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 12menu "General architecture-dependent options" 13 14config CRASH_CORE 15 bool 16 17config KEXEC_CORE 18 select CRASH_CORE 19 bool 20 21config KEXEC_ELF 22 bool 23 24config HAVE_IMA_KEXEC 25 bool 26 27config SET_FS 28 bool 29 30config HOTPLUG_SMT 31 bool 32 33config GENERIC_ENTRY 34 bool 35 36config KPROBES 37 bool "Kprobes" 38 depends on MODULES 39 depends on HAVE_KPROBES 40 select KALLSYMS 41 help 42 Kprobes allows you to trap at almost any kernel address and 43 execute a callback function. register_kprobe() establishes 44 a probepoint and specifies the callback. Kprobes is useful 45 for kernel debugging, non-intrusive instrumentation and testing. 46 If in doubt, say "N". 47 48config JUMP_LABEL 49 bool "Optimize very unlikely/likely branches" 50 depends on HAVE_ARCH_JUMP_LABEL 51 depends on CC_HAS_ASM_GOTO 52 help 53 This option enables a transparent branch optimization that 54 makes certain almost-always-true or almost-always-false branch 55 conditions even cheaper to execute within the kernel. 56 57 Certain performance-sensitive kernel code, such as trace points, 58 scheduler functionality, networking code and KVM have such 59 branches and include support for this optimization technique. 60 61 If it is detected that the compiler has support for "asm goto", 62 the kernel will compile such branches with just a nop 63 instruction. When the condition flag is toggled to true, the 64 nop will be converted to a jump instruction to execute the 65 conditional block of instructions. 66 67 This technique lowers overhead and stress on the branch prediction 68 of the processor and generally makes the kernel faster. The update 69 of the condition is slower, but those are always very rare. 70 71 ( On 32-bit x86, the necessary options added to the compiler 72 flags may increase the size of the kernel slightly. ) 73 74config STATIC_KEYS_SELFTEST 75 bool "Static key selftest" 76 depends on JUMP_LABEL 77 help 78 Boot time self-test of the branch patching code. 79 80config STATIC_CALL_SELFTEST 81 bool "Static call selftest" 82 depends on HAVE_STATIC_CALL 83 help 84 Boot time self-test of the call patching code. 85 86config OPTPROBES 87 def_bool y 88 depends on KPROBES && HAVE_OPTPROBES 89 select TASKS_RCU if PREEMPTION 90 91config KPROBES_ON_FTRACE 92 def_bool y 93 depends on KPROBES && HAVE_KPROBES_ON_FTRACE 94 depends on DYNAMIC_FTRACE_WITH_REGS 95 help 96 If function tracer is enabled and the arch supports full 97 passing of pt_regs to function tracing, then kprobes can 98 optimize on top of function tracing. 99 100config UPROBES 101 def_bool n 102 depends on ARCH_SUPPORTS_UPROBES 103 help 104 Uprobes is the user-space counterpart to kprobes: they 105 enable instrumentation applications (such as 'perf probe') 106 to establish unintrusive probes in user-space binaries and 107 libraries, by executing handler functions when the probes 108 are hit by user-space applications. 109 110 ( These probes come in the form of single-byte breakpoints, 111 managed by the kernel and kept transparent to the probed 112 application. ) 113 114config HAVE_64BIT_ALIGNED_ACCESS 115 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS 116 help 117 Some architectures require 64 bit accesses to be 64 bit 118 aligned, which also requires structs containing 64 bit values 119 to be 64 bit aligned too. This includes some 32 bit 120 architectures which can do 64 bit accesses, as well as 64 bit 121 architectures without unaligned access. 122 123 This symbol should be selected by an architecture if 64 bit 124 accesses are required to be 64 bit aligned in this way even 125 though it is not a 64 bit architecture. 126 127 See Documentation/core-api/unaligned-memory-access.rst for 128 more information on the topic of unaligned memory accesses. 129 130config HAVE_EFFICIENT_UNALIGNED_ACCESS 131 bool 132 help 133 Some architectures are unable to perform unaligned accesses 134 without the use of get_unaligned/put_unaligned. Others are 135 unable to perform such accesses efficiently (e.g. trap on 136 unaligned access and require fixing it up in the exception 137 handler.) 138 139 This symbol should be selected by an architecture if it can 140 perform unaligned accesses efficiently to allow different 141 code paths to be selected for these cases. Some network 142 drivers, for example, could opt to not fix up alignment 143 problems with received packets if doing so would not help 144 much. 145 146 See Documentation/core-api/unaligned-memory-access.rst for more 147 information on the topic of unaligned memory accesses. 148 149config ARCH_USE_BUILTIN_BSWAP 150 bool 151 help 152 Modern versions of GCC (since 4.4) have builtin functions 153 for handling byte-swapping. Using these, instead of the old 154 inline assembler that the architecture code provides in the 155 __arch_bswapXX() macros, allows the compiler to see what's 156 happening and offers more opportunity for optimisation. In 157 particular, the compiler will be able to combine the byteswap 158 with a nearby load or store and use load-and-swap or 159 store-and-swap instructions if the architecture has them. It 160 should almost *never* result in code which is worse than the 161 hand-coded assembler in <asm/swab.h>. But just in case it 162 does, the use of the builtins is optional. 163 164 Any architecture with load-and-swap or store-and-swap 165 instructions should set this. And it shouldn't hurt to set it 166 on architectures that don't have such instructions. 167 168config KRETPROBES 169 def_bool y 170 depends on KPROBES && HAVE_KRETPROBES 171 172config USER_RETURN_NOTIFIER 173 bool 174 depends on HAVE_USER_RETURN_NOTIFIER 175 help 176 Provide a kernel-internal notification when a cpu is about to 177 switch to user mode. 178 179config HAVE_IOREMAP_PROT 180 bool 181 182config HAVE_KPROBES 183 bool 184 185config HAVE_KRETPROBES 186 bool 187 188config HAVE_OPTPROBES 189 bool 190 191config HAVE_KPROBES_ON_FTRACE 192 bool 193 194config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE 195 bool 196 help 197 Since kretprobes modifies return address on the stack, the 198 stacktrace may see the kretprobe trampoline address instead 199 of correct one. If the architecture stacktrace code and 200 unwinder can adjust such entries, select this configuration. 201 202config HAVE_FUNCTION_ERROR_INJECTION 203 bool 204 205config HAVE_NMI 206 bool 207 208config HAVE_FUNCTION_DESCRIPTORS 209 bool 210 211config TRACE_IRQFLAGS_SUPPORT 212 bool 213 214# 215# An arch should select this if it provides all these things: 216# 217# task_pt_regs() in asm/processor.h or asm/ptrace.h 218# arch_has_single_step() if there is hardware single-step support 219# arch_has_block_step() if there is hardware block-step support 220# asm/syscall.h supplying asm-generic/syscall.h interface 221# linux/regset.h user_regset interfaces 222# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 223# TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit} 224# TIF_NOTIFY_RESUME calls tracehook_notify_resume() 225# signal delivery calls tracehook_signal_handler() 226# 227config HAVE_ARCH_TRACEHOOK 228 bool 229 230config HAVE_DMA_CONTIGUOUS 231 bool 232 233config GENERIC_SMP_IDLE_THREAD 234 bool 235 236config GENERIC_IDLE_POLL_SETUP 237 bool 238 239config ARCH_HAS_FORTIFY_SOURCE 240 bool 241 help 242 An architecture should select this when it can successfully 243 build and run with CONFIG_FORTIFY_SOURCE. 244 245# 246# Select if the arch provides a historic keepinit alias for the retain_initrd 247# command line option 248# 249config ARCH_HAS_KEEPINITRD 250 bool 251 252# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h 253config ARCH_HAS_SET_MEMORY 254 bool 255 256# Select if arch has all set_direct_map_invalid/default() functions 257config ARCH_HAS_SET_DIRECT_MAP 258 bool 259 260# 261# Select if the architecture provides the arch_dma_set_uncached symbol to 262# either provide an uncached segment alias for a DMA allocation, or 263# to remap the page tables in place. 264# 265config ARCH_HAS_DMA_SET_UNCACHED 266 bool 267 268# 269# Select if the architectures provides the arch_dma_clear_uncached symbol 270# to undo an in-place page table remap for uncached access. 271# 272config ARCH_HAS_DMA_CLEAR_UNCACHED 273 bool 274 275# Select if arch init_task must go in the __init_task_data section 276config ARCH_TASK_STRUCT_ON_STACK 277 bool 278 279# Select if arch has its private alloc_task_struct() function 280config ARCH_TASK_STRUCT_ALLOCATOR 281 bool 282 283config HAVE_ARCH_THREAD_STRUCT_WHITELIST 284 bool 285 depends on !ARCH_TASK_STRUCT_ALLOCATOR 286 help 287 An architecture should select this to provide hardened usercopy 288 knowledge about what region of the thread_struct should be 289 whitelisted for copying to userspace. Normally this is only the 290 FPU registers. Specifically, arch_thread_struct_whitelist() 291 should be implemented. Without this, the entire thread_struct 292 field in task_struct will be left whitelisted. 293 294# Select if arch has its private alloc_thread_stack() function 295config ARCH_THREAD_STACK_ALLOCATOR 296 bool 297 298# Select if arch wants to size task_struct dynamically via arch_task_struct_size: 299config ARCH_WANTS_DYNAMIC_TASK_STRUCT 300 bool 301 302config ARCH_WANTS_NO_INSTR 303 bool 304 help 305 An architecture should select this if the noinstr macro is being used on 306 functions to denote that the toolchain should avoid instrumenting such 307 functions and is required for correctness. 308 309config ARCH_32BIT_OFF_T 310 bool 311 depends on !64BIT 312 help 313 All new 32-bit architectures should have 64-bit off_t type on 314 userspace side which corresponds to the loff_t kernel type. This 315 is the requirement for modern ABIs. Some existing architectures 316 still support 32-bit off_t. This option is enabled for all such 317 architectures explicitly. 318 319# Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat 320config ARCH_32BIT_USTAT_F_TINODE 321 bool 322 323config HAVE_ASM_MODVERSIONS 324 bool 325 help 326 This symbol should be selected by an architecture if it provides 327 <asm/asm-prototypes.h> to support the module versioning for symbols 328 exported from assembly code. 329 330config HAVE_REGS_AND_STACK_ACCESS_API 331 bool 332 help 333 This symbol should be selected by an architecture if it supports 334 the API needed to access registers and stack entries from pt_regs, 335 declared in asm/ptrace.h 336 For example the kprobes-based event tracer needs this API. 337 338config HAVE_RSEQ 339 bool 340 depends on HAVE_REGS_AND_STACK_ACCESS_API 341 help 342 This symbol should be selected by an architecture if it 343 supports an implementation of restartable sequences. 344 345config HAVE_FUNCTION_ARG_ACCESS_API 346 bool 347 help 348 This symbol should be selected by an architecture if it supports 349 the API needed to access function arguments from pt_regs, 350 declared in asm/ptrace.h 351 352config HAVE_HW_BREAKPOINT 353 bool 354 depends on PERF_EVENTS 355 356config HAVE_MIXED_BREAKPOINTS_REGS 357 bool 358 depends on HAVE_HW_BREAKPOINT 359 help 360 Depending on the arch implementation of hardware breakpoints, 361 some of them have separate registers for data and instruction 362 breakpoints addresses, others have mixed registers to store 363 them but define the access type in a control register. 364 Select this option if your arch implements breakpoints under the 365 latter fashion. 366 367config HAVE_USER_RETURN_NOTIFIER 368 bool 369 370config HAVE_PERF_EVENTS_NMI 371 bool 372 help 373 System hardware can generate an NMI using the perf event 374 subsystem. Also has support for calculating CPU cycle events 375 to determine how many clock cycles in a given period. 376 377config HAVE_HARDLOCKUP_DETECTOR_PERF 378 bool 379 depends on HAVE_PERF_EVENTS_NMI 380 help 381 The arch chooses to use the generic perf-NMI-based hardlockup 382 detector. Must define HAVE_PERF_EVENTS_NMI. 383 384config HAVE_NMI_WATCHDOG 385 depends on HAVE_NMI 386 bool 387 help 388 The arch provides a low level NMI watchdog. It provides 389 asm/nmi.h, and defines its own arch_touch_nmi_watchdog(). 390 391config HAVE_HARDLOCKUP_DETECTOR_ARCH 392 bool 393 select HAVE_NMI_WATCHDOG 394 help 395 The arch chooses to provide its own hardlockup detector, which is 396 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config 397 interfaces and parameters provided by hardlockup detector subsystem. 398 399config HAVE_PERF_REGS 400 bool 401 help 402 Support selective register dumps for perf events. This includes 403 bit-mapping of each registers and a unique architecture id. 404 405config HAVE_PERF_USER_STACK_DUMP 406 bool 407 help 408 Support user stack dumps for perf event samples. This needs 409 access to the user stack pointer which is not unified across 410 architectures. 411 412config HAVE_ARCH_JUMP_LABEL 413 bool 414 415config HAVE_ARCH_JUMP_LABEL_RELATIVE 416 bool 417 418config MMU_GATHER_TABLE_FREE 419 bool 420 421config MMU_GATHER_RCU_TABLE_FREE 422 bool 423 select MMU_GATHER_TABLE_FREE 424 425config MMU_GATHER_PAGE_SIZE 426 bool 427 428config MMU_GATHER_NO_RANGE 429 bool 430 431config MMU_GATHER_NO_GATHER 432 bool 433 depends on MMU_GATHER_TABLE_FREE 434 435config ARCH_WANT_IRQS_OFF_ACTIVATE_MM 436 bool 437 help 438 Temporary select until all architectures can be converted to have 439 irqs disabled over activate_mm. Architectures that do IPI based TLB 440 shootdowns should enable this. 441 442config ARCH_HAVE_NMI_SAFE_CMPXCHG 443 bool 444 445config HAVE_ALIGNED_STRUCT_PAGE 446 bool 447 help 448 This makes sure that struct pages are double word aligned and that 449 e.g. the SLUB allocator can perform double word atomic operations 450 on a struct page for better performance. However selecting this 451 might increase the size of a struct page by a word. 452 453config HAVE_CMPXCHG_LOCAL 454 bool 455 456config HAVE_CMPXCHG_DOUBLE 457 bool 458 459config ARCH_WEAK_RELEASE_ACQUIRE 460 bool 461 462config ARCH_WANT_IPC_PARSE_VERSION 463 bool 464 465config ARCH_WANT_COMPAT_IPC_PARSE_VERSION 466 bool 467 468config ARCH_WANT_OLD_COMPAT_IPC 469 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 470 bool 471 472config HAVE_ARCH_SECCOMP 473 bool 474 help 475 An arch should select this symbol to support seccomp mode 1 (the fixed 476 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn, 477 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment: 478 - __NR_seccomp_read_32 479 - __NR_seccomp_write_32 480 - __NR_seccomp_exit_32 481 - __NR_seccomp_sigreturn_32 482 483config HAVE_ARCH_SECCOMP_FILTER 484 bool 485 select HAVE_ARCH_SECCOMP 486 help 487 An arch should select this symbol if it provides all of these things: 488 - all the requirements for HAVE_ARCH_SECCOMP 489 - syscall_get_arch() 490 - syscall_get_arguments() 491 - syscall_rollback() 492 - syscall_set_return_value() 493 - SIGSYS siginfo_t support 494 - secure_computing is called from a ptrace_event()-safe context 495 - secure_computing return value is checked and a return value of -1 496 results in the system call being skipped immediately. 497 - seccomp syscall wired up 498 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE, 499 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If 500 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too. 501 502config SECCOMP 503 prompt "Enable seccomp to safely execute untrusted bytecode" 504 def_bool y 505 depends on HAVE_ARCH_SECCOMP 506 help 507 This kernel feature is useful for number crunching applications 508 that may need to handle untrusted bytecode during their 509 execution. By using pipes or other transports made available 510 to the process as file descriptors supporting the read/write 511 syscalls, it's possible to isolate those applications in their 512 own address space using seccomp. Once seccomp is enabled via 513 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be 514 disabled and the task is only allowed to execute a few safe 515 syscalls defined by each seccomp mode. 516 517 If unsure, say Y. 518 519config SECCOMP_FILTER 520 def_bool y 521 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 522 help 523 Enable tasks to build secure computing environments defined 524 in terms of Berkeley Packet Filter programs which implement 525 task-defined system call filtering polices. 526 527 See Documentation/userspace-api/seccomp_filter.rst for details. 528 529config SECCOMP_CACHE_DEBUG 530 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache" 531 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR 532 depends on PROC_FS 533 help 534 This enables the /proc/pid/seccomp_cache interface to monitor 535 seccomp cache data. The file format is subject to change. Reading 536 the file requires CAP_SYS_ADMIN. 537 538 This option is for debugging only. Enabling presents the risk that 539 an adversary may be able to infer the seccomp filter logic. 540 541 If unsure, say N. 542 543config HAVE_ARCH_STACKLEAK 544 bool 545 help 546 An architecture should select this if it has the code which 547 fills the used part of the kernel stack with the STACKLEAK_POISON 548 value before returning from system calls. 549 550config HAVE_STACKPROTECTOR 551 bool 552 help 553 An arch should select this symbol if: 554 - it has implemented a stack canary (e.g. __stack_chk_guard) 555 556config STACKPROTECTOR 557 bool "Stack Protector buffer overflow detection" 558 depends on HAVE_STACKPROTECTOR 559 depends on $(cc-option,-fstack-protector) 560 default y 561 help 562 This option turns on the "stack-protector" GCC feature. This 563 feature puts, at the beginning of functions, a canary value on 564 the stack just before the return address, and validates 565 the value just before actually returning. Stack based buffer 566 overflows (that need to overwrite this return address) now also 567 overwrite the canary, which gets detected and the attack is then 568 neutralized via a kernel panic. 569 570 Functions will have the stack-protector canary logic added if they 571 have an 8-byte or larger character array on the stack. 572 573 This feature requires gcc version 4.2 or above, or a distribution 574 gcc with the feature backported ("-fstack-protector"). 575 576 On an x86 "defconfig" build, this feature adds canary checks to 577 about 3% of all kernel functions, which increases kernel code size 578 by about 0.3%. 579 580config STACKPROTECTOR_STRONG 581 bool "Strong Stack Protector" 582 depends on STACKPROTECTOR 583 depends on $(cc-option,-fstack-protector-strong) 584 default y 585 help 586 Functions will have the stack-protector canary logic added in any 587 of the following conditions: 588 589 - local variable's address used as part of the right hand side of an 590 assignment or function argument 591 - local variable is an array (or union containing an array), 592 regardless of array type or length 593 - uses register local variables 594 595 This feature requires gcc version 4.9 or above, or a distribution 596 gcc with the feature backported ("-fstack-protector-strong"). 597 598 On an x86 "defconfig" build, this feature adds canary checks to 599 about 20% of all kernel functions, which increases the kernel code 600 size by about 2%. 601 602config ARCH_SUPPORTS_SHADOW_CALL_STACK 603 bool 604 help 605 An architecture should select this if it supports Clang's Shadow 606 Call Stack and implements runtime support for shadow stack 607 switching. 608 609config SHADOW_CALL_STACK 610 bool "Clang Shadow Call Stack" 611 depends on CC_IS_CLANG && ARCH_SUPPORTS_SHADOW_CALL_STACK 612 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER 613 help 614 This option enables Clang's Shadow Call Stack, which uses a 615 shadow stack to protect function return addresses from being 616 overwritten by an attacker. More information can be found in 617 Clang's documentation: 618 619 https://clang.llvm.org/docs/ShadowCallStack.html 620 621 Note that security guarantees in the kernel differ from the 622 ones documented for user space. The kernel must store addresses 623 of shadow stacks in memory, which means an attacker capable of 624 reading and writing arbitrary memory may be able to locate them 625 and hijack control flow by modifying the stacks. 626 627config LTO 628 bool 629 help 630 Selected if the kernel will be built using the compiler's LTO feature. 631 632config LTO_CLANG 633 bool 634 select LTO 635 help 636 Selected if the kernel will be built using Clang's LTO feature. 637 638config ARCH_SUPPORTS_LTO_CLANG 639 bool 640 help 641 An architecture should select this option if it supports: 642 - compiling with Clang, 643 - compiling inline assembly with Clang's integrated assembler, 644 - and linking with LLD. 645 646config ARCH_SUPPORTS_LTO_CLANG_THIN 647 bool 648 help 649 An architecture should select this option if it can support Clang's 650 ThinLTO mode. 651 652config HAS_LTO_CLANG 653 def_bool y 654 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM 655 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm) 656 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm) 657 depends on ARCH_SUPPORTS_LTO_CLANG 658 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT 659 depends on !KASAN || KASAN_HW_TAGS 660 depends on !GCOV_KERNEL 661 help 662 The compiler and Kconfig options support building with Clang's 663 LTO. 664 665choice 666 prompt "Link Time Optimization (LTO)" 667 default LTO_NONE 668 help 669 This option enables Link Time Optimization (LTO), which allows the 670 compiler to optimize binaries globally. 671 672 If unsure, select LTO_NONE. Note that LTO is very resource-intensive 673 so it's disabled by default. 674 675config LTO_NONE 676 bool "None" 677 help 678 Build the kernel normally, without Link Time Optimization (LTO). 679 680config LTO_CLANG_FULL 681 bool "Clang Full LTO (EXPERIMENTAL)" 682 depends on HAS_LTO_CLANG 683 depends on !COMPILE_TEST 684 select LTO_CLANG 685 help 686 This option enables Clang's full Link Time Optimization (LTO), which 687 allows the compiler to optimize the kernel globally. If you enable 688 this option, the compiler generates LLVM bitcode instead of ELF 689 object files, and the actual compilation from bitcode happens at 690 the LTO link step, which may take several minutes depending on the 691 kernel configuration. More information can be found from LLVM's 692 documentation: 693 694 https://llvm.org/docs/LinkTimeOptimization.html 695 696 During link time, this option can use a large amount of RAM, and 697 may take much longer than the ThinLTO option. 698 699config LTO_CLANG_THIN 700 bool "Clang ThinLTO (EXPERIMENTAL)" 701 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN 702 select LTO_CLANG 703 help 704 This option enables Clang's ThinLTO, which allows for parallel 705 optimization and faster incremental compiles compared to the 706 CONFIG_LTO_CLANG_FULL option. More information can be found 707 from Clang's documentation: 708 709 https://clang.llvm.org/docs/ThinLTO.html 710 711 If unsure, say Y. 712endchoice 713 714config ARCH_SUPPORTS_CFI_CLANG 715 bool 716 help 717 An architecture should select this option if it can support Clang's 718 Control-Flow Integrity (CFI) checking. 719 720config CFI_CLANG 721 bool "Use Clang's Control Flow Integrity (CFI)" 722 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG 723 # Clang >= 12: 724 # - https://bugs.llvm.org/show_bug.cgi?id=46258 725 # - https://bugs.llvm.org/show_bug.cgi?id=47479 726 depends on CLANG_VERSION >= 120000 727 select KALLSYMS 728 help 729 This option enables Clang’s forward-edge Control Flow Integrity 730 (CFI) checking, where the compiler injects a runtime check to each 731 indirect function call to ensure the target is a valid function with 732 the correct static type. This restricts possible call targets and 733 makes it more difficult for an attacker to exploit bugs that allow 734 the modification of stored function pointers. More information can be 735 found from Clang's documentation: 736 737 https://clang.llvm.org/docs/ControlFlowIntegrity.html 738 739config CFI_CLANG_SHADOW 740 bool "Use CFI shadow to speed up cross-module checks" 741 default y 742 depends on CFI_CLANG && MODULES 743 help 744 If you select this option, the kernel builds a fast look-up table of 745 CFI check functions in loaded modules to reduce performance overhead. 746 747 If unsure, say Y. 748 749config CFI_PERMISSIVE 750 bool "Use CFI in permissive mode" 751 depends on CFI_CLANG 752 help 753 When selected, Control Flow Integrity (CFI) violations result in a 754 warning instead of a kernel panic. This option should only be used 755 for finding indirect call type mismatches during development. 756 757 If unsure, say N. 758 759config HAVE_ARCH_WITHIN_STACK_FRAMES 760 bool 761 help 762 An architecture should select this if it can walk the kernel stack 763 frames to determine if an object is part of either the arguments 764 or local variables (i.e. that it excludes saved return addresses, 765 and similar) by implementing an inline arch_within_stack_frames(), 766 which is used by CONFIG_HARDENED_USERCOPY. 767 768config HAVE_CONTEXT_TRACKING 769 bool 770 help 771 Provide kernel/user boundaries probes necessary for subsystems 772 that need it, such as userspace RCU extended quiescent state. 773 Syscalls need to be wrapped inside user_exit()-user_enter(), either 774 optimized behind static key or through the slow path using TIF_NOHZ 775 flag. Exceptions handlers must be wrapped as well. Irqs are already 776 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal 777 handling on irq exit still need to be protected. 778 779config HAVE_CONTEXT_TRACKING_OFFSTACK 780 bool 781 help 782 Architecture neither relies on exception_enter()/exception_exit() 783 nor on schedule_user(). Also preempt_schedule_notrace() and 784 preempt_schedule_irq() can't be called in a preemptible section 785 while context tracking is CONTEXT_USER. This feature reflects a sane 786 entry implementation where the following requirements are met on 787 critical entry code, ie: before user_exit() or after user_enter(): 788 789 - Critical entry code isn't preemptible (or better yet: 790 not interruptible). 791 - No use of RCU read side critical sections, unless rcu_nmi_enter() 792 got called. 793 - No use of instrumentation, unless instrumentation_begin() got 794 called. 795 796config HAVE_TIF_NOHZ 797 bool 798 help 799 Arch relies on TIF_NOHZ and syscall slow path to implement context 800 tracking calls to user_enter()/user_exit(). 801 802config HAVE_VIRT_CPU_ACCOUNTING 803 bool 804 805config HAVE_VIRT_CPU_ACCOUNTING_IDLE 806 bool 807 help 808 Architecture has its own way to account idle CPU time and therefore 809 doesn't implement vtime_account_idle(). 810 811config ARCH_HAS_SCALED_CPUTIME 812 bool 813 814config HAVE_VIRT_CPU_ACCOUNTING_GEN 815 bool 816 default y if 64BIT 817 help 818 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. 819 Before enabling this option, arch code must be audited 820 to ensure there are no races in concurrent read/write of 821 cputime_t. For example, reading/writing 64-bit cputime_t on 822 some 32-bit arches may require multiple accesses, so proper 823 locking is needed to protect against concurrent accesses. 824 825config HAVE_IRQ_TIME_ACCOUNTING 826 bool 827 help 828 Archs need to ensure they use a high enough resolution clock to 829 support irq time accounting and then call enable_sched_clock_irqtime(). 830 831config HAVE_MOVE_PUD 832 bool 833 help 834 Architectures that select this are able to move page tables at the 835 PUD level. If there are only 3 page table levels, the move effectively 836 happens at the PGD level. 837 838config HAVE_MOVE_PMD 839 bool 840 help 841 Archs that select this are able to move page tables at the PMD level. 842 843config HAVE_ARCH_TRANSPARENT_HUGEPAGE 844 bool 845 846config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD 847 bool 848 849config HAVE_ARCH_HUGE_VMAP 850 bool 851 852# 853# Archs that select this would be capable of PMD-sized vmaps (i.e., 854# arch_vmap_pmd_supported() returns true), and they must make no assumptions 855# that vmalloc memory is mapped with PAGE_SIZE ptes. The VM_NO_HUGE_VMAP flag 856# can be used to prohibit arch-specific allocations from using hugepages to 857# help with this (e.g., modules may require it). 858# 859config HAVE_ARCH_HUGE_VMALLOC 860 depends on HAVE_ARCH_HUGE_VMAP 861 bool 862 863config ARCH_WANT_HUGE_PMD_SHARE 864 bool 865 866config HAVE_ARCH_SOFT_DIRTY 867 bool 868 869config HAVE_MOD_ARCH_SPECIFIC 870 bool 871 help 872 The arch uses struct mod_arch_specific to store data. Many arches 873 just need a simple module loader without arch specific data - those 874 should not enable this. 875 876config MODULES_USE_ELF_RELA 877 bool 878 help 879 Modules only use ELF RELA relocations. Modules with ELF REL 880 relocations will give an error. 881 882config MODULES_USE_ELF_REL 883 bool 884 help 885 Modules only use ELF REL relocations. Modules with ELF RELA 886 relocations will give an error. 887 888config HAVE_IRQ_EXIT_ON_IRQ_STACK 889 bool 890 help 891 Architecture doesn't only execute the irq handler on the irq stack 892 but also irq_exit(). This way we can process softirqs on this irq 893 stack instead of switching to a new one when we call __do_softirq() 894 in the end of an hardirq. 895 This spares a stack switch and improves cache usage on softirq 896 processing. 897 898config HAVE_SOFTIRQ_ON_OWN_STACK 899 bool 900 help 901 Architecture provides a function to run __do_softirq() on a 902 separate stack. 903 904config PGTABLE_LEVELS 905 int 906 default 2 907 908config ARCH_HAS_ELF_RANDOMIZE 909 bool 910 help 911 An architecture supports choosing randomized locations for 912 stack, mmap, brk, and ET_DYN. Defined functions: 913 - arch_mmap_rnd() 914 - arch_randomize_brk() 915 916config HAVE_ARCH_MMAP_RND_BITS 917 bool 918 help 919 An arch should select this symbol if it supports setting a variable 920 number of bits for use in establishing the base address for mmap 921 allocations, has MMU enabled and provides values for both: 922 - ARCH_MMAP_RND_BITS_MIN 923 - ARCH_MMAP_RND_BITS_MAX 924 925config HAVE_EXIT_THREAD 926 bool 927 help 928 An architecture implements exit_thread. 929 930config ARCH_MMAP_RND_BITS_MIN 931 int 932 933config ARCH_MMAP_RND_BITS_MAX 934 int 935 936config ARCH_MMAP_RND_BITS_DEFAULT 937 int 938 939config ARCH_MMAP_RND_BITS 940 int "Number of bits to use for ASLR of mmap base address" if EXPERT 941 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX 942 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT 943 default ARCH_MMAP_RND_BITS_MIN 944 depends on HAVE_ARCH_MMAP_RND_BITS 945 help 946 This value can be used to select the number of bits to use to 947 determine the random offset to the base address of vma regions 948 resulting from mmap allocations. This value will be bounded 949 by the architecture's minimum and maximum supported values. 950 951 This value can be changed after boot using the 952 /proc/sys/vm/mmap_rnd_bits tunable 953 954config HAVE_ARCH_MMAP_RND_COMPAT_BITS 955 bool 956 help 957 An arch should select this symbol if it supports running applications 958 in compatibility mode, supports setting a variable number of bits for 959 use in establishing the base address for mmap allocations, has MMU 960 enabled and provides values for both: 961 - ARCH_MMAP_RND_COMPAT_BITS_MIN 962 - ARCH_MMAP_RND_COMPAT_BITS_MAX 963 964config ARCH_MMAP_RND_COMPAT_BITS_MIN 965 int 966 967config ARCH_MMAP_RND_COMPAT_BITS_MAX 968 int 969 970config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 971 int 972 973config ARCH_MMAP_RND_COMPAT_BITS 974 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT 975 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX 976 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 977 default ARCH_MMAP_RND_COMPAT_BITS_MIN 978 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS 979 help 980 This value can be used to select the number of bits to use to 981 determine the random offset to the base address of vma regions 982 resulting from mmap allocations for compatible applications This 983 value will be bounded by the architecture's minimum and maximum 984 supported values. 985 986 This value can be changed after boot using the 987 /proc/sys/vm/mmap_rnd_compat_bits tunable 988 989config HAVE_ARCH_COMPAT_MMAP_BASES 990 bool 991 help 992 This allows 64bit applications to invoke 32-bit mmap() syscall 993 and vice-versa 32-bit applications to call 64-bit mmap(). 994 Required for applications doing different bitness syscalls. 995 996config PAGE_SIZE_LESS_THAN_64KB 997 def_bool y 998 depends on !ARM64_64K_PAGES 999 depends on !IA64_PAGE_SIZE_64KB 1000 depends on !PAGE_SIZE_64KB 1001 depends on !PARISC_PAGE_SIZE_64KB 1002 depends on !PPC_64K_PAGES 1003 depends on PAGE_SIZE_LESS_THAN_256KB 1004 1005config PAGE_SIZE_LESS_THAN_256KB 1006 def_bool y 1007 depends on !PPC_256K_PAGES 1008 depends on !PAGE_SIZE_256KB 1009 1010# This allows to use a set of generic functions to determine mmap base 1011# address by giving priority to top-down scheme only if the process 1012# is not in legacy mode (compat task, unlimited stack size or 1013# sysctl_legacy_va_layout). 1014# Architecture that selects this option can provide its own version of: 1015# - STACK_RND_MASK 1016config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT 1017 bool 1018 depends on MMU 1019 select ARCH_HAS_ELF_RANDOMIZE 1020 1021config HAVE_STACK_VALIDATION 1022 bool 1023 help 1024 Architecture supports the 'objtool check' host tool command, which 1025 performs compile-time stack metadata validation. 1026 1027config HAVE_RELIABLE_STACKTRACE 1028 bool 1029 help 1030 Architecture has either save_stack_trace_tsk_reliable() or 1031 arch_stack_walk_reliable() function which only returns a stack trace 1032 if it can guarantee the trace is reliable. 1033 1034config HAVE_ARCH_HASH 1035 bool 1036 default n 1037 help 1038 If this is set, the architecture provides an <asm/hash.h> 1039 file which provides platform-specific implementations of some 1040 functions in <linux/hash.h> or fs/namei.c. 1041 1042config HAVE_ARCH_NVRAM_OPS 1043 bool 1044 1045config ISA_BUS_API 1046 def_bool ISA 1047 1048# 1049# ABI hall of shame 1050# 1051config CLONE_BACKWARDS 1052 bool 1053 help 1054 Architecture has tls passed as the 4th argument of clone(2), 1055 not the 5th one. 1056 1057config CLONE_BACKWARDS2 1058 bool 1059 help 1060 Architecture has the first two arguments of clone(2) swapped. 1061 1062config CLONE_BACKWARDS3 1063 bool 1064 help 1065 Architecture has tls passed as the 3rd argument of clone(2), 1066 not the 5th one. 1067 1068config ODD_RT_SIGACTION 1069 bool 1070 help 1071 Architecture has unusual rt_sigaction(2) arguments 1072 1073config OLD_SIGSUSPEND 1074 bool 1075 help 1076 Architecture has old sigsuspend(2) syscall, of one-argument variety 1077 1078config OLD_SIGSUSPEND3 1079 bool 1080 help 1081 Even weirder antique ABI - three-argument sigsuspend(2) 1082 1083config OLD_SIGACTION 1084 bool 1085 help 1086 Architecture has old sigaction(2) syscall. Nope, not the same 1087 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2), 1088 but fairly different variant of sigaction(2), thanks to OSF/1 1089 compatibility... 1090 1091config COMPAT_OLD_SIGACTION 1092 bool 1093 1094config COMPAT_32BIT_TIME 1095 bool "Provide system calls for 32-bit time_t" 1096 default !64BIT || COMPAT 1097 help 1098 This enables 32 bit time_t support in addition to 64 bit time_t support. 1099 This is relevant on all 32-bit architectures, and 64-bit architectures 1100 as part of compat syscall handling. 1101 1102config ARCH_NO_PREEMPT 1103 bool 1104 1105config ARCH_EPHEMERAL_INODES 1106 def_bool n 1107 help 1108 An arch should select this symbol if it doesn't keep track of inode 1109 instances on its own, but instead relies on something else (e.g. the 1110 host kernel for an UML kernel). 1111 1112config ARCH_SUPPORTS_RT 1113 bool 1114 1115config CPU_NO_EFFICIENT_FFS 1116 def_bool n 1117 1118config HAVE_ARCH_VMAP_STACK 1119 def_bool n 1120 help 1121 An arch should select this symbol if it can support kernel stacks 1122 in vmalloc space. This means: 1123 1124 - vmalloc space must be large enough to hold many kernel stacks. 1125 This may rule out many 32-bit architectures. 1126 1127 - Stacks in vmalloc space need to work reliably. For example, if 1128 vmap page tables are created on demand, either this mechanism 1129 needs to work while the stack points to a virtual address with 1130 unpopulated page tables or arch code (switch_to() and switch_mm(), 1131 most likely) needs to ensure that the stack's page table entries 1132 are populated before running on a possibly unpopulated stack. 1133 1134 - If the stack overflows into a guard page, something reasonable 1135 should happen. The definition of "reasonable" is flexible, but 1136 instantly rebooting without logging anything would be unfriendly. 1137 1138config VMAP_STACK 1139 default y 1140 bool "Use a virtually-mapped stack" 1141 depends on HAVE_ARCH_VMAP_STACK 1142 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC 1143 help 1144 Enable this if you want the use virtually-mapped kernel stacks 1145 with guard pages. This causes kernel stack overflows to be 1146 caught immediately rather than causing difficult-to-diagnose 1147 corruption. 1148 1149 To use this with software KASAN modes, the architecture must support 1150 backing virtual mappings with real shadow memory, and KASAN_VMALLOC 1151 must be enabled. 1152 1153config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1154 def_bool n 1155 help 1156 An arch should select this symbol if it can support kernel stack 1157 offset randomization with calls to add_random_kstack_offset() 1158 during syscall entry and choose_random_kstack_offset() during 1159 syscall exit. Careful removal of -fstack-protector-strong and 1160 -fstack-protector should also be applied to the entry code and 1161 closely examined, as the artificial stack bump looks like an array 1162 to the compiler, so it will attempt to add canary checks regardless 1163 of the static branch state. 1164 1165config RANDOMIZE_KSTACK_OFFSET_DEFAULT 1166 bool "Randomize kernel stack offset on syscall entry" 1167 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1168 help 1169 The kernel stack offset can be randomized (after pt_regs) by 1170 roughly 5 bits of entropy, frustrating memory corruption 1171 attacks that depend on stack address determinism or 1172 cross-syscall address exposures. This feature is controlled 1173 by kernel boot param "randomize_kstack_offset=on/off", and this 1174 config chooses the default boot state. 1175 1176config ARCH_OPTIONAL_KERNEL_RWX 1177 def_bool n 1178 1179config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1180 def_bool n 1181 1182config ARCH_HAS_STRICT_KERNEL_RWX 1183 def_bool n 1184 1185config STRICT_KERNEL_RWX 1186 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX 1187 depends on ARCH_HAS_STRICT_KERNEL_RWX 1188 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1189 help 1190 If this is set, kernel text and rodata memory will be made read-only, 1191 and non-text memory will be made non-executable. This provides 1192 protection against certain security exploits (e.g. executing the heap 1193 or modifying text) 1194 1195 These features are considered standard security practice these days. 1196 You should say Y here in almost all cases. 1197 1198config ARCH_HAS_STRICT_MODULE_RWX 1199 def_bool n 1200 1201config STRICT_MODULE_RWX 1202 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX 1203 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES 1204 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1205 help 1206 If this is set, module text and rodata memory will be made read-only, 1207 and non-text memory will be made non-executable. This provides 1208 protection against certain security exploits (e.g. writing to text) 1209 1210# select if the architecture provides an asm/dma-direct.h header 1211config ARCH_HAS_PHYS_TO_DMA 1212 bool 1213 1214config HAVE_ARCH_COMPILER_H 1215 bool 1216 help 1217 An architecture can select this if it provides an 1218 asm/compiler.h header that should be included after 1219 linux/compiler-*.h in order to override macro definitions that those 1220 headers generally provide. 1221 1222config HAVE_ARCH_PREL32_RELOCATIONS 1223 bool 1224 help 1225 May be selected by an architecture if it supports place-relative 1226 32-bit relocations, both in the toolchain and in the module loader, 1227 in which case relative references can be used in special sections 1228 for PCI fixup, initcalls etc which are only half the size on 64 bit 1229 architectures, and don't require runtime relocation on relocatable 1230 kernels. 1231 1232config ARCH_USE_MEMREMAP_PROT 1233 bool 1234 1235config LOCK_EVENT_COUNTS 1236 bool "Locking event counts collection" 1237 depends on DEBUG_FS 1238 help 1239 Enable light-weight counting of various locking related events 1240 in the system with minimal performance impact. This reduces 1241 the chance of application behavior change because of timing 1242 differences. The counts are reported via debugfs. 1243 1244# Select if the architecture has support for applying RELR relocations. 1245config ARCH_HAS_RELR 1246 bool 1247 1248config RELR 1249 bool "Use RELR relocation packing" 1250 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR 1251 default y 1252 help 1253 Store the kernel's dynamic relocations in the RELR relocation packing 1254 format. Requires a compatible linker (LLD supports this feature), as 1255 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy 1256 are compatible). 1257 1258config ARCH_HAS_MEM_ENCRYPT 1259 bool 1260 1261config ARCH_HAS_CC_PLATFORM 1262 bool 1263 1264config HAVE_SPARSE_SYSCALL_NR 1265 bool 1266 help 1267 An architecture should select this if its syscall numbering is sparse 1268 to save space. For example, MIPS architecture has a syscall array with 1269 entries at 4000, 5000 and 6000 locations. This option turns on syscall 1270 related optimizations for a given architecture. 1271 1272config ARCH_HAS_VDSO_DATA 1273 bool 1274 1275config HAVE_STATIC_CALL 1276 bool 1277 1278config HAVE_STATIC_CALL_INLINE 1279 bool 1280 depends on HAVE_STATIC_CALL 1281 1282config HAVE_PREEMPT_DYNAMIC 1283 bool 1284 depends on HAVE_STATIC_CALL 1285 depends on GENERIC_ENTRY 1286 help 1287 Select this if the architecture support boot time preempt setting 1288 on top of static calls. It is strongly advised to support inline 1289 static call to avoid any overhead. 1290 1291config ARCH_WANT_LD_ORPHAN_WARN 1292 bool 1293 help 1294 An arch should select this symbol once all linker sections are explicitly 1295 included, size-asserted, or discarded in the linker scripts. This is 1296 important because we never want expected sections to be placed heuristically 1297 by the linker, since the locations of such sections can change between linker 1298 versions. 1299 1300config HAVE_ARCH_PFN_VALID 1301 bool 1302 1303config ARCH_SUPPORTS_DEBUG_PAGEALLOC 1304 bool 1305 1306config ARCH_SUPPORTS_PAGE_TABLE_CHECK 1307 bool 1308 1309config ARCH_SPLIT_ARG64 1310 bool 1311 help 1312 If a 32-bit architecture requires 64-bit arguments to be split into 1313 pairs of 32-bit arguments, select this option. 1314 1315config ARCH_HAS_ELFCORE_COMPAT 1316 bool 1317 1318config ARCH_HAS_PARANOID_L1D_FLUSH 1319 bool 1320 1321config DYNAMIC_SIGFRAME 1322 bool 1323 1324# Select, if arch has a named attribute group bound to NUMA device nodes. 1325config HAVE_ARCH_NODE_DEV_GROUP 1326 bool 1327 1328source "kernel/gcov/Kconfig" 1329 1330source "scripts/gcc-plugins/Kconfig" 1331 1332endmenu 1333