1# SPDX-License-Identifier: GPL-2.0-only 2config DEFCONFIG_LIST 3 string 4 depends on !UML 5 option defconfig_list 6 default "/lib/modules/$(shell,uname -r)/.config" 7 default "/etc/kernel-config" 8 default "/boot/config-$(shell,uname -r)" 9 default ARCH_DEFCONFIG 10 default "arch/$(ARCH)/defconfig" 11 12config CC_IS_GCC 13 def_bool $(success,$(CC) --version | head -n 1 | grep -q gcc) 14 15config GCC_VERSION 16 int 17 default $(shell,$(srctree)/scripts/gcc-version.sh $(CC)) if CC_IS_GCC 18 default 0 19 20config CC_IS_CLANG 21 def_bool $(success,$(CC) --version | head -n 1 | grep -q clang) 22 23config CLANG_VERSION 24 int 25 default $(shell,$(srctree)/scripts/clang-version.sh $(CC)) 26 27config CC_CAN_LINK 28 def_bool $(success,$(srctree)/scripts/cc-can-link.sh $(CC)) 29 30config CC_HAS_ASM_GOTO 31 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC)) 32 33config CC_HAS_WARN_MAYBE_UNINITIALIZED 34 def_bool $(cc-option,-Wmaybe-uninitialized) 35 help 36 GCC >= 4.7 supports this option. 37 38config CC_DISABLE_WARN_MAYBE_UNINITIALIZED 39 bool 40 depends on CC_HAS_WARN_MAYBE_UNINITIALIZED 41 default CC_IS_GCC && GCC_VERSION < 40900 # unreliable for GCC < 4.9 42 help 43 GCC's -Wmaybe-uninitialized is not reliable by definition. 44 Lots of false positive warnings are produced in some cases. 45 46 If this option is enabled, -Wno-maybe-uninitialzed is passed 47 to the compiler to suppress maybe-uninitialized warnings. 48 49config CONSTRUCTORS 50 bool 51 52config IRQ_WORK 53 bool 54 55config BUILDTIME_EXTABLE_SORT 56 bool 57 58config THREAD_INFO_IN_TASK 59 bool 60 help 61 Select this to move thread_info off the stack into task_struct. To 62 make this work, an arch will need to remove all thread_info fields 63 except flags and fix any runtime bugs. 64 65 One subtle change that will be needed is to use try_get_task_stack() 66 and put_task_stack() in save_thread_stack_tsk() and get_wchan(). 67 68menu "General setup" 69 70config BROKEN 71 bool 72 73config BROKEN_ON_SMP 74 bool 75 depends on BROKEN || !SMP 76 default y 77 78config INIT_ENV_ARG_LIMIT 79 int 80 default 32 if !UML 81 default 128 if UML 82 help 83 Maximum of each of the number of arguments and environment 84 variables passed to init from the kernel command line. 85 86config COMPILE_TEST 87 bool "Compile also drivers which will not load" 88 depends on !UML 89 default n 90 help 91 Some drivers can be compiled on a different platform than they are 92 intended to be run on. Despite they cannot be loaded there (or even 93 when they load they cannot be used due to missing HW support), 94 developers still, opposing to distributors, might want to build such 95 drivers to compile-test them. 96 97 If you are a developer and want to build everything available, say Y 98 here. If you are a user/distributor, say N here to exclude useless 99 drivers to be distributed. 100 101config HEADER_TEST 102 bool "Compile test headers that should be standalone compilable" 103 help 104 Compile test headers listed in header-test-y target to ensure they are 105 self-contained, i.e. compilable as standalone units. 106 107 If you are a developer or tester and want to ensure the requested 108 headers are self-contained, say Y here. Otherwise, choose N. 109 110config KERNEL_HEADER_TEST 111 bool "Compile test kernel headers" 112 depends on HEADER_TEST 113 help 114 Headers in include/ are used to build external moduls. 115 Compile test them to ensure they are self-contained, i.e. 116 compilable as standalone units. 117 118 If you are a developer or tester and want to ensure the headers 119 in include/ are self-contained, say Y here. Otherwise, choose N. 120 121config UAPI_HEADER_TEST 122 bool "Compile test UAPI headers" 123 depends on HEADER_TEST && HEADERS_INSTALL && CC_CAN_LINK 124 help 125 Compile test headers exported to user-space to ensure they are 126 self-contained, i.e. compilable as standalone units. 127 128 If you are a developer or tester and want to ensure the exported 129 headers are self-contained, say Y here. Otherwise, choose N. 130 131config LOCALVERSION 132 string "Local version - append to kernel release" 133 help 134 Append an extra string to the end of your kernel version. 135 This will show up when you type uname, for example. 136 The string you set here will be appended after the contents of 137 any files with a filename matching localversion* in your 138 object and source tree, in that order. Your total string can 139 be a maximum of 64 characters. 140 141config LOCALVERSION_AUTO 142 bool "Automatically append version information to the version string" 143 default y 144 depends on !COMPILE_TEST 145 help 146 This will try to automatically determine if the current tree is a 147 release tree by looking for git tags that belong to the current 148 top of tree revision. 149 150 A string of the format -gxxxxxxxx will be added to the localversion 151 if a git-based tree is found. The string generated by this will be 152 appended after any matching localversion* files, and after the value 153 set in CONFIG_LOCALVERSION. 154 155 (The actual string used here is the first eight characters produced 156 by running the command: 157 158 $ git rev-parse --verify HEAD 159 160 which is done within the script "scripts/setlocalversion".) 161 162config BUILD_SALT 163 string "Build ID Salt" 164 default "" 165 help 166 The build ID is used to link binaries and their debug info. Setting 167 this option will use the value in the calculation of the build id. 168 This is mostly useful for distributions which want to ensure the 169 build is unique between builds. It's safe to leave the default. 170 171config HAVE_KERNEL_GZIP 172 bool 173 174config HAVE_KERNEL_BZIP2 175 bool 176 177config HAVE_KERNEL_LZMA 178 bool 179 180config HAVE_KERNEL_XZ 181 bool 182 183config HAVE_KERNEL_LZO 184 bool 185 186config HAVE_KERNEL_LZ4 187 bool 188 189config HAVE_KERNEL_UNCOMPRESSED 190 bool 191 192choice 193 prompt "Kernel compression mode" 194 default KERNEL_GZIP 195 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED 196 help 197 The linux kernel is a kind of self-extracting executable. 198 Several compression algorithms are available, which differ 199 in efficiency, compression and decompression speed. 200 Compression speed is only relevant when building a kernel. 201 Decompression speed is relevant at each boot. 202 203 If you have any problems with bzip2 or lzma compressed 204 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 205 version of this functionality (bzip2 only), for 2.4, was 206 supplied by Christian Ludwig) 207 208 High compression options are mostly useful for users, who 209 are low on disk space (embedded systems), but for whom ram 210 size matters less. 211 212 If in doubt, select 'gzip' 213 214config KERNEL_GZIP 215 bool "Gzip" 216 depends on HAVE_KERNEL_GZIP 217 help 218 The old and tried gzip compression. It provides a good balance 219 between compression ratio and decompression speed. 220 221config KERNEL_BZIP2 222 bool "Bzip2" 223 depends on HAVE_KERNEL_BZIP2 224 help 225 Its compression ratio and speed is intermediate. 226 Decompression speed is slowest among the choices. The kernel 227 size is about 10% smaller with bzip2, in comparison to gzip. 228 Bzip2 uses a large amount of memory. For modern kernels you 229 will need at least 8MB RAM or more for booting. 230 231config KERNEL_LZMA 232 bool "LZMA" 233 depends on HAVE_KERNEL_LZMA 234 help 235 This compression algorithm's ratio is best. Decompression speed 236 is between gzip and bzip2. Compression is slowest. 237 The kernel size is about 33% smaller with LZMA in comparison to gzip. 238 239config KERNEL_XZ 240 bool "XZ" 241 depends on HAVE_KERNEL_XZ 242 help 243 XZ uses the LZMA2 algorithm and instruction set specific 244 BCJ filters which can improve compression ratio of executable 245 code. The size of the kernel is about 30% smaller with XZ in 246 comparison to gzip. On architectures for which there is a BCJ 247 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 248 will create a few percent smaller kernel than plain LZMA. 249 250 The speed is about the same as with LZMA: The decompression 251 speed of XZ is better than that of bzip2 but worse than gzip 252 and LZO. Compression is slow. 253 254config KERNEL_LZO 255 bool "LZO" 256 depends on HAVE_KERNEL_LZO 257 help 258 Its compression ratio is the poorest among the choices. The kernel 259 size is about 10% bigger than gzip; however its speed 260 (both compression and decompression) is the fastest. 261 262config KERNEL_LZ4 263 bool "LZ4" 264 depends on HAVE_KERNEL_LZ4 265 help 266 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. 267 A preliminary version of LZ4 de/compression tool is available at 268 <https://code.google.com/p/lz4/>. 269 270 Its compression ratio is worse than LZO. The size of the kernel 271 is about 8% bigger than LZO. But the decompression speed is 272 faster than LZO. 273 274config KERNEL_UNCOMPRESSED 275 bool "None" 276 depends on HAVE_KERNEL_UNCOMPRESSED 277 help 278 Produce uncompressed kernel image. This option is usually not what 279 you want. It is useful for debugging the kernel in slow simulation 280 environments, where decompressing and moving the kernel is awfully 281 slow. This option allows early boot code to skip the decompressor 282 and jump right at uncompressed kernel image. 283 284endchoice 285 286config DEFAULT_HOSTNAME 287 string "Default hostname" 288 default "(none)" 289 help 290 This option determines the default system hostname before userspace 291 calls sethostname(2). The kernel traditionally uses "(none)" here, 292 but you may wish to use a different default here to make a minimal 293 system more usable with less configuration. 294 295# 296# For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can 297# add proper SWAP support to them, in which case this can be remove. 298# 299config ARCH_NO_SWAP 300 bool 301 302config SWAP 303 bool "Support for paging of anonymous memory (swap)" 304 depends on MMU && BLOCK && !ARCH_NO_SWAP 305 default y 306 help 307 This option allows you to choose whether you want to have support 308 for so called swap devices or swap files in your kernel that are 309 used to provide more virtual memory than the actual RAM present 310 in your computer. If unsure say Y. 311 312config SYSVIPC 313 bool "System V IPC" 314 ---help--- 315 Inter Process Communication is a suite of library functions and 316 system calls which let processes (running programs) synchronize and 317 exchange information. It is generally considered to be a good thing, 318 and some programs won't run unless you say Y here. In particular, if 319 you want to run the DOS emulator dosemu under Linux (read the 320 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 321 you'll need to say Y here. 322 323 You can find documentation about IPC with "info ipc" and also in 324 section 6.4 of the Linux Programmer's Guide, available from 325 <http://www.tldp.org/guides.html>. 326 327config SYSVIPC_SYSCTL 328 bool 329 depends on SYSVIPC 330 depends on SYSCTL 331 default y 332 333config POSIX_MQUEUE 334 bool "POSIX Message Queues" 335 depends on NET 336 ---help--- 337 POSIX variant of message queues is a part of IPC. In POSIX message 338 queues every message has a priority which decides about succession 339 of receiving it by a process. If you want to compile and run 340 programs written e.g. for Solaris with use of its POSIX message 341 queues (functions mq_*) say Y here. 342 343 POSIX message queues are visible as a filesystem called 'mqueue' 344 and can be mounted somewhere if you want to do filesystem 345 operations on message queues. 346 347 If unsure, say Y. 348 349config POSIX_MQUEUE_SYSCTL 350 bool 351 depends on POSIX_MQUEUE 352 depends on SYSCTL 353 default y 354 355config CROSS_MEMORY_ATTACH 356 bool "Enable process_vm_readv/writev syscalls" 357 depends on MMU 358 default y 359 help 360 Enabling this option adds the system calls process_vm_readv and 361 process_vm_writev which allow a process with the correct privileges 362 to directly read from or write to another process' address space. 363 See the man page for more details. 364 365config USELIB 366 bool "uselib syscall" 367 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION 368 help 369 This option enables the uselib syscall, a system call used in the 370 dynamic linker from libc5 and earlier. glibc does not use this 371 system call. If you intend to run programs built on libc5 or 372 earlier, you may need to enable this syscall. Current systems 373 running glibc can safely disable this. 374 375config AUDIT 376 bool "Auditing support" 377 depends on NET 378 help 379 Enable auditing infrastructure that can be used with another 380 kernel subsystem, such as SELinux (which requires this for 381 logging of avc messages output). System call auditing is included 382 on architectures which support it. 383 384config HAVE_ARCH_AUDITSYSCALL 385 bool 386 387config AUDITSYSCALL 388 def_bool y 389 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL 390 select FSNOTIFY 391 392source "kernel/irq/Kconfig" 393source "kernel/time/Kconfig" 394source "kernel/Kconfig.preempt" 395 396menu "CPU/Task time and stats accounting" 397 398config VIRT_CPU_ACCOUNTING 399 bool 400 401choice 402 prompt "Cputime accounting" 403 default TICK_CPU_ACCOUNTING if !PPC64 404 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 405 406# Kind of a stub config for the pure tick based cputime accounting 407config TICK_CPU_ACCOUNTING 408 bool "Simple tick based cputime accounting" 409 depends on !S390 && !NO_HZ_FULL 410 help 411 This is the basic tick based cputime accounting that maintains 412 statistics about user, system and idle time spent on per jiffies 413 granularity. 414 415 If unsure, say Y. 416 417config VIRT_CPU_ACCOUNTING_NATIVE 418 bool "Deterministic task and CPU time accounting" 419 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 420 select VIRT_CPU_ACCOUNTING 421 help 422 Select this option to enable more accurate task and CPU time 423 accounting. This is done by reading a CPU counter on each 424 kernel entry and exit and on transitions within the kernel 425 between system, softirq and hardirq state, so there is a 426 small performance impact. In the case of s390 or IBM POWER > 5, 427 this also enables accounting of stolen time on logically-partitioned 428 systems. 429 430config VIRT_CPU_ACCOUNTING_GEN 431 bool "Full dynticks CPU time accounting" 432 depends on HAVE_CONTEXT_TRACKING 433 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN 434 depends on GENERIC_CLOCKEVENTS 435 select VIRT_CPU_ACCOUNTING 436 select CONTEXT_TRACKING 437 help 438 Select this option to enable task and CPU time accounting on full 439 dynticks systems. This accounting is implemented by watching every 440 kernel-user boundaries using the context tracking subsystem. 441 The accounting is thus performed at the expense of some significant 442 overhead. 443 444 For now this is only useful if you are working on the full 445 dynticks subsystem development. 446 447 If unsure, say N. 448 449endchoice 450 451config IRQ_TIME_ACCOUNTING 452 bool "Fine granularity task level IRQ time accounting" 453 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE 454 help 455 Select this option to enable fine granularity task irq time 456 accounting. This is done by reading a timestamp on each 457 transitions between softirq and hardirq state, so there can be a 458 small performance impact. 459 460 If in doubt, say N here. 461 462config HAVE_SCHED_AVG_IRQ 463 def_bool y 464 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING 465 depends on SMP 466 467config BSD_PROCESS_ACCT 468 bool "BSD Process Accounting" 469 depends on MULTIUSER 470 help 471 If you say Y here, a user level program will be able to instruct the 472 kernel (via a special system call) to write process accounting 473 information to a file: whenever a process exits, information about 474 that process will be appended to the file by the kernel. The 475 information includes things such as creation time, owning user, 476 command name, memory usage, controlling terminal etc. (the complete 477 list is in the struct acct in <file:include/linux/acct.h>). It is 478 up to the user level program to do useful things with this 479 information. This is generally a good idea, so say Y. 480 481config BSD_PROCESS_ACCT_V3 482 bool "BSD Process Accounting version 3 file format" 483 depends on BSD_PROCESS_ACCT 484 default n 485 help 486 If you say Y here, the process accounting information is written 487 in a new file format that also logs the process IDs of each 488 process and its parent. Note that this file format is incompatible 489 with previous v0/v1/v2 file formats, so you will need updated tools 490 for processing it. A preliminary version of these tools is available 491 at <http://www.gnu.org/software/acct/>. 492 493config TASKSTATS 494 bool "Export task/process statistics through netlink" 495 depends on NET 496 depends on MULTIUSER 497 default n 498 help 499 Export selected statistics for tasks/processes through the 500 generic netlink interface. Unlike BSD process accounting, the 501 statistics are available during the lifetime of tasks/processes as 502 responses to commands. Like BSD accounting, they are sent to user 503 space on task exit. 504 505 Say N if unsure. 506 507config TASK_DELAY_ACCT 508 bool "Enable per-task delay accounting" 509 depends on TASKSTATS 510 select SCHED_INFO 511 help 512 Collect information on time spent by a task waiting for system 513 resources like cpu, synchronous block I/O completion and swapping 514 in pages. Such statistics can help in setting a task's priorities 515 relative to other tasks for cpu, io, rss limits etc. 516 517 Say N if unsure. 518 519config TASK_XACCT 520 bool "Enable extended accounting over taskstats" 521 depends on TASKSTATS 522 help 523 Collect extended task accounting data and send the data 524 to userland for processing over the taskstats interface. 525 526 Say N if unsure. 527 528config TASK_IO_ACCOUNTING 529 bool "Enable per-task storage I/O accounting" 530 depends on TASK_XACCT 531 help 532 Collect information on the number of bytes of storage I/O which this 533 task has caused. 534 535 Say N if unsure. 536 537config PSI 538 bool "Pressure stall information tracking" 539 help 540 Collect metrics that indicate how overcommitted the CPU, memory, 541 and IO capacity are in the system. 542 543 If you say Y here, the kernel will create /proc/pressure/ with the 544 pressure statistics files cpu, memory, and io. These will indicate 545 the share of walltime in which some or all tasks in the system are 546 delayed due to contention of the respective resource. 547 548 In kernels with cgroup support, cgroups (cgroup2 only) will 549 have cpu.pressure, memory.pressure, and io.pressure files, 550 which aggregate pressure stalls for the grouped tasks only. 551 552 For more details see Documentation/accounting/psi.rst. 553 554 Say N if unsure. 555 556config PSI_DEFAULT_DISABLED 557 bool "Require boot parameter to enable pressure stall information tracking" 558 default n 559 depends on PSI 560 help 561 If set, pressure stall information tracking will be disabled 562 per default but can be enabled through passing psi=1 on the 563 kernel commandline during boot. 564 565 This feature adds some code to the task wakeup and sleep 566 paths of the scheduler. The overhead is too low to affect 567 common scheduling-intense workloads in practice (such as 568 webservers, memcache), but it does show up in artificial 569 scheduler stress tests, such as hackbench. 570 571 If you are paranoid and not sure what the kernel will be 572 used for, say Y. 573 574 Say N if unsure. 575 576endmenu # "CPU/Task time and stats accounting" 577 578config CPU_ISOLATION 579 bool "CPU isolation" 580 depends on SMP || COMPILE_TEST 581 default y 582 help 583 Make sure that CPUs running critical tasks are not disturbed by 584 any source of "noise" such as unbound workqueues, timers, kthreads... 585 Unbound jobs get offloaded to housekeeping CPUs. This is driven by 586 the "isolcpus=" boot parameter. 587 588 Say Y if unsure. 589 590source "kernel/rcu/Kconfig" 591 592config BUILD_BIN2C 593 bool 594 default n 595 596config IKCONFIG 597 tristate "Kernel .config support" 598 ---help--- 599 This option enables the complete Linux kernel ".config" file 600 contents to be saved in the kernel. It provides documentation 601 of which kernel options are used in a running kernel or in an 602 on-disk kernel. This information can be extracted from the kernel 603 image file with the script scripts/extract-ikconfig and used as 604 input to rebuild the current kernel or to build another kernel. 605 It can also be extracted from a running kernel by reading 606 /proc/config.gz if enabled (below). 607 608config IKCONFIG_PROC 609 bool "Enable access to .config through /proc/config.gz" 610 depends on IKCONFIG && PROC_FS 611 ---help--- 612 This option enables access to the kernel configuration file 613 through /proc/config.gz. 614 615config IKHEADERS 616 tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz" 617 depends on SYSFS 618 help 619 This option enables access to the in-kernel headers that are generated during 620 the build process. These can be used to build eBPF tracing programs, 621 or similar programs. If you build the headers as a module, a module called 622 kheaders.ko is built which can be loaded on-demand to get access to headers. 623 624config LOG_BUF_SHIFT 625 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 626 range 12 25 627 default 17 628 depends on PRINTK 629 help 630 Select the minimal kernel log buffer size as a power of 2. 631 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config 632 parameter, see below. Any higher size also might be forced 633 by "log_buf_len" boot parameter. 634 635 Examples: 636 17 => 128 KB 637 16 => 64 KB 638 15 => 32 KB 639 14 => 16 KB 640 13 => 8 KB 641 12 => 4 KB 642 643config LOG_CPU_MAX_BUF_SHIFT 644 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" 645 depends on SMP 646 range 0 21 647 default 12 if !BASE_SMALL 648 default 0 if BASE_SMALL 649 depends on PRINTK 650 help 651 This option allows to increase the default ring buffer size 652 according to the number of CPUs. The value defines the contribution 653 of each CPU as a power of 2. The used space is typically only few 654 lines however it might be much more when problems are reported, 655 e.g. backtraces. 656 657 The increased size means that a new buffer has to be allocated and 658 the original static one is unused. It makes sense only on systems 659 with more CPUs. Therefore this value is used only when the sum of 660 contributions is greater than the half of the default kernel ring 661 buffer as defined by LOG_BUF_SHIFT. The default values are set 662 so that more than 64 CPUs are needed to trigger the allocation. 663 664 Also this option is ignored when "log_buf_len" kernel parameter is 665 used as it forces an exact (power of two) size of the ring buffer. 666 667 The number of possible CPUs is used for this computation ignoring 668 hotplugging making the computation optimal for the worst case 669 scenario while allowing a simple algorithm to be used from bootup. 670 671 Examples shift values and their meaning: 672 17 => 128 KB for each CPU 673 16 => 64 KB for each CPU 674 15 => 32 KB for each CPU 675 14 => 16 KB for each CPU 676 13 => 8 KB for each CPU 677 12 => 4 KB for each CPU 678 679config PRINTK_SAFE_LOG_BUF_SHIFT 680 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)" 681 range 10 21 682 default 13 683 depends on PRINTK 684 help 685 Select the size of an alternate printk per-CPU buffer where messages 686 printed from usafe contexts are temporary stored. One example would 687 be NMI messages, another one - printk recursion. The messages are 688 copied to the main log buffer in a safe context to avoid a deadlock. 689 The value defines the size as a power of 2. 690 691 Those messages are rare and limited. The largest one is when 692 a backtrace is printed. It usually fits into 4KB. Select 693 8KB if you want to be on the safe side. 694 695 Examples: 696 17 => 128 KB for each CPU 697 16 => 64 KB for each CPU 698 15 => 32 KB for each CPU 699 14 => 16 KB for each CPU 700 13 => 8 KB for each CPU 701 12 => 4 KB for each CPU 702 703# 704# Architectures with an unreliable sched_clock() should select this: 705# 706config HAVE_UNSTABLE_SCHED_CLOCK 707 bool 708 709config GENERIC_SCHED_CLOCK 710 bool 711 712menu "Scheduler features" 713 714config UCLAMP_TASK 715 bool "Enable utilization clamping for RT/FAIR tasks" 716 depends on CPU_FREQ_GOV_SCHEDUTIL 717 help 718 This feature enables the scheduler to track the clamped utilization 719 of each CPU based on RUNNABLE tasks scheduled on that CPU. 720 721 With this option, the user can specify the min and max CPU 722 utilization allowed for RUNNABLE tasks. The max utilization defines 723 the maximum frequency a task should use while the min utilization 724 defines the minimum frequency it should use. 725 726 Both min and max utilization clamp values are hints to the scheduler, 727 aiming at improving its frequency selection policy, but they do not 728 enforce or grant any specific bandwidth for tasks. 729 730 If in doubt, say N. 731 732config UCLAMP_BUCKETS_COUNT 733 int "Number of supported utilization clamp buckets" 734 range 5 20 735 default 5 736 depends on UCLAMP_TASK 737 help 738 Defines the number of clamp buckets to use. The range of each bucket 739 will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the 740 number of clamp buckets the finer their granularity and the higher 741 the precision of clamping aggregation and tracking at run-time. 742 743 For example, with the minimum configuration value we will have 5 744 clamp buckets tracking 20% utilization each. A 25% boosted tasks will 745 be refcounted in the [20..39]% bucket and will set the bucket clamp 746 effective value to 25%. 747 If a second 30% boosted task should be co-scheduled on the same CPU, 748 that task will be refcounted in the same bucket of the first task and 749 it will boost the bucket clamp effective value to 30%. 750 The clamp effective value of a bucket is reset to its nominal value 751 (20% in the example above) when there are no more tasks refcounted in 752 that bucket. 753 754 An additional boost/capping margin can be added to some tasks. In the 755 example above the 25% task will be boosted to 30% until it exits the 756 CPU. If that should be considered not acceptable on certain systems, 757 it's always possible to reduce the margin by increasing the number of 758 clamp buckets to trade off used memory for run-time tracking 759 precision. 760 761 If in doubt, use the default value. 762 763endmenu 764 765# 766# For architectures that want to enable the support for NUMA-affine scheduler 767# balancing logic: 768# 769config ARCH_SUPPORTS_NUMA_BALANCING 770 bool 771 772# 773# For architectures that prefer to flush all TLBs after a number of pages 774# are unmapped instead of sending one IPI per page to flush. The architecture 775# must provide guarantees on what happens if a clean TLB cache entry is 776# written after the unmap. Details are in mm/rmap.c near the check for 777# should_defer_flush. The architecture should also consider if the full flush 778# and the refill costs are offset by the savings of sending fewer IPIs. 779config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH 780 bool 781 782# 783# For architectures that know their GCC __int128 support is sound 784# 785config ARCH_SUPPORTS_INT128 786 bool 787 788# For architectures that (ab)use NUMA to represent different memory regions 789# all cpu-local but of different latencies, such as SuperH. 790# 791config ARCH_WANT_NUMA_VARIABLE_LOCALITY 792 bool 793 794config NUMA_BALANCING 795 bool "Memory placement aware NUMA scheduler" 796 depends on ARCH_SUPPORTS_NUMA_BALANCING 797 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 798 depends on SMP && NUMA && MIGRATION 799 help 800 This option adds support for automatic NUMA aware memory/task placement. 801 The mechanism is quite primitive and is based on migrating memory when 802 it has references to the node the task is running on. 803 804 This system will be inactive on UMA systems. 805 806config NUMA_BALANCING_DEFAULT_ENABLED 807 bool "Automatically enable NUMA aware memory/task placement" 808 default y 809 depends on NUMA_BALANCING 810 help 811 If set, automatic NUMA balancing will be enabled if running on a NUMA 812 machine. 813 814menuconfig CGROUPS 815 bool "Control Group support" 816 select KERNFS 817 help 818 This option adds support for grouping sets of processes together, for 819 use with process control subsystems such as Cpusets, CFS, memory 820 controls or device isolation. 821 See 822 - Documentation/scheduler/sched-design-CFS.rst (CFS) 823 - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation 824 and resource control) 825 826 Say N if unsure. 827 828if CGROUPS 829 830config PAGE_COUNTER 831 bool 832 833config MEMCG 834 bool "Memory controller" 835 select PAGE_COUNTER 836 select EVENTFD 837 help 838 Provides control over the memory footprint of tasks in a cgroup. 839 840config MEMCG_SWAP 841 bool "Swap controller" 842 depends on MEMCG && SWAP 843 help 844 Provides control over the swap space consumed by tasks in a cgroup. 845 846config MEMCG_SWAP_ENABLED 847 bool "Swap controller enabled by default" 848 depends on MEMCG_SWAP 849 default y 850 help 851 Memory Resource Controller Swap Extension comes with its price in 852 a bigger memory consumption. General purpose distribution kernels 853 which want to enable the feature but keep it disabled by default 854 and let the user enable it by swapaccount=1 boot command line 855 parameter should have this option unselected. 856 For those who want to have the feature enabled by default should 857 select this option (if, for some reason, they need to disable it 858 then swapaccount=0 does the trick). 859 860config MEMCG_KMEM 861 bool 862 depends on MEMCG && !SLOB 863 default y 864 865config BLK_CGROUP 866 bool "IO controller" 867 depends on BLOCK 868 default n 869 ---help--- 870 Generic block IO controller cgroup interface. This is the common 871 cgroup interface which should be used by various IO controlling 872 policies. 873 874 Currently, CFQ IO scheduler uses it to recognize task groups and 875 control disk bandwidth allocation (proportional time slice allocation) 876 to such task groups. It is also used by bio throttling logic in 877 block layer to implement upper limit in IO rates on a device. 878 879 This option only enables generic Block IO controller infrastructure. 880 One needs to also enable actual IO controlling logic/policy. For 881 enabling proportional weight division of disk bandwidth in CFQ, set 882 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 883 CONFIG_BLK_DEV_THROTTLING=y. 884 885 See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information. 886 887config CGROUP_WRITEBACK 888 bool 889 depends on MEMCG && BLK_CGROUP 890 default y 891 892menuconfig CGROUP_SCHED 893 bool "CPU controller" 894 default n 895 help 896 This feature lets CPU scheduler recognize task groups and control CPU 897 bandwidth allocation to such task groups. It uses cgroups to group 898 tasks. 899 900if CGROUP_SCHED 901config FAIR_GROUP_SCHED 902 bool "Group scheduling for SCHED_OTHER" 903 depends on CGROUP_SCHED 904 default CGROUP_SCHED 905 906config CFS_BANDWIDTH 907 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 908 depends on FAIR_GROUP_SCHED 909 default n 910 help 911 This option allows users to define CPU bandwidth rates (limits) for 912 tasks running within the fair group scheduler. Groups with no limit 913 set are considered to be unconstrained and will run with no 914 restriction. 915 See Documentation/scheduler/sched-bwc.rst for more information. 916 917config RT_GROUP_SCHED 918 bool "Group scheduling for SCHED_RR/FIFO" 919 depends on CGROUP_SCHED 920 default n 921 help 922 This feature lets you explicitly allocate real CPU bandwidth 923 to task groups. If enabled, it will also make it impossible to 924 schedule realtime tasks for non-root users until you allocate 925 realtime bandwidth for them. 926 See Documentation/scheduler/sched-rt-group.rst for more information. 927 928endif #CGROUP_SCHED 929 930config CGROUP_PIDS 931 bool "PIDs controller" 932 help 933 Provides enforcement of process number limits in the scope of a 934 cgroup. Any attempt to fork more processes than is allowed in the 935 cgroup will fail. PIDs are fundamentally a global resource because it 936 is fairly trivial to reach PID exhaustion before you reach even a 937 conservative kmemcg limit. As a result, it is possible to grind a 938 system to halt without being limited by other cgroup policies. The 939 PIDs controller is designed to stop this from happening. 940 941 It should be noted that organisational operations (such as attaching 942 to a cgroup hierarchy) will *not* be blocked by the PIDs controller, 943 since the PIDs limit only affects a process's ability to fork, not to 944 attach to a cgroup. 945 946config CGROUP_RDMA 947 bool "RDMA controller" 948 help 949 Provides enforcement of RDMA resources defined by IB stack. 950 It is fairly easy for consumers to exhaust RDMA resources, which 951 can result into resource unavailability to other consumers. 952 RDMA controller is designed to stop this from happening. 953 Attaching processes with active RDMA resources to the cgroup 954 hierarchy is allowed even if can cross the hierarchy's limit. 955 956config CGROUP_FREEZER 957 bool "Freezer controller" 958 help 959 Provides a way to freeze and unfreeze all tasks in a 960 cgroup. 961 962 This option affects the ORIGINAL cgroup interface. The cgroup2 memory 963 controller includes important in-kernel memory consumers per default. 964 965 If you're using cgroup2, say N. 966 967config CGROUP_HUGETLB 968 bool "HugeTLB controller" 969 depends on HUGETLB_PAGE 970 select PAGE_COUNTER 971 default n 972 help 973 Provides a cgroup controller for HugeTLB pages. 974 When you enable this, you can put a per cgroup limit on HugeTLB usage. 975 The limit is enforced during page fault. Since HugeTLB doesn't 976 support page reclaim, enforcing the limit at page fault time implies 977 that, the application will get SIGBUS signal if it tries to access 978 HugeTLB pages beyond its limit. This requires the application to know 979 beforehand how much HugeTLB pages it would require for its use. The 980 control group is tracked in the third page lru pointer. This means 981 that we cannot use the controller with huge page less than 3 pages. 982 983config CPUSETS 984 bool "Cpuset controller" 985 depends on SMP 986 help 987 This option will let you create and manage CPUSETs which 988 allow dynamically partitioning a system into sets of CPUs and 989 Memory Nodes and assigning tasks to run only within those sets. 990 This is primarily useful on large SMP or NUMA systems. 991 992 Say N if unsure. 993 994config PROC_PID_CPUSET 995 bool "Include legacy /proc/<pid>/cpuset file" 996 depends on CPUSETS 997 default y 998 999config CGROUP_DEVICE 1000 bool "Device controller" 1001 help 1002 Provides a cgroup controller implementing whitelists for 1003 devices which a process in the cgroup can mknod or open. 1004 1005config CGROUP_CPUACCT 1006 bool "Simple CPU accounting controller" 1007 help 1008 Provides a simple controller for monitoring the 1009 total CPU consumed by the tasks in a cgroup. 1010 1011config CGROUP_PERF 1012 bool "Perf controller" 1013 depends on PERF_EVENTS 1014 help 1015 This option extends the perf per-cpu mode to restrict monitoring 1016 to threads which belong to the cgroup specified and run on the 1017 designated cpu. 1018 1019 Say N if unsure. 1020 1021config CGROUP_BPF 1022 bool "Support for eBPF programs attached to cgroups" 1023 depends on BPF_SYSCALL 1024 select SOCK_CGROUP_DATA 1025 help 1026 Allow attaching eBPF programs to a cgroup using the bpf(2) 1027 syscall command BPF_PROG_ATTACH. 1028 1029 In which context these programs are accessed depends on the type 1030 of attachment. For instance, programs that are attached using 1031 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of 1032 inet sockets. 1033 1034config CGROUP_DEBUG 1035 bool "Debug controller" 1036 default n 1037 depends on DEBUG_KERNEL 1038 help 1039 This option enables a simple controller that exports 1040 debugging information about the cgroups framework. This 1041 controller is for control cgroup debugging only. Its 1042 interfaces are not stable. 1043 1044 Say N. 1045 1046config SOCK_CGROUP_DATA 1047 bool 1048 default n 1049 1050endif # CGROUPS 1051 1052menuconfig NAMESPACES 1053 bool "Namespaces support" if EXPERT 1054 depends on MULTIUSER 1055 default !EXPERT 1056 help 1057 Provides the way to make tasks work with different objects using 1058 the same id. For example same IPC id may refer to different objects 1059 or same user id or pid may refer to different tasks when used in 1060 different namespaces. 1061 1062if NAMESPACES 1063 1064config UTS_NS 1065 bool "UTS namespace" 1066 default y 1067 help 1068 In this namespace tasks see different info provided with the 1069 uname() system call 1070 1071config IPC_NS 1072 bool "IPC namespace" 1073 depends on (SYSVIPC || POSIX_MQUEUE) 1074 default y 1075 help 1076 In this namespace tasks work with IPC ids which correspond to 1077 different IPC objects in different namespaces. 1078 1079config USER_NS 1080 bool "User namespace" 1081 default n 1082 help 1083 This allows containers, i.e. vservers, to use user namespaces 1084 to provide different user info for different servers. 1085 1086 When user namespaces are enabled in the kernel it is 1087 recommended that the MEMCG option also be enabled and that 1088 user-space use the memory control groups to limit the amount 1089 of memory a memory unprivileged users can use. 1090 1091 If unsure, say N. 1092 1093config PID_NS 1094 bool "PID Namespaces" 1095 default y 1096 help 1097 Support process id namespaces. This allows having multiple 1098 processes with the same pid as long as they are in different 1099 pid namespaces. This is a building block of containers. 1100 1101config NET_NS 1102 bool "Network namespace" 1103 depends on NET 1104 default y 1105 help 1106 Allow user space to create what appear to be multiple instances 1107 of the network stack. 1108 1109endif # NAMESPACES 1110 1111config CHECKPOINT_RESTORE 1112 bool "Checkpoint/restore support" 1113 select PROC_CHILDREN 1114 default n 1115 help 1116 Enables additional kernel features in a sake of checkpoint/restore. 1117 In particular it adds auxiliary prctl codes to setup process text, 1118 data and heap segment sizes, and a few additional /proc filesystem 1119 entries. 1120 1121 If unsure, say N here. 1122 1123config SCHED_AUTOGROUP 1124 bool "Automatic process group scheduling" 1125 select CGROUPS 1126 select CGROUP_SCHED 1127 select FAIR_GROUP_SCHED 1128 help 1129 This option optimizes the scheduler for common desktop workloads by 1130 automatically creating and populating task groups. This separation 1131 of workloads isolates aggressive CPU burners (like build jobs) from 1132 desktop applications. Task group autogeneration is currently based 1133 upon task session. 1134 1135config SYSFS_DEPRECATED 1136 bool "Enable deprecated sysfs features to support old userspace tools" 1137 depends on SYSFS 1138 default n 1139 help 1140 This option adds code that switches the layout of the "block" class 1141 devices, to not show up in /sys/class/block/, but only in 1142 /sys/block/. 1143 1144 This switch is only active when the sysfs.deprecated=1 boot option is 1145 passed or the SYSFS_DEPRECATED_V2 option is set. 1146 1147 This option allows new kernels to run on old distributions and tools, 1148 which might get confused by /sys/class/block/. Since 2007/2008 all 1149 major distributions and tools handle this just fine. 1150 1151 Recent distributions and userspace tools after 2009/2010 depend on 1152 the existence of /sys/class/block/, and will not work with this 1153 option enabled. 1154 1155 Only if you are using a new kernel on an old distribution, you might 1156 need to say Y here. 1157 1158config SYSFS_DEPRECATED_V2 1159 bool "Enable deprecated sysfs features by default" 1160 default n 1161 depends on SYSFS 1162 depends on SYSFS_DEPRECATED 1163 help 1164 Enable deprecated sysfs by default. 1165 1166 See the CONFIG_SYSFS_DEPRECATED option for more details about this 1167 option. 1168 1169 Only if you are using a new kernel on an old distribution, you might 1170 need to say Y here. Even then, odds are you would not need it 1171 enabled, you can always pass the boot option if absolutely necessary. 1172 1173config RELAY 1174 bool "Kernel->user space relay support (formerly relayfs)" 1175 select IRQ_WORK 1176 help 1177 This option enables support for relay interface support in 1178 certain file systems (such as debugfs). 1179 It is designed to provide an efficient mechanism for tools and 1180 facilities to relay large amounts of data from kernel space to 1181 user space. 1182 1183 If unsure, say N. 1184 1185config BLK_DEV_INITRD 1186 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 1187 help 1188 The initial RAM filesystem is a ramfs which is loaded by the 1189 boot loader (loadlin or lilo) and that is mounted as root 1190 before the normal boot procedure. It is typically used to 1191 load modules needed to mount the "real" root file system, 1192 etc. See <file:Documentation/admin-guide/initrd.rst> for details. 1193 1194 If RAM disk support (BLK_DEV_RAM) is also included, this 1195 also enables initial RAM disk (initrd) support and adds 1196 15 Kbytes (more on some other architectures) to the kernel size. 1197 1198 If unsure say Y. 1199 1200if BLK_DEV_INITRD 1201 1202source "usr/Kconfig" 1203 1204endif 1205 1206choice 1207 prompt "Compiler optimization level" 1208 default CC_OPTIMIZE_FOR_PERFORMANCE 1209 1210config CC_OPTIMIZE_FOR_PERFORMANCE 1211 bool "Optimize for performance" 1212 help 1213 This is the default optimization level for the kernel, building 1214 with the "-O2" compiler flag for best performance and most 1215 helpful compile-time warnings. 1216 1217config CC_OPTIMIZE_FOR_SIZE 1218 bool "Optimize for size" 1219 imply CC_DISABLE_WARN_MAYBE_UNINITIALIZED # avoid false positives 1220 help 1221 Enabling this option will pass "-Os" instead of "-O2" to 1222 your compiler resulting in a smaller kernel. 1223 1224 If unsure, say N. 1225 1226endchoice 1227 1228config HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1229 bool 1230 help 1231 This requires that the arch annotates or otherwise protects 1232 its external entry points from being discarded. Linker scripts 1233 must also merge .text.*, .data.*, and .bss.* correctly into 1234 output sections. Care must be taken not to pull in unrelated 1235 sections (e.g., '.text.init'). Typically '.' in section names 1236 is used to distinguish them from label names / C identifiers. 1237 1238config LD_DEAD_CODE_DATA_ELIMINATION 1239 bool "Dead code and data elimination (EXPERIMENTAL)" 1240 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION 1241 depends on EXPERT 1242 depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800) 1243 depends on $(cc-option,-ffunction-sections -fdata-sections) 1244 depends on $(ld-option,--gc-sections) 1245 help 1246 Enable this if you want to do dead code and data elimination with 1247 the linker by compiling with -ffunction-sections -fdata-sections, 1248 and linking with --gc-sections. 1249 1250 This can reduce on disk and in-memory size of the kernel 1251 code and static data, particularly for small configs and 1252 on small systems. This has the possibility of introducing 1253 silently broken kernel if the required annotations are not 1254 present. This option is not well tested yet, so use at your 1255 own risk. 1256 1257config SYSCTL 1258 bool 1259 1260config HAVE_UID16 1261 bool 1262 1263config SYSCTL_EXCEPTION_TRACE 1264 bool 1265 help 1266 Enable support for /proc/sys/debug/exception-trace. 1267 1268config SYSCTL_ARCH_UNALIGN_NO_WARN 1269 bool 1270 help 1271 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1272 Allows arch to define/use @no_unaligned_warning to possibly warn 1273 about unaligned access emulation going on under the hood. 1274 1275config SYSCTL_ARCH_UNALIGN_ALLOW 1276 bool 1277 help 1278 Enable support for /proc/sys/kernel/unaligned-trap 1279 Allows arches to define/use @unaligned_enabled to runtime toggle 1280 the unaligned access emulation. 1281 see arch/parisc/kernel/unaligned.c for reference 1282 1283config HAVE_PCSPKR_PLATFORM 1284 bool 1285 1286# interpreter that classic socket filters depend on 1287config BPF 1288 bool 1289 1290menuconfig EXPERT 1291 bool "Configure standard kernel features (expert users)" 1292 # Unhide debug options, to make the on-by-default options visible 1293 select DEBUG_KERNEL 1294 help 1295 This option allows certain base kernel options and settings 1296 to be disabled or tweaked. This is for specialized 1297 environments which can tolerate a "non-standard" kernel. 1298 Only use this if you really know what you are doing. 1299 1300config UID16 1301 bool "Enable 16-bit UID system calls" if EXPERT 1302 depends on HAVE_UID16 && MULTIUSER 1303 default y 1304 help 1305 This enables the legacy 16-bit UID syscall wrappers. 1306 1307config MULTIUSER 1308 bool "Multiple users, groups and capabilities support" if EXPERT 1309 default y 1310 help 1311 This option enables support for non-root users, groups and 1312 capabilities. 1313 1314 If you say N here, all processes will run with UID 0, GID 0, and all 1315 possible capabilities. Saying N here also compiles out support for 1316 system calls related to UIDs, GIDs, and capabilities, such as setuid, 1317 setgid, and capset. 1318 1319 If unsure, say Y here. 1320 1321config SGETMASK_SYSCALL 1322 bool "sgetmask/ssetmask syscalls support" if EXPERT 1323 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH 1324 ---help--- 1325 sys_sgetmask and sys_ssetmask are obsolete system calls 1326 no longer supported in libc but still enabled by default in some 1327 architectures. 1328 1329 If unsure, leave the default option here. 1330 1331config SYSFS_SYSCALL 1332 bool "Sysfs syscall support" if EXPERT 1333 default y 1334 ---help--- 1335 sys_sysfs is an obsolete system call no longer supported in libc. 1336 Note that disabling this option is more secure but might break 1337 compatibility with some systems. 1338 1339 If unsure say Y here. 1340 1341config SYSCTL_SYSCALL 1342 bool "Sysctl syscall support" if EXPERT 1343 depends on PROC_SYSCTL 1344 default n 1345 select SYSCTL 1346 ---help--- 1347 sys_sysctl uses binary paths that have been found challenging 1348 to properly maintain and use. The interface in /proc/sys 1349 using paths with ascii names is now the primary path to this 1350 information. 1351 1352 Almost nothing using the binary sysctl interface so if you are 1353 trying to save some space it is probably safe to disable this, 1354 making your kernel marginally smaller. 1355 1356 If unsure say N here. 1357 1358config FHANDLE 1359 bool "open by fhandle syscalls" if EXPERT 1360 select EXPORTFS 1361 default y 1362 help 1363 If you say Y here, a user level program will be able to map 1364 file names to handle and then later use the handle for 1365 different file system operations. This is useful in implementing 1366 userspace file servers, which now track files using handles instead 1367 of names. The handle would remain the same even if file names 1368 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 1369 syscalls. 1370 1371config POSIX_TIMERS 1372 bool "Posix Clocks & timers" if EXPERT 1373 default y 1374 help 1375 This includes native support for POSIX timers to the kernel. 1376 Some embedded systems have no use for them and therefore they 1377 can be configured out to reduce the size of the kernel image. 1378 1379 When this option is disabled, the following syscalls won't be 1380 available: timer_create, timer_gettime: timer_getoverrun, 1381 timer_settime, timer_delete, clock_adjtime, getitimer, 1382 setitimer, alarm. Furthermore, the clock_settime, clock_gettime, 1383 clock_getres and clock_nanosleep syscalls will be limited to 1384 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only. 1385 1386 If unsure say y. 1387 1388config PRINTK 1389 default y 1390 bool "Enable support for printk" if EXPERT 1391 select IRQ_WORK 1392 help 1393 This option enables normal printk support. Removing it 1394 eliminates most of the message strings from the kernel image 1395 and makes the kernel more or less silent. As this makes it 1396 very difficult to diagnose system problems, saying N here is 1397 strongly discouraged. 1398 1399config PRINTK_NMI 1400 def_bool y 1401 depends on PRINTK 1402 depends on HAVE_NMI 1403 1404config BUG 1405 bool "BUG() support" if EXPERT 1406 default y 1407 help 1408 Disabling this option eliminates support for BUG and WARN, reducing 1409 the size of your kernel image and potentially quietly ignoring 1410 numerous fatal conditions. You should only consider disabling this 1411 option for embedded systems with no facilities for reporting errors. 1412 Just say Y. 1413 1414config ELF_CORE 1415 depends on COREDUMP 1416 default y 1417 bool "Enable ELF core dumps" if EXPERT 1418 help 1419 Enable support for generating core dumps. Disabling saves about 4k. 1420 1421 1422config PCSPKR_PLATFORM 1423 bool "Enable PC-Speaker support" if EXPERT 1424 depends on HAVE_PCSPKR_PLATFORM 1425 select I8253_LOCK 1426 default y 1427 help 1428 This option allows to disable the internal PC-Speaker 1429 support, saving some memory. 1430 1431config BASE_FULL 1432 default y 1433 bool "Enable full-sized data structures for core" if EXPERT 1434 help 1435 Disabling this option reduces the size of miscellaneous core 1436 kernel data structures. This saves memory on small machines, 1437 but may reduce performance. 1438 1439config FUTEX 1440 bool "Enable futex support" if EXPERT 1441 default y 1442 imply RT_MUTEXES 1443 help 1444 Disabling this option will cause the kernel to be built without 1445 support for "fast userspace mutexes". The resulting kernel may not 1446 run glibc-based applications correctly. 1447 1448config FUTEX_PI 1449 bool 1450 depends on FUTEX && RT_MUTEXES 1451 default y 1452 1453config HAVE_FUTEX_CMPXCHG 1454 bool 1455 depends on FUTEX 1456 help 1457 Architectures should select this if futex_atomic_cmpxchg_inatomic() 1458 is implemented and always working. This removes a couple of runtime 1459 checks. 1460 1461config EPOLL 1462 bool "Enable eventpoll support" if EXPERT 1463 default y 1464 help 1465 Disabling this option will cause the kernel to be built without 1466 support for epoll family of system calls. 1467 1468config SIGNALFD 1469 bool "Enable signalfd() system call" if EXPERT 1470 default y 1471 help 1472 Enable the signalfd() system call that allows to receive signals 1473 on a file descriptor. 1474 1475 If unsure, say Y. 1476 1477config TIMERFD 1478 bool "Enable timerfd() system call" if EXPERT 1479 default y 1480 help 1481 Enable the timerfd() system call that allows to receive timer 1482 events on a file descriptor. 1483 1484 If unsure, say Y. 1485 1486config EVENTFD 1487 bool "Enable eventfd() system call" if EXPERT 1488 default y 1489 help 1490 Enable the eventfd() system call that allows to receive both 1491 kernel notification (ie. KAIO) or userspace notifications. 1492 1493 If unsure, say Y. 1494 1495config SHMEM 1496 bool "Use full shmem filesystem" if EXPERT 1497 default y 1498 depends on MMU 1499 help 1500 The shmem is an internal filesystem used to manage shared memory. 1501 It is backed by swap and manages resource limits. It is also exported 1502 to userspace as tmpfs if TMPFS is enabled. Disabling this 1503 option replaces shmem and tmpfs with the much simpler ramfs code, 1504 which may be appropriate on small systems without swap. 1505 1506config AIO 1507 bool "Enable AIO support" if EXPERT 1508 default y 1509 help 1510 This option enables POSIX asynchronous I/O which may by used 1511 by some high performance threaded applications. Disabling 1512 this option saves about 7k. 1513 1514config IO_URING 1515 bool "Enable IO uring support" if EXPERT 1516 select ANON_INODES 1517 default y 1518 help 1519 This option enables support for the io_uring interface, enabling 1520 applications to submit and complete IO through submission and 1521 completion rings that are shared between the kernel and application. 1522 1523config ADVISE_SYSCALLS 1524 bool "Enable madvise/fadvise syscalls" if EXPERT 1525 default y 1526 help 1527 This option enables the madvise and fadvise syscalls, used by 1528 applications to advise the kernel about their future memory or file 1529 usage, improving performance. If building an embedded system where no 1530 applications use these syscalls, you can disable this option to save 1531 space. 1532 1533config MEMBARRIER 1534 bool "Enable membarrier() system call" if EXPERT 1535 default y 1536 help 1537 Enable the membarrier() system call that allows issuing memory 1538 barriers across all running threads, which can be used to distribute 1539 the cost of user-space memory barriers asymmetrically by transforming 1540 pairs of memory barriers into pairs consisting of membarrier() and a 1541 compiler barrier. 1542 1543 If unsure, say Y. 1544 1545config KALLSYMS 1546 bool "Load all symbols for debugging/ksymoops" if EXPERT 1547 default y 1548 help 1549 Say Y here to let the kernel print out symbolic crash information and 1550 symbolic stack backtraces. This increases the size of the kernel 1551 somewhat, as all symbols have to be loaded into the kernel image. 1552 1553config KALLSYMS_ALL 1554 bool "Include all symbols in kallsyms" 1555 depends on DEBUG_KERNEL && KALLSYMS 1556 help 1557 Normally kallsyms only contains the symbols of functions for nicer 1558 OOPS messages and backtraces (i.e., symbols from the text and inittext 1559 sections). This is sufficient for most cases. And only in very rare 1560 cases (e.g., when a debugger is used) all symbols are required (e.g., 1561 names of variables from the data sections, etc). 1562 1563 This option makes sure that all symbols are loaded into the kernel 1564 image (i.e., symbols from all sections) in cost of increased kernel 1565 size (depending on the kernel configuration, it may be 300KiB or 1566 something like this). 1567 1568 Say N unless you really need all symbols. 1569 1570config KALLSYMS_ABSOLUTE_PERCPU 1571 bool 1572 depends on KALLSYMS 1573 default X86_64 && SMP 1574 1575config KALLSYMS_BASE_RELATIVE 1576 bool 1577 depends on KALLSYMS 1578 default !IA64 1579 help 1580 Instead of emitting them as absolute values in the native word size, 1581 emit the symbol references in the kallsyms table as 32-bit entries, 1582 each containing a relative value in the range [base, base + U32_MAX] 1583 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either 1584 an absolute value in the range [0, S32_MAX] or a relative value in the 1585 range [base, base + S32_MAX], where base is the lowest relative symbol 1586 address encountered in the image. 1587 1588 On 64-bit builds, this reduces the size of the address table by 50%, 1589 but more importantly, it results in entries whose values are build 1590 time constants, and no relocation pass is required at runtime to fix 1591 up the entries based on the runtime load address of the kernel. 1592 1593# end of the "standard kernel features (expert users)" menu 1594 1595# syscall, maps, verifier 1596config BPF_SYSCALL 1597 bool "Enable bpf() system call" 1598 select BPF 1599 select IRQ_WORK 1600 default n 1601 help 1602 Enable the bpf() system call that allows to manipulate eBPF 1603 programs and maps via file descriptors. 1604 1605config BPF_JIT_ALWAYS_ON 1606 bool "Permanently enable BPF JIT and remove BPF interpreter" 1607 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT 1608 help 1609 Enables BPF JIT and removes BPF interpreter to avoid 1610 speculative execution of BPF instructions by the interpreter 1611 1612config USERFAULTFD 1613 bool "Enable userfaultfd() system call" 1614 depends on MMU 1615 help 1616 Enable the userfaultfd() system call that allows to intercept and 1617 handle page faults in userland. 1618 1619config ARCH_HAS_MEMBARRIER_CALLBACKS 1620 bool 1621 1622config ARCH_HAS_MEMBARRIER_SYNC_CORE 1623 bool 1624 1625config RSEQ 1626 bool "Enable rseq() system call" if EXPERT 1627 default y 1628 depends on HAVE_RSEQ 1629 select MEMBARRIER 1630 help 1631 Enable the restartable sequences system call. It provides a 1632 user-space cache for the current CPU number value, which 1633 speeds up getting the current CPU number from user-space, 1634 as well as an ABI to speed up user-space operations on 1635 per-CPU data. 1636 1637 If unsure, say Y. 1638 1639config DEBUG_RSEQ 1640 default n 1641 bool "Enabled debugging of rseq() system call" if EXPERT 1642 depends on RSEQ && DEBUG_KERNEL 1643 help 1644 Enable extra debugging checks for the rseq system call. 1645 1646 If unsure, say N. 1647 1648config EMBEDDED 1649 bool "Embedded system" 1650 option allnoconfig_y 1651 select EXPERT 1652 help 1653 This option should be enabled if compiling the kernel for 1654 an embedded system so certain expert options are available 1655 for configuration. 1656 1657config HAVE_PERF_EVENTS 1658 bool 1659 help 1660 See tools/perf/design.txt for details. 1661 1662config PERF_USE_VMALLOC 1663 bool 1664 help 1665 See tools/perf/design.txt for details 1666 1667config PC104 1668 bool "PC/104 support" if EXPERT 1669 help 1670 Expose PC/104 form factor device drivers and options available for 1671 selection and configuration. Enable this option if your target 1672 machine has a PC/104 bus. 1673 1674menu "Kernel Performance Events And Counters" 1675 1676config PERF_EVENTS 1677 bool "Kernel performance events and counters" 1678 default y if PROFILING 1679 depends on HAVE_PERF_EVENTS 1680 select IRQ_WORK 1681 select SRCU 1682 help 1683 Enable kernel support for various performance events provided 1684 by software and hardware. 1685 1686 Software events are supported either built-in or via the 1687 use of generic tracepoints. 1688 1689 Most modern CPUs support performance events via performance 1690 counter registers. These registers count the number of certain 1691 types of hw events: such as instructions executed, cachemisses 1692 suffered, or branches mis-predicted - without slowing down the 1693 kernel or applications. These registers can also trigger interrupts 1694 when a threshold number of events have passed - and can thus be 1695 used to profile the code that runs on that CPU. 1696 1697 The Linux Performance Event subsystem provides an abstraction of 1698 these software and hardware event capabilities, available via a 1699 system call and used by the "perf" utility in tools/perf/. It 1700 provides per task and per CPU counters, and it provides event 1701 capabilities on top of those. 1702 1703 Say Y if unsure. 1704 1705config DEBUG_PERF_USE_VMALLOC 1706 default n 1707 bool "Debug: use vmalloc to back perf mmap() buffers" 1708 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC 1709 select PERF_USE_VMALLOC 1710 help 1711 Use vmalloc memory to back perf mmap() buffers. 1712 1713 Mostly useful for debugging the vmalloc code on platforms 1714 that don't require it. 1715 1716 Say N if unsure. 1717 1718endmenu 1719 1720config VM_EVENT_COUNTERS 1721 default y 1722 bool "Enable VM event counters for /proc/vmstat" if EXPERT 1723 help 1724 VM event counters are needed for event counts to be shown. 1725 This option allows the disabling of the VM event counters 1726 on EXPERT systems. /proc/vmstat will only show page counts 1727 if VM event counters are disabled. 1728 1729config SLUB_DEBUG 1730 default y 1731 bool "Enable SLUB debugging support" if EXPERT 1732 depends on SLUB && SYSFS 1733 help 1734 SLUB has extensive debug support features. Disabling these can 1735 result in significant savings in code size. This also disables 1736 SLUB sysfs support. /sys/slab will not exist and there will be 1737 no support for cache validation etc. 1738 1739config SLUB_MEMCG_SYSFS_ON 1740 default n 1741 bool "Enable memcg SLUB sysfs support by default" if EXPERT 1742 depends on SLUB && SYSFS && MEMCG 1743 help 1744 SLUB creates a directory under /sys/kernel/slab for each 1745 allocation cache to host info and debug files. If memory 1746 cgroup is enabled, each cache can have per memory cgroup 1747 caches. SLUB can create the same sysfs directories for these 1748 caches under /sys/kernel/slab/CACHE/cgroup but it can lead 1749 to a very high number of debug files being created. This is 1750 controlled by slub_memcg_sysfs boot parameter and this 1751 config option determines the parameter's default value. 1752 1753config COMPAT_BRK 1754 bool "Disable heap randomization" 1755 default y 1756 help 1757 Randomizing heap placement makes heap exploits harder, but it 1758 also breaks ancient binaries (including anything libc5 based). 1759 This option changes the bootup default to heap randomization 1760 disabled, and can be overridden at runtime by setting 1761 /proc/sys/kernel/randomize_va_space to 2. 1762 1763 On non-ancient distros (post-2000 ones) N is usually a safe choice. 1764 1765choice 1766 prompt "Choose SLAB allocator" 1767 default SLUB 1768 help 1769 This option allows to select a slab allocator. 1770 1771config SLAB 1772 bool "SLAB" 1773 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1774 help 1775 The regular slab allocator that is established and known to work 1776 well in all environments. It organizes cache hot objects in 1777 per cpu and per node queues. 1778 1779config SLUB 1780 bool "SLUB (Unqueued Allocator)" 1781 select HAVE_HARDENED_USERCOPY_ALLOCATOR 1782 help 1783 SLUB is a slab allocator that minimizes cache line usage 1784 instead of managing queues of cached objects (SLAB approach). 1785 Per cpu caching is realized using slabs of objects instead 1786 of queues of objects. SLUB can use memory efficiently 1787 and has enhanced diagnostics. SLUB is the default choice for 1788 a slab allocator. 1789 1790config SLOB 1791 depends on EXPERT 1792 bool "SLOB (Simple Allocator)" 1793 help 1794 SLOB replaces the stock allocator with a drastically simpler 1795 allocator. SLOB is generally more space efficient but 1796 does not perform as well on large systems. 1797 1798endchoice 1799 1800config SLAB_MERGE_DEFAULT 1801 bool "Allow slab caches to be merged" 1802 default y 1803 help 1804 For reduced kernel memory fragmentation, slab caches can be 1805 merged when they share the same size and other characteristics. 1806 This carries a risk of kernel heap overflows being able to 1807 overwrite objects from merged caches (and more easily control 1808 cache layout), which makes such heap attacks easier to exploit 1809 by attackers. By keeping caches unmerged, these kinds of exploits 1810 can usually only damage objects in the same cache. To disable 1811 merging at runtime, "slab_nomerge" can be passed on the kernel 1812 command line. 1813 1814config SLAB_FREELIST_RANDOM 1815 default n 1816 depends on SLAB || SLUB 1817 bool "SLAB freelist randomization" 1818 help 1819 Randomizes the freelist order used on creating new pages. This 1820 security feature reduces the predictability of the kernel slab 1821 allocator against heap overflows. 1822 1823config SLAB_FREELIST_HARDENED 1824 bool "Harden slab freelist metadata" 1825 depends on SLUB 1826 help 1827 Many kernel heap attacks try to target slab cache metadata and 1828 other infrastructure. This options makes minor performance 1829 sacrifices to harden the kernel slab allocator against common 1830 freelist exploit methods. 1831 1832config SHUFFLE_PAGE_ALLOCATOR 1833 bool "Page allocator randomization" 1834 default SLAB_FREELIST_RANDOM && ACPI_NUMA 1835 help 1836 Randomization of the page allocator improves the average 1837 utilization of a direct-mapped memory-side-cache. See section 1838 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI 1839 6.2a specification for an example of how a platform advertises 1840 the presence of a memory-side-cache. There are also incidental 1841 security benefits as it reduces the predictability of page 1842 allocations to compliment SLAB_FREELIST_RANDOM, but the 1843 default granularity of shuffling on the "MAX_ORDER - 1" i.e, 1844 10th order of pages is selected based on cache utilization 1845 benefits on x86. 1846 1847 While the randomization improves cache utilization it may 1848 negatively impact workloads on platforms without a cache. For 1849 this reason, by default, the randomization is enabled only 1850 after runtime detection of a direct-mapped memory-side-cache. 1851 Otherwise, the randomization may be force enabled with the 1852 'page_alloc.shuffle' kernel command line parameter. 1853 1854 Say Y if unsure. 1855 1856config SLUB_CPU_PARTIAL 1857 default y 1858 depends on SLUB && SMP 1859 bool "SLUB per cpu partial cache" 1860 help 1861 Per cpu partial caches accelerate objects allocation and freeing 1862 that is local to a processor at the price of more indeterminism 1863 in the latency of the free. On overflow these caches will be cleared 1864 which requires the taking of locks that may cause latency spikes. 1865 Typically one would choose no for a realtime system. 1866 1867config MMAP_ALLOW_UNINITIALIZED 1868 bool "Allow mmapped anonymous memory to be uninitialized" 1869 depends on EXPERT && !MMU 1870 default n 1871 help 1872 Normally, and according to the Linux spec, anonymous memory obtained 1873 from mmap() has its contents cleared before it is passed to 1874 userspace. Enabling this config option allows you to request that 1875 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1876 providing a huge performance boost. If this option is not enabled, 1877 then the flag will be ignored. 1878 1879 This is taken advantage of by uClibc's malloc(), and also by 1880 ELF-FDPIC binfmt's brk and stack allocator. 1881 1882 Because of the obvious security issues, this option should only be 1883 enabled on embedded devices where you control what is run in 1884 userspace. Since that isn't generally a problem on no-MMU systems, 1885 it is normally safe to say Y here. 1886 1887 See Documentation/nommu-mmap.txt for more information. 1888 1889config SYSTEM_DATA_VERIFICATION 1890 def_bool n 1891 select SYSTEM_TRUSTED_KEYRING 1892 select KEYS 1893 select CRYPTO 1894 select CRYPTO_RSA 1895 select ASYMMETRIC_KEY_TYPE 1896 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1897 select ASN1 1898 select OID_REGISTRY 1899 select X509_CERTIFICATE_PARSER 1900 select PKCS7_MESSAGE_PARSER 1901 help 1902 Provide PKCS#7 message verification using the contents of the system 1903 trusted keyring to provide public keys. This then can be used for 1904 module verification, kexec image verification and firmware blob 1905 verification. 1906 1907config PROFILING 1908 bool "Profiling support" 1909 help 1910 Say Y here to enable the extended profiling support mechanisms used 1911 by profilers such as OProfile. 1912 1913# 1914# Place an empty function call at each tracepoint site. Can be 1915# dynamically changed for a probe function. 1916# 1917config TRACEPOINTS 1918 bool 1919 1920endmenu # General setup 1921 1922source "arch/Kconfig" 1923 1924config RT_MUTEXES 1925 bool 1926 1927config BASE_SMALL 1928 int 1929 default 0 if BASE_FULL 1930 default 1 if !BASE_FULL 1931 1932menuconfig MODULES 1933 bool "Enable loadable module support" 1934 option modules 1935 help 1936 Kernel modules are small pieces of compiled code which can 1937 be inserted in the running kernel, rather than being 1938 permanently built into the kernel. You use the "modprobe" 1939 tool to add (and sometimes remove) them. If you say Y here, 1940 many parts of the kernel can be built as modules (by 1941 answering M instead of Y where indicated): this is most 1942 useful for infrequently used options which are not required 1943 for booting. For more information, see the man pages for 1944 modprobe, lsmod, modinfo, insmod and rmmod. 1945 1946 If you say Y here, you will need to run "make 1947 modules_install" to put the modules under /lib/modules/ 1948 where modprobe can find them (you may need to be root to do 1949 this). 1950 1951 If unsure, say Y. 1952 1953if MODULES 1954 1955config MODULE_FORCE_LOAD 1956 bool "Forced module loading" 1957 default n 1958 help 1959 Allow loading of modules without version information (ie. modprobe 1960 --force). Forced module loading sets the 'F' (forced) taint flag and 1961 is usually a really bad idea. 1962 1963config MODULE_UNLOAD 1964 bool "Module unloading" 1965 help 1966 Without this option you will not be able to unload any 1967 modules (note that some modules may not be unloadable 1968 anyway), which makes your kernel smaller, faster 1969 and simpler. If unsure, say Y. 1970 1971config MODULE_FORCE_UNLOAD 1972 bool "Forced module unloading" 1973 depends on MODULE_UNLOAD 1974 help 1975 This option allows you to force a module to unload, even if the 1976 kernel believes it is unsafe: the kernel will remove the module 1977 without waiting for anyone to stop using it (using the -f option to 1978 rmmod). This is mainly for kernel developers and desperate users. 1979 If unsure, say N. 1980 1981config MODVERSIONS 1982 bool "Module versioning support" 1983 help 1984 Usually, you have to use modules compiled with your kernel. 1985 Saying Y here makes it sometimes possible to use modules 1986 compiled for different kernels, by adding enough information 1987 to the modules to (hopefully) spot any changes which would 1988 make them incompatible with the kernel you are running. If 1989 unsure, say N. 1990 1991config MODULE_REL_CRCS 1992 bool 1993 depends on MODVERSIONS 1994 1995config MODULE_SRCVERSION_ALL 1996 bool "Source checksum for all modules" 1997 help 1998 Modules which contain a MODULE_VERSION get an extra "srcversion" 1999 field inserted into their modinfo section, which contains a 2000 sum of the source files which made it. This helps maintainers 2001 see exactly which source was used to build a module (since 2002 others sometimes change the module source without updating 2003 the version). With this option, such a "srcversion" field 2004 will be created for all modules. If unsure, say N. 2005 2006config MODULE_SIG 2007 bool "Module signature verification" 2008 depends on MODULES 2009 select SYSTEM_DATA_VERIFICATION 2010 help 2011 Check modules for valid signatures upon load: the signature 2012 is simply appended to the module. For more information see 2013 <file:Documentation/admin-guide/module-signing.rst>. 2014 2015 Note that this option adds the OpenSSL development packages as a 2016 kernel build dependency so that the signing tool can use its crypto 2017 library. 2018 2019 !!!WARNING!!! If you enable this option, you MUST make sure that the 2020 module DOES NOT get stripped after being signed. This includes the 2021 debuginfo strip done by some packagers (such as rpmbuild) and 2022 inclusion into an initramfs that wants the module size reduced. 2023 2024config MODULE_SIG_FORCE 2025 bool "Require modules to be validly signed" 2026 depends on MODULE_SIG 2027 help 2028 Reject unsigned modules or signed modules for which we don't have a 2029 key. Without this, such modules will simply taint the kernel. 2030 2031config MODULE_SIG_ALL 2032 bool "Automatically sign all modules" 2033 default y 2034 depends on MODULE_SIG 2035 help 2036 Sign all modules during make modules_install. Without this option, 2037 modules must be signed manually, using the scripts/sign-file tool. 2038 2039comment "Do not forget to sign required modules with scripts/sign-file" 2040 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 2041 2042choice 2043 prompt "Which hash algorithm should modules be signed with?" 2044 depends on MODULE_SIG 2045 help 2046 This determines which sort of hashing algorithm will be used during 2047 signature generation. This algorithm _must_ be built into the kernel 2048 directly so that signature verification can take place. It is not 2049 possible to load a signed module containing the algorithm to check 2050 the signature on that module. 2051 2052config MODULE_SIG_SHA1 2053 bool "Sign modules with SHA-1" 2054 select CRYPTO_SHA1 2055 2056config MODULE_SIG_SHA224 2057 bool "Sign modules with SHA-224" 2058 select CRYPTO_SHA256 2059 2060config MODULE_SIG_SHA256 2061 bool "Sign modules with SHA-256" 2062 select CRYPTO_SHA256 2063 2064config MODULE_SIG_SHA384 2065 bool "Sign modules with SHA-384" 2066 select CRYPTO_SHA512 2067 2068config MODULE_SIG_SHA512 2069 bool "Sign modules with SHA-512" 2070 select CRYPTO_SHA512 2071 2072endchoice 2073 2074config MODULE_SIG_HASH 2075 string 2076 depends on MODULE_SIG 2077 default "sha1" if MODULE_SIG_SHA1 2078 default "sha224" if MODULE_SIG_SHA224 2079 default "sha256" if MODULE_SIG_SHA256 2080 default "sha384" if MODULE_SIG_SHA384 2081 default "sha512" if MODULE_SIG_SHA512 2082 2083config MODULE_COMPRESS 2084 bool "Compress modules on installation" 2085 depends on MODULES 2086 help 2087 2088 Compresses kernel modules when 'make modules_install' is run; gzip or 2089 xz depending on "Compression algorithm" below. 2090 2091 module-init-tools MAY support gzip, and kmod MAY support gzip and xz. 2092 2093 Out-of-tree kernel modules installed using Kbuild will also be 2094 compressed upon installation. 2095 2096 Note: for modules inside an initrd or initramfs, it's more efficient 2097 to compress the whole initrd or initramfs instead. 2098 2099 Note: This is fully compatible with signed modules. 2100 2101 If in doubt, say N. 2102 2103choice 2104 prompt "Compression algorithm" 2105 depends on MODULE_COMPRESS 2106 default MODULE_COMPRESS_GZIP 2107 help 2108 This determines which sort of compression will be used during 2109 'make modules_install'. 2110 2111 GZIP (default) and XZ are supported. 2112 2113config MODULE_COMPRESS_GZIP 2114 bool "GZIP" 2115 2116config MODULE_COMPRESS_XZ 2117 bool "XZ" 2118 2119endchoice 2120 2121config TRIM_UNUSED_KSYMS 2122 bool "Trim unused exported kernel symbols" 2123 depends on MODULES && !UNUSED_SYMBOLS 2124 help 2125 The kernel and some modules make many symbols available for 2126 other modules to use via EXPORT_SYMBOL() and variants. Depending 2127 on the set of modules being selected in your kernel configuration, 2128 many of those exported symbols might never be used. 2129 2130 This option allows for unused exported symbols to be dropped from 2131 the build. In turn, this provides the compiler more opportunities 2132 (especially when using LTO) for optimizing the code and reducing 2133 binary size. This might have some security advantages as well. 2134 2135 If unsure, or if you need to build out-of-tree modules, say N. 2136 2137endif # MODULES 2138 2139config MODULES_TREE_LOOKUP 2140 def_bool y 2141 depends on PERF_EVENTS || TRACING 2142 2143config INIT_ALL_POSSIBLE 2144 bool 2145 help 2146 Back when each arch used to define their own cpu_online_mask and 2147 cpu_possible_mask, some of them chose to initialize cpu_possible_mask 2148 with all 1s, and others with all 0s. When they were centralised, 2149 it was better to provide this option than to break all the archs 2150 and have several arch maintainers pursuing me down dark alleys. 2151 2152source "block/Kconfig" 2153 2154config PREEMPT_NOTIFIERS 2155 bool 2156 2157config PADATA 2158 depends on SMP 2159 bool 2160 2161config ASN1 2162 tristate 2163 help 2164 Build a simple ASN.1 grammar compiler that produces a bytecode output 2165 that can be interpreted by the ASN.1 stream decoder and used to 2166 inform it as to what tags are to be expected in a stream and what 2167 functions to call on what tags. 2168 2169source "kernel/Kconfig.locks" 2170 2171config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE 2172 bool 2173 2174# It may be useful for an architecture to override the definitions of the 2175# SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h> 2176# and the COMPAT_ variants in <linux/compat.h>, in particular to use a 2177# different calling convention for syscalls. They can also override the 2178# macros for not-implemented syscalls in kernel/sys_ni.c and 2179# kernel/time/posix-stubs.c. All these overrides need to be available in 2180# <asm/syscall_wrapper.h>. 2181config ARCH_HAS_SYSCALL_WRAPPER 2182 def_bool n 2183