1config DEFCONFIG_LIST 2 string 3 depends on !UML 4 option defconfig_list 5 default "/lib/modules/$(shell,uname --release)/.config" 6 default "/etc/kernel-config" 7 default "/boot/config-$(shell,uname --release)" 8 default ARCH_DEFCONFIG 9 default "arch/$(ARCH)/defconfig" 10 11config CONSTRUCTORS 12 bool 13 depends on !UML 14 15config IRQ_WORK 16 bool 17 18config BUILDTIME_EXTABLE_SORT 19 bool 20 21config THREAD_INFO_IN_TASK 22 bool 23 help 24 Select this to move thread_info off the stack into task_struct. To 25 make this work, an arch will need to remove all thread_info fields 26 except flags and fix any runtime bugs. 27 28 One subtle change that will be needed is to use try_get_task_stack() 29 and put_task_stack() in save_thread_stack_tsk() and get_wchan(). 30 31menu "General setup" 32 33config BROKEN 34 bool 35 36config BROKEN_ON_SMP 37 bool 38 depends on BROKEN || !SMP 39 default y 40 41config INIT_ENV_ARG_LIMIT 42 int 43 default 32 if !UML 44 default 128 if UML 45 help 46 Maximum of each of the number of arguments and environment 47 variables passed to init from the kernel command line. 48 49config COMPILE_TEST 50 bool "Compile also drivers which will not load" 51 depends on !UML 52 default n 53 help 54 Some drivers can be compiled on a different platform than they are 55 intended to be run on. Despite they cannot be loaded there (or even 56 when they load they cannot be used due to missing HW support), 57 developers still, opposing to distributors, might want to build such 58 drivers to compile-test them. 59 60 If you are a developer and want to build everything available, say Y 61 here. If you are a user/distributor, say N here to exclude useless 62 drivers to be distributed. 63 64config LOCALVERSION 65 string "Local version - append to kernel release" 66 help 67 Append an extra string to the end of your kernel version. 68 This will show up when you type uname, for example. 69 The string you set here will be appended after the contents of 70 any files with a filename matching localversion* in your 71 object and source tree, in that order. Your total string can 72 be a maximum of 64 characters. 73 74config LOCALVERSION_AUTO 75 bool "Automatically append version information to the version string" 76 default y 77 depends on !COMPILE_TEST 78 help 79 This will try to automatically determine if the current tree is a 80 release tree by looking for git tags that belong to the current 81 top of tree revision. 82 83 A string of the format -gxxxxxxxx will be added to the localversion 84 if a git-based tree is found. The string generated by this will be 85 appended after any matching localversion* files, and after the value 86 set in CONFIG_LOCALVERSION. 87 88 (The actual string used here is the first eight characters produced 89 by running the command: 90 91 $ git rev-parse --verify HEAD 92 93 which is done within the script "scripts/setlocalversion".) 94 95config HAVE_KERNEL_GZIP 96 bool 97 98config HAVE_KERNEL_BZIP2 99 bool 100 101config HAVE_KERNEL_LZMA 102 bool 103 104config HAVE_KERNEL_XZ 105 bool 106 107config HAVE_KERNEL_LZO 108 bool 109 110config HAVE_KERNEL_LZ4 111 bool 112 113choice 114 prompt "Kernel compression mode" 115 default KERNEL_GZIP 116 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 117 help 118 The linux kernel is a kind of self-extracting executable. 119 Several compression algorithms are available, which differ 120 in efficiency, compression and decompression speed. 121 Compression speed is only relevant when building a kernel. 122 Decompression speed is relevant at each boot. 123 124 If you have any problems with bzip2 or lzma compressed 125 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 126 version of this functionality (bzip2 only), for 2.4, was 127 supplied by Christian Ludwig) 128 129 High compression options are mostly useful for users, who 130 are low on disk space (embedded systems), but for whom ram 131 size matters less. 132 133 If in doubt, select 'gzip' 134 135config KERNEL_GZIP 136 bool "Gzip" 137 depends on HAVE_KERNEL_GZIP 138 help 139 The old and tried gzip compression. It provides a good balance 140 between compression ratio and decompression speed. 141 142config KERNEL_BZIP2 143 bool "Bzip2" 144 depends on HAVE_KERNEL_BZIP2 145 help 146 Its compression ratio and speed is intermediate. 147 Decompression speed is slowest among the choices. The kernel 148 size is about 10% smaller with bzip2, in comparison to gzip. 149 Bzip2 uses a large amount of memory. For modern kernels you 150 will need at least 8MB RAM or more for booting. 151 152config KERNEL_LZMA 153 bool "LZMA" 154 depends on HAVE_KERNEL_LZMA 155 help 156 This compression algorithm's ratio is best. Decompression speed 157 is between gzip and bzip2. Compression is slowest. 158 The kernel size is about 33% smaller with LZMA in comparison to gzip. 159 160config KERNEL_XZ 161 bool "XZ" 162 depends on HAVE_KERNEL_XZ 163 help 164 XZ uses the LZMA2 algorithm and instruction set specific 165 BCJ filters which can improve compression ratio of executable 166 code. The size of the kernel is about 30% smaller with XZ in 167 comparison to gzip. On architectures for which there is a BCJ 168 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 169 will create a few percent smaller kernel than plain LZMA. 170 171 The speed is about the same as with LZMA: The decompression 172 speed of XZ is better than that of bzip2 but worse than gzip 173 and LZO. Compression is slow. 174 175config KERNEL_LZO 176 bool "LZO" 177 depends on HAVE_KERNEL_LZO 178 help 179 Its compression ratio is the poorest among the choices. The kernel 180 size is about 10% bigger than gzip; however its speed 181 (both compression and decompression) is the fastest. 182 183config KERNEL_LZ4 184 bool "LZ4" 185 depends on HAVE_KERNEL_LZ4 186 help 187 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. 188 A preliminary version of LZ4 de/compression tool is available at 189 <https://code.google.com/p/lz4/>. 190 191 Its compression ratio is worse than LZO. The size of the kernel 192 is about 8% bigger than LZO. But the decompression speed is 193 faster than LZO. 194 195endchoice 196 197config DEFAULT_HOSTNAME 198 string "Default hostname" 199 default "(none)" 200 help 201 This option determines the default system hostname before userspace 202 calls sethostname(2). The kernel traditionally uses "(none)" here, 203 but you may wish to use a different default here to make a minimal 204 system more usable with less configuration. 205 206config SWAP 207 bool "Support for paging of anonymous memory (swap)" 208 depends on MMU && BLOCK 209 default y 210 help 211 This option allows you to choose whether you want to have support 212 for so called swap devices or swap files in your kernel that are 213 used to provide more virtual memory than the actual RAM present 214 in your computer. If unsure say Y. 215 216config SYSVIPC 217 bool "System V IPC" 218 ---help--- 219 Inter Process Communication is a suite of library functions and 220 system calls which let processes (running programs) synchronize and 221 exchange information. It is generally considered to be a good thing, 222 and some programs won't run unless you say Y here. In particular, if 223 you want to run the DOS emulator dosemu under Linux (read the 224 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 225 you'll need to say Y here. 226 227 You can find documentation about IPC with "info ipc" and also in 228 section 6.4 of the Linux Programmer's Guide, available from 229 <http://www.tldp.org/guides.html>. 230 231config SYSVIPC_SYSCTL 232 bool 233 depends on SYSVIPC 234 depends on SYSCTL 235 default y 236 237config POSIX_MQUEUE 238 bool "POSIX Message Queues" 239 depends on NET 240 ---help--- 241 POSIX variant of message queues is a part of IPC. In POSIX message 242 queues every message has a priority which decides about succession 243 of receiving it by a process. If you want to compile and run 244 programs written e.g. for Solaris with use of its POSIX message 245 queues (functions mq_*) say Y here. 246 247 POSIX message queues are visible as a filesystem called 'mqueue' 248 and can be mounted somewhere if you want to do filesystem 249 operations on message queues. 250 251 If unsure, say Y. 252 253config POSIX_MQUEUE_SYSCTL 254 bool 255 depends on POSIX_MQUEUE 256 depends on SYSCTL 257 default y 258 259config CROSS_MEMORY_ATTACH 260 bool "Enable process_vm_readv/writev syscalls" 261 depends on MMU 262 default y 263 help 264 Enabling this option adds the system calls process_vm_readv and 265 process_vm_writev which allow a process with the correct privileges 266 to directly read from or write to another process' address space. 267 See the man page for more details. 268 269config USELIB 270 bool "uselib syscall" 271 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION 272 help 273 This option enables the uselib syscall, a system call used in the 274 dynamic linker from libc5 and earlier. glibc does not use this 275 system call. If you intend to run programs built on libc5 or 276 earlier, you may need to enable this syscall. Current systems 277 running glibc can safely disable this. 278 279config AUDIT 280 bool "Auditing support" 281 depends on NET 282 help 283 Enable auditing infrastructure that can be used with another 284 kernel subsystem, such as SELinux (which requires this for 285 logging of avc messages output). System call auditing is included 286 on architectures which support it. 287 288config HAVE_ARCH_AUDITSYSCALL 289 bool 290 291config AUDITSYSCALL 292 def_bool y 293 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL 294 295config AUDIT_WATCH 296 def_bool y 297 depends on AUDITSYSCALL 298 select FSNOTIFY 299 300config AUDIT_TREE 301 def_bool y 302 depends on AUDITSYSCALL 303 select FSNOTIFY 304 305source "kernel/irq/Kconfig" 306source "kernel/time/Kconfig" 307 308menu "CPU/Task time and stats accounting" 309 310config VIRT_CPU_ACCOUNTING 311 bool 312 313choice 314 prompt "Cputime accounting" 315 default TICK_CPU_ACCOUNTING if !PPC64 316 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 317 318# Kind of a stub config for the pure tick based cputime accounting 319config TICK_CPU_ACCOUNTING 320 bool "Simple tick based cputime accounting" 321 depends on !S390 && !NO_HZ_FULL 322 help 323 This is the basic tick based cputime accounting that maintains 324 statistics about user, system and idle time spent on per jiffies 325 granularity. 326 327 If unsure, say Y. 328 329config VIRT_CPU_ACCOUNTING_NATIVE 330 bool "Deterministic task and CPU time accounting" 331 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 332 select VIRT_CPU_ACCOUNTING 333 help 334 Select this option to enable more accurate task and CPU time 335 accounting. This is done by reading a CPU counter on each 336 kernel entry and exit and on transitions within the kernel 337 between system, softirq and hardirq state, so there is a 338 small performance impact. In the case of s390 or IBM POWER > 5, 339 this also enables accounting of stolen time on logically-partitioned 340 systems. 341 342config VIRT_CPU_ACCOUNTING_GEN 343 bool "Full dynticks CPU time accounting" 344 depends on HAVE_CONTEXT_TRACKING 345 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN 346 select VIRT_CPU_ACCOUNTING 347 select CONTEXT_TRACKING 348 help 349 Select this option to enable task and CPU time accounting on full 350 dynticks systems. This accounting is implemented by watching every 351 kernel-user boundaries using the context tracking subsystem. 352 The accounting is thus performed at the expense of some significant 353 overhead. 354 355 For now this is only useful if you are working on the full 356 dynticks subsystem development. 357 358 If unsure, say N. 359 360endchoice 361 362config IRQ_TIME_ACCOUNTING 363 bool "Fine granularity task level IRQ time accounting" 364 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE 365 help 366 Select this option to enable fine granularity task irq time 367 accounting. This is done by reading a timestamp on each 368 transitions between softirq and hardirq state, so there can be a 369 small performance impact. 370 371 If in doubt, say N here. 372 373config BSD_PROCESS_ACCT 374 bool "BSD Process Accounting" 375 depends on MULTIUSER 376 help 377 If you say Y here, a user level program will be able to instruct the 378 kernel (via a special system call) to write process accounting 379 information to a file: whenever a process exits, information about 380 that process will be appended to the file by the kernel. The 381 information includes things such as creation time, owning user, 382 command name, memory usage, controlling terminal etc. (the complete 383 list is in the struct acct in <file:include/linux/acct.h>). It is 384 up to the user level program to do useful things with this 385 information. This is generally a good idea, so say Y. 386 387config BSD_PROCESS_ACCT_V3 388 bool "BSD Process Accounting version 3 file format" 389 depends on BSD_PROCESS_ACCT 390 default n 391 help 392 If you say Y here, the process accounting information is written 393 in a new file format that also logs the process IDs of each 394 process and it's parent. Note that this file format is incompatible 395 with previous v0/v1/v2 file formats, so you will need updated tools 396 for processing it. A preliminary version of these tools is available 397 at <http://www.gnu.org/software/acct/>. 398 399config TASKSTATS 400 bool "Export task/process statistics through netlink" 401 depends on NET 402 depends on MULTIUSER 403 default n 404 help 405 Export selected statistics for tasks/processes through the 406 generic netlink interface. Unlike BSD process accounting, the 407 statistics are available during the lifetime of tasks/processes as 408 responses to commands. Like BSD accounting, they are sent to user 409 space on task exit. 410 411 Say N if unsure. 412 413config TASK_DELAY_ACCT 414 bool "Enable per-task delay accounting" 415 depends on TASKSTATS 416 select SCHED_INFO 417 help 418 Collect information on time spent by a task waiting for system 419 resources like cpu, synchronous block I/O completion and swapping 420 in pages. Such statistics can help in setting a task's priorities 421 relative to other tasks for cpu, io, rss limits etc. 422 423 Say N if unsure. 424 425config TASK_XACCT 426 bool "Enable extended accounting over taskstats" 427 depends on TASKSTATS 428 help 429 Collect extended task accounting data and send the data 430 to userland for processing over the taskstats interface. 431 432 Say N if unsure. 433 434config TASK_IO_ACCOUNTING 435 bool "Enable per-task storage I/O accounting" 436 depends on TASK_XACCT 437 help 438 Collect information on the number of bytes of storage I/O which this 439 task has caused. 440 441 Say N if unsure. 442 443endmenu # "CPU/Task time and stats accounting" 444 445config CPU_ISOLATION 446 bool "CPU isolation" 447 depends on SMP || COMPILE_TEST 448 default y 449 help 450 Make sure that CPUs running critical tasks are not disturbed by 451 any source of "noise" such as unbound workqueues, timers, kthreads... 452 Unbound jobs get offloaded to housekeeping CPUs. This is driven by 453 the "isolcpus=" boot parameter. 454 455 Say Y if unsure. 456 457source "kernel/rcu/Kconfig" 458 459config BUILD_BIN2C 460 bool 461 default n 462 463config IKCONFIG 464 tristate "Kernel .config support" 465 select BUILD_BIN2C 466 ---help--- 467 This option enables the complete Linux kernel ".config" file 468 contents to be saved in the kernel. It provides documentation 469 of which kernel options are used in a running kernel or in an 470 on-disk kernel. This information can be extracted from the kernel 471 image file with the script scripts/extract-ikconfig and used as 472 input to rebuild the current kernel or to build another kernel. 473 It can also be extracted from a running kernel by reading 474 /proc/config.gz if enabled (below). 475 476config IKCONFIG_PROC 477 bool "Enable access to .config through /proc/config.gz" 478 depends on IKCONFIG && PROC_FS 479 ---help--- 480 This option enables access to the kernel configuration file 481 through /proc/config.gz. 482 483config LOG_BUF_SHIFT 484 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 485 range 12 25 486 default 17 487 depends on PRINTK 488 help 489 Select the minimal kernel log buffer size as a power of 2. 490 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config 491 parameter, see below. Any higher size also might be forced 492 by "log_buf_len" boot parameter. 493 494 Examples: 495 17 => 128 KB 496 16 => 64 KB 497 15 => 32 KB 498 14 => 16 KB 499 13 => 8 KB 500 12 => 4 KB 501 502config LOG_CPU_MAX_BUF_SHIFT 503 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" 504 depends on SMP 505 range 0 21 506 default 12 if !BASE_SMALL 507 default 0 if BASE_SMALL 508 depends on PRINTK 509 help 510 This option allows to increase the default ring buffer size 511 according to the number of CPUs. The value defines the contribution 512 of each CPU as a power of 2. The used space is typically only few 513 lines however it might be much more when problems are reported, 514 e.g. backtraces. 515 516 The increased size means that a new buffer has to be allocated and 517 the original static one is unused. It makes sense only on systems 518 with more CPUs. Therefore this value is used only when the sum of 519 contributions is greater than the half of the default kernel ring 520 buffer as defined by LOG_BUF_SHIFT. The default values are set 521 so that more than 64 CPUs are needed to trigger the allocation. 522 523 Also this option is ignored when "log_buf_len" kernel parameter is 524 used as it forces an exact (power of two) size of the ring buffer. 525 526 The number of possible CPUs is used for this computation ignoring 527 hotplugging making the computation optimal for the worst case 528 scenario while allowing a simple algorithm to be used from bootup. 529 530 Examples shift values and their meaning: 531 17 => 128 KB for each CPU 532 16 => 64 KB for each CPU 533 15 => 32 KB for each CPU 534 14 => 16 KB for each CPU 535 13 => 8 KB for each CPU 536 12 => 4 KB for each CPU 537 538config PRINTK_SAFE_LOG_BUF_SHIFT 539 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)" 540 range 10 21 541 default 13 542 depends on PRINTK 543 help 544 Select the size of an alternate printk per-CPU buffer where messages 545 printed from usafe contexts are temporary stored. One example would 546 be NMI messages, another one - printk recursion. The messages are 547 copied to the main log buffer in a safe context to avoid a deadlock. 548 The value defines the size as a power of 2. 549 550 Those messages are rare and limited. The largest one is when 551 a backtrace is printed. It usually fits into 4KB. Select 552 8KB if you want to be on the safe side. 553 554 Examples: 555 17 => 128 KB for each CPU 556 16 => 64 KB for each CPU 557 15 => 32 KB for each CPU 558 14 => 16 KB for each CPU 559 13 => 8 KB for each CPU 560 12 => 4 KB for each CPU 561 562# 563# Architectures with an unreliable sched_clock() should select this: 564# 565config HAVE_UNSTABLE_SCHED_CLOCK 566 bool 567 568config GENERIC_SCHED_CLOCK 569 bool 570 571# 572# For architectures that want to enable the support for NUMA-affine scheduler 573# balancing logic: 574# 575config ARCH_SUPPORTS_NUMA_BALANCING 576 bool 577 578# 579# For architectures that prefer to flush all TLBs after a number of pages 580# are unmapped instead of sending one IPI per page to flush. The architecture 581# must provide guarantees on what happens if a clean TLB cache entry is 582# written after the unmap. Details are in mm/rmap.c near the check for 583# should_defer_flush. The architecture should also consider if the full flush 584# and the refill costs are offset by the savings of sending fewer IPIs. 585config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH 586 bool 587 588# 589# For architectures that know their GCC __int128 support is sound 590# 591config ARCH_SUPPORTS_INT128 592 bool 593 594# For architectures that (ab)use NUMA to represent different memory regions 595# all cpu-local but of different latencies, such as SuperH. 596# 597config ARCH_WANT_NUMA_VARIABLE_LOCALITY 598 bool 599 600config NUMA_BALANCING 601 bool "Memory placement aware NUMA scheduler" 602 depends on ARCH_SUPPORTS_NUMA_BALANCING 603 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 604 depends on SMP && NUMA && MIGRATION 605 help 606 This option adds support for automatic NUMA aware memory/task placement. 607 The mechanism is quite primitive and is based on migrating memory when 608 it has references to the node the task is running on. 609 610 This system will be inactive on UMA systems. 611 612config NUMA_BALANCING_DEFAULT_ENABLED 613 bool "Automatically enable NUMA aware memory/task placement" 614 default y 615 depends on NUMA_BALANCING 616 help 617 If set, automatic NUMA balancing will be enabled if running on a NUMA 618 machine. 619 620menuconfig CGROUPS 621 bool "Control Group support" 622 select KERNFS 623 help 624 This option adds support for grouping sets of processes together, for 625 use with process control subsystems such as Cpusets, CFS, memory 626 controls or device isolation. 627 See 628 - Documentation/scheduler/sched-design-CFS.txt (CFS) 629 - Documentation/cgroup-v1/ (features for grouping, isolation 630 and resource control) 631 632 Say N if unsure. 633 634if CGROUPS 635 636config PAGE_COUNTER 637 bool 638 639config MEMCG 640 bool "Memory controller" 641 select PAGE_COUNTER 642 select EVENTFD 643 help 644 Provides control over the memory footprint of tasks in a cgroup. 645 646config MEMCG_SWAP 647 bool "Swap controller" 648 depends on MEMCG && SWAP 649 help 650 Provides control over the swap space consumed by tasks in a cgroup. 651 652config MEMCG_SWAP_ENABLED 653 bool "Swap controller enabled by default" 654 depends on MEMCG_SWAP 655 default y 656 help 657 Memory Resource Controller Swap Extension comes with its price in 658 a bigger memory consumption. General purpose distribution kernels 659 which want to enable the feature but keep it disabled by default 660 and let the user enable it by swapaccount=1 boot command line 661 parameter should have this option unselected. 662 For those who want to have the feature enabled by default should 663 select this option (if, for some reason, they need to disable it 664 then swapaccount=0 does the trick). 665 666config BLK_CGROUP 667 bool "IO controller" 668 depends on BLOCK 669 default n 670 ---help--- 671 Generic block IO controller cgroup interface. This is the common 672 cgroup interface which should be used by various IO controlling 673 policies. 674 675 Currently, CFQ IO scheduler uses it to recognize task groups and 676 control disk bandwidth allocation (proportional time slice allocation) 677 to such task groups. It is also used by bio throttling logic in 678 block layer to implement upper limit in IO rates on a device. 679 680 This option only enables generic Block IO controller infrastructure. 681 One needs to also enable actual IO controlling logic/policy. For 682 enabling proportional weight division of disk bandwidth in CFQ, set 683 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 684 CONFIG_BLK_DEV_THROTTLING=y. 685 686 See Documentation/cgroup-v1/blkio-controller.txt for more information. 687 688config DEBUG_BLK_CGROUP 689 bool "IO controller debugging" 690 depends on BLK_CGROUP 691 default n 692 ---help--- 693 Enable some debugging help. Currently it exports additional stat 694 files in a cgroup which can be useful for debugging. 695 696config CGROUP_WRITEBACK 697 bool 698 depends on MEMCG && BLK_CGROUP 699 default y 700 701menuconfig CGROUP_SCHED 702 bool "CPU controller" 703 default n 704 help 705 This feature lets CPU scheduler recognize task groups and control CPU 706 bandwidth allocation to such task groups. It uses cgroups to group 707 tasks. 708 709if CGROUP_SCHED 710config FAIR_GROUP_SCHED 711 bool "Group scheduling for SCHED_OTHER" 712 depends on CGROUP_SCHED 713 default CGROUP_SCHED 714 715config CFS_BANDWIDTH 716 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 717 depends on FAIR_GROUP_SCHED 718 default n 719 help 720 This option allows users to define CPU bandwidth rates (limits) for 721 tasks running within the fair group scheduler. Groups with no limit 722 set are considered to be unconstrained and will run with no 723 restriction. 724 See Documentation/scheduler/sched-bwc.txt for more information. 725 726config RT_GROUP_SCHED 727 bool "Group scheduling for SCHED_RR/FIFO" 728 depends on CGROUP_SCHED 729 default n 730 help 731 This feature lets you explicitly allocate real CPU bandwidth 732 to task groups. If enabled, it will also make it impossible to 733 schedule realtime tasks for non-root users until you allocate 734 realtime bandwidth for them. 735 See Documentation/scheduler/sched-rt-group.txt for more information. 736 737endif #CGROUP_SCHED 738 739config CGROUP_PIDS 740 bool "PIDs controller" 741 help 742 Provides enforcement of process number limits in the scope of a 743 cgroup. Any attempt to fork more processes than is allowed in the 744 cgroup will fail. PIDs are fundamentally a global resource because it 745 is fairly trivial to reach PID exhaustion before you reach even a 746 conservative kmemcg limit. As a result, it is possible to grind a 747 system to halt without being limited by other cgroup policies. The 748 PIDs controller is designed to stop this from happening. 749 750 It should be noted that organisational operations (such as attaching 751 to a cgroup hierarchy will *not* be blocked by the PIDs controller), 752 since the PIDs limit only affects a process's ability to fork, not to 753 attach to a cgroup. 754 755config CGROUP_RDMA 756 bool "RDMA controller" 757 help 758 Provides enforcement of RDMA resources defined by IB stack. 759 It is fairly easy for consumers to exhaust RDMA resources, which 760 can result into resource unavailability to other consumers. 761 RDMA controller is designed to stop this from happening. 762 Attaching processes with active RDMA resources to the cgroup 763 hierarchy is allowed even if can cross the hierarchy's limit. 764 765config CGROUP_FREEZER 766 bool "Freezer controller" 767 help 768 Provides a way to freeze and unfreeze all tasks in a 769 cgroup. 770 771 This option affects the ORIGINAL cgroup interface. The cgroup2 memory 772 controller includes important in-kernel memory consumers per default. 773 774 If you're using cgroup2, say N. 775 776config CGROUP_HUGETLB 777 bool "HugeTLB controller" 778 depends on HUGETLB_PAGE 779 select PAGE_COUNTER 780 default n 781 help 782 Provides a cgroup controller for HugeTLB pages. 783 When you enable this, you can put a per cgroup limit on HugeTLB usage. 784 The limit is enforced during page fault. Since HugeTLB doesn't 785 support page reclaim, enforcing the limit at page fault time implies 786 that, the application will get SIGBUS signal if it tries to access 787 HugeTLB pages beyond its limit. This requires the application to know 788 beforehand how much HugeTLB pages it would require for its use. The 789 control group is tracked in the third page lru pointer. This means 790 that we cannot use the controller with huge page less than 3 pages. 791 792config CPUSETS 793 bool "Cpuset controller" 794 depends on SMP 795 help 796 This option will let you create and manage CPUSETs which 797 allow dynamically partitioning a system into sets of CPUs and 798 Memory Nodes and assigning tasks to run only within those sets. 799 This is primarily useful on large SMP or NUMA systems. 800 801 Say N if unsure. 802 803config PROC_PID_CPUSET 804 bool "Include legacy /proc/<pid>/cpuset file" 805 depends on CPUSETS 806 default y 807 808config CGROUP_DEVICE 809 bool "Device controller" 810 help 811 Provides a cgroup controller implementing whitelists for 812 devices which a process in the cgroup can mknod or open. 813 814config CGROUP_CPUACCT 815 bool "Simple CPU accounting controller" 816 help 817 Provides a simple controller for monitoring the 818 total CPU consumed by the tasks in a cgroup. 819 820config CGROUP_PERF 821 bool "Perf controller" 822 depends on PERF_EVENTS 823 help 824 This option extends the perf per-cpu mode to restrict monitoring 825 to threads which belong to the cgroup specified and run on the 826 designated cpu. 827 828 Say N if unsure. 829 830config CGROUP_BPF 831 bool "Support for eBPF programs attached to cgroups" 832 depends on BPF_SYSCALL 833 select SOCK_CGROUP_DATA 834 help 835 Allow attaching eBPF programs to a cgroup using the bpf(2) 836 syscall command BPF_PROG_ATTACH. 837 838 In which context these programs are accessed depends on the type 839 of attachment. For instance, programs that are attached using 840 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of 841 inet sockets. 842 843config CGROUP_DEBUG 844 bool "Debug controller" 845 default n 846 depends on DEBUG_KERNEL 847 help 848 This option enables a simple controller that exports 849 debugging information about the cgroups framework. This 850 controller is for control cgroup debugging only. Its 851 interfaces are not stable. 852 853 Say N. 854 855config SOCK_CGROUP_DATA 856 bool 857 default n 858 859endif # CGROUPS 860 861menuconfig NAMESPACES 862 bool "Namespaces support" if EXPERT 863 depends on MULTIUSER 864 default !EXPERT 865 help 866 Provides the way to make tasks work with different objects using 867 the same id. For example same IPC id may refer to different objects 868 or same user id or pid may refer to different tasks when used in 869 different namespaces. 870 871if NAMESPACES 872 873config UTS_NS 874 bool "UTS namespace" 875 default y 876 help 877 In this namespace tasks see different info provided with the 878 uname() system call 879 880config IPC_NS 881 bool "IPC namespace" 882 depends on (SYSVIPC || POSIX_MQUEUE) 883 default y 884 help 885 In this namespace tasks work with IPC ids which correspond to 886 different IPC objects in different namespaces. 887 888config USER_NS 889 bool "User namespace" 890 default n 891 help 892 This allows containers, i.e. vservers, to use user namespaces 893 to provide different user info for different servers. 894 895 When user namespaces are enabled in the kernel it is 896 recommended that the MEMCG option also be enabled and that 897 user-space use the memory control groups to limit the amount 898 of memory a memory unprivileged users can use. 899 900 If unsure, say N. 901 902config PID_NS 903 bool "PID Namespaces" 904 default y 905 help 906 Support process id namespaces. This allows having multiple 907 processes with the same pid as long as they are in different 908 pid namespaces. This is a building block of containers. 909 910config NET_NS 911 bool "Network namespace" 912 depends on NET 913 default y 914 help 915 Allow user space to create what appear to be multiple instances 916 of the network stack. 917 918endif # NAMESPACES 919 920config SCHED_AUTOGROUP 921 bool "Automatic process group scheduling" 922 select CGROUPS 923 select CGROUP_SCHED 924 select FAIR_GROUP_SCHED 925 help 926 This option optimizes the scheduler for common desktop workloads by 927 automatically creating and populating task groups. This separation 928 of workloads isolates aggressive CPU burners (like build jobs) from 929 desktop applications. Task group autogeneration is currently based 930 upon task session. 931 932config SYSFS_DEPRECATED 933 bool "Enable deprecated sysfs features to support old userspace tools" 934 depends on SYSFS 935 default n 936 help 937 This option adds code that switches the layout of the "block" class 938 devices, to not show up in /sys/class/block/, but only in 939 /sys/block/. 940 941 This switch is only active when the sysfs.deprecated=1 boot option is 942 passed or the SYSFS_DEPRECATED_V2 option is set. 943 944 This option allows new kernels to run on old distributions and tools, 945 which might get confused by /sys/class/block/. Since 2007/2008 all 946 major distributions and tools handle this just fine. 947 948 Recent distributions and userspace tools after 2009/2010 depend on 949 the existence of /sys/class/block/, and will not work with this 950 option enabled. 951 952 Only if you are using a new kernel on an old distribution, you might 953 need to say Y here. 954 955config SYSFS_DEPRECATED_V2 956 bool "Enable deprecated sysfs features by default" 957 default n 958 depends on SYSFS 959 depends on SYSFS_DEPRECATED 960 help 961 Enable deprecated sysfs by default. 962 963 See the CONFIG_SYSFS_DEPRECATED option for more details about this 964 option. 965 966 Only if you are using a new kernel on an old distribution, you might 967 need to say Y here. Even then, odds are you would not need it 968 enabled, you can always pass the boot option if absolutely necessary. 969 970config RELAY 971 bool "Kernel->user space relay support (formerly relayfs)" 972 select IRQ_WORK 973 help 974 This option enables support for relay interface support in 975 certain file systems (such as debugfs). 976 It is designed to provide an efficient mechanism for tools and 977 facilities to relay large amounts of data from kernel space to 978 user space. 979 980 If unsure, say N. 981 982config BLK_DEV_INITRD 983 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 984 help 985 The initial RAM filesystem is a ramfs which is loaded by the 986 boot loader (loadlin or lilo) and that is mounted as root 987 before the normal boot procedure. It is typically used to 988 load modules needed to mount the "real" root file system, 989 etc. See <file:Documentation/admin-guide/initrd.rst> for details. 990 991 If RAM disk support (BLK_DEV_RAM) is also included, this 992 also enables initial RAM disk (initrd) support and adds 993 15 Kbytes (more on some other architectures) to the kernel size. 994 995 If unsure say Y. 996 997if BLK_DEV_INITRD 998 999source "usr/Kconfig" 1000 1001endif 1002 1003choice 1004 prompt "Compiler optimization level" 1005 default CC_OPTIMIZE_FOR_PERFORMANCE 1006 1007config CC_OPTIMIZE_FOR_PERFORMANCE 1008 bool "Optimize for performance" 1009 help 1010 This is the default optimization level for the kernel, building 1011 with the "-O2" compiler flag for best performance and most 1012 helpful compile-time warnings. 1013 1014config CC_OPTIMIZE_FOR_SIZE 1015 bool "Optimize for size" 1016 help 1017 Enabling this option will pass "-Os" instead of "-O2" to 1018 your compiler resulting in a smaller kernel. 1019 1020 If unsure, say N. 1021 1022endchoice 1023 1024config HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1025 bool 1026 help 1027 This requires that the arch annotates or otherwise protects 1028 its external entry points from being discarded. Linker scripts 1029 must also merge .text.*, .data.*, and .bss.* correctly into 1030 output sections. Care must be taken not to pull in unrelated 1031 sections (e.g., '.text.init'). Typically '.' in section names 1032 is used to distinguish them from label names / C identifiers. 1033 1034config LD_DEAD_CODE_DATA_ELIMINATION 1035 bool "Dead code and data elimination (EXPERIMENTAL)" 1036 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1037 depends on EXPERT 1038 help 1039 Select this if the architecture wants to do dead code and 1040 data elimination with the linker by compiling with 1041 -ffunction-sections -fdata-sections, and linking with 1042 --gc-sections. 1043 1044 This can reduce on disk and in-memory size of the kernel 1045 code and static data, particularly for small configs and 1046 on small systems. This has the possibility of introducing 1047 silently broken kernel if the required annotations are not 1048 present. This option is not well tested yet, so use at your 1049 own risk. 1050 1051config SYSCTL 1052 bool 1053 1054config ANON_INODES 1055 bool 1056 1057config HAVE_UID16 1058 bool 1059 1060config SYSCTL_EXCEPTION_TRACE 1061 bool 1062 help 1063 Enable support for /proc/sys/debug/exception-trace. 1064 1065config SYSCTL_ARCH_UNALIGN_NO_WARN 1066 bool 1067 help 1068 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1069 Allows arch to define/use @no_unaligned_warning to possibly warn 1070 about unaligned access emulation going on under the hood. 1071 1072config SYSCTL_ARCH_UNALIGN_ALLOW 1073 bool 1074 help 1075 Enable support for /proc/sys/kernel/unaligned-trap 1076 Allows arches to define/use @unaligned_enabled to runtime toggle 1077 the unaligned access emulation. 1078 see arch/parisc/kernel/unaligned.c for reference 1079 1080config HAVE_PCSPKR_PLATFORM 1081 bool 1082 1083# interpreter that classic socket filters depend on 1084config BPF 1085 bool 1086 1087menuconfig EXPERT 1088 bool "Configure standard kernel features (expert users)" 1089 # Unhide debug options, to make the on-by-default options visible 1090 select DEBUG_KERNEL 1091 help 1092 This option allows certain base kernel options and settings 1093 to be disabled or tweaked. This is for specialized 1094 environments which can tolerate a "non-standard" kernel. 1095 Only use this if you really know what you are doing. 1096 1097config UID16 1098 bool "Enable 16-bit UID system calls" if EXPERT 1099 depends on HAVE_UID16 && MULTIUSER 1100 default y 1101 help 1102 This enables the legacy 16-bit UID syscall wrappers. 1103 1104config MULTIUSER 1105 bool "Multiple users, groups and capabilities support" if EXPERT 1106 default y 1107 help 1108 This option enables support for non-root users, groups and 1109 capabilities. 1110 1111 If you say N here, all processes will run with UID 0, GID 0, and all 1112 possible capabilities. Saying N here also compiles out support for 1113 system calls related to UIDs, GIDs, and capabilities, such as setuid, 1114 setgid, and capset. 1115 1116 If unsure, say Y here. 1117 1118config SGETMASK_SYSCALL 1119 bool "sgetmask/ssetmask syscalls support" if EXPERT 1120 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH 1121 ---help--- 1122 sys_sgetmask and sys_ssetmask are obsolete system calls 1123 no longer supported in libc but still enabled by default in some 1124 architectures. 1125 1126 If unsure, leave the default option here. 1127 1128config SYSFS_SYSCALL 1129 bool "Sysfs syscall support" if EXPERT 1130 default y 1131 ---help--- 1132 sys_sysfs is an obsolete system call no longer supported in libc. 1133 Note that disabling this option is more secure but might break 1134 compatibility with some systems. 1135 1136 If unsure say Y here. 1137 1138config SYSCTL_SYSCALL 1139 bool "Sysctl syscall support" if EXPERT 1140 depends on PROC_SYSCTL 1141 default n 1142 select SYSCTL 1143 ---help--- 1144 sys_sysctl uses binary paths that have been found challenging 1145 to properly maintain and use. The interface in /proc/sys 1146 using paths with ascii names is now the primary path to this 1147 information. 1148 1149 Almost nothing using the binary sysctl interface so if you are 1150 trying to save some space it is probably safe to disable this, 1151 making your kernel marginally smaller. 1152 1153 If unsure say N here. 1154 1155config FHANDLE 1156 bool "open by fhandle syscalls" if EXPERT 1157 select EXPORTFS 1158 default y 1159 help 1160 If you say Y here, a user level program will be able to map 1161 file names to handle and then later use the handle for 1162 different file system operations. This is useful in implementing 1163 userspace file servers, which now track files using handles instead 1164 of names. The handle would remain the same even if file names 1165 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 1166 syscalls. 1167 1168config POSIX_TIMERS 1169 bool "Posix Clocks & timers" if EXPERT 1170 default y 1171 help 1172 This includes native support for POSIX timers to the kernel. 1173 Some embedded systems have no use for them and therefore they 1174 can be configured out to reduce the size of the kernel image. 1175 1176 When this option is disabled, the following syscalls won't be 1177 available: timer_create, timer_gettime: timer_getoverrun, 1178 timer_settime, timer_delete, clock_adjtime, getitimer, 1179 setitimer, alarm. Furthermore, the clock_settime, clock_gettime, 1180 clock_getres and clock_nanosleep syscalls will be limited to 1181 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only. 1182 1183 If unsure say y. 1184 1185config PRINTK 1186 default y 1187 bool "Enable support for printk" if EXPERT 1188 select IRQ_WORK 1189 help 1190 This option enables normal printk support. Removing it 1191 eliminates most of the message strings from the kernel image 1192 and makes the kernel more or less silent. As this makes it 1193 very difficult to diagnose system problems, saying N here is 1194 strongly discouraged. 1195 1196config PRINTK_NMI 1197 def_bool y 1198 depends on PRINTK 1199 depends on HAVE_NMI 1200 1201config BUG 1202 bool "BUG() support" if EXPERT 1203 default y 1204 help 1205 Disabling this option eliminates support for BUG and WARN, reducing 1206 the size of your kernel image and potentially quietly ignoring 1207 numerous fatal conditions. You should only consider disabling this 1208 option for embedded systems with no facilities for reporting errors. 1209 Just say Y. 1210 1211config ELF_CORE 1212 depends on COREDUMP 1213 default y 1214 bool "Enable ELF core dumps" if EXPERT 1215 help 1216 Enable support for generating core dumps. Disabling saves about 4k. 1217 1218 1219config PCSPKR_PLATFORM 1220 bool "Enable PC-Speaker support" if EXPERT 1221 depends on HAVE_PCSPKR_PLATFORM 1222 select I8253_LOCK 1223 default y 1224 help 1225 This option allows to disable the internal PC-Speaker 1226 support, saving some memory. 1227 1228config BASE_FULL 1229 default y 1230 bool "Enable full-sized data structures for core" if EXPERT 1231 help 1232 Disabling this option reduces the size of miscellaneous core 1233 kernel data structures. This saves memory on small machines, 1234 but may reduce performance. 1235 1236config FUTEX 1237 bool "Enable futex support" if EXPERT 1238 default y 1239 imply RT_MUTEXES 1240 help 1241 Disabling this option will cause the kernel to be built without 1242 support for "fast userspace mutexes". The resulting kernel may not 1243 run glibc-based applications correctly. 1244 1245config FUTEX_PI 1246 bool 1247 depends on FUTEX && RT_MUTEXES 1248 default y 1249 1250config HAVE_FUTEX_CMPXCHG 1251 bool 1252 depends on FUTEX 1253 help 1254 Architectures should select this if futex_atomic_cmpxchg_inatomic() 1255 is implemented and always working. This removes a couple of runtime 1256 checks. 1257 1258config EPOLL 1259 bool "Enable eventpoll support" if EXPERT 1260 default y 1261 select ANON_INODES 1262 help 1263 Disabling this option will cause the kernel to be built without 1264 support for epoll family of system calls. 1265 1266config SIGNALFD 1267 bool "Enable signalfd() system call" if EXPERT 1268 select ANON_INODES 1269 default y 1270 help 1271 Enable the signalfd() system call that allows to receive signals 1272 on a file descriptor. 1273 1274 If unsure, say Y. 1275 1276config TIMERFD 1277 bool "Enable timerfd() system call" if EXPERT 1278 select ANON_INODES 1279 default y 1280 help 1281 Enable the timerfd() system call that allows to receive timer 1282 events on a file descriptor. 1283 1284 If unsure, say Y. 1285 1286config EVENTFD 1287 bool "Enable eventfd() system call" if EXPERT 1288 select ANON_INODES 1289 default y 1290 help 1291 Enable the eventfd() system call that allows to receive both 1292 kernel notification (ie. KAIO) or userspace notifications. 1293 1294 If unsure, say Y. 1295 1296config SHMEM 1297 bool "Use full shmem filesystem" if EXPERT 1298 default y 1299 depends on MMU 1300 help 1301 The shmem is an internal filesystem used to manage shared memory. 1302 It is backed by swap and manages resource limits. It is also exported 1303 to userspace as tmpfs if TMPFS is enabled. Disabling this 1304 option replaces shmem and tmpfs with the much simpler ramfs code, 1305 which may be appropriate on small systems without swap. 1306 1307config AIO 1308 bool "Enable AIO support" if EXPERT 1309 default y 1310 help 1311 This option enables POSIX asynchronous I/O which may by used 1312 by some high performance threaded applications. Disabling 1313 this option saves about 7k. 1314 1315config ADVISE_SYSCALLS 1316 bool "Enable madvise/fadvise syscalls" if EXPERT 1317 default y 1318 help 1319 This option enables the madvise and fadvise syscalls, used by 1320 applications to advise the kernel about their future memory or file 1321 usage, improving performance. If building an embedded system where no 1322 applications use these syscalls, you can disable this option to save 1323 space. 1324 1325config MEMBARRIER 1326 bool "Enable membarrier() system call" if EXPERT 1327 default y 1328 help 1329 Enable the membarrier() system call that allows issuing memory 1330 barriers across all running threads, which can be used to distribute 1331 the cost of user-space memory barriers asymmetrically by transforming 1332 pairs of memory barriers into pairs consisting of membarrier() and a 1333 compiler barrier. 1334 1335 If unsure, say Y. 1336 1337config CHECKPOINT_RESTORE 1338 bool "Checkpoint/restore support" if EXPERT 1339 select PROC_CHILDREN 1340 default n 1341 help 1342 Enables additional kernel features in a sake of checkpoint/restore. 1343 In particular it adds auxiliary prctl codes to setup process text, 1344 data and heap segment sizes, and a few additional /proc filesystem 1345 entries. 1346 1347 If unsure, say N here. 1348 1349config KALLSYMS 1350 bool "Load all symbols for debugging/ksymoops" if EXPERT 1351 default y 1352 help 1353 Say Y here to let the kernel print out symbolic crash information and 1354 symbolic stack backtraces. This increases the size of the kernel 1355 somewhat, as all symbols have to be loaded into the kernel image. 1356 1357config KALLSYMS_ALL 1358 bool "Include all symbols in kallsyms" 1359 depends on DEBUG_KERNEL && KALLSYMS 1360 help 1361 Normally kallsyms only contains the symbols of functions for nicer 1362 OOPS messages and backtraces (i.e., symbols from the text and inittext 1363 sections). This is sufficient for most cases. And only in very rare 1364 cases (e.g., when a debugger is used) all symbols are required (e.g., 1365 names of variables from the data sections, etc). 1366 1367 This option makes sure that all symbols are loaded into the kernel 1368 image (i.e., symbols from all sections) in cost of increased kernel 1369 size (depending on the kernel configuration, it may be 300KiB or 1370 something like this). 1371 1372 Say N unless you really need all symbols. 1373 1374config KALLSYMS_ABSOLUTE_PERCPU 1375 bool 1376 depends on KALLSYMS 1377 default X86_64 && SMP 1378 1379config KALLSYMS_BASE_RELATIVE 1380 bool 1381 depends on KALLSYMS 1382 default !IA64 1383 help 1384 Instead of emitting them as absolute values in the native word size, 1385 emit the symbol references in the kallsyms table as 32-bit entries, 1386 each containing a relative value in the range [base, base + U32_MAX] 1387 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either 1388 an absolute value in the range [0, S32_MAX] or a relative value in the 1389 range [base, base + S32_MAX], where base is the lowest relative symbol 1390 address encountered in the image. 1391 1392 On 64-bit builds, this reduces the size of the address table by 50%, 1393 but more importantly, it results in entries whose values are build 1394 time constants, and no relocation pass is required at runtime to fix 1395 up the entries based on the runtime load address of the kernel. 1396 1397# end of the "standard kernel features (expert users)" menu 1398 1399# syscall, maps, verifier 1400config BPF_SYSCALL 1401 bool "Enable bpf() system call" 1402 select ANON_INODES 1403 select BPF 1404 select IRQ_WORK 1405 default n 1406 help 1407 Enable the bpf() system call that allows to manipulate eBPF 1408 programs and maps via file descriptors. 1409 1410config BPF_JIT_ALWAYS_ON 1411 bool "Permanently enable BPF JIT and remove BPF interpreter" 1412 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT 1413 help 1414 Enables BPF JIT and removes BPF interpreter to avoid 1415 speculative execution of BPF instructions by the interpreter 1416 1417config USERFAULTFD 1418 bool "Enable userfaultfd() system call" 1419 select ANON_INODES 1420 depends on MMU 1421 help 1422 Enable the userfaultfd() system call that allows to intercept and 1423 handle page faults in userland. 1424 1425config ARCH_HAS_MEMBARRIER_CALLBACKS 1426 bool 1427 1428config ARCH_HAS_MEMBARRIER_SYNC_CORE 1429 bool 1430 1431config EMBEDDED 1432 bool "Embedded system" 1433 option allnoconfig_y 1434 select EXPERT 1435 help 1436 This option should be enabled if compiling the kernel for 1437 an embedded system so certain expert options are available 1438 for configuration. 1439 1440config HAVE_PERF_EVENTS 1441 bool 1442 help 1443 See tools/perf/design.txt for details. 1444 1445config PERF_USE_VMALLOC 1446 bool 1447 help 1448 See tools/perf/design.txt for details 1449 1450config PC104 1451 bool "PC/104 support" if EXPERT 1452 help 1453 Expose PC/104 form factor device drivers and options available for 1454 selection and configuration. Enable this option if your target 1455 machine has a PC/104 bus. 1456 1457menu "Kernel Performance Events And Counters" 1458 1459config PERF_EVENTS 1460 bool "Kernel performance events and counters" 1461 default y if PROFILING 1462 depends on HAVE_PERF_EVENTS 1463 select ANON_INODES 1464 select IRQ_WORK 1465 select SRCU 1466 help 1467 Enable kernel support for various performance events provided 1468 by software and hardware. 1469 1470 Software events are supported either built-in or via the 1471 use of generic tracepoints. 1472 1473 Most modern CPUs support performance events via performance 1474 counter registers. These registers count the number of certain 1475 types of hw events: such as instructions executed, cachemisses 1476 suffered, or branches mis-predicted - without slowing down the 1477 kernel or applications. These registers can also trigger interrupts 1478 when a threshold number of events have passed - and can thus be 1479 used to profile the code that runs on that CPU. 1480 1481 The Linux Performance Event subsystem provides an abstraction of 1482 these software and hardware event capabilities, available via a 1483 system call and used by the "perf" utility in tools/perf/. It 1484 provides per task and per CPU counters, and it provides event 1485 capabilities on top of those. 1486 1487 Say Y if unsure. 1488 1489config DEBUG_PERF_USE_VMALLOC 1490 default n 1491 bool "Debug: use vmalloc to back perf mmap() buffers" 1492 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC 1493 select PERF_USE_VMALLOC 1494 help 1495 Use vmalloc memory to back perf mmap() buffers. 1496 1497 Mostly useful for debugging the vmalloc code on platforms 1498 that don't require it. 1499 1500 Say N if unsure. 1501 1502endmenu 1503 1504config VM_EVENT_COUNTERS 1505 default y 1506 bool "Enable VM event counters for /proc/vmstat" if EXPERT 1507 help 1508 VM event counters are needed for event counts to be shown. 1509 This option allows the disabling of the VM event counters 1510 on EXPERT systems. /proc/vmstat will only show page counts 1511 if VM event counters are disabled. 1512 1513config SLUB_DEBUG 1514 default y 1515 bool "Enable SLUB debugging support" if EXPERT 1516 depends on SLUB && SYSFS 1517 help 1518 SLUB has extensive debug support features. Disabling these can 1519 result in significant savings in code size. This also disables 1520 SLUB sysfs support. /sys/slab will not exist and there will be 1521 no support for cache validation etc. 1522 1523config SLUB_MEMCG_SYSFS_ON 1524 default n 1525 bool "Enable memcg SLUB sysfs support by default" if EXPERT 1526 depends on SLUB && SYSFS && MEMCG 1527 help 1528 SLUB creates a directory under /sys/kernel/slab for each 1529 allocation cache to host info and debug files. If memory 1530 cgroup is enabled, each cache can have per memory cgroup 1531 caches. SLUB can create the same sysfs directories for these 1532 caches under /sys/kernel/slab/CACHE/cgroup but it can lead 1533 to a very high number of debug files being created. This is 1534 controlled by slub_memcg_sysfs boot parameter and this 1535 config option determines the parameter's default value. 1536 1537config COMPAT_BRK 1538 bool "Disable heap randomization" 1539 default y 1540 help 1541 Randomizing heap placement makes heap exploits harder, but it 1542 also breaks ancient binaries (including anything libc5 based). 1543 This option changes the bootup default to heap randomization 1544 disabled, and can be overridden at runtime by setting 1545 /proc/sys/kernel/randomize_va_space to 2. 1546 1547 On non-ancient distros (post-2000 ones) N is usually a safe choice. 1548 1549choice 1550 prompt "Choose SLAB allocator" 1551 default SLUB 1552 help 1553 This option allows to select a slab allocator. 1554 1555config SLAB 1556 bool "SLAB" 1557 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1558 help 1559 The regular slab allocator that is established and known to work 1560 well in all environments. It organizes cache hot objects in 1561 per cpu and per node queues. 1562 1563config SLUB 1564 bool "SLUB (Unqueued Allocator)" 1565 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1566 help 1567 SLUB is a slab allocator that minimizes cache line usage 1568 instead of managing queues of cached objects (SLAB approach). 1569 Per cpu caching is realized using slabs of objects instead 1570 of queues of objects. SLUB can use memory efficiently 1571 and has enhanced diagnostics. SLUB is the default choice for 1572 a slab allocator. 1573 1574config SLOB 1575 depends on EXPERT 1576 bool "SLOB (Simple Allocator)" 1577 help 1578 SLOB replaces the stock allocator with a drastically simpler 1579 allocator. SLOB is generally more space efficient but 1580 does not perform as well on large systems. 1581 1582endchoice 1583 1584config SLAB_MERGE_DEFAULT 1585 bool "Allow slab caches to be merged" 1586 default y 1587 help 1588 For reduced kernel memory fragmentation, slab caches can be 1589 merged when they share the same size and other characteristics. 1590 This carries a risk of kernel heap overflows being able to 1591 overwrite objects from merged caches (and more easily control 1592 cache layout), which makes such heap attacks easier to exploit 1593 by attackers. By keeping caches unmerged, these kinds of exploits 1594 can usually only damage objects in the same cache. To disable 1595 merging at runtime, "slab_nomerge" can be passed on the kernel 1596 command line. 1597 1598config SLAB_FREELIST_RANDOM 1599 default n 1600 depends on SLAB || SLUB 1601 bool "SLAB freelist randomization" 1602 help 1603 Randomizes the freelist order used on creating new pages. This 1604 security feature reduces the predictability of the kernel slab 1605 allocator against heap overflows. 1606 1607config SLAB_FREELIST_HARDENED 1608 bool "Harden slab freelist metadata" 1609 depends on SLUB 1610 help 1611 Many kernel heap attacks try to target slab cache metadata and 1612 other infrastructure. This options makes minor performance 1613 sacrifies to harden the kernel slab allocator against common 1614 freelist exploit methods. 1615 1616config SLUB_CPU_PARTIAL 1617 default y 1618 depends on SLUB && SMP 1619 bool "SLUB per cpu partial cache" 1620 help 1621 Per cpu partial caches accellerate objects allocation and freeing 1622 that is local to a processor at the price of more indeterminism 1623 in the latency of the free. On overflow these caches will be cleared 1624 which requires the taking of locks that may cause latency spikes. 1625 Typically one would choose no for a realtime system. 1626 1627config MMAP_ALLOW_UNINITIALIZED 1628 bool "Allow mmapped anonymous memory to be uninitialized" 1629 depends on EXPERT && !MMU 1630 default n 1631 help 1632 Normally, and according to the Linux spec, anonymous memory obtained 1633 from mmap() has it's contents cleared before it is passed to 1634 userspace. Enabling this config option allows you to request that 1635 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1636 providing a huge performance boost. If this option is not enabled, 1637 then the flag will be ignored. 1638 1639 This is taken advantage of by uClibc's malloc(), and also by 1640 ELF-FDPIC binfmt's brk and stack allocator. 1641 1642 Because of the obvious security issues, this option should only be 1643 enabled on embedded devices where you control what is run in 1644 userspace. Since that isn't generally a problem on no-MMU systems, 1645 it is normally safe to say Y here. 1646 1647 See Documentation/nommu-mmap.txt for more information. 1648 1649config SYSTEM_DATA_VERIFICATION 1650 def_bool n 1651 select SYSTEM_TRUSTED_KEYRING 1652 select KEYS 1653 select CRYPTO 1654 select CRYPTO_RSA 1655 select ASYMMETRIC_KEY_TYPE 1656 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1657 select ASN1 1658 select OID_REGISTRY 1659 select X509_CERTIFICATE_PARSER 1660 select PKCS7_MESSAGE_PARSER 1661 help 1662 Provide PKCS#7 message verification using the contents of the system 1663 trusted keyring to provide public keys. This then can be used for 1664 module verification, kexec image verification and firmware blob 1665 verification. 1666 1667config PROFILING 1668 bool "Profiling support" 1669 help 1670 Say Y here to enable the extended profiling support mechanisms used 1671 by profilers such as OProfile. 1672 1673# 1674# Place an empty function call at each tracepoint site. Can be 1675# dynamically changed for a probe function. 1676# 1677config TRACEPOINTS 1678 bool 1679 1680source "arch/Kconfig" 1681 1682endmenu # General setup 1683 1684config HAVE_GENERIC_DMA_COHERENT 1685 bool 1686 default n 1687 1688config RT_MUTEXES 1689 bool 1690 1691config BASE_SMALL 1692 int 1693 default 0 if BASE_FULL 1694 default 1 if !BASE_FULL 1695 1696menuconfig MODULES 1697 bool "Enable loadable module support" 1698 option modules 1699 help 1700 Kernel modules are small pieces of compiled code which can 1701 be inserted in the running kernel, rather than being 1702 permanently built into the kernel. You use the "modprobe" 1703 tool to add (and sometimes remove) them. If you say Y here, 1704 many parts of the kernel can be built as modules (by 1705 answering M instead of Y where indicated): this is most 1706 useful for infrequently used options which are not required 1707 for booting. For more information, see the man pages for 1708 modprobe, lsmod, modinfo, insmod and rmmod. 1709 1710 If you say Y here, you will need to run "make 1711 modules_install" to put the modules under /lib/modules/ 1712 where modprobe can find them (you may need to be root to do 1713 this). 1714 1715 If unsure, say Y. 1716 1717if MODULES 1718 1719config MODULE_FORCE_LOAD 1720 bool "Forced module loading" 1721 default n 1722 help 1723 Allow loading of modules without version information (ie. modprobe 1724 --force). Forced module loading sets the 'F' (forced) taint flag and 1725 is usually a really bad idea. 1726 1727config MODULE_UNLOAD 1728 bool "Module unloading" 1729 help 1730 Without this option you will not be able to unload any 1731 modules (note that some modules may not be unloadable 1732 anyway), which makes your kernel smaller, faster 1733 and simpler. If unsure, say Y. 1734 1735config MODULE_FORCE_UNLOAD 1736 bool "Forced module unloading" 1737 depends on MODULE_UNLOAD 1738 help 1739 This option allows you to force a module to unload, even if the 1740 kernel believes it is unsafe: the kernel will remove the module 1741 without waiting for anyone to stop using it (using the -f option to 1742 rmmod). This is mainly for kernel developers and desperate users. 1743 If unsure, say N. 1744 1745config MODVERSIONS 1746 bool "Module versioning support" 1747 help 1748 Usually, you have to use modules compiled with your kernel. 1749 Saying Y here makes it sometimes possible to use modules 1750 compiled for different kernels, by adding enough information 1751 to the modules to (hopefully) spot any changes which would 1752 make them incompatible with the kernel you are running. If 1753 unsure, say N. 1754 1755config MODULE_REL_CRCS 1756 bool 1757 depends on MODVERSIONS 1758 1759config MODULE_SRCVERSION_ALL 1760 bool "Source checksum for all modules" 1761 help 1762 Modules which contain a MODULE_VERSION get an extra "srcversion" 1763 field inserted into their modinfo section, which contains a 1764 sum of the source files which made it. This helps maintainers 1765 see exactly which source was used to build a module (since 1766 others sometimes change the module source without updating 1767 the version). With this option, such a "srcversion" field 1768 will be created for all modules. If unsure, say N. 1769 1770config MODULE_SIG 1771 bool "Module signature verification" 1772 depends on MODULES 1773 select SYSTEM_DATA_VERIFICATION 1774 help 1775 Check modules for valid signatures upon load: the signature 1776 is simply appended to the module. For more information see 1777 <file:Documentation/admin-guide/module-signing.rst>. 1778 1779 Note that this option adds the OpenSSL development packages as a 1780 kernel build dependency so that the signing tool can use its crypto 1781 library. 1782 1783 !!!WARNING!!! If you enable this option, you MUST make sure that the 1784 module DOES NOT get stripped after being signed. This includes the 1785 debuginfo strip done by some packagers (such as rpmbuild) and 1786 inclusion into an initramfs that wants the module size reduced. 1787 1788config MODULE_SIG_FORCE 1789 bool "Require modules to be validly signed" 1790 depends on MODULE_SIG 1791 help 1792 Reject unsigned modules or signed modules for which we don't have a 1793 key. Without this, such modules will simply taint the kernel. 1794 1795config MODULE_SIG_ALL 1796 bool "Automatically sign all modules" 1797 default y 1798 depends on MODULE_SIG 1799 help 1800 Sign all modules during make modules_install. Without this option, 1801 modules must be signed manually, using the scripts/sign-file tool. 1802 1803comment "Do not forget to sign required modules with scripts/sign-file" 1804 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 1805 1806choice 1807 prompt "Which hash algorithm should modules be signed with?" 1808 depends on MODULE_SIG 1809 help 1810 This determines which sort of hashing algorithm will be used during 1811 signature generation. This algorithm _must_ be built into the kernel 1812 directly so that signature verification can take place. It is not 1813 possible to load a signed module containing the algorithm to check 1814 the signature on that module. 1815 1816config MODULE_SIG_SHA1 1817 bool "Sign modules with SHA-1" 1818 select CRYPTO_SHA1 1819 1820config MODULE_SIG_SHA224 1821 bool "Sign modules with SHA-224" 1822 select CRYPTO_SHA256 1823 1824config MODULE_SIG_SHA256 1825 bool "Sign modules with SHA-256" 1826 select CRYPTO_SHA256 1827 1828config MODULE_SIG_SHA384 1829 bool "Sign modules with SHA-384" 1830 select CRYPTO_SHA512 1831 1832config MODULE_SIG_SHA512 1833 bool "Sign modules with SHA-512" 1834 select CRYPTO_SHA512 1835 1836endchoice 1837 1838config MODULE_SIG_HASH 1839 string 1840 depends on MODULE_SIG 1841 default "sha1" if MODULE_SIG_SHA1 1842 default "sha224" if MODULE_SIG_SHA224 1843 default "sha256" if MODULE_SIG_SHA256 1844 default "sha384" if MODULE_SIG_SHA384 1845 default "sha512" if MODULE_SIG_SHA512 1846 1847config MODULE_COMPRESS 1848 bool "Compress modules on installation" 1849 depends on MODULES 1850 help 1851 1852 Compresses kernel modules when 'make modules_install' is run; gzip or 1853 xz depending on "Compression algorithm" below. 1854 1855 module-init-tools MAY support gzip, and kmod MAY support gzip and xz. 1856 1857 Out-of-tree kernel modules installed using Kbuild will also be 1858 compressed upon installation. 1859 1860 Note: for modules inside an initrd or initramfs, it's more efficient 1861 to compress the whole initrd or initramfs instead. 1862 1863 Note: This is fully compatible with signed modules. 1864 1865 If in doubt, say N. 1866 1867choice 1868 prompt "Compression algorithm" 1869 depends on MODULE_COMPRESS 1870 default MODULE_COMPRESS_GZIP 1871 help 1872 This determines which sort of compression will be used during 1873 'make modules_install'. 1874 1875 GZIP (default) and XZ are supported. 1876 1877config MODULE_COMPRESS_GZIP 1878 bool "GZIP" 1879 1880config MODULE_COMPRESS_XZ 1881 bool "XZ" 1882 1883endchoice 1884 1885config TRIM_UNUSED_KSYMS 1886 bool "Trim unused exported kernel symbols" 1887 depends on MODULES && !UNUSED_SYMBOLS 1888 help 1889 The kernel and some modules make many symbols available for 1890 other modules to use via EXPORT_SYMBOL() and variants. Depending 1891 on the set of modules being selected in your kernel configuration, 1892 many of those exported symbols might never be used. 1893 1894 This option allows for unused exported symbols to be dropped from 1895 the build. In turn, this provides the compiler more opportunities 1896 (especially when using LTO) for optimizing the code and reducing 1897 binary size. This might have some security advantages as well. 1898 1899 If unsure, or if you need to build out-of-tree modules, say N. 1900 1901endif # MODULES 1902 1903config MODULES_TREE_LOOKUP 1904 def_bool y 1905 depends on PERF_EVENTS || TRACING 1906 1907config INIT_ALL_POSSIBLE 1908 bool 1909 help 1910 Back when each arch used to define their own cpu_online_mask and 1911 cpu_possible_mask, some of them chose to initialize cpu_possible_mask 1912 with all 1s, and others with all 0s. When they were centralised, 1913 it was better to provide this option than to break all the archs 1914 and have several arch maintainers pursuing me down dark alleys. 1915 1916source "block/Kconfig" 1917 1918config PREEMPT_NOTIFIERS 1919 bool 1920 1921config PADATA 1922 depends on SMP 1923 bool 1924 1925config ASN1 1926 tristate 1927 help 1928 Build a simple ASN.1 grammar compiler that produces a bytecode output 1929 that can be interpreted by the ASN.1 stream decoder and used to 1930 inform it as to what tags are to be expected in a stream and what 1931 functions to call on what tags. 1932 1933source "kernel/Kconfig.locks" 1934 1935config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE 1936 bool 1937 1938# It may be useful for an architecture to override the definitions of the 1939# SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h> 1940# and the COMPAT_ variants in <linux/compat.h>, in particular to use a 1941# different calling convention for syscalls. They can also override the 1942# macros for not-implemented syscalls in kernel/sys_ni.c and 1943# kernel/time/posix-stubs.c. All these overrides need to be available in 1944# <asm/syscall_wrapper.h>. 1945config ARCH_HAS_SYSCALL_WRAPPER 1946 def_bool n 1947