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 ARCH_HAS_SUBPAGE_FAULTS 15 bool 16 help 17 Select if the architecture can check permissions at sub-page 18 granularity (e.g. arm64 MTE). The probe_user_*() functions 19 must be implemented. 20 21config HOTPLUG_SMT 22 bool 23 24config SMT_NUM_THREADS_DYNAMIC 25 bool 26 27# Selected by HOTPLUG_CORE_SYNC_DEAD or HOTPLUG_CORE_SYNC_FULL 28config HOTPLUG_CORE_SYNC 29 bool 30 31# Basic CPU dead synchronization selected by architecture 32config HOTPLUG_CORE_SYNC_DEAD 33 bool 34 select HOTPLUG_CORE_SYNC 35 36# Full CPU synchronization with alive state selected by architecture 37config HOTPLUG_CORE_SYNC_FULL 38 bool 39 select HOTPLUG_CORE_SYNC_DEAD if HOTPLUG_CPU 40 select HOTPLUG_CORE_SYNC 41 42config HOTPLUG_SPLIT_STARTUP 43 bool 44 select HOTPLUG_CORE_SYNC_FULL 45 46config HOTPLUG_PARALLEL 47 bool 48 select HOTPLUG_SPLIT_STARTUP 49 50config GENERIC_ENTRY 51 bool 52 53config KPROBES 54 bool "Kprobes" 55 depends on MODULES 56 depends on HAVE_KPROBES 57 select KALLSYMS 58 select TASKS_RCU if PREEMPTION 59 help 60 Kprobes allows you to trap at almost any kernel address and 61 execute a callback function. register_kprobe() establishes 62 a probepoint and specifies the callback. Kprobes is useful 63 for kernel debugging, non-intrusive instrumentation and testing. 64 If in doubt, say "N". 65 66config JUMP_LABEL 67 bool "Optimize very unlikely/likely branches" 68 depends on HAVE_ARCH_JUMP_LABEL 69 select OBJTOOL if HAVE_JUMP_LABEL_HACK 70 help 71 This option enables a transparent branch optimization that 72 makes certain almost-always-true or almost-always-false branch 73 conditions even cheaper to execute within the kernel. 74 75 Certain performance-sensitive kernel code, such as trace points, 76 scheduler functionality, networking code and KVM have such 77 branches and include support for this optimization technique. 78 79 If it is detected that the compiler has support for "asm goto", 80 the kernel will compile such branches with just a nop 81 instruction. When the condition flag is toggled to true, the 82 nop will be converted to a jump instruction to execute the 83 conditional block of instructions. 84 85 This technique lowers overhead and stress on the branch prediction 86 of the processor and generally makes the kernel faster. The update 87 of the condition is slower, but those are always very rare. 88 89 ( On 32-bit x86, the necessary options added to the compiler 90 flags may increase the size of the kernel slightly. ) 91 92config STATIC_KEYS_SELFTEST 93 bool "Static key selftest" 94 depends on JUMP_LABEL 95 help 96 Boot time self-test of the branch patching code. 97 98config STATIC_CALL_SELFTEST 99 bool "Static call selftest" 100 depends on HAVE_STATIC_CALL 101 help 102 Boot time self-test of the call patching code. 103 104config OPTPROBES 105 def_bool y 106 depends on KPROBES && HAVE_OPTPROBES 107 select TASKS_RCU if PREEMPTION 108 109config KPROBES_ON_FTRACE 110 def_bool y 111 depends on KPROBES && HAVE_KPROBES_ON_FTRACE 112 depends on DYNAMIC_FTRACE_WITH_REGS 113 help 114 If function tracer is enabled and the arch supports full 115 passing of pt_regs to function tracing, then kprobes can 116 optimize on top of function tracing. 117 118config UPROBES 119 def_bool n 120 depends on ARCH_SUPPORTS_UPROBES 121 help 122 Uprobes is the user-space counterpart to kprobes: they 123 enable instrumentation applications (such as 'perf probe') 124 to establish unintrusive probes in user-space binaries and 125 libraries, by executing handler functions when the probes 126 are hit by user-space applications. 127 128 ( These probes come in the form of single-byte breakpoints, 129 managed by the kernel and kept transparent to the probed 130 application. ) 131 132config HAVE_64BIT_ALIGNED_ACCESS 133 def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS 134 help 135 Some architectures require 64 bit accesses to be 64 bit 136 aligned, which also requires structs containing 64 bit values 137 to be 64 bit aligned too. This includes some 32 bit 138 architectures which can do 64 bit accesses, as well as 64 bit 139 architectures without unaligned access. 140 141 This symbol should be selected by an architecture if 64 bit 142 accesses are required to be 64 bit aligned in this way even 143 though it is not a 64 bit architecture. 144 145 See Documentation/core-api/unaligned-memory-access.rst for 146 more information on the topic of unaligned memory accesses. 147 148config HAVE_EFFICIENT_UNALIGNED_ACCESS 149 bool 150 help 151 Some architectures are unable to perform unaligned accesses 152 without the use of get_unaligned/put_unaligned. Others are 153 unable to perform such accesses efficiently (e.g. trap on 154 unaligned access and require fixing it up in the exception 155 handler.) 156 157 This symbol should be selected by an architecture if it can 158 perform unaligned accesses efficiently to allow different 159 code paths to be selected for these cases. Some network 160 drivers, for example, could opt to not fix up alignment 161 problems with received packets if doing so would not help 162 much. 163 164 See Documentation/core-api/unaligned-memory-access.rst for more 165 information on the topic of unaligned memory accesses. 166 167config ARCH_USE_BUILTIN_BSWAP 168 bool 169 help 170 Modern versions of GCC (since 4.4) have builtin functions 171 for handling byte-swapping. Using these, instead of the old 172 inline assembler that the architecture code provides in the 173 __arch_bswapXX() macros, allows the compiler to see what's 174 happening and offers more opportunity for optimisation. In 175 particular, the compiler will be able to combine the byteswap 176 with a nearby load or store and use load-and-swap or 177 store-and-swap instructions if the architecture has them. It 178 should almost *never* result in code which is worse than the 179 hand-coded assembler in <asm/swab.h>. But just in case it 180 does, the use of the builtins is optional. 181 182 Any architecture with load-and-swap or store-and-swap 183 instructions should set this. And it shouldn't hurt to set it 184 on architectures that don't have such instructions. 185 186config KRETPROBES 187 def_bool y 188 depends on KPROBES && (HAVE_KRETPROBES || HAVE_RETHOOK) 189 190config KRETPROBE_ON_RETHOOK 191 def_bool y 192 depends on HAVE_RETHOOK 193 depends on KRETPROBES 194 select RETHOOK 195 196config USER_RETURN_NOTIFIER 197 bool 198 depends on HAVE_USER_RETURN_NOTIFIER 199 help 200 Provide a kernel-internal notification when a cpu is about to 201 switch to user mode. 202 203config HAVE_IOREMAP_PROT 204 bool 205 206config HAVE_KPROBES 207 bool 208 209config HAVE_KRETPROBES 210 bool 211 212config HAVE_OPTPROBES 213 bool 214 215config HAVE_KPROBES_ON_FTRACE 216 bool 217 218config ARCH_CORRECT_STACKTRACE_ON_KRETPROBE 219 bool 220 help 221 Since kretprobes modifies return address on the stack, the 222 stacktrace may see the kretprobe trampoline address instead 223 of correct one. If the architecture stacktrace code and 224 unwinder can adjust such entries, select this configuration. 225 226config HAVE_FUNCTION_ERROR_INJECTION 227 bool 228 229config HAVE_NMI 230 bool 231 232config HAVE_FUNCTION_DESCRIPTORS 233 bool 234 235config TRACE_IRQFLAGS_SUPPORT 236 bool 237 238config TRACE_IRQFLAGS_NMI_SUPPORT 239 bool 240 241# 242# An arch should select this if it provides all these things: 243# 244# task_pt_regs() in asm/processor.h or asm/ptrace.h 245# arch_has_single_step() if there is hardware single-step support 246# arch_has_block_step() if there is hardware block-step support 247# asm/syscall.h supplying asm-generic/syscall.h interface 248# linux/regset.h user_regset interfaces 249# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 250# TIF_SYSCALL_TRACE calls ptrace_report_syscall_{entry,exit} 251# TIF_NOTIFY_RESUME calls resume_user_mode_work() 252# 253config HAVE_ARCH_TRACEHOOK 254 bool 255 256config HAVE_DMA_CONTIGUOUS 257 bool 258 259config GENERIC_SMP_IDLE_THREAD 260 bool 261 262config GENERIC_IDLE_POLL_SETUP 263 bool 264 265config ARCH_HAS_FORTIFY_SOURCE 266 bool 267 help 268 An architecture should select this when it can successfully 269 build and run with CONFIG_FORTIFY_SOURCE. 270 271# 272# Select if the arch provides a historic keepinit alias for the retain_initrd 273# command line option 274# 275config ARCH_HAS_KEEPINITRD 276 bool 277 278# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h 279config ARCH_HAS_SET_MEMORY 280 bool 281 282# Select if arch has all set_direct_map_invalid/default() functions 283config ARCH_HAS_SET_DIRECT_MAP 284 bool 285 286# 287# Select if the architecture provides the arch_dma_set_uncached symbol to 288# either provide an uncached segment alias for a DMA allocation, or 289# to remap the page tables in place. 290# 291config ARCH_HAS_DMA_SET_UNCACHED 292 bool 293 294# 295# Select if the architectures provides the arch_dma_clear_uncached symbol 296# to undo an in-place page table remap for uncached access. 297# 298config ARCH_HAS_DMA_CLEAR_UNCACHED 299 bool 300 301config ARCH_HAS_CPU_FINALIZE_INIT 302 bool 303 304# The architecture has a per-task state that includes the mm's PASID 305config ARCH_HAS_CPU_PASID 306 bool 307 select IOMMU_MM_DATA 308 309# Select if arch init_task must go in the __init_task_data section 310config ARCH_TASK_STRUCT_ON_STACK 311 bool 312 313# Select if arch has its private alloc_task_struct() function 314config ARCH_TASK_STRUCT_ALLOCATOR 315 bool 316 317config HAVE_ARCH_THREAD_STRUCT_WHITELIST 318 bool 319 depends on !ARCH_TASK_STRUCT_ALLOCATOR 320 help 321 An architecture should select this to provide hardened usercopy 322 knowledge about what region of the thread_struct should be 323 whitelisted for copying to userspace. Normally this is only the 324 FPU registers. Specifically, arch_thread_struct_whitelist() 325 should be implemented. Without this, the entire thread_struct 326 field in task_struct will be left whitelisted. 327 328# Select if arch has its private alloc_thread_stack() function 329config ARCH_THREAD_STACK_ALLOCATOR 330 bool 331 332# Select if arch wants to size task_struct dynamically via arch_task_struct_size: 333config ARCH_WANTS_DYNAMIC_TASK_STRUCT 334 bool 335 336config ARCH_WANTS_NO_INSTR 337 bool 338 help 339 An architecture should select this if the noinstr macro is being used on 340 functions to denote that the toolchain should avoid instrumenting such 341 functions and is required for correctness. 342 343config ARCH_32BIT_OFF_T 344 bool 345 depends on !64BIT 346 help 347 All new 32-bit architectures should have 64-bit off_t type on 348 userspace side which corresponds to the loff_t kernel type. This 349 is the requirement for modern ABIs. Some existing architectures 350 still support 32-bit off_t. This option is enabled for all such 351 architectures explicitly. 352 353# Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat 354config ARCH_32BIT_USTAT_F_TINODE 355 bool 356 357config HAVE_ASM_MODVERSIONS 358 bool 359 help 360 This symbol should be selected by an architecture if it provides 361 <asm/asm-prototypes.h> to support the module versioning for symbols 362 exported from assembly code. 363 364config HAVE_REGS_AND_STACK_ACCESS_API 365 bool 366 help 367 This symbol should be selected by an architecture if it supports 368 the API needed to access registers and stack entries from pt_regs, 369 declared in asm/ptrace.h 370 For example the kprobes-based event tracer needs this API. 371 372config HAVE_RSEQ 373 bool 374 depends on HAVE_REGS_AND_STACK_ACCESS_API 375 help 376 This symbol should be selected by an architecture if it 377 supports an implementation of restartable sequences. 378 379config HAVE_RUST 380 bool 381 help 382 This symbol should be selected by an architecture if it 383 supports Rust. 384 385config HAVE_FUNCTION_ARG_ACCESS_API 386 bool 387 help 388 This symbol should be selected by an architecture if it supports 389 the API needed to access function arguments from pt_regs, 390 declared in asm/ptrace.h 391 392config HAVE_HW_BREAKPOINT 393 bool 394 depends on PERF_EVENTS 395 396config HAVE_MIXED_BREAKPOINTS_REGS 397 bool 398 depends on HAVE_HW_BREAKPOINT 399 help 400 Depending on the arch implementation of hardware breakpoints, 401 some of them have separate registers for data and instruction 402 breakpoints addresses, others have mixed registers to store 403 them but define the access type in a control register. 404 Select this option if your arch implements breakpoints under the 405 latter fashion. 406 407config HAVE_USER_RETURN_NOTIFIER 408 bool 409 410config HAVE_PERF_EVENTS_NMI 411 bool 412 help 413 System hardware can generate an NMI using the perf event 414 subsystem. Also has support for calculating CPU cycle events 415 to determine how many clock cycles in a given period. 416 417config HAVE_HARDLOCKUP_DETECTOR_PERF 418 bool 419 depends on HAVE_PERF_EVENTS_NMI 420 help 421 The arch chooses to use the generic perf-NMI-based hardlockup 422 detector. Must define HAVE_PERF_EVENTS_NMI. 423 424config HAVE_HARDLOCKUP_DETECTOR_ARCH 425 bool 426 help 427 The arch provides its own hardlockup detector implementation instead 428 of the generic ones. 429 430 It uses the same command line parameters, and sysctl interface, 431 as the generic hardlockup detectors. 432 433config HAVE_PERF_REGS 434 bool 435 help 436 Support selective register dumps for perf events. This includes 437 bit-mapping of each registers and a unique architecture id. 438 439config HAVE_PERF_USER_STACK_DUMP 440 bool 441 help 442 Support user stack dumps for perf event samples. This needs 443 access to the user stack pointer which is not unified across 444 architectures. 445 446config HAVE_ARCH_JUMP_LABEL 447 bool 448 449config HAVE_ARCH_JUMP_LABEL_RELATIVE 450 bool 451 452config MMU_GATHER_TABLE_FREE 453 bool 454 455config MMU_GATHER_RCU_TABLE_FREE 456 bool 457 select MMU_GATHER_TABLE_FREE 458 459config MMU_GATHER_PAGE_SIZE 460 bool 461 462config MMU_GATHER_NO_RANGE 463 bool 464 select MMU_GATHER_MERGE_VMAS 465 466config MMU_GATHER_NO_FLUSH_CACHE 467 bool 468 469config MMU_GATHER_MERGE_VMAS 470 bool 471 472config MMU_GATHER_NO_GATHER 473 bool 474 depends on MMU_GATHER_TABLE_FREE 475 476config ARCH_WANT_IRQS_OFF_ACTIVATE_MM 477 bool 478 help 479 Temporary select until all architectures can be converted to have 480 irqs disabled over activate_mm. Architectures that do IPI based TLB 481 shootdowns should enable this. 482 483# Use normal mm refcounting for MMU_LAZY_TLB kernel thread references. 484# MMU_LAZY_TLB_REFCOUNT=n can improve the scalability of context switching 485# to/from kernel threads when the same mm is running on a lot of CPUs (a large 486# multi-threaded application), by reducing contention on the mm refcount. 487# 488# This can be disabled if the architecture ensures no CPUs are using an mm as a 489# "lazy tlb" beyond its final refcount (i.e., by the time __mmdrop frees the mm 490# or its kernel page tables). This could be arranged by arch_exit_mmap(), or 491# final exit(2) TLB flush, for example. 492# 493# To implement this, an arch *must*: 494# Ensure the _lazy_tlb variants of mmgrab/mmdrop are used when manipulating 495# the lazy tlb reference of a kthread's ->active_mm (non-arch code has been 496# converted already). 497config MMU_LAZY_TLB_REFCOUNT 498 def_bool y 499 depends on !MMU_LAZY_TLB_SHOOTDOWN 500 501# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an 502# mm as a lazy tlb beyond its last reference count, by shooting down these 503# users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may 504# be using the mm as a lazy tlb, so that they may switch themselves to using 505# init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs 506# may be using mm as a lazy tlb mm. 507# 508# To implement this, an arch *must*: 509# - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains 510# at least all possible CPUs in which the mm is lazy. 511# - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above). 512config MMU_LAZY_TLB_SHOOTDOWN 513 bool 514 515config ARCH_HAVE_NMI_SAFE_CMPXCHG 516 bool 517 518config ARCH_HAS_NMI_SAFE_THIS_CPU_OPS 519 bool 520 521config HAVE_ALIGNED_STRUCT_PAGE 522 bool 523 help 524 This makes sure that struct pages are double word aligned and that 525 e.g. the SLUB allocator can perform double word atomic operations 526 on a struct page for better performance. However selecting this 527 might increase the size of a struct page by a word. 528 529config HAVE_CMPXCHG_LOCAL 530 bool 531 532config HAVE_CMPXCHG_DOUBLE 533 bool 534 535config ARCH_WEAK_RELEASE_ACQUIRE 536 bool 537 538config ARCH_WANT_IPC_PARSE_VERSION 539 bool 540 541config ARCH_WANT_COMPAT_IPC_PARSE_VERSION 542 bool 543 544config ARCH_WANT_OLD_COMPAT_IPC 545 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 546 bool 547 548config HAVE_ARCH_SECCOMP 549 bool 550 help 551 An arch should select this symbol to support seccomp mode 1 (the fixed 552 syscall policy), and must provide an overrides for __NR_seccomp_sigreturn, 553 and compat syscalls if the asm-generic/seccomp.h defaults need adjustment: 554 - __NR_seccomp_read_32 555 - __NR_seccomp_write_32 556 - __NR_seccomp_exit_32 557 - __NR_seccomp_sigreturn_32 558 559config HAVE_ARCH_SECCOMP_FILTER 560 bool 561 select HAVE_ARCH_SECCOMP 562 help 563 An arch should select this symbol if it provides all of these things: 564 - all the requirements for HAVE_ARCH_SECCOMP 565 - syscall_get_arch() 566 - syscall_get_arguments() 567 - syscall_rollback() 568 - syscall_set_return_value() 569 - SIGSYS siginfo_t support 570 - secure_computing is called from a ptrace_event()-safe context 571 - secure_computing return value is checked and a return value of -1 572 results in the system call being skipped immediately. 573 - seccomp syscall wired up 574 - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE, 575 SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If 576 COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too. 577 578config SECCOMP 579 prompt "Enable seccomp to safely execute untrusted bytecode" 580 def_bool y 581 depends on HAVE_ARCH_SECCOMP 582 help 583 This kernel feature is useful for number crunching applications 584 that may need to handle untrusted bytecode during their 585 execution. By using pipes or other transports made available 586 to the process as file descriptors supporting the read/write 587 syscalls, it's possible to isolate those applications in their 588 own address space using seccomp. Once seccomp is enabled via 589 prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be 590 disabled and the task is only allowed to execute a few safe 591 syscalls defined by each seccomp mode. 592 593 If unsure, say Y. 594 595config SECCOMP_FILTER 596 def_bool y 597 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 598 help 599 Enable tasks to build secure computing environments defined 600 in terms of Berkeley Packet Filter programs which implement 601 task-defined system call filtering polices. 602 603 See Documentation/userspace-api/seccomp_filter.rst for details. 604 605config SECCOMP_CACHE_DEBUG 606 bool "Show seccomp filter cache status in /proc/pid/seccomp_cache" 607 depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR 608 depends on PROC_FS 609 help 610 This enables the /proc/pid/seccomp_cache interface to monitor 611 seccomp cache data. The file format is subject to change. Reading 612 the file requires CAP_SYS_ADMIN. 613 614 This option is for debugging only. Enabling presents the risk that 615 an adversary may be able to infer the seccomp filter logic. 616 617 If unsure, say N. 618 619config HAVE_ARCH_STACKLEAK 620 bool 621 help 622 An architecture should select this if it has the code which 623 fills the used part of the kernel stack with the STACKLEAK_POISON 624 value before returning from system calls. 625 626config HAVE_STACKPROTECTOR 627 bool 628 help 629 An arch should select this symbol if: 630 - it has implemented a stack canary (e.g. __stack_chk_guard) 631 632config STACKPROTECTOR 633 bool "Stack Protector buffer overflow detection" 634 depends on HAVE_STACKPROTECTOR 635 depends on $(cc-option,-fstack-protector) 636 default y 637 help 638 This option turns on the "stack-protector" GCC feature. This 639 feature puts, at the beginning of functions, a canary value on 640 the stack just before the return address, and validates 641 the value just before actually returning. Stack based buffer 642 overflows (that need to overwrite this return address) now also 643 overwrite the canary, which gets detected and the attack is then 644 neutralized via a kernel panic. 645 646 Functions will have the stack-protector canary logic added if they 647 have an 8-byte or larger character array on the stack. 648 649 This feature requires gcc version 4.2 or above, or a distribution 650 gcc with the feature backported ("-fstack-protector"). 651 652 On an x86 "defconfig" build, this feature adds canary checks to 653 about 3% of all kernel functions, which increases kernel code size 654 by about 0.3%. 655 656config STACKPROTECTOR_STRONG 657 bool "Strong Stack Protector" 658 depends on STACKPROTECTOR 659 depends on $(cc-option,-fstack-protector-strong) 660 default y 661 help 662 Functions will have the stack-protector canary logic added in any 663 of the following conditions: 664 665 - local variable's address used as part of the right hand side of an 666 assignment or function argument 667 - local variable is an array (or union containing an array), 668 regardless of array type or length 669 - uses register local variables 670 671 This feature requires gcc version 4.9 or above, or a distribution 672 gcc with the feature backported ("-fstack-protector-strong"). 673 674 On an x86 "defconfig" build, this feature adds canary checks to 675 about 20% of all kernel functions, which increases the kernel code 676 size by about 2%. 677 678config ARCH_SUPPORTS_SHADOW_CALL_STACK 679 bool 680 help 681 An architecture should select this if it supports the compiler's 682 Shadow Call Stack and implements runtime support for shadow stack 683 switching. 684 685config SHADOW_CALL_STACK 686 bool "Shadow Call Stack" 687 depends on ARCH_SUPPORTS_SHADOW_CALL_STACK 688 depends on DYNAMIC_FTRACE_WITH_ARGS || DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER 689 help 690 This option enables the compiler's Shadow Call Stack, which 691 uses a shadow stack to protect function return addresses from 692 being overwritten by an attacker. More information can be found 693 in the compiler's documentation: 694 695 - Clang: https://clang.llvm.org/docs/ShadowCallStack.html 696 - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options 697 698 Note that security guarantees in the kernel differ from the 699 ones documented for user space. The kernel must store addresses 700 of shadow stacks in memory, which means an attacker capable of 701 reading and writing arbitrary memory may be able to locate them 702 and hijack control flow by modifying the stacks. 703 704config DYNAMIC_SCS 705 bool 706 help 707 Set by the arch code if it relies on code patching to insert the 708 shadow call stack push and pop instructions rather than on the 709 compiler. 710 711config LTO 712 bool 713 help 714 Selected if the kernel will be built using the compiler's LTO feature. 715 716config LTO_CLANG 717 bool 718 select LTO 719 help 720 Selected if the kernel will be built using Clang's LTO feature. 721 722config ARCH_SUPPORTS_LTO_CLANG 723 bool 724 help 725 An architecture should select this option if it supports: 726 - compiling with Clang, 727 - compiling inline assembly with Clang's integrated assembler, 728 - and linking with LLD. 729 730config ARCH_SUPPORTS_LTO_CLANG_THIN 731 bool 732 help 733 An architecture should select this option if it can support Clang's 734 ThinLTO mode. 735 736config HAS_LTO_CLANG 737 def_bool y 738 depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM 739 depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm) 740 depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm) 741 depends on ARCH_SUPPORTS_LTO_CLANG 742 depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT 743 # https://github.com/ClangBuiltLinux/linux/issues/1721 744 depends on (!KASAN || KASAN_HW_TAGS || CLANG_VERSION >= 170000) || !DEBUG_INFO 745 depends on (!KCOV || CLANG_VERSION >= 170000) || !DEBUG_INFO 746 depends on !GCOV_KERNEL 747 help 748 The compiler and Kconfig options support building with Clang's 749 LTO. 750 751choice 752 prompt "Link Time Optimization (LTO)" 753 default LTO_NONE 754 help 755 This option enables Link Time Optimization (LTO), which allows the 756 compiler to optimize binaries globally. 757 758 If unsure, select LTO_NONE. Note that LTO is very resource-intensive 759 so it's disabled by default. 760 761config LTO_NONE 762 bool "None" 763 help 764 Build the kernel normally, without Link Time Optimization (LTO). 765 766config LTO_CLANG_FULL 767 bool "Clang Full LTO (EXPERIMENTAL)" 768 depends on HAS_LTO_CLANG 769 depends on !COMPILE_TEST 770 select LTO_CLANG 771 help 772 This option enables Clang's full Link Time Optimization (LTO), which 773 allows the compiler to optimize the kernel globally. If you enable 774 this option, the compiler generates LLVM bitcode instead of ELF 775 object files, and the actual compilation from bitcode happens at 776 the LTO link step, which may take several minutes depending on the 777 kernel configuration. More information can be found from LLVM's 778 documentation: 779 780 https://llvm.org/docs/LinkTimeOptimization.html 781 782 During link time, this option can use a large amount of RAM, and 783 may take much longer than the ThinLTO option. 784 785config LTO_CLANG_THIN 786 bool "Clang ThinLTO (EXPERIMENTAL)" 787 depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN 788 select LTO_CLANG 789 help 790 This option enables Clang's ThinLTO, which allows for parallel 791 optimization and faster incremental compiles compared to the 792 CONFIG_LTO_CLANG_FULL option. More information can be found 793 from Clang's documentation: 794 795 https://clang.llvm.org/docs/ThinLTO.html 796 797 If unsure, say Y. 798endchoice 799 800config ARCH_SUPPORTS_CFI_CLANG 801 bool 802 help 803 An architecture should select this option if it can support Clang's 804 Control-Flow Integrity (CFI) checking. 805 806config ARCH_USES_CFI_TRAPS 807 bool 808 809config CFI_CLANG 810 bool "Use Clang's Control Flow Integrity (CFI)" 811 depends on ARCH_SUPPORTS_CFI_CLANG 812 depends on $(cc-option,-fsanitize=kcfi) 813 help 814 This option enables Clang’s forward-edge Control Flow Integrity 815 (CFI) checking, where the compiler injects a runtime check to each 816 indirect function call to ensure the target is a valid function with 817 the correct static type. This restricts possible call targets and 818 makes it more difficult for an attacker to exploit bugs that allow 819 the modification of stored function pointers. More information can be 820 found from Clang's documentation: 821 822 https://clang.llvm.org/docs/ControlFlowIntegrity.html 823 824config CFI_PERMISSIVE 825 bool "Use CFI in permissive mode" 826 depends on CFI_CLANG 827 help 828 When selected, Control Flow Integrity (CFI) violations result in a 829 warning instead of a kernel panic. This option should only be used 830 for finding indirect call type mismatches during development. 831 832 If unsure, say N. 833 834config HAVE_ARCH_WITHIN_STACK_FRAMES 835 bool 836 help 837 An architecture should select this if it can walk the kernel stack 838 frames to determine if an object is part of either the arguments 839 or local variables (i.e. that it excludes saved return addresses, 840 and similar) by implementing an inline arch_within_stack_frames(), 841 which is used by CONFIG_HARDENED_USERCOPY. 842 843config HAVE_CONTEXT_TRACKING_USER 844 bool 845 help 846 Provide kernel/user boundaries probes necessary for subsystems 847 that need it, such as userspace RCU extended quiescent state. 848 Syscalls need to be wrapped inside user_exit()-user_enter(), either 849 optimized behind static key or through the slow path using TIF_NOHZ 850 flag. Exceptions handlers must be wrapped as well. Irqs are already 851 protected inside ct_irq_enter/ct_irq_exit() but preemption or signal 852 handling on irq exit still need to be protected. 853 854config HAVE_CONTEXT_TRACKING_USER_OFFSTACK 855 bool 856 help 857 Architecture neither relies on exception_enter()/exception_exit() 858 nor on schedule_user(). Also preempt_schedule_notrace() and 859 preempt_schedule_irq() can't be called in a preemptible section 860 while context tracking is CONTEXT_USER. This feature reflects a sane 861 entry implementation where the following requirements are met on 862 critical entry code, ie: before user_exit() or after user_enter(): 863 864 - Critical entry code isn't preemptible (or better yet: 865 not interruptible). 866 - No use of RCU read side critical sections, unless ct_nmi_enter() 867 got called. 868 - No use of instrumentation, unless instrumentation_begin() got 869 called. 870 871config HAVE_TIF_NOHZ 872 bool 873 help 874 Arch relies on TIF_NOHZ and syscall slow path to implement context 875 tracking calls to user_enter()/user_exit(). 876 877config HAVE_VIRT_CPU_ACCOUNTING 878 bool 879 880config HAVE_VIRT_CPU_ACCOUNTING_IDLE 881 bool 882 help 883 Architecture has its own way to account idle CPU time and therefore 884 doesn't implement vtime_account_idle(). 885 886config ARCH_HAS_SCALED_CPUTIME 887 bool 888 889config HAVE_VIRT_CPU_ACCOUNTING_GEN 890 bool 891 default y if 64BIT 892 help 893 With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. 894 Before enabling this option, arch code must be audited 895 to ensure there are no races in concurrent read/write of 896 cputime_t. For example, reading/writing 64-bit cputime_t on 897 some 32-bit arches may require multiple accesses, so proper 898 locking is needed to protect against concurrent accesses. 899 900config HAVE_IRQ_TIME_ACCOUNTING 901 bool 902 help 903 Archs need to ensure they use a high enough resolution clock to 904 support irq time accounting and then call enable_sched_clock_irqtime(). 905 906config HAVE_MOVE_PUD 907 bool 908 help 909 Architectures that select this are able to move page tables at the 910 PUD level. If there are only 3 page table levels, the move effectively 911 happens at the PGD level. 912 913config HAVE_MOVE_PMD 914 bool 915 help 916 Archs that select this are able to move page tables at the PMD level. 917 918config HAVE_ARCH_TRANSPARENT_HUGEPAGE 919 bool 920 921config HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD 922 bool 923 924config HAVE_ARCH_HUGE_VMAP 925 bool 926 927# 928# Archs that select this would be capable of PMD-sized vmaps (i.e., 929# arch_vmap_pmd_supported() returns true). The VM_ALLOW_HUGE_VMAP flag 930# must be used to enable allocations to use hugepages. 931# 932config HAVE_ARCH_HUGE_VMALLOC 933 depends on HAVE_ARCH_HUGE_VMAP 934 bool 935 936config ARCH_WANT_HUGE_PMD_SHARE 937 bool 938 939# Archs that want to use pmd_mkwrite on kernel memory need it defined even 940# if there are no userspace memory management features that use it 941config ARCH_WANT_KERNEL_PMD_MKWRITE 942 bool 943 944config ARCH_WANT_PMD_MKWRITE 945 def_bool TRANSPARENT_HUGEPAGE || ARCH_WANT_KERNEL_PMD_MKWRITE 946 947config HAVE_ARCH_SOFT_DIRTY 948 bool 949 950config HAVE_MOD_ARCH_SPECIFIC 951 bool 952 help 953 The arch uses struct mod_arch_specific to store data. Many arches 954 just need a simple module loader without arch specific data - those 955 should not enable this. 956 957config MODULES_USE_ELF_RELA 958 bool 959 help 960 Modules only use ELF RELA relocations. Modules with ELF REL 961 relocations will give an error. 962 963config MODULES_USE_ELF_REL 964 bool 965 help 966 Modules only use ELF REL relocations. Modules with ELF RELA 967 relocations will give an error. 968 969config ARCH_WANTS_MODULES_DATA_IN_VMALLOC 970 bool 971 help 972 For architectures like powerpc/32 which have constraints on module 973 allocation and need to allocate module data outside of module area. 974 975config HAVE_IRQ_EXIT_ON_IRQ_STACK 976 bool 977 help 978 Architecture doesn't only execute the irq handler on the irq stack 979 but also irq_exit(). This way we can process softirqs on this irq 980 stack instead of switching to a new one when we call __do_softirq() 981 in the end of an hardirq. 982 This spares a stack switch and improves cache usage on softirq 983 processing. 984 985config HAVE_SOFTIRQ_ON_OWN_STACK 986 bool 987 help 988 Architecture provides a function to run __do_softirq() on a 989 separate stack. 990 991config SOFTIRQ_ON_OWN_STACK 992 def_bool HAVE_SOFTIRQ_ON_OWN_STACK && !PREEMPT_RT 993 994config ALTERNATE_USER_ADDRESS_SPACE 995 bool 996 help 997 Architectures set this when the CPU uses separate address 998 spaces for kernel and user space pointers. In this case, the 999 access_ok() check on a __user pointer is skipped. 1000 1001config PGTABLE_LEVELS 1002 int 1003 default 2 1004 1005config ARCH_HAS_ELF_RANDOMIZE 1006 bool 1007 help 1008 An architecture supports choosing randomized locations for 1009 stack, mmap, brk, and ET_DYN. Defined functions: 1010 - arch_mmap_rnd() 1011 - arch_randomize_brk() 1012 1013config HAVE_ARCH_MMAP_RND_BITS 1014 bool 1015 help 1016 An arch should select this symbol if it supports setting a variable 1017 number of bits for use in establishing the base address for mmap 1018 allocations, has MMU enabled and provides values for both: 1019 - ARCH_MMAP_RND_BITS_MIN 1020 - ARCH_MMAP_RND_BITS_MAX 1021 1022config HAVE_EXIT_THREAD 1023 bool 1024 help 1025 An architecture implements exit_thread. 1026 1027config ARCH_MMAP_RND_BITS_MIN 1028 int 1029 1030config ARCH_MMAP_RND_BITS_MAX 1031 int 1032 1033config ARCH_MMAP_RND_BITS_DEFAULT 1034 int 1035 1036config ARCH_MMAP_RND_BITS 1037 int "Number of bits to use for ASLR of mmap base address" if EXPERT 1038 range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX 1039 default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT 1040 default ARCH_MMAP_RND_BITS_MIN 1041 depends on HAVE_ARCH_MMAP_RND_BITS 1042 help 1043 This value can be used to select the number of bits to use to 1044 determine the random offset to the base address of vma regions 1045 resulting from mmap allocations. This value will be bounded 1046 by the architecture's minimum and maximum supported values. 1047 1048 This value can be changed after boot using the 1049 /proc/sys/vm/mmap_rnd_bits tunable 1050 1051config HAVE_ARCH_MMAP_RND_COMPAT_BITS 1052 bool 1053 help 1054 An arch should select this symbol if it supports running applications 1055 in compatibility mode, supports setting a variable number of bits for 1056 use in establishing the base address for mmap allocations, has MMU 1057 enabled and provides values for both: 1058 - ARCH_MMAP_RND_COMPAT_BITS_MIN 1059 - ARCH_MMAP_RND_COMPAT_BITS_MAX 1060 1061config ARCH_MMAP_RND_COMPAT_BITS_MIN 1062 int 1063 1064config ARCH_MMAP_RND_COMPAT_BITS_MAX 1065 int 1066 1067config ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1068 int 1069 1070config ARCH_MMAP_RND_COMPAT_BITS 1071 int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT 1072 range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX 1073 default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1074 default ARCH_MMAP_RND_COMPAT_BITS_MIN 1075 depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS 1076 help 1077 This value can be used to select the number of bits to use to 1078 determine the random offset to the base address of vma regions 1079 resulting from mmap allocations for compatible applications This 1080 value will be bounded by the architecture's minimum and maximum 1081 supported values. 1082 1083 This value can be changed after boot using the 1084 /proc/sys/vm/mmap_rnd_compat_bits tunable 1085 1086config HAVE_ARCH_COMPAT_MMAP_BASES 1087 bool 1088 help 1089 This allows 64bit applications to invoke 32-bit mmap() syscall 1090 and vice-versa 32-bit applications to call 64-bit mmap(). 1091 Required for applications doing different bitness syscalls. 1092 1093config PAGE_SIZE_LESS_THAN_64KB 1094 def_bool y 1095 depends on !ARM64_64K_PAGES 1096 depends on !PAGE_SIZE_64KB 1097 depends on !PARISC_PAGE_SIZE_64KB 1098 depends on PAGE_SIZE_LESS_THAN_256KB 1099 1100config PAGE_SIZE_LESS_THAN_256KB 1101 def_bool y 1102 depends on !PAGE_SIZE_256KB 1103 1104# This allows to use a set of generic functions to determine mmap base 1105# address by giving priority to top-down scheme only if the process 1106# is not in legacy mode (compat task, unlimited stack size or 1107# sysctl_legacy_va_layout). 1108# Architecture that selects this option can provide its own version of: 1109# - STACK_RND_MASK 1110config ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT 1111 bool 1112 depends on MMU 1113 select ARCH_HAS_ELF_RANDOMIZE 1114 1115config HAVE_OBJTOOL 1116 bool 1117 1118config HAVE_JUMP_LABEL_HACK 1119 bool 1120 1121config HAVE_NOINSTR_HACK 1122 bool 1123 1124config HAVE_NOINSTR_VALIDATION 1125 bool 1126 1127config HAVE_UACCESS_VALIDATION 1128 bool 1129 select OBJTOOL 1130 1131config HAVE_STACK_VALIDATION 1132 bool 1133 help 1134 Architecture supports objtool compile-time frame pointer rule 1135 validation. 1136 1137config HAVE_RELIABLE_STACKTRACE 1138 bool 1139 help 1140 Architecture has either save_stack_trace_tsk_reliable() or 1141 arch_stack_walk_reliable() function which only returns a stack trace 1142 if it can guarantee the trace is reliable. 1143 1144config HAVE_ARCH_HASH 1145 bool 1146 default n 1147 help 1148 If this is set, the architecture provides an <asm/hash.h> 1149 file which provides platform-specific implementations of some 1150 functions in <linux/hash.h> or fs/namei.c. 1151 1152config HAVE_ARCH_NVRAM_OPS 1153 bool 1154 1155config ISA_BUS_API 1156 def_bool ISA 1157 1158# 1159# ABI hall of shame 1160# 1161config CLONE_BACKWARDS 1162 bool 1163 help 1164 Architecture has tls passed as the 4th argument of clone(2), 1165 not the 5th one. 1166 1167config CLONE_BACKWARDS2 1168 bool 1169 help 1170 Architecture has the first two arguments of clone(2) swapped. 1171 1172config CLONE_BACKWARDS3 1173 bool 1174 help 1175 Architecture has tls passed as the 3rd argument of clone(2), 1176 not the 5th one. 1177 1178config ODD_RT_SIGACTION 1179 bool 1180 help 1181 Architecture has unusual rt_sigaction(2) arguments 1182 1183config OLD_SIGSUSPEND 1184 bool 1185 help 1186 Architecture has old sigsuspend(2) syscall, of one-argument variety 1187 1188config OLD_SIGSUSPEND3 1189 bool 1190 help 1191 Even weirder antique ABI - three-argument sigsuspend(2) 1192 1193config OLD_SIGACTION 1194 bool 1195 help 1196 Architecture has old sigaction(2) syscall. Nope, not the same 1197 as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2), 1198 but fairly different variant of sigaction(2), thanks to OSF/1 1199 compatibility... 1200 1201config COMPAT_OLD_SIGACTION 1202 bool 1203 1204config COMPAT_32BIT_TIME 1205 bool "Provide system calls for 32-bit time_t" 1206 default !64BIT || COMPAT 1207 help 1208 This enables 32 bit time_t support in addition to 64 bit time_t support. 1209 This is relevant on all 32-bit architectures, and 64-bit architectures 1210 as part of compat syscall handling. 1211 1212config ARCH_NO_PREEMPT 1213 bool 1214 1215config ARCH_SUPPORTS_RT 1216 bool 1217 1218config CPU_NO_EFFICIENT_FFS 1219 def_bool n 1220 1221config HAVE_ARCH_VMAP_STACK 1222 def_bool n 1223 help 1224 An arch should select this symbol if it can support kernel stacks 1225 in vmalloc space. This means: 1226 1227 - vmalloc space must be large enough to hold many kernel stacks. 1228 This may rule out many 32-bit architectures. 1229 1230 - Stacks in vmalloc space need to work reliably. For example, if 1231 vmap page tables are created on demand, either this mechanism 1232 needs to work while the stack points to a virtual address with 1233 unpopulated page tables or arch code (switch_to() and switch_mm(), 1234 most likely) needs to ensure that the stack's page table entries 1235 are populated before running on a possibly unpopulated stack. 1236 1237 - If the stack overflows into a guard page, something reasonable 1238 should happen. The definition of "reasonable" is flexible, but 1239 instantly rebooting without logging anything would be unfriendly. 1240 1241config VMAP_STACK 1242 default y 1243 bool "Use a virtually-mapped stack" 1244 depends on HAVE_ARCH_VMAP_STACK 1245 depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC 1246 help 1247 Enable this if you want the use virtually-mapped kernel stacks 1248 with guard pages. This causes kernel stack overflows to be 1249 caught immediately rather than causing difficult-to-diagnose 1250 corruption. 1251 1252 To use this with software KASAN modes, the architecture must support 1253 backing virtual mappings with real shadow memory, and KASAN_VMALLOC 1254 must be enabled. 1255 1256config HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1257 def_bool n 1258 help 1259 An arch should select this symbol if it can support kernel stack 1260 offset randomization with calls to add_random_kstack_offset() 1261 during syscall entry and choose_random_kstack_offset() during 1262 syscall exit. Careful removal of -fstack-protector-strong and 1263 -fstack-protector should also be applied to the entry code and 1264 closely examined, as the artificial stack bump looks like an array 1265 to the compiler, so it will attempt to add canary checks regardless 1266 of the static branch state. 1267 1268config RANDOMIZE_KSTACK_OFFSET 1269 bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT 1270 default y 1271 depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1272 depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000 1273 help 1274 The kernel stack offset can be randomized (after pt_regs) by 1275 roughly 5 bits of entropy, frustrating memory corruption 1276 attacks that depend on stack address determinism or 1277 cross-syscall address exposures. 1278 1279 The feature is controlled via the "randomize_kstack_offset=on/off" 1280 kernel boot param, and if turned off has zero overhead due to its use 1281 of static branches (see JUMP_LABEL). 1282 1283 If unsure, say Y. 1284 1285config RANDOMIZE_KSTACK_OFFSET_DEFAULT 1286 bool "Default state of kernel stack offset randomization" 1287 depends on RANDOMIZE_KSTACK_OFFSET 1288 help 1289 Kernel stack offset randomization is controlled by kernel boot param 1290 "randomize_kstack_offset=on/off", and this config chooses the default 1291 boot state. 1292 1293config ARCH_OPTIONAL_KERNEL_RWX 1294 def_bool n 1295 1296config ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1297 def_bool n 1298 1299config ARCH_HAS_STRICT_KERNEL_RWX 1300 def_bool n 1301 1302config STRICT_KERNEL_RWX 1303 bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX 1304 depends on ARCH_HAS_STRICT_KERNEL_RWX 1305 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1306 help 1307 If this is set, kernel text and rodata memory will be made read-only, 1308 and non-text memory will be made non-executable. This provides 1309 protection against certain security exploits (e.g. executing the heap 1310 or modifying text) 1311 1312 These features are considered standard security practice these days. 1313 You should say Y here in almost all cases. 1314 1315config ARCH_HAS_STRICT_MODULE_RWX 1316 def_bool n 1317 1318config STRICT_MODULE_RWX 1319 bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX 1320 depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES 1321 default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1322 help 1323 If this is set, module text and rodata memory will be made read-only, 1324 and non-text memory will be made non-executable. This provides 1325 protection against certain security exploits (e.g. writing to text) 1326 1327# select if the architecture provides an asm/dma-direct.h header 1328config ARCH_HAS_PHYS_TO_DMA 1329 bool 1330 1331config HAVE_ARCH_COMPILER_H 1332 bool 1333 help 1334 An architecture can select this if it provides an 1335 asm/compiler.h header that should be included after 1336 linux/compiler-*.h in order to override macro definitions that those 1337 headers generally provide. 1338 1339config HAVE_ARCH_PREL32_RELOCATIONS 1340 bool 1341 help 1342 May be selected by an architecture if it supports place-relative 1343 32-bit relocations, both in the toolchain and in the module loader, 1344 in which case relative references can be used in special sections 1345 for PCI fixup, initcalls etc which are only half the size on 64 bit 1346 architectures, and don't require runtime relocation on relocatable 1347 kernels. 1348 1349config ARCH_USE_MEMREMAP_PROT 1350 bool 1351 1352config LOCK_EVENT_COUNTS 1353 bool "Locking event counts collection" 1354 depends on DEBUG_FS 1355 help 1356 Enable light-weight counting of various locking related events 1357 in the system with minimal performance impact. This reduces 1358 the chance of application behavior change because of timing 1359 differences. The counts are reported via debugfs. 1360 1361# Select if the architecture has support for applying RELR relocations. 1362config ARCH_HAS_RELR 1363 bool 1364 1365config RELR 1366 bool "Use RELR relocation packing" 1367 depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR 1368 default y 1369 help 1370 Store the kernel's dynamic relocations in the RELR relocation packing 1371 format. Requires a compatible linker (LLD supports this feature), as 1372 well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy 1373 are compatible). 1374 1375config ARCH_HAS_MEM_ENCRYPT 1376 bool 1377 1378config ARCH_HAS_CC_PLATFORM 1379 bool 1380 1381config HAVE_SPARSE_SYSCALL_NR 1382 bool 1383 help 1384 An architecture should select this if its syscall numbering is sparse 1385 to save space. For example, MIPS architecture has a syscall array with 1386 entries at 4000, 5000 and 6000 locations. This option turns on syscall 1387 related optimizations for a given architecture. 1388 1389config ARCH_HAS_VDSO_DATA 1390 bool 1391 1392config HAVE_STATIC_CALL 1393 bool 1394 1395config HAVE_STATIC_CALL_INLINE 1396 bool 1397 depends on HAVE_STATIC_CALL 1398 select OBJTOOL 1399 1400config HAVE_PREEMPT_DYNAMIC 1401 bool 1402 1403config HAVE_PREEMPT_DYNAMIC_CALL 1404 bool 1405 depends on HAVE_STATIC_CALL 1406 select HAVE_PREEMPT_DYNAMIC 1407 help 1408 An architecture should select this if it can handle the preemption 1409 model being selected at boot time using static calls. 1410 1411 Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a 1412 preemption function will be patched directly. 1413 1414 Where an architecture does not select HAVE_STATIC_CALL_INLINE, any 1415 call to a preemption function will go through a trampoline, and the 1416 trampoline will be patched. 1417 1418 It is strongly advised to support inline static call to avoid any 1419 overhead. 1420 1421config HAVE_PREEMPT_DYNAMIC_KEY 1422 bool 1423 depends on HAVE_ARCH_JUMP_LABEL 1424 select HAVE_PREEMPT_DYNAMIC 1425 help 1426 An architecture should select this if it can handle the preemption 1427 model being selected at boot time using static keys. 1428 1429 Each preemption function will be given an early return based on a 1430 static key. This should have slightly lower overhead than non-inline 1431 static calls, as this effectively inlines each trampoline into the 1432 start of its callee. This may avoid redundant work, and may 1433 integrate better with CFI schemes. 1434 1435 This will have greater overhead than using inline static calls as 1436 the call to the preemption function cannot be entirely elided. 1437 1438config ARCH_WANT_LD_ORPHAN_WARN 1439 bool 1440 help 1441 An arch should select this symbol once all linker sections are explicitly 1442 included, size-asserted, or discarded in the linker scripts. This is 1443 important because we never want expected sections to be placed heuristically 1444 by the linker, since the locations of such sections can change between linker 1445 versions. 1446 1447config HAVE_ARCH_PFN_VALID 1448 bool 1449 1450config ARCH_SUPPORTS_DEBUG_PAGEALLOC 1451 bool 1452 1453config ARCH_SUPPORTS_PAGE_TABLE_CHECK 1454 bool 1455 1456config ARCH_SPLIT_ARG64 1457 bool 1458 help 1459 If a 32-bit architecture requires 64-bit arguments to be split into 1460 pairs of 32-bit arguments, select this option. 1461 1462config ARCH_HAS_ELFCORE_COMPAT 1463 bool 1464 1465config ARCH_HAS_PARANOID_L1D_FLUSH 1466 bool 1467 1468config ARCH_HAVE_TRACE_MMIO_ACCESS 1469 bool 1470 1471config DYNAMIC_SIGFRAME 1472 bool 1473 1474# Select, if arch has a named attribute group bound to NUMA device nodes. 1475config HAVE_ARCH_NODE_DEV_GROUP 1476 bool 1477 1478config ARCH_HAS_NONLEAF_PMD_YOUNG 1479 bool 1480 help 1481 Architectures that select this option are capable of setting the 1482 accessed bit in non-leaf PMD entries when using them as part of linear 1483 address translations. Page table walkers that clear the accessed bit 1484 may use this capability to reduce their search space. 1485 1486source "kernel/gcov/Kconfig" 1487 1488source "scripts/gcc-plugins/Kconfig" 1489 1490config FUNCTION_ALIGNMENT_4B 1491 bool 1492 1493config FUNCTION_ALIGNMENT_8B 1494 bool 1495 1496config FUNCTION_ALIGNMENT_16B 1497 bool 1498 1499config FUNCTION_ALIGNMENT_32B 1500 bool 1501 1502config FUNCTION_ALIGNMENT_64B 1503 bool 1504 1505config FUNCTION_ALIGNMENT 1506 int 1507 default 64 if FUNCTION_ALIGNMENT_64B 1508 default 32 if FUNCTION_ALIGNMENT_32B 1509 default 16 if FUNCTION_ALIGNMENT_16B 1510 default 8 if FUNCTION_ALIGNMENT_8B 1511 default 4 if FUNCTION_ALIGNMENT_4B 1512 default 0 1513 1514endmenu 1515