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 CC_IS_GCC 12 def_bool $(success,$(CC) --version | head -n 1 | grep -q gcc) 13 14config GCC_VERSION 15 int 16 default $(shell,$(srctree)/scripts/gcc-version.sh -p $(CC) | sed 's/^0*//') if CC_IS_GCC 17 default 0 18 19config CC_IS_CLANG 20 def_bool $(success,$(CC) --version | head -n 1 | grep -q clang) 21 22config CLANG_VERSION 23 int 24 default $(shell,$(srctree)/scripts/clang-version.sh $(CC)) 25 26config CONSTRUCTORS 27 bool 28 depends on !UML 29 30config IRQ_WORK 31 bool 32 33config BUILDTIME_EXTABLE_SORT 34 bool 35 36config THREAD_INFO_IN_TASK 37 bool 38 help 39 Select this to move thread_info off the stack into task_struct. To 40 make this work, an arch will need to remove all thread_info fields 41 except flags and fix any runtime bugs. 42 43 One subtle change that will be needed is to use try_get_task_stack() 44 and put_task_stack() in save_thread_stack_tsk() and get_wchan(). 45 46menu "General setup" 47 48config BROKEN 49 bool 50 51config BROKEN_ON_SMP 52 bool 53 depends on BROKEN || !SMP 54 default y 55 56config INIT_ENV_ARG_LIMIT 57 int 58 default 32 if !UML 59 default 128 if UML 60 help 61 Maximum of each of the number of arguments and environment 62 variables passed to init from the kernel command line. 63 64config COMPILE_TEST 65 bool "Compile also drivers which will not load" 66 depends on !UML 67 default n 68 help 69 Some drivers can be compiled on a different platform than they are 70 intended to be run on. Despite they cannot be loaded there (or even 71 when they load they cannot be used due to missing HW support), 72 developers still, opposing to distributors, might want to build such 73 drivers to compile-test them. 74 75 If you are a developer and want to build everything available, say Y 76 here. If you are a user/distributor, say N here to exclude useless 77 drivers to be distributed. 78 79config LOCALVERSION 80 string "Local version - append to kernel release" 81 help 82 Append an extra string to the end of your kernel version. 83 This will show up when you type uname, for example. 84 The string you set here will be appended after the contents of 85 any files with a filename matching localversion* in your 86 object and source tree, in that order. Your total string can 87 be a maximum of 64 characters. 88 89config LOCALVERSION_AUTO 90 bool "Automatically append version information to the version string" 91 default y 92 depends on !COMPILE_TEST 93 help 94 This will try to automatically determine if the current tree is a 95 release tree by looking for git tags that belong to the current 96 top of tree revision. 97 98 A string of the format -gxxxxxxxx will be added to the localversion 99 if a git-based tree is found. The string generated by this will be 100 appended after any matching localversion* files, and after the value 101 set in CONFIG_LOCALVERSION. 102 103 (The actual string used here is the first eight characters produced 104 by running the command: 105 106 $ git rev-parse --verify HEAD 107 108 which is done within the script "scripts/setlocalversion".) 109 110config BUILD_SALT 111 string "Build ID Salt" 112 default "" 113 help 114 The build ID is used to link binaries and their debug info. Setting 115 this option will use the value in the calculation of the build id. 116 This is mostly useful for distributions which want to ensure the 117 build is unique between builds. It's safe to leave the default. 118 119config HAVE_KERNEL_GZIP 120 bool 121 122config HAVE_KERNEL_BZIP2 123 bool 124 125config HAVE_KERNEL_LZMA 126 bool 127 128config HAVE_KERNEL_XZ 129 bool 130 131config HAVE_KERNEL_LZO 132 bool 133 134config HAVE_KERNEL_LZ4 135 bool 136 137choice 138 prompt "Kernel compression mode" 139 default KERNEL_GZIP 140 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 141 help 142 The linux kernel is a kind of self-extracting executable. 143 Several compression algorithms are available, which differ 144 in efficiency, compression and decompression speed. 145 Compression speed is only relevant when building a kernel. 146 Decompression speed is relevant at each boot. 147 148 If you have any problems with bzip2 or lzma compressed 149 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 150 version of this functionality (bzip2 only), for 2.4, was 151 supplied by Christian Ludwig) 152 153 High compression options are mostly useful for users, who 154 are low on disk space (embedded systems), but for whom ram 155 size matters less. 156 157 If in doubt, select 'gzip' 158 159config KERNEL_GZIP 160 bool "Gzip" 161 depends on HAVE_KERNEL_GZIP 162 help 163 The old and tried gzip compression. It provides a good balance 164 between compression ratio and decompression speed. 165 166config KERNEL_BZIP2 167 bool "Bzip2" 168 depends on HAVE_KERNEL_BZIP2 169 help 170 Its compression ratio and speed is intermediate. 171 Decompression speed is slowest among the choices. The kernel 172 size is about 10% smaller with bzip2, in comparison to gzip. 173 Bzip2 uses a large amount of memory. For modern kernels you 174 will need at least 8MB RAM or more for booting. 175 176config KERNEL_LZMA 177 bool "LZMA" 178 depends on HAVE_KERNEL_LZMA 179 help 180 This compression algorithm's ratio is best. Decompression speed 181 is between gzip and bzip2. Compression is slowest. 182 The kernel size is about 33% smaller with LZMA in comparison to gzip. 183 184config KERNEL_XZ 185 bool "XZ" 186 depends on HAVE_KERNEL_XZ 187 help 188 XZ uses the LZMA2 algorithm and instruction set specific 189 BCJ filters which can improve compression ratio of executable 190 code. The size of the kernel is about 30% smaller with XZ in 191 comparison to gzip. On architectures for which there is a BCJ 192 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 193 will create a few percent smaller kernel than plain LZMA. 194 195 The speed is about the same as with LZMA: The decompression 196 speed of XZ is better than that of bzip2 but worse than gzip 197 and LZO. Compression is slow. 198 199config KERNEL_LZO 200 bool "LZO" 201 depends on HAVE_KERNEL_LZO 202 help 203 Its compression ratio is the poorest among the choices. The kernel 204 size is about 10% bigger than gzip; however its speed 205 (both compression and decompression) is the fastest. 206 207config KERNEL_LZ4 208 bool "LZ4" 209 depends on HAVE_KERNEL_LZ4 210 help 211 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. 212 A preliminary version of LZ4 de/compression tool is available at 213 <https://code.google.com/p/lz4/>. 214 215 Its compression ratio is worse than LZO. The size of the kernel 216 is about 8% bigger than LZO. But the decompression speed is 217 faster than LZO. 218 219endchoice 220 221config DEFAULT_HOSTNAME 222 string "Default hostname" 223 default "(none)" 224 help 225 This option determines the default system hostname before userspace 226 calls sethostname(2). The kernel traditionally uses "(none)" here, 227 but you may wish to use a different default here to make a minimal 228 system more usable with less configuration. 229 230config SWAP 231 bool "Support for paging of anonymous memory (swap)" 232 depends on MMU && BLOCK 233 default y 234 help 235 This option allows you to choose whether you want to have support 236 for so called swap devices or swap files in your kernel that are 237 used to provide more virtual memory than the actual RAM present 238 in your computer. If unsure say Y. 239 240config SYSVIPC 241 bool "System V IPC" 242 ---help--- 243 Inter Process Communication is a suite of library functions and 244 system calls which let processes (running programs) synchronize and 245 exchange information. It is generally considered to be a good thing, 246 and some programs won't run unless you say Y here. In particular, if 247 you want to run the DOS emulator dosemu under Linux (read the 248 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 249 you'll need to say Y here. 250 251 You can find documentation about IPC with "info ipc" and also in 252 section 6.4 of the Linux Programmer's Guide, available from 253 <http://www.tldp.org/guides.html>. 254 255config SYSVIPC_SYSCTL 256 bool 257 depends on SYSVIPC 258 depends on SYSCTL 259 default y 260 261config POSIX_MQUEUE 262 bool "POSIX Message Queues" 263 depends on NET 264 ---help--- 265 POSIX variant of message queues is a part of IPC. In POSIX message 266 queues every message has a priority which decides about succession 267 of receiving it by a process. If you want to compile and run 268 programs written e.g. for Solaris with use of its POSIX message 269 queues (functions mq_*) say Y here. 270 271 POSIX message queues are visible as a filesystem called 'mqueue' 272 and can be mounted somewhere if you want to do filesystem 273 operations on message queues. 274 275 If unsure, say Y. 276 277config POSIX_MQUEUE_SYSCTL 278 bool 279 depends on POSIX_MQUEUE 280 depends on SYSCTL 281 default y 282 283config CROSS_MEMORY_ATTACH 284 bool "Enable process_vm_readv/writev syscalls" 285 depends on MMU 286 default y 287 help 288 Enabling this option adds the system calls process_vm_readv and 289 process_vm_writev which allow a process with the correct privileges 290 to directly read from or write to another process' address space. 291 See the man page for more details. 292 293config USELIB 294 bool "uselib syscall" 295 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION 296 help 297 This option enables the uselib syscall, a system call used in the 298 dynamic linker from libc5 and earlier. glibc does not use this 299 system call. If you intend to run programs built on libc5 or 300 earlier, you may need to enable this syscall. Current systems 301 running glibc can safely disable this. 302 303config AUDIT 304 bool "Auditing support" 305 depends on NET 306 help 307 Enable auditing infrastructure that can be used with another 308 kernel subsystem, such as SELinux (which requires this for 309 logging of avc messages output). System call auditing is included 310 on architectures which support it. 311 312config HAVE_ARCH_AUDITSYSCALL 313 bool 314 315config AUDITSYSCALL 316 def_bool y 317 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL 318 319config AUDIT_WATCH 320 def_bool y 321 depends on AUDITSYSCALL 322 select FSNOTIFY 323 324config AUDIT_TREE 325 def_bool y 326 depends on AUDITSYSCALL 327 select FSNOTIFY 328 329source "kernel/irq/Kconfig" 330source "kernel/time/Kconfig" 331 332menu "CPU/Task time and stats accounting" 333 334config VIRT_CPU_ACCOUNTING 335 bool 336 337choice 338 prompt "Cputime accounting" 339 default TICK_CPU_ACCOUNTING if !PPC64 340 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 341 342# Kind of a stub config for the pure tick based cputime accounting 343config TICK_CPU_ACCOUNTING 344 bool "Simple tick based cputime accounting" 345 depends on !S390 && !NO_HZ_FULL 346 help 347 This is the basic tick based cputime accounting that maintains 348 statistics about user, system and idle time spent on per jiffies 349 granularity. 350 351 If unsure, say Y. 352 353config VIRT_CPU_ACCOUNTING_NATIVE 354 bool "Deterministic task and CPU time accounting" 355 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 356 select VIRT_CPU_ACCOUNTING 357 help 358 Select this option to enable more accurate task and CPU time 359 accounting. This is done by reading a CPU counter on each 360 kernel entry and exit and on transitions within the kernel 361 between system, softirq and hardirq state, so there is a 362 small performance impact. In the case of s390 or IBM POWER > 5, 363 this also enables accounting of stolen time on logically-partitioned 364 systems. 365 366config VIRT_CPU_ACCOUNTING_GEN 367 bool "Full dynticks CPU time accounting" 368 depends on HAVE_CONTEXT_TRACKING 369 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN 370 select VIRT_CPU_ACCOUNTING 371 select CONTEXT_TRACKING 372 help 373 Select this option to enable task and CPU time accounting on full 374 dynticks systems. This accounting is implemented by watching every 375 kernel-user boundaries using the context tracking subsystem. 376 The accounting is thus performed at the expense of some significant 377 overhead. 378 379 For now this is only useful if you are working on the full 380 dynticks subsystem development. 381 382 If unsure, say N. 383 384endchoice 385 386config IRQ_TIME_ACCOUNTING 387 bool "Fine granularity task level IRQ time accounting" 388 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE 389 help 390 Select this option to enable fine granularity task irq time 391 accounting. This is done by reading a timestamp on each 392 transitions between softirq and hardirq state, so there can be a 393 small performance impact. 394 395 If in doubt, say N here. 396 397config BSD_PROCESS_ACCT 398 bool "BSD Process Accounting" 399 depends on MULTIUSER 400 help 401 If you say Y here, a user level program will be able to instruct the 402 kernel (via a special system call) to write process accounting 403 information to a file: whenever a process exits, information about 404 that process will be appended to the file by the kernel. The 405 information includes things such as creation time, owning user, 406 command name, memory usage, controlling terminal etc. (the complete 407 list is in the struct acct in <file:include/linux/acct.h>). It is 408 up to the user level program to do useful things with this 409 information. This is generally a good idea, so say Y. 410 411config BSD_PROCESS_ACCT_V3 412 bool "BSD Process Accounting version 3 file format" 413 depends on BSD_PROCESS_ACCT 414 default n 415 help 416 If you say Y here, the process accounting information is written 417 in a new file format that also logs the process IDs of each 418 process and it's parent. Note that this file format is incompatible 419 with previous v0/v1/v2 file formats, so you will need updated tools 420 for processing it. A preliminary version of these tools is available 421 at <http://www.gnu.org/software/acct/>. 422 423config TASKSTATS 424 bool "Export task/process statistics through netlink" 425 depends on NET 426 depends on MULTIUSER 427 default n 428 help 429 Export selected statistics for tasks/processes through the 430 generic netlink interface. Unlike BSD process accounting, the 431 statistics are available during the lifetime of tasks/processes as 432 responses to commands. Like BSD accounting, they are sent to user 433 space on task exit. 434 435 Say N if unsure. 436 437config TASK_DELAY_ACCT 438 bool "Enable per-task delay accounting" 439 depends on TASKSTATS 440 select SCHED_INFO 441 help 442 Collect information on time spent by a task waiting for system 443 resources like cpu, synchronous block I/O completion and swapping 444 in pages. Such statistics can help in setting a task's priorities 445 relative to other tasks for cpu, io, rss limits etc. 446 447 Say N if unsure. 448 449config TASK_XACCT 450 bool "Enable extended accounting over taskstats" 451 depends on TASKSTATS 452 help 453 Collect extended task accounting data and send the data 454 to userland for processing over the taskstats interface. 455 456 Say N if unsure. 457 458config TASK_IO_ACCOUNTING 459 bool "Enable per-task storage I/O accounting" 460 depends on TASK_XACCT 461 help 462 Collect information on the number of bytes of storage I/O which this 463 task has caused. 464 465 Say N if unsure. 466 467endmenu # "CPU/Task time and stats accounting" 468 469config CPU_ISOLATION 470 bool "CPU isolation" 471 depends on SMP || COMPILE_TEST 472 default y 473 help 474 Make sure that CPUs running critical tasks are not disturbed by 475 any source of "noise" such as unbound workqueues, timers, kthreads... 476 Unbound jobs get offloaded to housekeeping CPUs. This is driven by 477 the "isolcpus=" boot parameter. 478 479 Say Y if unsure. 480 481source "kernel/rcu/Kconfig" 482 483config BUILD_BIN2C 484 bool 485 default n 486 487config IKCONFIG 488 tristate "Kernel .config support" 489 select BUILD_BIN2C 490 ---help--- 491 This option enables the complete Linux kernel ".config" file 492 contents to be saved in the kernel. It provides documentation 493 of which kernel options are used in a running kernel or in an 494 on-disk kernel. This information can be extracted from the kernel 495 image file with the script scripts/extract-ikconfig and used as 496 input to rebuild the current kernel or to build another kernel. 497 It can also be extracted from a running kernel by reading 498 /proc/config.gz if enabled (below). 499 500config IKCONFIG_PROC 501 bool "Enable access to .config through /proc/config.gz" 502 depends on IKCONFIG && PROC_FS 503 ---help--- 504 This option enables access to the kernel configuration file 505 through /proc/config.gz. 506 507config LOG_BUF_SHIFT 508 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 509 range 12 25 510 default 17 511 depends on PRINTK 512 help 513 Select the minimal kernel log buffer size as a power of 2. 514 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config 515 parameter, see below. Any higher size also might be forced 516 by "log_buf_len" boot parameter. 517 518 Examples: 519 17 => 128 KB 520 16 => 64 KB 521 15 => 32 KB 522 14 => 16 KB 523 13 => 8 KB 524 12 => 4 KB 525 526config LOG_CPU_MAX_BUF_SHIFT 527 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" 528 depends on SMP 529 range 0 21 530 default 12 if !BASE_SMALL 531 default 0 if BASE_SMALL 532 depends on PRINTK 533 help 534 This option allows to increase the default ring buffer size 535 according to the number of CPUs. The value defines the contribution 536 of each CPU as a power of 2. The used space is typically only few 537 lines however it might be much more when problems are reported, 538 e.g. backtraces. 539 540 The increased size means that a new buffer has to be allocated and 541 the original static one is unused. It makes sense only on systems 542 with more CPUs. Therefore this value is used only when the sum of 543 contributions is greater than the half of the default kernel ring 544 buffer as defined by LOG_BUF_SHIFT. The default values are set 545 so that more than 64 CPUs are needed to trigger the allocation. 546 547 Also this option is ignored when "log_buf_len" kernel parameter is 548 used as it forces an exact (power of two) size of the ring buffer. 549 550 The number of possible CPUs is used for this computation ignoring 551 hotplugging making the computation optimal for the worst case 552 scenario while allowing a simple algorithm to be used from bootup. 553 554 Examples shift values and their meaning: 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 562config PRINTK_SAFE_LOG_BUF_SHIFT 563 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)" 564 range 10 21 565 default 13 566 depends on PRINTK 567 help 568 Select the size of an alternate printk per-CPU buffer where messages 569 printed from usafe contexts are temporary stored. One example would 570 be NMI messages, another one - printk recursion. The messages are 571 copied to the main log buffer in a safe context to avoid a deadlock. 572 The value defines the size as a power of 2. 573 574 Those messages are rare and limited. The largest one is when 575 a backtrace is printed. It usually fits into 4KB. Select 576 8KB if you want to be on the safe side. 577 578 Examples: 579 17 => 128 KB for each CPU 580 16 => 64 KB for each CPU 581 15 => 32 KB for each CPU 582 14 => 16 KB for each CPU 583 13 => 8 KB for each CPU 584 12 => 4 KB for each CPU 585 586# 587# Architectures with an unreliable sched_clock() should select this: 588# 589config HAVE_UNSTABLE_SCHED_CLOCK 590 bool 591 592config GENERIC_SCHED_CLOCK 593 bool 594 595# 596# For architectures that want to enable the support for NUMA-affine scheduler 597# balancing logic: 598# 599config ARCH_SUPPORTS_NUMA_BALANCING 600 bool 601 602# 603# For architectures that prefer to flush all TLBs after a number of pages 604# are unmapped instead of sending one IPI per page to flush. The architecture 605# must provide guarantees on what happens if a clean TLB cache entry is 606# written after the unmap. Details are in mm/rmap.c near the check for 607# should_defer_flush. The architecture should also consider if the full flush 608# and the refill costs are offset by the savings of sending fewer IPIs. 609config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH 610 bool 611 612# 613# For architectures that know their GCC __int128 support is sound 614# 615config ARCH_SUPPORTS_INT128 616 bool 617 618# For architectures that (ab)use NUMA to represent different memory regions 619# all cpu-local but of different latencies, such as SuperH. 620# 621config ARCH_WANT_NUMA_VARIABLE_LOCALITY 622 bool 623 624config NUMA_BALANCING 625 bool "Memory placement aware NUMA scheduler" 626 depends on ARCH_SUPPORTS_NUMA_BALANCING 627 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 628 depends on SMP && NUMA && MIGRATION 629 help 630 This option adds support for automatic NUMA aware memory/task placement. 631 The mechanism is quite primitive and is based on migrating memory when 632 it has references to the node the task is running on. 633 634 This system will be inactive on UMA systems. 635 636config NUMA_BALANCING_DEFAULT_ENABLED 637 bool "Automatically enable NUMA aware memory/task placement" 638 default y 639 depends on NUMA_BALANCING 640 help 641 If set, automatic NUMA balancing will be enabled if running on a NUMA 642 machine. 643 644menuconfig CGROUPS 645 bool "Control Group support" 646 select KERNFS 647 help 648 This option adds support for grouping sets of processes together, for 649 use with process control subsystems such as Cpusets, CFS, memory 650 controls or device isolation. 651 See 652 - Documentation/scheduler/sched-design-CFS.txt (CFS) 653 - Documentation/cgroup-v1/ (features for grouping, isolation 654 and resource control) 655 656 Say N if unsure. 657 658if CGROUPS 659 660config PAGE_COUNTER 661 bool 662 663config MEMCG 664 bool "Memory controller" 665 select PAGE_COUNTER 666 select EVENTFD 667 help 668 Provides control over the memory footprint of tasks in a cgroup. 669 670config MEMCG_SWAP 671 bool "Swap controller" 672 depends on MEMCG && SWAP 673 help 674 Provides control over the swap space consumed by tasks in a cgroup. 675 676config MEMCG_SWAP_ENABLED 677 bool "Swap controller enabled by default" 678 depends on MEMCG_SWAP 679 default y 680 help 681 Memory Resource Controller Swap Extension comes with its price in 682 a bigger memory consumption. General purpose distribution kernels 683 which want to enable the feature but keep it disabled by default 684 and let the user enable it by swapaccount=1 boot command line 685 parameter should have this option unselected. 686 For those who want to have the feature enabled by default should 687 select this option (if, for some reason, they need to disable it 688 then swapaccount=0 does the trick). 689 690config BLK_CGROUP 691 bool "IO controller" 692 depends on BLOCK 693 default n 694 ---help--- 695 Generic block IO controller cgroup interface. This is the common 696 cgroup interface which should be used by various IO controlling 697 policies. 698 699 Currently, CFQ IO scheduler uses it to recognize task groups and 700 control disk bandwidth allocation (proportional time slice allocation) 701 to such task groups. It is also used by bio throttling logic in 702 block layer to implement upper limit in IO rates on a device. 703 704 This option only enables generic Block IO controller infrastructure. 705 One needs to also enable actual IO controlling logic/policy. For 706 enabling proportional weight division of disk bandwidth in CFQ, set 707 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 708 CONFIG_BLK_DEV_THROTTLING=y. 709 710 See Documentation/cgroup-v1/blkio-controller.txt for more information. 711 712config DEBUG_BLK_CGROUP 713 bool "IO controller debugging" 714 depends on BLK_CGROUP 715 default n 716 ---help--- 717 Enable some debugging help. Currently it exports additional stat 718 files in a cgroup which can be useful for debugging. 719 720config CGROUP_WRITEBACK 721 bool 722 depends on MEMCG && BLK_CGROUP 723 default y 724 725menuconfig CGROUP_SCHED 726 bool "CPU controller" 727 default n 728 help 729 This feature lets CPU scheduler recognize task groups and control CPU 730 bandwidth allocation to such task groups. It uses cgroups to group 731 tasks. 732 733if CGROUP_SCHED 734config FAIR_GROUP_SCHED 735 bool "Group scheduling for SCHED_OTHER" 736 depends on CGROUP_SCHED 737 default CGROUP_SCHED 738 739config CFS_BANDWIDTH 740 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 741 depends on FAIR_GROUP_SCHED 742 default n 743 help 744 This option allows users to define CPU bandwidth rates (limits) for 745 tasks running within the fair group scheduler. Groups with no limit 746 set are considered to be unconstrained and will run with no 747 restriction. 748 See Documentation/scheduler/sched-bwc.txt for more information. 749 750config RT_GROUP_SCHED 751 bool "Group scheduling for SCHED_RR/FIFO" 752 depends on CGROUP_SCHED 753 default n 754 help 755 This feature lets you explicitly allocate real CPU bandwidth 756 to task groups. If enabled, it will also make it impossible to 757 schedule realtime tasks for non-root users until you allocate 758 realtime bandwidth for them. 759 See Documentation/scheduler/sched-rt-group.txt for more information. 760 761endif #CGROUP_SCHED 762 763config CGROUP_PIDS 764 bool "PIDs controller" 765 help 766 Provides enforcement of process number limits in the scope of a 767 cgroup. Any attempt to fork more processes than is allowed in the 768 cgroup will fail. PIDs are fundamentally a global resource because it 769 is fairly trivial to reach PID exhaustion before you reach even a 770 conservative kmemcg limit. As a result, it is possible to grind a 771 system to halt without being limited by other cgroup policies. The 772 PIDs controller is designed to stop this from happening. 773 774 It should be noted that organisational operations (such as attaching 775 to a cgroup hierarchy will *not* be blocked by the PIDs controller), 776 since the PIDs limit only affects a process's ability to fork, not to 777 attach to a cgroup. 778 779config CGROUP_RDMA 780 bool "RDMA controller" 781 help 782 Provides enforcement of RDMA resources defined by IB stack. 783 It is fairly easy for consumers to exhaust RDMA resources, which 784 can result into resource unavailability to other consumers. 785 RDMA controller is designed to stop this from happening. 786 Attaching processes with active RDMA resources to the cgroup 787 hierarchy is allowed even if can cross the hierarchy's limit. 788 789config CGROUP_FREEZER 790 bool "Freezer controller" 791 help 792 Provides a way to freeze and unfreeze all tasks in a 793 cgroup. 794 795 This option affects the ORIGINAL cgroup interface. The cgroup2 memory 796 controller includes important in-kernel memory consumers per default. 797 798 If you're using cgroup2, say N. 799 800config CGROUP_HUGETLB 801 bool "HugeTLB controller" 802 depends on HUGETLB_PAGE 803 select PAGE_COUNTER 804 default n 805 help 806 Provides a cgroup controller for HugeTLB pages. 807 When you enable this, you can put a per cgroup limit on HugeTLB usage. 808 The limit is enforced during page fault. Since HugeTLB doesn't 809 support page reclaim, enforcing the limit at page fault time implies 810 that, the application will get SIGBUS signal if it tries to access 811 HugeTLB pages beyond its limit. This requires the application to know 812 beforehand how much HugeTLB pages it would require for its use. The 813 control group is tracked in the third page lru pointer. This means 814 that we cannot use the controller with huge page less than 3 pages. 815 816config CPUSETS 817 bool "Cpuset controller" 818 depends on SMP 819 help 820 This option will let you create and manage CPUSETs which 821 allow dynamically partitioning a system into sets of CPUs and 822 Memory Nodes and assigning tasks to run only within those sets. 823 This is primarily useful on large SMP or NUMA systems. 824 825 Say N if unsure. 826 827config PROC_PID_CPUSET 828 bool "Include legacy /proc/<pid>/cpuset file" 829 depends on CPUSETS 830 default y 831 832config CGROUP_DEVICE 833 bool "Device controller" 834 help 835 Provides a cgroup controller implementing whitelists for 836 devices which a process in the cgroup can mknod or open. 837 838config CGROUP_CPUACCT 839 bool "Simple CPU accounting controller" 840 help 841 Provides a simple controller for monitoring the 842 total CPU consumed by the tasks in a cgroup. 843 844config CGROUP_PERF 845 bool "Perf controller" 846 depends on PERF_EVENTS 847 help 848 This option extends the perf per-cpu mode to restrict monitoring 849 to threads which belong to the cgroup specified and run on the 850 designated cpu. 851 852 Say N if unsure. 853 854config CGROUP_BPF 855 bool "Support for eBPF programs attached to cgroups" 856 depends on BPF_SYSCALL 857 select SOCK_CGROUP_DATA 858 help 859 Allow attaching eBPF programs to a cgroup using the bpf(2) 860 syscall command BPF_PROG_ATTACH. 861 862 In which context these programs are accessed depends on the type 863 of attachment. For instance, programs that are attached using 864 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of 865 inet sockets. 866 867config CGROUP_DEBUG 868 bool "Debug controller" 869 default n 870 depends on DEBUG_KERNEL 871 help 872 This option enables a simple controller that exports 873 debugging information about the cgroups framework. This 874 controller is for control cgroup debugging only. Its 875 interfaces are not stable. 876 877 Say N. 878 879config SOCK_CGROUP_DATA 880 bool 881 default n 882 883endif # CGROUPS 884 885menuconfig NAMESPACES 886 bool "Namespaces support" if EXPERT 887 depends on MULTIUSER 888 default !EXPERT 889 help 890 Provides the way to make tasks work with different objects using 891 the same id. For example same IPC id may refer to different objects 892 or same user id or pid may refer to different tasks when used in 893 different namespaces. 894 895if NAMESPACES 896 897config UTS_NS 898 bool "UTS namespace" 899 default y 900 help 901 In this namespace tasks see different info provided with the 902 uname() system call 903 904config IPC_NS 905 bool "IPC namespace" 906 depends on (SYSVIPC || POSIX_MQUEUE) 907 default y 908 help 909 In this namespace tasks work with IPC ids which correspond to 910 different IPC objects in different namespaces. 911 912config USER_NS 913 bool "User namespace" 914 default n 915 help 916 This allows containers, i.e. vservers, to use user namespaces 917 to provide different user info for different servers. 918 919 When user namespaces are enabled in the kernel it is 920 recommended that the MEMCG option also be enabled and that 921 user-space use the memory control groups to limit the amount 922 of memory a memory unprivileged users can use. 923 924 If unsure, say N. 925 926config PID_NS 927 bool "PID Namespaces" 928 default y 929 help 930 Support process id namespaces. This allows having multiple 931 processes with the same pid as long as they are in different 932 pid namespaces. This is a building block of containers. 933 934config NET_NS 935 bool "Network namespace" 936 depends on NET 937 default y 938 help 939 Allow user space to create what appear to be multiple instances 940 of the network stack. 941 942endif # NAMESPACES 943 944config SCHED_AUTOGROUP 945 bool "Automatic process group scheduling" 946 select CGROUPS 947 select CGROUP_SCHED 948 select FAIR_GROUP_SCHED 949 help 950 This option optimizes the scheduler for common desktop workloads by 951 automatically creating and populating task groups. This separation 952 of workloads isolates aggressive CPU burners (like build jobs) from 953 desktop applications. Task group autogeneration is currently based 954 upon task session. 955 956config SYSFS_DEPRECATED 957 bool "Enable deprecated sysfs features to support old userspace tools" 958 depends on SYSFS 959 default n 960 help 961 This option adds code that switches the layout of the "block" class 962 devices, to not show up in /sys/class/block/, but only in 963 /sys/block/. 964 965 This switch is only active when the sysfs.deprecated=1 boot option is 966 passed or the SYSFS_DEPRECATED_V2 option is set. 967 968 This option allows new kernels to run on old distributions and tools, 969 which might get confused by /sys/class/block/. Since 2007/2008 all 970 major distributions and tools handle this just fine. 971 972 Recent distributions and userspace tools after 2009/2010 depend on 973 the existence of /sys/class/block/, and will not work with this 974 option enabled. 975 976 Only if you are using a new kernel on an old distribution, you might 977 need to say Y here. 978 979config SYSFS_DEPRECATED_V2 980 bool "Enable deprecated sysfs features by default" 981 default n 982 depends on SYSFS 983 depends on SYSFS_DEPRECATED 984 help 985 Enable deprecated sysfs by default. 986 987 See the CONFIG_SYSFS_DEPRECATED option for more details about this 988 option. 989 990 Only if you are using a new kernel on an old distribution, you might 991 need to say Y here. Even then, odds are you would not need it 992 enabled, you can always pass the boot option if absolutely necessary. 993 994config RELAY 995 bool "Kernel->user space relay support (formerly relayfs)" 996 select IRQ_WORK 997 help 998 This option enables support for relay interface support in 999 certain file systems (such as debugfs). 1000 It is designed to provide an efficient mechanism for tools and 1001 facilities to relay large amounts of data from kernel space to 1002 user space. 1003 1004 If unsure, say N. 1005 1006config BLK_DEV_INITRD 1007 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 1008 help 1009 The initial RAM filesystem is a ramfs which is loaded by the 1010 boot loader (loadlin or lilo) and that is mounted as root 1011 before the normal boot procedure. It is typically used to 1012 load modules needed to mount the "real" root file system, 1013 etc. See <file:Documentation/admin-guide/initrd.rst> for details. 1014 1015 If RAM disk support (BLK_DEV_RAM) is also included, this 1016 also enables initial RAM disk (initrd) support and adds 1017 15 Kbytes (more on some other architectures) to the kernel size. 1018 1019 If unsure say Y. 1020 1021if BLK_DEV_INITRD 1022 1023source "usr/Kconfig" 1024 1025endif 1026 1027choice 1028 prompt "Compiler optimization level" 1029 default CC_OPTIMIZE_FOR_PERFORMANCE 1030 1031config CC_OPTIMIZE_FOR_PERFORMANCE 1032 bool "Optimize for performance" 1033 help 1034 This is the default optimization level for the kernel, building 1035 with the "-O2" compiler flag for best performance and most 1036 helpful compile-time warnings. 1037 1038config CC_OPTIMIZE_FOR_SIZE 1039 bool "Optimize for size" 1040 help 1041 Enabling this option will pass "-Os" instead of "-O2" to 1042 your compiler resulting in a smaller kernel. 1043 1044 If unsure, say N. 1045 1046endchoice 1047 1048config HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1049 bool 1050 help 1051 This requires that the arch annotates or otherwise protects 1052 its external entry points from being discarded. Linker scripts 1053 must also merge .text.*, .data.*, and .bss.* correctly into 1054 output sections. Care must be taken not to pull in unrelated 1055 sections (e.g., '.text.init'). Typically '.' in section names 1056 is used to distinguish them from label names / C identifiers. 1057 1058config LD_DEAD_CODE_DATA_ELIMINATION 1059 bool "Dead code and data elimination (EXPERIMENTAL)" 1060 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1061 depends on EXPERT 1062 help 1063 Enable this if you want to do dead code and data elimination with 1064 the linker by compiling with -ffunction-sections -fdata-sections, 1065 and linking with --gc-sections. 1066 1067 This can reduce on disk and in-memory size of the kernel 1068 code and static data, particularly for small configs and 1069 on small systems. This has the possibility of introducing 1070 silently broken kernel if the required annotations are not 1071 present. This option is not well tested yet, so use at your 1072 own risk. 1073 1074config SYSCTL 1075 bool 1076 1077config ANON_INODES 1078 bool 1079 1080config HAVE_UID16 1081 bool 1082 1083config SYSCTL_EXCEPTION_TRACE 1084 bool 1085 help 1086 Enable support for /proc/sys/debug/exception-trace. 1087 1088config SYSCTL_ARCH_UNALIGN_NO_WARN 1089 bool 1090 help 1091 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1092 Allows arch to define/use @no_unaligned_warning to possibly warn 1093 about unaligned access emulation going on under the hood. 1094 1095config SYSCTL_ARCH_UNALIGN_ALLOW 1096 bool 1097 help 1098 Enable support for /proc/sys/kernel/unaligned-trap 1099 Allows arches to define/use @unaligned_enabled to runtime toggle 1100 the unaligned access emulation. 1101 see arch/parisc/kernel/unaligned.c for reference 1102 1103config HAVE_PCSPKR_PLATFORM 1104 bool 1105 1106# interpreter that classic socket filters depend on 1107config BPF 1108 bool 1109 1110menuconfig EXPERT 1111 bool "Configure standard kernel features (expert users)" 1112 # Unhide debug options, to make the on-by-default options visible 1113 select DEBUG_KERNEL 1114 help 1115 This option allows certain base kernel options and settings 1116 to be disabled or tweaked. This is for specialized 1117 environments which can tolerate a "non-standard" kernel. 1118 Only use this if you really know what you are doing. 1119 1120config UID16 1121 bool "Enable 16-bit UID system calls" if EXPERT 1122 depends on HAVE_UID16 && MULTIUSER 1123 default y 1124 help 1125 This enables the legacy 16-bit UID syscall wrappers. 1126 1127config MULTIUSER 1128 bool "Multiple users, groups and capabilities support" if EXPERT 1129 default y 1130 help 1131 This option enables support for non-root users, groups and 1132 capabilities. 1133 1134 If you say N here, all processes will run with UID 0, GID 0, and all 1135 possible capabilities. Saying N here also compiles out support for 1136 system calls related to UIDs, GIDs, and capabilities, such as setuid, 1137 setgid, and capset. 1138 1139 If unsure, say Y here. 1140 1141config SGETMASK_SYSCALL 1142 bool "sgetmask/ssetmask syscalls support" if EXPERT 1143 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH 1144 ---help--- 1145 sys_sgetmask and sys_ssetmask are obsolete system calls 1146 no longer supported in libc but still enabled by default in some 1147 architectures. 1148 1149 If unsure, leave the default option here. 1150 1151config SYSFS_SYSCALL 1152 bool "Sysfs syscall support" if EXPERT 1153 default y 1154 ---help--- 1155 sys_sysfs is an obsolete system call no longer supported in libc. 1156 Note that disabling this option is more secure but might break 1157 compatibility with some systems. 1158 1159 If unsure say Y here. 1160 1161config SYSCTL_SYSCALL 1162 bool "Sysctl syscall support" if EXPERT 1163 depends on PROC_SYSCTL 1164 default n 1165 select SYSCTL 1166 ---help--- 1167 sys_sysctl uses binary paths that have been found challenging 1168 to properly maintain and use. The interface in /proc/sys 1169 using paths with ascii names is now the primary path to this 1170 information. 1171 1172 Almost nothing using the binary sysctl interface so if you are 1173 trying to save some space it is probably safe to disable this, 1174 making your kernel marginally smaller. 1175 1176 If unsure say N here. 1177 1178config FHANDLE 1179 bool "open by fhandle syscalls" if EXPERT 1180 select EXPORTFS 1181 default y 1182 help 1183 If you say Y here, a user level program will be able to map 1184 file names to handle and then later use the handle for 1185 different file system operations. This is useful in implementing 1186 userspace file servers, which now track files using handles instead 1187 of names. The handle would remain the same even if file names 1188 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 1189 syscalls. 1190 1191config POSIX_TIMERS 1192 bool "Posix Clocks & timers" if EXPERT 1193 default y 1194 help 1195 This includes native support for POSIX timers to the kernel. 1196 Some embedded systems have no use for them and therefore they 1197 can be configured out to reduce the size of the kernel image. 1198 1199 When this option is disabled, the following syscalls won't be 1200 available: timer_create, timer_gettime: timer_getoverrun, 1201 timer_settime, timer_delete, clock_adjtime, getitimer, 1202 setitimer, alarm. Furthermore, the clock_settime, clock_gettime, 1203 clock_getres and clock_nanosleep syscalls will be limited to 1204 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only. 1205 1206 If unsure say y. 1207 1208config PRINTK 1209 default y 1210 bool "Enable support for printk" if EXPERT 1211 select IRQ_WORK 1212 help 1213 This option enables normal printk support. Removing it 1214 eliminates most of the message strings from the kernel image 1215 and makes the kernel more or less silent. As this makes it 1216 very difficult to diagnose system problems, saying N here is 1217 strongly discouraged. 1218 1219config PRINTK_NMI 1220 def_bool y 1221 depends on PRINTK 1222 depends on HAVE_NMI 1223 1224config BUG 1225 bool "BUG() support" if EXPERT 1226 default y 1227 help 1228 Disabling this option eliminates support for BUG and WARN, reducing 1229 the size of your kernel image and potentially quietly ignoring 1230 numerous fatal conditions. You should only consider disabling this 1231 option for embedded systems with no facilities for reporting errors. 1232 Just say Y. 1233 1234config ELF_CORE 1235 depends on COREDUMP 1236 default y 1237 bool "Enable ELF core dumps" if EXPERT 1238 help 1239 Enable support for generating core dumps. Disabling saves about 4k. 1240 1241 1242config PCSPKR_PLATFORM 1243 bool "Enable PC-Speaker support" if EXPERT 1244 depends on HAVE_PCSPKR_PLATFORM 1245 select I8253_LOCK 1246 default y 1247 help 1248 This option allows to disable the internal PC-Speaker 1249 support, saving some memory. 1250 1251config BASE_FULL 1252 default y 1253 bool "Enable full-sized data structures for core" if EXPERT 1254 help 1255 Disabling this option reduces the size of miscellaneous core 1256 kernel data structures. This saves memory on small machines, 1257 but may reduce performance. 1258 1259config FUTEX 1260 bool "Enable futex support" if EXPERT 1261 default y 1262 imply RT_MUTEXES 1263 help 1264 Disabling this option will cause the kernel to be built without 1265 support for "fast userspace mutexes". The resulting kernel may not 1266 run glibc-based applications correctly. 1267 1268config FUTEX_PI 1269 bool 1270 depends on FUTEX && RT_MUTEXES 1271 default y 1272 1273config HAVE_FUTEX_CMPXCHG 1274 bool 1275 depends on FUTEX 1276 help 1277 Architectures should select this if futex_atomic_cmpxchg_inatomic() 1278 is implemented and always working. This removes a couple of runtime 1279 checks. 1280 1281config EPOLL 1282 bool "Enable eventpoll support" if EXPERT 1283 default y 1284 select ANON_INODES 1285 help 1286 Disabling this option will cause the kernel to be built without 1287 support for epoll family of system calls. 1288 1289config SIGNALFD 1290 bool "Enable signalfd() system call" if EXPERT 1291 select ANON_INODES 1292 default y 1293 help 1294 Enable the signalfd() system call that allows to receive signals 1295 on a file descriptor. 1296 1297 If unsure, say Y. 1298 1299config TIMERFD 1300 bool "Enable timerfd() system call" if EXPERT 1301 select ANON_INODES 1302 default y 1303 help 1304 Enable the timerfd() system call that allows to receive timer 1305 events on a file descriptor. 1306 1307 If unsure, say Y. 1308 1309config EVENTFD 1310 bool "Enable eventfd() system call" if EXPERT 1311 select ANON_INODES 1312 default y 1313 help 1314 Enable the eventfd() system call that allows to receive both 1315 kernel notification (ie. KAIO) or userspace notifications. 1316 1317 If unsure, say Y. 1318 1319config SHMEM 1320 bool "Use full shmem filesystem" if EXPERT 1321 default y 1322 depends on MMU 1323 help 1324 The shmem is an internal filesystem used to manage shared memory. 1325 It is backed by swap and manages resource limits. It is also exported 1326 to userspace as tmpfs if TMPFS is enabled. Disabling this 1327 option replaces shmem and tmpfs with the much simpler ramfs code, 1328 which may be appropriate on small systems without swap. 1329 1330config AIO 1331 bool "Enable AIO support" if EXPERT 1332 default y 1333 help 1334 This option enables POSIX asynchronous I/O which may by used 1335 by some high performance threaded applications. Disabling 1336 this option saves about 7k. 1337 1338config ADVISE_SYSCALLS 1339 bool "Enable madvise/fadvise syscalls" if EXPERT 1340 default y 1341 help 1342 This option enables the madvise and fadvise syscalls, used by 1343 applications to advise the kernel about their future memory or file 1344 usage, improving performance. If building an embedded system where no 1345 applications use these syscalls, you can disable this option to save 1346 space. 1347 1348config MEMBARRIER 1349 bool "Enable membarrier() system call" if EXPERT 1350 default y 1351 help 1352 Enable the membarrier() system call that allows issuing memory 1353 barriers across all running threads, which can be used to distribute 1354 the cost of user-space memory barriers asymmetrically by transforming 1355 pairs of memory barriers into pairs consisting of membarrier() and a 1356 compiler barrier. 1357 1358 If unsure, say Y. 1359 1360config CHECKPOINT_RESTORE 1361 bool "Checkpoint/restore support" if EXPERT 1362 select PROC_CHILDREN 1363 default n 1364 help 1365 Enables additional kernel features in a sake of checkpoint/restore. 1366 In particular it adds auxiliary prctl codes to setup process text, 1367 data and heap segment sizes, and a few additional /proc filesystem 1368 entries. 1369 1370 If unsure, say N here. 1371 1372config KALLSYMS 1373 bool "Load all symbols for debugging/ksymoops" if EXPERT 1374 default y 1375 help 1376 Say Y here to let the kernel print out symbolic crash information and 1377 symbolic stack backtraces. This increases the size of the kernel 1378 somewhat, as all symbols have to be loaded into the kernel image. 1379 1380config KALLSYMS_ALL 1381 bool "Include all symbols in kallsyms" 1382 depends on DEBUG_KERNEL && KALLSYMS 1383 help 1384 Normally kallsyms only contains the symbols of functions for nicer 1385 OOPS messages and backtraces (i.e., symbols from the text and inittext 1386 sections). This is sufficient for most cases. And only in very rare 1387 cases (e.g., when a debugger is used) all symbols are required (e.g., 1388 names of variables from the data sections, etc). 1389 1390 This option makes sure that all symbols are loaded into the kernel 1391 image (i.e., symbols from all sections) in cost of increased kernel 1392 size (depending on the kernel configuration, it may be 300KiB or 1393 something like this). 1394 1395 Say N unless you really need all symbols. 1396 1397config KALLSYMS_ABSOLUTE_PERCPU 1398 bool 1399 depends on KALLSYMS 1400 default X86_64 && SMP 1401 1402config KALLSYMS_BASE_RELATIVE 1403 bool 1404 depends on KALLSYMS 1405 default !IA64 1406 help 1407 Instead of emitting them as absolute values in the native word size, 1408 emit the symbol references in the kallsyms table as 32-bit entries, 1409 each containing a relative value in the range [base, base + U32_MAX] 1410 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either 1411 an absolute value in the range [0, S32_MAX] or a relative value in the 1412 range [base, base + S32_MAX], where base is the lowest relative symbol 1413 address encountered in the image. 1414 1415 On 64-bit builds, this reduces the size of the address table by 50%, 1416 but more importantly, it results in entries whose values are build 1417 time constants, and no relocation pass is required at runtime to fix 1418 up the entries based on the runtime load address of the kernel. 1419 1420# end of the "standard kernel features (expert users)" menu 1421 1422# syscall, maps, verifier 1423config BPF_SYSCALL 1424 bool "Enable bpf() system call" 1425 select ANON_INODES 1426 select BPF 1427 select IRQ_WORK 1428 default n 1429 help 1430 Enable the bpf() system call that allows to manipulate eBPF 1431 programs and maps via file descriptors. 1432 1433config BPF_JIT_ALWAYS_ON 1434 bool "Permanently enable BPF JIT and remove BPF interpreter" 1435 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT 1436 help 1437 Enables BPF JIT and removes BPF interpreter to avoid 1438 speculative execution of BPF instructions by the interpreter 1439 1440config USERFAULTFD 1441 bool "Enable userfaultfd() system call" 1442 select ANON_INODES 1443 depends on MMU 1444 help 1445 Enable the userfaultfd() system call that allows to intercept and 1446 handle page faults in userland. 1447 1448config ARCH_HAS_MEMBARRIER_CALLBACKS 1449 bool 1450 1451config ARCH_HAS_MEMBARRIER_SYNC_CORE 1452 bool 1453 1454config RSEQ 1455 bool "Enable rseq() system call" if EXPERT 1456 default y 1457 depends on HAVE_RSEQ 1458 select MEMBARRIER 1459 help 1460 Enable the restartable sequences system call. It provides a 1461 user-space cache for the current CPU number value, which 1462 speeds up getting the current CPU number from user-space, 1463 as well as an ABI to speed up user-space operations on 1464 per-CPU data. 1465 1466 If unsure, say Y. 1467 1468config DEBUG_RSEQ 1469 default n 1470 bool "Enabled debugging of rseq() system call" if EXPERT 1471 depends on RSEQ && DEBUG_KERNEL 1472 help 1473 Enable extra debugging checks for the rseq system call. 1474 1475 If unsure, say N. 1476 1477config EMBEDDED 1478 bool "Embedded system" 1479 option allnoconfig_y 1480 select EXPERT 1481 help 1482 This option should be enabled if compiling the kernel for 1483 an embedded system so certain expert options are available 1484 for configuration. 1485 1486config HAVE_PERF_EVENTS 1487 bool 1488 help 1489 See tools/perf/design.txt for details. 1490 1491config PERF_USE_VMALLOC 1492 bool 1493 help 1494 See tools/perf/design.txt for details 1495 1496config PC104 1497 bool "PC/104 support" if EXPERT 1498 help 1499 Expose PC/104 form factor device drivers and options available for 1500 selection and configuration. Enable this option if your target 1501 machine has a PC/104 bus. 1502 1503menu "Kernel Performance Events And Counters" 1504 1505config PERF_EVENTS 1506 bool "Kernel performance events and counters" 1507 default y if PROFILING 1508 depends on HAVE_PERF_EVENTS 1509 select ANON_INODES 1510 select IRQ_WORK 1511 select SRCU 1512 help 1513 Enable kernel support for various performance events provided 1514 by software and hardware. 1515 1516 Software events are supported either built-in or via the 1517 use of generic tracepoints. 1518 1519 Most modern CPUs support performance events via performance 1520 counter registers. These registers count the number of certain 1521 types of hw events: such as instructions executed, cachemisses 1522 suffered, or branches mis-predicted - without slowing down the 1523 kernel or applications. These registers can also trigger interrupts 1524 when a threshold number of events have passed - and can thus be 1525 used to profile the code that runs on that CPU. 1526 1527 The Linux Performance Event subsystem provides an abstraction of 1528 these software and hardware event capabilities, available via a 1529 system call and used by the "perf" utility in tools/perf/. It 1530 provides per task and per CPU counters, and it provides event 1531 capabilities on top of those. 1532 1533 Say Y if unsure. 1534 1535config DEBUG_PERF_USE_VMALLOC 1536 default n 1537 bool "Debug: use vmalloc to back perf mmap() buffers" 1538 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC 1539 select PERF_USE_VMALLOC 1540 help 1541 Use vmalloc memory to back perf mmap() buffers. 1542 1543 Mostly useful for debugging the vmalloc code on platforms 1544 that don't require it. 1545 1546 Say N if unsure. 1547 1548endmenu 1549 1550config VM_EVENT_COUNTERS 1551 default y 1552 bool "Enable VM event counters for /proc/vmstat" if EXPERT 1553 help 1554 VM event counters are needed for event counts to be shown. 1555 This option allows the disabling of the VM event counters 1556 on EXPERT systems. /proc/vmstat will only show page counts 1557 if VM event counters are disabled. 1558 1559config SLUB_DEBUG 1560 default y 1561 bool "Enable SLUB debugging support" if EXPERT 1562 depends on SLUB && SYSFS 1563 help 1564 SLUB has extensive debug support features. Disabling these can 1565 result in significant savings in code size. This also disables 1566 SLUB sysfs support. /sys/slab will not exist and there will be 1567 no support for cache validation etc. 1568 1569config SLUB_MEMCG_SYSFS_ON 1570 default n 1571 bool "Enable memcg SLUB sysfs support by default" if EXPERT 1572 depends on SLUB && SYSFS && MEMCG 1573 help 1574 SLUB creates a directory under /sys/kernel/slab for each 1575 allocation cache to host info and debug files. If memory 1576 cgroup is enabled, each cache can have per memory cgroup 1577 caches. SLUB can create the same sysfs directories for these 1578 caches under /sys/kernel/slab/CACHE/cgroup but it can lead 1579 to a very high number of debug files being created. This is 1580 controlled by slub_memcg_sysfs boot parameter and this 1581 config option determines the parameter's default value. 1582 1583config COMPAT_BRK 1584 bool "Disable heap randomization" 1585 default y 1586 help 1587 Randomizing heap placement makes heap exploits harder, but it 1588 also breaks ancient binaries (including anything libc5 based). 1589 This option changes the bootup default to heap randomization 1590 disabled, and can be overridden at runtime by setting 1591 /proc/sys/kernel/randomize_va_space to 2. 1592 1593 On non-ancient distros (post-2000 ones) N is usually a safe choice. 1594 1595choice 1596 prompt "Choose SLAB allocator" 1597 default SLUB 1598 help 1599 This option allows to select a slab allocator. 1600 1601config SLAB 1602 bool "SLAB" 1603 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1604 help 1605 The regular slab allocator that is established and known to work 1606 well in all environments. It organizes cache hot objects in 1607 per cpu and per node queues. 1608 1609config SLUB 1610 bool "SLUB (Unqueued Allocator)" 1611 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1612 help 1613 SLUB is a slab allocator that minimizes cache line usage 1614 instead of managing queues of cached objects (SLAB approach). 1615 Per cpu caching is realized using slabs of objects instead 1616 of queues of objects. SLUB can use memory efficiently 1617 and has enhanced diagnostics. SLUB is the default choice for 1618 a slab allocator. 1619 1620config SLOB 1621 depends on EXPERT 1622 bool "SLOB (Simple Allocator)" 1623 help 1624 SLOB replaces the stock allocator with a drastically simpler 1625 allocator. SLOB is generally more space efficient but 1626 does not perform as well on large systems. 1627 1628endchoice 1629 1630config SLAB_MERGE_DEFAULT 1631 bool "Allow slab caches to be merged" 1632 default y 1633 help 1634 For reduced kernel memory fragmentation, slab caches can be 1635 merged when they share the same size and other characteristics. 1636 This carries a risk of kernel heap overflows being able to 1637 overwrite objects from merged caches (and more easily control 1638 cache layout), which makes such heap attacks easier to exploit 1639 by attackers. By keeping caches unmerged, these kinds of exploits 1640 can usually only damage objects in the same cache. To disable 1641 merging at runtime, "slab_nomerge" can be passed on the kernel 1642 command line. 1643 1644config SLAB_FREELIST_RANDOM 1645 default n 1646 depends on SLAB || SLUB 1647 bool "SLAB freelist randomization" 1648 help 1649 Randomizes the freelist order used on creating new pages. This 1650 security feature reduces the predictability of the kernel slab 1651 allocator against heap overflows. 1652 1653config SLAB_FREELIST_HARDENED 1654 bool "Harden slab freelist metadata" 1655 depends on SLUB 1656 help 1657 Many kernel heap attacks try to target slab cache metadata and 1658 other infrastructure. This options makes minor performance 1659 sacrifies to harden the kernel slab allocator against common 1660 freelist exploit methods. 1661 1662config SLUB_CPU_PARTIAL 1663 default y 1664 depends on SLUB && SMP 1665 bool "SLUB per cpu partial cache" 1666 help 1667 Per cpu partial caches accellerate objects allocation and freeing 1668 that is local to a processor at the price of more indeterminism 1669 in the latency of the free. On overflow these caches will be cleared 1670 which requires the taking of locks that may cause latency spikes. 1671 Typically one would choose no for a realtime system. 1672 1673config MMAP_ALLOW_UNINITIALIZED 1674 bool "Allow mmapped anonymous memory to be uninitialized" 1675 depends on EXPERT && !MMU 1676 default n 1677 help 1678 Normally, and according to the Linux spec, anonymous memory obtained 1679 from mmap() has it's contents cleared before it is passed to 1680 userspace. Enabling this config option allows you to request that 1681 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1682 providing a huge performance boost. If this option is not enabled, 1683 then the flag will be ignored. 1684 1685 This is taken advantage of by uClibc's malloc(), and also by 1686 ELF-FDPIC binfmt's brk and stack allocator. 1687 1688 Because of the obvious security issues, this option should only be 1689 enabled on embedded devices where you control what is run in 1690 userspace. Since that isn't generally a problem on no-MMU systems, 1691 it is normally safe to say Y here. 1692 1693 See Documentation/nommu-mmap.txt for more information. 1694 1695config SYSTEM_DATA_VERIFICATION 1696 def_bool n 1697 select SYSTEM_TRUSTED_KEYRING 1698 select KEYS 1699 select CRYPTO 1700 select CRYPTO_RSA 1701 select ASYMMETRIC_KEY_TYPE 1702 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1703 select ASN1 1704 select OID_REGISTRY 1705 select X509_CERTIFICATE_PARSER 1706 select PKCS7_MESSAGE_PARSER 1707 help 1708 Provide PKCS#7 message verification using the contents of the system 1709 trusted keyring to provide public keys. This then can be used for 1710 module verification, kexec image verification and firmware blob 1711 verification. 1712 1713config PROFILING 1714 bool "Profiling support" 1715 help 1716 Say Y here to enable the extended profiling support mechanisms used 1717 by profilers such as OProfile. 1718 1719# 1720# Place an empty function call at each tracepoint site. Can be 1721# dynamically changed for a probe function. 1722# 1723config TRACEPOINTS 1724 bool 1725 1726source "arch/Kconfig" 1727 1728endmenu # General setup 1729 1730config RT_MUTEXES 1731 bool 1732 1733config BASE_SMALL 1734 int 1735 default 0 if BASE_FULL 1736 default 1 if !BASE_FULL 1737 1738menuconfig MODULES 1739 bool "Enable loadable module support" 1740 option modules 1741 help 1742 Kernel modules are small pieces of compiled code which can 1743 be inserted in the running kernel, rather than being 1744 permanently built into the kernel. You use the "modprobe" 1745 tool to add (and sometimes remove) them. If you say Y here, 1746 many parts of the kernel can be built as modules (by 1747 answering M instead of Y where indicated): this is most 1748 useful for infrequently used options which are not required 1749 for booting. For more information, see the man pages for 1750 modprobe, lsmod, modinfo, insmod and rmmod. 1751 1752 If you say Y here, you will need to run "make 1753 modules_install" to put the modules under /lib/modules/ 1754 where modprobe can find them (you may need to be root to do 1755 this). 1756 1757 If unsure, say Y. 1758 1759if MODULES 1760 1761config MODULE_FORCE_LOAD 1762 bool "Forced module loading" 1763 default n 1764 help 1765 Allow loading of modules without version information (ie. modprobe 1766 --force). Forced module loading sets the 'F' (forced) taint flag and 1767 is usually a really bad idea. 1768 1769config MODULE_UNLOAD 1770 bool "Module unloading" 1771 help 1772 Without this option you will not be able to unload any 1773 modules (note that some modules may not be unloadable 1774 anyway), which makes your kernel smaller, faster 1775 and simpler. If unsure, say Y. 1776 1777config MODULE_FORCE_UNLOAD 1778 bool "Forced module unloading" 1779 depends on MODULE_UNLOAD 1780 help 1781 This option allows you to force a module to unload, even if the 1782 kernel believes it is unsafe: the kernel will remove the module 1783 without waiting for anyone to stop using it (using the -f option to 1784 rmmod). This is mainly for kernel developers and desperate users. 1785 If unsure, say N. 1786 1787config MODVERSIONS 1788 bool "Module versioning support" 1789 help 1790 Usually, you have to use modules compiled with your kernel. 1791 Saying Y here makes it sometimes possible to use modules 1792 compiled for different kernels, by adding enough information 1793 to the modules to (hopefully) spot any changes which would 1794 make them incompatible with the kernel you are running. If 1795 unsure, say N. 1796 1797config MODULE_REL_CRCS 1798 bool 1799 depends on MODVERSIONS 1800 1801config MODULE_SRCVERSION_ALL 1802 bool "Source checksum for all modules" 1803 help 1804 Modules which contain a MODULE_VERSION get an extra "srcversion" 1805 field inserted into their modinfo section, which contains a 1806 sum of the source files which made it. This helps maintainers 1807 see exactly which source was used to build a module (since 1808 others sometimes change the module source without updating 1809 the version). With this option, such a "srcversion" field 1810 will be created for all modules. If unsure, say N. 1811 1812config MODULE_SIG 1813 bool "Module signature verification" 1814 depends on MODULES 1815 select SYSTEM_DATA_VERIFICATION 1816 help 1817 Check modules for valid signatures upon load: the signature 1818 is simply appended to the module. For more information see 1819 <file:Documentation/admin-guide/module-signing.rst>. 1820 1821 Note that this option adds the OpenSSL development packages as a 1822 kernel build dependency so that the signing tool can use its crypto 1823 library. 1824 1825 !!!WARNING!!! If you enable this option, you MUST make sure that the 1826 module DOES NOT get stripped after being signed. This includes the 1827 debuginfo strip done by some packagers (such as rpmbuild) and 1828 inclusion into an initramfs that wants the module size reduced. 1829 1830config MODULE_SIG_FORCE 1831 bool "Require modules to be validly signed" 1832 depends on MODULE_SIG 1833 help 1834 Reject unsigned modules or signed modules for which we don't have a 1835 key. Without this, such modules will simply taint the kernel. 1836 1837config MODULE_SIG_ALL 1838 bool "Automatically sign all modules" 1839 default y 1840 depends on MODULE_SIG 1841 help 1842 Sign all modules during make modules_install. Without this option, 1843 modules must be signed manually, using the scripts/sign-file tool. 1844 1845comment "Do not forget to sign required modules with scripts/sign-file" 1846 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 1847 1848choice 1849 prompt "Which hash algorithm should modules be signed with?" 1850 depends on MODULE_SIG 1851 help 1852 This determines which sort of hashing algorithm will be used during 1853 signature generation. This algorithm _must_ be built into the kernel 1854 directly so that signature verification can take place. It is not 1855 possible to load a signed module containing the algorithm to check 1856 the signature on that module. 1857 1858config MODULE_SIG_SHA1 1859 bool "Sign modules with SHA-1" 1860 select CRYPTO_SHA1 1861 1862config MODULE_SIG_SHA224 1863 bool "Sign modules with SHA-224" 1864 select CRYPTO_SHA256 1865 1866config MODULE_SIG_SHA256 1867 bool "Sign modules with SHA-256" 1868 select CRYPTO_SHA256 1869 1870config MODULE_SIG_SHA384 1871 bool "Sign modules with SHA-384" 1872 select CRYPTO_SHA512 1873 1874config MODULE_SIG_SHA512 1875 bool "Sign modules with SHA-512" 1876 select CRYPTO_SHA512 1877 1878endchoice 1879 1880config MODULE_SIG_HASH 1881 string 1882 depends on MODULE_SIG 1883 default "sha1" if MODULE_SIG_SHA1 1884 default "sha224" if MODULE_SIG_SHA224 1885 default "sha256" if MODULE_SIG_SHA256 1886 default "sha384" if MODULE_SIG_SHA384 1887 default "sha512" if MODULE_SIG_SHA512 1888 1889config MODULE_COMPRESS 1890 bool "Compress modules on installation" 1891 depends on MODULES 1892 help 1893 1894 Compresses kernel modules when 'make modules_install' is run; gzip or 1895 xz depending on "Compression algorithm" below. 1896 1897 module-init-tools MAY support gzip, and kmod MAY support gzip and xz. 1898 1899 Out-of-tree kernel modules installed using Kbuild will also be 1900 compressed upon installation. 1901 1902 Note: for modules inside an initrd or initramfs, it's more efficient 1903 to compress the whole initrd or initramfs instead. 1904 1905 Note: This is fully compatible with signed modules. 1906 1907 If in doubt, say N. 1908 1909choice 1910 prompt "Compression algorithm" 1911 depends on MODULE_COMPRESS 1912 default MODULE_COMPRESS_GZIP 1913 help 1914 This determines which sort of compression will be used during 1915 'make modules_install'. 1916 1917 GZIP (default) and XZ are supported. 1918 1919config MODULE_COMPRESS_GZIP 1920 bool "GZIP" 1921 1922config MODULE_COMPRESS_XZ 1923 bool "XZ" 1924 1925endchoice 1926 1927config TRIM_UNUSED_KSYMS 1928 bool "Trim unused exported kernel symbols" 1929 depends on MODULES && !UNUSED_SYMBOLS 1930 help 1931 The kernel and some modules make many symbols available for 1932 other modules to use via EXPORT_SYMBOL() and variants. Depending 1933 on the set of modules being selected in your kernel configuration, 1934 many of those exported symbols might never be used. 1935 1936 This option allows for unused exported symbols to be dropped from 1937 the build. In turn, this provides the compiler more opportunities 1938 (especially when using LTO) for optimizing the code and reducing 1939 binary size. This might have some security advantages as well. 1940 1941 If unsure, or if you need to build out-of-tree modules, say N. 1942 1943endif # MODULES 1944 1945config MODULES_TREE_LOOKUP 1946 def_bool y 1947 depends on PERF_EVENTS || TRACING 1948 1949config INIT_ALL_POSSIBLE 1950 bool 1951 help 1952 Back when each arch used to define their own cpu_online_mask and 1953 cpu_possible_mask, some of them chose to initialize cpu_possible_mask 1954 with all 1s, and others with all 0s. When they were centralised, 1955 it was better to provide this option than to break all the archs 1956 and have several arch maintainers pursuing me down dark alleys. 1957 1958source "block/Kconfig" 1959 1960config PREEMPT_NOTIFIERS 1961 bool 1962 1963config PADATA 1964 depends on SMP 1965 bool 1966 1967config ASN1 1968 tristate 1969 help 1970 Build a simple ASN.1 grammar compiler that produces a bytecode output 1971 that can be interpreted by the ASN.1 stream decoder and used to 1972 inform it as to what tags are to be expected in a stream and what 1973 functions to call on what tags. 1974 1975source "kernel/Kconfig.locks" 1976 1977config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE 1978 bool 1979 1980# It may be useful for an architecture to override the definitions of the 1981# SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h> 1982# and the COMPAT_ variants in <linux/compat.h>, in particular to use a 1983# different calling convention for syscalls. They can also override the 1984# macros for not-implemented syscalls in kernel/sys_ni.c and 1985# kernel/time/posix-stubs.c. All these overrides need to be available in 1986# <asm/syscall_wrapper.h>. 1987config ARCH_HAS_SYSCALL_WRAPPER 1988 def_bool n 1989