1# 2# General architecture dependent options 3# 4 5config OPROFILE 6 tristate "OProfile system profiling" 7 depends on PROFILING 8 depends on HAVE_OPROFILE 9 select RING_BUFFER 10 select RING_BUFFER_ALLOW_SWAP 11 help 12 OProfile is a profiling system capable of profiling the 13 whole system, include the kernel, kernel modules, libraries, 14 and applications. 15 16 If unsure, say N. 17 18config OPROFILE_EVENT_MULTIPLEX 19 bool "OProfile multiplexing support (EXPERIMENTAL)" 20 default n 21 depends on OPROFILE && X86 22 help 23 The number of hardware counters is limited. The multiplexing 24 feature enables OProfile to gather more events than counters 25 are provided by the hardware. This is realized by switching 26 between events at an user specified time interval. 27 28 If unsure, say N. 29 30config HAVE_OPROFILE 31 bool 32 33config OPROFILE_NMI_TIMER 34 def_bool y 35 depends on PERF_EVENTS && HAVE_PERF_EVENTS_NMI 36 37config KPROBES 38 bool "Kprobes" 39 depends on MODULES 40 depends on HAVE_KPROBES 41 select KALLSYMS 42 help 43 Kprobes allows you to trap at almost any kernel address and 44 execute a callback function. register_kprobe() establishes 45 a probepoint and specifies the callback. Kprobes is useful 46 for kernel debugging, non-intrusive instrumentation and testing. 47 If in doubt, say "N". 48 49config JUMP_LABEL 50 bool "Optimize very unlikely/likely branches" 51 depends on HAVE_ARCH_JUMP_LABEL 52 help 53 This option enables a transparent branch optimization that 54 makes certain almost-always-true or almost-always-false branch 55 conditions even cheaper to execute within the kernel. 56 57 Certain performance-sensitive kernel code, such as trace points, 58 scheduler functionality, networking code and KVM have such 59 branches and include support for this optimization technique. 60 61 If it is detected that the compiler has support for "asm goto", 62 the kernel will compile such branches with just a nop 63 instruction. When the condition flag is toggled to true, the 64 nop will be converted to a jump instruction to execute the 65 conditional block of instructions. 66 67 This technique lowers overhead and stress on the branch prediction 68 of the processor and generally makes the kernel faster. The update 69 of the condition is slower, but those are always very rare. 70 71 ( On 32-bit x86, the necessary options added to the compiler 72 flags may increase the size of the kernel slightly. ) 73 74config OPTPROBES 75 def_bool y 76 depends on KPROBES && HAVE_OPTPROBES 77 depends on !PREEMPT 78 79config UPROBES 80 bool "Transparent user-space probes (EXPERIMENTAL)" 81 depends on UPROBE_EVENT && PERF_EVENTS 82 default n 83 select PERCPU_RWSEM 84 help 85 Uprobes is the user-space counterpart to kprobes: they 86 enable instrumentation applications (such as 'perf probe') 87 to establish unintrusive probes in user-space binaries and 88 libraries, by executing handler functions when the probes 89 are hit by user-space applications. 90 91 ( These probes come in the form of single-byte breakpoints, 92 managed by the kernel and kept transparent to the probed 93 application. ) 94 95 If in doubt, say "N". 96 97config HAVE_EFFICIENT_UNALIGNED_ACCESS 98 bool 99 help 100 Some architectures are unable to perform unaligned accesses 101 without the use of get_unaligned/put_unaligned. Others are 102 unable to perform such accesses efficiently (e.g. trap on 103 unaligned access and require fixing it up in the exception 104 handler.) 105 106 This symbol should be selected by an architecture if it can 107 perform unaligned accesses efficiently to allow different 108 code paths to be selected for these cases. Some network 109 drivers, for example, could opt to not fix up alignment 110 problems with received packets if doing so would not help 111 much. 112 113 See Documentation/unaligned-memory-access.txt for more 114 information on the topic of unaligned memory accesses. 115 116config ARCH_USE_BUILTIN_BSWAP 117 bool 118 help 119 Modern versions of GCC (since 4.4) have builtin functions 120 for handling byte-swapping. Using these, instead of the old 121 inline assembler that the architecture code provides in the 122 __arch_bswapXX() macros, allows the compiler to see what's 123 happening and offers more opportunity for optimisation. In 124 particular, the compiler will be able to combine the byteswap 125 with a nearby load or store and use load-and-swap or 126 store-and-swap instructions if the architecture has them. It 127 should almost *never* result in code which is worse than the 128 hand-coded assembler in <asm/swab.h>. But just in case it 129 does, the use of the builtins is optional. 130 131 Any architecture with load-and-swap or store-and-swap 132 instructions should set this. And it shouldn't hurt to set it 133 on architectures that don't have such instructions. 134 135config HAVE_SYSCALL_WRAPPERS 136 bool 137 138config KRETPROBES 139 def_bool y 140 depends on KPROBES && HAVE_KRETPROBES 141 142config USER_RETURN_NOTIFIER 143 bool 144 depends on HAVE_USER_RETURN_NOTIFIER 145 help 146 Provide a kernel-internal notification when a cpu is about to 147 switch to user mode. 148 149config HAVE_IOREMAP_PROT 150 bool 151 152config HAVE_KPROBES 153 bool 154 155config HAVE_KRETPROBES 156 bool 157 158config HAVE_OPTPROBES 159 bool 160 161config HAVE_NMI_WATCHDOG 162 bool 163# 164# An arch should select this if it provides all these things: 165# 166# task_pt_regs() in asm/processor.h or asm/ptrace.h 167# arch_has_single_step() if there is hardware single-step support 168# arch_has_block_step() if there is hardware block-step support 169# asm/syscall.h supplying asm-generic/syscall.h interface 170# linux/regset.h user_regset interfaces 171# CORE_DUMP_USE_REGSET #define'd in linux/elf.h 172# TIF_SYSCALL_TRACE calls tracehook_report_syscall_{entry,exit} 173# TIF_NOTIFY_RESUME calls tracehook_notify_resume() 174# signal delivery calls tracehook_signal_handler() 175# 176config HAVE_ARCH_TRACEHOOK 177 bool 178 179config HAVE_DMA_ATTRS 180 bool 181 182config HAVE_DMA_CONTIGUOUS 183 bool 184 185config USE_GENERIC_SMP_HELPERS 186 bool 187 188config GENERIC_SMP_IDLE_THREAD 189 bool 190 191# Select if arch init_task initializer is different to init/init_task.c 192config ARCH_INIT_TASK 193 bool 194 195# Select if arch has its private alloc_task_struct() function 196config ARCH_TASK_STRUCT_ALLOCATOR 197 bool 198 199# Select if arch has its private alloc_thread_info() function 200config ARCH_THREAD_INFO_ALLOCATOR 201 bool 202 203config HAVE_REGS_AND_STACK_ACCESS_API 204 bool 205 help 206 This symbol should be selected by an architecure if it supports 207 the API needed to access registers and stack entries from pt_regs, 208 declared in asm/ptrace.h 209 For example the kprobes-based event tracer needs this API. 210 211config HAVE_CLK 212 bool 213 help 214 The <linux/clk.h> calls support software clock gating and 215 thus are a key power management tool on many systems. 216 217config HAVE_DMA_API_DEBUG 218 bool 219 220config HAVE_HW_BREAKPOINT 221 bool 222 depends on PERF_EVENTS 223 224config HAVE_MIXED_BREAKPOINTS_REGS 225 bool 226 depends on HAVE_HW_BREAKPOINT 227 help 228 Depending on the arch implementation of hardware breakpoints, 229 some of them have separate registers for data and instruction 230 breakpoints addresses, others have mixed registers to store 231 them but define the access type in a control register. 232 Select this option if your arch implements breakpoints under the 233 latter fashion. 234 235config HAVE_USER_RETURN_NOTIFIER 236 bool 237 238config HAVE_PERF_EVENTS_NMI 239 bool 240 help 241 System hardware can generate an NMI using the perf event 242 subsystem. Also has support for calculating CPU cycle events 243 to determine how many clock cycles in a given period. 244 245config HAVE_PERF_REGS 246 bool 247 help 248 Support selective register dumps for perf events. This includes 249 bit-mapping of each registers and a unique architecture id. 250 251config HAVE_PERF_USER_STACK_DUMP 252 bool 253 help 254 Support user stack dumps for perf event samples. This needs 255 access to the user stack pointer which is not unified across 256 architectures. 257 258config HAVE_ARCH_JUMP_LABEL 259 bool 260 261config HAVE_ARCH_MUTEX_CPU_RELAX 262 bool 263 264config HAVE_RCU_TABLE_FREE 265 bool 266 267config ARCH_HAVE_NMI_SAFE_CMPXCHG 268 bool 269 270config HAVE_ALIGNED_STRUCT_PAGE 271 bool 272 help 273 This makes sure that struct pages are double word aligned and that 274 e.g. the SLUB allocator can perform double word atomic operations 275 on a struct page for better performance. However selecting this 276 might increase the size of a struct page by a word. 277 278config HAVE_CMPXCHG_LOCAL 279 bool 280 281config HAVE_CMPXCHG_DOUBLE 282 bool 283 284config ARCH_WANT_IPC_PARSE_VERSION 285 bool 286 287config ARCH_WANT_COMPAT_IPC_PARSE_VERSION 288 bool 289 290config ARCH_WANT_OLD_COMPAT_IPC 291 select ARCH_WANT_COMPAT_IPC_PARSE_VERSION 292 bool 293 294config GENERIC_KERNEL_THREAD 295 bool 296 297config GENERIC_KERNEL_EXECVE 298 bool 299 300config HAVE_ARCH_SECCOMP_FILTER 301 bool 302 help 303 An arch should select this symbol if it provides all of these things: 304 - syscall_get_arch() 305 - syscall_get_arguments() 306 - syscall_rollback() 307 - syscall_set_return_value() 308 - SIGSYS siginfo_t support 309 - secure_computing is called from a ptrace_event()-safe context 310 - secure_computing return value is checked and a return value of -1 311 results in the system call being skipped immediately. 312 313config SECCOMP_FILTER 314 def_bool y 315 depends on HAVE_ARCH_SECCOMP_FILTER && SECCOMP && NET 316 help 317 Enable tasks to build secure computing environments defined 318 in terms of Berkeley Packet Filter programs which implement 319 task-defined system call filtering polices. 320 321 See Documentation/prctl/seccomp_filter.txt for details. 322 323config HAVE_CONTEXT_TRACKING 324 bool 325 help 326 Provide kernel/user boundaries probes necessary for subsystems 327 that need it, such as userspace RCU extended quiescent state. 328 Syscalls need to be wrapped inside user_exit()-user_enter() through 329 the slow path using TIF_NOHZ flag. Exceptions handlers must be 330 wrapped as well. Irqs are already protected inside 331 rcu_irq_enter/rcu_irq_exit() but preemption or signal handling on 332 irq exit still need to be protected. 333 334config HAVE_VIRT_CPU_ACCOUNTING 335 bool 336 337config HAVE_IRQ_TIME_ACCOUNTING 338 bool 339 help 340 Archs need to ensure they use a high enough resolution clock to 341 support irq time accounting and then call enable_sched_clock_irqtime(). 342 343config HAVE_ARCH_TRANSPARENT_HUGEPAGE 344 bool 345 346config HAVE_MOD_ARCH_SPECIFIC 347 bool 348 help 349 The arch uses struct mod_arch_specific to store data. Many arches 350 just need a simple module loader without arch specific data - those 351 should not enable this. 352 353config MODULES_USE_ELF_RELA 354 bool 355 help 356 Modules only use ELF RELA relocations. Modules with ELF REL 357 relocations will give an error. 358 359config MODULES_USE_ELF_REL 360 bool 361 help 362 Modules only use ELF REL relocations. Modules with ELF RELA 363 relocations will give an error. 364 365# 366# ABI hall of shame 367# 368config CLONE_BACKWARDS 369 bool 370 help 371 Architecture has tls passed as the 4th argument of clone(2), 372 not the 5th one. 373 374config CLONE_BACKWARDS2 375 bool 376 help 377 Architecture has the first two arguments of clone(2) swapped. 378 379source "kernel/gcov/Kconfig" 380