xref: /openbmc/linux/arch/Kconfig (revision 22471e13)
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
21467d2782SThiago Jung Bauermannconfig HAVE_IMA_KEXEC
22467d2782SThiago Jung Bauermann	bool
23467d2782SThiago Jung Bauermann
24125e5645SMathieu Desnoyersconfig OPROFILE
25b309a294SRobert Richter	tristate "OProfile system profiling"
26125e5645SMathieu Desnoyers	depends on PROFILING
27125e5645SMathieu Desnoyers	depends on HAVE_OPROFILE
28d69d59f4SIngo Molnar	select RING_BUFFER
299a5963ebSChristian Borntraeger	select RING_BUFFER_ALLOW_SWAP
30125e5645SMathieu Desnoyers	help
31125e5645SMathieu Desnoyers	  OProfile is a profiling system capable of profiling the
32125e5645SMathieu Desnoyers	  whole system, include the kernel, kernel modules, libraries,
33125e5645SMathieu Desnoyers	  and applications.
34125e5645SMathieu Desnoyers
35125e5645SMathieu Desnoyers	  If unsure, say N.
36125e5645SMathieu Desnoyers
374d4036e0SJason Yehconfig OPROFILE_EVENT_MULTIPLEX
384d4036e0SJason Yeh	bool "OProfile multiplexing support (EXPERIMENTAL)"
394d4036e0SJason Yeh	default n
404d4036e0SJason Yeh	depends on OPROFILE && X86
414d4036e0SJason Yeh	help
424d4036e0SJason Yeh	  The number of hardware counters is limited. The multiplexing
434d4036e0SJason Yeh	  feature enables OProfile to gather more events than counters
444d4036e0SJason Yeh	  are provided by the hardware. This is realized by switching
459332ef9dSMasahiro Yamada	  between events at a user specified time interval.
464d4036e0SJason Yeh
474d4036e0SJason Yeh	  If unsure, say N.
484d4036e0SJason Yeh
49125e5645SMathieu Desnoyersconfig HAVE_OPROFILE
509ba16087SJan Beulich	bool
51125e5645SMathieu Desnoyers
52dcfce4a0SRobert Richterconfig OPROFILE_NMI_TIMER
53dcfce4a0SRobert Richter	def_bool y
54af9feebeSAnton Blanchard	depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI && !PPC64
55dcfce4a0SRobert Richter
56125e5645SMathieu Desnoyersconfig KPROBES
57125e5645SMathieu Desnoyers	bool "Kprobes"
5805ed160eSMasami Hiramatsu	depends on MODULES
59125e5645SMathieu Desnoyers	depends on HAVE_KPROBES
6005ed160eSMasami Hiramatsu	select KALLSYMS
61125e5645SMathieu Desnoyers	help
62125e5645SMathieu Desnoyers	  Kprobes allows you to trap at almost any kernel address and
63125e5645SMathieu Desnoyers	  execute a callback function.  register_kprobe() establishes
64125e5645SMathieu Desnoyers	  a probepoint and specifies the callback.  Kprobes is useful
65125e5645SMathieu Desnoyers	  for kernel debugging, non-intrusive instrumentation and testing.
66125e5645SMathieu Desnoyers	  If in doubt, say "N".
67125e5645SMathieu Desnoyers
6845f81b1cSSteven Rostedtconfig JUMP_LABEL
69c5905afbSIngo Molnar       bool "Optimize very unlikely/likely branches"
7045f81b1cSSteven Rostedt       depends on HAVE_ARCH_JUMP_LABEL
7145f81b1cSSteven Rostedt       help
72c5905afbSIngo Molnar         This option enables a transparent branch optimization that
73c5905afbSIngo Molnar	 makes certain almost-always-true or almost-always-false branch
74c5905afbSIngo Molnar	 conditions even cheaper to execute within the kernel.
7545f81b1cSSteven Rostedt
76c5905afbSIngo Molnar	 Certain performance-sensitive kernel code, such as trace points,
77c5905afbSIngo Molnar	 scheduler functionality, networking code and KVM have such
78c5905afbSIngo Molnar	 branches and include support for this optimization technique.
79c5905afbSIngo Molnar
80c5905afbSIngo Molnar         If it is detected that the compiler has support for "asm goto",
81c5905afbSIngo Molnar	 the kernel will compile such branches with just a nop
82c5905afbSIngo Molnar	 instruction. When the condition flag is toggled to true, the
83c5905afbSIngo Molnar	 nop will be converted to a jump instruction to execute the
84c5905afbSIngo Molnar	 conditional block of instructions.
85c5905afbSIngo Molnar
86c5905afbSIngo Molnar	 This technique lowers overhead and stress on the branch prediction
87c5905afbSIngo Molnar	 of the processor and generally makes the kernel faster. The update
88c5905afbSIngo Molnar	 of the condition is slower, but those are always very rare.
89c5905afbSIngo Molnar
90c5905afbSIngo Molnar	 ( On 32-bit x86, the necessary options added to the compiler
91c5905afbSIngo Molnar	   flags may increase the size of the kernel slightly. )
9245f81b1cSSteven Rostedt
931987c947SPeter Zijlstraconfig STATIC_KEYS_SELFTEST
941987c947SPeter Zijlstra	bool "Static key selftest"
951987c947SPeter Zijlstra	depends on JUMP_LABEL
961987c947SPeter Zijlstra	help
971987c947SPeter Zijlstra	  Boot time self-test of the branch patching code.
981987c947SPeter Zijlstra
99afd66255SMasami Hiramatsuconfig OPTPROBES
1005cc718b9SMasami Hiramatsu	def_bool y
1015cc718b9SMasami Hiramatsu	depends on KPROBES && HAVE_OPTPROBES
102a30b85dfSMasami Hiramatsu	select TASKS_RCU if PREEMPT
103afd66255SMasami Hiramatsu
104e7dbfe34SMasami Hiramatsuconfig KPROBES_ON_FTRACE
105e7dbfe34SMasami Hiramatsu	def_bool y
106e7dbfe34SMasami Hiramatsu	depends on KPROBES && HAVE_KPROBES_ON_FTRACE
107e7dbfe34SMasami Hiramatsu	depends on DYNAMIC_FTRACE_WITH_REGS
108e7dbfe34SMasami Hiramatsu	help
109e7dbfe34SMasami Hiramatsu	 If function tracer is enabled and the arch supports full
110e7dbfe34SMasami Hiramatsu	 passing of pt_regs to function tracing, then kprobes can
111e7dbfe34SMasami Hiramatsu	 optimize on top of function tracing.
112e7dbfe34SMasami Hiramatsu
1132b144498SSrikar Dronamrajuconfig UPROBES
11409294e31SDavid A. Long	def_bool n
115e8f4aa60SAllen Pais	depends on ARCH_SUPPORTS_UPROBES
1162b144498SSrikar Dronamraju	help
1177b2d81d4SIngo Molnar	  Uprobes is the user-space counterpart to kprobes: they
1187b2d81d4SIngo Molnar	  enable instrumentation applications (such as 'perf probe')
1197b2d81d4SIngo Molnar	  to establish unintrusive probes in user-space binaries and
1207b2d81d4SIngo Molnar	  libraries, by executing handler functions when the probes
1217b2d81d4SIngo Molnar	  are hit by user-space applications.
1227b2d81d4SIngo Molnar
1237b2d81d4SIngo Molnar	  ( These probes come in the form of single-byte breakpoints,
1247b2d81d4SIngo Molnar	    managed by the kernel and kept transparent to the probed
1257b2d81d4SIngo Molnar	    application. )
1262b144498SSrikar Dronamraju
127c19fa94aSJames Hoganconfig HAVE_64BIT_ALIGNED_ACCESS
128c19fa94aSJames Hogan	def_bool 64BIT && !HAVE_EFFICIENT_UNALIGNED_ACCESS
129c19fa94aSJames Hogan	help
130c19fa94aSJames Hogan	  Some architectures require 64 bit accesses to be 64 bit
131c19fa94aSJames Hogan	  aligned, which also requires structs containing 64 bit values
132c19fa94aSJames Hogan	  to be 64 bit aligned too. This includes some 32 bit
133c19fa94aSJames Hogan	  architectures which can do 64 bit accesses, as well as 64 bit
134c19fa94aSJames Hogan	  architectures without unaligned access.
135c19fa94aSJames Hogan
136c19fa94aSJames Hogan	  This symbol should be selected by an architecture if 64 bit
137c19fa94aSJames Hogan	  accesses are required to be 64 bit aligned in this way even
138c19fa94aSJames Hogan	  though it is not a 64 bit architecture.
139c19fa94aSJames Hogan
140c19fa94aSJames Hogan	  See Documentation/unaligned-memory-access.txt for more
141c19fa94aSJames Hogan	  information on the topic of unaligned memory accesses.
142c19fa94aSJames Hogan
14358340a07SJohannes Bergconfig HAVE_EFFICIENT_UNALIGNED_ACCESS
1449ba16087SJan Beulich	bool
14558340a07SJohannes Berg	help
14658340a07SJohannes Berg	  Some architectures are unable to perform unaligned accesses
14758340a07SJohannes Berg	  without the use of get_unaligned/put_unaligned. Others are
14858340a07SJohannes Berg	  unable to perform such accesses efficiently (e.g. trap on
14958340a07SJohannes Berg	  unaligned access and require fixing it up in the exception
15058340a07SJohannes Berg	  handler.)
15158340a07SJohannes Berg
15258340a07SJohannes Berg	  This symbol should be selected by an architecture if it can
15358340a07SJohannes Berg	  perform unaligned accesses efficiently to allow different
15458340a07SJohannes Berg	  code paths to be selected for these cases. Some network
15558340a07SJohannes Berg	  drivers, for example, could opt to not fix up alignment
15658340a07SJohannes Berg	  problems with received packets if doing so would not help
15758340a07SJohannes Berg	  much.
15858340a07SJohannes Berg
15958340a07SJohannes Berg	  See Documentation/unaligned-memory-access.txt for more
16058340a07SJohannes Berg	  information on the topic of unaligned memory accesses.
16158340a07SJohannes Berg
162cf66bb93SDavid Woodhouseconfig ARCH_USE_BUILTIN_BSWAP
163cf66bb93SDavid Woodhouse       bool
164cf66bb93SDavid Woodhouse       help
165cf66bb93SDavid Woodhouse	 Modern versions of GCC (since 4.4) have builtin functions
166cf66bb93SDavid Woodhouse	 for handling byte-swapping. Using these, instead of the old
167cf66bb93SDavid Woodhouse	 inline assembler that the architecture code provides in the
168cf66bb93SDavid Woodhouse	 __arch_bswapXX() macros, allows the compiler to see what's
169cf66bb93SDavid Woodhouse	 happening and offers more opportunity for optimisation. In
170cf66bb93SDavid Woodhouse	 particular, the compiler will be able to combine the byteswap
171cf66bb93SDavid Woodhouse	 with a nearby load or store and use load-and-swap or
172cf66bb93SDavid Woodhouse	 store-and-swap instructions if the architecture has them. It
173cf66bb93SDavid Woodhouse	 should almost *never* result in code which is worse than the
174cf66bb93SDavid Woodhouse	 hand-coded assembler in <asm/swab.h>.  But just in case it
175cf66bb93SDavid Woodhouse	 does, the use of the builtins is optional.
176cf66bb93SDavid Woodhouse
177cf66bb93SDavid Woodhouse	 Any architecture with load-and-swap or store-and-swap
178cf66bb93SDavid Woodhouse	 instructions should set this. And it shouldn't hurt to set it
179cf66bb93SDavid Woodhouse	 on architectures that don't have such instructions.
180cf66bb93SDavid Woodhouse
1819edddaa2SAnanth N Mavinakayanahalliconfig KRETPROBES
1829edddaa2SAnanth N Mavinakayanahalli	def_bool y
1839edddaa2SAnanth N Mavinakayanahalli	depends on KPROBES && HAVE_KRETPROBES
1849edddaa2SAnanth N Mavinakayanahalli
1857c68af6eSAvi Kivityconfig USER_RETURN_NOTIFIER
1867c68af6eSAvi Kivity	bool
1877c68af6eSAvi Kivity	depends on HAVE_USER_RETURN_NOTIFIER
1887c68af6eSAvi Kivity	help
1897c68af6eSAvi Kivity	  Provide a kernel-internal notification when a cpu is about to
1907c68af6eSAvi Kivity	  switch to user mode.
1917c68af6eSAvi Kivity
19228b2ee20SRik van Rielconfig HAVE_IOREMAP_PROT
1939ba16087SJan Beulich	bool
19428b2ee20SRik van Riel
195125e5645SMathieu Desnoyersconfig HAVE_KPROBES
1969ba16087SJan Beulich	bool
1979edddaa2SAnanth N Mavinakayanahalli
1989edddaa2SAnanth N Mavinakayanahalliconfig HAVE_KRETPROBES
1999ba16087SJan Beulich	bool
20074bc7ceeSArthur Kepner
201afd66255SMasami Hiramatsuconfig HAVE_OPTPROBES
202afd66255SMasami Hiramatsu	bool
203d314d74cSCong Wang
204e7dbfe34SMasami Hiramatsuconfig HAVE_KPROBES_ON_FTRACE
205e7dbfe34SMasami Hiramatsu	bool
206e7dbfe34SMasami Hiramatsu
207540adea3SMasami Hiramatsuconfig HAVE_FUNCTION_ERROR_INJECTION
2089802d865SJosef Bacik	bool
2099802d865SJosef Bacik
21042a0bb3fSPetr Mladekconfig HAVE_NMI
21142a0bb3fSPetr Mladek	bool
21242a0bb3fSPetr Mladek
2131f5a4ad9SRoland McGrath#
2141f5a4ad9SRoland McGrath# An arch should select this if it provides all these things:
2151f5a4ad9SRoland McGrath#
2161f5a4ad9SRoland McGrath#	task_pt_regs()		in asm/processor.h or asm/ptrace.h
2171f5a4ad9SRoland McGrath#	arch_has_single_step()	if there is hardware single-step support
2181f5a4ad9SRoland McGrath#	arch_has_block_step()	if there is hardware block-step support
2191f5a4ad9SRoland McGrath#	asm/syscall.h		supplying asm-generic/syscall.h interface
2201f5a4ad9SRoland McGrath#	linux/regset.h		user_regset interfaces
2211f5a4ad9SRoland McGrath#	CORE_DUMP_USE_REGSET	#define'd in linux/elf.h
2221f5a4ad9SRoland McGrath#	TIF_SYSCALL_TRACE	calls tracehook_report_syscall_{entry,exit}
2231f5a4ad9SRoland McGrath#	TIF_NOTIFY_RESUME	calls tracehook_notify_resume()
2241f5a4ad9SRoland McGrath#	signal delivery		calls tracehook_signal_handler()
2251f5a4ad9SRoland McGrath#
2261f5a4ad9SRoland McGrathconfig HAVE_ARCH_TRACEHOOK
2279ba16087SJan Beulich	bool
2281f5a4ad9SRoland McGrath
229c64be2bbSMarek Szyprowskiconfig HAVE_DMA_CONTIGUOUS
230c64be2bbSMarek Szyprowski	bool
231c64be2bbSMarek Szyprowski
23229d5e047SThomas Gleixnerconfig GENERIC_SMP_IDLE_THREAD
23329d5e047SThomas Gleixner       bool
23429d5e047SThomas Gleixner
235485cf5daSKevin Hilmanconfig GENERIC_IDLE_POLL_SETUP
236485cf5daSKevin Hilman       bool
237485cf5daSKevin Hilman
2386974f0c4SDaniel Micayconfig ARCH_HAS_FORTIFY_SOURCE
2396974f0c4SDaniel Micay	bool
2406974f0c4SDaniel Micay	help
2416974f0c4SDaniel Micay	  An architecture should select this when it can successfully
2426974f0c4SDaniel Micay	  build and run with CONFIG_FORTIFY_SOURCE.
2436974f0c4SDaniel Micay
244d2852a22SDaniel Borkmann# Select if arch has all set_memory_ro/rw/x/nx() functions in asm/cacheflush.h
245d2852a22SDaniel Borkmannconfig ARCH_HAS_SET_MEMORY
246d2852a22SDaniel Borkmann	bool
247d2852a22SDaniel Borkmann
2480500871fSDavid Howells# Select if arch init_task must go in the __init_task_data section
2490500871fSDavid Howellsconfig ARCH_TASK_STRUCT_ON_STACK
250a4a2eb49SThomas Gleixner       bool
251a4a2eb49SThomas Gleixner
252f5e10287SThomas Gleixner# Select if arch has its private alloc_task_struct() function
253f5e10287SThomas Gleixnerconfig ARCH_TASK_STRUCT_ALLOCATOR
254f5e10287SThomas Gleixner	bool
255f5e10287SThomas Gleixner
2565905429aSKees Cookconfig HAVE_ARCH_THREAD_STRUCT_WHITELIST
2575905429aSKees Cook	bool
2585905429aSKees Cook	depends on !ARCH_TASK_STRUCT_ALLOCATOR
2595905429aSKees Cook	help
2605905429aSKees Cook	  An architecture should select this to provide hardened usercopy
2615905429aSKees Cook	  knowledge about what region of the thread_struct should be
2625905429aSKees Cook	  whitelisted for copying to userspace. Normally this is only the
2635905429aSKees Cook	  FPU registers. Specifically, arch_thread_struct_whitelist()
2645905429aSKees Cook	  should be implemented. Without this, the entire thread_struct
2655905429aSKees Cook	  field in task_struct will be left whitelisted.
2665905429aSKees Cook
267b235beeaSLinus Torvalds# Select if arch has its private alloc_thread_stack() function
268b235beeaSLinus Torvaldsconfig ARCH_THREAD_STACK_ALLOCATOR
269f5e10287SThomas Gleixner	bool
270f5e10287SThomas Gleixner
2715aaeb5c0SIngo Molnar# Select if arch wants to size task_struct dynamically via arch_task_struct_size:
2725aaeb5c0SIngo Molnarconfig ARCH_WANTS_DYNAMIC_TASK_STRUCT
2735aaeb5c0SIngo Molnar	bool
2745aaeb5c0SIngo Molnar
275f850c30cSHeiko Carstensconfig HAVE_REGS_AND_STACK_ACCESS_API
276f850c30cSHeiko Carstens	bool
277e01292b1SHeiko Carstens	help
278e01292b1SHeiko Carstens	  This symbol should be selected by an architecure if it supports
279e01292b1SHeiko Carstens	  the API needed to access registers and stack entries from pt_regs,
280e01292b1SHeiko Carstens	  declared in asm/ptrace.h
281e01292b1SHeiko Carstens	  For example the kprobes-based event tracer needs this API.
282f850c30cSHeiko Carstens
283d7822b1eSMathieu Desnoyersconfig HAVE_RSEQ
284d7822b1eSMathieu Desnoyers	bool
285d7822b1eSMathieu Desnoyers	depends on HAVE_REGS_AND_STACK_ACCESS_API
286d7822b1eSMathieu Desnoyers	help
287d7822b1eSMathieu Desnoyers	  This symbol should be selected by an architecture if it
288d7822b1eSMathieu Desnoyers	  supports an implementation of restartable sequences.
289d7822b1eSMathieu Desnoyers
2909483a578SDavid Brownellconfig HAVE_CLK
2919ba16087SJan Beulich	bool
2929483a578SDavid Brownell	help
2939483a578SDavid Brownell	  The <linux/clk.h> calls support software clock gating and
2949483a578SDavid Brownell	  thus are a key power management tool on many systems.
2959483a578SDavid Brownell
29662a038d3SK.Prasadconfig HAVE_HW_BREAKPOINT
29762a038d3SK.Prasad	bool
29899e8c5a3SFrederic Weisbecker	depends on PERF_EVENTS
29962a038d3SK.Prasad
3000102752eSFrederic Weisbeckerconfig HAVE_MIXED_BREAKPOINTS_REGS
3010102752eSFrederic Weisbecker	bool
3020102752eSFrederic Weisbecker	depends on HAVE_HW_BREAKPOINT
3030102752eSFrederic Weisbecker	help
3040102752eSFrederic Weisbecker	  Depending on the arch implementation of hardware breakpoints,
3050102752eSFrederic Weisbecker	  some of them have separate registers for data and instruction
3060102752eSFrederic Weisbecker	  breakpoints addresses, others have mixed registers to store
3070102752eSFrederic Weisbecker	  them but define the access type in a control register.
3080102752eSFrederic Weisbecker	  Select this option if your arch implements breakpoints under the
3090102752eSFrederic Weisbecker	  latter fashion.
3100102752eSFrederic Weisbecker
3117c68af6eSAvi Kivityconfig HAVE_USER_RETURN_NOTIFIER
3127c68af6eSAvi Kivity	bool
313a1922ed6SIngo Molnar
314c01d4323SFrederic Weisbeckerconfig HAVE_PERF_EVENTS_NMI
315c01d4323SFrederic Weisbecker	bool
31623637d47SFrederic Weisbecker	help
31723637d47SFrederic Weisbecker	  System hardware can generate an NMI using the perf event
31823637d47SFrederic Weisbecker	  subsystem.  Also has support for calculating CPU cycle events
31923637d47SFrederic Weisbecker	  to determine how many clock cycles in a given period.
320c01d4323SFrederic Weisbecker
32105a4a952SNicholas Pigginconfig HAVE_HARDLOCKUP_DETECTOR_PERF
32205a4a952SNicholas Piggin	bool
32305a4a952SNicholas Piggin	depends on HAVE_PERF_EVENTS_NMI
32405a4a952SNicholas Piggin	help
32505a4a952SNicholas Piggin	  The arch chooses to use the generic perf-NMI-based hardlockup
32605a4a952SNicholas Piggin	  detector. Must define HAVE_PERF_EVENTS_NMI.
32705a4a952SNicholas Piggin
32805a4a952SNicholas Pigginconfig HAVE_NMI_WATCHDOG
32905a4a952SNicholas Piggin	depends on HAVE_NMI
33005a4a952SNicholas Piggin	bool
33105a4a952SNicholas Piggin	help
33205a4a952SNicholas Piggin	  The arch provides a low level NMI watchdog. It provides
33305a4a952SNicholas Piggin	  asm/nmi.h, and defines its own arch_touch_nmi_watchdog().
33405a4a952SNicholas Piggin
33505a4a952SNicholas Pigginconfig HAVE_HARDLOCKUP_DETECTOR_ARCH
33605a4a952SNicholas Piggin	bool
33705a4a952SNicholas Piggin	select HAVE_NMI_WATCHDOG
33805a4a952SNicholas Piggin	help
33905a4a952SNicholas Piggin	  The arch chooses to provide its own hardlockup detector, which is
34005a4a952SNicholas Piggin	  a superset of the HAVE_NMI_WATCHDOG. It also conforms to config
34105a4a952SNicholas Piggin	  interfaces and parameters provided by hardlockup detector subsystem.
34205a4a952SNicholas Piggin
343c5e63197SJiri Olsaconfig HAVE_PERF_REGS
344c5e63197SJiri Olsa	bool
345c5e63197SJiri Olsa	help
346c5e63197SJiri Olsa	  Support selective register dumps for perf events. This includes
347c5e63197SJiri Olsa	  bit-mapping of each registers and a unique architecture id.
348c5e63197SJiri Olsa
349c5ebcedbSJiri Olsaconfig HAVE_PERF_USER_STACK_DUMP
350c5ebcedbSJiri Olsa	bool
351c5ebcedbSJiri Olsa	help
352c5ebcedbSJiri Olsa	  Support user stack dumps for perf event samples. This needs
353c5ebcedbSJiri Olsa	  access to the user stack pointer which is not unified across
354c5ebcedbSJiri Olsa	  architectures.
355c5ebcedbSJiri Olsa
356bf5438fcSJason Baronconfig HAVE_ARCH_JUMP_LABEL
357bf5438fcSJason Baron	bool
358bf5438fcSJason Baron
35926723911SPeter Zijlstraconfig HAVE_RCU_TABLE_FREE
36026723911SPeter Zijlstra	bool
36126723911SPeter Zijlstra
362df013ffbSHuang Yingconfig ARCH_HAVE_NMI_SAFE_CMPXCHG
363df013ffbSHuang Ying	bool
364df013ffbSHuang Ying
36543570fd2SHeiko Carstensconfig HAVE_ALIGNED_STRUCT_PAGE
36643570fd2SHeiko Carstens	bool
36743570fd2SHeiko Carstens	help
36843570fd2SHeiko Carstens	  This makes sure that struct pages are double word aligned and that
36943570fd2SHeiko Carstens	  e.g. the SLUB allocator can perform double word atomic operations
37043570fd2SHeiko Carstens	  on a struct page for better performance. However selecting this
37143570fd2SHeiko Carstens	  might increase the size of a struct page by a word.
37243570fd2SHeiko Carstens
3734156153cSHeiko Carstensconfig HAVE_CMPXCHG_LOCAL
3744156153cSHeiko Carstens	bool
3754156153cSHeiko Carstens
3762565409fSHeiko Carstensconfig HAVE_CMPXCHG_DOUBLE
3772565409fSHeiko Carstens	bool
3782565409fSHeiko Carstens
37977e58496SPaul E. McKenneyconfig ARCH_WEAK_RELEASE_ACQUIRE
38077e58496SPaul E. McKenney	bool
38177e58496SPaul E. McKenney
382c1d7e01dSWill Deaconconfig ARCH_WANT_IPC_PARSE_VERSION
383c1d7e01dSWill Deacon	bool
384c1d7e01dSWill Deacon
385c1d7e01dSWill Deaconconfig ARCH_WANT_COMPAT_IPC_PARSE_VERSION
386c1d7e01dSWill Deacon	bool
387c1d7e01dSWill Deacon
38848b25c43SChris Metcalfconfig ARCH_WANT_OLD_COMPAT_IPC
389c1d7e01dSWill Deacon	select ARCH_WANT_COMPAT_IPC_PARSE_VERSION
39048b25c43SChris Metcalf	bool
39148b25c43SChris Metcalf
392e2cfabdfSWill Drewryconfig HAVE_ARCH_SECCOMP_FILTER
393e2cfabdfSWill Drewry	bool
394e2cfabdfSWill Drewry	help
395fb0fadf9SWill Drewry	  An arch should select this symbol if it provides all of these things:
396bb6ea430SWill Drewry	  - syscall_get_arch()
397bb6ea430SWill Drewry	  - syscall_get_arguments()
398bb6ea430SWill Drewry	  - syscall_rollback()
399bb6ea430SWill Drewry	  - syscall_set_return_value()
400fb0fadf9SWill Drewry	  - SIGSYS siginfo_t support
401fb0fadf9SWill Drewry	  - secure_computing is called from a ptrace_event()-safe context
402fb0fadf9SWill Drewry	  - secure_computing return value is checked and a return value of -1
403fb0fadf9SWill Drewry	    results in the system call being skipped immediately.
40448dc92b9SKees Cook	  - seccomp syscall wired up
405e2cfabdfSWill Drewry
406e2cfabdfSWill Drewryconfig SECCOMP_FILTER
407e2cfabdfSWill Drewry	def_bool y
408e2cfabdfSWill Drewry	depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET
409e2cfabdfSWill Drewry	help
410e2cfabdfSWill Drewry	  Enable tasks to build secure computing environments defined
411e2cfabdfSWill Drewry	  in terms of Berkeley Packet Filter programs which implement
412e2cfabdfSWill Drewry	  task-defined system call filtering polices.
413e2cfabdfSWill Drewry
4145fb94e9cSMauro Carvalho Chehab	  See Documentation/userspace-api/seccomp_filter.rst for details.
415e2cfabdfSWill Drewry
41659f53855SMasahiro Yamadapreferred-plugin-hostcc := $(if-success,[ $(gcc-version) -ge 40800 ],$(HOSTCXX),$(HOSTCC))
41759f53855SMasahiro Yamada
41859f53855SMasahiro Yamadaconfig PLUGIN_HOSTCC
41959f53855SMasahiro Yamada	string
42059f53855SMasahiro Yamada	default "$(shell,$(srctree)/scripts/gcc-plugin.sh "$(preferred-plugin-hostcc)" "$(HOSTCXX)" "$(CC)")"
42159f53855SMasahiro Yamada	help
42259f53855SMasahiro Yamada	  Host compiler used to build GCC plugins.  This can be $(HOSTCXX),
42359f53855SMasahiro Yamada	  $(HOSTCC), or a null string if GCC plugin is unsupported.
42459f53855SMasahiro Yamada
4256b90bd4bSEmese Revfyconfig HAVE_GCC_PLUGINS
4266b90bd4bSEmese Revfy	bool
4276b90bd4bSEmese Revfy	help
4286b90bd4bSEmese Revfy	  An arch should select this symbol if it supports building with
4296b90bd4bSEmese Revfy	  GCC plugins.
4306b90bd4bSEmese Revfy
4316b90bd4bSEmese Revfymenuconfig GCC_PLUGINS
4326b90bd4bSEmese Revfy	bool "GCC plugins"
4336b90bd4bSEmese Revfy	depends on HAVE_GCC_PLUGINS
43459f53855SMasahiro Yamada	depends on PLUGIN_HOSTCC != ""
4356b90bd4bSEmese Revfy	help
4366b90bd4bSEmese Revfy	  GCC plugins are loadable modules that provide extra features to the
4376b90bd4bSEmese Revfy	  compiler. They are useful for runtime instrumentation and static analysis.
4386b90bd4bSEmese Revfy
4396b90bd4bSEmese Revfy	  See Documentation/gcc-plugins.txt for details.
4406b90bd4bSEmese Revfy
4410dae776cSEmese Revfyconfig GCC_PLUGIN_CYC_COMPLEXITY
442215e2aa6SKees Cook	bool "Compute the cyclomatic complexity of a function" if EXPERT
4430dae776cSEmese Revfy	depends on GCC_PLUGINS
4441658dceeSMasahiro Yamada	depends on !COMPILE_TEST	# too noisy
4450dae776cSEmese Revfy	help
4460dae776cSEmese Revfy	  The complexity M of a function's control flow graph is defined as:
4470dae776cSEmese Revfy	   M = E - N + 2P
4480dae776cSEmese Revfy	  where
4490dae776cSEmese Revfy
4500dae776cSEmese Revfy	  E = the number of edges
4510dae776cSEmese Revfy	  N = the number of nodes
4520dae776cSEmese Revfy	  P = the number of connected components (exit nodes).
4530dae776cSEmese Revfy
454215e2aa6SKees Cook	  Enabling this plugin reports the complexity to stderr during the
455215e2aa6SKees Cook	  build. It mainly serves as a simple example of how to create a
456215e2aa6SKees Cook	  gcc plugin for the kernel.
457215e2aa6SKees Cook
458543c37cbSEmese Revfyconfig GCC_PLUGIN_SANCOV
459543c37cbSEmese Revfy	bool
460543c37cbSEmese Revfy	depends on GCC_PLUGINS
461543c37cbSEmese Revfy	help
462543c37cbSEmese Revfy	  This plugin inserts a __sanitizer_cov_trace_pc() call at the start of
463543c37cbSEmese Revfy	  basic blocks. It supports all gcc versions with plugin support (from
464543c37cbSEmese Revfy	  gcc-4.5 on). It is based on the commit "Add fuzzing coverage support"
465543c37cbSEmese Revfy	  by Dmitry Vyukov <dvyukov@google.com>.
466543c37cbSEmese Revfy
46738addce8SEmese Revfyconfig GCC_PLUGIN_LATENT_ENTROPY
46838addce8SEmese Revfy	bool "Generate some entropy during boot and runtime"
46938addce8SEmese Revfy	depends on GCC_PLUGINS
47038addce8SEmese Revfy	help
47138addce8SEmese Revfy	  By saying Y here the kernel will instrument some kernel code to
47238addce8SEmese Revfy	  extract some entropy from both original and artificially created
47338addce8SEmese Revfy	  program state.  This will help especially embedded systems where
47438addce8SEmese Revfy	  there is little 'natural' source of entropy normally.  The cost
47538addce8SEmese Revfy	  is some slowdown of the boot process (about 0.5%) and fork and
47638addce8SEmese Revfy	  irq processing.
47738addce8SEmese Revfy
47838addce8SEmese Revfy	  Note that entropy extracted this way is not cryptographically
47938addce8SEmese Revfy	  secure!
48038addce8SEmese Revfy
48138addce8SEmese Revfy	  This plugin was ported from grsecurity/PaX. More information at:
48238addce8SEmese Revfy	   * https://grsecurity.net/
48338addce8SEmese Revfy	   * https://pax.grsecurity.net/
48438addce8SEmese Revfy
485c61f13eaSKees Cookconfig GCC_PLUGIN_STRUCTLEAK
486c61f13eaSKees Cook	bool "Force initialization of variables containing userspace addresses"
487c61f13eaSKees Cook	depends on GCC_PLUGINS
488c9cf87eaSDmitry Vyukov	# Currently STRUCTLEAK inserts initialization out of live scope of
489c9cf87eaSDmitry Vyukov	# variables from KASAN point of view. This leads to KASAN false
490c9cf87eaSDmitry Vyukov	# positive reports. Prohibit this combination for now.
491c9cf87eaSDmitry Vyukov	depends on !KASAN_EXTRA
492c61f13eaSKees Cook	help
493f136e090SJean Delvare	  This plugin zero-initializes any structures containing a
494c61f13eaSKees Cook	  __user attribute. This can prevent some classes of information
495c61f13eaSKees Cook	  exposures.
496c61f13eaSKees Cook
497c61f13eaSKees Cook	  This plugin was ported from grsecurity/PaX. More information at:
498c61f13eaSKees Cook	   * https://grsecurity.net/
499c61f13eaSKees Cook	   * https://pax.grsecurity.net/
500c61f13eaSKees Cook
501f7dd2507SArd Biesheuvelconfig GCC_PLUGIN_STRUCTLEAK_BYREF_ALL
502f7dd2507SArd Biesheuvel	bool "Force initialize all struct type variables passed by reference"
503f7dd2507SArd Biesheuvel	depends on GCC_PLUGIN_STRUCTLEAK
504caa91ba5SMasahiro Yamada	depends on !COMPILE_TEST
505f7dd2507SArd Biesheuvel	help
506f7dd2507SArd Biesheuvel	  Zero initialize any struct type local variable that may be passed by
507f7dd2507SArd Biesheuvel	  reference without having been initialized.
508f7dd2507SArd Biesheuvel
509c61f13eaSKees Cookconfig GCC_PLUGIN_STRUCTLEAK_VERBOSE
510c61f13eaSKees Cook	bool "Report forcefully initialized variables"
511c61f13eaSKees Cook	depends on GCC_PLUGIN_STRUCTLEAK
5121658dceeSMasahiro Yamada	depends on !COMPILE_TEST	# too noisy
513c61f13eaSKees Cook	help
514c61f13eaSKees Cook	  This option will cause a warning to be printed each time the
515c61f13eaSKees Cook	  structleak plugin finds a variable it thinks needs to be
516c61f13eaSKees Cook	  initialized. Since not all existing initializers are detected
517c61f13eaSKees Cook	  by the plugin, this can produce false positive warnings.
518c61f13eaSKees Cook
519313dd1b6SKees Cookconfig GCC_PLUGIN_RANDSTRUCT
520313dd1b6SKees Cook	bool "Randomize layout of sensitive kernel structures"
521313dd1b6SKees Cook	depends on GCC_PLUGINS
522313dd1b6SKees Cook	select MODVERSIONS if MODULES
523313dd1b6SKees Cook	help
5249225331bSKees Cook	  If you say Y here, the layouts of structures that are entirely
5259225331bSKees Cook	  function pointers (and have not been manually annotated with
5269225331bSKees Cook	  __no_randomize_layout), or structures that have been explicitly
5279225331bSKees Cook	  marked with __randomize_layout, will be randomized at compile-time.
5289225331bSKees Cook	  This can introduce the requirement of an additional information
5299225331bSKees Cook	  exposure vulnerability for exploits targeting these structure
5309225331bSKees Cook	  types.
531313dd1b6SKees Cook
532313dd1b6SKees Cook	  Enabling this feature will introduce some performance impact,
533313dd1b6SKees Cook	  slightly increase memory usage, and prevent the use of forensic
534313dd1b6SKees Cook	  tools like Volatility against the system (unless the kernel
535313dd1b6SKees Cook	  source tree isn't cleaned after kernel installation).
536313dd1b6SKees Cook
537313dd1b6SKees Cook	  The seed used for compilation is located at
538313dd1b6SKees Cook	  scripts/gcc-plgins/randomize_layout_seed.h.  It remains after
539313dd1b6SKees Cook	  a make clean to allow for external modules to be compiled with
540313dd1b6SKees Cook	  the existing seed and will be removed by a make mrproper or
541313dd1b6SKees Cook	  make distclean.
542313dd1b6SKees Cook
543313dd1b6SKees Cook	  Note that the implementation requires gcc 4.7 or newer.
544313dd1b6SKees Cook
545313dd1b6SKees Cook	  This plugin was ported from grsecurity/PaX. More information at:
546313dd1b6SKees Cook	   * https://grsecurity.net/
547313dd1b6SKees Cook	   * https://pax.grsecurity.net/
548313dd1b6SKees Cook
549313dd1b6SKees Cookconfig GCC_PLUGIN_RANDSTRUCT_PERFORMANCE
550313dd1b6SKees Cook	bool "Use cacheline-aware structure randomization"
551313dd1b6SKees Cook	depends on GCC_PLUGIN_RANDSTRUCT
5521658dceeSMasahiro Yamada	depends on !COMPILE_TEST	# do not reduce test coverage
553313dd1b6SKees Cook	help
554313dd1b6SKees Cook	  If you say Y here, the RANDSTRUCT randomization will make a
555313dd1b6SKees Cook	  best effort at restricting randomization to cacheline-sized
556313dd1b6SKees Cook	  groups of elements.  It will further not randomize bitfields
557313dd1b6SKees Cook	  in structures.  This reduces the performance hit of RANDSTRUCT
558313dd1b6SKees Cook	  at the cost of weakened randomization.
559313dd1b6SKees Cook
560d148eac0SMasahiro Yamadaconfig HAVE_STACKPROTECTOR
56119952a92SKees Cook	bool
56219952a92SKees Cook	help
56319952a92SKees Cook	  An arch should select this symbol if:
56419952a92SKees Cook	  - it has implemented a stack canary (e.g. __stack_chk_guard)
56519952a92SKees Cook
5662a61f474SMasahiro Yamadaconfig CC_HAS_STACKPROTECTOR_NONE
5672a61f474SMasahiro Yamada	def_bool $(cc-option,-fno-stack-protector)
5682a61f474SMasahiro Yamada
569050e9baaSLinus Torvaldsconfig STACKPROTECTOR
5702a61f474SMasahiro Yamada	bool "Stack Protector buffer overflow detection"
571d148eac0SMasahiro Yamada	depends on HAVE_STACKPROTECTOR
5722a61f474SMasahiro Yamada	depends on $(cc-option,-fstack-protector)
5732a61f474SMasahiro Yamada	default y
5748779657dSKees Cook	help
5758779657dSKees Cook	  This option turns on the "stack-protector" GCC feature. This
57619952a92SKees Cook	  feature puts, at the beginning of functions, a canary value on
57719952a92SKees Cook	  the stack just before the return address, and validates
57819952a92SKees Cook	  the value just before actually returning.  Stack based buffer
57919952a92SKees Cook	  overflows (that need to overwrite this return address) now also
58019952a92SKees Cook	  overwrite the canary, which gets detected and the attack is then
58119952a92SKees Cook	  neutralized via a kernel panic.
58219952a92SKees Cook
5838779657dSKees Cook	  Functions will have the stack-protector canary logic added if they
5848779657dSKees Cook	  have an 8-byte or larger character array on the stack.
5858779657dSKees Cook
58619952a92SKees Cook	  This feature requires gcc version 4.2 or above, or a distribution
5878779657dSKees Cook	  gcc with the feature backported ("-fstack-protector").
5888779657dSKees Cook
5898779657dSKees Cook	  On an x86 "defconfig" build, this feature adds canary checks to
5908779657dSKees Cook	  about 3% of all kernel functions, which increases kernel code size
5918779657dSKees Cook	  by about 0.3%.
5928779657dSKees Cook
593050e9baaSLinus Torvaldsconfig STACKPROTECTOR_STRONG
5942a61f474SMasahiro Yamada	bool "Strong Stack Protector"
595050e9baaSLinus Torvalds	depends on STACKPROTECTOR
5962a61f474SMasahiro Yamada	depends on $(cc-option,-fstack-protector-strong)
5972a61f474SMasahiro Yamada	default y
5988779657dSKees Cook	help
5998779657dSKees Cook	  Functions will have the stack-protector canary logic added in any
6008779657dSKees Cook	  of the following conditions:
6018779657dSKees Cook
6028779657dSKees Cook	  - local variable's address used as part of the right hand side of an
6038779657dSKees Cook	    assignment or function argument
6048779657dSKees Cook	  - local variable is an array (or union containing an array),
6058779657dSKees Cook	    regardless of array type or length
6068779657dSKees Cook	  - uses register local variables
6078779657dSKees Cook
6088779657dSKees Cook	  This feature requires gcc version 4.9 or above, or a distribution
6098779657dSKees Cook	  gcc with the feature backported ("-fstack-protector-strong").
6108779657dSKees Cook
6118779657dSKees Cook	  On an x86 "defconfig" build, this feature adds canary checks to
6128779657dSKees Cook	  about 20% of all kernel functions, which increases the kernel code
6138779657dSKees Cook	  size by about 2%.
6148779657dSKees Cook
6150f60a8efSKees Cookconfig HAVE_ARCH_WITHIN_STACK_FRAMES
6160f60a8efSKees Cook	bool
6170f60a8efSKees Cook	help
6180f60a8efSKees Cook	  An architecture should select this if it can walk the kernel stack
6190f60a8efSKees Cook	  frames to determine if an object is part of either the arguments
6200f60a8efSKees Cook	  or local variables (i.e. that it excludes saved return addresses,
6210f60a8efSKees Cook	  and similar) by implementing an inline arch_within_stack_frames(),
6220f60a8efSKees Cook	  which is used by CONFIG_HARDENED_USERCOPY.
6230f60a8efSKees Cook
62491d1aa43SFrederic Weisbeckerconfig HAVE_CONTEXT_TRACKING
6252b1d5024SFrederic Weisbecker	bool
6262b1d5024SFrederic Weisbecker	help
62791d1aa43SFrederic Weisbecker	  Provide kernel/user boundaries probes necessary for subsystems
62891d1aa43SFrederic Weisbecker	  that need it, such as userspace RCU extended quiescent state.
62991d1aa43SFrederic Weisbecker	  Syscalls need to be wrapped inside user_exit()-user_enter() through
63091d1aa43SFrederic Weisbecker	  the slow path using TIF_NOHZ flag. Exceptions handlers must be
63191d1aa43SFrederic Weisbecker	  wrapped as well. Irqs are already protected inside
63291d1aa43SFrederic Weisbecker	  rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on
63391d1aa43SFrederic Weisbecker	  irq exit still need to be protected.
6342b1d5024SFrederic Weisbecker
635b952741cSFrederic Weisbeckerconfig HAVE_VIRT_CPU_ACCOUNTING
636b952741cSFrederic Weisbecker	bool
637b952741cSFrederic Weisbecker
63840565b5aSStanislaw Gruszkaconfig ARCH_HAS_SCALED_CPUTIME
63940565b5aSStanislaw Gruszka	bool
64040565b5aSStanislaw Gruszka
641554b0004SKevin Hilmanconfig HAVE_VIRT_CPU_ACCOUNTING_GEN
642554b0004SKevin Hilman	bool
643554b0004SKevin Hilman	default y if 64BIT
644554b0004SKevin Hilman	help
645554b0004SKevin Hilman	  With VIRT_CPU_ACCOUNTING_GEN, cputime_t becomes 64-bit.
646554b0004SKevin Hilman	  Before enabling this option, arch code must be audited
647554b0004SKevin Hilman	  to ensure there are no races in concurrent read/write of
648554b0004SKevin Hilman	  cputime_t. For example, reading/writing 64-bit cputime_t on
649554b0004SKevin Hilman	  some 32-bit arches may require multiple accesses, so proper
650554b0004SKevin Hilman	  locking is needed to protect against concurrent accesses.
651554b0004SKevin Hilman
652554b0004SKevin Hilman
653fdf9c356SFrederic Weisbeckerconfig HAVE_IRQ_TIME_ACCOUNTING
654fdf9c356SFrederic Weisbecker	bool
655fdf9c356SFrederic Weisbecker	help
656fdf9c356SFrederic Weisbecker	  Archs need to ensure they use a high enough resolution clock to
657fdf9c356SFrederic Weisbecker	  support irq time accounting and then call enable_sched_clock_irqtime().
658fdf9c356SFrederic Weisbecker
65915626062SGerald Schaeferconfig HAVE_ARCH_TRANSPARENT_HUGEPAGE
66015626062SGerald Schaefer	bool
66115626062SGerald Schaefer
662a00cc7d9SMatthew Wilcoxconfig HAVE_ARCH_TRANSPARENT_HUGEPAGE_PUD
663a00cc7d9SMatthew Wilcox	bool
664a00cc7d9SMatthew Wilcox
6650ddab1d2SToshi Kaniconfig HAVE_ARCH_HUGE_VMAP
6660ddab1d2SToshi Kani	bool
6670ddab1d2SToshi Kani
6680f8975ecSPavel Emelyanovconfig HAVE_ARCH_SOFT_DIRTY
6690f8975ecSPavel Emelyanov	bool
6700f8975ecSPavel Emelyanov
671786d35d4SDavid Howellsconfig HAVE_MOD_ARCH_SPECIFIC
672786d35d4SDavid Howells	bool
673786d35d4SDavid Howells	help
674786d35d4SDavid Howells	  The arch uses struct mod_arch_specific to store data.  Many arches
675786d35d4SDavid Howells	  just need a simple module loader without arch specific data - those
676786d35d4SDavid Howells	  should not enable this.
677786d35d4SDavid Howells
678786d35d4SDavid Howellsconfig MODULES_USE_ELF_RELA
679786d35d4SDavid Howells	bool
680786d35d4SDavid Howells	help
681786d35d4SDavid Howells	  Modules only use ELF RELA relocations.  Modules with ELF REL
682786d35d4SDavid Howells	  relocations will give an error.
683786d35d4SDavid Howells
684786d35d4SDavid Howellsconfig MODULES_USE_ELF_REL
685786d35d4SDavid Howells	bool
686786d35d4SDavid Howells	help
687786d35d4SDavid Howells	  Modules only use ELF REL relocations.  Modules with ELF RELA
688786d35d4SDavid Howells	  relocations will give an error.
689786d35d4SDavid Howells
690cc1f0274SFrederic Weisbeckerconfig HAVE_IRQ_EXIT_ON_IRQ_STACK
691cc1f0274SFrederic Weisbecker	bool
692cc1f0274SFrederic Weisbecker	help
693cc1f0274SFrederic Weisbecker	  Architecture doesn't only execute the irq handler on the irq stack
694cc1f0274SFrederic Weisbecker	  but also irq_exit(). This way we can process softirqs on this irq
695cc1f0274SFrederic Weisbecker	  stack instead of switching to a new one when we call __do_softirq()
696cc1f0274SFrederic Weisbecker	  in the end of an hardirq.
697cc1f0274SFrederic Weisbecker	  This spares a stack switch and improves cache usage on softirq
698cc1f0274SFrederic Weisbecker	  processing.
699cc1f0274SFrederic Weisbecker
700235a8f02SKirill A. Shutemovconfig PGTABLE_LEVELS
701235a8f02SKirill A. Shutemov	int
702235a8f02SKirill A. Shutemov	default 2
703235a8f02SKirill A. Shutemov
7042b68f6caSKees Cookconfig ARCH_HAS_ELF_RANDOMIZE
7052b68f6caSKees Cook	bool
7062b68f6caSKees Cook	help
7072b68f6caSKees Cook	  An architecture supports choosing randomized locations for
7082b68f6caSKees Cook	  stack, mmap, brk, and ET_DYN. Defined functions:
7092b68f6caSKees Cook	  - arch_mmap_rnd()
710204db6edSKees Cook	  - arch_randomize_brk()
7112b68f6caSKees Cook
712d07e2259SDaniel Cashmanconfig HAVE_ARCH_MMAP_RND_BITS
713d07e2259SDaniel Cashman	bool
714d07e2259SDaniel Cashman	help
715d07e2259SDaniel Cashman	  An arch should select this symbol if it supports setting a variable
716d07e2259SDaniel Cashman	  number of bits for use in establishing the base address for mmap
717d07e2259SDaniel Cashman	  allocations, has MMU enabled and provides values for both:
718d07e2259SDaniel Cashman	  - ARCH_MMAP_RND_BITS_MIN
719d07e2259SDaniel Cashman	  - ARCH_MMAP_RND_BITS_MAX
720d07e2259SDaniel Cashman
7215f56a5dfSJiri Slabyconfig HAVE_EXIT_THREAD
7225f56a5dfSJiri Slaby	bool
7235f56a5dfSJiri Slaby	help
7245f56a5dfSJiri Slaby	  An architecture implements exit_thread.
7255f56a5dfSJiri Slaby
726d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_MIN
727d07e2259SDaniel Cashman	int
728d07e2259SDaniel Cashman
729d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_MAX
730d07e2259SDaniel Cashman	int
731d07e2259SDaniel Cashman
732d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS_DEFAULT
733d07e2259SDaniel Cashman	int
734d07e2259SDaniel Cashman
735d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_BITS
736d07e2259SDaniel Cashman	int "Number of bits to use for ASLR of mmap base address" if EXPERT
737d07e2259SDaniel Cashman	range ARCH_MMAP_RND_BITS_MIN ARCH_MMAP_RND_BITS_MAX
738d07e2259SDaniel Cashman	default ARCH_MMAP_RND_BITS_DEFAULT if ARCH_MMAP_RND_BITS_DEFAULT
739d07e2259SDaniel Cashman	default ARCH_MMAP_RND_BITS_MIN
740d07e2259SDaniel Cashman	depends on HAVE_ARCH_MMAP_RND_BITS
741d07e2259SDaniel Cashman	help
742d07e2259SDaniel Cashman	  This value can be used to select the number of bits to use to
743d07e2259SDaniel Cashman	  determine the random offset to the base address of vma regions
744d07e2259SDaniel Cashman	  resulting from mmap allocations. This value will be bounded
745d07e2259SDaniel Cashman	  by the architecture's minimum and maximum supported values.
746d07e2259SDaniel Cashman
747d07e2259SDaniel Cashman	  This value can be changed after boot using the
748d07e2259SDaniel Cashman	  /proc/sys/vm/mmap_rnd_bits tunable
749d07e2259SDaniel Cashman
750d07e2259SDaniel Cashmanconfig HAVE_ARCH_MMAP_RND_COMPAT_BITS
751d07e2259SDaniel Cashman	bool
752d07e2259SDaniel Cashman	help
753d07e2259SDaniel Cashman	  An arch should select this symbol if it supports running applications
754d07e2259SDaniel Cashman	  in compatibility mode, supports setting a variable number of bits for
755d07e2259SDaniel Cashman	  use in establishing the base address for mmap allocations, has MMU
756d07e2259SDaniel Cashman	  enabled and provides values for both:
757d07e2259SDaniel Cashman	  - ARCH_MMAP_RND_COMPAT_BITS_MIN
758d07e2259SDaniel Cashman	  - ARCH_MMAP_RND_COMPAT_BITS_MAX
759d07e2259SDaniel Cashman
760d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_MIN
761d07e2259SDaniel Cashman	int
762d07e2259SDaniel Cashman
763d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_MAX
764d07e2259SDaniel Cashman	int
765d07e2259SDaniel Cashman
766d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
767d07e2259SDaniel Cashman	int
768d07e2259SDaniel Cashman
769d07e2259SDaniel Cashmanconfig ARCH_MMAP_RND_COMPAT_BITS
770d07e2259SDaniel Cashman	int "Number of bits to use for ASLR of mmap base address for compatible applications" if EXPERT
771d07e2259SDaniel Cashman	range ARCH_MMAP_RND_COMPAT_BITS_MIN ARCH_MMAP_RND_COMPAT_BITS_MAX
772d07e2259SDaniel Cashman	default ARCH_MMAP_RND_COMPAT_BITS_DEFAULT if ARCH_MMAP_RND_COMPAT_BITS_DEFAULT
773d07e2259SDaniel Cashman	default ARCH_MMAP_RND_COMPAT_BITS_MIN
774d07e2259SDaniel Cashman	depends on HAVE_ARCH_MMAP_RND_COMPAT_BITS
775d07e2259SDaniel Cashman	help
776d07e2259SDaniel Cashman	  This value can be used to select the number of bits to use to
777d07e2259SDaniel Cashman	  determine the random offset to the base address of vma regions
778d07e2259SDaniel Cashman	  resulting from mmap allocations for compatible applications This
779d07e2259SDaniel Cashman	  value will be bounded by the architecture's minimum and maximum
780d07e2259SDaniel Cashman	  supported values.
781d07e2259SDaniel Cashman
782d07e2259SDaniel Cashman	  This value can be changed after boot using the
783d07e2259SDaniel Cashman	  /proc/sys/vm/mmap_rnd_compat_bits tunable
784d07e2259SDaniel Cashman
7851b028f78SDmitry Safonovconfig HAVE_ARCH_COMPAT_MMAP_BASES
7861b028f78SDmitry Safonov	bool
7871b028f78SDmitry Safonov	help
7881b028f78SDmitry Safonov	  This allows 64bit applications to invoke 32-bit mmap() syscall
7891b028f78SDmitry Safonov	  and vice-versa 32-bit applications to call 64-bit mmap().
7901b028f78SDmitry Safonov	  Required for applications doing different bitness syscalls.
7911b028f78SDmitry Safonov
7923033f14aSJosh Triplettconfig HAVE_COPY_THREAD_TLS
7933033f14aSJosh Triplett	bool
7943033f14aSJosh Triplett	help
7953033f14aSJosh Triplett	  Architecture provides copy_thread_tls to accept tls argument via
7963033f14aSJosh Triplett	  normal C parameter passing, rather than extracting the syscall
7973033f14aSJosh Triplett	  argument from pt_regs.
7983033f14aSJosh Triplett
799b9ab5ebbSJosh Poimboeufconfig HAVE_STACK_VALIDATION
800b9ab5ebbSJosh Poimboeuf	bool
801b9ab5ebbSJosh Poimboeuf	help
802b9ab5ebbSJosh Poimboeuf	  Architecture supports the 'objtool check' host tool command, which
803b9ab5ebbSJosh Poimboeuf	  performs compile-time stack metadata validation.
804b9ab5ebbSJosh Poimboeuf
805af085d90SJosh Poimboeufconfig HAVE_RELIABLE_STACKTRACE
806af085d90SJosh Poimboeuf	bool
807af085d90SJosh Poimboeuf	help
808af085d90SJosh Poimboeuf	  Architecture has a save_stack_trace_tsk_reliable() function which
809af085d90SJosh Poimboeuf	  only returns a stack trace if it can guarantee the trace is reliable.
810af085d90SJosh Poimboeuf
811468a9428SGeorge Spelvinconfig HAVE_ARCH_HASH
812468a9428SGeorge Spelvin	bool
813468a9428SGeorge Spelvin	default n
814468a9428SGeorge Spelvin	help
815468a9428SGeorge Spelvin	  If this is set, the architecture provides an <asm/hash.h>
816468a9428SGeorge Spelvin	  file which provides platform-specific implementations of some
817468a9428SGeorge Spelvin	  functions in <linux/hash.h> or fs/namei.c.
818468a9428SGeorge Spelvin
8193a495511SWilliam Breathitt Grayconfig ISA_BUS_API
8203a495511SWilliam Breathitt Gray	def_bool ISA
8213a495511SWilliam Breathitt Gray
822d2125043SAl Viro#
823d2125043SAl Viro# ABI hall of shame
824d2125043SAl Viro#
825d2125043SAl Viroconfig CLONE_BACKWARDS
826d2125043SAl Viro	bool
827d2125043SAl Viro	help
828d2125043SAl Viro	  Architecture has tls passed as the 4th argument of clone(2),
829d2125043SAl Viro	  not the 5th one.
830d2125043SAl Viro
831d2125043SAl Viroconfig CLONE_BACKWARDS2
832d2125043SAl Viro	bool
833d2125043SAl Viro	help
834d2125043SAl Viro	  Architecture has the first two arguments of clone(2) swapped.
835d2125043SAl Viro
836dfa9771aSMichal Simekconfig CLONE_BACKWARDS3
837dfa9771aSMichal Simek	bool
838dfa9771aSMichal Simek	help
839dfa9771aSMichal Simek	  Architecture has tls passed as the 3rd argument of clone(2),
840dfa9771aSMichal Simek	  not the 5th one.
841dfa9771aSMichal Simek
842eaca6eaeSAl Viroconfig ODD_RT_SIGACTION
843eaca6eaeSAl Viro	bool
844eaca6eaeSAl Viro	help
845eaca6eaeSAl Viro	  Architecture has unusual rt_sigaction(2) arguments
846eaca6eaeSAl Viro
8470a0e8cdfSAl Viroconfig OLD_SIGSUSPEND
8480a0e8cdfSAl Viro	bool
8490a0e8cdfSAl Viro	help
8500a0e8cdfSAl Viro	  Architecture has old sigsuspend(2) syscall, of one-argument variety
8510a0e8cdfSAl Viro
8520a0e8cdfSAl Viroconfig OLD_SIGSUSPEND3
8530a0e8cdfSAl Viro	bool
8540a0e8cdfSAl Viro	help
8550a0e8cdfSAl Viro	  Even weirder antique ABI - three-argument sigsuspend(2)
8560a0e8cdfSAl Viro
857495dfbf7SAl Viroconfig OLD_SIGACTION
858495dfbf7SAl Viro	bool
859495dfbf7SAl Viro	help
860495dfbf7SAl Viro	  Architecture has old sigaction(2) syscall.  Nope, not the same
861495dfbf7SAl Viro	  as OLD_SIGSUSPEND | OLD_SIGSUSPEND3 - alpha has sigsuspend(2),
862495dfbf7SAl Viro	  but fairly different variant of sigaction(2), thanks to OSF/1
863495dfbf7SAl Viro	  compatibility...
864495dfbf7SAl Viro
865495dfbf7SAl Viroconfig COMPAT_OLD_SIGACTION
866495dfbf7SAl Viro	bool
867495dfbf7SAl Viro
868d4703ddaSDeepa Dinamaniconfig 64BIT_TIME
869d4703ddaSDeepa Dinamani	def_bool ARCH_HAS_64BIT_TIME
870d4703ddaSDeepa Dinamani	help
871d4703ddaSDeepa Dinamani	  This should be selected by all architectures that need to support
872d4703ddaSDeepa Dinamani	  new system calls with a 64-bit time_t. This is relevant on all 32-bit
873d4703ddaSDeepa Dinamani	  architectures, and 64-bit architectures as part of compat syscall
874d4703ddaSDeepa Dinamani	  handling.
875d4703ddaSDeepa Dinamani
87617435e5fSDeepa Dinamaniconfig COMPAT_32BIT_TIME
87717435e5fSDeepa Dinamani	def_bool (!64BIT && 64BIT_TIME) || COMPAT
87817435e5fSDeepa Dinamani	help
87917435e5fSDeepa Dinamani	  This enables 32 bit time_t support in addition to 64 bit time_t support.
88017435e5fSDeepa Dinamani	  This is relevant on all 32-bit architectures, and 64-bit architectures
88117435e5fSDeepa Dinamani	  as part of compat syscall handling.
88217435e5fSDeepa Dinamani
8830d4a619bSChristoph Hellwigconfig ARCH_NO_COHERENT_DMA_MMAP
8840d4a619bSChristoph Hellwig	bool
8850d4a619bSChristoph Hellwig
88687a4c375SChristoph Hellwigconfig ARCH_NO_PREEMPT
88787a4c375SChristoph Hellwig	bool
88887a4c375SChristoph Hellwig
889fff7fb0bSZhaoxiu Zengconfig CPU_NO_EFFICIENT_FFS
890fff7fb0bSZhaoxiu Zeng	def_bool n
891fff7fb0bSZhaoxiu Zeng
892ba14a194SAndy Lutomirskiconfig HAVE_ARCH_VMAP_STACK
893ba14a194SAndy Lutomirski	def_bool n
894ba14a194SAndy Lutomirski	help
895ba14a194SAndy Lutomirski	  An arch should select this symbol if it can support kernel stacks
896ba14a194SAndy Lutomirski	  in vmalloc space.  This means:
897ba14a194SAndy Lutomirski
898ba14a194SAndy Lutomirski	  - vmalloc space must be large enough to hold many kernel stacks.
899ba14a194SAndy Lutomirski	    This may rule out many 32-bit architectures.
900ba14a194SAndy Lutomirski
901ba14a194SAndy Lutomirski	  - Stacks in vmalloc space need to work reliably.  For example, if
902ba14a194SAndy Lutomirski	    vmap page tables are created on demand, either this mechanism
903ba14a194SAndy Lutomirski	    needs to work while the stack points to a virtual address with
904ba14a194SAndy Lutomirski	    unpopulated page tables or arch code (switch_to() and switch_mm(),
905ba14a194SAndy Lutomirski	    most likely) needs to ensure that the stack's page table entries
906ba14a194SAndy Lutomirski	    are populated before running on a possibly unpopulated stack.
907ba14a194SAndy Lutomirski
908ba14a194SAndy Lutomirski	  - If the stack overflows into a guard page, something reasonable
909ba14a194SAndy Lutomirski	    should happen.  The definition of "reasonable" is flexible, but
910ba14a194SAndy Lutomirski	    instantly rebooting without logging anything would be unfriendly.
911ba14a194SAndy Lutomirski
912ba14a194SAndy Lutomirskiconfig VMAP_STACK
913ba14a194SAndy Lutomirski	default y
914ba14a194SAndy Lutomirski	bool "Use a virtually-mapped stack"
915ba14a194SAndy Lutomirski	depends on HAVE_ARCH_VMAP_STACK && !KASAN
916ba14a194SAndy Lutomirski	---help---
917ba14a194SAndy Lutomirski	  Enable this if you want the use virtually-mapped kernel stacks
918ba14a194SAndy Lutomirski	  with guard pages.  This causes kernel stack overflows to be
919ba14a194SAndy Lutomirski	  caught immediately rather than causing difficult-to-diagnose
920ba14a194SAndy Lutomirski	  corruption.
921ba14a194SAndy Lutomirski
922ba14a194SAndy Lutomirski	  This is presently incompatible with KASAN because KASAN expects
923ba14a194SAndy Lutomirski	  the stack to map directly to the KASAN shadow map using a formula
924ba14a194SAndy Lutomirski	  that is incorrect if the stack is in vmalloc space.
925ba14a194SAndy Lutomirski
926ad21fc4fSLaura Abbottconfig ARCH_OPTIONAL_KERNEL_RWX
927ad21fc4fSLaura Abbott	def_bool n
928ad21fc4fSLaura Abbott
929ad21fc4fSLaura Abbottconfig ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
930ad21fc4fSLaura Abbott	def_bool n
931ad21fc4fSLaura Abbott
932ad21fc4fSLaura Abbottconfig ARCH_HAS_STRICT_KERNEL_RWX
933ad21fc4fSLaura Abbott	def_bool n
934ad21fc4fSLaura Abbott
9350f5bf6d0SLaura Abbottconfig STRICT_KERNEL_RWX
936ad21fc4fSLaura Abbott	bool "Make kernel text and rodata read-only" if ARCH_OPTIONAL_KERNEL_RWX
937ad21fc4fSLaura Abbott	depends on ARCH_HAS_STRICT_KERNEL_RWX
938ad21fc4fSLaura Abbott	default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
939ad21fc4fSLaura Abbott	help
940ad21fc4fSLaura Abbott	  If this is set, kernel text and rodata memory will be made read-only,
941ad21fc4fSLaura Abbott	  and non-text memory will be made non-executable. This provides
942ad21fc4fSLaura Abbott	  protection against certain security exploits (e.g. executing the heap
943ad21fc4fSLaura Abbott	  or modifying text)
944ad21fc4fSLaura Abbott
945ad21fc4fSLaura Abbott	  These features are considered standard security practice these days.
946ad21fc4fSLaura Abbott	  You should say Y here in almost all cases.
947ad21fc4fSLaura Abbott
948ad21fc4fSLaura Abbottconfig ARCH_HAS_STRICT_MODULE_RWX
949ad21fc4fSLaura Abbott	def_bool n
950ad21fc4fSLaura Abbott
9510f5bf6d0SLaura Abbottconfig STRICT_MODULE_RWX
952ad21fc4fSLaura Abbott	bool "Set loadable kernel module data as NX and text as RO" if ARCH_OPTIONAL_KERNEL_RWX
953ad21fc4fSLaura Abbott	depends on ARCH_HAS_STRICT_MODULE_RWX && MODULES
954ad21fc4fSLaura Abbott	default !ARCH_OPTIONAL_KERNEL_RWX || ARCH_OPTIONAL_KERNEL_RWX_DEFAULT
955ad21fc4fSLaura Abbott	help
956ad21fc4fSLaura Abbott	  If this is set, module text and rodata memory will be made read-only,
957ad21fc4fSLaura Abbott	  and non-text memory will be made non-executable. This provides
958ad21fc4fSLaura Abbott	  protection against certain security exploits (e.g. writing to text)
959ad21fc4fSLaura Abbott
960ea8c64acSChristoph Hellwig# select if the architecture provides an asm/dma-direct.h header
961ea8c64acSChristoph Hellwigconfig ARCH_HAS_PHYS_TO_DMA
962ea8c64acSChristoph Hellwig	bool
963ea8c64acSChristoph Hellwig
9647a46ec0eSKees Cookconfig ARCH_HAS_REFCOUNT
9657a46ec0eSKees Cook	bool
9667a46ec0eSKees Cook	help
9677a46ec0eSKees Cook	  An architecture selects this when it has implemented refcount_t
9687a46ec0eSKees Cook	  using open coded assembly primitives that provide an optimized
9697a46ec0eSKees Cook	  refcount_t implementation, possibly at the expense of some full
9707a46ec0eSKees Cook	  refcount state checks of CONFIG_REFCOUNT_FULL=y.
9717a46ec0eSKees Cook
9727a46ec0eSKees Cook	  The refcount overflow check behavior, however, must be retained.
9737a46ec0eSKees Cook	  Catching overflows is the primary security concern for protecting
9747a46ec0eSKees Cook	  against bugs in reference counts.
9757a46ec0eSKees Cook
976fd25d19fSKees Cookconfig REFCOUNT_FULL
977fd25d19fSKees Cook	bool "Perform full reference count validation at the expense of speed"
978fd25d19fSKees Cook	help
979fd25d19fSKees Cook	  Enabling this switches the refcounting infrastructure from a fast
980fd25d19fSKees Cook	  unchecked atomic_t implementation to a fully state checked
981fd25d19fSKees Cook	  implementation, which can be (slightly) slower but provides protections
982fd25d19fSKees Cook	  against various use-after-free conditions that can be used in
983fd25d19fSKees Cook	  security flaw exploits.
984fd25d19fSKees Cook
9852521f2c2SPeter Oberparleitersource "kernel/gcov/Kconfig"
98622471e13SRandy Dunlap
98722471e13SRandy Dunlapendmenu
988