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 ARCH_HAS_SUBPAGE_FAULTS 28 bool 29 help 30 Select if the architecture can check permissions at sub-page 31 granularity (e.g. arm64 MTE). The probe_user_*() functions 32 must be implemented. 33 34config SET_FS 35 bool 36 37config HOTPLUG_SMT 38 bool 39 40config GENERIC_ENTRY 41 bool 42 43config KPROBES 44 bool "Kprobes" 45 depends on MODULES 46 depends on HAVE_KPROBES 47 select KALLSYMS 48 help 49 Kprobes allows you to trap at almost any kernel address and 50 execute a callback function. register_kprobe() establishes 51 a probepoint and specifies the callback. Kprobes is useful 52 for kernel debugging, non-intrusive instrumentation and testing. 53 If in doubt, say "N". 54 55config JUMP_LABEL 56 bool "Optimize very unlikely/likely branches" 57 depends on HAVE_ARCH_JUMP_LABEL 58 depends on CC_HAS_ASM_GOTO 59 help 60 This option enables a transparent branch optimization that 61 makes certain almost-always-true or almost-always-false branch 62 conditions even cheaper to execute within the kernel. 63 64 Certain performance-sensitive kernel code, such as trace points, 65 scheduler functionality, networking code and KVM have such 66 branches and include support for this optimization technique. 67 68 If it is detected that the compiler has support for "asm goto", 69 the kernel will compile such branches with just a nop 70 instruction. When the condition flag is toggled to true, the 71 nop will be converted to a jump instruction to execute the 72 conditional block of instructions. 73 74 This technique lowers overhead and stress on the branch prediction 75 of the processor and generally makes the kernel faster. The update 76 of the condition is slower, but those are always very rare. 77 78 ( On 32-bit x86, the necessary options added to the compiler 79 flags may increase the size of the kernel slightly. ) 80 81config STATIC_KEYS_SELFTEST 82 bool "Static key selftest" 83 depends on JUMP_LABEL 84 help 85 Boot time self-test of the branch patching code. 86 87config STATIC_CALL_SELFTEST 88 bool "Static call selftest" 89 depends on HAVE_STATIC_CALL 90 help 91 Boot time self-test of the call patching code. 92 93config OPTPROBES 94 def_bool y 95 depends on KPROBES && HAVE_OPTPROBES 96 select TASKS_RCU if PREEMPTION 97 98config KPROBES_ON_FTRACE 99 def_bool y 100 depends on KPROBES && HAVE_KPROBES_ON_FTRACE 101 depends on DYNAMIC_FTRACE_WITH_REGS 102 help 103 If function tracer is enabled and the arch supports full 104 passing of pt_regs to function tracing, then kprobes can 105 optimize on top of function tracing. 106 107config UPROBES 108 def_bool n 109 depends on ARCH_SUPPORTS_UPROBES 110 help 111 Uprobes is the user-space counterpart to kprobes: they 112 enable instrumentation applications (such as 'perf probe') 113 to establish unintrusive probes in user-space binaries and 114 libraries, by executing handler functions when the probes 115 are hit by user-space applications. 116 117 ( These probes come in the form of single-byte breakpoints, 118 managed by the kernel and kept transparent to the probed 119 application. ) 120 121config HAVE_64BIT_ALIGNED_ACCESS 122 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS 123 help 124 Some architectures require 64 bit accesses to be 64 bit 125 aligned, which also requires structs containing 64 bit values 126 to be 64 bit aligned too. This includes some 32 bit 127 architectures which can do 64 bit accesses, as well as 64 bit 128 architectures without unaligned access. 129 130 This symbol should be selected by an architecture if 64 bit 131 accesses are required to be 64 bit aligned in this way even 132 though it is not a 64 bit architecture. 133 134 See Documentation/core-api/unaligned-memory-access.rst for 135 more information on the topic of unaligned memory accesses. 136 137config HAVE_EFFICIENT_UNALIGNED_ACCESS 138 bool 139 help 140 Some architectures are unable to perform unaligned accesses 141 without the use of get_unaligned/put_unaligned. Others are 142 unable to perform such accesses efficiently (e.g. trap on 143 unaligned access and require fixing it up in the exception 144 handler.) 145 146 This symbol should be selected by an architecture if it can 147 perform unaligned accesses efficiently to allow different 148 code paths to be selected for these cases. Some network 149 drivers, for example, could opt to not fix up alignment 150 problems with received packets if doing so would not help 151 much. 152 153 See Documentation/core-api/unaligned-memory-access.rst for more 154 information on the topic of unaligned memory accesses. 155 156config ARCH_USE_BUILTIN_BSWAP 157 bool 158 help 159 Modern versions of GCC (since 4.4) have builtin functions 160 for handling byte-swapping. Using these, instead of the old 161 inline assembler that the architecture code provides in the 162 __arch_bswapXX() macros, allows the compiler to see what's 163 happening and offers more opportunity for optimisation. In 164 particular, the compiler will be able to combine the byteswap 165 with a nearby load or store and use load-and-swap or 166 store-and-swap instructions if the architecture has them. It 167 should almost *never* result in code which is worse than the 168 hand-coded assembler in <asm/swab.h>. But just in case it 169 does, the use of the builtins is optional. 170 171 Any architecture with load-and-swap or store-and-swap 172 instructions should set this. And it shouldn't hurt to set it 173 on architectures that don't have such instructions. 174 175config KRETPROBES 176 def_bool y 177 depends on KPROBES && HAVE_KRETPROBES 178 179config USER_RETURN_NOTIFIER 180 bool 181 depends on HAVE_USER_RETURN_NOTIFIER 182 help 183 Provide a kernel-internal notification when a cpu is about to 184 switch to user mode. 185 186config HAVE_IOREMAP_PROT 187 bool 188 189config HAVE_KPROBES 190 bool 191 192config HAVE_KRETPROBES 193 bool 194 195config HAVE_OPTPROBES 196 bool 197 198config HAVE_KPROBES_ON_FTRACE 199 bool 200 201config HAVE_FUNCTION_ERROR_INJECTION 202 bool 203 204config HAVE_NMI 205 bool 206 207config TRACE_IRQFLAGS_SUPPORT 208 bool 209 210config TRACE_IRQFLAGS_NMI_SUPPORT 211 bool 212 213# 214# An arch should select this if it provides all these things: 215# 216# task_pt_regs() in asm/processor.h or asm/ptrace.h 217# arch_has_single_step() if there is hardware single-step support 218# arch_has_block_step() if there is hardware block-step support 219# asm/syscall.h supplying asm-generic/syscall.h interface 220# linux/regset.h user_regset interfaces 221# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 222# TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit} 223# TIF_NOTIFY_RESUME calls tracehook_notify_resume() 224# signal delivery calls tracehook_signal_handler() 225# 226config HAVE_ARCH_TRACEHOOK 227 bool 228 229config HAVE_DMA_CONTIGUOUS 230 bool 231 232config GENERIC_SMP_IDLE_THREAD 233 bool 234 235config GENERIC_IDLE_POLL_SETUP 236 bool 237 238config ARCH_HAS_FORTIFY_SOURCE 239 bool 240 help 241 An architecture should select this when it can successfully 242 build and run with CONFIG_FORTIFY_SOURCE. 243 244# 245# Select if the arch provides a historic keepinit alias for the retain_initrd 246# command line option 247# 248config ARCH_HAS_KEEPINITRD 249 bool 250 251# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h 252config ARCH_HAS_SET_MEMORY 253 bool 254 255# Select if arch has all set_direct_map_invalid/default() functions 256config ARCH_HAS_SET_DIRECT_MAP 257 bool 258 259# 260# Select if the architecture provides the arch_dma_set_uncached symbol to 261# either provide an uncached segment alias for a DMA allocation, or 262# to remap the page tables in place. 263# 264config ARCH_HAS_DMA_SET_UNCACHED 265 bool 266 267# 268# Select if the architectures provides the arch_dma_clear_uncached symbol 269# to undo an in-place page table remap for uncached access. 270# 271config ARCH_HAS_DMA_CLEAR_UNCACHED 272 bool 273 274config ARCH_HAS_CPU_FINALIZE_INIT 275 bool 276 277# Select if arch init_task must go in the __init_task_data section 278config ARCH_TASK_STRUCT_ON_STACK 279 bool 280 281# Select if arch has its private alloc_task_struct() function 282config ARCH_TASK_STRUCT_ALLOCATOR 283 bool 284 285config HAVE_ARCH_THREAD_STRUCT_WHITELIST 286 bool 287 depends on !ARCH_TASK_STRUCT_ALLOCATOR 288 help 289 An architecture should select this to provide hardened usercopy 290 knowledge about what region of the thread_struct should be 291 whitelisted for copying to userspace. Normally this is only the 292 FPU registers. Specifically, arch_thread_struct_whitelist() 293 should be implemented. Without this, the entire thread_struct 294 field in task_struct will be left whitelisted. 295 296# Select if arch has its private alloc_thread_stack() function 297config ARCH_THREAD_STACK_ALLOCATOR 298 bool 299 300# Select if arch wants to size task_struct dynamically via arch_task_struct_size: 301config ARCH_WANTS_DYNAMIC_TASK_STRUCT 302 bool 303 304config ARCH_WANTS_NO_INSTR 305 bool 306 help 307 An architecture should select this if the noinstr macro is being used on 308 functions to denote that the toolchain should avoid instrumenting such 309 functions and is required for correctness. 310 311config ARCH_32BIT_OFF_T 312 bool 313 depends on !64BIT 314 help 315 All new 32-bit architectures should have 64-bit off_t type on 316 userspace side which corresponds to the loff_t kernel type. This 317 is the requirement for modern ABIs. Some existing architectures 318 still support 32-bit off_t. This option is enabled for all such 319 architectures explicitly. 320 321# Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat 322config ARCH_32BIT_USTAT_F_TINODE 323 bool 324 325config HAVE_ASM_MODVERSIONS 326 bool 327 help 328 This symbol should be selected by an architecture if it provides 329 <asm/asm-prototypes.h> to support the module versioning for symbols 330 exported from assembly code. 331 332config HAVE_REGS_AND_STACK_ACCESS_API 333 bool 334 help 335 This symbol should be selected by an architecture if it supports 336 the API needed to access registers and stack entries from pt_regs, 337 declared in asm/ptrace.h 338 For example the kprobes-based event tracer needs this API. 339 340config HAVE_RSEQ 341 bool 342 depends on HAVE_REGS_AND_STACK_ACCESS_API 343 help 344 This symbol should be selected by an architecture if it 345 supports an implementation of restartable sequences. 346 347config HAVE_FUNCTION_ARG_ACCESS_API 348 bool 349 help 350 This symbol should be selected by an architecture if it supports 351 the API needed to access function arguments from pt_regs, 352 declared in asm/ptrace.h 353 354config HAVE_HW_BREAKPOINT 355 bool 356 depends on PERF_EVENTS 357 358config HAVE_MIXED_BREAKPOINTS_REGS 359 bool 360 depends on HAVE_HW_BREAKPOINT 361 help 362 Depending on the arch implementation of hardware breakpoints, 363 some of them have separate registers for data and instruction 364 breakpoints addresses, others have mixed registers to store 365 them but define the access type in a control register. 366 Select this option if your arch implements breakpoints under the 367 latter fashion. 368 369config HAVE_USER_RETURN_NOTIFIER 370 bool 371 372config HAVE_PERF_EVENTS_NMI 373 bool 374 help 375 System hardware can generate an NMI using the perf event 376 subsystem. Also has support for calculating CPU cycle events 377 to determine how many clock cycles in a given period. 378 379config HAVE_HARDLOCKUP_DETECTOR_PERF 380 bool 381 depends on HAVE_PERF_EVENTS_NMI 382 help 383 The arch chooses to use the generic perf-NMI-based hardlockup 384 detector. Must define HAVE_PERF_EVENTS_NMI. 385 386config HAVE_NMI_WATCHDOG 387 depends on HAVE_NMI 388 bool 389 help 390 The arch provides a low level NMI watchdog. It provides 391 asm/nmi.h, and defines its own arch_touch_nmi_watchdog(). 392 393config HAVE_HARDLOCKUP_DETECTOR_ARCH 394 bool 395 select HAVE_NMI_WATCHDOG 396 help 397 The arch chooses to provide its own hardlockup detector, which is 398 a superset of the HAVE_NMI_WATCHDOG. It also conforms to config 399 interfaces and parameters provided by hardlockup detector subsystem. 400 401config HAVE_PERF_REGS 402 bool 403 help 404 Support selective register dumps for perf events. This includes 405 bit-mapping of each registers and a unique architecture id. 406 407config HAVE_PERF_USER_STACK_DUMP 408 bool 409 help 410 Support user stack dumps for perf event samples. This needs 411 access to the user stack pointer which is not unified across 412 architectures. 413 414config HAVE_ARCH_JUMP_LABEL 415 bool 416 417config HAVE_ARCH_JUMP_LABEL_RELATIVE 418 bool 419 420config MMU_GATHER_TABLE_FREE 421 bool 422 423config MMU_GATHER_RCU_TABLE_FREE 424 bool 425 select MMU_GATHER_TABLE_FREE 426 427config MMU_GATHER_PAGE_SIZE 428 bool 429 430config MMU_GATHER_NO_RANGE 431 bool 432 433config MMU_GATHER_NO_GATHER 434 bool 435 depends on MMU_GATHER_TABLE_FREE 436 437config ARCH_WANT_IRQS_OFF_ACTIVATE_MM 438 bool 439 help 440 Temporary select until all architectures can be converted to have 441 irqs disabled over activate_mm. Architectures that do IPI based TLB 442 shootdowns should enable this. 443 444config ARCH_HAVE_NMI_SAFE_CMPXCHG 445 bool 446 447config HAVE_ALIGNED_STRUCT_PAGE 448 bool 449 help 450 This makes sure that struct pages are double word aligned and that 451 e.g. the SLUB allocator can perform double word atomic operations 452 on a struct page for better performance. However selecting this 453 might increase the size of a struct page by a word. 454 455config HAVE_CMPXCHG_LOCAL 456 bool 457 458config HAVE_CMPXCHG_DOUBLE 459 bool 460 461config ARCH_WEAK_RELEASE_ACQUIRE 462 bool 463 464config ARCH_WANT_IPC_PARSE_VERSION 465 bool 466 467config ARCH_WANT_COMPAT_IPC_PARSE_VERSION 468 bool 469 470config ARCH_WANT_OLD_COMPAT_IPC 471 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 472 bool 473 474config HAVE_ARCH_SECCOMP 475 bool 476 help 477 An arch should select this symbol to support seccomp mode 1 (the fixed 478 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn, 479 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment: 480 - __NR_seccomp_read_32 481 - __NR_seccomp_write_32 482 - __NR_seccomp_exit_32 483 - __NR_seccomp_sigreturn_32 484 485config HAVE_ARCH_SECCOMP_FILTER 486 bool 487 select HAVE_ARCH_SECCOMP 488 help 489 An arch should select this symbol if it provides all of these things: 490 - all the requirements for HAVE_ARCH_SECCOMP 491 - syscall_get_arch() 492 - syscall_get_arguments() 493 - syscall_rollback() 494 - syscall_set_return_value() 495 - SIGSYS siginfo_t support 496 - secure_computing is called from a ptrace_event()-safe context 497 - secure_computing return value is checked and a return value of -1 498 results in the system call being skipped immediately. 499 - seccomp syscall wired up 500 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE, 501 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If 502 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too. 503 504config SECCOMP 505 prompt "Enable seccomp to safely execute untrusted bytecode" 506 def_bool y 507 depends on HAVE_ARCH_SECCOMP 508 help 509 This kernel feature is useful for number crunching applications 510 that may need to handle untrusted bytecode during their 511 execution. By using pipes or other transports made available 512 to the process as file descriptors supporting the read/write 513 syscalls, it's possible to isolate those applications in their 514 own address space using seccomp. Once seccomp is enabled via 515 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be 516 disabled and the task is only allowed to execute a few safe 517 syscalls defined by each seccomp mode. 518 519 If unsure, say Y. 520 521config SECCOMP_FILTER 522 def_bool y 523 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 524 help 525 Enable tasks to build secure computing environments defined 526 in terms of Berkeley Packet Filter programs which implement 527 task-defined system call filtering polices. 528 529 See Documentation/userspace-api/seccomp_filter.rst for details. 530 531config SECCOMP_CACHE_DEBUG 532 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache" 533 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR 534 depends on PROC_FS 535 help 536 This enables the /proc/pid/seccomp_cache interface to monitor 537 seccomp cache data. The file format is subject to change. Reading 538 the file requires CAP_SYS_ADMIN. 539 540 This option is for debugging only. Enabling presents the risk that 541 an adversary may be able to infer the seccomp filter logic. 542 543 If unsure, say N. 544 545config HAVE_ARCH_STACKLEAK 546 bool 547 help 548 An architecture should select this if it has the code which 549 fills the used part of the kernel stack with the STACKLEAK_POISON 550 value before returning from system calls. 551 552config HAVE_STACKPROTECTOR 553 bool 554 help 555 An arch should select this symbol if: 556 - it has implemented a stack canary (e.g. __stack_chk_guard) 557 558config STACKPROTECTOR 559 bool "Stack Protector buffer overflow detection" 560 depends on HAVE_STACKPROTECTOR 561 depends on $(cc-option,-fstack-protector) 562 default y 563 help 564 This option turns on the "stack-protector" GCC feature. This 565 feature puts, at the beginning of functions, a canary value on 566 the stack just before the return address, and validates 567 the value just before actually returning. Stack based buffer 568 overflows (that need to overwrite this return address) now also 569 overwrite the canary, which gets detected and the attack is then 570 neutralized via a kernel panic. 571 572 Functions will have the stack-protector canary logic added if they 573 have an 8-byte or larger character array on the stack. 574 575 This feature requires gcc version 4.2 or above, or a distribution 576 gcc with the feature backported ("-fstack-protector"). 577 578 On an x86 "defconfig" build, this feature adds canary checks to 579 about 3% of all kernel functions, which increases kernel code size 580 by about 0.3%. 581 582config STACKPROTECTOR_STRONG 583 bool "Strong Stack Protector" 584 depends on STACKPROTECTOR 585 depends on $(cc-option,-fstack-protector-strong) 586 default y 587 help 588 Functions will have the stack-protector canary logic added in any 589 of the following conditions: 590 591 - local variable's address used as part of the right hand side of an 592 assignment or function argument 593 - local variable is an array (or union containing an array), 594 regardless of array type or length 595 - uses register local variables 596 597 This feature requires gcc version 4.9 or above, or a distribution 598 gcc with the feature backported ("-fstack-protector-strong"). 599 600 On an x86 "defconfig" build, this feature adds canary checks to 601 about 20% of all kernel functions, which increases the kernel code 602 size by about 2%. 603 604config ARCH_SUPPORTS_SHADOW_CALL_STACK 605 bool 606 help 607 An architecture should select this if it supports Clang's Shadow 608 Call Stack and implements runtime support for shadow stack 609 switching. 610 611config SHADOW_CALL_STACK 612 bool "Clang Shadow Call Stack" 613 depends on CC_IS_CLANG && ARCH_SUPPORTS_SHADOW_CALL_STACK 614 depends on DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER 615 depends on MMU 616 help 617 This option enables Clang's Shadow Call Stack, which uses a 618 shadow stack to protect function return addresses from being 619 overwritten by an attacker. More information can be found in 620 Clang's documentation: 621 622 https://clang.llvm.org/docs/ShadowCallStack.html 623 624 Note that security guarantees in the kernel differ from the 625 ones documented for user space. The kernel must store addresses 626 of shadow stacks in memory, which means an attacker capable of 627 reading and writing arbitrary memory may be able to locate them 628 and hijack control flow by modifying the stacks. 629 630config DYNAMIC_SCS 631 bool 632 help 633 Set by the arch code if it relies on code patching to insert the 634 shadow call stack push and pop instructions rather than on the 635 compiler. 636 637config LTO 638 bool 639 help 640 Selected if the kernel will be built using the compiler's LTO feature. 641 642config LTO_CLANG 643 bool 644 select LTO 645 help 646 Selected if the kernel will be built using Clang's LTO feature. 647 648config ARCH_SUPPORTS_LTO_CLANG 649 bool 650 help 651 An architecture should select this option if it supports: 652 - compiling with Clang, 653 - compiling inline assembly with Clang's integrated assembler, 654 - and linking with LLD. 655 656config ARCH_SUPPORTS_LTO_CLANG_THIN 657 bool 658 help 659 An architecture should select this option if it can support Clang's 660 ThinLTO mode. 661 662config HAS_LTO_CLANG 663 def_bool y 664 # Clang >= 11: https://github.com/ClangBuiltLinux/linux/issues/510 665 depends on CC_IS_CLANG && CLANG_VERSION >= 110000 && LD_IS_LLD && AS_IS_LLVM 666 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm) 667 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm) 668 depends on ARCH_SUPPORTS_LTO_CLANG 669 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT 670 depends on !KASAN || KASAN_HW_TAGS 671 depends on !GCOV_KERNEL 672 help 673 The compiler and Kconfig options support building with Clang's 674 LTO. 675 676choice 677 prompt "Link Time Optimization (LTO)" 678 default LTO_NONE 679 help 680 This option enables Link Time Optimization (LTO), which allows the 681 compiler to optimize binaries globally. 682 683 If unsure, select LTO_NONE. Note that LTO is very resource-intensive 684 so it's disabled by default. 685 686config LTO_NONE 687 bool "None" 688 help 689 Build the kernel normally, without Link Time Optimization (LTO). 690 691config LTO_CLANG_FULL 692 bool "Clang Full LTO (EXPERIMENTAL)" 693 depends on HAS_LTO_CLANG 694 depends on !COMPILE_TEST 695 select LTO_CLANG 696 help 697 This option enables Clang's full Link Time Optimization (LTO), which 698 allows the compiler to optimize the kernel globally. If you enable 699 this option, the compiler generates LLVM bitcode instead of ELF 700 object files, and the actual compilation from bitcode happens at 701 the LTO link step, which may take several minutes depending on the 702 kernel configuration. More information can be found from LLVM's 703 documentation: 704 705 https://llvm.org/docs/LinkTimeOptimization.html 706 707 During link time, this option can use a large amount of RAM, and 708 may take much longer than the ThinLTO option. 709 710config LTO_CLANG_THIN 711 bool "Clang ThinLTO (EXPERIMENTAL)" 712 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN 713 select LTO_CLANG 714 help 715 This option enables Clang's ThinLTO, which allows for parallel 716 optimization and faster incremental compiles compared to the 717 CONFIG_LTO_CLANG_FULL option. More information can be found 718 from Clang's documentation: 719 720 https://clang.llvm.org/docs/ThinLTO.html 721 722 If unsure, say Y. 723endchoice 724 725config ARCH_SUPPORTS_CFI_CLANG 726 bool 727 help 728 An architecture should select this option if it can support Clang's 729 Control-Flow Integrity (CFI) checking. 730 731config CFI_CLANG 732 bool "Use Clang's Control Flow Integrity (CFI)" 733 depends on LTO_CLANG && ARCH_SUPPORTS_CFI_CLANG 734 depends on CLANG_VERSION >= 140000 735 select KALLSYMS 736 help 737 This option enables Clang’s forward-edge Control Flow Integrity 738 (CFI) checking, where the compiler injects a runtime check to each 739 indirect function call to ensure the target is a valid function with 740 the correct static type. This restricts possible call targets and 741 makes it more difficult for an attacker to exploit bugs that allow 742 the modification of stored function pointers. More information can be 743 found from Clang's documentation: 744 745 https://clang.llvm.org/docs/ControlFlowIntegrity.html 746 747config CFI_CLANG_SHADOW 748 bool "Use CFI shadow to speed up cross-module checks" 749 default y 750 depends on CFI_CLANG && MODULES 751 help 752 If you select this option, the kernel builds a fast look-up table of 753 CFI check functions in loaded modules to reduce performance overhead. 754 755 If unsure, say Y. 756 757config CFI_PERMISSIVE 758 bool "Use CFI in permissive mode" 759 depends on CFI_CLANG 760 help 761 When selected, Control Flow Integrity (CFI) violations result in a 762 warning instead of a kernel panic. This option should only be used 763 for finding indirect call type mismatches during development. 764 765 If unsure, say N. 766 767config HAVE_ARCH_WITHIN_STACK_FRAMES 768 bool 769 help 770 An architecture should select this if it can walk the kernel stack 771 frames to determine if an object is part of either the arguments 772 or local variables (i.e. that it excludes saved return addresses, 773 and similar) by implementing an inline arch_within_stack_frames(), 774 which is used by CONFIG_HARDENED_USERCOPY. 775 776config HAVE_CONTEXT_TRACKING 777 bool 778 help 779 Provide kernel/user boundaries probes necessary for subsystems 780 that need it, such as userspace RCU extended quiescent state. 781 Syscalls need to be wrapped inside user_exit()-user_enter(), either 782 optimized behind static key or through the slow path using TIF_NOHZ 783 flag. Exceptions handlers must be wrapped as well. Irqs are already 784 protected inside rcu_irq_enter/rcu_irq_exit() but preemption or signal 785 handling on irq exit still need to be protected. 786 787config HAVE_CONTEXT_TRACKING_OFFSTACK 788 bool 789 help 790 Architecture neither relies on exception_enter()/exception_exit() 791 nor on schedule_user(). Also preempt_schedule_notrace() and 792 preempt_schedule_irq() can't be called in a preemptible section 793 while context tracking is CONTEXT_USER. This feature reflects a sane 794 entry implementation where the following requirements are met on 795 critical entry code, ie: before user_exit() or after user_enter(): 796 797 - Critical entry code isn't preemptible (or better yet: 798 not interruptible). 799 - No use of RCU read side critical sections, unless rcu_nmi_enter() 800 got called. 801 - No use of instrumentation, unless instrumentation_begin() got 802 called. 803 804config HAVE_TIF_NOHZ 805 bool 806 help 807 Arch relies on TIF_NOHZ and syscall slow path to implement context 808 tracking calls to user_enter()/user_exit(). 809 810config HAVE_VIRT_CPU_ACCOUNTING 811 bool 812 813config HAVE_VIRT_CPU_ACCOUNTING_IDLE 814 bool 815 help 816 Architecture has its own way to account idle CPU time and therefore 817 doesn't implement vtime_account_idle(). 818 819config ARCH_HAS_SCALED_CPUTIME 820 bool 821 822config HAVE_VIRT_CPU_ACCOUNTING_GEN 823 bool 824 default y if 64BIT 825 help 826 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. 827 Before enabling this option, arch code must be audited 828 to ensure there are no races in concurrent read/write of 829 cputime_t. For example, reading/writing 64-bit cputime_t on 830 some 32-bit arches may require multiple accesses, so proper 831 locking is needed to protect against concurrent accesses. 832 833config HAVE_IRQ_TIME_ACCOUNTING 834 bool 835 help 836 Archs need to ensure they use a high enough resolution clock to 837 support irq time accounting and then call enable_sched_clock_irqtime(). 838 839config HAVE_MOVE_PUD 840 bool 841 help 842 Architectures that select this are able to move page tables at the 843 PUD level. If there are only 3 page table levels, the move effectively 844 happens at the PGD level. 845 846config HAVE_MOVE_PMD 847 bool 848 help 849 Archs that select this are able to move page tables at the PMD level. 850 851config HAVE_ARCH_TRANSPARENT_HUGEPAGE 852 bool 853 854config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD 855 bool 856 857config HAVE_ARCH_HUGE_VMAP 858 bool 859 860# 861# Archs that select this would be capable of PMD-sized vmaps (i.e., 862# arch_vmap_pmd_supported() returns true), and they must make no assumptions 863# that vmalloc memory is mapped with PAGE_SIZE ptes. The VM_NO_HUGE_VMAP flag 864# can be used to prohibit arch-specific allocations from using hugepages to 865# help with this (e.g., modules may require it). 866# 867config HAVE_ARCH_HUGE_VMALLOC 868 depends on HAVE_ARCH_HUGE_VMAP 869 bool 870 871config ARCH_WANT_HUGE_PMD_SHARE 872 bool 873 874config HAVE_ARCH_SOFT_DIRTY 875 bool 876 877config HAVE_MOD_ARCH_SPECIFIC 878 bool 879 help 880 The arch uses struct mod_arch_specific to store data. Many arches 881 just need a simple module loader without arch specific data - those 882 should not enable this. 883 884config MODULES_USE_ELF_RELA 885 bool 886 help 887 Modules only use ELF RELA relocations. Modules with ELF REL 888 relocations will give an error. 889 890config MODULES_USE_ELF_REL 891 bool 892 help 893 Modules only use ELF REL relocations. Modules with ELF RELA 894 relocations will give an error. 895 896config HAVE_IRQ_EXIT_ON_IRQ_STACK 897 bool 898 help 899 Architecture doesn't only execute the irq handler on the irq stack 900 but also irq_exit(). This way we can process softirqs on this irq 901 stack instead of switching to a new one when we call __do_softirq() 902 in the end of an hardirq. 903 This spares a stack switch and improves cache usage on softirq 904 processing. 905 906config HAVE_SOFTIRQ_ON_OWN_STACK 907 bool 908 help 909 Architecture provides a function to run __do_softirq() on a 910 separate stack. 911 912config PGTABLE_LEVELS 913 int 914 default 2 915 916config ARCH_HAS_ELF_RANDOMIZE 917 bool 918 help 919 An architecture supports choosing randomized locations for 920 stack, mmap, brk, and ET_DYN. Defined functions: 921 - arch_mmap_rnd() 922 - arch_randomize_brk() 923 924config HAVE_ARCH_MMAP_RND_BITS 925 bool 926 help 927 An arch should select this symbol if it supports setting a variable 928 number of bits for use in establishing the base address for mmap 929 allocations, has MMU enabled and provides values for both: 930 - ARCH_MMAP_RND_BITS_MIN 931 - ARCH_MMAP_RND_BITS_MAX 932 933config HAVE_EXIT_THREAD 934 bool 935 help 936 An architecture implements exit_thread. 937 938config ARCH_MMAP_RND_BITS_MIN 939 int 940 941config ARCH_MMAP_RND_BITS_MAX 942 int 943 944config ARCH_MMAP_RND_BITS_DEFAULT 945 int 946 947config ARCH_MMAP_RND_BITS 948 int "Number of bits to use for ASLR of mmap base address" if EXPERT 949 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX 950 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT 951 default ARCH_MMAP_RND_BITS_MIN 952 depends on HAVE_ARCH_MMAP_RND_BITS 953 help 954 This value can be used to select the number of bits to use to 955 determine the random offset to the base address of vma regions 956 resulting from mmap allocations. This value will be bounded 957 by the architecture's minimum and maximum supported values. 958 959 This value can be changed after boot using the 960 /proc/sys/vm/mmap_rnd_bits tunable 961 962config HAVE_ARCH_MMAP_RND_COMPAT_BITS 963 bool 964 help 965 An arch should select this symbol if it supports running applications 966 in compatibility mode, supports setting a variable number of bits for 967 use in establishing the base address for mmap allocations, has MMU 968 enabled and provides values for both: 969 - ARCH_MMAP_RND_COMPAT_BITS_MIN 970 - ARCH_MMAP_RND_COMPAT_BITS_MAX 971 972config ARCH_MMAP_RND_COMPAT_BITS_MIN 973 int 974 975config ARCH_MMAP_RND_COMPAT_BITS_MAX 976 int 977 978config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 979 int 980 981config ARCH_MMAP_RND_COMPAT_BITS 982 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT 983 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX 984 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 985 default ARCH_MMAP_RND_COMPAT_BITS_MIN 986 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS 987 help 988 This value can be used to select the number of bits to use to 989 determine the random offset to the base address of vma regions 990 resulting from mmap allocations for compatible applications This 991 value will be bounded by the architecture's minimum and maximum 992 supported values. 993 994 This value can be changed after boot using the 995 /proc/sys/vm/mmap_rnd_compat_bits tunable 996 997config HAVE_ARCH_COMPAT_MMAP_BASES 998 bool 999 help 1000 This allows 64bit applications to invoke 32-bit mmap() syscall 1001 and vice-versa 32-bit applications to call 64-bit mmap(). 1002 Required for applications doing different bitness syscalls. 1003 1004# This allows to use a set of generic functions to determine mmap base 1005# address by giving priority to top-down scheme only if the process 1006# is not in legacy mode (compat task, unlimited stack size or 1007# sysctl_legacy_va_layout). 1008# Architecture that selects this option can provide its own version of: 1009# - STACK_RND_MASK 1010config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT 1011 bool 1012 depends on MMU 1013 select ARCH_HAS_ELF_RANDOMIZE 1014 1015config HAVE_STACK_VALIDATION 1016 bool 1017 help 1018 Architecture supports the 'objtool check' host tool command, which 1019 performs compile-time stack metadata validation. 1020 1021config HAVE_RELIABLE_STACKTRACE 1022 bool 1023 help 1024 Architecture has either save_stack_trace_tsk_reliable() or 1025 arch_stack_walk_reliable() function which only returns a stack trace 1026 if it can guarantee the trace is reliable. 1027 1028config HAVE_ARCH_HASH 1029 bool 1030 default n 1031 help 1032 If this is set, the architecture provides an <asm/hash.h> 1033 file which provides platform-specific implementations of some 1034 functions in <linux/hash.h> or fs/namei.c. 1035 1036config HAVE_ARCH_NVRAM_OPS 1037 bool 1038 1039config ISA_BUS_API 1040 def_bool ISA 1041 1042# 1043# ABI hall of shame 1044# 1045config CLONE_BACKWARDS 1046 bool 1047 help 1048 Architecture has tls passed as the 4th argument of clone(2), 1049 not the 5th one. 1050 1051config CLONE_BACKWARDS2 1052 bool 1053 help 1054 Architecture has the first two arguments of clone(2) swapped. 1055 1056config CLONE_BACKWARDS3 1057 bool 1058 help 1059 Architecture has tls passed as the 3rd argument of clone(2), 1060 not the 5th one. 1061 1062config ODD_RT_SIGACTION 1063 bool 1064 help 1065 Architecture has unusual rt_sigaction(2) arguments 1066 1067config OLD_SIGSUSPEND 1068 bool 1069 help 1070 Architecture has old sigsuspend(2) syscall, of one-argument variety 1071 1072config OLD_SIGSUSPEND3 1073 bool 1074 help 1075 Even weirder antique ABI - three-argument sigsuspend(2) 1076 1077config OLD_SIGACTION 1078 bool 1079 help 1080 Architecture has old sigaction(2) syscall. Nope, not the same 1081 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2), 1082 but fairly different variant of sigaction(2), thanks to OSF/1 1083 compatibility... 1084 1085config COMPAT_OLD_SIGACTION 1086 bool 1087 1088config COMPAT_32BIT_TIME 1089 bool "Provide system calls for 32-bit time_t" 1090 default !64BIT || COMPAT 1091 help 1092 This enables 32 bit time_t support in addition to 64 bit time_t support. 1093 This is relevant on all 32-bit architectures, and 64-bit architectures 1094 as part of compat syscall handling. 1095 1096config ARCH_NO_PREEMPT 1097 bool 1098 1099config ARCH_EPHEMERAL_INODES 1100 def_bool n 1101 help 1102 An arch should select this symbol if it doesn't keep track of inode 1103 instances on its own, but instead relies on something else (e.g. the 1104 host kernel for an UML kernel). 1105 1106config ARCH_SUPPORTS_RT 1107 bool 1108 1109config CPU_NO_EFFICIENT_FFS 1110 def_bool n 1111 1112config HAVE_ARCH_VMAP_STACK 1113 def_bool n 1114 help 1115 An arch should select this symbol if it can support kernel stacks 1116 in vmalloc space. This means: 1117 1118 - vmalloc space must be large enough to hold many kernel stacks. 1119 This may rule out many 32-bit architectures. 1120 1121 - Stacks in vmalloc space need to work reliably. For example, if 1122 vmap page tables are created on demand, either this mechanism 1123 needs to work while the stack points to a virtual address with 1124 unpopulated page tables or arch code (switch_to() and switch_mm(), 1125 most likely) needs to ensure that the stack's page table entries 1126 are populated before running on a possibly unpopulated stack. 1127 1128 - If the stack overflows into a guard page, something reasonable 1129 should happen. The definition of "reasonable" is flexible, but 1130 instantly rebooting without logging anything would be unfriendly. 1131 1132config VMAP_STACK 1133 default y 1134 bool "Use a virtually-mapped stack" 1135 depends on HAVE_ARCH_VMAP_STACK 1136 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC 1137 help 1138 Enable this if you want the use virtually-mapped kernel stacks 1139 with guard pages. This causes kernel stack overflows to be 1140 caught immediately rather than causing difficult-to-diagnose 1141 corruption. 1142 1143 To use this with software KASAN modes, the architecture must support 1144 backing virtual mappings with real shadow memory, and KASAN_VMALLOC 1145 must be enabled. 1146 1147config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1148 def_bool n 1149 help 1150 An arch should select this symbol if it can support kernel stack 1151 offset randomization with calls to add_random_kstack_offset() 1152 during syscall entry and choose_random_kstack_offset() during 1153 syscall exit. Careful removal of -fstack-protector-strong and 1154 -fstack-protector should also be applied to the entry code and 1155 closely examined, as the artificial stack bump looks like an array 1156 to the compiler, so it will attempt to add canary checks regardless 1157 of the static branch state. 1158 1159config RANDOMIZE_KSTACK_OFFSET_DEFAULT 1160 bool "Randomize kernel stack offset on syscall entry" 1161 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1162 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000 1163 help 1164 The kernel stack offset can be randomized (after pt_regs) by 1165 roughly 5 bits of entropy, frustrating memory corruption 1166 attacks that depend on stack address determinism or 1167 cross-syscall address exposures. This feature is controlled 1168 by kernel boot param "randomize_kstack_offset=on/off", and this 1169 config chooses the default boot state. 1170 1171config ARCH_OPTIONAL_KERNEL_RWX 1172 def_bool n 1173 1174config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1175 def_bool n 1176 1177config ARCH_HAS_STRICT_KERNEL_RWX 1178 def_bool n 1179 1180config STRICT_KERNEL_RWX 1181 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX 1182 depends on ARCH_HAS_STRICT_KERNEL_RWX 1183 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1184 help 1185 If this is set, kernel text and rodata memory will be made read-only, 1186 and non-text memory will be made non-executable. This provides 1187 protection against certain security exploits (e.g. executing the heap 1188 or modifying text) 1189 1190 These features are considered standard security practice these days. 1191 You should say Y here in almost all cases. 1192 1193config ARCH_HAS_STRICT_MODULE_RWX 1194 def_bool n 1195 1196config STRICT_MODULE_RWX 1197 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX 1198 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES 1199 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1200 help 1201 If this is set, module text and rodata memory will be made read-only, 1202 and non-text memory will be made non-executable. This provides 1203 protection against certain security exploits (e.g. writing to text) 1204 1205# select if the architecture provides an asm/dma-direct.h header 1206config ARCH_HAS_PHYS_TO_DMA 1207 bool 1208 1209config HAVE_ARCH_COMPILER_H 1210 bool 1211 help 1212 An architecture can select this if it provides an 1213 asm/compiler.h header that should be included after 1214 linux/compiler-*.h in order to override macro definitions that those 1215 headers generally provide. 1216 1217config HAVE_ARCH_PREL32_RELOCATIONS 1218 bool 1219 help 1220 May be selected by an architecture if it supports place-relative 1221 32-bit relocations, both in the toolchain and in the module loader, 1222 in which case relative references can be used in special sections 1223 for PCI fixup, initcalls etc which are only half the size on 64 bit 1224 architectures, and don't require runtime relocation on relocatable 1225 kernels. 1226 1227config ARCH_USE_MEMREMAP_PROT 1228 bool 1229 1230config LOCK_EVENT_COUNTS 1231 bool "Locking event counts collection" 1232 depends on DEBUG_FS 1233 help 1234 Enable light-weight counting of various locking related events 1235 in the system with minimal performance impact. This reduces 1236 the chance of application behavior change because of timing 1237 differences. The counts are reported via debugfs. 1238 1239# Select if the architecture has support for applying RELR relocations. 1240config ARCH_HAS_RELR 1241 bool 1242 1243config RELR 1244 bool "Use RELR relocation packing" 1245 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR 1246 default y 1247 help 1248 Store the kernel's dynamic relocations in the RELR relocation packing 1249 format. Requires a compatible linker (LLD supports this feature), as 1250 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy 1251 are compatible). 1252 1253config ARCH_HAS_MEM_ENCRYPT 1254 bool 1255 1256config ARCH_HAS_MEM_RELINQUISH 1257 bool 1258 1259config ARCH_HAS_CC_PLATFORM 1260 bool 1261 1262config HAVE_SPARSE_SYSCALL_NR 1263 bool 1264 help 1265 An architecture should select this if its syscall numbering is sparse 1266 to save space. For example, MIPS architecture has a syscall array with 1267 entries at 4000, 5000 and 6000 locations. This option turns on syscall 1268 related optimizations for a given architecture. 1269 1270config ARCH_HAS_VDSO_DATA 1271 bool 1272 1273config HAVE_STATIC_CALL 1274 bool 1275 1276config HAVE_STATIC_CALL_INLINE 1277 bool 1278 depends on HAVE_STATIC_CALL 1279 1280config HAVE_PREEMPT_DYNAMIC 1281 bool 1282 depends on HAVE_STATIC_CALL 1283 depends on GENERIC_ENTRY 1284 help 1285 Select this if the architecture support boot time preempt setting 1286 on top of static calls. It is strongly advised to support inline 1287 static call to avoid any overhead. 1288 1289config ARCH_WANT_LD_ORPHAN_WARN 1290 bool 1291 help 1292 An arch should select this symbol once all linker sections are explicitly 1293 included, size-asserted, or discarded in the linker scripts. This is 1294 important because we never want expected sections to be placed heuristically 1295 by the linker, since the locations of such sections can change between linker 1296 versions. 1297 1298config HAVE_ARCH_PFN_VALID 1299 bool 1300 1301config ARCH_SUPPORTS_DEBUG_PAGEALLOC 1302 bool 1303 1304config ARCH_SPLIT_ARG64 1305 bool 1306 help 1307 If a 32-bit architecture requires 64-bit arguments to be split into 1308 pairs of 32-bit arguments, select this option. 1309 1310config ARCH_HAS_ELFCORE_COMPAT 1311 bool 1312 1313config ARCH_HAS_PARANOID_L1D_FLUSH 1314 bool 1315 1316config ARCH_HAVE_TRACE_MMIO_ACCESS 1317 bool 1318 1319config ARCH_HAS_NONLEAF_PMD_YOUNG 1320 bool 1321 help 1322 Architectures that select this option are capable of setting the 1323 accessed bit in non-leaf PMD entries when using them as part of linear 1324 address translations. Page table walkers that clear the accessed bit 1325 may use this capability to reduce their search space. 1326 1327source "kernel/gcov/Kconfig" 1328 1329source "scripts/gcc-plugins/Kconfig" 1330 1331endmenu 1332