1config ARCH 2 string 3 option env="ARCH" 4 5config KERNELVERSION 6 string 7 option env="KERNELVERSION" 8 9config DEFCONFIG_LIST 10 string 11 depends on !UML 12 option defconfig_list 13 default "/lib/modules/$UNAME_RELEASE/.config" 14 default "/etc/kernel-config" 15 default "/boot/config-$UNAME_RELEASE" 16 default "$ARCH_DEFCONFIG" 17 default "arch/$ARCH/defconfig" 18 19config CONSTRUCTORS 20 bool 21 depends on !UML 22 23config IRQ_WORK 24 bool 25 26config BUILDTIME_EXTABLE_SORT 27 bool 28 29menu "General setup" 30 31config BROKEN 32 bool 33 34config BROKEN_ON_SMP 35 bool 36 depends on BROKEN || !SMP 37 default y 38 39config INIT_ENV_ARG_LIMIT 40 int 41 default 32 if !UML 42 default 128 if UML 43 help 44 Maximum of each of the number of arguments and environment 45 variables passed to init from the kernel command line. 46 47 48config CROSS_COMPILE 49 string "Cross-compiler tool prefix" 50 help 51 Same as running 'make CROSS_COMPILE=prefix-' but stored for 52 default make runs in this kernel build directory. You don't 53 need to set this unless you want the configured kernel build 54 directory to select the cross-compiler automatically. 55 56config LOCALVERSION 57 string "Local version - append to kernel release" 58 help 59 Append an extra string to the end of your kernel version. 60 This will show up when you type uname, for example. 61 The string you set here will be appended after the contents of 62 any files with a filename matching localversion* in your 63 object and source tree, in that order. Your total string can 64 be a maximum of 64 characters. 65 66config LOCALVERSION_AUTO 67 bool "Automatically append version information to the version string" 68 default y 69 help 70 This will try to automatically determine if the current tree is a 71 release tree by looking for git tags that belong to the current 72 top of tree revision. 73 74 A string of the format -gxxxxxxxx will be added to the localversion 75 if a git-based tree is found. The string generated by this will be 76 appended after any matching localversion* files, and after the value 77 set in CONFIG_LOCALVERSION. 78 79 (The actual string used here is the first eight characters produced 80 by running the command: 81 82 $ git rev-parse --verify HEAD 83 84 which is done within the script "scripts/setlocalversion".) 85 86config HAVE_KERNEL_GZIP 87 bool 88 89config HAVE_KERNEL_BZIP2 90 bool 91 92config HAVE_KERNEL_LZMA 93 bool 94 95config HAVE_KERNEL_XZ 96 bool 97 98config HAVE_KERNEL_LZO 99 bool 100 101choice 102 prompt "Kernel compression mode" 103 default KERNEL_GZIP 104 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO 105 help 106 The linux kernel is a kind of self-extracting executable. 107 Several compression algorithms are available, which differ 108 in efficiency, compression and decompression speed. 109 Compression speed is only relevant when building a kernel. 110 Decompression speed is relevant at each boot. 111 112 If you have any problems with bzip2 or lzma compressed 113 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 114 version of this functionality (bzip2 only), for 2.4, was 115 supplied by Christian Ludwig) 116 117 High compression options are mostly useful for users, who 118 are low on disk space (embedded systems), but for whom ram 119 size matters less. 120 121 If in doubt, select 'gzip' 122 123config KERNEL_GZIP 124 bool "Gzip" 125 depends on HAVE_KERNEL_GZIP 126 help 127 The old and tried gzip compression. It provides a good balance 128 between compression ratio and decompression speed. 129 130config KERNEL_BZIP2 131 bool "Bzip2" 132 depends on HAVE_KERNEL_BZIP2 133 help 134 Its compression ratio and speed is intermediate. 135 Decompression speed is slowest among the choices. The kernel 136 size is about 10% smaller with bzip2, in comparison to gzip. 137 Bzip2 uses a large amount of memory. For modern kernels you 138 will need at least 8MB RAM or more for booting. 139 140config KERNEL_LZMA 141 bool "LZMA" 142 depends on HAVE_KERNEL_LZMA 143 help 144 This compression algorithm's ratio is best. Decompression speed 145 is between gzip and bzip2. Compression is slowest. 146 The kernel size is about 33% smaller with LZMA in comparison to gzip. 147 148config KERNEL_XZ 149 bool "XZ" 150 depends on HAVE_KERNEL_XZ 151 help 152 XZ uses the LZMA2 algorithm and instruction set specific 153 BCJ filters which can improve compression ratio of executable 154 code. The size of the kernel is about 30% smaller with XZ in 155 comparison to gzip. On architectures for which there is a BCJ 156 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 157 will create a few percent smaller kernel than plain LZMA. 158 159 The speed is about the same as with LZMA: The decompression 160 speed of XZ is better than that of bzip2 but worse than gzip 161 and LZO. Compression is slow. 162 163config KERNEL_LZO 164 bool "LZO" 165 depends on HAVE_KERNEL_LZO 166 help 167 Its compression ratio is the poorest among the choices. The kernel 168 size is about 10% bigger than gzip; however its speed 169 (both compression and decompression) is the fastest. 170 171endchoice 172 173config DEFAULT_HOSTNAME 174 string "Default hostname" 175 default "(none)" 176 help 177 This option determines the default system hostname before userspace 178 calls sethostname(2). The kernel traditionally uses "(none)" here, 179 but you may wish to use a different default here to make a minimal 180 system more usable with less configuration. 181 182config SWAP 183 bool "Support for paging of anonymous memory (swap)" 184 depends on MMU && BLOCK 185 default y 186 help 187 This option allows you to choose whether you want to have support 188 for so called swap devices or swap files in your kernel that are 189 used to provide more virtual memory than the actual RAM present 190 in your computer. If unsure say Y. 191 192config SYSVIPC 193 bool "System V IPC" 194 ---help--- 195 Inter Process Communication is a suite of library functions and 196 system calls which let processes (running programs) synchronize and 197 exchange information. It is generally considered to be a good thing, 198 and some programs won't run unless you say Y here. In particular, if 199 you want to run the DOS emulator dosemu under Linux (read the 200 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 201 you'll need to say Y here. 202 203 You can find documentation about IPC with "info ipc" and also in 204 section 6.4 of the Linux Programmer's Guide, available from 205 <http://www.tldp.org/guides.html>. 206 207config SYSVIPC_SYSCTL 208 bool 209 depends on SYSVIPC 210 depends on SYSCTL 211 default y 212 213config POSIX_MQUEUE 214 bool "POSIX Message Queues" 215 depends on NET 216 ---help--- 217 POSIX variant of message queues is a part of IPC. In POSIX message 218 queues every message has a priority which decides about succession 219 of receiving it by a process. If you want to compile and run 220 programs written e.g. for Solaris with use of its POSIX message 221 queues (functions mq_*) say Y here. 222 223 POSIX message queues are visible as a filesystem called 'mqueue' 224 and can be mounted somewhere if you want to do filesystem 225 operations on message queues. 226 227 If unsure, say Y. 228 229config POSIX_MQUEUE_SYSCTL 230 bool 231 depends on POSIX_MQUEUE 232 depends on SYSCTL 233 default y 234 235config FHANDLE 236 bool "open by fhandle syscalls" 237 select EXPORTFS 238 help 239 If you say Y here, a user level program will be able to map 240 file names to handle and then later use the handle for 241 different file system operations. This is useful in implementing 242 userspace file servers, which now track files using handles instead 243 of names. The handle would remain the same even if file names 244 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 245 syscalls. 246 247config AUDIT 248 bool "Auditing support" 249 depends on NET 250 help 251 Enable auditing infrastructure that can be used with another 252 kernel subsystem, such as SELinux (which requires this for 253 logging of avc messages output). Does not do system-call 254 auditing without CONFIG_AUDITSYSCALL. 255 256config AUDITSYSCALL 257 bool "Enable system-call auditing support" 258 depends on AUDIT && (X86 || PPC || S390 || IA64 || UML || SPARC64 || SUPERH || (ARM && AEABI && !OABI_COMPAT)) 259 default y if SECURITY_SELINUX 260 help 261 Enable low-overhead system-call auditing infrastructure that 262 can be used independently or with another kernel subsystem, 263 such as SELinux. 264 265config AUDIT_WATCH 266 def_bool y 267 depends on AUDITSYSCALL 268 select FSNOTIFY 269 270config AUDIT_TREE 271 def_bool y 272 depends on AUDITSYSCALL 273 select FSNOTIFY 274 275config AUDIT_LOGINUID_IMMUTABLE 276 bool "Make audit loginuid immutable" 277 depends on AUDIT 278 help 279 The config option toggles if a task setting its loginuid requires 280 CAP_SYS_AUDITCONTROL or if that task should require no special permissions 281 but should instead only allow setting its loginuid if it was never 282 previously set. On systems which use systemd or a similar central 283 process to restart login services this should be set to true. On older 284 systems in which an admin would typically have to directly stop and 285 start processes this should be set to false. Setting this to true allows 286 one to drop potentially dangerous capabilites from the login tasks, 287 but may not be backwards compatible with older init systems. 288 289source "kernel/irq/Kconfig" 290source "kernel/time/Kconfig" 291 292menu "CPU/Task time and stats accounting" 293 294config VIRT_CPU_ACCOUNTING 295 bool 296 297choice 298 prompt "Cputime accounting" 299 default TICK_CPU_ACCOUNTING if !PPC64 300 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 301 302# Kind of a stub config for the pure tick based cputime accounting 303config TICK_CPU_ACCOUNTING 304 bool "Simple tick based cputime accounting" 305 depends on !S390 && !NO_HZ_FULL 306 help 307 This is the basic tick based cputime accounting that maintains 308 statistics about user, system and idle time spent on per jiffies 309 granularity. 310 311 If unsure, say Y. 312 313config VIRT_CPU_ACCOUNTING_NATIVE 314 bool "Deterministic task and CPU time accounting" 315 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 316 select VIRT_CPU_ACCOUNTING 317 help 318 Select this option to enable more accurate task and CPU time 319 accounting. This is done by reading a CPU counter on each 320 kernel entry and exit and on transitions within the kernel 321 between system, softirq and hardirq state, so there is a 322 small performance impact. In the case of s390 or IBM POWER > 5, 323 this also enables accounting of stolen time on logically-partitioned 324 systems. 325 326config VIRT_CPU_ACCOUNTING_GEN 327 bool "Full dynticks CPU time accounting" 328 depends on HAVE_CONTEXT_TRACKING && 64BIT 329 select VIRT_CPU_ACCOUNTING 330 select CONTEXT_TRACKING 331 help 332 Select this option to enable task and CPU time accounting on full 333 dynticks systems. This accounting is implemented by watching every 334 kernel-user boundaries using the context tracking subsystem. 335 The accounting is thus performed at the expense of some significant 336 overhead. 337 338 For now this is only useful if you are working on the full 339 dynticks subsystem development. 340 341 If unsure, say N. 342 343config IRQ_TIME_ACCOUNTING 344 bool "Fine granularity task level IRQ time accounting" 345 depends on HAVE_IRQ_TIME_ACCOUNTING && !NO_HZ_FULL 346 help 347 Select this option to enable fine granularity task irq time 348 accounting. This is done by reading a timestamp on each 349 transitions between softirq and hardirq state, so there can be a 350 small performance impact. 351 352 If in doubt, say N here. 353 354endchoice 355 356config BSD_PROCESS_ACCT 357 bool "BSD Process Accounting" 358 help 359 If you say Y here, a user level program will be able to instruct the 360 kernel (via a special system call) to write process accounting 361 information to a file: whenever a process exits, information about 362 that process will be appended to the file by the kernel. The 363 information includes things such as creation time, owning user, 364 command name, memory usage, controlling terminal etc. (the complete 365 list is in the struct acct in <file:include/linux/acct.h>). It is 366 up to the user level program to do useful things with this 367 information. This is generally a good idea, so say Y. 368 369config BSD_PROCESS_ACCT_V3 370 bool "BSD Process Accounting version 3 file format" 371 depends on BSD_PROCESS_ACCT 372 default n 373 help 374 If you say Y here, the process accounting information is written 375 in a new file format that also logs the process IDs of each 376 process and it's parent. Note that this file format is incompatible 377 with previous v0/v1/v2 file formats, so you will need updated tools 378 for processing it. A preliminary version of these tools is available 379 at <http://www.gnu.org/software/acct/>. 380 381config TASKSTATS 382 bool "Export task/process statistics through netlink" 383 depends on NET 384 default n 385 help 386 Export selected statistics for tasks/processes through the 387 generic netlink interface. Unlike BSD process accounting, the 388 statistics are available during the lifetime of tasks/processes as 389 responses to commands. Like BSD accounting, they are sent to user 390 space on task exit. 391 392 Say N if unsure. 393 394config TASK_DELAY_ACCT 395 bool "Enable per-task delay accounting" 396 depends on TASKSTATS 397 help 398 Collect information on time spent by a task waiting for system 399 resources like cpu, synchronous block I/O completion and swapping 400 in pages. Such statistics can help in setting a task's priorities 401 relative to other tasks for cpu, io, rss limits etc. 402 403 Say N if unsure. 404 405config TASK_XACCT 406 bool "Enable extended accounting over taskstats" 407 depends on TASKSTATS 408 help 409 Collect extended task accounting data and send the data 410 to userland for processing over the taskstats interface. 411 412 Say N if unsure. 413 414config TASK_IO_ACCOUNTING 415 bool "Enable per-task storage I/O accounting" 416 depends on TASK_XACCT 417 help 418 Collect information on the number of bytes of storage I/O which this 419 task has caused. 420 421 Say N if unsure. 422 423endmenu # "CPU/Task time and stats accounting" 424 425menu "RCU Subsystem" 426 427choice 428 prompt "RCU Implementation" 429 default TREE_RCU 430 431config TREE_RCU 432 bool "Tree-based hierarchical RCU" 433 depends on !PREEMPT && SMP 434 help 435 This option selects the RCU implementation that is 436 designed for very large SMP system with hundreds or 437 thousands of CPUs. It also scales down nicely to 438 smaller systems. 439 440config TREE_PREEMPT_RCU 441 bool "Preemptible tree-based hierarchical RCU" 442 depends on PREEMPT 443 help 444 This option selects the RCU implementation that is 445 designed for very large SMP systems with hundreds or 446 thousands of CPUs, but for which real-time response 447 is also required. It also scales down nicely to 448 smaller systems. 449 450 Select this option if you are unsure. 451 452config TINY_RCU 453 bool "UP-only small-memory-footprint RCU" 454 depends on !PREEMPT && !SMP 455 help 456 This option selects the RCU implementation that is 457 designed for UP systems from which real-time response 458 is not required. This option greatly reduces the 459 memory footprint of RCU. 460 461config TINY_PREEMPT_RCU 462 bool "Preemptible UP-only small-memory-footprint RCU" 463 depends on PREEMPT && !SMP 464 help 465 This option selects the RCU implementation that is designed 466 for real-time UP systems. This option greatly reduces the 467 memory footprint of RCU. 468 469endchoice 470 471config PREEMPT_RCU 472 def_bool ( TREE_PREEMPT_RCU || TINY_PREEMPT_RCU ) 473 help 474 This option enables preemptible-RCU code that is common between 475 the TREE_PREEMPT_RCU and TINY_PREEMPT_RCU implementations. 476 477config RCU_STALL_COMMON 478 def_bool ( TREE_RCU || TREE_PREEMPT_RCU || RCU_TRACE ) 479 help 480 This option enables RCU CPU stall code that is common between 481 the TINY and TREE variants of RCU. The purpose is to allow 482 the tiny variants to disable RCU CPU stall warnings, while 483 making these warnings mandatory for the tree variants. 484 485config CONTEXT_TRACKING 486 bool 487 488config RCU_USER_QS 489 bool "Consider userspace as in RCU extended quiescent state" 490 depends on HAVE_CONTEXT_TRACKING && SMP 491 select CONTEXT_TRACKING 492 help 493 This option sets hooks on kernel / userspace boundaries and 494 puts RCU in extended quiescent state when the CPU runs in 495 userspace. It means that when a CPU runs in userspace, it is 496 excluded from the global RCU state machine and thus doesn't 497 try to keep the timer tick on for RCU. 498 499 Unless you want to hack and help the development of the full 500 dynticks mode, you shouldn't enable this option. It also 501 adds unnecessary overhead. 502 503 If unsure say N 504 505config CONTEXT_TRACKING_FORCE 506 bool "Force context tracking" 507 depends on CONTEXT_TRACKING 508 default CONTEXT_TRACKING 509 help 510 Probe on user/kernel boundaries by default in order to 511 test the features that rely on it such as userspace RCU extended 512 quiescent states. 513 This test is there for debugging until we have a real user like the 514 full dynticks mode. 515 516config RCU_FANOUT 517 int "Tree-based hierarchical RCU fanout value" 518 range 2 64 if 64BIT 519 range 2 32 if !64BIT 520 depends on TREE_RCU || TREE_PREEMPT_RCU 521 default 64 if 64BIT 522 default 32 if !64BIT 523 help 524 This option controls the fanout of hierarchical implementations 525 of RCU, allowing RCU to work efficiently on machines with 526 large numbers of CPUs. This value must be at least the fourth 527 root of NR_CPUS, which allows NR_CPUS to be insanely large. 528 The default value of RCU_FANOUT should be used for production 529 systems, but if you are stress-testing the RCU implementation 530 itself, small RCU_FANOUT values allow you to test large-system 531 code paths on small(er) systems. 532 533 Select a specific number if testing RCU itself. 534 Take the default if unsure. 535 536config RCU_FANOUT_LEAF 537 int "Tree-based hierarchical RCU leaf-level fanout value" 538 range 2 RCU_FANOUT if 64BIT 539 range 2 RCU_FANOUT if !64BIT 540 depends on TREE_RCU || TREE_PREEMPT_RCU 541 default 16 542 help 543 This option controls the leaf-level fanout of hierarchical 544 implementations of RCU, and allows trading off cache misses 545 against lock contention. Systems that synchronize their 546 scheduling-clock interrupts for energy-efficiency reasons will 547 want the default because the smaller leaf-level fanout keeps 548 lock contention levels acceptably low. Very large systems 549 (hundreds or thousands of CPUs) will instead want to set this 550 value to the maximum value possible in order to reduce the 551 number of cache misses incurred during RCU's grace-period 552 initialization. These systems tend to run CPU-bound, and thus 553 are not helped by synchronized interrupts, and thus tend to 554 skew them, which reduces lock contention enough that large 555 leaf-level fanouts work well. 556 557 Select a specific number if testing RCU itself. 558 559 Select the maximum permissible value for large systems. 560 561 Take the default if unsure. 562 563config RCU_FANOUT_EXACT 564 bool "Disable tree-based hierarchical RCU auto-balancing" 565 depends on TREE_RCU || TREE_PREEMPT_RCU 566 default n 567 help 568 This option forces use of the exact RCU_FANOUT value specified, 569 regardless of imbalances in the hierarchy. This is useful for 570 testing RCU itself, and might one day be useful on systems with 571 strong NUMA behavior. 572 573 Without RCU_FANOUT_EXACT, the code will balance the hierarchy. 574 575 Say N if unsure. 576 577config RCU_FAST_NO_HZ 578 bool "Accelerate last non-dyntick-idle CPU's grace periods" 579 depends on NO_HZ_COMMON && SMP 580 default n 581 help 582 This option permits CPUs to enter dynticks-idle state even if 583 they have RCU callbacks queued, and prevents RCU from waking 584 these CPUs up more than roughly once every four jiffies (by 585 default, you can adjust this using the rcutree.rcu_idle_gp_delay 586 parameter), thus improving energy efficiency. On the other 587 hand, this option increases the duration of RCU grace periods, 588 for example, slowing down synchronize_rcu(). 589 590 Say Y if energy efficiency is critically important, and you 591 don't care about increased grace-period durations. 592 593 Say N if you are unsure. 594 595config TREE_RCU_TRACE 596 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU ) 597 select DEBUG_FS 598 help 599 This option provides tracing for the TREE_RCU and 600 TREE_PREEMPT_RCU implementations, permitting Makefile to 601 trivially select kernel/rcutree_trace.c. 602 603config RCU_BOOST 604 bool "Enable RCU priority boosting" 605 depends on RT_MUTEXES && PREEMPT_RCU 606 default n 607 help 608 This option boosts the priority of preempted RCU readers that 609 block the current preemptible RCU grace period for too long. 610 This option also prevents heavy loads from blocking RCU 611 callback invocation for all flavors of RCU. 612 613 Say Y here if you are working with real-time apps or heavy loads 614 Say N here if you are unsure. 615 616config RCU_BOOST_PRIO 617 int "Real-time priority to boost RCU readers to" 618 range 1 99 619 depends on RCU_BOOST 620 default 1 621 help 622 This option specifies the real-time priority to which long-term 623 preempted RCU readers are to be boosted. If you are working 624 with a real-time application that has one or more CPU-bound 625 threads running at a real-time priority level, you should set 626 RCU_BOOST_PRIO to a priority higher then the highest-priority 627 real-time CPU-bound thread. The default RCU_BOOST_PRIO value 628 of 1 is appropriate in the common case, which is real-time 629 applications that do not have any CPU-bound threads. 630 631 Some real-time applications might not have a single real-time 632 thread that saturates a given CPU, but instead might have 633 multiple real-time threads that, taken together, fully utilize 634 that CPU. In this case, you should set RCU_BOOST_PRIO to 635 a priority higher than the lowest-priority thread that is 636 conspiring to prevent the CPU from running any non-real-time 637 tasks. For example, if one thread at priority 10 and another 638 thread at priority 5 are between themselves fully consuming 639 the CPU time on a given CPU, then RCU_BOOST_PRIO should be 640 set to priority 6 or higher. 641 642 Specify the real-time priority, or take the default if unsure. 643 644config RCU_BOOST_DELAY 645 int "Milliseconds to delay boosting after RCU grace-period start" 646 range 0 3000 647 depends on RCU_BOOST 648 default 500 649 help 650 This option specifies the time to wait after the beginning of 651 a given grace period before priority-boosting preempted RCU 652 readers blocking that grace period. Note that any RCU reader 653 blocking an expedited RCU grace period is boosted immediately. 654 655 Accept the default if unsure. 656 657config RCU_NOCB_CPU 658 bool "Offload RCU callback processing from boot-selected CPUs (EXPERIMENTAL" 659 depends on TREE_RCU || TREE_PREEMPT_RCU 660 default n 661 help 662 Use this option to reduce OS jitter for aggressive HPC or 663 real-time workloads. It can also be used to offload RCU 664 callback invocation to energy-efficient CPUs in battery-powered 665 asymmetric multiprocessors. 666 667 This option offloads callback invocation from the set of 668 CPUs specified at boot time by the rcu_nocbs parameter. 669 For each such CPU, a kthread ("rcuox/N") will be created to 670 invoke callbacks, where the "N" is the CPU being offloaded, 671 and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and 672 "s" for RCU-sched. Nothing prevents this kthread from running 673 on the specified CPUs, but (1) the kthreads may be preempted 674 between each callback, and (2) affinity or cgroups can be used 675 to force the kthreads to run on whatever set of CPUs is desired. 676 677 Say Y here if you want to help to debug reduced OS jitter. 678 Say N here if you are unsure. 679 680choice 681 prompt "Build-forced no-CBs CPUs" 682 default RCU_NOCB_CPU_NONE 683 help 684 This option allows no-CBs CPUs to be specified at build time. 685 Additional no-CBs CPUs may be specified by the rcu_nocbs= 686 boot parameter. 687 688config RCU_NOCB_CPU_NONE 689 bool "No build_forced no-CBs CPUs" 690 depends on RCU_NOCB_CPU && !NO_HZ_FULL 691 help 692 This option does not force any of the CPUs to be no-CBs CPUs. 693 Only CPUs designated by the rcu_nocbs= boot parameter will be 694 no-CBs CPUs. 695 696config RCU_NOCB_CPU_ZERO 697 bool "CPU 0 is a build_forced no-CBs CPU" 698 depends on RCU_NOCB_CPU && !NO_HZ_FULL 699 help 700 This option forces CPU 0 to be a no-CBs CPU. Additional CPUs 701 may be designated as no-CBs CPUs using the rcu_nocbs= boot 702 parameter will be no-CBs CPUs. 703 704 Select this if CPU 0 needs to be a no-CBs CPU for real-time 705 or energy-efficiency reasons. 706 707config RCU_NOCB_CPU_ALL 708 bool "All CPUs are build_forced no-CBs CPUs" 709 depends on RCU_NOCB_CPU 710 help 711 This option forces all CPUs to be no-CBs CPUs. The rcu_nocbs= 712 boot parameter will be ignored. 713 714 Select this if all CPUs need to be no-CBs CPUs for real-time 715 or energy-efficiency reasons. 716 717endchoice 718 719endmenu # "RCU Subsystem" 720 721config IKCONFIG 722 tristate "Kernel .config support" 723 ---help--- 724 This option enables the complete Linux kernel ".config" file 725 contents to be saved in the kernel. It provides documentation 726 of which kernel options are used in a running kernel or in an 727 on-disk kernel. This information can be extracted from the kernel 728 image file with the script scripts/extract-ikconfig and used as 729 input to rebuild the current kernel or to build another kernel. 730 It can also be extracted from a running kernel by reading 731 /proc/config.gz if enabled (below). 732 733config IKCONFIG_PROC 734 bool "Enable access to .config through /proc/config.gz" 735 depends on IKCONFIG && PROC_FS 736 ---help--- 737 This option enables access to the kernel configuration file 738 through /proc/config.gz. 739 740config LOG_BUF_SHIFT 741 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 742 range 12 21 743 default 17 744 help 745 Select kernel log buffer size as a power of 2. 746 Examples: 747 17 => 128 KB 748 16 => 64 KB 749 15 => 32 KB 750 14 => 16 KB 751 13 => 8 KB 752 12 => 4 KB 753 754# 755# Architectures with an unreliable sched_clock() should select this: 756# 757config HAVE_UNSTABLE_SCHED_CLOCK 758 bool 759 760# 761# For architectures that want to enable the support for NUMA-affine scheduler 762# balancing logic: 763# 764config ARCH_SUPPORTS_NUMA_BALANCING 765 bool 766 767# For architectures that (ab)use NUMA to represent different memory regions 768# all cpu-local but of different latencies, such as SuperH. 769# 770config ARCH_WANT_NUMA_VARIABLE_LOCALITY 771 bool 772 773# 774# For architectures that are willing to define _PAGE_NUMA as _PAGE_PROTNONE 775config ARCH_WANTS_PROT_NUMA_PROT_NONE 776 bool 777 778config ARCH_USES_NUMA_PROT_NONE 779 bool 780 default y 781 depends on ARCH_WANTS_PROT_NUMA_PROT_NONE 782 depends on NUMA_BALANCING 783 784config NUMA_BALANCING_DEFAULT_ENABLED 785 bool "Automatically enable NUMA aware memory/task placement" 786 default y 787 depends on NUMA_BALANCING 788 help 789 If set, autonumic NUMA balancing will be enabled if running on a NUMA 790 machine. 791 792config NUMA_BALANCING 793 bool "Memory placement aware NUMA scheduler" 794 depends on ARCH_SUPPORTS_NUMA_BALANCING 795 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 796 depends on SMP && NUMA && MIGRATION 797 help 798 This option adds support for automatic NUMA aware memory/task placement. 799 The mechanism is quite primitive and is based on migrating memory when 800 it is references to the node the task is running on. 801 802 This system will be inactive on UMA systems. 803 804menuconfig CGROUPS 805 boolean "Control Group support" 806 depends on EVENTFD 807 help 808 This option adds support for grouping sets of processes together, for 809 use with process control subsystems such as Cpusets, CFS, memory 810 controls or device isolation. 811 See 812 - Documentation/scheduler/sched-design-CFS.txt (CFS) 813 - Documentation/cgroups/ (features for grouping, isolation 814 and resource control) 815 816 Say N if unsure. 817 818if CGROUPS 819 820config CGROUP_DEBUG 821 bool "Example debug cgroup subsystem" 822 default n 823 help 824 This option enables a simple cgroup subsystem that 825 exports useful debugging information about the cgroups 826 framework. 827 828 Say N if unsure. 829 830config CGROUP_FREEZER 831 bool "Freezer cgroup subsystem" 832 help 833 Provides a way to freeze and unfreeze all tasks in a 834 cgroup. 835 836config CGROUP_DEVICE 837 bool "Device controller for cgroups" 838 help 839 Provides a cgroup implementing whitelists for devices which 840 a process in the cgroup can mknod or open. 841 842config CPUSETS 843 bool "Cpuset support" 844 help 845 This option will let you create and manage CPUSETs which 846 allow dynamically partitioning a system into sets of CPUs and 847 Memory Nodes and assigning tasks to run only within those sets. 848 This is primarily useful on large SMP or NUMA systems. 849 850 Say N if unsure. 851 852config PROC_PID_CPUSET 853 bool "Include legacy /proc/<pid>/cpuset file" 854 depends on CPUSETS 855 default y 856 857config CGROUP_CPUACCT 858 bool "Simple CPU accounting cgroup subsystem" 859 help 860 Provides a simple Resource Controller for monitoring the 861 total CPU consumed by the tasks in a cgroup. 862 863config RESOURCE_COUNTERS 864 bool "Resource counters" 865 help 866 This option enables controller independent resource accounting 867 infrastructure that works with cgroups. 868 869config MEMCG 870 bool "Memory Resource Controller for Control Groups" 871 depends on RESOURCE_COUNTERS 872 select MM_OWNER 873 help 874 Provides a memory resource controller that manages both anonymous 875 memory and page cache. (See Documentation/cgroups/memory.txt) 876 877 Note that setting this option increases fixed memory overhead 878 associated with each page of memory in the system. By this, 879 20(40)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory 880 usage tracking struct at boot. Total amount of this is printed out 881 at boot. 882 883 Only enable when you're ok with these trade offs and really 884 sure you need the memory resource controller. Even when you enable 885 this, you can set "cgroup_disable=memory" at your boot option to 886 disable memory resource controller and you can avoid overheads. 887 (and lose benefits of memory resource controller) 888 889 This config option also selects MM_OWNER config option, which 890 could in turn add some fork/exit overhead. 891 892config MEMCG_SWAP 893 bool "Memory Resource Controller Swap Extension" 894 depends on MEMCG && SWAP 895 help 896 Add swap management feature to memory resource controller. When you 897 enable this, you can limit mem+swap usage per cgroup. In other words, 898 when you disable this, memory resource controller has no cares to 899 usage of swap...a process can exhaust all of the swap. This extension 900 is useful when you want to avoid exhaustion swap but this itself 901 adds more overheads and consumes memory for remembering information. 902 Especially if you use 32bit system or small memory system, please 903 be careful about enabling this. When memory resource controller 904 is disabled by boot option, this will be automatically disabled and 905 there will be no overhead from this. Even when you set this config=y, 906 if boot option "swapaccount=0" is set, swap will not be accounted. 907 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page 908 size is 4096bytes, 512k per 1Gbytes of swap. 909config MEMCG_SWAP_ENABLED 910 bool "Memory Resource Controller Swap Extension enabled by default" 911 depends on MEMCG_SWAP 912 default y 913 help 914 Memory Resource Controller Swap Extension comes with its price in 915 a bigger memory consumption. General purpose distribution kernels 916 which want to enable the feature but keep it disabled by default 917 and let the user enable it by swapaccount boot command line 918 parameter should have this option unselected. 919 For those who want to have the feature enabled by default should 920 select this option (if, for some reason, they need to disable it 921 then swapaccount=0 does the trick). 922config MEMCG_KMEM 923 bool "Memory Resource Controller Kernel Memory accounting" 924 depends on MEMCG 925 depends on SLUB || SLAB 926 help 927 The Kernel Memory extension for Memory Resource Controller can limit 928 the amount of memory used by kernel objects in the system. Those are 929 fundamentally different from the entities handled by the standard 930 Memory Controller, which are page-based, and can be swapped. Users of 931 the kmem extension can use it to guarantee that no group of processes 932 will ever exhaust kernel resources alone. 933 934config CGROUP_HUGETLB 935 bool "HugeTLB Resource Controller for Control Groups" 936 depends on RESOURCE_COUNTERS && HUGETLB_PAGE 937 default n 938 help 939 Provides a cgroup Resource Controller for HugeTLB pages. 940 When you enable this, you can put a per cgroup limit on HugeTLB usage. 941 The limit is enforced during page fault. Since HugeTLB doesn't 942 support page reclaim, enforcing the limit at page fault time implies 943 that, the application will get SIGBUS signal if it tries to access 944 HugeTLB pages beyond its limit. This requires the application to know 945 beforehand how much HugeTLB pages it would require for its use. The 946 control group is tracked in the third page lru pointer. This means 947 that we cannot use the controller with huge page less than 3 pages. 948 949config CGROUP_PERF 950 bool "Enable perf_event per-cpu per-container group (cgroup) monitoring" 951 depends on PERF_EVENTS && CGROUPS 952 help 953 This option extends the per-cpu mode to restrict monitoring to 954 threads which belong to the cgroup specified and run on the 955 designated cpu. 956 957 Say N if unsure. 958 959menuconfig CGROUP_SCHED 960 bool "Group CPU scheduler" 961 default n 962 help 963 This feature lets CPU scheduler recognize task groups and control CPU 964 bandwidth allocation to such task groups. It uses cgroups to group 965 tasks. 966 967if CGROUP_SCHED 968config FAIR_GROUP_SCHED 969 bool "Group scheduling for SCHED_OTHER" 970 depends on CGROUP_SCHED 971 default CGROUP_SCHED 972 973config CFS_BANDWIDTH 974 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 975 depends on FAIR_GROUP_SCHED 976 default n 977 help 978 This option allows users to define CPU bandwidth rates (limits) for 979 tasks running within the fair group scheduler. Groups with no limit 980 set are considered to be unconstrained and will run with no 981 restriction. 982 See tip/Documentation/scheduler/sched-bwc.txt for more information. 983 984config RT_GROUP_SCHED 985 bool "Group scheduling for SCHED_RR/FIFO" 986 depends on CGROUP_SCHED 987 default n 988 help 989 This feature lets you explicitly allocate real CPU bandwidth 990 to task groups. If enabled, it will also make it impossible to 991 schedule realtime tasks for non-root users until you allocate 992 realtime bandwidth for them. 993 See Documentation/scheduler/sched-rt-group.txt for more information. 994 995endif #CGROUP_SCHED 996 997config BLK_CGROUP 998 bool "Block IO controller" 999 depends on BLOCK 1000 default n 1001 ---help--- 1002 Generic block IO controller cgroup interface. This is the common 1003 cgroup interface which should be used by various IO controlling 1004 policies. 1005 1006 Currently, CFQ IO scheduler uses it to recognize task groups and 1007 control disk bandwidth allocation (proportional time slice allocation) 1008 to such task groups. It is also used by bio throttling logic in 1009 block layer to implement upper limit in IO rates on a device. 1010 1011 This option only enables generic Block IO controller infrastructure. 1012 One needs to also enable actual IO controlling logic/policy. For 1013 enabling proportional weight division of disk bandwidth in CFQ, set 1014 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 1015 CONFIG_BLK_DEV_THROTTLING=y. 1016 1017 See Documentation/cgroups/blkio-controller.txt for more information. 1018 1019config DEBUG_BLK_CGROUP 1020 bool "Enable Block IO controller debugging" 1021 depends on BLK_CGROUP 1022 default n 1023 ---help--- 1024 Enable some debugging help. Currently it exports additional stat 1025 files in a cgroup which can be useful for debugging. 1026 1027endif # CGROUPS 1028 1029config CHECKPOINT_RESTORE 1030 bool "Checkpoint/restore support" if EXPERT 1031 default n 1032 help 1033 Enables additional kernel features in a sake of checkpoint/restore. 1034 In particular it adds auxiliary prctl codes to setup process text, 1035 data and heap segment sizes, and a few additional /proc filesystem 1036 entries. 1037 1038 If unsure, say N here. 1039 1040menuconfig NAMESPACES 1041 bool "Namespaces support" if EXPERT 1042 default !EXPERT 1043 help 1044 Provides the way to make tasks work with different objects using 1045 the same id. For example same IPC id may refer to different objects 1046 or same user id or pid may refer to different tasks when used in 1047 different namespaces. 1048 1049if NAMESPACES 1050 1051config UTS_NS 1052 bool "UTS namespace" 1053 default y 1054 help 1055 In this namespace tasks see different info provided with the 1056 uname() system call 1057 1058config IPC_NS 1059 bool "IPC namespace" 1060 depends on (SYSVIPC || POSIX_MQUEUE) 1061 default y 1062 help 1063 In this namespace tasks work with IPC ids which correspond to 1064 different IPC objects in different namespaces. 1065 1066config USER_NS 1067 bool "User namespace" 1068 depends on UIDGID_CONVERTED 1069 select UIDGID_STRICT_TYPE_CHECKS 1070 1071 default n 1072 help 1073 This allows containers, i.e. vservers, to use user namespaces 1074 to provide different user info for different servers. 1075 1076 When user namespaces are enabled in the kernel it is 1077 recommended that the MEMCG and MEMCG_KMEM options also be 1078 enabled and that user-space use the memory control groups to 1079 limit the amount of memory a memory unprivileged users can 1080 use. 1081 1082 If unsure, say N. 1083 1084config PID_NS 1085 bool "PID Namespaces" 1086 default y 1087 help 1088 Support process id namespaces. This allows having multiple 1089 processes with the same pid as long as they are in different 1090 pid namespaces. This is a building block of containers. 1091 1092config NET_NS 1093 bool "Network namespace" 1094 depends on NET 1095 default y 1096 help 1097 Allow user space to create what appear to be multiple instances 1098 of the network stack. 1099 1100endif # NAMESPACES 1101 1102config UIDGID_CONVERTED 1103 # True if all of the selected software conmponents are known 1104 # to have uid_t and gid_t converted to kuid_t and kgid_t 1105 # where appropriate and are otherwise safe to use with 1106 # the user namespace. 1107 bool 1108 default y 1109 1110 # Filesystems 1111 depends on XFS_FS = n 1112 1113config UIDGID_STRICT_TYPE_CHECKS 1114 bool "Require conversions between uid/gids and their internal representation" 1115 depends on UIDGID_CONVERTED 1116 default n 1117 help 1118 While the nececessary conversions are being added to all subsystems this option allows 1119 the code to continue to build for unconverted subsystems. 1120 1121 Say Y here if you want the strict type checking enabled 1122 1123config SCHED_AUTOGROUP 1124 bool "Automatic process group scheduling" 1125 select EVENTFD 1126 select CGROUPS 1127 select CGROUP_SCHED 1128 select FAIR_GROUP_SCHED 1129 help 1130 This option optimizes the scheduler for common desktop workloads by 1131 automatically creating and populating task groups. This separation 1132 of workloads isolates aggressive CPU burners (like build jobs) from 1133 desktop applications. Task group autogeneration is currently based 1134 upon task session. 1135 1136config MM_OWNER 1137 bool 1138 1139config SYSFS_DEPRECATED 1140 bool "Enable deprecated sysfs features to support old userspace tools" 1141 depends on SYSFS 1142 default n 1143 help 1144 This option adds code that switches the layout of the "block" class 1145 devices, to not show up in /sys/class/block/, but only in 1146 /sys/block/. 1147 1148 This switch is only active when the sysfs.deprecated=1 boot option is 1149 passed or the SYSFS_DEPRECATED_V2 option is set. 1150 1151 This option allows new kernels to run on old distributions and tools, 1152 which might get confused by /sys/class/block/. Since 2007/2008 all 1153 major distributions and tools handle this just fine. 1154 1155 Recent distributions and userspace tools after 2009/2010 depend on 1156 the existence of /sys/class/block/, and will not work with this 1157 option enabled. 1158 1159 Only if you are using a new kernel on an old distribution, you might 1160 need to say Y here. 1161 1162config SYSFS_DEPRECATED_V2 1163 bool "Enable deprecated sysfs features by default" 1164 default n 1165 depends on SYSFS 1166 depends on SYSFS_DEPRECATED 1167 help 1168 Enable deprecated sysfs by default. 1169 1170 See the CONFIG_SYSFS_DEPRECATED option for more details about this 1171 option. 1172 1173 Only if you are using a new kernel on an old distribution, you might 1174 need to say Y here. Even then, odds are you would not need it 1175 enabled, you can always pass the boot option if absolutely necessary. 1176 1177config RELAY 1178 bool "Kernel->user space relay support (formerly relayfs)" 1179 help 1180 This option enables support for relay interface support in 1181 certain file systems (such as debugfs). 1182 It is designed to provide an efficient mechanism for tools and 1183 facilities to relay large amounts of data from kernel space to 1184 user space. 1185 1186 If unsure, say N. 1187 1188config BLK_DEV_INITRD 1189 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 1190 depends on BROKEN || !FRV 1191 help 1192 The initial RAM filesystem is a ramfs which is loaded by the 1193 boot loader (loadlin or lilo) and that is mounted as root 1194 before the normal boot procedure. It is typically used to 1195 load modules needed to mount the "real" root file system, 1196 etc. See <file:Documentation/initrd.txt> for details. 1197 1198 If RAM disk support (BLK_DEV_RAM) is also included, this 1199 also enables initial RAM disk (initrd) support and adds 1200 15 Kbytes (more on some other architectures) to the kernel size. 1201 1202 If unsure say Y. 1203 1204if BLK_DEV_INITRD 1205 1206source "usr/Kconfig" 1207 1208endif 1209 1210config CC_OPTIMIZE_FOR_SIZE 1211 bool "Optimize for size" 1212 help 1213 Enabling this option will pass "-Os" instead of "-O2" to gcc 1214 resulting in a smaller kernel. 1215 1216 If unsure, say N. 1217 1218config SYSCTL 1219 bool 1220 1221config ANON_INODES 1222 bool 1223 1224config HAVE_UID16 1225 bool 1226 1227config SYSCTL_EXCEPTION_TRACE 1228 bool 1229 help 1230 Enable support for /proc/sys/debug/exception-trace. 1231 1232config SYSCTL_ARCH_UNALIGN_NO_WARN 1233 bool 1234 help 1235 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1236 Allows arch to define/use @no_unaligned_warning to possibly warn 1237 about unaligned access emulation going on under the hood. 1238 1239config SYSCTL_ARCH_UNALIGN_ALLOW 1240 bool 1241 help 1242 Enable support for /proc/sys/kernel/unaligned-trap 1243 Allows arches to define/use @unaligned_enabled to runtime toggle 1244 the unaligned access emulation. 1245 see arch/parisc/kernel/unaligned.c for reference 1246 1247config HOTPLUG 1248 def_bool y 1249 1250config HAVE_PCSPKR_PLATFORM 1251 bool 1252 1253menuconfig EXPERT 1254 bool "Configure standard kernel features (expert users)" 1255 # Unhide debug options, to make the on-by-default options visible 1256 select DEBUG_KERNEL 1257 help 1258 This option allows certain base kernel options and settings 1259 to be disabled or tweaked. This is for specialized 1260 environments which can tolerate a "non-standard" kernel. 1261 Only use this if you really know what you are doing. 1262 1263config UID16 1264 bool "Enable 16-bit UID system calls" if EXPERT 1265 depends on HAVE_UID16 1266 default y 1267 help 1268 This enables the legacy 16-bit UID syscall wrappers. 1269 1270config SYSCTL_SYSCALL 1271 bool "Sysctl syscall support" if EXPERT 1272 depends on PROC_SYSCTL 1273 default n 1274 select SYSCTL 1275 ---help--- 1276 sys_sysctl uses binary paths that have been found challenging 1277 to properly maintain and use. The interface in /proc/sys 1278 using paths with ascii names is now the primary path to this 1279 information. 1280 1281 Almost nothing using the binary sysctl interface so if you are 1282 trying to save some space it is probably safe to disable this, 1283 making your kernel marginally smaller. 1284 1285 If unsure say N here. 1286 1287config KALLSYMS 1288 bool "Load all symbols for debugging/ksymoops" if EXPERT 1289 default y 1290 help 1291 Say Y here to let the kernel print out symbolic crash information and 1292 symbolic stack backtraces. This increases the size of the kernel 1293 somewhat, as all symbols have to be loaded into the kernel image. 1294 1295config KALLSYMS_ALL 1296 bool "Include all symbols in kallsyms" 1297 depends on DEBUG_KERNEL && KALLSYMS 1298 help 1299 Normally kallsyms only contains the symbols of functions for nicer 1300 OOPS messages and backtraces (i.e., symbols from the text and inittext 1301 sections). This is sufficient for most cases. And only in very rare 1302 cases (e.g., when a debugger is used) all symbols are required (e.g., 1303 names of variables from the data sections, etc). 1304 1305 This option makes sure that all symbols are loaded into the kernel 1306 image (i.e., symbols from all sections) in cost of increased kernel 1307 size (depending on the kernel configuration, it may be 300KiB or 1308 something like this). 1309 1310 Say N unless you really need all symbols. 1311 1312config PRINTK 1313 default y 1314 bool "Enable support for printk" if EXPERT 1315 select IRQ_WORK 1316 help 1317 This option enables normal printk support. Removing it 1318 eliminates most of the message strings from the kernel image 1319 and makes the kernel more or less silent. As this makes it 1320 very difficult to diagnose system problems, saying N here is 1321 strongly discouraged. 1322 1323config BUG 1324 bool "BUG() support" if EXPERT 1325 default y 1326 help 1327 Disabling this option eliminates support for BUG and WARN, reducing 1328 the size of your kernel image and potentially quietly ignoring 1329 numerous fatal conditions. You should only consider disabling this 1330 option for embedded systems with no facilities for reporting errors. 1331 Just say Y. 1332 1333config ELF_CORE 1334 depends on COREDUMP 1335 default y 1336 bool "Enable ELF core dumps" if EXPERT 1337 help 1338 Enable support for generating core dumps. Disabling saves about 4k. 1339 1340 1341config PCSPKR_PLATFORM 1342 bool "Enable PC-Speaker support" if EXPERT 1343 depends on HAVE_PCSPKR_PLATFORM 1344 select I8253_LOCK 1345 default y 1346 help 1347 This option allows to disable the internal PC-Speaker 1348 support, saving some memory. 1349 1350config BASE_FULL 1351 default y 1352 bool "Enable full-sized data structures for core" if EXPERT 1353 help 1354 Disabling this option reduces the size of miscellaneous core 1355 kernel data structures. This saves memory on small machines, 1356 but may reduce performance. 1357 1358config FUTEX 1359 bool "Enable futex support" if EXPERT 1360 default y 1361 select RT_MUTEXES 1362 help 1363 Disabling this option will cause the kernel to be built without 1364 support for "fast userspace mutexes". The resulting kernel may not 1365 run glibc-based applications correctly. 1366 1367config EPOLL 1368 bool "Enable eventpoll support" if EXPERT 1369 default y 1370 select ANON_INODES 1371 help 1372 Disabling this option will cause the kernel to be built without 1373 support for epoll family of system calls. 1374 1375config SIGNALFD 1376 bool "Enable signalfd() system call" if EXPERT 1377 select ANON_INODES 1378 default y 1379 help 1380 Enable the signalfd() system call that allows to receive signals 1381 on a file descriptor. 1382 1383 If unsure, say Y. 1384 1385config TIMERFD 1386 bool "Enable timerfd() system call" if EXPERT 1387 select ANON_INODES 1388 default y 1389 help 1390 Enable the timerfd() system call that allows to receive timer 1391 events on a file descriptor. 1392 1393 If unsure, say Y. 1394 1395config EVENTFD 1396 bool "Enable eventfd() system call" if EXPERT 1397 select ANON_INODES 1398 default y 1399 help 1400 Enable the eventfd() system call that allows to receive both 1401 kernel notification (ie. KAIO) or userspace notifications. 1402 1403 If unsure, say Y. 1404 1405config SHMEM 1406 bool "Use full shmem filesystem" if EXPERT 1407 default y 1408 depends on MMU 1409 help 1410 The shmem is an internal filesystem used to manage shared memory. 1411 It is backed by swap and manages resource limits. It is also exported 1412 to userspace as tmpfs if TMPFS is enabled. Disabling this 1413 option replaces shmem and tmpfs with the much simpler ramfs code, 1414 which may be appropriate on small systems without swap. 1415 1416config AIO 1417 bool "Enable AIO support" if EXPERT 1418 default y 1419 help 1420 This option enables POSIX asynchronous I/O which may by used 1421 by some high performance threaded applications. Disabling 1422 this option saves about 7k. 1423 1424config PCI_QUIRKS 1425 default y 1426 bool "Enable PCI quirk workarounds" if EXPERT 1427 depends on PCI 1428 help 1429 This enables workarounds for various PCI chipset 1430 bugs/quirks. Disable this only if your target machine is 1431 unaffected by PCI quirks. 1432 1433config EMBEDDED 1434 bool "Embedded system" 1435 select EXPERT 1436 help 1437 This option should be enabled if compiling the kernel for 1438 an embedded system so certain expert options are available 1439 for configuration. 1440 1441config HAVE_PERF_EVENTS 1442 bool 1443 help 1444 See tools/perf/design.txt for details. 1445 1446config PERF_USE_VMALLOC 1447 bool 1448 help 1449 See tools/perf/design.txt for details 1450 1451menu "Kernel Performance Events And Counters" 1452 1453config PERF_EVENTS 1454 bool "Kernel performance events and counters" 1455 default y if PROFILING 1456 depends on HAVE_PERF_EVENTS 1457 select ANON_INODES 1458 select IRQ_WORK 1459 help 1460 Enable kernel support for various performance events provided 1461 by software and hardware. 1462 1463 Software events are supported either built-in or via the 1464 use of generic tracepoints. 1465 1466 Most modern CPUs support performance events via performance 1467 counter registers. These registers count the number of certain 1468 types of hw events: such as instructions executed, cachemisses 1469 suffered, or branches mis-predicted - without slowing down the 1470 kernel or applications. These registers can also trigger interrupts 1471 when a threshold number of events have passed - and can thus be 1472 used to profile the code that runs on that CPU. 1473 1474 The Linux Performance Event subsystem provides an abstraction of 1475 these software and hardware event capabilities, available via a 1476 system call and used by the "perf" utility in tools/perf/. It 1477 provides per task and per CPU counters, and it provides event 1478 capabilities on top of those. 1479 1480 Say Y if unsure. 1481 1482config DEBUG_PERF_USE_VMALLOC 1483 default n 1484 bool "Debug: use vmalloc to back perf mmap() buffers" 1485 depends on PERF_EVENTS && DEBUG_KERNEL 1486 select PERF_USE_VMALLOC 1487 help 1488 Use vmalloc memory to back perf mmap() buffers. 1489 1490 Mostly useful for debugging the vmalloc code on platforms 1491 that don't require it. 1492 1493 Say N if unsure. 1494 1495endmenu 1496 1497config VM_EVENT_COUNTERS 1498 default y 1499 bool "Enable VM event counters for /proc/vmstat" if EXPERT 1500 help 1501 VM event counters are needed for event counts to be shown. 1502 This option allows the disabling of the VM event counters 1503 on EXPERT systems. /proc/vmstat will only show page counts 1504 if VM event counters are disabled. 1505 1506config SLUB_DEBUG 1507 default y 1508 bool "Enable SLUB debugging support" if EXPERT 1509 depends on SLUB && SYSFS 1510 help 1511 SLUB has extensive debug support features. Disabling these can 1512 result in significant savings in code size. This also disables 1513 SLUB sysfs support. /sys/slab will not exist and there will be 1514 no support for cache validation etc. 1515 1516config COMPAT_BRK 1517 bool "Disable heap randomization" 1518 default y 1519 help 1520 Randomizing heap placement makes heap exploits harder, but it 1521 also breaks ancient binaries (including anything libc5 based). 1522 This option changes the bootup default to heap randomization 1523 disabled, and can be overridden at runtime by setting 1524 /proc/sys/kernel/randomize_va_space to 2. 1525 1526 On non-ancient distros (post-2000 ones) N is usually a safe choice. 1527 1528choice 1529 prompt "Choose SLAB allocator" 1530 default SLUB 1531 help 1532 This option allows to select a slab allocator. 1533 1534config SLAB 1535 bool "SLAB" 1536 help 1537 The regular slab allocator that is established and known to work 1538 well in all environments. It organizes cache hot objects in 1539 per cpu and per node queues. 1540 1541config SLUB 1542 bool "SLUB (Unqueued Allocator)" 1543 help 1544 SLUB is a slab allocator that minimizes cache line usage 1545 instead of managing queues of cached objects (SLAB approach). 1546 Per cpu caching is realized using slabs of objects instead 1547 of queues of objects. SLUB can use memory efficiently 1548 and has enhanced diagnostics. SLUB is the default choice for 1549 a slab allocator. 1550 1551config SLOB 1552 depends on EXPERT 1553 bool "SLOB (Simple Allocator)" 1554 help 1555 SLOB replaces the stock allocator with a drastically simpler 1556 allocator. SLOB is generally more space efficient but 1557 does not perform as well on large systems. 1558 1559endchoice 1560 1561config MMAP_ALLOW_UNINITIALIZED 1562 bool "Allow mmapped anonymous memory to be uninitialized" 1563 depends on EXPERT && !MMU 1564 default n 1565 help 1566 Normally, and according to the Linux spec, anonymous memory obtained 1567 from mmap() has it's contents cleared before it is passed to 1568 userspace. Enabling this config option allows you to request that 1569 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1570 providing a huge performance boost. If this option is not enabled, 1571 then the flag will be ignored. 1572 1573 This is taken advantage of by uClibc's malloc(), and also by 1574 ELF-FDPIC binfmt's brk and stack allocator. 1575 1576 Because of the obvious security issues, this option should only be 1577 enabled on embedded devices where you control what is run in 1578 userspace. Since that isn't generally a problem on no-MMU systems, 1579 it is normally safe to say Y here. 1580 1581 See Documentation/nommu-mmap.txt for more information. 1582 1583config PROFILING 1584 bool "Profiling support" 1585 help 1586 Say Y here to enable the extended profiling support mechanisms used 1587 by profilers such as OProfile. 1588 1589# 1590# Place an empty function call at each tracepoint site. Can be 1591# dynamically changed for a probe function. 1592# 1593config TRACEPOINTS 1594 bool 1595 1596source "arch/Kconfig" 1597 1598endmenu # General setup 1599 1600config HAVE_GENERIC_DMA_COHERENT 1601 bool 1602 default n 1603 1604config SLABINFO 1605 bool 1606 depends on PROC_FS 1607 depends on SLAB || SLUB_DEBUG 1608 default y 1609 1610config RT_MUTEXES 1611 boolean 1612 1613config BASE_SMALL 1614 int 1615 default 0 if BASE_FULL 1616 default 1 if !BASE_FULL 1617 1618menuconfig MODULES 1619 bool "Enable loadable module support" 1620 help 1621 Kernel modules are small pieces of compiled code which can 1622 be inserted in the running kernel, rather than being 1623 permanently built into the kernel. You use the "modprobe" 1624 tool to add (and sometimes remove) them. If you say Y here, 1625 many parts of the kernel can be built as modules (by 1626 answering M instead of Y where indicated): this is most 1627 useful for infrequently used options which are not required 1628 for booting. For more information, see the man pages for 1629 modprobe, lsmod, modinfo, insmod and rmmod. 1630 1631 If you say Y here, you will need to run "make 1632 modules_install" to put the modules under /lib/modules/ 1633 where modprobe can find them (you may need to be root to do 1634 this). 1635 1636 If unsure, say Y. 1637 1638if MODULES 1639 1640config MODULE_FORCE_LOAD 1641 bool "Forced module loading" 1642 default n 1643 help 1644 Allow loading of modules without version information (ie. modprobe 1645 --force). Forced module loading sets the 'F' (forced) taint flag and 1646 is usually a really bad idea. 1647 1648config MODULE_UNLOAD 1649 bool "Module unloading" 1650 help 1651 Without this option you will not be able to unload any 1652 modules (note that some modules may not be unloadable 1653 anyway), which makes your kernel smaller, faster 1654 and simpler. If unsure, say Y. 1655 1656config MODULE_FORCE_UNLOAD 1657 bool "Forced module unloading" 1658 depends on MODULE_UNLOAD 1659 help 1660 This option allows you to force a module to unload, even if the 1661 kernel believes it is unsafe: the kernel will remove the module 1662 without waiting for anyone to stop using it (using the -f option to 1663 rmmod). This is mainly for kernel developers and desperate users. 1664 If unsure, say N. 1665 1666config MODVERSIONS 1667 bool "Module versioning support" 1668 help 1669 Usually, you have to use modules compiled with your kernel. 1670 Saying Y here makes it sometimes possible to use modules 1671 compiled for different kernels, by adding enough information 1672 to the modules to (hopefully) spot any changes which would 1673 make them incompatible with the kernel you are running. If 1674 unsure, say N. 1675 1676config MODULE_SRCVERSION_ALL 1677 bool "Source checksum for all modules" 1678 help 1679 Modules which contain a MODULE_VERSION get an extra "srcversion" 1680 field inserted into their modinfo section, which contains a 1681 sum of the source files which made it. This helps maintainers 1682 see exactly which source was used to build a module (since 1683 others sometimes change the module source without updating 1684 the version). With this option, such a "srcversion" field 1685 will be created for all modules. If unsure, say N. 1686 1687config MODULE_SIG 1688 bool "Module signature verification" 1689 depends on MODULES 1690 select KEYS 1691 select CRYPTO 1692 select ASYMMETRIC_KEY_TYPE 1693 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1694 select PUBLIC_KEY_ALGO_RSA 1695 select ASN1 1696 select OID_REGISTRY 1697 select X509_CERTIFICATE_PARSER 1698 help 1699 Check modules for valid signatures upon load: the signature 1700 is simply appended to the module. For more information see 1701 Documentation/module-signing.txt. 1702 1703 !!!WARNING!!! If you enable this option, you MUST make sure that the 1704 module DOES NOT get stripped after being signed. This includes the 1705 debuginfo strip done by some packagers (such as rpmbuild) and 1706 inclusion into an initramfs that wants the module size reduced. 1707 1708config MODULE_SIG_FORCE 1709 bool "Require modules to be validly signed" 1710 depends on MODULE_SIG 1711 help 1712 Reject unsigned modules or signed modules for which we don't have a 1713 key. Without this, such modules will simply taint the kernel. 1714 1715config MODULE_SIG_ALL 1716 bool "Automatically sign all modules" 1717 default y 1718 depends on MODULE_SIG 1719 help 1720 Sign all modules during make modules_install. Without this option, 1721 modules must be signed manually, using the scripts/sign-file tool. 1722 1723comment "Do not forget to sign required modules with scripts/sign-file" 1724 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 1725 1726choice 1727 prompt "Which hash algorithm should modules be signed with?" 1728 depends on MODULE_SIG 1729 help 1730 This determines which sort of hashing algorithm will be used during 1731 signature generation. This algorithm _must_ be built into the kernel 1732 directly so that signature verification can take place. It is not 1733 possible to load a signed module containing the algorithm to check 1734 the signature on that module. 1735 1736config MODULE_SIG_SHA1 1737 bool "Sign modules with SHA-1" 1738 select CRYPTO_SHA1 1739 1740config MODULE_SIG_SHA224 1741 bool "Sign modules with SHA-224" 1742 select CRYPTO_SHA256 1743 1744config MODULE_SIG_SHA256 1745 bool "Sign modules with SHA-256" 1746 select CRYPTO_SHA256 1747 1748config MODULE_SIG_SHA384 1749 bool "Sign modules with SHA-384" 1750 select CRYPTO_SHA512 1751 1752config MODULE_SIG_SHA512 1753 bool "Sign modules with SHA-512" 1754 select CRYPTO_SHA512 1755 1756endchoice 1757 1758config MODULE_SIG_HASH 1759 string 1760 depends on MODULE_SIG 1761 default "sha1" if MODULE_SIG_SHA1 1762 default "sha224" if MODULE_SIG_SHA224 1763 default "sha256" if MODULE_SIG_SHA256 1764 default "sha384" if MODULE_SIG_SHA384 1765 default "sha512" if MODULE_SIG_SHA512 1766 1767endif # MODULES 1768 1769config INIT_ALL_POSSIBLE 1770 bool 1771 help 1772 Back when each arch used to define their own cpu_online_mask and 1773 cpu_possible_mask, some of them chose to initialize cpu_possible_mask 1774 with all 1s, and others with all 0s. When they were centralised, 1775 it was better to provide this option than to break all the archs 1776 and have several arch maintainers pursuing me down dark alleys. 1777 1778config STOP_MACHINE 1779 bool 1780 default y 1781 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU 1782 help 1783 Need stop_machine() primitive. 1784 1785source "block/Kconfig" 1786 1787config PREEMPT_NOTIFIERS 1788 bool 1789 1790config PADATA 1791 depends on SMP 1792 bool 1793 1794# Can be selected by architectures with broken toolchains 1795# that get confused by correct const<->read_only section 1796# mappings 1797config BROKEN_RODATA 1798 bool 1799 1800config ASN1 1801 tristate 1802 help 1803 Build a simple ASN.1 grammar compiler that produces a bytecode output 1804 that can be interpreted by the ASN.1 stream decoder and used to 1805 inform it as to what tags are to be expected in a stream and what 1806 functions to call on what tags. 1807 1808source "kernel/Kconfig.locks" 1809