1b2441318SGreg Kroah-Hartman# SPDX-License-Identifier: GPL-2.0 2fb32e03fSMathieu Desnoyers# 3fb32e03fSMathieu Desnoyers# General architecture dependent options 4fb32e03fSMathieu Desnoyers# 5125e5645SMathieu Desnoyers 61572497cSChristoph Hellwig# 71572497cSChristoph Hellwig# Note: arch/$(SRCARCH)/Kconfig needs to be included first so that it can 81572497cSChristoph Hellwig# override the default values in this file. 91572497cSChristoph Hellwig# 101572497cSChristoph Hellwigsource "arch/$(SRCARCH)/Kconfig" 111572497cSChristoph Hellwig 1222471e13SRandy Dunlapmenu "General architecture-dependent options" 1322471e13SRandy Dunlap 14692f66f2SHari Bathiniconfig CRASH_CORE 15692f66f2SHari Bathini bool 16692f66f2SHari Bathini 172965faa5SDave Youngconfig KEXEC_CORE 18692f66f2SHari Bathini select CRASH_CORE 192965faa5SDave Young bool 202965faa5SDave Young 21175fca3bSSven Schnelleconfig KEXEC_ELF 22175fca3bSSven Schnelle bool 23175fca3bSSven Schnelle 24467d2782SThiago Jung Bauermannconfig HAVE_IMA_KEXEC 25467d2782SThiago Jung Bauermann bool 26467d2782SThiago Jung Bauermann 27da32b581SCatalin Marinasconfig ARCH_HAS_SUBPAGE_FAULTS 28da32b581SCatalin Marinas bool 29da32b581SCatalin Marinas help 30da32b581SCatalin Marinas Select if the architecture can check permissions at sub-page 31da32b581SCatalin Marinas granularity (e.g. arm64 MTE). The probe_user_*() functions 32da32b581SCatalin Marinas must be implemented. 33da32b581SCatalin Marinas 3405736e4aSThomas Gleixnerconfig HOTPLUG_SMT 3505736e4aSThomas Gleixner bool 3605736e4aSThomas Gleixner 376f062123SThomas Gleixner# Selected by HOTPLUG_CORE_SYNC_DEAD or HOTPLUG_CORE_SYNC_FULL 386f062123SThomas Gleixnerconfig HOTPLUG_CORE_SYNC 396f062123SThomas Gleixner bool 406f062123SThomas Gleixner 416f062123SThomas Gleixner# Basic CPU dead synchronization selected by architecture 426f062123SThomas Gleixnerconfig HOTPLUG_CORE_SYNC_DEAD 436f062123SThomas Gleixner bool 446f062123SThomas Gleixner select HOTPLUG_CORE_SYNC 456f062123SThomas Gleixner 466f062123SThomas Gleixner# Full CPU synchronization with alive state selected by architecture 476f062123SThomas Gleixnerconfig HOTPLUG_CORE_SYNC_FULL 486f062123SThomas Gleixner bool 496f062123SThomas Gleixner select HOTPLUG_CORE_SYNC_DEAD if HOTPLUG_CPU 506f062123SThomas Gleixner select HOTPLUG_CORE_SYNC 516f062123SThomas Gleixner 52*a631be92SThomas Gleixnerconfig HOTPLUG_SPLIT_STARTUP 53*a631be92SThomas Gleixner bool 54*a631be92SThomas Gleixner select HOTPLUG_CORE_SYNC_FULL 55*a631be92SThomas Gleixner 56142781e1SThomas Gleixnerconfig GENERIC_ENTRY 57142781e1SThomas Gleixner bool 58142781e1SThomas Gleixner 59125e5645SMathieu Desnoyersconfig KPROBES 60125e5645SMathieu Desnoyers bool "Kprobes" 6105ed160eSMasami Hiramatsu depends on MODULES 62125e5645SMathieu Desnoyers depends on HAVE_KPROBES 6305ed160eSMasami Hiramatsu select KALLSYMS 64835f14edSPaul E. McKenney select TASKS_RCU if PREEMPTION 65125e5645SMathieu Desnoyers help 66125e5645SMathieu Desnoyers Kprobes allows you to trap at almost any kernel address and 67125e5645SMathieu Desnoyers execute a callback function. register_kprobe() establishes 68125e5645SMathieu Desnoyers a probepoint and specifies the callback. Kprobes is useful 69125e5645SMathieu Desnoyers for kernel debugging, non-intrusive instrumentation and testing. 70125e5645SMathieu Desnoyers If in doubt, say "N". 71125e5645SMathieu Desnoyers 7245f81b1cSSteven Rostedtconfig JUMP_LABEL 73c5905afbSIngo Molnar bool "Optimize very unlikely/likely branches" 7445f81b1cSSteven Rostedt depends on HAVE_ARCH_JUMP_LABEL 754ab7674fSJosh Poimboeuf select OBJTOOL if HAVE_JUMP_LABEL_HACK 7645f81b1cSSteven Rostedt help 77c5905afbSIngo Molnar This option enables a transparent branch optimization that 78c5905afbSIngo Molnar makes certain almost-always-true or almost-always-false branch 79c5905afbSIngo Molnar conditions even cheaper to execute within the kernel. 8045f81b1cSSteven Rostedt 81c5905afbSIngo Molnar Certain performance-sensitive kernel code, such as trace points, 82c5905afbSIngo Molnar scheduler functionality, networking code and KVM have such 83c5905afbSIngo Molnar branches and include support for this optimization technique. 84c5905afbSIngo Molnar 85c5905afbSIngo Molnar If it is detected that the compiler has support for "asm goto", 86c5905afbSIngo Molnar the kernel will compile such branches with just a nop 87c5905afbSIngo Molnar instruction. When the condition flag is toggled to true, the 88c5905afbSIngo Molnar nop will be converted to a jump instruction to execute the 89c5905afbSIngo Molnar conditional block of instructions. 90c5905afbSIngo Molnar 91c5905afbSIngo Molnar This technique lowers overhead and stress on the branch prediction 92c5905afbSIngo Molnar of the processor and generally makes the kernel faster. The update 93c5905afbSIngo Molnar of the condition is slower, but those are always very rare. 94c5905afbSIngo Molnar 95c5905afbSIngo Molnar ( On 32-bit x86, the necessary options added to the compiler 96c5905afbSIngo Molnar flags may increase the size of the kernel slightly. ) 9745f81b1cSSteven Rostedt 981987c947SPeter Zijlstraconfig STATIC_KEYS_SELFTEST 991987c947SPeter Zijlstra bool "Static key selftest" 1001987c947SPeter Zijlstra depends on JUMP_LABEL 1011987c947SPeter Zijlstra help 1021987c947SPeter Zijlstra Boot time self-test of the branch patching code. 1031987c947SPeter Zijlstra 104f03c4129SPeter Zijlstraconfig STATIC_CALL_SELFTEST 105f03c4129SPeter Zijlstra bool "Static call selftest" 106f03c4129SPeter Zijlstra depends on HAVE_STATIC_CALL 107f03c4129SPeter Zijlstra help 108f03c4129SPeter Zijlstra Boot time self-test of the call patching code. 109f03c4129SPeter Zijlstra 110afd66255SMasami Hiramatsuconfig OPTPROBES 1115cc718b9SMasami Hiramatsu def_bool y 1125cc718b9SMasami Hiramatsu depends on KPROBES && HAVE_OPTPROBES 11301b1d88bSThomas Gleixner select TASKS_RCU if PREEMPTION 114afd66255SMasami Hiramatsu 115e7dbfe34SMasami Hiramatsuconfig KPROBES_ON_FTRACE 116e7dbfe34SMasami Hiramatsu def_bool y 117e7dbfe34SMasami Hiramatsu depends on KPROBES && HAVE_KPROBES_ON_FTRACE 118e7dbfe34SMasami Hiramatsu depends on DYNAMIC_FTRACE_WITH_REGS 119e7dbfe34SMasami Hiramatsu help 120e7dbfe34SMasami Hiramatsu If function tracer is enabled and the arch supports full 121e7dbfe34SMasami Hiramatsu passing of pt_regs to function tracing, then kprobes can 122e7dbfe34SMasami Hiramatsu optimize on top of function tracing. 123e7dbfe34SMasami Hiramatsu 1242b144498SSrikar Dronamrajuconfig UPROBES 12509294e31SDavid A. Long def_bool n 126e8f4aa60SAllen Pais depends on ARCH_SUPPORTS_UPROBES 1272b144498SSrikar Dronamraju help 1287b2d81d4SIngo Molnar Uprobes is the user-space counterpart to kprobes: they 1297b2d81d4SIngo Molnar enable instrumentation applications (such as 'perf probe') 1307b2d81d4SIngo Molnar to establish unintrusive probes in user-space binaries and 1317b2d81d4SIngo Molnar libraries, by executing handler functions when the probes 1327b2d81d4SIngo Molnar are hit by user-space applications. 1337b2d81d4SIngo Molnar 1347b2d81d4SIngo Molnar ( These probes come in the form of single-byte breakpoints, 1357b2d81d4SIngo Molnar managed by the kernel and kept transparent to the probed 1367b2d81d4SIngo Molnar application. ) 1372b144498SSrikar Dronamraju 138adab66b7SSteven Rostedt (VMware)config HAVE_64BIT_ALIGNED_ACCESS 139adab66b7SSteven Rostedt (VMware) def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS 140adab66b7SSteven Rostedt (VMware) help 141adab66b7SSteven Rostedt (VMware) Some architectures require 64 bit accesses to be 64 bit 142adab66b7SSteven Rostedt (VMware) aligned, which also requires structs containing 64 bit values 143adab66b7SSteven Rostedt (VMware) to be 64 bit aligned too. This includes some 32 bit 144adab66b7SSteven Rostedt (VMware) architectures which can do 64 bit accesses, as well as 64 bit 145adab66b7SSteven Rostedt (VMware) architectures without unaligned access. 146adab66b7SSteven Rostedt (VMware) 147adab66b7SSteven Rostedt (VMware) This symbol should be selected by an architecture if 64 bit 148adab66b7SSteven Rostedt (VMware) accesses are required to be 64 bit aligned in this way even 149adab66b7SSteven Rostedt (VMware) though it is not a 64 bit architecture. 150adab66b7SSteven Rostedt (VMware) 151ba1a297dSLukas Bulwahn See Documentation/core-api/unaligned-memory-access.rst for 152ba1a297dSLukas Bulwahn more information on the topic of unaligned memory accesses. 153adab66b7SSteven Rostedt (VMware) 15458340a07SJohannes Bergconfig HAVE_EFFICIENT_UNALIGNED_ACCESS 1559ba16087SJan Beulich bool 15658340a07SJohannes Berg help 15758340a07SJohannes Berg Some architectures are unable to perform unaligned accesses 15858340a07SJohannes Berg without the use of get_unaligned/put_unaligned. Others are 15958340a07SJohannes Berg unable to perform such accesses efficiently (e.g. trap on 16058340a07SJohannes Berg unaligned access and require fixing it up in the exception 16158340a07SJohannes Berg handler.) 16258340a07SJohannes Berg 16358340a07SJohannes Berg This symbol should be selected by an architecture if it can 16458340a07SJohannes Berg perform unaligned accesses efficiently to allow different 16558340a07SJohannes Berg code paths to be selected for these cases. Some network 16658340a07SJohannes Berg drivers, for example, could opt to not fix up alignment 16758340a07SJohannes Berg problems with received packets if doing so would not help 16858340a07SJohannes Berg much. 16958340a07SJohannes Berg 170c9b54d6fSMauro Carvalho Chehab See Documentation/core-api/unaligned-memory-access.rst for more 17158340a07SJohannes Berg information on the topic of unaligned memory accesses. 17258340a07SJohannes Berg 173cf66bb93SDavid Woodhouseconfig ARCH_USE_BUILTIN_BSWAP 174cf66bb93SDavid Woodhouse bool 175cf66bb93SDavid Woodhouse help 176cf66bb93SDavid Woodhouse Modern versions of GCC (since 4.4) have builtin functions 177cf66bb93SDavid Woodhouse for handling byte-swapping. Using these, instead of the old 178cf66bb93SDavid Woodhouse inline assembler that the architecture code provides in the 179cf66bb93SDavid Woodhouse __arch_bswapXX() macros, allows the compiler to see what's 180cf66bb93SDavid Woodhouse happening and offers more opportunity for optimisation. In 181cf66bb93SDavid Woodhouse particular, the compiler will be able to combine the byteswap 182cf66bb93SDavid Woodhouse with a nearby load or store and use load-and-swap or 183cf66bb93SDavid Woodhouse store-and-swap instructions if the architecture has them. It 184cf66bb93SDavid Woodhouse should almost *never* result in code which is worse than the 185cf66bb93SDavid Woodhouse hand-coded assembler in <asm/swab.h>. But just in case it 186cf66bb93SDavid Woodhouse does, the use of the builtins is optional. 187cf66bb93SDavid Woodhouse 188cf66bb93SDavid Woodhouse Any architecture with load-and-swap or store-and-swap 189cf66bb93SDavid Woodhouse instructions should set this. And it shouldn't hurt to set it 190cf66bb93SDavid Woodhouse on architectures that don't have such instructions. 191cf66bb93SDavid Woodhouse 1929edddaa2SAnanth N Mavinakayanahalliconfig KRETPROBES 1939edddaa2SAnanth N Mavinakayanahalli def_bool y 19473f9b911SMasami Hiramatsu depends on KPROBES && (HAVE_KRETPROBES || HAVE_RETHOOK) 19573f9b911SMasami Hiramatsu 19673f9b911SMasami Hiramatsuconfig KRETPROBE_ON_RETHOOK 19773f9b911SMasami Hiramatsu def_bool y 19873f9b911SMasami Hiramatsu depends on HAVE_RETHOOK 19973f9b911SMasami Hiramatsu depends on KRETPROBES 20073f9b911SMasami Hiramatsu select RETHOOK 2019edddaa2SAnanth N Mavinakayanahalli 2027c68af6eSAvi Kivityconfig USER_RETURN_NOTIFIER 2037c68af6eSAvi Kivity bool 2047c68af6eSAvi Kivity depends on HAVE_USER_RETURN_NOTIFIER 2057c68af6eSAvi Kivity help 2067c68af6eSAvi Kivity Provide a kernel-internal notification when a cpu is about to 2077c68af6eSAvi Kivity switch to user mode. 2087c68af6eSAvi Kivity 20928b2ee20SRik van Rielconfig HAVE_IOREMAP_PROT 2109ba16087SJan Beulich bool 21128b2ee20SRik van Riel 212125e5645SMathieu Desnoyersconfig HAVE_KPROBES 2139ba16087SJan Beulich bool 2149edddaa2SAnanth N Mavinakayanahalli 2159edddaa2SAnanth N Mavinakayanahalliconfig HAVE_KRETPROBES 2169ba16087SJan Beulich bool 21774bc7ceeSArthur Kepner 218afd66255SMasami Hiramatsuconfig HAVE_OPTPROBES 219afd66255SMasami Hiramatsu bool 220d314d74cSCong Wang 221e7dbfe34SMasami Hiramatsuconfig HAVE_KPROBES_ON_FTRACE 222e7dbfe34SMasami Hiramatsu bool 223e7dbfe34SMasami Hiramatsu 2241f6d3a8fSMasami Hiramatsuconfig ARCH_CORRECT_STACKTRACE_ON_KRETPROBE 2251f6d3a8fSMasami Hiramatsu bool 2261f6d3a8fSMasami Hiramatsu help 2271f6d3a8fSMasami Hiramatsu Since kretprobes modifies return address on the stack, the 2281f6d3a8fSMasami Hiramatsu stacktrace may see the kretprobe trampoline address instead 2291f6d3a8fSMasami Hiramatsu of correct one. If the architecture stacktrace code and 2301f6d3a8fSMasami Hiramatsu unwinder can adjust such entries, select this configuration. 2311f6d3a8fSMasami Hiramatsu 232540adea3SMasami Hiramatsuconfig HAVE_FUNCTION_ERROR_INJECTION 2339802d865SJosef Bacik bool 2349802d865SJosef Bacik 23542a0bb3fSPetr Mladekconfig HAVE_NMI 23642a0bb3fSPetr Mladek bool 23742a0bb3fSPetr Mladek 238a257caccSChristophe Leroyconfig HAVE_FUNCTION_DESCRIPTORS 239a257caccSChristophe Leroy bool 240a257caccSChristophe Leroy 2414aae683fSMasahiro Yamadaconfig TRACE_IRQFLAGS_SUPPORT 2424aae683fSMasahiro Yamada bool 2434aae683fSMasahiro Yamada 2444510bffbSMark Rutlandconfig TRACE_IRQFLAGS_NMI_SUPPORT 2454510bffbSMark Rutland bool 2464510bffbSMark Rutland 2471f5a4ad9SRoland McGrath# 2481f5a4ad9SRoland McGrath# An arch should select this if it provides all these things: 2491f5a4ad9SRoland McGrath# 2501f5a4ad9SRoland McGrath# task_pt_regs() in asm/processor.h or asm/ptrace.h 2511f5a4ad9SRoland McGrath# arch_has_single_step() if there is hardware single-step support 2521f5a4ad9SRoland McGrath# arch_has_block_step() if there is hardware block-step support 2531f5a4ad9SRoland McGrath# asm/syscall.h supplying asm-generic/syscall.h interface 2541f5a4ad9SRoland McGrath# linux/regset.h user_regset interfaces 2551f5a4ad9SRoland McGrath# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 256153474baSEric W. Biederman# TIF_SYSCALL_TRACE calls ptrace_report_syscall_{entry,exit} 25703248addSEric W. Biederman# TIF_NOTIFY_RESUME calls resume_user_mode_work() 2581f5a4ad9SRoland McGrath# 2591f5a4ad9SRoland McGrathconfig HAVE_ARCH_TRACEHOOK 2609ba16087SJan Beulich bool 2611f5a4ad9SRoland McGrath 262c64be2bbSMarek Szyprowskiconfig HAVE_DMA_CONTIGUOUS 263c64be2bbSMarek Szyprowski bool 264c64be2bbSMarek Szyprowski 26529d5e047SThomas Gleixnerconfig GENERIC_SMP_IDLE_THREAD 26629d5e047SThomas Gleixner bool 26729d5e047SThomas Gleixner 268485cf5daSKevin Hilmanconfig GENERIC_IDLE_POLL_SETUP 269485cf5daSKevin Hilman bool 270485cf5daSKevin Hilman 2716974f0c4SDaniel Micayconfig ARCH_HAS_FORTIFY_SOURCE 2726974f0c4SDaniel Micay bool 2736974f0c4SDaniel Micay help 2746974f0c4SDaniel Micay An architecture should select this when it can successfully 2756974f0c4SDaniel Micay build and run with CONFIG_FORTIFY_SOURCE. 2766974f0c4SDaniel Micay 277d8ae8a37SChristoph Hellwig# 278d8ae8a37SChristoph Hellwig# Select if the arch provides a historic keepinit alias for the retain_initrd 279d8ae8a37SChristoph Hellwig# command line option 280d8ae8a37SChristoph Hellwig# 281d8ae8a37SChristoph Hellwigconfig ARCH_HAS_KEEPINITRD 282d8ae8a37SChristoph Hellwig bool 283d8ae8a37SChristoph Hellwig 284d2852a22SDaniel Borkmann# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h 285d2852a22SDaniel Borkmannconfig ARCH_HAS_SET_MEMORY 286d2852a22SDaniel Borkmann bool 287d2852a22SDaniel Borkmann 288d253ca0cSRick Edgecombe# Select if arch has all set_direct_map_invalid/default() functions 289d253ca0cSRick Edgecombeconfig ARCH_HAS_SET_DIRECT_MAP 290d253ca0cSRick Edgecombe bool 291d253ca0cSRick Edgecombe 292c30700dbSChristoph Hellwig# 293fa7e2247SChristoph Hellwig# Select if the architecture provides the arch_dma_set_uncached symbol to 294a86ecfa6SColin Ian King# either provide an uncached segment alias for a DMA allocation, or 295fa7e2247SChristoph Hellwig# to remap the page tables in place. 296c30700dbSChristoph Hellwig# 297fa7e2247SChristoph Hellwigconfig ARCH_HAS_DMA_SET_UNCACHED 298c30700dbSChristoph Hellwig bool 299c30700dbSChristoph Hellwig 300999a5d12SChristoph Hellwig# 301999a5d12SChristoph Hellwig# Select if the architectures provides the arch_dma_clear_uncached symbol 302999a5d12SChristoph Hellwig# to undo an in-place page table remap for uncached access. 303999a5d12SChristoph Hellwig# 304999a5d12SChristoph Hellwigconfig ARCH_HAS_DMA_CLEAR_UNCACHED 305f5e10287SThomas Gleixner bool 306f5e10287SThomas Gleixner 3075905429aSKees Cook# Select if arch init_task must go in the __init_task_data section 3085905429aSKees Cookconfig ARCH_TASK_STRUCT_ON_STACK 3095905429aSKees Cook bool 3105905429aSKees Cook 3115905429aSKees Cook# Select if arch has its private alloc_task_struct() function 3125905429aSKees Cookconfig ARCH_TASK_STRUCT_ALLOCATOR 3135905429aSKees Cook bool 3145905429aSKees Cook 3155905429aSKees Cookconfig HAVE_ARCH_THREAD_STRUCT_WHITELIST 3165905429aSKees Cook bool 3175905429aSKees Cook depends on !ARCH_TASK_STRUCT_ALLOCATOR 318b235beeaSLinus Torvalds help 319b235beeaSLinus Torvalds An architecture should select this to provide hardened usercopy 320f5e10287SThomas Gleixner knowledge about what region of the thread_struct should be 321f5e10287SThomas Gleixner whitelisted for copying to userspace. Normally this is only the 3225aaeb5c0SIngo Molnar FPU registers. Specifically, arch_thread_struct_whitelist() 3235aaeb5c0SIngo Molnar should be implemented. Without this, the entire thread_struct 3245aaeb5c0SIngo Molnar field in task_struct will be left whitelisted. 3255aaeb5c0SIngo Molnar 326942fa985SYury Norov# Select if arch has its private alloc_thread_stack() function 327942fa985SYury Norovconfig ARCH_THREAD_STACK_ALLOCATOR 328942fa985SYury Norov bool 329942fa985SYury Norov 330942fa985SYury Norov# Select if arch wants to size task_struct dynamically via arch_task_struct_size: 331942fa985SYury Norovconfig ARCH_WANTS_DYNAMIC_TASK_STRUCT 332942fa985SYury Norov bool 333942fa985SYury Norov 33451c2ee6dSNick Desaulniersconfig ARCH_WANTS_NO_INSTR 33551c2ee6dSNick Desaulniers bool 33651c2ee6dSNick Desaulniers help 33751c2ee6dSNick Desaulniers An architecture should select this if the noinstr macro is being used on 33851c2ee6dSNick Desaulniers functions to denote that the toolchain should avoid instrumenting such 33951c2ee6dSNick Desaulniers functions and is required for correctness. 34051c2ee6dSNick Desaulniers 341942fa985SYury Norovconfig ARCH_32BIT_OFF_T 342942fa985SYury Norov bool 343942fa985SYury Norov depends on !64BIT 344942fa985SYury Norov help 345942fa985SYury Norov All new 32-bit architectures should have 64-bit off_t type on 346942fa985SYury Norov userspace side which corresponds to the loff_t kernel type. This 347942fa985SYury Norov is the requirement for modern ABIs. Some existing architectures 348942fa985SYury Norov still support 32-bit off_t. This option is enabled for all such 349942fa985SYury Norov architectures explicitly. 350942fa985SYury Norov 35196c0a6a7SHeiko Carstens# Selected by 64 bit architectures which have a 32 bit f_tinode in struct ustat 35296c0a6a7SHeiko Carstensconfig ARCH_32BIT_USTAT_F_TINODE 35396c0a6a7SHeiko Carstens bool 35496c0a6a7SHeiko Carstens 3552ff2b7ecSMasahiro Yamadaconfig HAVE_ASM_MODVERSIONS 3562ff2b7ecSMasahiro Yamada bool 3572ff2b7ecSMasahiro Yamada help 358a86ecfa6SColin Ian King This symbol should be selected by an architecture if it provides 3592ff2b7ecSMasahiro Yamada <asm/asm-prototypes.h> to support the module versioning for symbols 3602ff2b7ecSMasahiro Yamada exported from assembly code. 3612ff2b7ecSMasahiro Yamada 362f850c30cSHeiko Carstensconfig HAVE_REGS_AND_STACK_ACCESS_API 363f850c30cSHeiko Carstens bool 364e01292b1SHeiko Carstens help 365a86ecfa6SColin Ian King This symbol should be selected by an architecture if it supports 366e01292b1SHeiko Carstens the API needed to access registers and stack entries from pt_regs, 367e01292b1SHeiko Carstens declared in asm/ptrace.h 368e01292b1SHeiko Carstens For example the kprobes-based event tracer needs this API. 369f850c30cSHeiko Carstens 370d7822b1eSMathieu Desnoyersconfig HAVE_RSEQ 371d7822b1eSMathieu Desnoyers bool 372d7822b1eSMathieu Desnoyers depends on HAVE_REGS_AND_STACK_ACCESS_API 373d7822b1eSMathieu Desnoyers help 374d7822b1eSMathieu Desnoyers This symbol should be selected by an architecture if it 375d7822b1eSMathieu Desnoyers supports an implementation of restartable sequences. 376d7822b1eSMathieu Desnoyers 3772f7ab126SMiguel Ojedaconfig HAVE_RUST 3782f7ab126SMiguel Ojeda bool 3792f7ab126SMiguel Ojeda help 3802f7ab126SMiguel Ojeda This symbol should be selected by an architecture if it 3812f7ab126SMiguel Ojeda supports Rust. 3822f7ab126SMiguel Ojeda 3833c88ee19SMasami Hiramatsuconfig HAVE_FUNCTION_ARG_ACCESS_API 3843c88ee19SMasami Hiramatsu bool 3853c88ee19SMasami Hiramatsu help 386a86ecfa6SColin Ian King This symbol should be selected by an architecture if it supports 3873c88ee19SMasami Hiramatsu the API needed to access function arguments from pt_regs, 3883c88ee19SMasami Hiramatsu declared in asm/ptrace.h 3893c88ee19SMasami Hiramatsu 39062a038d3SK.Prasadconfig HAVE_HW_BREAKPOINT 39162a038d3SK.Prasad bool 39299e8c5a3SFrederic Weisbecker depends on PERF_EVENTS 39362a038d3SK.Prasad 3940102752eSFrederic Weisbeckerconfig HAVE_MIXED_BREAKPOINTS_REGS 3950102752eSFrederic Weisbecker bool 3960102752eSFrederic Weisbecker depends on HAVE_HW_BREAKPOINT 3970102752eSFrederic Weisbecker help 3980102752eSFrederic Weisbecker Depending on the arch implementation of hardware breakpoints, 3990102752eSFrederic Weisbecker some of them have separate registers for data and instruction 4000102752eSFrederic Weisbecker breakpoints addresses, others have mixed registers to store 4010102752eSFrederic Weisbecker them but define the access type in a control register. 4020102752eSFrederic Weisbecker Select this option if your arch implements breakpoints under the 4030102752eSFrederic Weisbecker latter fashion. 4040102752eSFrederic Weisbecker 4057c68af6eSAvi Kivityconfig HAVE_USER_RETURN_NOTIFIER 4067c68af6eSAvi Kivity bool 407a1922ed6SIngo Molnar 408c01d4323SFrederic Weisbeckerconfig HAVE_PERF_EVENTS_NMI 409c01d4323SFrederic Weisbecker bool 41023637d47SFrederic Weisbecker help 41123637d47SFrederic Weisbecker System hardware can generate an NMI using the perf event 41223637d47SFrederic Weisbecker subsystem. Also has support for calculating CPU cycle events 41323637d47SFrederic Weisbecker to determine how many clock cycles in a given period. 414c01d4323SFrederic Weisbecker 41505a4a952SNicholas Pigginconfig HAVE_HARDLOCKUP_DETECTOR_PERF 41605a4a952SNicholas Piggin bool 41705a4a952SNicholas Piggin depends on HAVE_PERF_EVENTS_NMI 41805a4a952SNicholas Piggin help 41905a4a952SNicholas Piggin The arch chooses to use the generic perf-NMI-based hardlockup 42005a4a952SNicholas Piggin detector. Must define HAVE_PERF_EVENTS_NMI. 42105a4a952SNicholas Piggin 42205a4a952SNicholas Pigginconfig HAVE_NMI_WATCHDOG 42305a4a952SNicholas Piggin depends on HAVE_NMI 42405a4a952SNicholas Piggin bool 42505a4a952SNicholas Piggin help 42605a4a952SNicholas Piggin The arch provides a low level NMI watchdog. It provides 42705a4a952SNicholas Piggin asm/nmi.h, and defines its own arch_touch_nmi_watchdog(). 42805a4a952SNicholas Piggin 42905a4a952SNicholas Pigginconfig HAVE_HARDLOCKUP_DETECTOR_ARCH 43005a4a952SNicholas Piggin bool 43105a4a952SNicholas Piggin select HAVE_NMI_WATCHDOG 43205a4a952SNicholas Piggin help 43305a4a952SNicholas Piggin The arch chooses to provide its own hardlockup detector, which is 43405a4a952SNicholas Piggin a superset of the HAVE_NMI_WATCHDOG. It also conforms to config 43505a4a952SNicholas Piggin interfaces and parameters provided by hardlockup detector subsystem. 43605a4a952SNicholas Piggin 437c5e63197SJiri Olsaconfig HAVE_PERF_REGS 438c5e63197SJiri Olsa bool 439c5e63197SJiri Olsa help 440c5e63197SJiri Olsa Support selective register dumps for perf events. This includes 441c5e63197SJiri Olsa bit-mapping of each registers and a unique architecture id. 442c5e63197SJiri Olsa 443c5ebcedbSJiri Olsaconfig HAVE_PERF_USER_STACK_DUMP 444c5ebcedbSJiri Olsa bool 445c5ebcedbSJiri Olsa help 446c5ebcedbSJiri Olsa Support user stack dumps for perf event samples. This needs 447c5ebcedbSJiri Olsa access to the user stack pointer which is not unified across 448c5ebcedbSJiri Olsa architectures. 449c5ebcedbSJiri Olsa 450bf5438fcSJason Baronconfig HAVE_ARCH_JUMP_LABEL 451bf5438fcSJason Baron bool 452bf5438fcSJason Baron 45350ff18abSArd Biesheuvelconfig HAVE_ARCH_JUMP_LABEL_RELATIVE 45450ff18abSArd Biesheuvel bool 45550ff18abSArd Biesheuvel 4560d6e24d4SPeter Zijlstraconfig MMU_GATHER_TABLE_FREE 4570d6e24d4SPeter Zijlstra bool 4580d6e24d4SPeter Zijlstra 459ff2e6d72SPeter Zijlstraconfig MMU_GATHER_RCU_TABLE_FREE 46026723911SPeter Zijlstra bool 4610d6e24d4SPeter Zijlstra select MMU_GATHER_TABLE_FREE 46226723911SPeter Zijlstra 4633af4bd03SPeter Zijlstraconfig MMU_GATHER_PAGE_SIZE 464ed6a7935SPeter Zijlstra bool 465ed6a7935SPeter Zijlstra 46627796d03SPeter Zijlstraconfig MMU_GATHER_NO_RANGE 46727796d03SPeter Zijlstra bool 4681e9fdf21SPeter Zijlstra select MMU_GATHER_MERGE_VMAS 4691e9fdf21SPeter Zijlstra 4701e9fdf21SPeter Zijlstraconfig MMU_GATHER_NO_FLUSH_CACHE 4711e9fdf21SPeter Zijlstra bool 4721e9fdf21SPeter Zijlstra 4731e9fdf21SPeter Zijlstraconfig MMU_GATHER_MERGE_VMAS 4741e9fdf21SPeter Zijlstra bool 47527796d03SPeter Zijlstra 476580a586cSPeter Zijlstraconfig MMU_GATHER_NO_GATHER 477952a31c9SMartin Schwidefsky bool 4780d6e24d4SPeter Zijlstra depends on MMU_GATHER_TABLE_FREE 479952a31c9SMartin Schwidefsky 480d53c3dfbSNicholas Pigginconfig ARCH_WANT_IRQS_OFF_ACTIVATE_MM 481d53c3dfbSNicholas Piggin bool 482d53c3dfbSNicholas Piggin help 483d53c3dfbSNicholas Piggin Temporary select until all architectures can be converted to have 484d53c3dfbSNicholas Piggin irqs disabled over activate_mm. Architectures that do IPI based TLB 485d53c3dfbSNicholas Piggin shootdowns should enable this. 486d53c3dfbSNicholas Piggin 48788e3009bSNicholas Piggin# Use normal mm refcounting for MMU_LAZY_TLB kernel thread references. 48888e3009bSNicholas Piggin# MMU_LAZY_TLB_REFCOUNT=n can improve the scalability of context switching 48988e3009bSNicholas Piggin# to/from kernel threads when the same mm is running on a lot of CPUs (a large 49088e3009bSNicholas Piggin# multi-threaded application), by reducing contention on the mm refcount. 49188e3009bSNicholas Piggin# 49288e3009bSNicholas Piggin# This can be disabled if the architecture ensures no CPUs are using an mm as a 49388e3009bSNicholas Piggin# "lazy tlb" beyond its final refcount (i.e., by the time __mmdrop frees the mm 49488e3009bSNicholas Piggin# or its kernel page tables). This could be arranged by arch_exit_mmap(), or 49588e3009bSNicholas Piggin# final exit(2) TLB flush, for example. 49688e3009bSNicholas Piggin# 49788e3009bSNicholas Piggin# To implement this, an arch *must*: 49888e3009bSNicholas Piggin# Ensure the _lazy_tlb variants of mmgrab/mmdrop are used when manipulating 49988e3009bSNicholas Piggin# the lazy tlb reference of a kthread's ->active_mm (non-arch code has been 50088e3009bSNicholas Piggin# converted already). 50188e3009bSNicholas Pigginconfig MMU_LAZY_TLB_REFCOUNT 50288e3009bSNicholas Piggin def_bool y 5032655421aSNicholas Piggin depends on !MMU_LAZY_TLB_SHOOTDOWN 5042655421aSNicholas Piggin 5052655421aSNicholas Piggin# This option allows MMU_LAZY_TLB_REFCOUNT=n. It ensures no CPUs are using an 5062655421aSNicholas Piggin# mm as a lazy tlb beyond its last reference count, by shooting down these 5072655421aSNicholas Piggin# users before the mm is deallocated. __mmdrop() first IPIs all CPUs that may 5082655421aSNicholas Piggin# be using the mm as a lazy tlb, so that they may switch themselves to using 5092655421aSNicholas Piggin# init_mm for their active mm. mm_cpumask(mm) is used to determine which CPUs 5102655421aSNicholas Piggin# may be using mm as a lazy tlb mm. 5112655421aSNicholas Piggin# 5122655421aSNicholas Piggin# To implement this, an arch *must*: 5132655421aSNicholas Piggin# - At the time of the final mmdrop of the mm, ensure mm_cpumask(mm) contains 5142655421aSNicholas Piggin# at least all possible CPUs in which the mm is lazy. 5152655421aSNicholas Piggin# - It must meet the requirements for MMU_LAZY_TLB_REFCOUNT=n (see above). 5162655421aSNicholas Pigginconfig MMU_LAZY_TLB_SHOOTDOWN 5172655421aSNicholas Piggin bool 51888e3009bSNicholas Piggin 519df013ffbSHuang Yingconfig ARCH_HAVE_NMI_SAFE_CMPXCHG 520df013ffbSHuang Ying bool 521df013ffbSHuang Ying 5222e83b879SPaul E. McKenneyconfig ARCH_HAS_NMI_SAFE_THIS_CPU_OPS 5232e83b879SPaul E. McKenney bool 5242e83b879SPaul E. McKenney 52543570fd2SHeiko Carstensconfig HAVE_ALIGNED_STRUCT_PAGE 52643570fd2SHeiko Carstens bool 52743570fd2SHeiko Carstens help 52843570fd2SHeiko Carstens This makes sure that struct pages are double word aligned and that 52943570fd2SHeiko Carstens e.g. the SLUB allocator can perform double word atomic operations 53043570fd2SHeiko Carstens on a struct page for better performance. However selecting this 53143570fd2SHeiko Carstens might increase the size of a struct page by a word. 53243570fd2SHeiko Carstens 5334156153cSHeiko Carstensconfig HAVE_CMPXCHG_LOCAL 5344156153cSHeiko Carstens bool 5354156153cSHeiko Carstens 5362565409fSHeiko Carstensconfig HAVE_CMPXCHG_DOUBLE 5372565409fSHeiko Carstens bool 5382565409fSHeiko Carstens 53977e58496SPaul E. McKenneyconfig ARCH_WEAK_RELEASE_ACQUIRE 54077e58496SPaul E. McKenney bool 54177e58496SPaul E. McKenney 542c1d7e01dSWill Deaconconfig ARCH_WANT_IPC_PARSE_VERSION 543c1d7e01dSWill Deacon bool 544c1d7e01dSWill Deacon 545c1d7e01dSWill Deaconconfig ARCH_WANT_COMPAT_IPC_PARSE_VERSION 546c1d7e01dSWill Deacon bool 547c1d7e01dSWill Deacon 54848b25c43SChris Metcalfconfig ARCH_WANT_OLD_COMPAT_IPC 549c1d7e01dSWill Deacon select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 55048b25c43SChris Metcalf bool 55148b25c43SChris Metcalf 552282a181bSYiFei Zhuconfig HAVE_ARCH_SECCOMP 553e2cfabdfSWill Drewry bool 554e2cfabdfSWill Drewry help 555282a181bSYiFei Zhu An arch should select this symbol to support seccomp mode 1 (the fixed 556282a181bSYiFei Zhu syscall policy), and must provide an overrides for __NR_seccomp_sigreturn, 557282a181bSYiFei Zhu and compat syscalls if the asm-generic/seccomp.h defaults need adjustment: 558282a181bSYiFei Zhu - __NR_seccomp_read_32 559282a181bSYiFei Zhu - __NR_seccomp_write_32 560282a181bSYiFei Zhu - __NR_seccomp_exit_32 561282a181bSYiFei Zhu - __NR_seccomp_sigreturn_32 562282a181bSYiFei Zhu 563282a181bSYiFei Zhuconfig HAVE_ARCH_SECCOMP_FILTER 564282a181bSYiFei Zhu bool 565282a181bSYiFei Zhu select HAVE_ARCH_SECCOMP 566282a181bSYiFei Zhu help 567fb0fadf9SWill Drewry An arch should select this symbol if it provides all of these things: 568282a181bSYiFei Zhu - all the requirements for HAVE_ARCH_SECCOMP 569bb6ea430SWill Drewry - syscall_get_arch() 570bb6ea430SWill Drewry - syscall_get_arguments() 571bb6ea430SWill Drewry - syscall_rollback() 572bb6ea430SWill Drewry - syscall_set_return_value() 573fb0fadf9SWill Drewry - SIGSYS siginfo_t support 574fb0fadf9SWill Drewry - secure_computing is called from a ptrace_event()-safe context 575fb0fadf9SWill Drewry - secure_computing return value is checked and a return value of -1 576fb0fadf9SWill Drewry results in the system call being skipped immediately. 57748dc92b9SKees Cook - seccomp syscall wired up 5780d8315ddSYiFei Zhu - if !HAVE_SPARSE_SYSCALL_NR, have SECCOMP_ARCH_NATIVE, 5790d8315ddSYiFei Zhu SECCOMP_ARCH_NATIVE_NR, SECCOMP_ARCH_NATIVE_NAME defined. If 5800d8315ddSYiFei Zhu COMPAT is supported, have the SECCOMP_ARCH_COMPAT* defines too. 581e2cfabdfSWill Drewry 582282a181bSYiFei Zhuconfig SECCOMP 583282a181bSYiFei Zhu prompt "Enable seccomp to safely execute untrusted bytecode" 584282a181bSYiFei Zhu def_bool y 585282a181bSYiFei Zhu depends on HAVE_ARCH_SECCOMP 586282a181bSYiFei Zhu help 587282a181bSYiFei Zhu This kernel feature is useful for number crunching applications 588282a181bSYiFei Zhu that may need to handle untrusted bytecode during their 589282a181bSYiFei Zhu execution. By using pipes or other transports made available 590282a181bSYiFei Zhu to the process as file descriptors supporting the read/write 591282a181bSYiFei Zhu syscalls, it's possible to isolate those applications in their 592282a181bSYiFei Zhu own address space using seccomp. Once seccomp is enabled via 593282a181bSYiFei Zhu prctl(PR_SET_SECCOMP) or the seccomp() syscall, it cannot be 594282a181bSYiFei Zhu disabled and the task is only allowed to execute a few safe 595282a181bSYiFei Zhu syscalls defined by each seccomp mode. 596282a181bSYiFei Zhu 597282a181bSYiFei Zhu If unsure, say Y. 598282a181bSYiFei Zhu 599e2cfabdfSWill Drewryconfig SECCOMP_FILTER 600e2cfabdfSWill Drewry def_bool y 601e2cfabdfSWill Drewry depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 602e2cfabdfSWill Drewry help 603e2cfabdfSWill Drewry Enable tasks to build secure computing environments defined 604e2cfabdfSWill Drewry in terms of Berkeley Packet Filter programs which implement 605e2cfabdfSWill Drewry task-defined system call filtering polices. 606e2cfabdfSWill Drewry 6075fb94e9cSMauro Carvalho Chehab See Documentation/userspace-api/seccomp_filter.rst for details. 608e2cfabdfSWill Drewry 6090d8315ddSYiFei Zhuconfig SECCOMP_CACHE_DEBUG 6100d8315ddSYiFei Zhu bool "Show seccomp filter cache status in /proc/pid/seccomp_cache" 6110d8315ddSYiFei Zhu depends on SECCOMP_FILTER && !HAVE_SPARSE_SYSCALL_NR 6120d8315ddSYiFei Zhu depends on PROC_FS 6130d8315ddSYiFei Zhu help 6140d8315ddSYiFei Zhu This enables the /proc/pid/seccomp_cache interface to monitor 6150d8315ddSYiFei Zhu seccomp cache data. The file format is subject to change. Reading 6160d8315ddSYiFei Zhu the file requires CAP_SYS_ADMIN. 6170d8315ddSYiFei Zhu 6180d8315ddSYiFei Zhu This option is for debugging only. Enabling presents the risk that 6190d8315ddSYiFei Zhu an adversary may be able to infer the seccomp filter logic. 6200d8315ddSYiFei Zhu 6210d8315ddSYiFei Zhu If unsure, say N. 6220d8315ddSYiFei Zhu 623afaef01cSAlexander Popovconfig HAVE_ARCH_STACKLEAK 624afaef01cSAlexander Popov bool 625afaef01cSAlexander Popov help 626afaef01cSAlexander Popov An architecture should select this if it has the code which 627afaef01cSAlexander Popov fills the used part of the kernel stack with the STACKLEAK_POISON 628afaef01cSAlexander Popov value before returning from system calls. 629afaef01cSAlexander Popov 630d148eac0SMasahiro Yamadaconfig HAVE_STACKPROTECTOR 63119952a92SKees Cook bool 63219952a92SKees Cook help 63319952a92SKees Cook An arch should select this symbol if: 63419952a92SKees Cook - it has implemented a stack canary (e.g. __stack_chk_guard) 63519952a92SKees Cook 636050e9baaSLinus Torvaldsconfig STACKPROTECTOR 6372a61f474SMasahiro Yamada bool "Stack Protector buffer overflow detection" 638d148eac0SMasahiro Yamada depends on HAVE_STACKPROTECTOR 6392a61f474SMasahiro Yamada depends on $(cc-option,-fstack-protector) 6402a61f474SMasahiro Yamada default y 6418779657dSKees Cook help 6428779657dSKees Cook This option turns on the "stack-protector" GCC feature. This 64319952a92SKees Cook feature puts, at the beginning of functions, a canary value on 64419952a92SKees Cook the stack just before the return address, and validates 64519952a92SKees Cook the value just before actually returning. Stack based buffer 64619952a92SKees Cook overflows (that need to overwrite this return address) now also 64719952a92SKees Cook overwrite the canary, which gets detected and the attack is then 64819952a92SKees Cook neutralized via a kernel panic. 64919952a92SKees Cook 6508779657dSKees Cook Functions will have the stack-protector canary logic added if they 6518779657dSKees Cook have an 8-byte or larger character array on the stack. 6528779657dSKees Cook 65319952a92SKees Cook This feature requires gcc version 4.2 or above, or a distribution 6548779657dSKees Cook gcc with the feature backported ("-fstack-protector"). 6558779657dSKees Cook 6568779657dSKees Cook On an x86 "defconfig" build, this feature adds canary checks to 6578779657dSKees Cook about 3% of all kernel functions, which increases kernel code size 6588779657dSKees Cook by about 0.3%. 6598779657dSKees Cook 660050e9baaSLinus Torvaldsconfig STACKPROTECTOR_STRONG 6612a61f474SMasahiro Yamada bool "Strong Stack Protector" 662050e9baaSLinus Torvalds depends on STACKPROTECTOR 6632a61f474SMasahiro Yamada depends on $(cc-option,-fstack-protector-strong) 6642a61f474SMasahiro Yamada default y 6658779657dSKees Cook help 6668779657dSKees Cook Functions will have the stack-protector canary logic added in any 6678779657dSKees Cook of the following conditions: 6688779657dSKees Cook 6698779657dSKees Cook - local variable's address used as part of the right hand side of an 6708779657dSKees Cook assignment or function argument 6718779657dSKees Cook - local variable is an array (or union containing an array), 6728779657dSKees Cook regardless of array type or length 6738779657dSKees Cook - uses register local variables 6748779657dSKees Cook 6758779657dSKees Cook This feature requires gcc version 4.9 or above, or a distribution 6768779657dSKees Cook gcc with the feature backported ("-fstack-protector-strong"). 6778779657dSKees Cook 6788779657dSKees Cook On an x86 "defconfig" build, this feature adds canary checks to 6798779657dSKees Cook about 20% of all kernel functions, which increases the kernel code 6808779657dSKees Cook size by about 2%. 6818779657dSKees Cook 682d08b9f0cSSami Tolvanenconfig ARCH_SUPPORTS_SHADOW_CALL_STACK 683d08b9f0cSSami Tolvanen bool 684d08b9f0cSSami Tolvanen help 685afcf5441SDan Li An architecture should select this if it supports the compiler's 686afcf5441SDan Li Shadow Call Stack and implements runtime support for shadow stack 687aa7a65aeSWill Deacon switching. 688d08b9f0cSSami Tolvanen 689d08b9f0cSSami Tolvanenconfig SHADOW_CALL_STACK 690afcf5441SDan Li bool "Shadow Call Stack" 691afcf5441SDan Li depends on ARCH_SUPPORTS_SHADOW_CALL_STACK 69238792972SArd Biesheuvel depends on DYNAMIC_FTRACE_WITH_ARGS || DYNAMIC_FTRACE_WITH_REGS || !FUNCTION_GRAPH_TRACER 693d08b9f0cSSami Tolvanen help 694afcf5441SDan Li This option enables the compiler's Shadow Call Stack, which 695afcf5441SDan Li uses a shadow stack to protect function return addresses from 696afcf5441SDan Li being overwritten by an attacker. More information can be found 697afcf5441SDan Li in the compiler's documentation: 698d08b9f0cSSami Tolvanen 699afcf5441SDan Li - Clang: https://clang.llvm.org/docs/ShadowCallStack.html 700afcf5441SDan Li - GCC: https://gcc.gnu.org/onlinedocs/gcc/Instrumentation-Options.html#Instrumentation-Options 701d08b9f0cSSami Tolvanen 702d08b9f0cSSami Tolvanen Note that security guarantees in the kernel differ from the 703d08b9f0cSSami Tolvanen ones documented for user space. The kernel must store addresses 704d08b9f0cSSami Tolvanen of shadow stacks in memory, which means an attacker capable of 705d08b9f0cSSami Tolvanen reading and writing arbitrary memory may be able to locate them 706d08b9f0cSSami Tolvanen and hijack control flow by modifying the stacks. 707d08b9f0cSSami Tolvanen 7089beccca0SArd Biesheuvelconfig DYNAMIC_SCS 7099beccca0SArd Biesheuvel bool 7109beccca0SArd Biesheuvel help 7119beccca0SArd Biesheuvel Set by the arch code if it relies on code patching to insert the 7129beccca0SArd Biesheuvel shadow call stack push and pop instructions rather than on the 7139beccca0SArd Biesheuvel compiler. 7149beccca0SArd Biesheuvel 715dc5723b0SSami Tolvanenconfig LTO 716dc5723b0SSami Tolvanen bool 717dc5723b0SSami Tolvanen help 718dc5723b0SSami Tolvanen Selected if the kernel will be built using the compiler's LTO feature. 719dc5723b0SSami Tolvanen 720dc5723b0SSami Tolvanenconfig LTO_CLANG 721dc5723b0SSami Tolvanen bool 722dc5723b0SSami Tolvanen select LTO 723dc5723b0SSami Tolvanen help 724dc5723b0SSami Tolvanen Selected if the kernel will be built using Clang's LTO feature. 725dc5723b0SSami Tolvanen 726dc5723b0SSami Tolvanenconfig ARCH_SUPPORTS_LTO_CLANG 727dc5723b0SSami Tolvanen bool 728dc5723b0SSami Tolvanen help 729dc5723b0SSami Tolvanen An architecture should select this option if it supports: 730dc5723b0SSami Tolvanen - compiling with Clang, 731dc5723b0SSami Tolvanen - compiling inline assembly with Clang's integrated assembler, 732dc5723b0SSami Tolvanen - and linking with LLD. 733dc5723b0SSami Tolvanen 734dc5723b0SSami Tolvanenconfig ARCH_SUPPORTS_LTO_CLANG_THIN 735dc5723b0SSami Tolvanen bool 736dc5723b0SSami Tolvanen help 737dc5723b0SSami Tolvanen An architecture should select this option if it can support Clang's 738dc5723b0SSami Tolvanen ThinLTO mode. 739dc5723b0SSami Tolvanen 740dc5723b0SSami Tolvanenconfig HAS_LTO_CLANG 741dc5723b0SSami Tolvanen def_bool y 7421e68a8afSNathan Chancellor depends on CC_IS_CLANG && LD_IS_LLD && AS_IS_LLVM 743dc5723b0SSami Tolvanen depends on $(success,$(NM) --help | head -n 1 | grep -qi llvm) 744dc5723b0SSami Tolvanen depends on $(success,$(AR) --help | head -n 1 | grep -qi llvm) 745dc5723b0SSami Tolvanen depends on ARCH_SUPPORTS_LTO_CLANG 746dc5723b0SSami Tolvanen depends on !FTRACE_MCOUNT_USE_RECORDMCOUNT 747bf3c2551SSami Tolvanen depends on !KASAN || KASAN_HW_TAGS 748dc5723b0SSami Tolvanen depends on !GCOV_KERNEL 749dc5723b0SSami Tolvanen help 750dc5723b0SSami Tolvanen The compiler and Kconfig options support building with Clang's 751dc5723b0SSami Tolvanen LTO. 752dc5723b0SSami Tolvanen 753dc5723b0SSami Tolvanenchoice 754dc5723b0SSami Tolvanen prompt "Link Time Optimization (LTO)" 755dc5723b0SSami Tolvanen default LTO_NONE 756dc5723b0SSami Tolvanen help 757dc5723b0SSami Tolvanen This option enables Link Time Optimization (LTO), which allows the 758dc5723b0SSami Tolvanen compiler to optimize binaries globally. 759dc5723b0SSami Tolvanen 760dc5723b0SSami Tolvanen If unsure, select LTO_NONE. Note that LTO is very resource-intensive 761dc5723b0SSami Tolvanen so it's disabled by default. 762dc5723b0SSami Tolvanen 763dc5723b0SSami Tolvanenconfig LTO_NONE 764dc5723b0SSami Tolvanen bool "None" 765dc5723b0SSami Tolvanen help 766dc5723b0SSami Tolvanen Build the kernel normally, without Link Time Optimization (LTO). 767dc5723b0SSami Tolvanen 768dc5723b0SSami Tolvanenconfig LTO_CLANG_FULL 769dc5723b0SSami Tolvanen bool "Clang Full LTO (EXPERIMENTAL)" 770dc5723b0SSami Tolvanen depends on HAS_LTO_CLANG 771dc5723b0SSami Tolvanen depends on !COMPILE_TEST 772dc5723b0SSami Tolvanen select LTO_CLANG 773dc5723b0SSami Tolvanen help 774dc5723b0SSami Tolvanen This option enables Clang's full Link Time Optimization (LTO), which 775dc5723b0SSami Tolvanen allows the compiler to optimize the kernel globally. If you enable 776dc5723b0SSami Tolvanen this option, the compiler generates LLVM bitcode instead of ELF 777dc5723b0SSami Tolvanen object files, and the actual compilation from bitcode happens at 778dc5723b0SSami Tolvanen the LTO link step, which may take several minutes depending on the 779dc5723b0SSami Tolvanen kernel configuration. More information can be found from LLVM's 780dc5723b0SSami Tolvanen documentation: 781dc5723b0SSami Tolvanen 782dc5723b0SSami Tolvanen https://llvm.org/docs/LinkTimeOptimization.html 783dc5723b0SSami Tolvanen 784dc5723b0SSami Tolvanen During link time, this option can use a large amount of RAM, and 785dc5723b0SSami Tolvanen may take much longer than the ThinLTO option. 786dc5723b0SSami Tolvanen 787dc5723b0SSami Tolvanenconfig LTO_CLANG_THIN 788dc5723b0SSami Tolvanen bool "Clang ThinLTO (EXPERIMENTAL)" 789dc5723b0SSami Tolvanen depends on HAS_LTO_CLANG && ARCH_SUPPORTS_LTO_CLANG_THIN 790dc5723b0SSami Tolvanen select LTO_CLANG 791dc5723b0SSami Tolvanen help 792dc5723b0SSami Tolvanen This option enables Clang's ThinLTO, which allows for parallel 793dc5723b0SSami Tolvanen optimization and faster incremental compiles compared to the 794dc5723b0SSami Tolvanen CONFIG_LTO_CLANG_FULL option. More information can be found 795dc5723b0SSami Tolvanen from Clang's documentation: 796dc5723b0SSami Tolvanen 797dc5723b0SSami Tolvanen https://clang.llvm.org/docs/ThinLTO.html 798dc5723b0SSami Tolvanen 799dc5723b0SSami Tolvanen If unsure, say Y. 800dc5723b0SSami Tolvanenendchoice 801dc5723b0SSami Tolvanen 802cf68fffbSSami Tolvanenconfig ARCH_SUPPORTS_CFI_CLANG 803cf68fffbSSami Tolvanen bool 804cf68fffbSSami Tolvanen help 805cf68fffbSSami Tolvanen An architecture should select this option if it can support Clang's 806cf68fffbSSami Tolvanen Control-Flow Integrity (CFI) checking. 807cf68fffbSSami Tolvanen 80889245600SSami Tolvanenconfig ARCH_USES_CFI_TRAPS 80989245600SSami Tolvanen bool 81089245600SSami Tolvanen 811cf68fffbSSami Tolvanenconfig CFI_CLANG 812cf68fffbSSami Tolvanen bool "Use Clang's Control Flow Integrity (CFI)" 81389245600SSami Tolvanen depends on ARCH_SUPPORTS_CFI_CLANG 81489245600SSami Tolvanen depends on $(cc-option,-fsanitize=kcfi) 815cf68fffbSSami Tolvanen help 816cf68fffbSSami Tolvanen This option enables Clang’s forward-edge Control Flow Integrity 817cf68fffbSSami Tolvanen (CFI) checking, where the compiler injects a runtime check to each 818cf68fffbSSami Tolvanen indirect function call to ensure the target is a valid function with 819cf68fffbSSami Tolvanen the correct static type. This restricts possible call targets and 820cf68fffbSSami Tolvanen makes it more difficult for an attacker to exploit bugs that allow 821cf68fffbSSami Tolvanen the modification of stored function pointers. More information can be 822cf68fffbSSami Tolvanen found from Clang's documentation: 823cf68fffbSSami Tolvanen 824cf68fffbSSami Tolvanen https://clang.llvm.org/docs/ControlFlowIntegrity.html 825cf68fffbSSami Tolvanen 826cf68fffbSSami Tolvanenconfig CFI_PERMISSIVE 827cf68fffbSSami Tolvanen bool "Use CFI in permissive mode" 828cf68fffbSSami Tolvanen depends on CFI_CLANG 829cf68fffbSSami Tolvanen help 830cf68fffbSSami Tolvanen When selected, Control Flow Integrity (CFI) violations result in a 831cf68fffbSSami Tolvanen warning instead of a kernel panic. This option should only be used 832cf68fffbSSami Tolvanen for finding indirect call type mismatches during development. 833cf68fffbSSami Tolvanen 834cf68fffbSSami Tolvanen If unsure, say N. 835cf68fffbSSami Tolvanen 8360f60a8efSKees Cookconfig HAVE_ARCH_WITHIN_STACK_FRAMES 8370f60a8efSKees Cook bool 8380f60a8efSKees Cook help 8390f60a8efSKees Cook An architecture should select this if it can walk the kernel stack 8400f60a8efSKees Cook frames to determine if an object is part of either the arguments 8410f60a8efSKees Cook or local variables (i.e. that it excludes saved return addresses, 8420f60a8efSKees Cook and similar) by implementing an inline arch_within_stack_frames(), 8430f60a8efSKees Cook which is used by CONFIG_HARDENED_USERCOPY. 8440f60a8efSKees Cook 84524a9c541SFrederic Weisbeckerconfig HAVE_CONTEXT_TRACKING_USER 8462b1d5024SFrederic Weisbecker bool 8472b1d5024SFrederic Weisbecker help 84891d1aa43SFrederic Weisbecker Provide kernel/user boundaries probes necessary for subsystems 84991d1aa43SFrederic Weisbecker that need it, such as userspace RCU extended quiescent state. 850490f561bSFrederic Weisbecker Syscalls need to be wrapped inside user_exit()-user_enter(), either 851490f561bSFrederic Weisbecker optimized behind static key or through the slow path using TIF_NOHZ 852490f561bSFrederic Weisbecker flag. Exceptions handlers must be wrapped as well. Irqs are already 8536f0e6c15SFrederic Weisbecker protected inside ct_irq_enter/ct_irq_exit() but preemption or signal 854490f561bSFrederic Weisbecker handling on irq exit still need to be protected. 855490f561bSFrederic Weisbecker 85624a9c541SFrederic Weisbeckerconfig HAVE_CONTEXT_TRACKING_USER_OFFSTACK 85783c2da2eSFrederic Weisbecker bool 85883c2da2eSFrederic Weisbecker help 85983c2da2eSFrederic Weisbecker Architecture neither relies on exception_enter()/exception_exit() 86083c2da2eSFrederic Weisbecker nor on schedule_user(). Also preempt_schedule_notrace() and 86183c2da2eSFrederic Weisbecker preempt_schedule_irq() can't be called in a preemptible section 86283c2da2eSFrederic Weisbecker while context tracking is CONTEXT_USER. This feature reflects a sane 86383c2da2eSFrederic Weisbecker entry implementation where the following requirements are met on 86483c2da2eSFrederic Weisbecker critical entry code, ie: before user_exit() or after user_enter(): 86583c2da2eSFrederic Weisbecker 86683c2da2eSFrederic Weisbecker - Critical entry code isn't preemptible (or better yet: 86783c2da2eSFrederic Weisbecker not interruptible). 868493c1822SFrederic Weisbecker - No use of RCU read side critical sections, unless ct_nmi_enter() 86983c2da2eSFrederic Weisbecker got called. 87083c2da2eSFrederic Weisbecker - No use of instrumentation, unless instrumentation_begin() got 87183c2da2eSFrederic Weisbecker called. 87283c2da2eSFrederic Weisbecker 873490f561bSFrederic Weisbeckerconfig HAVE_TIF_NOHZ 874490f561bSFrederic Weisbecker bool 875490f561bSFrederic Weisbecker help 876490f561bSFrederic Weisbecker Arch relies on TIF_NOHZ and syscall slow path to implement context 877490f561bSFrederic Weisbecker tracking calls to user_enter()/user_exit(). 8782b1d5024SFrederic Weisbecker 879b952741cSFrederic Weisbeckerconfig HAVE_VIRT_CPU_ACCOUNTING 880b952741cSFrederic Weisbecker bool 881b952741cSFrederic Weisbecker 8822b91ec9fSFrederic Weisbeckerconfig HAVE_VIRT_CPU_ACCOUNTING_IDLE 8832b91ec9fSFrederic Weisbecker bool 8842b91ec9fSFrederic Weisbecker help 8852b91ec9fSFrederic Weisbecker Architecture has its own way to account idle CPU time and therefore 8862b91ec9fSFrederic Weisbecker doesn't implement vtime_account_idle(). 8872b91ec9fSFrederic Weisbecker 88840565b5aSStanislaw Gruszkaconfig ARCH_HAS_SCALED_CPUTIME 88940565b5aSStanislaw Gruszka bool 89040565b5aSStanislaw Gruszka 891554b0004SKevin Hilmanconfig HAVE_VIRT_CPU_ACCOUNTING_GEN 892554b0004SKevin Hilman bool 893554b0004SKevin Hilman default y if 64BIT 894554b0004SKevin Hilman help 895554b0004SKevin Hilman With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit. 896554b0004SKevin Hilman Before enabling this option, arch code must be audited 897554b0004SKevin Hilman to ensure there are no races in concurrent read/write of 898554b0004SKevin Hilman cputime_t. For example, reading/writing 64-bit cputime_t on 899554b0004SKevin Hilman some 32-bit arches may require multiple accesses, so proper 900554b0004SKevin Hilman locking is needed to protect against concurrent accesses. 901554b0004SKevin Hilman 902fdf9c356SFrederic Weisbeckerconfig HAVE_IRQ_TIME_ACCOUNTING 903fdf9c356SFrederic Weisbecker bool 904fdf9c356SFrederic Weisbecker help 905fdf9c356SFrederic Weisbecker Archs need to ensure they use a high enough resolution clock to 906fdf9c356SFrederic Weisbecker support irq time accounting and then call enable_sched_clock_irqtime(). 907fdf9c356SFrederic Weisbecker 908c49dd340SKalesh Singhconfig HAVE_MOVE_PUD 909c49dd340SKalesh Singh bool 910c49dd340SKalesh Singh help 911c49dd340SKalesh Singh Architectures that select this are able to move page tables at the 912c49dd340SKalesh Singh PUD level. If there are only 3 page table levels, the move effectively 913c49dd340SKalesh Singh happens at the PGD level. 914c49dd340SKalesh Singh 9152c91bd4aSJoel Fernandes (Google)config HAVE_MOVE_PMD 9162c91bd4aSJoel Fernandes (Google) bool 9172c91bd4aSJoel Fernandes (Google) help 9182c91bd4aSJoel Fernandes (Google) Archs that select this are able to move page tables at the PMD level. 9192c91bd4aSJoel Fernandes (Google) 92015626062SGerald Schaeferconfig HAVE_ARCH_TRANSPARENT_HUGEPAGE 92115626062SGerald Schaefer bool 92215626062SGerald Schaefer 923a00cc7d9SMatthew Wilcoxconfig HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD 924a00cc7d9SMatthew Wilcox bool 925a00cc7d9SMatthew Wilcox 9260ddab1d2SToshi Kaniconfig HAVE_ARCH_HUGE_VMAP 9270ddab1d2SToshi Kani bool 9280ddab1d2SToshi Kani 929121e6f32SNicholas Piggin# 930121e6f32SNicholas Piggin# Archs that select this would be capable of PMD-sized vmaps (i.e., 931559089e0SSong Liu# arch_vmap_pmd_supported() returns true). The VM_ALLOW_HUGE_VMAP flag 932559089e0SSong Liu# must be used to enable allocations to use hugepages. 933121e6f32SNicholas Piggin# 934121e6f32SNicholas Pigginconfig HAVE_ARCH_HUGE_VMALLOC 935121e6f32SNicholas Piggin depends on HAVE_ARCH_HUGE_VMAP 936121e6f32SNicholas Piggin bool 937121e6f32SNicholas Piggin 9383876d4a3SAlexandre Ghiticonfig ARCH_WANT_HUGE_PMD_SHARE 9393876d4a3SAlexandre Ghiti bool 9403876d4a3SAlexandre Ghiti 9410f8975ecSPavel Emelyanovconfig HAVE_ARCH_SOFT_DIRTY 9420f8975ecSPavel Emelyanov bool 9430f8975ecSPavel Emelyanov 944786d35d4SDavid Howellsconfig HAVE_MOD_ARCH_SPECIFIC 945786d35d4SDavid Howells bool 946786d35d4SDavid Howells help 947786d35d4SDavid Howells The arch uses struct mod_arch_specific to store data. Many arches 948786d35d4SDavid Howells just need a simple module loader without arch specific data - those 949786d35d4SDavid Howells should not enable this. 950786d35d4SDavid Howells 951786d35d4SDavid Howellsconfig MODULES_USE_ELF_RELA 952786d35d4SDavid Howells bool 953786d35d4SDavid Howells help 954786d35d4SDavid Howells Modules only use ELF RELA relocations. Modules with ELF REL 955786d35d4SDavid Howells relocations will give an error. 956786d35d4SDavid Howells 957786d35d4SDavid Howellsconfig MODULES_USE_ELF_REL 958786d35d4SDavid Howells bool 959786d35d4SDavid Howells help 960786d35d4SDavid Howells Modules only use ELF REL relocations. Modules with ELF RELA 961786d35d4SDavid Howells relocations will give an error. 962786d35d4SDavid Howells 96301dc0386SChristophe Leroyconfig ARCH_WANTS_MODULES_DATA_IN_VMALLOC 96401dc0386SChristophe Leroy bool 96501dc0386SChristophe Leroy help 96601dc0386SChristophe Leroy For architectures like powerpc/32 which have constraints on module 96701dc0386SChristophe Leroy allocation and need to allocate module data outside of module area. 96801dc0386SChristophe Leroy 969cc1f0274SFrederic Weisbeckerconfig HAVE_IRQ_EXIT_ON_IRQ_STACK 970cc1f0274SFrederic Weisbecker bool 971cc1f0274SFrederic Weisbecker help 972cc1f0274SFrederic Weisbecker Architecture doesn't only execute the irq handler on the irq stack 973cc1f0274SFrederic Weisbecker but also irq_exit(). This way we can process softirqs on this irq 974cc1f0274SFrederic Weisbecker stack instead of switching to a new one when we call __do_softirq() 975cc1f0274SFrederic Weisbecker in the end of an hardirq. 976cc1f0274SFrederic Weisbecker This spares a stack switch and improves cache usage on softirq 977cc1f0274SFrederic Weisbecker processing. 978cc1f0274SFrederic Weisbecker 979cd1a41ceSThomas Gleixnerconfig HAVE_SOFTIRQ_ON_OWN_STACK 980cd1a41ceSThomas Gleixner bool 981cd1a41ceSThomas Gleixner help 982cd1a41ceSThomas Gleixner Architecture provides a function to run __do_softirq() on a 983c226bc3cSColin Ian King separate stack. 984cd1a41ceSThomas Gleixner 9858cbb2b50SSebastian Andrzej Siewiorconfig SOFTIRQ_ON_OWN_STACK 9868cbb2b50SSebastian Andrzej Siewior def_bool HAVE_SOFTIRQ_ON_OWN_STACK && !PREEMPT_RT 9878cbb2b50SSebastian Andrzej Siewior 98812700c17SArnd Bergmannconfig ALTERNATE_USER_ADDRESS_SPACE 98912700c17SArnd Bergmann bool 99012700c17SArnd Bergmann help 99112700c17SArnd Bergmann Architectures set this when the CPU uses separate address 99212700c17SArnd Bergmann spaces for kernel and user space pointers. In this case, the 99312700c17SArnd Bergmann access_ok() check on a __user pointer is skipped. 99412700c17SArnd Bergmann 995235a8f02SKirill A. Shutemovconfig PGTABLE_LEVELS 996235a8f02SKirill A. Shutemov int 997235a8f02SKirill A. Shutemov default 2 998235a8f02SKirill A. Shutemov 9992b68f6caSKees Cookconfig ARCH_HAS_ELF_RANDOMIZE 10002b68f6caSKees Cook bool 10012b68f6caSKees Cook help 10022b68f6caSKees Cook An architecture supports choosing randomized locations for 10032b68f6caSKees Cook stack, mmap, brk, and ET_DYN. Defined functions: 10042b68f6caSKees Cook - arch_mmap_rnd() 1005204db6edSKees Cook - arch_randomize_brk() 10062b68f6caSKees Cook 1007d07e2259SDaniel Cashmanconfig HAVE_ARCH_MMAP_RND_BITS 1008d07e2259SDaniel Cashman bool 1009d07e2259SDaniel Cashman help 1010d07e2259SDaniel Cashman An arch should select this symbol if it supports setting a variable 1011d07e2259SDaniel Cashman number of bits for use in establishing the base address for mmap 1012d07e2259SDaniel Cashman allocations, has MMU enabled and provides values for both: 1013d07e2259SDaniel Cashman - ARCH_MMAP_RND_BITS_MIN 1014d07e2259SDaniel Cashman - ARCH_MMAP_RND_BITS_MAX 1015d07e2259SDaniel Cashman 10165f56a5dfSJiri Slabyconfig HAVE_EXIT_THREAD 10175f56a5dfSJiri Slaby bool 10185f56a5dfSJiri Slaby help 10195f56a5dfSJiri Slaby An architecture implements exit_thread. 10205f56a5dfSJiri Slaby 1021d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_MIN 1022d07e2259SDaniel Cashman int 1023d07e2259SDaniel Cashman 1024d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_MAX 1025d07e2259SDaniel Cashman int 1026d07e2259SDaniel Cashman 1027d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_DEFAULT 1028d07e2259SDaniel Cashman int 1029d07e2259SDaniel Cashman 1030d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS 1031d07e2259SDaniel Cashman int "Number of bits to use for ASLR of mmap base address" if EXPERT 1032d07e2259SDaniel Cashman range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX 1033d07e2259SDaniel Cashman default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT 1034d07e2259SDaniel Cashman default ARCH_MMAP_RND_BITS_MIN 1035d07e2259SDaniel Cashman depends on HAVE_ARCH_MMAP_RND_BITS 1036d07e2259SDaniel Cashman help 1037d07e2259SDaniel Cashman This value can be used to select the number of bits to use to 1038d07e2259SDaniel Cashman determine the random offset to the base address of vma regions 1039d07e2259SDaniel Cashman resulting from mmap allocations. This value will be bounded 1040d07e2259SDaniel Cashman by the architecture's minimum and maximum supported values. 1041d07e2259SDaniel Cashman 1042d07e2259SDaniel Cashman This value can be changed after boot using the 1043d07e2259SDaniel Cashman /proc/sys/vm/mmap_rnd_bits tunable 1044d07e2259SDaniel Cashman 1045d07e2259SDaniel Cashmanconfig HAVE_ARCH_MMAP_RND_COMPAT_BITS 1046d07e2259SDaniel Cashman bool 1047d07e2259SDaniel Cashman help 1048d07e2259SDaniel Cashman An arch should select this symbol if it supports running applications 1049d07e2259SDaniel Cashman in compatibility mode, supports setting a variable number of bits for 1050d07e2259SDaniel Cashman use in establishing the base address for mmap allocations, has MMU 1051d07e2259SDaniel Cashman enabled and provides values for both: 1052d07e2259SDaniel Cashman - ARCH_MMAP_RND_COMPAT_BITS_MIN 1053d07e2259SDaniel Cashman - ARCH_MMAP_RND_COMPAT_BITS_MAX 1054d07e2259SDaniel Cashman 1055d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_MIN 1056d07e2259SDaniel Cashman int 1057d07e2259SDaniel Cashman 1058d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_MAX 1059d07e2259SDaniel Cashman int 1060d07e2259SDaniel Cashman 1061d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1062d07e2259SDaniel Cashman int 1063d07e2259SDaniel Cashman 1064d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS 1065d07e2259SDaniel Cashman int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT 1066d07e2259SDaniel Cashman range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX 1067d07e2259SDaniel Cashman default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT 1068d07e2259SDaniel Cashman default ARCH_MMAP_RND_COMPAT_BITS_MIN 1069d07e2259SDaniel Cashman depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS 1070d07e2259SDaniel Cashman help 1071d07e2259SDaniel Cashman This value can be used to select the number of bits to use to 1072d07e2259SDaniel Cashman determine the random offset to the base address of vma regions 1073d07e2259SDaniel Cashman resulting from mmap allocations for compatible applications This 1074d07e2259SDaniel Cashman value will be bounded by the architecture's minimum and maximum 1075d07e2259SDaniel Cashman supported values. 1076d07e2259SDaniel Cashman 1077d07e2259SDaniel Cashman This value can be changed after boot using the 1078d07e2259SDaniel Cashman /proc/sys/vm/mmap_rnd_compat_bits tunable 1079d07e2259SDaniel Cashman 10801b028f78SDmitry Safonovconfig HAVE_ARCH_COMPAT_MMAP_BASES 10811b028f78SDmitry Safonov bool 10821b028f78SDmitry Safonov help 10831b028f78SDmitry Safonov This allows 64bit applications to invoke 32-bit mmap() syscall 10841b028f78SDmitry Safonov and vice-versa 32-bit applications to call 64-bit mmap(). 10851b028f78SDmitry Safonov Required for applications doing different bitness syscalls. 10861b028f78SDmitry Safonov 10871f0e290cSGuenter Roeckconfig PAGE_SIZE_LESS_THAN_64KB 10881f0e290cSGuenter Roeck def_bool y 10891f0e290cSGuenter Roeck depends on !ARM64_64K_PAGES 10901f0e290cSGuenter Roeck depends on !IA64_PAGE_SIZE_64KB 10911f0e290cSGuenter Roeck depends on !PAGE_SIZE_64KB 10921f0e290cSGuenter Roeck depends on !PARISC_PAGE_SIZE_64KB 1093e4bbd20dSNathan Chancellor depends on PAGE_SIZE_LESS_THAN_256KB 1094e4bbd20dSNathan Chancellor 1095e4bbd20dSNathan Chancellorconfig PAGE_SIZE_LESS_THAN_256KB 1096e4bbd20dSNathan Chancellor def_bool y 10971f0e290cSGuenter Roeck depends on !PAGE_SIZE_256KB 10981f0e290cSGuenter Roeck 109967f3977fSAlexandre Ghiti# This allows to use a set of generic functions to determine mmap base 110067f3977fSAlexandre Ghiti# address by giving priority to top-down scheme only if the process 110167f3977fSAlexandre Ghiti# is not in legacy mode (compat task, unlimited stack size or 110267f3977fSAlexandre Ghiti# sysctl_legacy_va_layout). 110367f3977fSAlexandre Ghiti# Architecture that selects this option can provide its own version of: 110467f3977fSAlexandre Ghiti# - STACK_RND_MASK 110567f3977fSAlexandre Ghiticonfig ARCH_WANT_DEFAULT_TOPDOWN_MMAP_LAYOUT 110667f3977fSAlexandre Ghiti bool 110767f3977fSAlexandre Ghiti depends on MMU 1108e7142bf5SAlexandre Ghiti select ARCH_HAS_ELF_RANDOMIZE 110967f3977fSAlexandre Ghiti 111003f16cd0SJosh Poimboeufconfig HAVE_OBJTOOL 111103f16cd0SJosh Poimboeuf bool 111203f16cd0SJosh Poimboeuf 11134ab7674fSJosh Poimboeufconfig HAVE_JUMP_LABEL_HACK 11144ab7674fSJosh Poimboeuf bool 11154ab7674fSJosh Poimboeuf 111622102f45SJosh Poimboeufconfig HAVE_NOINSTR_HACK 111722102f45SJosh Poimboeuf bool 111822102f45SJosh Poimboeuf 1119489e355bSJosh Poimboeufconfig HAVE_NOINSTR_VALIDATION 1120489e355bSJosh Poimboeuf bool 1121489e355bSJosh Poimboeuf 11225f3da8c0SJosh Poimboeufconfig HAVE_UACCESS_VALIDATION 11235f3da8c0SJosh Poimboeuf bool 11245f3da8c0SJosh Poimboeuf select OBJTOOL 11255f3da8c0SJosh Poimboeuf 1126b9ab5ebbSJosh Poimboeufconfig HAVE_STACK_VALIDATION 1127b9ab5ebbSJosh Poimboeuf bool 1128b9ab5ebbSJosh Poimboeuf help 112903f16cd0SJosh Poimboeuf Architecture supports objtool compile-time frame pointer rule 113003f16cd0SJosh Poimboeuf validation. 1131b9ab5ebbSJosh Poimboeuf 1132af085d90SJosh Poimboeufconfig HAVE_RELIABLE_STACKTRACE 1133af085d90SJosh Poimboeuf bool 1134af085d90SJosh Poimboeuf help 1135140d7e88SMiroslav Benes Architecture has either save_stack_trace_tsk_reliable() or 1136140d7e88SMiroslav Benes arch_stack_walk_reliable() function which only returns a stack trace 1137140d7e88SMiroslav Benes if it can guarantee the trace is reliable. 1138af085d90SJosh Poimboeuf 1139468a9428SGeorge Spelvinconfig HAVE_ARCH_HASH 1140468a9428SGeorge Spelvin bool 1141468a9428SGeorge Spelvin default n 1142468a9428SGeorge Spelvin help 1143468a9428SGeorge Spelvin If this is set, the architecture provides an <asm/hash.h> 1144468a9428SGeorge Spelvin file which provides platform-specific implementations of some 1145468a9428SGeorge Spelvin functions in <linux/hash.h> or fs/namei.c. 1146468a9428SGeorge Spelvin 1147666047feSFinn Thainconfig HAVE_ARCH_NVRAM_OPS 1148666047feSFinn Thain bool 1149666047feSFinn Thain 11503a495511SWilliam Breathitt Grayconfig ISA_BUS_API 11513a495511SWilliam Breathitt Gray def_bool ISA 11523a495511SWilliam Breathitt Gray 1153d2125043SAl Viro# 1154d2125043SAl Viro# ABI hall of shame 1155d2125043SAl Viro# 1156d2125043SAl Viroconfig CLONE_BACKWARDS 1157d2125043SAl Viro bool 1158d2125043SAl Viro help 1159d2125043SAl Viro Architecture has tls passed as the 4th argument of clone(2), 1160d2125043SAl Viro not the 5th one. 1161d2125043SAl Viro 1162d2125043SAl Viroconfig CLONE_BACKWARDS2 1163d2125043SAl Viro bool 1164d2125043SAl Viro help 1165d2125043SAl Viro Architecture has the first two arguments of clone(2) swapped. 1166d2125043SAl Viro 1167dfa9771aSMichal Simekconfig CLONE_BACKWARDS3 1168dfa9771aSMichal Simek bool 1169dfa9771aSMichal Simek help 1170dfa9771aSMichal Simek Architecture has tls passed as the 3rd argument of clone(2), 1171dfa9771aSMichal Simek not the 5th one. 1172dfa9771aSMichal Simek 1173eaca6eaeSAl Viroconfig ODD_RT_SIGACTION 1174eaca6eaeSAl Viro bool 1175eaca6eaeSAl Viro help 1176eaca6eaeSAl Viro Architecture has unusual rt_sigaction(2) arguments 1177eaca6eaeSAl Viro 11780a0e8cdfSAl Viroconfig OLD_SIGSUSPEND 11790a0e8cdfSAl Viro bool 11800a0e8cdfSAl Viro help 11810a0e8cdfSAl Viro Architecture has old sigsuspend(2) syscall, of one-argument variety 11820a0e8cdfSAl Viro 11830a0e8cdfSAl Viroconfig OLD_SIGSUSPEND3 11840a0e8cdfSAl Viro bool 11850a0e8cdfSAl Viro help 11860a0e8cdfSAl Viro Even weirder antique ABI - three-argument sigsuspend(2) 11870a0e8cdfSAl Viro 1188495dfbf7SAl Viroconfig OLD_SIGACTION 1189495dfbf7SAl Viro bool 1190495dfbf7SAl Viro help 1191495dfbf7SAl Viro Architecture has old sigaction(2) syscall. Nope, not the same 1192495dfbf7SAl Viro as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2), 1193495dfbf7SAl Viro but fairly different variant of sigaction(2), thanks to OSF/1 1194495dfbf7SAl Viro compatibility... 1195495dfbf7SAl Viro 1196495dfbf7SAl Viroconfig COMPAT_OLD_SIGACTION 1197495dfbf7SAl Viro bool 1198495dfbf7SAl Viro 119917435e5fSDeepa Dinamaniconfig COMPAT_32BIT_TIME 1200942437c9SArnd Bergmann bool "Provide system calls for 32-bit time_t" 1201942437c9SArnd Bergmann default !64BIT || COMPAT 120217435e5fSDeepa Dinamani help 120317435e5fSDeepa Dinamani This enables 32 bit time_t support in addition to 64 bit time_t support. 120417435e5fSDeepa Dinamani This is relevant on all 32-bit architectures, and 64-bit architectures 120517435e5fSDeepa Dinamani as part of compat syscall handling. 120617435e5fSDeepa Dinamani 120787a4c375SChristoph Hellwigconfig ARCH_NO_PREEMPT 120887a4c375SChristoph Hellwig bool 120987a4c375SChristoph Hellwig 1210cb2c7d1aSMickaël Salaünconfig ARCH_EPHEMERAL_INODES 1211cb2c7d1aSMickaël Salaün def_bool n 1212cb2c7d1aSMickaël Salaün help 1213cb2c7d1aSMickaël Salaün An arch should select this symbol if it doesn't keep track of inode 1214cb2c7d1aSMickaël Salaün instances on its own, but instead relies on something else (e.g. the 1215cb2c7d1aSMickaël Salaün host kernel for an UML kernel). 1216cb2c7d1aSMickaël Salaün 1217a50a3f4bSThomas Gleixnerconfig ARCH_SUPPORTS_RT 1218a50a3f4bSThomas Gleixner bool 1219a50a3f4bSThomas Gleixner 1220fff7fb0bSZhaoxiu Zengconfig CPU_NO_EFFICIENT_FFS 1221fff7fb0bSZhaoxiu Zeng def_bool n 1222fff7fb0bSZhaoxiu Zeng 1223ba14a194SAndy Lutomirskiconfig HAVE_ARCH_VMAP_STACK 1224ba14a194SAndy Lutomirski def_bool n 1225ba14a194SAndy Lutomirski help 1226ba14a194SAndy Lutomirski An arch should select this symbol if it can support kernel stacks 1227ba14a194SAndy Lutomirski in vmalloc space. This means: 1228ba14a194SAndy Lutomirski 1229ba14a194SAndy Lutomirski - vmalloc space must be large enough to hold many kernel stacks. 1230ba14a194SAndy Lutomirski This may rule out many 32-bit architectures. 1231ba14a194SAndy Lutomirski 1232ba14a194SAndy Lutomirski - Stacks in vmalloc space need to work reliably. For example, if 1233ba14a194SAndy Lutomirski vmap page tables are created on demand, either this mechanism 1234ba14a194SAndy Lutomirski needs to work while the stack points to a virtual address with 1235ba14a194SAndy Lutomirski unpopulated page tables or arch code (switch_to() and switch_mm(), 1236ba14a194SAndy Lutomirski most likely) needs to ensure that the stack's page table entries 1237ba14a194SAndy Lutomirski are populated before running on a possibly unpopulated stack. 1238ba14a194SAndy Lutomirski 1239ba14a194SAndy Lutomirski - If the stack overflows into a guard page, something reasonable 1240ba14a194SAndy Lutomirski should happen. The definition of "reasonable" is flexible, but 1241ba14a194SAndy Lutomirski instantly rebooting without logging anything would be unfriendly. 1242ba14a194SAndy Lutomirski 1243ba14a194SAndy Lutomirskiconfig VMAP_STACK 1244ba14a194SAndy Lutomirski default y 1245ba14a194SAndy Lutomirski bool "Use a virtually-mapped stack" 1246eafb149eSDaniel Axtens depends on HAVE_ARCH_VMAP_STACK 124738dd767dSAndrey Konovalov depends on !KASAN || KASAN_HW_TAGS || KASAN_VMALLOC 1248a7f7f624SMasahiro Yamada help 1249ba14a194SAndy Lutomirski Enable this if you want the use virtually-mapped kernel stacks 1250ba14a194SAndy Lutomirski with guard pages. This causes kernel stack overflows to be 1251ba14a194SAndy Lutomirski caught immediately rather than causing difficult-to-diagnose 1252ba14a194SAndy Lutomirski corruption. 1253ba14a194SAndy Lutomirski 125438dd767dSAndrey Konovalov To use this with software KASAN modes, the architecture must support 125538dd767dSAndrey Konovalov backing virtual mappings with real shadow memory, and KASAN_VMALLOC 125638dd767dSAndrey Konovalov must be enabled. 1257ba14a194SAndy Lutomirski 125839218ff4SKees Cookconfig HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 125939218ff4SKees Cook def_bool n 126039218ff4SKees Cook help 126139218ff4SKees Cook An arch should select this symbol if it can support kernel stack 126239218ff4SKees Cook offset randomization with calls to add_random_kstack_offset() 126339218ff4SKees Cook during syscall entry and choose_random_kstack_offset() during 126439218ff4SKees Cook syscall exit. Careful removal of -fstack-protector-strong and 126539218ff4SKees Cook -fstack-protector should also be applied to the entry code and 126639218ff4SKees Cook closely examined, as the artificial stack bump looks like an array 126739218ff4SKees Cook to the compiler, so it will attempt to add canary checks regardless 126839218ff4SKees Cook of the static branch state. 126939218ff4SKees Cook 12708cb37a59SMarco Elverconfig RANDOMIZE_KSTACK_OFFSET 12718cb37a59SMarco Elver bool "Support for randomizing kernel stack offset on syscall entry" if EXPERT 12728cb37a59SMarco Elver default y 127339218ff4SKees Cook depends on HAVE_ARCH_RANDOMIZE_KSTACK_OFFSET 1274efa90c11SMarco Elver depends on INIT_STACK_NONE || !CC_IS_CLANG || CLANG_VERSION >= 140000 127539218ff4SKees Cook help 127639218ff4SKees Cook The kernel stack offset can be randomized (after pt_regs) by 127739218ff4SKees Cook roughly 5 bits of entropy, frustrating memory corruption 127839218ff4SKees Cook attacks that depend on stack address determinism or 12798cb37a59SMarco Elver cross-syscall address exposures. 12808cb37a59SMarco Elver 12818cb37a59SMarco Elver The feature is controlled via the "randomize_kstack_offset=on/off" 12828cb37a59SMarco Elver kernel boot param, and if turned off has zero overhead due to its use 12838cb37a59SMarco Elver of static branches (see JUMP_LABEL). 12848cb37a59SMarco Elver 12858cb37a59SMarco Elver If unsure, say Y. 12868cb37a59SMarco Elver 12878cb37a59SMarco Elverconfig RANDOMIZE_KSTACK_OFFSET_DEFAULT 12888cb37a59SMarco Elver bool "Default state of kernel stack offset randomization" 12898cb37a59SMarco Elver depends on RANDOMIZE_KSTACK_OFFSET 12908cb37a59SMarco Elver help 12918cb37a59SMarco Elver Kernel stack offset randomization is controlled by kernel boot param 12928cb37a59SMarco Elver "randomize_kstack_offset=on/off", and this config chooses the default 12938cb37a59SMarco Elver boot state. 129439218ff4SKees Cook 1295ad21fc4fSLaura Abbottconfig ARCH_OPTIONAL_KERNEL_RWX 1296ad21fc4fSLaura Abbott def_bool n 1297ad21fc4fSLaura Abbott 1298ad21fc4fSLaura Abbottconfig ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1299ad21fc4fSLaura Abbott def_bool n 1300ad21fc4fSLaura Abbott 1301ad21fc4fSLaura Abbottconfig ARCH_HAS_STRICT_KERNEL_RWX 1302ad21fc4fSLaura Abbott def_bool n 1303ad21fc4fSLaura Abbott 13040f5bf6d0SLaura Abbottconfig STRICT_KERNEL_RWX 1305ad21fc4fSLaura Abbott bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX 1306ad21fc4fSLaura Abbott depends on ARCH_HAS_STRICT_KERNEL_RWX 1307ad21fc4fSLaura Abbott default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1308ad21fc4fSLaura Abbott help 1309ad21fc4fSLaura Abbott If this is set, kernel text and rodata memory will be made read-only, 1310ad21fc4fSLaura Abbott and non-text memory will be made non-executable. This provides 1311ad21fc4fSLaura Abbott protection against certain security exploits (e.g. executing the heap 1312ad21fc4fSLaura Abbott or modifying text) 1313ad21fc4fSLaura Abbott 1314ad21fc4fSLaura Abbott These features are considered standard security practice these days. 1315ad21fc4fSLaura Abbott You should say Y here in almost all cases. 1316ad21fc4fSLaura Abbott 1317ad21fc4fSLaura Abbottconfig ARCH_HAS_STRICT_MODULE_RWX 1318ad21fc4fSLaura Abbott def_bool n 1319ad21fc4fSLaura Abbott 13200f5bf6d0SLaura Abbottconfig STRICT_MODULE_RWX 1321ad21fc4fSLaura Abbott bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX 1322ad21fc4fSLaura Abbott depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES 1323ad21fc4fSLaura Abbott default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT 1324ad21fc4fSLaura Abbott help 1325ad21fc4fSLaura Abbott If this is set, module text and rodata memory will be made read-only, 1326ad21fc4fSLaura Abbott and non-text memory will be made non-executable. This provides 1327ad21fc4fSLaura Abbott protection against certain security exploits (e.g. writing to text) 1328ad21fc4fSLaura Abbott 1329ea8c64acSChristoph Hellwig# select if the architecture provides an asm/dma-direct.h header 1330ea8c64acSChristoph Hellwigconfig ARCH_HAS_PHYS_TO_DMA 1331ea8c64acSChristoph Hellwig bool 1332ea8c64acSChristoph Hellwig 133304f264d3SPaul Burtonconfig HAVE_ARCH_COMPILER_H 133404f264d3SPaul Burton bool 133504f264d3SPaul Burton help 133604f264d3SPaul Burton An architecture can select this if it provides an 133704f264d3SPaul Burton asm/compiler.h header that should be included after 133804f264d3SPaul Burton linux/compiler-*.h in order to override macro definitions that those 133904f264d3SPaul Burton headers generally provide. 134004f264d3SPaul Burton 1341271ca788SArd Biesheuvelconfig HAVE_ARCH_PREL32_RELOCATIONS 1342271ca788SArd Biesheuvel bool 1343271ca788SArd Biesheuvel help 1344271ca788SArd Biesheuvel May be selected by an architecture if it supports place-relative 1345271ca788SArd Biesheuvel 32-bit relocations, both in the toolchain and in the module loader, 1346271ca788SArd Biesheuvel in which case relative references can be used in special sections 1347271ca788SArd Biesheuvel for PCI fixup, initcalls etc which are only half the size on 64 bit 1348271ca788SArd Biesheuvel architectures, and don't require runtime relocation on relocatable 1349271ca788SArd Biesheuvel kernels. 1350271ca788SArd Biesheuvel 1351ce9084baSArd Biesheuvelconfig ARCH_USE_MEMREMAP_PROT 1352ce9084baSArd Biesheuvel bool 1353ce9084baSArd Biesheuvel 1354fb346fd9SWaiman Longconfig LOCK_EVENT_COUNTS 1355fb346fd9SWaiman Long bool "Locking event counts collection" 1356fb346fd9SWaiman Long depends on DEBUG_FS 1357a7f7f624SMasahiro Yamada help 1358fb346fd9SWaiman Long Enable light-weight counting of various locking related events 1359fb346fd9SWaiman Long in the system with minimal performance impact. This reduces 1360fb346fd9SWaiman Long the chance of application behavior change because of timing 1361fb346fd9SWaiman Long differences. The counts are reported via debugfs. 1362fb346fd9SWaiman Long 13635cf896fbSPeter Collingbourne# Select if the architecture has support for applying RELR relocations. 13645cf896fbSPeter Collingbourneconfig ARCH_HAS_RELR 13655cf896fbSPeter Collingbourne bool 13665cf896fbSPeter Collingbourne 13675cf896fbSPeter Collingbourneconfig RELR 13685cf896fbSPeter Collingbourne bool "Use RELR relocation packing" 13695cf896fbSPeter Collingbourne depends on ARCH_HAS_RELR && TOOLS_SUPPORT_RELR 13705cf896fbSPeter Collingbourne default y 13715cf896fbSPeter Collingbourne help 13725cf896fbSPeter Collingbourne Store the kernel's dynamic relocations in the RELR relocation packing 13735cf896fbSPeter Collingbourne format. Requires a compatible linker (LLD supports this feature), as 13745cf896fbSPeter Collingbourne well as compatible NM and OBJCOPY utilities (llvm-nm and llvm-objcopy 13755cf896fbSPeter Collingbourne are compatible). 13765cf896fbSPeter Collingbourne 13770c9c1d56SThiago Jung Bauermannconfig ARCH_HAS_MEM_ENCRYPT 13780c9c1d56SThiago Jung Bauermann bool 13790c9c1d56SThiago Jung Bauermann 138046b49b12STom Lendackyconfig ARCH_HAS_CC_PLATFORM 138146b49b12STom Lendacky bool 138246b49b12STom Lendacky 13830e242208SHassan Naveedconfig HAVE_SPARSE_SYSCALL_NR 13840e242208SHassan Naveed bool 13850e242208SHassan Naveed help 13860e242208SHassan Naveed An architecture should select this if its syscall numbering is sparse 13870e242208SHassan Naveed to save space. For example, MIPS architecture has a syscall array with 13880e242208SHassan Naveed entries at 4000, 5000 and 6000 locations. This option turns on syscall 13890e242208SHassan Naveed related optimizations for a given architecture. 13900e242208SHassan Naveed 1391d60d7de3SSven Schnelleconfig ARCH_HAS_VDSO_DATA 1392d60d7de3SSven Schnelle bool 1393d60d7de3SSven Schnelle 1394115284d8SJosh Poimboeufconfig HAVE_STATIC_CALL 1395115284d8SJosh Poimboeuf bool 1396115284d8SJosh Poimboeuf 13979183c3f9SJosh Poimboeufconfig HAVE_STATIC_CALL_INLINE 13989183c3f9SJosh Poimboeuf bool 13999183c3f9SJosh Poimboeuf depends on HAVE_STATIC_CALL 140003f16cd0SJosh Poimboeuf select OBJTOOL 14019183c3f9SJosh Poimboeuf 14026ef869e0SMichal Hockoconfig HAVE_PREEMPT_DYNAMIC 14036ef869e0SMichal Hocko bool 140499cf983cSMark Rutland 140599cf983cSMark Rutlandconfig HAVE_PREEMPT_DYNAMIC_CALL 140699cf983cSMark Rutland bool 14076ef869e0SMichal Hocko depends on HAVE_STATIC_CALL 140899cf983cSMark Rutland select HAVE_PREEMPT_DYNAMIC 14096ef869e0SMichal Hocko help 141099cf983cSMark Rutland An architecture should select this if it can handle the preemption 141199cf983cSMark Rutland model being selected at boot time using static calls. 141299cf983cSMark Rutland 141399cf983cSMark Rutland Where an architecture selects HAVE_STATIC_CALL_INLINE, any call to a 141499cf983cSMark Rutland preemption function will be patched directly. 141599cf983cSMark Rutland 141699cf983cSMark Rutland Where an architecture does not select HAVE_STATIC_CALL_INLINE, any 141799cf983cSMark Rutland call to a preemption function will go through a trampoline, and the 141899cf983cSMark Rutland trampoline will be patched. 141999cf983cSMark Rutland 142099cf983cSMark Rutland It is strongly advised to support inline static call to avoid any 142199cf983cSMark Rutland overhead. 142299cf983cSMark Rutland 142399cf983cSMark Rutlandconfig HAVE_PREEMPT_DYNAMIC_KEY 142499cf983cSMark Rutland bool 1425a0a12c3eSNick Desaulniers depends on HAVE_ARCH_JUMP_LABEL 142699cf983cSMark Rutland select HAVE_PREEMPT_DYNAMIC 142799cf983cSMark Rutland help 142899cf983cSMark Rutland An architecture should select this if it can handle the preemption 142999cf983cSMark Rutland model being selected at boot time using static keys. 143099cf983cSMark Rutland 143199cf983cSMark Rutland Each preemption function will be given an early return based on a 143299cf983cSMark Rutland static key. This should have slightly lower overhead than non-inline 143399cf983cSMark Rutland static calls, as this effectively inlines each trampoline into the 143499cf983cSMark Rutland start of its callee. This may avoid redundant work, and may 143599cf983cSMark Rutland integrate better with CFI schemes. 143699cf983cSMark Rutland 143799cf983cSMark Rutland This will have greater overhead than using inline static calls as 143899cf983cSMark Rutland the call to the preemption function cannot be entirely elided. 14396ef869e0SMichal Hocko 144059612b24SNathan Chancellorconfig ARCH_WANT_LD_ORPHAN_WARN 144159612b24SNathan Chancellor bool 144259612b24SNathan Chancellor help 144359612b24SNathan Chancellor An arch should select this symbol once all linker sections are explicitly 144459612b24SNathan Chancellor included, size-asserted, or discarded in the linker scripts. This is 144559612b24SNathan Chancellor important because we never want expected sections to be placed heuristically 144659612b24SNathan Chancellor by the linker, since the locations of such sections can change between linker 144759612b24SNathan Chancellor versions. 144859612b24SNathan Chancellor 14494f5b0c17SMike Rapoportconfig HAVE_ARCH_PFN_VALID 14504f5b0c17SMike Rapoport bool 14514f5b0c17SMike Rapoport 14525d6ad668SMike Rapoportconfig ARCH_SUPPORTS_DEBUG_PAGEALLOC 14535d6ad668SMike Rapoport bool 14545d6ad668SMike Rapoport 1455df4e817bSPasha Tatashinconfig ARCH_SUPPORTS_PAGE_TABLE_CHECK 1456df4e817bSPasha Tatashin bool 1457df4e817bSPasha Tatashin 14582ca408d9SBrian Gerstconfig ARCH_SPLIT_ARG64 14592ca408d9SBrian Gerst bool 14602ca408d9SBrian Gerst help 14612ca408d9SBrian Gerst If a 32-bit architecture requires 64-bit arguments to be split into 14622ca408d9SBrian Gerst pairs of 32-bit arguments, select this option. 14632ca408d9SBrian Gerst 14647facdc42SAl Viroconfig ARCH_HAS_ELFCORE_COMPAT 14657facdc42SAl Viro bool 14667facdc42SAl Viro 146758e106e7SBalbir Singhconfig ARCH_HAS_PARANOID_L1D_FLUSH 146858e106e7SBalbir Singh bool 146958e106e7SBalbir Singh 1470d593d64fSPrasad Sodagudiconfig ARCH_HAVE_TRACE_MMIO_ACCESS 1471d593d64fSPrasad Sodagudi bool 1472d593d64fSPrasad Sodagudi 14731bdda24cSThomas Gleixnerconfig DYNAMIC_SIGFRAME 14741bdda24cSThomas Gleixner bool 14751bdda24cSThomas Gleixner 147650468e43SJarkko Sakkinen# Select, if arch has a named attribute group bound to NUMA device nodes. 147750468e43SJarkko Sakkinenconfig HAVE_ARCH_NODE_DEV_GROUP 147850468e43SJarkko Sakkinen bool 147950468e43SJarkko Sakkinen 1480eed9a328SYu Zhaoconfig ARCH_HAS_NONLEAF_PMD_YOUNG 1481eed9a328SYu Zhao bool 1482eed9a328SYu Zhao help 1483eed9a328SYu Zhao Architectures that select this option are capable of setting the 1484eed9a328SYu Zhao accessed bit in non-leaf PMD entries when using them as part of linear 1485eed9a328SYu Zhao address translations. Page table walkers that clear the accessed bit 1486eed9a328SYu Zhao may use this capability to reduce their search space. 1487eed9a328SYu Zhao 14882521f2c2SPeter Oberparleitersource "kernel/gcov/Kconfig" 148945332b1bSMasahiro Yamada 149045332b1bSMasahiro Yamadasource "scripts/gcc-plugins/Kconfig" 1491fa1b5d09SLinus Torvalds 1492d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT_4B 1493d49a0626SPeter Zijlstra bool 1494d49a0626SPeter Zijlstra 1495d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT_8B 1496d49a0626SPeter Zijlstra bool 1497d49a0626SPeter Zijlstra 1498d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT_16B 1499d49a0626SPeter Zijlstra bool 1500d49a0626SPeter Zijlstra 1501d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT_32B 1502d49a0626SPeter Zijlstra bool 1503d49a0626SPeter Zijlstra 1504d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT_64B 1505d49a0626SPeter Zijlstra bool 1506d49a0626SPeter Zijlstra 1507d49a0626SPeter Zijlstraconfig FUNCTION_ALIGNMENT 1508d49a0626SPeter Zijlstra int 1509d49a0626SPeter Zijlstra default 64 if FUNCTION_ALIGNMENT_64B 1510d49a0626SPeter Zijlstra default 32 if FUNCTION_ALIGNMENT_32B 1511d49a0626SPeter Zijlstra default 16 if FUNCTION_ALIGNMENT_16B 1512d49a0626SPeter Zijlstra default 8 if FUNCTION_ALIGNMENT_8B 1513d49a0626SPeter Zijlstra default 4 if FUNCTION_ALIGNMENT_4B 1514d49a0626SPeter Zijlstra default 0 1515d49a0626SPeter Zijlstra 151622471e13SRandy Dunlapendmenu 1517