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 COMPILE_TEST 57 bool "Compile also drivers which will not load" 58 default n 59 help 60 Some drivers can be compiled on a different platform than they are 61 intended to be run on. Despite they cannot be loaded there (or even 62 when they load they cannot be used due to missing HW support), 63 developers still, opposing to distributors, might want to build such 64 drivers to compile-test them. 65 66 If you are a developer and want to build everything available, say Y 67 here. If you are a user/distributor, say N here to exclude useless 68 drivers to be distributed. 69 70config LOCALVERSION 71 string "Local version - append to kernel release" 72 help 73 Append an extra string to the end of your kernel version. 74 This will show up when you type uname, for example. 75 The string you set here will be appended after the contents of 76 any files with a filename matching localversion* in your 77 object and source tree, in that order. Your total string can 78 be a maximum of 64 characters. 79 80config LOCALVERSION_AUTO 81 bool "Automatically append version information to the version string" 82 default y 83 help 84 This will try to automatically determine if the current tree is a 85 release tree by looking for git tags that belong to the current 86 top of tree revision. 87 88 A string of the format -gxxxxxxxx will be added to the localversion 89 if a git-based tree is found. The string generated by this will be 90 appended after any matching localversion* files, and after the value 91 set in CONFIG_LOCALVERSION. 92 93 (The actual string used here is the first eight characters produced 94 by running the command: 95 96 $ git rev-parse --verify HEAD 97 98 which is done within the script "scripts/setlocalversion".) 99 100config HAVE_KERNEL_GZIP 101 bool 102 103config HAVE_KERNEL_BZIP2 104 bool 105 106config HAVE_KERNEL_LZMA 107 bool 108 109config HAVE_KERNEL_XZ 110 bool 111 112config HAVE_KERNEL_LZO 113 bool 114 115config HAVE_KERNEL_LZ4 116 bool 117 118choice 119 prompt "Kernel compression mode" 120 default KERNEL_GZIP 121 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 122 help 123 The linux kernel is a kind of self-extracting executable. 124 Several compression algorithms are available, which differ 125 in efficiency, compression and decompression speed. 126 Compression speed is only relevant when building a kernel. 127 Decompression speed is relevant at each boot. 128 129 If you have any problems with bzip2 or lzma compressed 130 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 131 version of this functionality (bzip2 only), for 2.4, was 132 supplied by Christian Ludwig) 133 134 High compression options are mostly useful for users, who 135 are low on disk space (embedded systems), but for whom ram 136 size matters less. 137 138 If in doubt, select 'gzip' 139 140config KERNEL_GZIP 141 bool "Gzip" 142 depends on HAVE_KERNEL_GZIP 143 help 144 The old and tried gzip compression. It provides a good balance 145 between compression ratio and decompression speed. 146 147config KERNEL_BZIP2 148 bool "Bzip2" 149 depends on HAVE_KERNEL_BZIP2 150 help 151 Its compression ratio and speed is intermediate. 152 Decompression speed is slowest among the choices. The kernel 153 size is about 10% smaller with bzip2, in comparison to gzip. 154 Bzip2 uses a large amount of memory. For modern kernels you 155 will need at least 8MB RAM or more for booting. 156 157config KERNEL_LZMA 158 bool "LZMA" 159 depends on HAVE_KERNEL_LZMA 160 help 161 This compression algorithm's ratio is best. Decompression speed 162 is between gzip and bzip2. Compression is slowest. 163 The kernel size is about 33% smaller with LZMA in comparison to gzip. 164 165config KERNEL_XZ 166 bool "XZ" 167 depends on HAVE_KERNEL_XZ 168 help 169 XZ uses the LZMA2 algorithm and instruction set specific 170 BCJ filters which can improve compression ratio of executable 171 code. The size of the kernel is about 30% smaller with XZ in 172 comparison to gzip. On architectures for which there is a BCJ 173 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 174 will create a few percent smaller kernel than plain LZMA. 175 176 The speed is about the same as with LZMA: The decompression 177 speed of XZ is better than that of bzip2 but worse than gzip 178 and LZO. Compression is slow. 179 180config KERNEL_LZO 181 bool "LZO" 182 depends on HAVE_KERNEL_LZO 183 help 184 Its compression ratio is the poorest among the choices. The kernel 185 size is about 10% bigger than gzip; however its speed 186 (both compression and decompression) is the fastest. 187 188config KERNEL_LZ4 189 bool "LZ4" 190 depends on HAVE_KERNEL_LZ4 191 help 192 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. 193 A preliminary version of LZ4 de/compression tool is available at 194 <https://code.google.com/p/lz4/>. 195 196 Its compression ratio is worse than LZO. The size of the kernel 197 is about 8% bigger than LZO. But the decompression speed is 198 faster than LZO. 199 200endchoice 201 202config DEFAULT_HOSTNAME 203 string "Default hostname" 204 default "(none)" 205 help 206 This option determines the default system hostname before userspace 207 calls sethostname(2). The kernel traditionally uses "(none)" here, 208 but you may wish to use a different default here to make a minimal 209 system more usable with less configuration. 210 211config SWAP 212 bool "Support for paging of anonymous memory (swap)" 213 depends on MMU && BLOCK 214 default y 215 help 216 This option allows you to choose whether you want to have support 217 for so called swap devices or swap files in your kernel that are 218 used to provide more virtual memory than the actual RAM present 219 in your computer. If unsure say Y. 220 221config SYSVIPC 222 bool "System V IPC" 223 ---help--- 224 Inter Process Communication is a suite of library functions and 225 system calls which let processes (running programs) synchronize and 226 exchange information. It is generally considered to be a good thing, 227 and some programs won't run unless you say Y here. In particular, if 228 you want to run the DOS emulator dosemu under Linux (read the 229 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 230 you'll need to say Y here. 231 232 You can find documentation about IPC with "info ipc" and also in 233 section 6.4 of the Linux Programmer's Guide, available from 234 <http://www.tldp.org/guides.html>. 235 236config SYSVIPC_SYSCTL 237 bool 238 depends on SYSVIPC 239 depends on SYSCTL 240 default y 241 242config POSIX_MQUEUE 243 bool "POSIX Message Queues" 244 depends on NET 245 ---help--- 246 POSIX variant of message queues is a part of IPC. In POSIX message 247 queues every message has a priority which decides about succession 248 of receiving it by a process. If you want to compile and run 249 programs written e.g. for Solaris with use of its POSIX message 250 queues (functions mq_*) say Y here. 251 252 POSIX message queues are visible as a filesystem called 'mqueue' 253 and can be mounted somewhere if you want to do filesystem 254 operations on message queues. 255 256 If unsure, say Y. 257 258config POSIX_MQUEUE_SYSCTL 259 bool 260 depends on POSIX_MQUEUE 261 depends on SYSCTL 262 default y 263 264config CROSS_MEMORY_ATTACH 265 bool "Enable process_vm_readv/writev syscalls" 266 depends on MMU 267 default y 268 help 269 Enabling this option adds the system calls process_vm_readv and 270 process_vm_writev which allow a process with the correct privileges 271 to directly read from or write to another process' address space. 272 See the man page for more details. 273 274config FHANDLE 275 bool "open by fhandle syscalls" 276 select EXPORTFS 277 help 278 If you say Y here, a user level program will be able to map 279 file names to handle and then later use the handle for 280 different file system operations. This is useful in implementing 281 userspace file servers, which now track files using handles instead 282 of names. The handle would remain the same even if file names 283 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 284 syscalls. 285 286config USELIB 287 bool "uselib syscall" 288 default y 289 help 290 This option enables the uselib syscall, a system call used in the 291 dynamic linker from libc5 and earlier. glibc does not use this 292 system call. If you intend to run programs built on libc5 or 293 earlier, you may need to enable this syscall. Current systems 294 running glibc can safely disable this. 295 296config AUDIT 297 bool "Auditing support" 298 depends on NET 299 help 300 Enable auditing infrastructure that can be used with another 301 kernel subsystem, such as SELinux (which requires this for 302 logging of avc messages output). Does not do system-call 303 auditing without CONFIG_AUDITSYSCALL. 304 305config HAVE_ARCH_AUDITSYSCALL 306 bool 307 308config AUDITSYSCALL 309 bool "Enable system-call auditing support" 310 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL 311 default y if SECURITY_SELINUX 312 help 313 Enable low-overhead system-call auditing infrastructure that 314 can be used independently or with another kernel subsystem, 315 such as SELinux. 316 317config AUDIT_WATCH 318 def_bool y 319 depends on AUDITSYSCALL 320 select FSNOTIFY 321 322config AUDIT_TREE 323 def_bool y 324 depends on AUDITSYSCALL 325 select FSNOTIFY 326 327source "kernel/irq/Kconfig" 328source "kernel/time/Kconfig" 329 330menu "CPU/Task time and stats accounting" 331 332config VIRT_CPU_ACCOUNTING 333 bool 334 335choice 336 prompt "Cputime accounting" 337 default TICK_CPU_ACCOUNTING if !PPC64 338 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 339 340# Kind of a stub config for the pure tick based cputime accounting 341config TICK_CPU_ACCOUNTING 342 bool "Simple tick based cputime accounting" 343 depends on !S390 && !NO_HZ_FULL 344 help 345 This is the basic tick based cputime accounting that maintains 346 statistics about user, system and idle time spent on per jiffies 347 granularity. 348 349 If unsure, say Y. 350 351config VIRT_CPU_ACCOUNTING_NATIVE 352 bool "Deterministic task and CPU time accounting" 353 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 354 select VIRT_CPU_ACCOUNTING 355 help 356 Select this option to enable more accurate task and CPU time 357 accounting. This is done by reading a CPU counter on each 358 kernel entry and exit and on transitions within the kernel 359 between system, softirq and hardirq state, so there is a 360 small performance impact. In the case of s390 or IBM POWER > 5, 361 this also enables accounting of stolen time on logically-partitioned 362 systems. 363 364config VIRT_CPU_ACCOUNTING_GEN 365 bool "Full dynticks CPU time accounting" 366 depends on HAVE_CONTEXT_TRACKING 367 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN 368 select VIRT_CPU_ACCOUNTING 369 select CONTEXT_TRACKING 370 help 371 Select this option to enable task and CPU time accounting on full 372 dynticks systems. This accounting is implemented by watching every 373 kernel-user boundaries using the context tracking subsystem. 374 The accounting is thus performed at the expense of some significant 375 overhead. 376 377 For now this is only useful if you are working on the full 378 dynticks subsystem development. 379 380 If unsure, say N. 381 382config IRQ_TIME_ACCOUNTING 383 bool "Fine granularity task level IRQ time accounting" 384 depends on HAVE_IRQ_TIME_ACCOUNTING && !NO_HZ_FULL 385 help 386 Select this option to enable fine granularity task irq time 387 accounting. This is done by reading a timestamp on each 388 transitions between softirq and hardirq state, so there can be a 389 small performance impact. 390 391 If in doubt, say N here. 392 393endchoice 394 395config BSD_PROCESS_ACCT 396 bool "BSD Process Accounting" 397 help 398 If you say Y here, a user level program will be able to instruct the 399 kernel (via a special system call) to write process accounting 400 information to a file: whenever a process exits, information about 401 that process will be appended to the file by the kernel. The 402 information includes things such as creation time, owning user, 403 command name, memory usage, controlling terminal etc. (the complete 404 list is in the struct acct in <file:include/linux/acct.h>). It is 405 up to the user level program to do useful things with this 406 information. This is generally a good idea, so say Y. 407 408config BSD_PROCESS_ACCT_V3 409 bool "BSD Process Accounting version 3 file format" 410 depends on BSD_PROCESS_ACCT 411 default n 412 help 413 If you say Y here, the process accounting information is written 414 in a new file format that also logs the process IDs of each 415 process and it's parent. Note that this file format is incompatible 416 with previous v0/v1/v2 file formats, so you will need updated tools 417 for processing it. A preliminary version of these tools is available 418 at <http://www.gnu.org/software/acct/>. 419 420config TASKSTATS 421 bool "Export task/process statistics through netlink" 422 depends on NET 423 default n 424 help 425 Export selected statistics for tasks/processes through the 426 generic netlink interface. Unlike BSD process accounting, the 427 statistics are available during the lifetime of tasks/processes as 428 responses to commands. Like BSD accounting, they are sent to user 429 space on task exit. 430 431 Say N if unsure. 432 433config TASK_DELAY_ACCT 434 bool "Enable per-task delay accounting" 435 depends on TASKSTATS 436 help 437 Collect information on time spent by a task waiting for system 438 resources like cpu, synchronous block I/O completion and swapping 439 in pages. Such statistics can help in setting a task's priorities 440 relative to other tasks for cpu, io, rss limits etc. 441 442 Say N if unsure. 443 444config TASK_XACCT 445 bool "Enable extended accounting over taskstats" 446 depends on TASKSTATS 447 help 448 Collect extended task accounting data and send the data 449 to userland for processing over the taskstats interface. 450 451 Say N if unsure. 452 453config TASK_IO_ACCOUNTING 454 bool "Enable per-task storage I/O accounting" 455 depends on TASK_XACCT 456 help 457 Collect information on the number of bytes of storage I/O which this 458 task has caused. 459 460 Say N if unsure. 461 462endmenu # "CPU/Task time and stats accounting" 463 464menu "RCU Subsystem" 465 466choice 467 prompt "RCU Implementation" 468 default TREE_RCU 469 470config TREE_RCU 471 bool "Tree-based hierarchical RCU" 472 depends on !PREEMPT && SMP 473 select IRQ_WORK 474 help 475 This option selects the RCU implementation that is 476 designed for very large SMP system with hundreds or 477 thousands of CPUs. It also scales down nicely to 478 smaller systems. 479 480config TREE_PREEMPT_RCU 481 bool "Preemptible tree-based hierarchical RCU" 482 depends on PREEMPT 483 select IRQ_WORK 484 help 485 This option selects the RCU implementation that is 486 designed for very large SMP systems with hundreds or 487 thousands of CPUs, but for which real-time response 488 is also required. It also scales down nicely to 489 smaller systems. 490 491 Select this option if you are unsure. 492 493config TINY_RCU 494 bool "UP-only small-memory-footprint RCU" 495 depends on !PREEMPT && !SMP 496 help 497 This option selects the RCU implementation that is 498 designed for UP systems from which real-time response 499 is not required. This option greatly reduces the 500 memory footprint of RCU. 501 502endchoice 503 504config PREEMPT_RCU 505 def_bool TREE_PREEMPT_RCU 506 help 507 This option enables preemptible-RCU code that is common between 508 TREE_PREEMPT_RCU and, in the old days, TINY_PREEMPT_RCU. 509 510config TASKS_RCU 511 bool "Task_based RCU implementation using voluntary context switch" 512 default n 513 help 514 This option enables a task-based RCU implementation that uses 515 only voluntary context switch (not preemption!), idle, and 516 user-mode execution as quiescent states. 517 518 If unsure, say N. 519 520config RCU_STALL_COMMON 521 def_bool ( TREE_RCU || TREE_PREEMPT_RCU || RCU_TRACE ) 522 help 523 This option enables RCU CPU stall code that is common between 524 the TINY and TREE variants of RCU. The purpose is to allow 525 the tiny variants to disable RCU CPU stall warnings, while 526 making these warnings mandatory for the tree variants. 527 528config CONTEXT_TRACKING 529 bool 530 531config RCU_USER_QS 532 bool "Consider userspace as in RCU extended quiescent state" 533 depends on HAVE_CONTEXT_TRACKING && SMP 534 select CONTEXT_TRACKING 535 help 536 This option sets hooks on kernel / userspace boundaries and 537 puts RCU in extended quiescent state when the CPU runs in 538 userspace. It means that when a CPU runs in userspace, it is 539 excluded from the global RCU state machine and thus doesn't 540 try to keep the timer tick on for RCU. 541 542 Unless you want to hack and help the development of the full 543 dynticks mode, you shouldn't enable this option. It also 544 adds unnecessary overhead. 545 546 If unsure say N 547 548config CONTEXT_TRACKING_FORCE 549 bool "Force context tracking" 550 depends on CONTEXT_TRACKING 551 default y if !NO_HZ_FULL 552 help 553 The major pre-requirement for full dynticks to work is to 554 support the context tracking subsystem. But there are also 555 other dependencies to provide in order to make the full 556 dynticks working. 557 558 This option stands for testing when an arch implements the 559 context tracking backend but doesn't yet fullfill all the 560 requirements to make the full dynticks feature working. 561 Without the full dynticks, there is no way to test the support 562 for context tracking and the subsystems that rely on it: RCU 563 userspace extended quiescent state and tickless cputime 564 accounting. This option copes with the absence of the full 565 dynticks subsystem by forcing the context tracking on all 566 CPUs in the system. 567 568 Say Y only if you're working on the development of an 569 architecture backend for the context tracking. 570 571 Say N otherwise, this option brings an overhead that you 572 don't want in production. 573 574 575config RCU_FANOUT 576 int "Tree-based hierarchical RCU fanout value" 577 range 2 64 if 64BIT 578 range 2 32 if !64BIT 579 depends on TREE_RCU || TREE_PREEMPT_RCU 580 default 64 if 64BIT 581 default 32 if !64BIT 582 help 583 This option controls the fanout of hierarchical implementations 584 of RCU, allowing RCU to work efficiently on machines with 585 large numbers of CPUs. This value must be at least the fourth 586 root of NR_CPUS, which allows NR_CPUS to be insanely large. 587 The default value of RCU_FANOUT should be used for production 588 systems, but if you are stress-testing the RCU implementation 589 itself, small RCU_FANOUT values allow you to test large-system 590 code paths on small(er) systems. 591 592 Select a specific number if testing RCU itself. 593 Take the default if unsure. 594 595config RCU_FANOUT_LEAF 596 int "Tree-based hierarchical RCU leaf-level fanout value" 597 range 2 RCU_FANOUT if 64BIT 598 range 2 RCU_FANOUT if !64BIT 599 depends on TREE_RCU || TREE_PREEMPT_RCU 600 default 16 601 help 602 This option controls the leaf-level fanout of hierarchical 603 implementations of RCU, and allows trading off cache misses 604 against lock contention. Systems that synchronize their 605 scheduling-clock interrupts for energy-efficiency reasons will 606 want the default because the smaller leaf-level fanout keeps 607 lock contention levels acceptably low. Very large systems 608 (hundreds or thousands of CPUs) will instead want to set this 609 value to the maximum value possible in order to reduce the 610 number of cache misses incurred during RCU's grace-period 611 initialization. These systems tend to run CPU-bound, and thus 612 are not helped by synchronized interrupts, and thus tend to 613 skew them, which reduces lock contention enough that large 614 leaf-level fanouts work well. 615 616 Select a specific number if testing RCU itself. 617 618 Select the maximum permissible value for large systems. 619 620 Take the default if unsure. 621 622config RCU_FANOUT_EXACT 623 bool "Disable tree-based hierarchical RCU auto-balancing" 624 depends on TREE_RCU || TREE_PREEMPT_RCU 625 default n 626 help 627 This option forces use of the exact RCU_FANOUT value specified, 628 regardless of imbalances in the hierarchy. This is useful for 629 testing RCU itself, and might one day be useful on systems with 630 strong NUMA behavior. 631 632 Without RCU_FANOUT_EXACT, the code will balance the hierarchy. 633 634 Say N if unsure. 635 636config RCU_FAST_NO_HZ 637 bool "Accelerate last non-dyntick-idle CPU's grace periods" 638 depends on NO_HZ_COMMON && SMP 639 default n 640 help 641 This option permits CPUs to enter dynticks-idle state even if 642 they have RCU callbacks queued, and prevents RCU from waking 643 these CPUs up more than roughly once every four jiffies (by 644 default, you can adjust this using the rcutree.rcu_idle_gp_delay 645 parameter), thus improving energy efficiency. On the other 646 hand, this option increases the duration of RCU grace periods, 647 for example, slowing down synchronize_rcu(). 648 649 Say Y if energy efficiency is critically important, and you 650 don't care about increased grace-period durations. 651 652 Say N if you are unsure. 653 654config TREE_RCU_TRACE 655 def_bool RCU_TRACE && ( TREE_RCU || TREE_PREEMPT_RCU ) 656 select DEBUG_FS 657 help 658 This option provides tracing for the TREE_RCU and 659 TREE_PREEMPT_RCU implementations, permitting Makefile to 660 trivially select kernel/rcutree_trace.c. 661 662config RCU_BOOST 663 bool "Enable RCU priority boosting" 664 depends on RT_MUTEXES && PREEMPT_RCU 665 default n 666 help 667 This option boosts the priority of preempted RCU readers that 668 block the current preemptible RCU grace period for too long. 669 This option also prevents heavy loads from blocking RCU 670 callback invocation for all flavors of RCU. 671 672 Say Y here if you are working with real-time apps or heavy loads 673 Say N here if you are unsure. 674 675config RCU_BOOST_PRIO 676 int "Real-time priority to boost RCU readers to" 677 range 1 99 678 depends on RCU_BOOST 679 default 1 680 help 681 This option specifies the real-time priority to which long-term 682 preempted RCU readers are to be boosted. If you are working 683 with a real-time application that has one or more CPU-bound 684 threads running at a real-time priority level, you should set 685 RCU_BOOST_PRIO to a priority higher then the highest-priority 686 real-time CPU-bound thread. The default RCU_BOOST_PRIO value 687 of 1 is appropriate in the common case, which is real-time 688 applications that do not have any CPU-bound threads. 689 690 Some real-time applications might not have a single real-time 691 thread that saturates a given CPU, but instead might have 692 multiple real-time threads that, taken together, fully utilize 693 that CPU. In this case, you should set RCU_BOOST_PRIO to 694 a priority higher than the lowest-priority thread that is 695 conspiring to prevent the CPU from running any non-real-time 696 tasks. For example, if one thread at priority 10 and another 697 thread at priority 5 are between themselves fully consuming 698 the CPU time on a given CPU, then RCU_BOOST_PRIO should be 699 set to priority 6 or higher. 700 701 Specify the real-time priority, or take the default if unsure. 702 703config RCU_BOOST_DELAY 704 int "Milliseconds to delay boosting after RCU grace-period start" 705 range 0 3000 706 depends on RCU_BOOST 707 default 500 708 help 709 This option specifies the time to wait after the beginning of 710 a given grace period before priority-boosting preempted RCU 711 readers blocking that grace period. Note that any RCU reader 712 blocking an expedited RCU grace period is boosted immediately. 713 714 Accept the default if unsure. 715 716config RCU_NOCB_CPU 717 bool "Offload RCU callback processing from boot-selected CPUs" 718 depends on TREE_RCU || TREE_PREEMPT_RCU 719 default n 720 help 721 Use this option to reduce OS jitter for aggressive HPC or 722 real-time workloads. It can also be used to offload RCU 723 callback invocation to energy-efficient CPUs in battery-powered 724 asymmetric multiprocessors. 725 726 This option offloads callback invocation from the set of 727 CPUs specified at boot time by the rcu_nocbs parameter. 728 For each such CPU, a kthread ("rcuox/N") will be created to 729 invoke callbacks, where the "N" is the CPU being offloaded, 730 and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and 731 "s" for RCU-sched. Nothing prevents this kthread from running 732 on the specified CPUs, but (1) the kthreads may be preempted 733 between each callback, and (2) affinity or cgroups can be used 734 to force the kthreads to run on whatever set of CPUs is desired. 735 736 Say Y here if you want to help to debug reduced OS jitter. 737 Say N here if you are unsure. 738 739choice 740 prompt "Build-forced no-CBs CPUs" 741 default RCU_NOCB_CPU_NONE 742 help 743 This option allows no-CBs CPUs (whose RCU callbacks are invoked 744 from kthreads rather than from softirq context) to be specified 745 at build time. Additional no-CBs CPUs may be specified by 746 the rcu_nocbs= boot parameter. 747 748config RCU_NOCB_CPU_NONE 749 bool "No build_forced no-CBs CPUs" 750 depends on RCU_NOCB_CPU 751 help 752 This option does not force any of the CPUs to be no-CBs CPUs. 753 Only CPUs designated by the rcu_nocbs= boot parameter will be 754 no-CBs CPUs, whose RCU callbacks will be invoked by per-CPU 755 kthreads whose names begin with "rcuo". All other CPUs will 756 invoke their own RCU callbacks in softirq context. 757 758 Select this option if you want to choose no-CBs CPUs at 759 boot time, for example, to allow testing of different no-CBs 760 configurations without having to rebuild the kernel each time. 761 762config RCU_NOCB_CPU_ZERO 763 bool "CPU 0 is a build_forced no-CBs CPU" 764 depends on RCU_NOCB_CPU 765 help 766 This option forces CPU 0 to be a no-CBs CPU, so that its RCU 767 callbacks are invoked by a per-CPU kthread whose name begins 768 with "rcuo". Additional CPUs may be designated as no-CBs 769 CPUs using the rcu_nocbs= boot parameter will be no-CBs CPUs. 770 All other CPUs will invoke their own RCU callbacks in softirq 771 context. 772 773 Select this if CPU 0 needs to be a no-CBs CPU for real-time 774 or energy-efficiency reasons, but the real reason it exists 775 is to ensure that randconfig testing covers mixed systems. 776 777config RCU_NOCB_CPU_ALL 778 bool "All CPUs are build_forced no-CBs CPUs" 779 depends on RCU_NOCB_CPU 780 help 781 This option forces all CPUs to be no-CBs CPUs. The rcu_nocbs= 782 boot parameter will be ignored. All CPUs' RCU callbacks will 783 be executed in the context of per-CPU rcuo kthreads created for 784 this purpose. Assuming that the kthreads whose names start with 785 "rcuo" are bound to "housekeeping" CPUs, this reduces OS jitter 786 on the remaining CPUs, but might decrease memory locality during 787 RCU-callback invocation, thus potentially degrading throughput. 788 789 Select this if all CPUs need to be no-CBs CPUs for real-time 790 or energy-efficiency reasons. 791 792endchoice 793 794endmenu # "RCU Subsystem" 795 796config BUILD_BIN2C 797 bool 798 default n 799 800config IKCONFIG 801 tristate "Kernel .config support" 802 select BUILD_BIN2C 803 ---help--- 804 This option enables the complete Linux kernel ".config" file 805 contents to be saved in the kernel. It provides documentation 806 of which kernel options are used in a running kernel or in an 807 on-disk kernel. This information can be extracted from the kernel 808 image file with the script scripts/extract-ikconfig and used as 809 input to rebuild the current kernel or to build another kernel. 810 It can also be extracted from a running kernel by reading 811 /proc/config.gz if enabled (below). 812 813config IKCONFIG_PROC 814 bool "Enable access to .config through /proc/config.gz" 815 depends on IKCONFIG && PROC_FS 816 ---help--- 817 This option enables access to the kernel configuration file 818 through /proc/config.gz. 819 820config LOG_BUF_SHIFT 821 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 822 range 12 21 823 default 17 824 depends on PRINTK 825 help 826 Select the minimal kernel log buffer size as a power of 2. 827 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config 828 parameter, see below. Any higher size also might be forced 829 by "log_buf_len" boot parameter. 830 831 Examples: 832 17 => 128 KB 833 16 => 64 KB 834 15 => 32 KB 835 14 => 16 KB 836 13 => 8 KB 837 12 => 4 KB 838 839config LOG_CPU_MAX_BUF_SHIFT 840 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" 841 depends on SMP 842 range 0 21 843 default 12 if !BASE_SMALL 844 default 0 if BASE_SMALL 845 depends on PRINTK 846 help 847 This option allows to increase the default ring buffer size 848 according to the number of CPUs. The value defines the contribution 849 of each CPU as a power of 2. The used space is typically only few 850 lines however it might be much more when problems are reported, 851 e.g. backtraces. 852 853 The increased size means that a new buffer has to be allocated and 854 the original static one is unused. It makes sense only on systems 855 with more CPUs. Therefore this value is used only when the sum of 856 contributions is greater than the half of the default kernel ring 857 buffer as defined by LOG_BUF_SHIFT. The default values are set 858 so that more than 64 CPUs are needed to trigger the allocation. 859 860 Also this option is ignored when "log_buf_len" kernel parameter is 861 used as it forces an exact (power of two) size of the ring buffer. 862 863 The number of possible CPUs is used for this computation ignoring 864 hotplugging making the compuation optimal for the the worst case 865 scenerio while allowing a simple algorithm to be used from bootup. 866 867 Examples shift values and their meaning: 868 17 => 128 KB for each CPU 869 16 => 64 KB for each CPU 870 15 => 32 KB for each CPU 871 14 => 16 KB for each CPU 872 13 => 8 KB for each CPU 873 12 => 4 KB for each CPU 874 875# 876# Architectures with an unreliable sched_clock() should select this: 877# 878config HAVE_UNSTABLE_SCHED_CLOCK 879 bool 880 881config GENERIC_SCHED_CLOCK 882 bool 883 884# 885# For architectures that want to enable the support for NUMA-affine scheduler 886# balancing logic: 887# 888config ARCH_SUPPORTS_NUMA_BALANCING 889 bool 890 891# 892# For architectures that know their GCC __int128 support is sound 893# 894config ARCH_SUPPORTS_INT128 895 bool 896 897# For architectures that (ab)use NUMA to represent different memory regions 898# all cpu-local but of different latencies, such as SuperH. 899# 900config ARCH_WANT_NUMA_VARIABLE_LOCALITY 901 bool 902 903config NUMA_BALANCING_DEFAULT_ENABLED 904 bool "Automatically enable NUMA aware memory/task placement" 905 default y 906 depends on NUMA_BALANCING 907 help 908 If set, automatic NUMA balancing will be enabled if running on a NUMA 909 machine. 910 911config NUMA_BALANCING 912 bool "Memory placement aware NUMA scheduler" 913 depends on ARCH_SUPPORTS_NUMA_BALANCING 914 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 915 depends on SMP && NUMA && MIGRATION 916 help 917 This option adds support for automatic NUMA aware memory/task placement. 918 The mechanism is quite primitive and is based on migrating memory when 919 it has references to the node the task is running on. 920 921 This system will be inactive on UMA systems. 922 923menuconfig CGROUPS 924 boolean "Control Group support" 925 select KERNFS 926 help 927 This option adds support for grouping sets of processes together, for 928 use with process control subsystems such as Cpusets, CFS, memory 929 controls or device isolation. 930 See 931 - Documentation/scheduler/sched-design-CFS.txt (CFS) 932 - Documentation/cgroups/ (features for grouping, isolation 933 and resource control) 934 935 Say N if unsure. 936 937if CGROUPS 938 939config CGROUP_DEBUG 940 bool "Example debug cgroup subsystem" 941 default n 942 help 943 This option enables a simple cgroup subsystem that 944 exports useful debugging information about the cgroups 945 framework. 946 947 Say N if unsure. 948 949config CGROUP_FREEZER 950 bool "Freezer cgroup subsystem" 951 help 952 Provides a way to freeze and unfreeze all tasks in a 953 cgroup. 954 955config CGROUP_DEVICE 956 bool "Device controller for cgroups" 957 help 958 Provides a cgroup implementing whitelists for devices which 959 a process in the cgroup can mknod or open. 960 961config CPUSETS 962 bool "Cpuset support" 963 help 964 This option will let you create and manage CPUSETs which 965 allow dynamically partitioning a system into sets of CPUs and 966 Memory Nodes and assigning tasks to run only within those sets. 967 This is primarily useful on large SMP or NUMA systems. 968 969 Say N if unsure. 970 971config PROC_PID_CPUSET 972 bool "Include legacy /proc/<pid>/cpuset file" 973 depends on CPUSETS 974 default y 975 976config CGROUP_CPUACCT 977 bool "Simple CPU accounting cgroup subsystem" 978 help 979 Provides a simple Resource Controller for monitoring the 980 total CPU consumed by the tasks in a cgroup. 981 982config RESOURCE_COUNTERS 983 bool "Resource counters" 984 help 985 This option enables controller independent resource accounting 986 infrastructure that works with cgroups. 987 988config MEMCG 989 bool "Memory Resource Controller for Control Groups" 990 depends on RESOURCE_COUNTERS 991 select EVENTFD 992 help 993 Provides a memory resource controller that manages both anonymous 994 memory and page cache. (See Documentation/cgroups/memory.txt) 995 996 Note that setting this option increases fixed memory overhead 997 associated with each page of memory in the system. By this, 998 8(16)bytes/PAGE_SIZE on 32(64)bit system will be occupied by memory 999 usage tracking struct at boot. Total amount of this is printed out 1000 at boot. 1001 1002 Only enable when you're ok with these trade offs and really 1003 sure you need the memory resource controller. Even when you enable 1004 this, you can set "cgroup_disable=memory" at your boot option to 1005 disable memory resource controller and you can avoid overheads. 1006 (and lose benefits of memory resource controller) 1007 1008config MEMCG_SWAP 1009 bool "Memory Resource Controller Swap Extension" 1010 depends on MEMCG && SWAP 1011 help 1012 Add swap management feature to memory resource controller. When you 1013 enable this, you can limit mem+swap usage per cgroup. In other words, 1014 when you disable this, memory resource controller has no cares to 1015 usage of swap...a process can exhaust all of the swap. This extension 1016 is useful when you want to avoid exhaustion swap but this itself 1017 adds more overheads and consumes memory for remembering information. 1018 Especially if you use 32bit system or small memory system, please 1019 be careful about enabling this. When memory resource controller 1020 is disabled by boot option, this will be automatically disabled and 1021 there will be no overhead from this. Even when you set this config=y, 1022 if boot option "swapaccount=0" is set, swap will not be accounted. 1023 Now, memory usage of swap_cgroup is 2 bytes per entry. If swap page 1024 size is 4096bytes, 512k per 1Gbytes of swap. 1025config MEMCG_SWAP_ENABLED 1026 bool "Memory Resource Controller Swap Extension enabled by default" 1027 depends on MEMCG_SWAP 1028 default y 1029 help 1030 Memory Resource Controller Swap Extension comes with its price in 1031 a bigger memory consumption. General purpose distribution kernels 1032 which want to enable the feature but keep it disabled by default 1033 and let the user enable it by swapaccount=1 boot command line 1034 parameter should have this option unselected. 1035 For those who want to have the feature enabled by default should 1036 select this option (if, for some reason, they need to disable it 1037 then swapaccount=0 does the trick). 1038config MEMCG_KMEM 1039 bool "Memory Resource Controller Kernel Memory accounting" 1040 depends on MEMCG 1041 depends on SLUB || SLAB 1042 help 1043 The Kernel Memory extension for Memory Resource Controller can limit 1044 the amount of memory used by kernel objects in the system. Those are 1045 fundamentally different from the entities handled by the standard 1046 Memory Controller, which are page-based, and can be swapped. Users of 1047 the kmem extension can use it to guarantee that no group of processes 1048 will ever exhaust kernel resources alone. 1049 1050 WARNING: Current implementation lacks reclaim support. That means 1051 allocation attempts will fail when close to the limit even if there 1052 are plenty of kmem available for reclaim. That makes this option 1053 unusable in real life so DO NOT SELECT IT unless for development 1054 purposes. 1055 1056config CGROUP_HUGETLB 1057 bool "HugeTLB Resource Controller for Control Groups" 1058 depends on RESOURCE_COUNTERS && HUGETLB_PAGE 1059 default n 1060 help 1061 Provides a cgroup Resource Controller for HugeTLB pages. 1062 When you enable this, you can put a per cgroup limit on HugeTLB usage. 1063 The limit is enforced during page fault. Since HugeTLB doesn't 1064 support page reclaim, enforcing the limit at page fault time implies 1065 that, the application will get SIGBUS signal if it tries to access 1066 HugeTLB pages beyond its limit. This requires the application to know 1067 beforehand how much HugeTLB pages it would require for its use. The 1068 control group is tracked in the third page lru pointer. This means 1069 that we cannot use the controller with huge page less than 3 pages. 1070 1071config CGROUP_PERF 1072 bool "Enable perf_event per-cpu per-container group (cgroup) monitoring" 1073 depends on PERF_EVENTS && CGROUPS 1074 help 1075 This option extends the per-cpu mode to restrict monitoring to 1076 threads which belong to the cgroup specified and run on the 1077 designated cpu. 1078 1079 Say N if unsure. 1080 1081menuconfig CGROUP_SCHED 1082 bool "Group CPU scheduler" 1083 default n 1084 help 1085 This feature lets CPU scheduler recognize task groups and control CPU 1086 bandwidth allocation to such task groups. It uses cgroups to group 1087 tasks. 1088 1089if CGROUP_SCHED 1090config FAIR_GROUP_SCHED 1091 bool "Group scheduling for SCHED_OTHER" 1092 depends on CGROUP_SCHED 1093 default CGROUP_SCHED 1094 1095config CFS_BANDWIDTH 1096 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 1097 depends on FAIR_GROUP_SCHED 1098 default n 1099 help 1100 This option allows users to define CPU bandwidth rates (limits) for 1101 tasks running within the fair group scheduler. Groups with no limit 1102 set are considered to be unconstrained and will run with no 1103 restriction. 1104 See tip/Documentation/scheduler/sched-bwc.txt for more information. 1105 1106config RT_GROUP_SCHED 1107 bool "Group scheduling for SCHED_RR/FIFO" 1108 depends on CGROUP_SCHED 1109 default n 1110 help 1111 This feature lets you explicitly allocate real CPU bandwidth 1112 to task groups. If enabled, it will also make it impossible to 1113 schedule realtime tasks for non-root users until you allocate 1114 realtime bandwidth for them. 1115 See Documentation/scheduler/sched-rt-group.txt for more information. 1116 1117endif #CGROUP_SCHED 1118 1119config BLK_CGROUP 1120 bool "Block IO controller" 1121 depends on BLOCK 1122 default n 1123 ---help--- 1124 Generic block IO controller cgroup interface. This is the common 1125 cgroup interface which should be used by various IO controlling 1126 policies. 1127 1128 Currently, CFQ IO scheduler uses it to recognize task groups and 1129 control disk bandwidth allocation (proportional time slice allocation) 1130 to such task groups. It is also used by bio throttling logic in 1131 block layer to implement upper limit in IO rates on a device. 1132 1133 This option only enables generic Block IO controller infrastructure. 1134 One needs to also enable actual IO controlling logic/policy. For 1135 enabling proportional weight division of disk bandwidth in CFQ, set 1136 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 1137 CONFIG_BLK_DEV_THROTTLING=y. 1138 1139 See Documentation/cgroups/blkio-controller.txt for more information. 1140 1141config DEBUG_BLK_CGROUP 1142 bool "Enable Block IO controller debugging" 1143 depends on BLK_CGROUP 1144 default n 1145 ---help--- 1146 Enable some debugging help. Currently it exports additional stat 1147 files in a cgroup which can be useful for debugging. 1148 1149endif # CGROUPS 1150 1151config CHECKPOINT_RESTORE 1152 bool "Checkpoint/restore support" if EXPERT 1153 default n 1154 help 1155 Enables additional kernel features in a sake of checkpoint/restore. 1156 In particular it adds auxiliary prctl codes to setup process text, 1157 data and heap segment sizes, and a few additional /proc filesystem 1158 entries. 1159 1160 If unsure, say N here. 1161 1162menuconfig NAMESPACES 1163 bool "Namespaces support" if EXPERT 1164 default !EXPERT 1165 help 1166 Provides the way to make tasks work with different objects using 1167 the same id. For example same IPC id may refer to different objects 1168 or same user id or pid may refer to different tasks when used in 1169 different namespaces. 1170 1171if NAMESPACES 1172 1173config UTS_NS 1174 bool "UTS namespace" 1175 default y 1176 help 1177 In this namespace tasks see different info provided with the 1178 uname() system call 1179 1180config IPC_NS 1181 bool "IPC namespace" 1182 depends on (SYSVIPC || POSIX_MQUEUE) 1183 default y 1184 help 1185 In this namespace tasks work with IPC ids which correspond to 1186 different IPC objects in different namespaces. 1187 1188config USER_NS 1189 bool "User namespace" 1190 default n 1191 help 1192 This allows containers, i.e. vservers, to use user namespaces 1193 to provide different user info for different servers. 1194 1195 When user namespaces are enabled in the kernel it is 1196 recommended that the MEMCG and MEMCG_KMEM options also be 1197 enabled and that user-space use the memory control groups to 1198 limit the amount of memory a memory unprivileged users can 1199 use. 1200 1201 If unsure, say N. 1202 1203config PID_NS 1204 bool "PID Namespaces" 1205 default y 1206 help 1207 Support process id namespaces. This allows having multiple 1208 processes with the same pid as long as they are in different 1209 pid namespaces. This is a building block of containers. 1210 1211config NET_NS 1212 bool "Network namespace" 1213 depends on NET 1214 default y 1215 help 1216 Allow user space to create what appear to be multiple instances 1217 of the network stack. 1218 1219endif # NAMESPACES 1220 1221config SCHED_AUTOGROUP 1222 bool "Automatic process group scheduling" 1223 select CGROUPS 1224 select CGROUP_SCHED 1225 select FAIR_GROUP_SCHED 1226 help 1227 This option optimizes the scheduler for common desktop workloads by 1228 automatically creating and populating task groups. This separation 1229 of workloads isolates aggressive CPU burners (like build jobs) from 1230 desktop applications. Task group autogeneration is currently based 1231 upon task session. 1232 1233config SYSFS_DEPRECATED 1234 bool "Enable deprecated sysfs features to support old userspace tools" 1235 depends on SYSFS 1236 default n 1237 help 1238 This option adds code that switches the layout of the "block" class 1239 devices, to not show up in /sys/class/block/, but only in 1240 /sys/block/. 1241 1242 This switch is only active when the sysfs.deprecated=1 boot option is 1243 passed or the SYSFS_DEPRECATED_V2 option is set. 1244 1245 This option allows new kernels to run on old distributions and tools, 1246 which might get confused by /sys/class/block/. Since 2007/2008 all 1247 major distributions and tools handle this just fine. 1248 1249 Recent distributions and userspace tools after 2009/2010 depend on 1250 the existence of /sys/class/block/, and will not work with this 1251 option enabled. 1252 1253 Only if you are using a new kernel on an old distribution, you might 1254 need to say Y here. 1255 1256config SYSFS_DEPRECATED_V2 1257 bool "Enable deprecated sysfs features by default" 1258 default n 1259 depends on SYSFS 1260 depends on SYSFS_DEPRECATED 1261 help 1262 Enable deprecated sysfs by default. 1263 1264 See the CONFIG_SYSFS_DEPRECATED option for more details about this 1265 option. 1266 1267 Only if you are using a new kernel on an old distribution, you might 1268 need to say Y here. Even then, odds are you would not need it 1269 enabled, you can always pass the boot option if absolutely necessary. 1270 1271config RELAY 1272 bool "Kernel->user space relay support (formerly relayfs)" 1273 help 1274 This option enables support for relay interface support in 1275 certain file systems (such as debugfs). 1276 It is designed to provide an efficient mechanism for tools and 1277 facilities to relay large amounts of data from kernel space to 1278 user space. 1279 1280 If unsure, say N. 1281 1282config BLK_DEV_INITRD 1283 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 1284 depends on BROKEN || !FRV 1285 help 1286 The initial RAM filesystem is a ramfs which is loaded by the 1287 boot loader (loadlin or lilo) and that is mounted as root 1288 before the normal boot procedure. It is typically used to 1289 load modules needed to mount the "real" root file system, 1290 etc. See <file:Documentation/initrd.txt> for details. 1291 1292 If RAM disk support (BLK_DEV_RAM) is also included, this 1293 also enables initial RAM disk (initrd) support and adds 1294 15 Kbytes (more on some other architectures) to the kernel size. 1295 1296 If unsure say Y. 1297 1298if BLK_DEV_INITRD 1299 1300source "usr/Kconfig" 1301 1302endif 1303 1304config CC_OPTIMIZE_FOR_SIZE 1305 bool "Optimize for size" 1306 help 1307 Enabling this option will pass "-Os" instead of "-O2" to gcc 1308 resulting in a smaller kernel. 1309 1310 If unsure, say N. 1311 1312config SYSCTL 1313 bool 1314 1315config ANON_INODES 1316 bool 1317 1318config HAVE_UID16 1319 bool 1320 1321config SYSCTL_EXCEPTION_TRACE 1322 bool 1323 help 1324 Enable support for /proc/sys/debug/exception-trace. 1325 1326config SYSCTL_ARCH_UNALIGN_NO_WARN 1327 bool 1328 help 1329 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1330 Allows arch to define/use @no_unaligned_warning to possibly warn 1331 about unaligned access emulation going on under the hood. 1332 1333config SYSCTL_ARCH_UNALIGN_ALLOW 1334 bool 1335 help 1336 Enable support for /proc/sys/kernel/unaligned-trap 1337 Allows arches to define/use @unaligned_enabled to runtime toggle 1338 the unaligned access emulation. 1339 see arch/parisc/kernel/unaligned.c for reference 1340 1341config HAVE_PCSPKR_PLATFORM 1342 bool 1343 1344# interpreter that classic socket filters depend on 1345config BPF 1346 bool 1347 1348menuconfig EXPERT 1349 bool "Configure standard kernel features (expert users)" 1350 # Unhide debug options, to make the on-by-default options visible 1351 select DEBUG_KERNEL 1352 help 1353 This option allows certain base kernel options and settings 1354 to be disabled or tweaked. This is for specialized 1355 environments which can tolerate a "non-standard" kernel. 1356 Only use this if you really know what you are doing. 1357 1358config UID16 1359 bool "Enable 16-bit UID system calls" if EXPERT 1360 depends on HAVE_UID16 1361 default y 1362 help 1363 This enables the legacy 16-bit UID syscall wrappers. 1364 1365config SGETMASK_SYSCALL 1366 bool "sgetmask/ssetmask syscalls support" if EXPERT 1367 def_bool PARISC || MN10300 || BLACKFIN || M68K || PPC || MIPS || X86 || SPARC || CRIS || MICROBLAZE || SUPERH 1368 ---help--- 1369 sys_sgetmask and sys_ssetmask are obsolete system calls 1370 no longer supported in libc but still enabled by default in some 1371 architectures. 1372 1373 If unsure, leave the default option here. 1374 1375config SYSFS_SYSCALL 1376 bool "Sysfs syscall support" if EXPERT 1377 default y 1378 ---help--- 1379 sys_sysfs is an obsolete system call no longer supported in libc. 1380 Note that disabling this option is more secure but might break 1381 compatibility with some systems. 1382 1383 If unsure say Y here. 1384 1385config SYSCTL_SYSCALL 1386 bool "Sysctl syscall support" if EXPERT 1387 depends on PROC_SYSCTL 1388 default n 1389 select SYSCTL 1390 ---help--- 1391 sys_sysctl uses binary paths that have been found challenging 1392 to properly maintain and use. The interface in /proc/sys 1393 using paths with ascii names is now the primary path to this 1394 information. 1395 1396 Almost nothing using the binary sysctl interface so if you are 1397 trying to save some space it is probably safe to disable this, 1398 making your kernel marginally smaller. 1399 1400 If unsure say N here. 1401 1402config KALLSYMS 1403 bool "Load all symbols for debugging/ksymoops" if EXPERT 1404 default y 1405 help 1406 Say Y here to let the kernel print out symbolic crash information and 1407 symbolic stack backtraces. This increases the size of the kernel 1408 somewhat, as all symbols have to be loaded into the kernel image. 1409 1410config KALLSYMS_ALL 1411 bool "Include all symbols in kallsyms" 1412 depends on DEBUG_KERNEL && KALLSYMS 1413 help 1414 Normally kallsyms only contains the symbols of functions for nicer 1415 OOPS messages and backtraces (i.e., symbols from the text and inittext 1416 sections). This is sufficient for most cases. And only in very rare 1417 cases (e.g., when a debugger is used) all symbols are required (e.g., 1418 names of variables from the data sections, etc). 1419 1420 This option makes sure that all symbols are loaded into the kernel 1421 image (i.e., symbols from all sections) in cost of increased kernel 1422 size (depending on the kernel configuration, it may be 300KiB or 1423 something like this). 1424 1425 Say N unless you really need all symbols. 1426 1427config PRINTK 1428 default y 1429 bool "Enable support for printk" if EXPERT 1430 select IRQ_WORK 1431 help 1432 This option enables normal printk support. Removing it 1433 eliminates most of the message strings from the kernel image 1434 and makes the kernel more or less silent. As this makes it 1435 very difficult to diagnose system problems, saying N here is 1436 strongly discouraged. 1437 1438config BUG 1439 bool "BUG() support" if EXPERT 1440 default y 1441 help 1442 Disabling this option eliminates support for BUG and WARN, reducing 1443 the size of your kernel image and potentially quietly ignoring 1444 numerous fatal conditions. You should only consider disabling this 1445 option for embedded systems with no facilities for reporting errors. 1446 Just say Y. 1447 1448config ELF_CORE 1449 depends on COREDUMP 1450 default y 1451 bool "Enable ELF core dumps" if EXPERT 1452 help 1453 Enable support for generating core dumps. Disabling saves about 4k. 1454 1455 1456config PCSPKR_PLATFORM 1457 bool "Enable PC-Speaker support" if EXPERT 1458 depends on HAVE_PCSPKR_PLATFORM 1459 select I8253_LOCK 1460 default y 1461 help 1462 This option allows to disable the internal PC-Speaker 1463 support, saving some memory. 1464 1465config BASE_FULL 1466 default y 1467 bool "Enable full-sized data structures for core" if EXPERT 1468 help 1469 Disabling this option reduces the size of miscellaneous core 1470 kernel data structures. This saves memory on small machines, 1471 but may reduce performance. 1472 1473config FUTEX 1474 bool "Enable futex support" if EXPERT 1475 default y 1476 select RT_MUTEXES 1477 help 1478 Disabling this option will cause the kernel to be built without 1479 support for "fast userspace mutexes". The resulting kernel may not 1480 run glibc-based applications correctly. 1481 1482config HAVE_FUTEX_CMPXCHG 1483 bool 1484 depends on FUTEX 1485 help 1486 Architectures should select this if futex_atomic_cmpxchg_inatomic() 1487 is implemented and always working. This removes a couple of runtime 1488 checks. 1489 1490config EPOLL 1491 bool "Enable eventpoll support" if EXPERT 1492 default y 1493 select ANON_INODES 1494 help 1495 Disabling this option will cause the kernel to be built without 1496 support for epoll family of system calls. 1497 1498config SIGNALFD 1499 bool "Enable signalfd() system call" if EXPERT 1500 select ANON_INODES 1501 default y 1502 help 1503 Enable the signalfd() system call that allows to receive signals 1504 on a file descriptor. 1505 1506 If unsure, say Y. 1507 1508config TIMERFD 1509 bool "Enable timerfd() system call" if EXPERT 1510 select ANON_INODES 1511 default y 1512 help 1513 Enable the timerfd() system call that allows to receive timer 1514 events on a file descriptor. 1515 1516 If unsure, say Y. 1517 1518config EVENTFD 1519 bool "Enable eventfd() system call" if EXPERT 1520 select ANON_INODES 1521 default y 1522 help 1523 Enable the eventfd() system call that allows to receive both 1524 kernel notification (ie. KAIO) or userspace notifications. 1525 1526 If unsure, say Y. 1527 1528# syscall, maps, verifier 1529config BPF_SYSCALL 1530 bool "Enable bpf() system call" if EXPERT 1531 select ANON_INODES 1532 select BPF 1533 default n 1534 help 1535 Enable the bpf() system call that allows to manipulate eBPF 1536 programs and maps via file descriptors. 1537 1538config SHMEM 1539 bool "Use full shmem filesystem" if EXPERT 1540 default y 1541 depends on MMU 1542 help 1543 The shmem is an internal filesystem used to manage shared memory. 1544 It is backed by swap and manages resource limits. It is also exported 1545 to userspace as tmpfs if TMPFS is enabled. Disabling this 1546 option replaces shmem and tmpfs with the much simpler ramfs code, 1547 which may be appropriate on small systems without swap. 1548 1549config AIO 1550 bool "Enable AIO support" if EXPERT 1551 default y 1552 help 1553 This option enables POSIX asynchronous I/O which may by used 1554 by some high performance threaded applications. Disabling 1555 this option saves about 7k. 1556 1557config ADVISE_SYSCALLS 1558 bool "Enable madvise/fadvise syscalls" if EXPERT 1559 default y 1560 help 1561 This option enables the madvise and fadvise syscalls, used by 1562 applications to advise the kernel about their future memory or file 1563 usage, improving performance. If building an embedded system where no 1564 applications use these syscalls, you can disable this option to save 1565 space. 1566 1567config PCI_QUIRKS 1568 default y 1569 bool "Enable PCI quirk workarounds" if EXPERT 1570 depends on PCI 1571 help 1572 This enables workarounds for various PCI chipset 1573 bugs/quirks. Disable this only if your target machine is 1574 unaffected by PCI quirks. 1575 1576config EMBEDDED 1577 bool "Embedded system" 1578 option allnoconfig_y 1579 select EXPERT 1580 help 1581 This option should be enabled if compiling the kernel for 1582 an embedded system so certain expert options are available 1583 for configuration. 1584 1585config HAVE_PERF_EVENTS 1586 bool 1587 help 1588 See tools/perf/design.txt for details. 1589 1590config PERF_USE_VMALLOC 1591 bool 1592 help 1593 See tools/perf/design.txt for details 1594 1595menu "Kernel Performance Events And Counters" 1596 1597config PERF_EVENTS 1598 bool "Kernel performance events and counters" 1599 default y if PROFILING 1600 depends on HAVE_PERF_EVENTS 1601 select ANON_INODES 1602 select IRQ_WORK 1603 help 1604 Enable kernel support for various performance events provided 1605 by software and hardware. 1606 1607 Software events are supported either built-in or via the 1608 use of generic tracepoints. 1609 1610 Most modern CPUs support performance events via performance 1611 counter registers. These registers count the number of certain 1612 types of hw events: such as instructions executed, cachemisses 1613 suffered, or branches mis-predicted - without slowing down the 1614 kernel or applications. These registers can also trigger interrupts 1615 when a threshold number of events have passed - and can thus be 1616 used to profile the code that runs on that CPU. 1617 1618 The Linux Performance Event subsystem provides an abstraction of 1619 these software and hardware event capabilities, available via a 1620 system call and used by the "perf" utility in tools/perf/. It 1621 provides per task and per CPU counters, and it provides event 1622 capabilities on top of those. 1623 1624 Say Y if unsure. 1625 1626config DEBUG_PERF_USE_VMALLOC 1627 default n 1628 bool "Debug: use vmalloc to back perf mmap() buffers" 1629 depends on PERF_EVENTS && DEBUG_KERNEL 1630 select PERF_USE_VMALLOC 1631 help 1632 Use vmalloc memory to back perf mmap() buffers. 1633 1634 Mostly useful for debugging the vmalloc code on platforms 1635 that don't require it. 1636 1637 Say N if unsure. 1638 1639endmenu 1640 1641config VM_EVENT_COUNTERS 1642 default y 1643 bool "Enable VM event counters for /proc/vmstat" if EXPERT 1644 help 1645 VM event counters are needed for event counts to be shown. 1646 This option allows the disabling of the VM event counters 1647 on EXPERT systems. /proc/vmstat will only show page counts 1648 if VM event counters are disabled. 1649 1650config SLUB_DEBUG 1651 default y 1652 bool "Enable SLUB debugging support" if EXPERT 1653 depends on SLUB && SYSFS 1654 help 1655 SLUB has extensive debug support features. Disabling these can 1656 result in significant savings in code size. This also disables 1657 SLUB sysfs support. /sys/slab will not exist and there will be 1658 no support for cache validation etc. 1659 1660config COMPAT_BRK 1661 bool "Disable heap randomization" 1662 default y 1663 help 1664 Randomizing heap placement makes heap exploits harder, but it 1665 also breaks ancient binaries (including anything libc5 based). 1666 This option changes the bootup default to heap randomization 1667 disabled, and can be overridden at runtime by setting 1668 /proc/sys/kernel/randomize_va_space to 2. 1669 1670 On non-ancient distros (post-2000 ones) N is usually a safe choice. 1671 1672choice 1673 prompt "Choose SLAB allocator" 1674 default SLUB 1675 help 1676 This option allows to select a slab allocator. 1677 1678config SLAB 1679 bool "SLAB" 1680 help 1681 The regular slab allocator that is established and known to work 1682 well in all environments. It organizes cache hot objects in 1683 per cpu and per node queues. 1684 1685config SLUB 1686 bool "SLUB (Unqueued Allocator)" 1687 help 1688 SLUB is a slab allocator that minimizes cache line usage 1689 instead of managing queues of cached objects (SLAB approach). 1690 Per cpu caching is realized using slabs of objects instead 1691 of queues of objects. SLUB can use memory efficiently 1692 and has enhanced diagnostics. SLUB is the default choice for 1693 a slab allocator. 1694 1695config SLOB 1696 depends on EXPERT 1697 bool "SLOB (Simple Allocator)" 1698 help 1699 SLOB replaces the stock allocator with a drastically simpler 1700 allocator. SLOB is generally more space efficient but 1701 does not perform as well on large systems. 1702 1703endchoice 1704 1705config SLUB_CPU_PARTIAL 1706 default y 1707 depends on SLUB && SMP 1708 bool "SLUB per cpu partial cache" 1709 help 1710 Per cpu partial caches accellerate objects allocation and freeing 1711 that is local to a processor at the price of more indeterminism 1712 in the latency of the free. On overflow these caches will be cleared 1713 which requires the taking of locks that may cause latency spikes. 1714 Typically one would choose no for a realtime system. 1715 1716config MMAP_ALLOW_UNINITIALIZED 1717 bool "Allow mmapped anonymous memory to be uninitialized" 1718 depends on EXPERT && !MMU 1719 default n 1720 help 1721 Normally, and according to the Linux spec, anonymous memory obtained 1722 from mmap() has it's contents cleared before it is passed to 1723 userspace. Enabling this config option allows you to request that 1724 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1725 providing a huge performance boost. If this option is not enabled, 1726 then the flag will be ignored. 1727 1728 This is taken advantage of by uClibc's malloc(), and also by 1729 ELF-FDPIC binfmt's brk and stack allocator. 1730 1731 Because of the obvious security issues, this option should only be 1732 enabled on embedded devices where you control what is run in 1733 userspace. Since that isn't generally a problem on no-MMU systems, 1734 it is normally safe to say Y here. 1735 1736 See Documentation/nommu-mmap.txt for more information. 1737 1738config SYSTEM_TRUSTED_KEYRING 1739 bool "Provide system-wide ring of trusted keys" 1740 depends on KEYS 1741 help 1742 Provide a system keyring to which trusted keys can be added. Keys in 1743 the keyring are considered to be trusted. Keys may be added at will 1744 by the kernel from compiled-in data and from hardware key stores, but 1745 userspace may only add extra keys if those keys can be verified by 1746 keys already in the keyring. 1747 1748 Keys in this keyring are used by module signature checking. 1749 1750config PROFILING 1751 bool "Profiling support" 1752 help 1753 Say Y here to enable the extended profiling support mechanisms used 1754 by profilers such as OProfile. 1755 1756# 1757# Place an empty function call at each tracepoint site. Can be 1758# dynamically changed for a probe function. 1759# 1760config TRACEPOINTS 1761 bool 1762 1763source "arch/Kconfig" 1764 1765endmenu # General setup 1766 1767config HAVE_GENERIC_DMA_COHERENT 1768 bool 1769 default n 1770 1771config SLABINFO 1772 bool 1773 depends on PROC_FS 1774 depends on SLAB || SLUB_DEBUG 1775 default y 1776 1777config RT_MUTEXES 1778 boolean 1779 1780config BASE_SMALL 1781 int 1782 default 0 if BASE_FULL 1783 default 1 if !BASE_FULL 1784 1785menuconfig MODULES 1786 bool "Enable loadable module support" 1787 option modules 1788 help 1789 Kernel modules are small pieces of compiled code which can 1790 be inserted in the running kernel, rather than being 1791 permanently built into the kernel. You use the "modprobe" 1792 tool to add (and sometimes remove) them. If you say Y here, 1793 many parts of the kernel can be built as modules (by 1794 answering M instead of Y where indicated): this is most 1795 useful for infrequently used options which are not required 1796 for booting. For more information, see the man pages for 1797 modprobe, lsmod, modinfo, insmod and rmmod. 1798 1799 If you say Y here, you will need to run "make 1800 modules_install" to put the modules under /lib/modules/ 1801 where modprobe can find them (you may need to be root to do 1802 this). 1803 1804 If unsure, say Y. 1805 1806if MODULES 1807 1808config MODULE_FORCE_LOAD 1809 bool "Forced module loading" 1810 default n 1811 help 1812 Allow loading of modules without version information (ie. modprobe 1813 --force). Forced module loading sets the 'F' (forced) taint flag and 1814 is usually a really bad idea. 1815 1816config MODULE_UNLOAD 1817 bool "Module unloading" 1818 help 1819 Without this option you will not be able to unload any 1820 modules (note that some modules may not be unloadable 1821 anyway), which makes your kernel smaller, faster 1822 and simpler. If unsure, say Y. 1823 1824config MODULE_FORCE_UNLOAD 1825 bool "Forced module unloading" 1826 depends on MODULE_UNLOAD 1827 help 1828 This option allows you to force a module to unload, even if the 1829 kernel believes it is unsafe: the kernel will remove the module 1830 without waiting for anyone to stop using it (using the -f option to 1831 rmmod). This is mainly for kernel developers and desperate users. 1832 If unsure, say N. 1833 1834config MODVERSIONS 1835 bool "Module versioning support" 1836 help 1837 Usually, you have to use modules compiled with your kernel. 1838 Saying Y here makes it sometimes possible to use modules 1839 compiled for different kernels, by adding enough information 1840 to the modules to (hopefully) spot any changes which would 1841 make them incompatible with the kernel you are running. If 1842 unsure, say N. 1843 1844config MODULE_SRCVERSION_ALL 1845 bool "Source checksum for all modules" 1846 help 1847 Modules which contain a MODULE_VERSION get an extra "srcversion" 1848 field inserted into their modinfo section, which contains a 1849 sum of the source files which made it. This helps maintainers 1850 see exactly which source was used to build a module (since 1851 others sometimes change the module source without updating 1852 the version). With this option, such a "srcversion" field 1853 will be created for all modules. If unsure, say N. 1854 1855config MODULE_SIG 1856 bool "Module signature verification" 1857 depends on MODULES 1858 select SYSTEM_TRUSTED_KEYRING 1859 select KEYS 1860 select CRYPTO 1861 select ASYMMETRIC_KEY_TYPE 1862 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1863 select PUBLIC_KEY_ALGO_RSA 1864 select ASN1 1865 select OID_REGISTRY 1866 select X509_CERTIFICATE_PARSER 1867 help 1868 Check modules for valid signatures upon load: the signature 1869 is simply appended to the module. For more information see 1870 Documentation/module-signing.txt. 1871 1872 !!!WARNING!!! If you enable this option, you MUST make sure that the 1873 module DOES NOT get stripped after being signed. This includes the 1874 debuginfo strip done by some packagers (such as rpmbuild) and 1875 inclusion into an initramfs that wants the module size reduced. 1876 1877config MODULE_SIG_FORCE 1878 bool "Require modules to be validly signed" 1879 depends on MODULE_SIG 1880 help 1881 Reject unsigned modules or signed modules for which we don't have a 1882 key. Without this, such modules will simply taint the kernel. 1883 1884config MODULE_SIG_ALL 1885 bool "Automatically sign all modules" 1886 default y 1887 depends on MODULE_SIG 1888 help 1889 Sign all modules during make modules_install. Without this option, 1890 modules must be signed manually, using the scripts/sign-file tool. 1891 1892comment "Do not forget to sign required modules with scripts/sign-file" 1893 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 1894 1895choice 1896 prompt "Which hash algorithm should modules be signed with?" 1897 depends on MODULE_SIG 1898 help 1899 This determines which sort of hashing algorithm will be used during 1900 signature generation. This algorithm _must_ be built into the kernel 1901 directly so that signature verification can take place. It is not 1902 possible to load a signed module containing the algorithm to check 1903 the signature on that module. 1904 1905config MODULE_SIG_SHA1 1906 bool "Sign modules with SHA-1" 1907 select CRYPTO_SHA1 1908 1909config MODULE_SIG_SHA224 1910 bool "Sign modules with SHA-224" 1911 select CRYPTO_SHA256 1912 1913config MODULE_SIG_SHA256 1914 bool "Sign modules with SHA-256" 1915 select CRYPTO_SHA256 1916 1917config MODULE_SIG_SHA384 1918 bool "Sign modules with SHA-384" 1919 select CRYPTO_SHA512 1920 1921config MODULE_SIG_SHA512 1922 bool "Sign modules with SHA-512" 1923 select CRYPTO_SHA512 1924 1925endchoice 1926 1927config MODULE_SIG_HASH 1928 string 1929 depends on MODULE_SIG 1930 default "sha1" if MODULE_SIG_SHA1 1931 default "sha224" if MODULE_SIG_SHA224 1932 default "sha256" if MODULE_SIG_SHA256 1933 default "sha384" if MODULE_SIG_SHA384 1934 default "sha512" if MODULE_SIG_SHA512 1935 1936config MODULE_COMPRESS 1937 bool "Compress modules on installation" 1938 depends on MODULES 1939 help 1940 This option compresses the kernel modules when 'make 1941 modules_install' is run. 1942 1943 The modules will be compressed either using gzip or xz depend on the 1944 choice made in "Compression algorithm". 1945 1946 module-init-tools has support for gzip format while kmod handle gzip 1947 and xz compressed modules. 1948 1949 When a kernel module is installed from outside of the main kernel 1950 source and uses the Kbuild system for installing modules then that 1951 kernel module will also be compressed when it is installed. 1952 1953 This option provides little benefit when the modules are to be used inside 1954 an initrd or initramfs, it generally is more efficient to compress the whole 1955 initrd or initramfs instead. 1956 1957 This is fully compatible with signed modules while the signed module is 1958 compressed. module-init-tools or kmod handles decompression and provide to 1959 other layer the uncompressed but signed payload. 1960 1961choice 1962 prompt "Compression algorithm" 1963 depends on MODULE_COMPRESS 1964 default MODULE_COMPRESS_GZIP 1965 help 1966 This determines which sort of compression will be used during 1967 'make modules_install'. 1968 1969 GZIP (default) and XZ are supported. 1970 1971config MODULE_COMPRESS_GZIP 1972 bool "GZIP" 1973 1974config MODULE_COMPRESS_XZ 1975 bool "XZ" 1976 1977endchoice 1978 1979endif # MODULES 1980 1981config INIT_ALL_POSSIBLE 1982 bool 1983 help 1984 Back when each arch used to define their own cpu_online_mask and 1985 cpu_possible_mask, some of them chose to initialize cpu_possible_mask 1986 with all 1s, and others with all 0s. When they were centralised, 1987 it was better to provide this option than to break all the archs 1988 and have several arch maintainers pursuing me down dark alleys. 1989 1990config STOP_MACHINE 1991 bool 1992 default y 1993 depends on (SMP && MODULE_UNLOAD) || HOTPLUG_CPU 1994 help 1995 Need stop_machine() primitive. 1996 1997source "block/Kconfig" 1998 1999config PREEMPT_NOTIFIERS 2000 bool 2001 2002config PADATA 2003 depends on SMP 2004 bool 2005 2006# Can be selected by architectures with broken toolchains 2007# that get confused by correct const<->read_only section 2008# mappings 2009config BROKEN_RODATA 2010 bool 2011 2012config ASN1 2013 tristate 2014 help 2015 Build a simple ASN.1 grammar compiler that produces a bytecode output 2016 that can be interpreted by the ASN.1 stream decoder and used to 2017 inform it as to what tags are to be expected in a stream and what 2018 functions to call on what tags. 2019 2020source "kernel/Kconfig.locks" 2021