAVE_STATIC_CALL 109 help 110 Boot time self-test of the call patching code. 111 112config OPTPROBES 113 def_bool y 114 depends on KPROBES && HAVE_OPTPROBES 115 select NEED_TASKS_RCU 116 117config KPROBES_ON_FTRACE 118 def_bool y 119 depends on KPROBES && HAVE_KPROBES_ON_FTRACE 120 depends on DYNAMIC_FTRACE_WITH_REGS 121 help 122 If function tracer is enabled and the arch supports full 123 passing of pt_regs to function tracing, then kprobes can 124 optimize on top of function tracing. 125 126config UPROBES 127 def_bool n 128 depends on ARCH_SUPPORTS_UPROBES 129 help 130 Uprobes is the user-space counterpart to kprobes: they 131 enable instrumentation applications (such as 'perf probe') 132 to establish unintrusive probes in user-space binaries and 133 libraries, by executing handler functions when the probes 134 are hit by user-space applications. 135 136 ( These probes come in the form of single-byte breakpoints, 137 managed by the kernel and kept transparent to the probed 138 application. ) 139 140config HAVE_64BIT_ALIGNED_ACCESS 141 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS 142 help 143 Some architectures require 64 bit accesses to be 64 bit 144 aligned, which also requires structs containing 64 bit values 145 to be 64 bit aligned too. This includes some 32 bit 146 architectures which can do 64 bit accesses, as well as 64 bit 147 architectures without unaligned access. 148 149 This symbol should be selected by an architecture if 64 bit 150 accesses are required to be 64 bit aligned in this way even 151 though it is not a 64 bit architecture. 152 153 See Documentation/core-api/unaligned-memory-access.rst for 154 more information on the topic of unaligned memory accesses. 155 156config HAVE_EFFICIENT_UNALIGNED_ACCESS 157 bool 158 help 159 Some architectures are unable to perform unaligned accesses 160 without the use of get_unaligned/put_unaligned. Others are 161 unable to perform such accesses efficiently (e.g. trap on 162 unaligned access and require fixing it up in the exception 163 handler.) 164 165 This symbol should be selected by an architecture if it can 166 perform unaligned accesses efficiently to allow different 167 code paths to be selected for these cases. Some network 168 drivers, for example, could opt to not fix up alignment 169 problems with received packets if doing so would not help 170 much. 171 172 See Documentation/core-api/unaligned-memory-access.rst for more 173 information on the topic of unaligned memory accesses. 174 175config ARCH_USE_BUILTIN_BSWAP 176 bool 177 help 178 Modern versions of GCC (since 4.4) have builtin functions 179 for handling byte-swapping. Using these, instead of the old 180 inline assembler that the architecture code provides in the 181 __arch_bswapXX() macros, allows the compiler to see what's 182 happening and offers more opportunity for optimisation. In 183 particular, the compiler will be able to combine the byteswap 184 with a nearby load or store and use load-and-swap or 185 store-and-swap instructions if the architecture has them. It 186 should almost *never* result in code which is worse than the 187 hand-coded assembler in <asm/swab.h>. But just in case it 188 does, the use of the builtins is optional. 189 190 Any architecture with load-and-swap or store-and-swap 191 instructions should set this. And it shouldn't hurt to set it 192 on architectures that don't have such instructions. 193 194config KRETPROBES 195 def_bool y 196 depends on KPROBES && (HAVE_KRETPROBES || HAVE_RETHOOK) 197 198config KRETPROBE_ON_RETHOOK 199 def_bool y 200 depends on HAVE_RETHOOK 201 depends on KRETPROBES 202 select RETHOOK 203 204config USER_RETURN_NOTIFIER 205 bool 206 depends on HAVE_USER_RETURN_NOTIFIER 207 help 208 Provide a kernel-internal notification when a cpu is about to 209 switch to user mode. 210 211config HAVE_IOREMAP_PROT 212 bool 213 214config HAVE_KPROBES 215 bool 216 217config HAVE_KRETPROBES 218 bool 219 220config HAVE_OPTPROBES 221 bool 222 223config HAVE_KPROBES_ON_FTRACE 224 bool 225 226config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE 227 bool 228 help 229 Since kretprobes modifies return address on the stack, the 230 stacktrace may see the kretprobe trampoline address instead 231 of correct one. If the architecture stacktrace code and 232 unwinder can adjust such entries, select this configuration. 233 234config HAVE_FUNCTION_ERROR_INJECTION 235 bool 236 237config HAVE_NMI 238 bool 239 240config HAVE_FUNCTION_DESCRIPTORS 241 bool 242 243config TRACE_IRQFLAGS_SUPPORT 244 bool 245 246config TRACE_IRQFLAGS_NMI_SUPPORT 247 bool 248 249# 250# An arch should select this if it provides all these things: 251# 252# task_pt_regs() in asm/processor.h or asm/ptrace.h 253# arch_has_single_step() if there is hardware single-step support 254# arch_has_block_step() if there is hardware block-step support 255# asm/syscall.h supplying asm-generic/syscall.h interface 256# linux/regset.h user_regset interfaces 257# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 258# TIF_SYSCALL_TRACE calls ptrace_report_syscall_{entry,exit} 259# TIF_NOTIFY_RESUME calls resume_user_mode_work() 260# 261config HAVE_ARCH_TRACEHOOK 262 bool 263 264config HAVE_DMA_CONTIGUOUS 265 bool 266 267config GENERIC_SMP_IDLE_THREAD 268 bool 269 270config GENERIC_IDLE_POLL_SETUP 271 bool 272 273config ARCH_HAS_FORTIFY_SOURCE 274 bool 275 help 276 An architecture should select this when it can successfully 277 build and run with CONFIG_FORTIFY_SOURCE. 278 279# 280# Select if the arch provides a historic keepinit alias for the retain_initrd 281# command line option 282# 283config ARCH_HAS_KEEPINITRD 284 bool 285 286# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h 287config ARCH_HAS_SET_MEMORY 288 bool 289 290# Select if arch has all set_direct_map_invalid/default() functions 291config ARCH_HAS_SET_DIRECT_MAP 292 bool 293 294# 295# Select if the architecture provides the arch_dma_set_uncached symbol to 296# either provide an uncached segment alias for a DMA allocation, or 297# to remap the page tables in place. 298# 299config ARCH_HAS_DMA_SET_UNCACHED 300 bool 301 302# 303# Select if the architectures provides the arch_dma_clear_uncached symbol 304# to undo an in-place page table remap for uncached access. 305# 306config ARCH_HAS_DMA_CLEAR_UNCACHED 307 bool 308 309config ARCH_HAS_CPU_FINALIZE_INIT 310 bool 311 312# The architecture has a per-task state that includes the mm's PASID 313config ARCH_HAS_CPU_PASID 314 bool 315 select IOMMU_MM_DATA 316 317config HAVE_ARCH_THREAD_STRUCT_WHITELIST 318 bool 319 help 320 An architecture should select this to provide hardened usercopy 321 knowledge about what region of the thread_struct should be 322 whitelisted for copying to userspace. Normally this is only the 323 FPU registers. Specifically, arch_thread_struct_whitelist() 324 should be implemented. Without this, the entire thread_struct 325 field in task_struct will be left whitelisted. 326 327# Select if arch wants to size task_struct dynamically via arch_task_struct_size: 328config ARCH_WANTS_DYNAMIC_TASK_STRUCT 329 bool 330 331config ARCH_WANTS_NO_INSTR 332 bool 333 help 334 An architecture should select this if the noinstr macro is being used on 335 functions to denote that the toolchain should avoid instrumenting such 336 functions and is required for correctness. 337 338config ARCH_32BIT_OFF_T 339 bool 340 depends on !64BIT 341 help 342 All new 32-bit architectures should have 64-bit off_t type on 343 userspace side which corresponds to the loff_t kernel type. This 344 is the requirement for modern ABIs. Some existing architectures 345 still support 32-bit off_t. This option is enabled for all such 346 architectures explicitly. 347 348# Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat 349config ARCH_32BIT_USTAT_F_TINODE 350 bool 351 352config HAVE_ASM_MODVERSIONS 353 bool 354 help 355 This symbol should be selected by an architecture if it provides 356 <asm/asm-prototypes.h> to support the module versioning for symbols 357 exported from assembly code. 358 359config HAVE_REGS_AND_STACK_ACCESS_API 360 bool 361 help 362 This symbol should be selected by an architecture if it supports 363 the API needed to access registers and stack entries from pt_regs, 364 declared in asm/ptrace.h 365 For example the kprobes-based event tracer needs this API. 366 367config HAVE_RSEQ 368 bool 369 depends on HAVE_REGS_AND_STACK_ACCESS_API 370 help 371 This symbol should be selected by an architecture if it 372 supports an implementation of restartable sequences. 373 374config HAVE_RUST 375 bool 376 help 377 This symbol should be selected by an architecture if it 378 supports Rust. 379 380config HAVE_FUNCTION_ARG_ACCESS_API 381 bool 382 help 383 This symbol should be selected by an architecture if it supports 384 the API needed to access function arguments from pt_regs, 385 declared in asm/ptrace.h 386 387config HAVE_HW_BREAKPOINT 388 bool 389 depends on PERF_EVENTS 390 391config HAVE_MIXED_BREAKPOINTS_REGS 392 bool 393 depends on HAVE_HW_BREAKPOINT 394 help 395 Depending on the arch implementation of hardware breakpoints, 396 some of them have separate registers for data and instruction 397 breakpoints addresses, others have mixed registers to store 398 them but define the access type in a control register. 399 Select this option if your arch implements breakpoints under the 400 latter fashion. 401 402config HAVE_USER_RETURN_NOTIFIER 403 bool 404 405config HAVE_PERF_EVENTS_NMI 406 bool 407 help 408 System hardware can generate an NMI using the perf event 409 subsystem. Also has support for calculating CPU cycle events 410 to determine how many clock cycles in a given period. 411 412config HAVE_HARDLOCKUP_DETECTOR_PERF 413 bool 414 depends on HAVE_PERF_EVENTS_NMI 415 help 416 The arch chooses to use the generic perf-NMI-based hardlockup 417 detector. Must define HAVE_PERF_EVENTS_NMI. 418 419config HAVE_HARDLOCKUP_DETECTOR_ARCH 420 bool 421 help 422 The arch provides its own hardlockup detector implementation instead 423 of the generic ones. 424 425 It uses the same command line parameters, and sysctl interface, 426 as the generic hardlockup detectors. 427 428config HAVE_PERF_REGS 429 bool 430 help 431 Support selective register dumps for perf events. This includes 432 bit-mapping of each registers and a unique architecture id. 433 434config HAVE_PERF_USER_STACK_DUMP 435 bool 436 help 437 Support user stack dumps for perf event samples. This needs 438 access to the user stack pointer which is not unified across 439 architectures. 440 441config HAVE_ARCH_JUMP_LABEL 442 bool 443 444config HAVE_ARCH_JUMP_LABEL_RELATIVE 445 bool 446 447config MMU_GATHER_TABLE_FREE 448 bool 449 450config MMU_GATHER_RCU_TABLE_FREE 451 bool 452 select MMU_GATHER_TABLE_FREE 453 454config MMU_GATHER_PAGE_SIZE 455 bool 456 457config MMU_GATHER_NO_RANGE 458 bool 459 select MMU_GATHER_MERGE_VMAS 460 461config MMU_GATHER_NO_FLUSH_CACHE 462 bool 463 464config MMU_GATHER_MERGE_VMAS 465 bool 466 467config MMU_GATHER_NO_GATHER 468 bool 469 depends on MMU_GATHER_TABLE_FREE 470 471config ARCH_WANT_IRQS_OFF_ACTIVATE_MM 472 bool 473 help 474 Temporary select until all architectures can be converted to have 475 irqs disabled over activate_mm. Architectures that do IPI based TLB 476 shootdowns should enable this. 477 478# Use normal mm refcounting for MMU_LAZY_TLB kernel thread references. 479# MMU_LAZY_TLB_REFCOUNT=n can improve the scalability of context switching 480# to/from kernel threads when the same mm is running on a lot of CPUs (a large 481# multi-threaded application), by reducing contention on the mm refcount. 482# 483# This can be disabled if the architecture ensures no CPUs are using an mm as a 484# "lazy tlb" beyond its final refcount (i.e., by the time __mmdrop frees the mm 485# or its kernel page tables). This could be arranged by arch_exit_mmap(), or 486# final exit(2) TLB flush, for example. 487# 488# To implement this, an arch *must*: 489# Ensure the _lazy_tlb variants of mmgrab/mmdrop are used when manipulating 490# the lazy tlb reference of a kthread's ->active_mm (non-arch code has been 491# converted already). 492config MMU_LAZY_TLB_REFCOUNT 493 def_bool y 494 depends on !MMU_LAZY_TLB_SHOOTDOWN 495 496# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an 497# mm as a lazy tlb beyond its last reference count, by shooting down these 498# users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may 499# be using the mm as a lazy tlb, so that they may switch themselves to using 500# init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs 501# may be using mm as a lazy tlb mm. 502# 503# To implement this, an arch *must*: 504# - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains 505# at least all possible CPUs in which the mm is lazy. 506# - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above). 507config MMU_LAZY_TLB_SHOOTDOWN 508 bool 509 510config ARCH_HAVE_NMI_SAFE_CMPXCHG 511 bool 512 513config ARCH_HAVE_EXTRA_ELF_NOTES 514 bool 515 help 516 An architecture should select this in order to enable adding an 517 arch-specific ELF note section to core files. It must provide two 518 functions: elf_coredump_extra_notes_size() and 519 elf_coredump_extra_notes_write() which are invoked by the ELF core 520 dumper. 521 522config ARCH_HAS_NMI_SAFE_THIS_CPU_OPS 523 bool 524 525config HAVE_ALIGNED_STRUCT_PAGE 526 bool 527 help 528 This makes sure that struct pages are double word aligned and that 529 e.g. the SLUB allocator can perform double word atomic operations 530 on a struct page for better performance. However selecting this 531 might increase the size of a struct page by a word. 532 533config HAVE_CMPXCHG_LOCAL 534 bool 535 536config HAVE_CMPXCHG_DOUBLE 537 bool 538 539config ARCH_WEAK_RELEASE_ACQUIRE 540 bool 541 542config ARCH_WANT_IPC_PARSE_VERSION 543 bool 544 545config ARCH_WANT_COMPAT_IPC_PARSE_VERSION 546 bool 547 548config ARCH_WANT_OLD_COMPAT_IPC 549 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 550 bool 551 552config HAVE_ARCH_SECCOMP 553 bool 554 help 555 An arch should select this symbol to support seccomp mode 1 (the fixed 556 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn, 557 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment: 558 - __NR_seccomp_read_32 559 - __NR_seccomp_write_32 560 - __NR_seccomp_exit_32 561 - __NR_seccomp_sigreturn_32 562 563config HAVE_ARCH_SECCOMP_FILTER 564 bool 565 select HAVE_ARCH_SECCOMP 566 help 567 An arch should select this symbol if it provides all of these things: 568 - all the requirements for HAVE_ARCH_SECCOMP 569 - syscall_get_arch() 570 - syscall_get_arguments() 571 - syscall_rollback() 572 - syscall_set_return_value() 573 - SIGSYS siginfo_t support 574 - secure_computing is called from a ptrace_event()-safe context 575 - secure_computing return value is checked and a return value of -1 576 results in the system call being skipped immediately. 577 - seccomp syscall wired up 578 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE, 579 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If 580 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too. 581 582config SECCOMP 583 prompt "Enable seccomp to safely execute untrusted bytecode" 584 def_bool y 585 depends on HAVE_ARCH_SECCOMP 586 help 587 This kernel feature is useful for number crunching applications 588 that may need to handle untrusted bytecode during their 589 execution. By using pipes or other transports made available 590 to the process as file descriptors supporting the read/write 591 syscalls, it's possible to isolate those applications in their 592 own address space using seccomp. Once seccomp is enabled via 593 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be 594 disabled and the task is only allowed to execute a few safe 595 syscalls defined by each seccomp mode. 596 597 If unsure, say Y. 598 599config SECCOMP_FILTER 600 def_bool y 601 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 602 help 603 Enable tasks to build secure computing environments defined 604 in terms of Berkeley Packet Filter programs which implement 605 task-defined system call filtering polices. 606 607 See Documentation/userspace-api/seccomp_filter.rst for details. 608 609config SECCOMP_CACHE_DEBUG 610 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache" 611 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR 612 depends on PROC_FS 613 help 614 This enables the /proc/pid/seccomp_cache interface to monitor 615 seccomp cache data. The file format is subject to change. Reading 616 the file requires CAP_SYS_ADMIN. 617 618 This option is for debugging only. Enabling presents the risk that 619 an adversary may be able to infer the seccomp filter logic. 620 621 If unsure, say N. 622 623config HAVE_ARCH_STACKLEAK 624 bool 625 help 626 An architecture should select this if it has the code which 627 fills the used part of the kernel stack with the STACKLEAK_POISON 628 value before returning from system calls. 629 630config HAVE_STACKPROTECTOR 631 bool 632 help 633 An arch should select this symbol if: 634 - it has implemented a stack canary (e.g. __stack_chk_guard) 635 636config STACKPROTECTOR 637 bool "Stack Protector buffer overflow detection" 638 depends on HAVE_STACKPROTECTOR 639 depends on $(cc-option,-fstack-protector) 640 default y 641 help 642 This option turns on the "stack-protector" GCC feature. This 643 feature puts, at the beginning of functions, a canary value on 644 the stack just before the return address, and validates 645 the value just before actually returning. Stack based buffer 646 overflows (that need to overwrite this return address) now also 647 overwrite the canary, which gets detected and the attack is then 648 neutralized via a kernel panic. 649 650 Functions will have the stack-protector canary logic added if they 651 have an 8-byte or larger character array on the stack. 652 653 This feature requires gcc version 4.2 or above, or a distribution 654 gcc with the feature backported ("-fstack-protector"). 655 656 On an x86 "defconfig" build, this feature adds canary checks to 657 about 3% of all kernel functions, which increases kernel code size 658 by about 0.3%. 659 660config STACKPROTECTOR_STRONG 661 bool "Strong Stack Protector" 662 depends on STACKPROTECTOR 663 depends on $(cc-option,-fstack-protector-strong) 664 default y 665 help 666 Functions will have the stack-protector canary logic added in any 667 of the following conditions: 668 669 - local variable's address used as part of the right hand side of an 670 assignment or function argument 671 - local variable is an array (or union containing an array), 672 regardless of array type or length 673 - uses register local variables 674 675 This feature requires gcc version 4.9 or above, or a distribution 676 gcc with the feature backported ("-fstack-protector-strong"). 677 678 On an x86 "defconfig" build, this feature adds canary checks to 679 about 20% of all kernel functions, which increases the kernel code 680 size by about 2%. 681 682config ARCH_SUPPORTS_SHADOW_CALL_STACK 683 bool 684 help 685 An architecture should select this if it supports the compiler's 686 Shadow Call Stack and implements runtime support for shadow stack 687 switching. 688 689config SHADOW_CALL_STACK 690 bool "Shadow Call Stack" 691 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK 692 depends on DYNAMIC_FTRACE_WITH_ARGS || DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER 693 depends on MMU 694 help 695 This option enables the compiler's Shadow Call Stack, which 696 uses a shadow stack to protect function return addresses from 697 being overwritten by an attacker. More information can be found 698 in the compiler's documentation: 699 700 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html 701 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options 702 703 Note that security guarantees in the kernel differ from the 704 ones documented for user space. The kernel must store addresses 705 of shadow stacks in memory, which means an attacker capable of 706 reading and writing arbitrary memory may be able to locate them 707 and hijack control flow by modifying the stacks. 708 709config DYNAMIC_SCS 710 bool 711 help 712 Set by the arch code if it relies on code patching to insert the 713 shadow call stack push and pop instructions rather than on the 714 compiler. 715 716config LTO 717 bool 718 help 719 Selected if the kernel will be built using the compiler's LTO feature. 720 721config LTO_CLANG 722 bool 723 select LTO 724 help 725 Selected if the kernel will be built using Clang's LTO feature. 726 727config ARCH_SUPPORTS_LTO_CLANG 728 bool 729 help 730 An architecture should select this option if it supports: 731 - compiling with Clang, 732 - compiling inline assembly with Clang's integrated assembler, 733 - and linking with LLD. 734 735config ARCH_SUPPORTS_LTO_CLANG_THIN 736 bool 737 help 738 An architecture should select this option if it can support Clang's 739 ThinLTO mode. 740 741config HAS_LTO_CLANG 742 def_bool y 743 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM 744 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm) 745 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm) 746 depends on ARCH_SUPPORTS_LTO_CLANG 747 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT 748 # https://github.com/ClangBuiltLinux/linux/issues/1721 749 depends on (!KASAN || KASAN_HW_TAGS || CLANG_VERSION >= 170000) || !DEBUG_INFO 750 depends on (!KCOV || CLANG_VERSION >= 170000) || !DEBUG_INFO 751 depends on !GCOV_KERNEL 752 help 753 The compiler and Kconfig options support building with Clang's 754 LTO. 755 756choice 757 prompt "Link Time Optimization (LTO)" 758 default LTO_NONE 759 help 760 This option enables Link Time Optimization (LTO), which allows the 761 compiler to optimize binaries globally. 762 763 If unsure, select LTO_NONE. Note that LTO is very resource-intensive 764 so it's disabled by default. 765 766config LTO_NONE 767 bool "None" 768 help 769 Build the kernel normally, without Link Time Optimization (LTO). 770 771config LTO_CLANG_FULL 772 bool "Clang Full LTO (EXPERIMENTAL)" 773 depends on HAS_LTO_CLANG 774 depends on !COMPILE_TEST 775 select LTO_CLANG 776 help 777 This option enables Clang's full Link Time Optimization (LTO), which 778 allows the compiler to optimize the kernel globally. If you enable 779 this option, the compiler generates LLVM bitcode instead of ELF 780 object files, and the actual compilation from bitcode happens at 781 the LTO link step, which may take several minutes depending on the 782 kernel configuration. More information can be found from LLVM's 783 documentation: 784 785 https://llvm.org/docs/LinkTimeOptimization.html 786 787 During link time, this option can use a large amount of RAM, and 788 may take much longer than the ThinLTO option. 789 790config LTO_CLANG_THIN 791 bool "Clang ThinLTO (EXPERIMENTAL)" 792 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN 793 select LTO_CLANG 794 help 795 This option enables Clang's ThinLTO, which allows for parallel 796 optimization and faster incremental compiles compared to the 797 CONFIG_LTO_CLANG_FULL option. More information can be found 798 from Clang's documentation: 799 800 https://clang.llvm.org/docs/ThinLTO.html 801 802 If unsure, say Y. 803endchoice 804 805config ARCH_SUPPORTS_CFI_CLANG 806 bool 807 help 808 An architecture should select this option if it can support Clang's 809 Control-Flow Integrity (CFI) checking. 810 811config ARCH_USES_CFI_TRAPS 812 bool 813 814config CFI_CLANG 815 bool "Use Clang's Control Flow Integrity (CFI)" 816 depends on ARCH_SUPPORTS_CFI_CLANG 817 depends on $(cc-option,-fsanitize=kcfi) 818 help 819 This option enables Clang's forward-edge Control Flow Integrity 820 (CFI) checking, where the compiler injects a runtime check to each 821 indirect function call to ensure the target is a valid function with 822 the correct static type. This restricts possible call targets and 823 makes it more difficult for an attacker to exploit bugs that allow 824 the modification of stored function pointers. More information can be 825 found from Clang's documentation: 826 827 https://clang.llvm.org/docs/ControlFlowIntegrity.html 828 829config CFI_PERMISSIVE 830 bool "Use CFI in permissive mode" 831 depends on CFI_CLANG 832 help 833 When selected, Control Flow Integrity (CFI) violations result in a 834 warning instead of a kernel panic. This option should only be used 835 for finding indirect call type mismatches during development. 836 837 If unsure, say N. 838 839config HAVE_ARCH_WITHIN_STACK_FRAMES 840 bool 841 help 842 An architecture should select this if it can walk the kernel stack 843 frames to determine if an object is part of either the arguments 844 or local variables (i.e. that it excludes saved return addresses, 845 and similar) by implementing an inline arch_within_stack_frames(), 846 which is used by CONFIG_HARDENED_USERCOPY. 847 848config HAVE_CONTEXT_TRACKING_USER 849 bool 850 help 851 Provide kernel/user boundaries probes necessary for subsystems 852 that need it, such as userspace RCU extended quiescent state. 853 Syscalls need to be wrapped inside user_exit()-user_enter(), either 854 optimized behind static key or through the slow path using TIF_NOHZ 855 flag. Exceptions handlers must be wrapped as well. Irqs are already 856 protected inside ct_irq_enter/ct_irq_exit() but preemption or signal 857 handling on irq exit still need to be protected. 858 859config HAVE_CONTEXT_TRACKING_USER_OFFSTACK 860 bool 861 help 862 Architecture neither relies on exception_enter()/exception_exit() 863 nor on schedule_user(). Also preempt_schedule_notrace() and 864 preempt_schedule_irq() can't be called in a preemptible section 865 while context tracking is CONTEXT_USER. This feature reflects a sane 866 entry implementation where the following requirements are met on 867 critical entry code, ie: before user_exit() or after user_enter(): 868 869 - Critical entry code isn't preemptible (or better yet: 870 not interruptible). 871 - No use of RCU read side critical sections, unless ct_nmi_enter() 872 got called. 873 - No use of instrumentation, unless instrumentation_begin() got 874 called. 875 876config HAVE_TIF_NOHZ 877 bool 878 help 879 Arch relies on TIF_NOHZ and syscall slow path to implement context 880 tracking calls to user_enter()/user_exit(). 881 882config HAVE_VIRT_CPU_ACCOUNTING 883 bool 884 885config HAVE_VIRT_CPU_ACCOUNTING_IDLE 886 bool 887 help 888 Architecture has its own way to account idle CPU time and therefore 889 doesn't implement vtime_account_idle(). 890 891config ARCH_HAS_SCALED_CPUTIME 892 bool 893 894config HAVE_VIRT_CPU_ACCOUNTING_GEN 895 bool 896 default y if 64BIT 897 help 898 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. 899 Before enabling this option, arch code must be audited 900 to ensure there are no races in concurrent read/write of 901 cputime_t. For example, reading/writing 64-bit cputime_t on 902 some 32-bit arches may require multiple accesses, so proper 903 locking is needed to protect against concurrent accesses. 904 905config HAVE_IRQ_TIME_ACCOUNTING 906 bool 907 help 908 Archs need to ensure they use a high enough resolution clock to 909 support irq time accounting and then call enable_sched_clock_irqtime(). 910 911config HAVE_MOVE_PUD 912 bool 913 help 914 Architectures that select this are able to move page tables at the 915 PUD level. If there are only 3 page table levels, the move effectively 916 happens at the PGD level. 917 918config HAVE_MOVE_PMD 919 bool 920 help 921 Archs that select this are able to move page tables at the PMD level. 922 923config HAVE_ARCH_TRANSPARENT_HUGEPAGE 924 bool 925 926config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD 927 bool 928 929config HAVE_ARCH_HUGE_VMAP 930 bool 931 932# 933# Archs that select this would be capable of PMD-sized vmaps (i.e., 934# arch_vmap_pmd_supported() returns true). The VM_ALLOW_HUGE_VMAP flag 935# must be used to enable allocations to use hugepages. 936# 937config HAVE_ARCH_HUGE_VMALLOC 938 depends on HAVE_ARCH_HUGE_VMAP 939 bool 940 941config ARCH_WANT_HUGE_PMD_SHARE 942 bool 943 944# Archs that want to use pmd_mkwrite on kernel memory need it defined even 945# if there are no userspace memory management features that use it 946config ARCH_WANT_KERNEL_PMD_MKWRITE 947 bool 948 949config ARCH_WANT_PMD_MKWRITE 950 def_bool TRANSPARENT_HUGEPAGE || ARCH_WANT_KERNEL_PMD_MKWRITE 951 952config HAVE_ARCH_SOFT_DIRTY 953 bool 954 955config HAVE_MOD_ARCH_SPECIFIC 956 bool 957 help 958 The arch uses struct mod_arch_specific to store data. Many arches 959 just need a simple module loader without arch specific data - those 960 should not enable this. 961 962config MODULES_USE_ELF_RELA 963 bool 964 help 965 Modules only use ELF RELA relocations. Modules with ELF REL 966 relocations will give an error. 967 968config MODULES_USE_ELF_REL 969 bool 970 help 971 Modules only use ELF REL relocations. Modules with ELF RELA 972 relocations will give an error. 973 974config ARCH_WANTS_MODULES_DATA_IN_VMALLOC 975 bool 976 help 977 For architectures like powerpc/32 which have constraints on module 978 allocation and need to allocate module data outside of module area. 979 980config ARCH_WANTS_EXECMEM_LATE 981 bool 982 help 983 For architectures that do not allocate executable memory early on 984 boot, but rather require its initialization late when there is 985 enough entropy for module space randomization, for instance 986 arm64. 987 988config HAVE_IRQ_EXIT_ON_IRQ_STACK 989 bool 990 help 991 Architecture doesn't only execute the irq handler on the irq stack 992 but also irq_exit(). This way we can process softirqs on this irq 993 stack instead of switching to a new one when we call __do_softirq() 994 in the end of an hardirq. 995 This spares a stack switch and improves cache usage on softirq 996 processing. 997 998config HAVE_SOFTIRQ_ON_OWN_STACK 999 bool 1000 help 1001 Architecture provides a function to run __do_softirq() on a 1002 separate stack. 1003 1004config SOFTIRQ_ON_OWN_STACK 1005 def_bool HAVE_SOFTIRQ_ON_OWN_STACK && !PREEMPT_RT 1006 1007config ALTERNATE_USER_ADDRESS_SPACE 1008 bool 1009 help 1010 Architectures set this when the CPU uses separate address 1011 spaces for kernel and user space pointers. In this case, the 1012 access_ok() check on a __user pointer is skipped. 1013 1014config PGTABLE_LEVELS 1015 int 1016 default 2 1017 1018config ARCH_HAS_ELF_RANDOMIZE 1019 bool 1020 help 1021 An architecture supports choosing randomized locations for 1022 stack, mmap, brk, and ET_DYN. Defined functions: 1023 - arch_mmap_rnd() 1024 - arch_randomize_brk() 1025 1026config HAVE_ARCH_MMAP_RND_BITS 1027 bool 1028 help 1029 An arch should select this symbol if it supports setting a variable 1030 number of bits for use in establishing the base address for mmap 1031 allocations, has MMU enabled and provides values for both: 1032 - ARCH_MMAP_RND_BITS_MIN 1033 - ARCH_MMAP_RND_BITS_MAX 1034 1035config HAVE_EXIT_THREAD 1036 bool 1037 help 1038 An architecture implements exit_thread. 1039 1040config ARCH_MMAP_RND_BITS_MIN 1041 int 1042 1043config ARCH_MMAP_RND_BITS_MAX 1044 int 1045 1046config ARCH_MMAP_RND_BITS_DEFAULT 1047 int 1048 1049config FORCE_MAX_MMAP_RND_BITS 1050 bool "Force maximum number of bits to use for ASLR of mmap base address" 1051 default y if !64BIT 1052 help 1053 ARCH_MMAP_RND_BITS and ARCH_MMAP_RND_COMPAT_BITS represent the number 1054 of bits to use for ASLR and if no custom value is assigned (EXPERT) 1055 then the architecture's lower bound (minimum) value is assumed. 1056 This toggle changes that default assumption to assume the arch upper 1057 bound (maximum) value instead. 1058 1059config ARCH_MMAP_RND_BITS 1060 int "Number of bits to use for ASLR of mmap base address" if EXPERT 1061 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX 1062 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT 1063 default ARCH_MMAP_RND_BITS_MAX if FORCE_MAX_MMAP_RND_BITS 1064 default ARCH_MMAP_RND_BITS_MIN 1065 depends on HAVE_ARCH_MMAP_RND_BITS 1066 help 1067 This value can be used to select the number of bits to use to 1068 determine the random offset to the base address of vma regions 1069 resulting from mmap allocations. This value will be bounded 1070 by the architecture's minimum and maximum supported values. 1071 1072 This value can be changed after boot using the 1073 /proc/sys/vm/mmap_rnd_bits tunable 1074 1075config HAVE_ARCH_MMAP_RND_COMPAT_BITS 1076 bool 1077 help 1078 An arch should select this symbol if it supports running applications 1079 in compatibility mode, supports setting a variable number of bits for 1080 use in establishing the base address for mmap allocations, has MMU 1081 enabled and provides values for both: 1082 - ARCH_MMAP_RND_COMPAT_BITS_MIN 1083 - ARCH_MMAP_RND_COMPAT_BITS_MAX 1084 1085config ARCH_MMAP_RND_COMPAT_BITS_MIN 1086 int 1087 1088config ARCH_MMAP_RND_COMPAT_BITS_MAX 1089 int 1090 1091config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1092 int 1093 1094config ARCH_MMAP_RND_COMPAT_BITS 1095 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT 1096 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX 1097 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1098 default ARCH_MMAP_RND_COMPAT_BITS_MAX if FORCE_MAX_MMAP_RND_BITS 1099 default ARCH_MMAP_RND_COMPAT_BITS_MIN 1100 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS 1101 help 1102 This value can be used to select the number of bits to use to 1103 determine the random offset to the base address of vma regions 1104 resulting from mmap allocations for compatible applications This 1105 value will be bounded by the architecture's minimum and maximum 1106 supported values. 1107 1108 This value can be changed after boot using the 1109 /proc/sys/vm/mmap_rnd_compat_bits tunable 1110 1111config HAVE_ARCH_COMPAT_MMAP_BASES 1112 bool 1113 help 1114 This allows 64bit applications to invoke 32-bit mmap() syscall 1115 and vice-versa 32-bit applications to call 64-bit mmap(). 1116 Required for applications doing different bitness syscalls. 1117 1118config HAVE_PAGE_SIZE_4KB 1119 bool 1120 1121config HAVE_PAGE_SIZE_8KB 1122 bool 1123 1124config HAVE_PAGE_SIZE_16KB 1125 bool 1126 1127config HAVE_PAGE_SIZE_32KB 1128 bool 1129 1130config HAVE_PAGE_SIZE_64KB 1131 bool 1132 1133config HAVE_PAGE_SIZE_256KB 1134 bool 1135 1136choice 1137 prompt "MMU page size" 1138 1139config PAGE_SIZE_4KB 1140 bool "4KiB pages" 1141 depends on HAVE_PAGE_SIZE_4KB 1142 help 1143 This option select the standard 4KiB Linux page size and the only 1144 available option on many architectures. Using 4KiB page size will 1145 minimize memory consumption and is therefore recommended for low 1146 memory systems. 1147 Some software that is written for x86 systems makes incorrect 1148 assumptions about the page size and only runs on 4KiB pages. 1149 1150config PAGE_SIZE_8KB 1151 bool "8KiB pages" 1152 depends on HAVE_PAGE_SIZE_8KB 1153 help 1154 This option is the only supported page size on a few older 1155 processors, and can be slightly faster than 4KiB pages. 1156 1157config PAGE_SIZE_16KB 1158 bool "16KiB pages" 1159 depends on HAVE_PAGE_SIZE_16KB 1160 help 1161 This option is usually a good compromise between memory 1162 consumption and performance for typical desktop and server 1163 workloads, often saving a level of page table lookups compared 1164 to 4KB pages as well as reducing TLB pressure and overhead of 1165 per-page operations in the kernel at the expense of a larger 1166 page cache. 1167 1168config PAGE_SIZE_32KB 1169 bool "32KiB pages" 1170 depends on HAVE_PAGE_SIZE_32KB 1171 help 1172 Using 32KiB page size will result in slightly higher performance 1173 kernel at the price of higher memory consumption compared to 1174 16KiB pages. This option is available only on cnMIPS cores. 1175 Note that you will need a suitable Linux distribution to 1176 support this. 1177 1178config PAGE_SIZE_64KB 1179 bool "64KiB pages" 1180 depends on HAVE_PAGE_SIZE_64KB 1181 help 1182 Using 64KiB page size will result in slightly higher performance 1183 kernel at the price of much higher memory consumption compared to 1184 4KiB or 16KiB pages. 1185 This is not suitable for general-purpose workloads but the 1186 better performance may be worth the cost for certain types of 1187 supercomputing or database applications that work mostly with 1188 large in-memory data rather than small files. 1189 1190config PAGE_SIZE_256KB 1191 bool "256KiB pages" 1192 depends on HAVE_PAGE_SIZE_256KB 1193 help 1194 256KiB pages have little practical value due to their extreme 1195 memory usage. The kernel will only be able to run applications 1196 that have been compiled with '-zmax-page-size' set to 256KiB 1197 (the default is 64KiB or 4KiB on most architectures). 1198 1199endchoice 1200 1201config PAGE_SIZE_LESS_THAN_64KB 1202 def_bool y 1203 depends on !PAGE_SIZE_64KB 1204 depends on PAGE_SIZE_LESS_THAN_256KB 1205 1206config PAGE_SIZE_LESS_THAN_256KB 1207 def_bool y 1208 depends on !PAGE_SIZE_256KB 1209 1210config PAGE_SHIFT 1211 int 1212 default 12 if PAGE_SIZE_4KB 1213 default 13 if PAGE_SIZE_8KB 1214 default 14 if PAGE_SIZE_16KB 1215 default 15 if PAGE_SIZE_32KB 1216 default 16 if PAGE_SIZE_64KB 1217 default 18 if PAGE_SIZE_256KB 1218 1219# This allows to use a set of generic functions to determine mmap base 1220# address by giving priority to top-down scheme only if the process 1221# is not in legacy mode (compat task, unlimited stack size or 1222# sysctl_legacy_va_layout). 1223# Architecture that selects this option can provide its own version of: 1224# - STACK_RND_MASK 1225config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT 1226 bool 1227 depends on MMU 1228 select ARCH_HAS_ELF_RANDOMIZE 1229 1230config HAVE_OBJTOOL 1231 bool 1232 1233config HAVE_JUMP_LABEL_HACK 1234 bool 1235 1236config HAVE_NOINSTR_HACK 1237 bool 1238 1239config HAVE_NOINSTR_VALIDATION 1240 bool 1241 1242config HAVE_UACCESS_VALIDATION 1243 bool 1244 select OBJTOOL 1245 1246config HAVE_STACK_VALIDATION 1247 bool 1248 help 1249 Architecture supports objtool compile-time frame pointer rule 1250 validation. 1251 1252config HAVE_RELIABLE_STACKTRACE 1253 bool 1254 help 1255 Architecture has either save_stack_trace_tsk_reliable() or 1256 arch_stack_walk_reliable() function which only returns a stack trace 1257 if it can guarantee the trace is reliable. 1258 1259config HAVE_ARCH_HASH 1260 bool 1261 default n 1262 help 1263 If this is set, the architecture provides an <asm/hash.h> 1264 file which provides platform-specific implementations of some 1265 functions in <linux/hash.h> or fs/namei.c. 1266 1267config HAVE_ARCH_NVRAM_OPS 1268 bool 1269 1270config ISA_BUS_API 1271 def_bool ISA 1272 1273# 1274# ABI hall of shame 1275# 1276config CLONE_BACKWARDS 1277 bool 1278 help 1279 Architecture has tls passed as the 4th argument of clone(2), 1280 not the 5th one. 1281 1282config CLONE_BACKWARDS2 1283 bool 1284 help 1285 Architecture has the first two arguments of clone(2) swapped. 1286 1287config CLONE_BACKWARDS3 1288 bool 1289 help 1290 Architecture has tls passed as the 3rd argument of clone(2), 1291 not the 5th one. 1292 1293config ODD_RT_SIGACTION 1294 bool 1295 help 1296 Architecture has unusual rt_sigaction(2) arguments 1297 1298config OLD_SIGSUSPEND 1299 bool 1300 help 1301 Architecture has old sigsuspend(2) syscall, of one-argument variety 1302 1303config OLD_SIGSUSPEND3 1304 bool 1305 help 1306 Even weirder antique ABI - three-argument sigsuspend(2) 1307 1308config OLD_SIGACTION 1309 bool 1310 help 1311 Architecture has old sigaction(2) syscall. Nope, not the same 1312 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2), 1313 but fairly different variant of sigaction(2), thanks to OSF/1 1314 compatibility... 1315 1316config COMPAT_OLD_SIGACTION 1317 bool 1318 1319config COMPAT_32BIT_TIME 1320 bool "Provide system calls for 32-bit time_t" 1321 default !64BIT || COMPAT 1322 help 1323 This enables 32 bit time_t support in addition to 64 bit time_t support. 1324 This is relevant on all 32-bit architectures, and 64-bit architectures 1325 as part of compat syscall handling. 1326 1327config ARCH_NO_PREEMPT 1328 bool 1329 1330config ARCH_SUPPORTS_RT 1331 bool 1332 1333config CPU_NO_EFFICIENT_FFS 1334 def_bool n 1335 1336config HAVE_ARCH_VMAP_STACK 1337 def_bool n 1338 help 1339 An arch should select this symbol if it can support kernel stacks 1340 in vmalloc space. This means: 1341 1342 - vmalloc space must be large enough to hold many kernel stacks. 1343 This may rule out many 32-bit architectures. 1344 1345 - Stacks in vmalloc space need to work reliably. For example, if 1346 vmap page tables are created on demand, either this mechanism 1347 needs to work while the stack points to a virtual address with 1348 unpopulated page tables or arch code (switch_to() and switch_mm(), 1349 most likely) needs to ensure that the stack's page table entries 1350 are populated before running on a possibly unpopulated stack. 1351 1352 - If the stack overflows into a guard page, something reasonable 1353 should happen. The definition of "reasonable" is flexible, but 1354 instantly rebooting without logging anything would be unfriendly. 1355 1356config VMAP_STACK 1357 default y 1358 bool "Use a virtually-mapped stack" 1359 depends on HAVE_ARCH_VMAP_STACK 1360 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC 1361 help 1362 Enable this if you want the use virtually-mapped kernel stacks 1363 with guard pages. This causes kernel stack overflows to be 1364 caught immediately rather than causing difficult-to-diagnose 1365 corruption. 1366 1367 To use this with software KASAN modes, the architecture must support 1368 backing virtual mappings with real shadow memory, and KASAN_VMALLOC 1369 must be enabled. 1370 1371config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1372 def_bool n 1373 help 1374 An arch should select this symbol if it can support kernel stack 1375 offset randomization with calls to add_random_kstack_offset() 1376 during syscall entry and choose_random_kstack_offset() during 1377 syscall exit. Careful removal of -fstack-protector-strong and 1378 -fstack-protector should also be applied to the entry code and 1379 closely examined, as the artificial stack bump looks like an array 1380 to the compiler, so it will attempt to add canary checks regardless 1381 of the static branch state. 1382 1383config RANDOMIZE_KSTACK_OFFSET 1384 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT 1385 default y 1386 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1387 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000 1388 help 1389 The kernel stack offset can be randomized (after pt_regs) by 1390 roughly 5 bits of entropy, frustrating memory corruption 1391 attacks that depend on stack address determinism or 1392 cross-syscall address exposures. 1393 1394 The feature is controlled via the "randomize_kstack_offset=on/off" 1395 kernel boot param, and if turned off has zero overhead due to its use 1396 of static branches (see JUMP_LABEL). 1397 1398 If unsure, say Y. 1399 1400config RANDOMIZE_KSTACK_OFFSET_DEFAULT 1401 bool "Default state of kernel stack offset randomization" 1402 depends on RANDOMIZE_KSTACK_OFFSET 1403 help 1404 Kernel stack offset randomization is controlled by kernel boot param 1405 "randomize_kstack_offset=on/off", and this config chooses the default 1406 boot state. 1407 1408config ARCH_OPTIONAL_KERNEL_RWX 1409 def_bool n 1410 1411config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1412 def_bool n 1413 1414config ARCH_HAS_STRICT_KERNEL_RWX 1415 def_bool n 1416 1417config STRICT_KERNEL_RWX 1418 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX 1419 depends on ARCH_HAS_STRICT_KERNEL_RWX 1420 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1421 help 1422 If this is set, kernel text and rodata memory will be made read-only, 1423 and non-text memory will be made non-executable. This provides 1424 protection against certain security exploits (e.g. executing the heap 1425 or modifying text) 1426 1427 These features are considered standard security practice these days. 1428 You should say Y here in almost all cases. 1429 1430config ARCH_HAS_STRICT_MODULE_RWX 1431 def_bool n 1432 1433config STRICT_MODULE_RWX 1434 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX 1435 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES 1436 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1437 help 1438 If this is set, module text and rodata memory will be made read-only, 1439 and non-text memory will be made non-executable. This provides 1440 protection against certain security exploits (e.g. writing to text) 1441 1442# select if the architecture provides an asm/dma-direct.h header 1443config ARCH_HAS_PHYS_TO_DMA 1444 bool 1445 1446config HAVE_ARCH_COMPILER_H 1447 bool 1448 help 1449 An architecture can select this if it provides an 1450 asm/compiler.h header that should be included after 1451 linux/compiler-*.h in order to override macro definitions that those 1452 headers generally provide. 1453 1454config HAVE_ARCH_PREL32_RELOCATIONS 1455 bool 1456 help 1457 May be selected by an architecture if it supports place-relative 1458 32-bit relocations, both in the toolchain and in the module loader, 1459 in which case relative references can be used in special sections 1460 for PCI fixup, initcalls etc which are only half the size on 64 bit 1461 architectures, and don't require runtime relocation on relocatable 1462 kernels. 1463 1464config ARCH_USE_MEMREMAP_PROT 1465 bool 1466 1467config LOCK_EVENT_COUNTS 1468 bool "Locking event counts collection" 1469 depends on DEBUG_FS 1470 help 1471 Enable light-weight counting of various locking related events 1472 in the system with minimal performance impact. This reduces 1473 the chance of application behavior change because of timing 1474 differences. The counts are reported via debugfs. 1475 1476# Select if the architecture has support for applying RELR relocations. 1477config ARCH_HAS_RELR 1478 bool 1479 1480config RELR 1481 bool "Use RELR relocation packing" 1482 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR 1483 default y 1484 help 1485 Store the kernel's dynamic relocations in the RELR relocation packing 1486 format. Requires a compatible linker (LLD supports this feature), as 1487 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy 1488 are compatible). 1489 1490config ARCH_HAS_MEM_ENCRYPT 1491 bool 1492 1493config ARCH_HAS_CC_PLATFORM 1494 bool 1495 1496config HAVE_SPARSE_SYSCALL_NR 1497 bool 1498 help 1499 An architecture should select this if its syscall numbering is sparse 1500 to save space. For example, MIPS architecture has a syscall array with 1501 entries at 4000, 5000 and 6000 locations. This option turns on syscall 1502 related optimizations for a given architecture. 1503 1504config ARCH_HAS_VDSO_DATA 1505 bool 1506 1507config HAVE_STATIC_CALL 1508 bool 1509 1510config HAVE_STATIC_CALL_INLINE 1511 bool 1512 depends on HAVE_STATIC_CALL 1513 select OBJTOOL 1514 1515config HAVE_PREEMPT_DYNAMIC 1516 bool 1517 1518config HAVE_PREEMPT_DYNAMIC_CALL 1519 bool 1520 depends on HAVE_STATIC_CALL 1521 select HAVE_PREEMPT_DYNAMIC 1522 help 1523 An architecture should select this if it can handle the preemption 1524 model being selected at boot time using static calls. 1525 1526 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a 1527 preemption function will be patched directly. 1528 1529 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any 1530 call to a preemption function will go through a trampoline, and the 1531 trampoline will be patched. 1532 1533 It is strongly advised to support inline static call to avoid any 1534 overhead. 1535 1536config HAVE_PREEMPT_DYNAMIC_KEY 1537 bool 1538 depends on HAVE_ARCH_JUMP_LABEL 1539 select HAVE_PREEMPT_DYNAMIC 1540 help 1541 An architecture should select this if it can handle the preemption 1542 model being selected at boot time using static keys. 1543 1544 Each preemption function will be given an early return based on a 1545 static key. This should have slightly lower overhead than non-inline 1546 static calls, as this effectively inlines each trampoline into the 1547 start of its callee. This may avoid redundant work, and may 1548 integrate better with CFI schemes. 1549 1550 This will have greater overhead than using inline static calls as 1551 the call to the preemption function cannot be entirely elided. 1552 1553config ARCH_WANT_LD_ORPHAN_WARN 1554 bool 1555 help 1556 An arch should select this symbol once all linker sections are explicitly 1557 included, size-asserted, or discarded in the linker scripts. This is 1558 important because we never want expected sections to be placed heuristically 1559 by the linker, since the locations of such sections can change between linker 1560 versions. 1561 1562config HAVE_ARCH_PFN_VALID 1563 bool 1564 1565config ARCH_SUPPORTS_DEBUG_PAGEALLOC 1566 bool 1567 1568config ARCH_SUPPORTS_PAGE_TABLE_CHECK 1569 bool 1570 1571config ARCH_SPLIT_ARG64 1572 bool 1573 help 1574 If a 32-bit architecture requires 64-bit arguments to be split into 1575 pairs of 32-bit arguments, select this option. 1576 1577config ARCH_HAS_ELFCORE_COMPAT 1578 bool 1579 1580config ARCH_HAS_PARANOID_L1D_FLUSH 1581 bool 1582 1583config ARCH_HAVE_TRACE_MMIO_ACCESS 1584 bool 1585 1586config DYNAMIC_SIGFRAME 1587 bool 1588 1589# Select, if arch has a named attribute group bound to NUMA device nodes. 1590config HAVE_ARCH_NODE_DEV_GROUP 1591 bool 1592 1593config ARCH_HAS_HW_PTE_YOUNG 1594 bool 1595 help 1596 Architectures that select this option are capable of setting the 1597 accessed bit in PTE entries when using them as part of linear address 1598 translations. Architectures that require runtime check should select 1599 this option and override arch_has_hw_pte_young(). 1600 1601config ARCH_HAS_NONLEAF_PMD_YOUNG 1602 bool 1603 help 1604 Architectures that select this option are capable of setting the 1605 accessed bit in non-leaf PMD entries when using them as part of linear 1606 address translations. Page table walkers that clear the accessed bit 1607 may use this capability to reduce their search space. 1608 1609config ARCH_HAS_KERNEL_FPU_SUPPORT 1610 bool 1611 help 1612 Architectures that select this option can run floating-point code in 1613 the kernel, as described in Documentation/core-api/floating-point.rst. 1614 1615source "kernel/gcov/Kconfig" 1616 1617source "scripts/gcc-plugins/Kconfig" 1618 1619config FUNCTION_ALIGNMENT_4B 1620 bool 1621 1622config FUNCTION_ALIGNMENT_8B 1623 bool 1624 1625config FUNCTION_ALIGNMENT_16B 1626 bool 1627 1628config FUNCTION_ALIGNMENT_32B 1629 bool 1630 1631config FUNCTION_ALIGNMENT_64B 1632 bool 1633 1634config FUNCTION_ALIGNMENT 1635 int 1636 default 64 if FUNCTION_ALIGNMENT_64B 1637 default 32 if FUNCTION_ALIGNMENT_32B 1638 default 16 if FUNCTION_ALIGNMENT_16B 1639 default 8 if FUNCTION_ALIGNMENT_8B 1640 default 4 if FUNCTION_ALIGNMENT_4B 1641 default 0 1642 1643config CC_HAS_MIN_FUNCTION_ALIGNMENT 1644 # Detect availability of the GCC option -fmin-function-alignment which 1645 # guarantees minimal alignment for all functions, unlike 1646 # -falign-functions which the compiler ignores for cold functions. 1647 def_bool $(cc-option, -fmin-function-alignment=8) 1648 1649config CC_HAS_SANE_FUNCTION_ALIGNMENT 1650 # Set if the guaranteed alignment with -fmin-function-alignment is 1651 # available or extra care is required in the kernel. Clang provides 1652 # strict alignment always, even with -falign-functions. 1653 def_bool CC_HAS_MIN_FUNCTION_ALIGNMENT || CC_IS_CLANG 1654 1655config ARCH_NEED_CMPXCHG_1_EMU 1656 bool 1657 1658endmenu 1659