1ec8f24b7SThomas Gleixner# SPDX-License-Identifier: GPL-2.0-only 2face4374SRoman Zippelconfig DEFCONFIG_LIST 3face4374SRoman Zippel string 4b2670eacSPaolo 'Blaisorblade' Giarrusso depends on !UML 5face4374SRoman Zippel option defconfig_list 647f38ae0SRob Landley default "/lib/modules/$(shell,uname -r)/.config" 7face4374SRoman Zippel default "/etc/kernel-config" 847f38ae0SRob Landley default "/boot/config-$(shell,uname -r)" 92a86f661SMasahiro Yamada default "arch/$(SRCARCH)/configs/$(KBUILD_DEFCONFIG)" 10face4374SRoman Zippel 11a4353898SMasahiro Yamadaconfig CC_IS_GCC 12a4353898SMasahiro Yamada def_bool $(success,$(CC) --version | head -n 1 | grep -q gcc) 13a4353898SMasahiro Yamada 14a4353898SMasahiro Yamadaconfig GCC_VERSION 15a4353898SMasahiro Yamada int 16fa7295abSMasahiro Yamada default $(shell,$(srctree)/scripts/gcc-version.sh $(CC)) if CC_IS_GCC 17a4353898SMasahiro Yamada default 0 18a4353898SMasahiro Yamada 199553d16fSAmit Daniel Kachhapconfig LD_VERSION 209553d16fSAmit Daniel Kachhap int 219553d16fSAmit Daniel Kachhap default $(shell,$(LD) --version | $(srctree)/scripts/ld-version.sh) 229553d16fSAmit Daniel Kachhap 23469cb737SMasahiro Yamadaconfig CC_IS_CLANG 24469cb737SMasahiro Yamada def_bool $(success,$(CC) --version | head -n 1 | grep -q clang) 25469cb737SMasahiro Yamada 26*b744b43fSSami Tolvanenconfig LD_IS_LLD 27*b744b43fSSami Tolvanen def_bool $(success,$(LD) -v | head -n 1 | grep -q LLD) 28*b744b43fSSami Tolvanen 29469cb737SMasahiro Yamadaconfig CLANG_VERSION 30469cb737SMasahiro Yamada int 31469cb737SMasahiro Yamada default $(shell,$(srctree)/scripts/clang-version.sh $(CC)) 32469cb737SMasahiro Yamada 331a927fd3SMasahiro Yamadaconfig CC_CAN_LINK 341a927fd3SMasahiro Yamada def_bool $(success,$(srctree)/scripts/cc-can-link.sh $(CC)) 351a927fd3SMasahiro Yamada 36e9666d10SMasahiro Yamadaconfig CC_HAS_ASM_GOTO 37e9666d10SMasahiro Yamada def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC)) 38e9666d10SMasahiro Yamada 395cf896fbSPeter Collingbourneconfig TOOLS_SUPPORT_RELR 402d122942SWill Deacon def_bool $(success,env "CC=$(CC)" "LD=$(LD)" "NM=$(NM)" "OBJCOPY=$(OBJCOPY)" $(srctree)/scripts/tools-support-relr.sh) 415cf896fbSPeter Collingbourne 42eb111869SRasmus Villemoesconfig CC_HAS_ASM_INLINE 43eb111869SRasmus Villemoes def_bool $(success,echo 'void foo(void) { asm inline (""); }' | $(CC) -x c - -c -o /dev/null) 44eb111869SRasmus Villemoes 45b303c6dfSMasahiro Yamadaconfig CC_HAS_WARN_MAYBE_UNINITIALIZED 46b303c6dfSMasahiro Yamada def_bool $(cc-option,-Wmaybe-uninitialized) 47b303c6dfSMasahiro Yamada help 48b303c6dfSMasahiro Yamada GCC >= 4.7 supports this option. 49b303c6dfSMasahiro Yamada 50b303c6dfSMasahiro Yamadaconfig CC_DISABLE_WARN_MAYBE_UNINITIALIZED 51b303c6dfSMasahiro Yamada bool 52b303c6dfSMasahiro Yamada depends on CC_HAS_WARN_MAYBE_UNINITIALIZED 53b303c6dfSMasahiro Yamada default CC_IS_GCC && GCC_VERSION < 40900 # unreliable for GCC < 4.9 54b303c6dfSMasahiro Yamada help 55b303c6dfSMasahiro Yamada GCC's -Wmaybe-uninitialized is not reliable by definition. 56b303c6dfSMasahiro Yamada Lots of false positive warnings are produced in some cases. 57b303c6dfSMasahiro Yamada 58b303c6dfSMasahiro Yamada If this option is enabled, -Wno-maybe-uninitialzed is passed 59b303c6dfSMasahiro Yamada to the compiler to suppress maybe-uninitialized warnings. 60b303c6dfSMasahiro Yamada 61b99b87f7SPeter Oberparleiterconfig CONSTRUCTORS 62b99b87f7SPeter Oberparleiter bool 6387c9366eSJohannes Berg depends on !UML 64b99b87f7SPeter Oberparleiter 65e360adbeSPeter Zijlstraconfig IRQ_WORK 66e360adbeSPeter Zijlstra bool 67e360adbeSPeter Zijlstra 6810916706SShile Zhangconfig BUILDTIME_TABLE_SORT 691dbdc6f1SDavid Daney bool 701dbdc6f1SDavid Daney 71c65eacbeSAndy Lutomirskiconfig THREAD_INFO_IN_TASK 72c65eacbeSAndy Lutomirski bool 73c65eacbeSAndy Lutomirski help 74c65eacbeSAndy Lutomirski Select this to move thread_info off the stack into task_struct. To 75c65eacbeSAndy Lutomirski make this work, an arch will need to remove all thread_info fields 76c65eacbeSAndy Lutomirski except flags and fix any runtime bugs. 77c65eacbeSAndy Lutomirski 78c6c314a6SAndy Lutomirski One subtle change that will be needed is to use try_get_task_stack() 79c6c314a6SAndy Lutomirski and put_task_stack() in save_thread_stack_tsk() and get_wchan(). 80c6c314a6SAndy Lutomirski 81ff0cfc66SAl Boldimenu "General setup" 821da177e4SLinus Torvalds 831da177e4SLinus Torvaldsconfig BROKEN 841da177e4SLinus Torvalds bool 851da177e4SLinus Torvalds 861da177e4SLinus Torvaldsconfig BROKEN_ON_SMP 871da177e4SLinus Torvalds bool 881da177e4SLinus Torvalds depends on BROKEN || !SMP 891da177e4SLinus Torvalds default y 901da177e4SLinus Torvalds 911da177e4SLinus Torvaldsconfig INIT_ENV_ARG_LIMIT 921da177e4SLinus Torvalds int 93dd673bcaSAdrian Bunk default 32 if !UML 94dd673bcaSAdrian Bunk default 128 if UML 951da177e4SLinus Torvalds help 9634ad92c2SRandy Dunlap Maximum of each of the number of arguments and environment 9734ad92c2SRandy Dunlap variables passed to init from the kernel command line. 981da177e4SLinus Torvalds 994bb16672SJiri Slabyconfig COMPILE_TEST 1004bb16672SJiri Slaby bool "Compile also drivers which will not load" 101bc083a64SRichard Weinberger depends on !UML 1024bb16672SJiri Slaby default n 1034bb16672SJiri Slaby help 1044bb16672SJiri Slaby Some drivers can be compiled on a different platform than they are 1054bb16672SJiri Slaby intended to be run on. Despite they cannot be loaded there (or even 1064bb16672SJiri Slaby when they load they cannot be used due to missing HW support), 1074bb16672SJiri Slaby developers still, opposing to distributors, might want to build such 1084bb16672SJiri Slaby drivers to compile-test them. 1094bb16672SJiri Slaby 1104bb16672SJiri Slaby If you are a developer and want to build everything available, say Y 1114bb16672SJiri Slaby here. If you are a user/distributor, say N here to exclude useless 1124bb16672SJiri Slaby drivers to be distributed. 1134bb16672SJiri Slaby 114d6fc9fcbSMasahiro Yamadaconfig UAPI_HEADER_TEST 115d6fc9fcbSMasahiro Yamada bool "Compile test UAPI headers" 116fcbb8461SMasahiro Yamada depends on HEADERS_INSTALL && CC_CAN_LINK 117d6fc9fcbSMasahiro Yamada help 118d6fc9fcbSMasahiro Yamada Compile test headers exported to user-space to ensure they are 119d6fc9fcbSMasahiro Yamada self-contained, i.e. compilable as standalone units. 120d6fc9fcbSMasahiro Yamada 121d6fc9fcbSMasahiro Yamada If you are a developer or tester and want to ensure the exported 122d6fc9fcbSMasahiro Yamada headers are self-contained, say Y here. Otherwise, choose N. 123d6fc9fcbSMasahiro Yamada 1241da177e4SLinus Torvaldsconfig LOCALVERSION 1251da177e4SLinus Torvalds string "Local version - append to kernel release" 1261da177e4SLinus Torvalds help 1271da177e4SLinus Torvalds Append an extra string to the end of your kernel version. 1281da177e4SLinus Torvalds This will show up when you type uname, for example. 1291da177e4SLinus Torvalds The string you set here will be appended after the contents of 1301da177e4SLinus Torvalds any files with a filename matching localversion* in your 1311da177e4SLinus Torvalds object and source tree, in that order. Your total string can 1321da177e4SLinus Torvalds be a maximum of 64 characters. 1331da177e4SLinus Torvalds 134aaebf433SRyan Andersonconfig LOCALVERSION_AUTO 135aaebf433SRyan Anderson bool "Automatically append version information to the version string" 136aaebf433SRyan Anderson default y 137ac3339baSAlexey Dobriyan depends on !COMPILE_TEST 138aaebf433SRyan Anderson help 139aaebf433SRyan Anderson This will try to automatically determine if the current tree is a 1406e5a5420SRobert P. J. Day release tree by looking for git tags that belong to the current 1416e5a5420SRobert P. J. Day top of tree revision. 142aaebf433SRyan Anderson 143aaebf433SRyan Anderson A string of the format -gxxxxxxxx will be added to the localversion 1446e5a5420SRobert P. J. Day if a git-based tree is found. The string generated by this will be 145aaebf433SRyan Anderson appended after any matching localversion* files, and after the value 1466e5a5420SRobert P. J. Day set in CONFIG_LOCALVERSION. 147aaebf433SRyan Anderson 1486e5a5420SRobert P. J. Day (The actual string used here is the first eight characters produced 1496e5a5420SRobert P. J. Day by running the command: 1506e5a5420SRobert P. J. Day 1516e5a5420SRobert P. J. Day $ git rev-parse --verify HEAD 1526e5a5420SRobert P. J. Day 1536e5a5420SRobert P. J. Day which is done within the script "scripts/setlocalversion".) 154aaebf433SRyan Anderson 1559afb719eSLaura Abbottconfig BUILD_SALT 1569afb719eSLaura Abbott string "Build ID Salt" 1579afb719eSLaura Abbott default "" 1589afb719eSLaura Abbott help 1599afb719eSLaura Abbott The build ID is used to link binaries and their debug info. Setting 1609afb719eSLaura Abbott this option will use the value in the calculation of the build id. 1619afb719eSLaura Abbott This is mostly useful for distributions which want to ensure the 1629afb719eSLaura Abbott build is unique between builds. It's safe to leave the default. 1639afb719eSLaura Abbott 1642e9f3bddSH. Peter Anvinconfig HAVE_KERNEL_GZIP 1652e9f3bddSH. Peter Anvin bool 1662e9f3bddSH. Peter Anvin 1672e9f3bddSH. Peter Anvinconfig HAVE_KERNEL_BZIP2 1682e9f3bddSH. Peter Anvin bool 1692e9f3bddSH. Peter Anvin 1702e9f3bddSH. Peter Anvinconfig HAVE_KERNEL_LZMA 1712e9f3bddSH. Peter Anvin bool 1722e9f3bddSH. Peter Anvin 1733ebe1243SLasse Collinconfig HAVE_KERNEL_XZ 1743ebe1243SLasse Collin bool 1753ebe1243SLasse Collin 1767dd65febSAlbin Tonnerreconfig HAVE_KERNEL_LZO 1777dd65febSAlbin Tonnerre bool 1787dd65febSAlbin Tonnerre 179e76e1fdfSKyungsik Leeconfig HAVE_KERNEL_LZ4 180e76e1fdfSKyungsik Lee bool 181e76e1fdfSKyungsik Lee 182f16466afSVasily Gorbikconfig HAVE_KERNEL_UNCOMPRESSED 183f16466afSVasily Gorbik bool 184f16466afSVasily Gorbik 18530d65dbfSAlain Knaffchoice 18630d65dbfSAlain Knaff prompt "Kernel compression mode" 18730d65dbfSAlain Knaff default KERNEL_GZIP 188f16466afSVasily Gorbik depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED 18930d65dbfSAlain Knaff help 19030d65dbfSAlain Knaff The linux kernel is a kind of self-extracting executable. 19130d65dbfSAlain Knaff Several compression algorithms are available, which differ 19230d65dbfSAlain Knaff in efficiency, compression and decompression speed. 19330d65dbfSAlain Knaff Compression speed is only relevant when building a kernel. 19430d65dbfSAlain Knaff Decompression speed is relevant at each boot. 19530d65dbfSAlain Knaff 19630d65dbfSAlain Knaff If you have any problems with bzip2 or lzma compressed 19730d65dbfSAlain Knaff kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older 19830d65dbfSAlain Knaff version of this functionality (bzip2 only), for 2.4, was 19930d65dbfSAlain Knaff supplied by Christian Ludwig) 20030d65dbfSAlain Knaff 20130d65dbfSAlain Knaff High compression options are mostly useful for users, who 20230d65dbfSAlain Knaff are low on disk space (embedded systems), but for whom ram 20330d65dbfSAlain Knaff size matters less. 20430d65dbfSAlain Knaff 20530d65dbfSAlain Knaff If in doubt, select 'gzip' 20630d65dbfSAlain Knaff 20730d65dbfSAlain Knaffconfig KERNEL_GZIP 20830d65dbfSAlain Knaff bool "Gzip" 2092e9f3bddSH. Peter Anvin depends on HAVE_KERNEL_GZIP 21030d65dbfSAlain Knaff help 2117dd65febSAlbin Tonnerre The old and tried gzip compression. It provides a good balance 2127dd65febSAlbin Tonnerre between compression ratio and decompression speed. 21330d65dbfSAlain Knaff 21430d65dbfSAlain Knaffconfig KERNEL_BZIP2 21530d65dbfSAlain Knaff bool "Bzip2" 2162e9f3bddSH. Peter Anvin depends on HAVE_KERNEL_BZIP2 21730d65dbfSAlain Knaff help 21830d65dbfSAlain Knaff Its compression ratio and speed is intermediate. 2190a4dd35cSRandy Dunlap Decompression speed is slowest among the choices. The kernel 2202e9f3bddSH. Peter Anvin size is about 10% smaller with bzip2, in comparison to gzip. 2212e9f3bddSH. Peter Anvin Bzip2 uses a large amount of memory. For modern kernels you 2222e9f3bddSH. Peter Anvin will need at least 8MB RAM or more for booting. 22330d65dbfSAlain Knaff 22430d65dbfSAlain Knaffconfig KERNEL_LZMA 22530d65dbfSAlain Knaff bool "LZMA" 2262e9f3bddSH. Peter Anvin depends on HAVE_KERNEL_LZMA 22730d65dbfSAlain Knaff help 2280a4dd35cSRandy Dunlap This compression algorithm's ratio is best. Decompression speed 2290a4dd35cSRandy Dunlap is between gzip and bzip2. Compression is slowest. 2300a4dd35cSRandy Dunlap The kernel size is about 33% smaller with LZMA in comparison to gzip. 23130d65dbfSAlain Knaff 2323ebe1243SLasse Collinconfig KERNEL_XZ 2333ebe1243SLasse Collin bool "XZ" 2343ebe1243SLasse Collin depends on HAVE_KERNEL_XZ 2353ebe1243SLasse Collin help 2363ebe1243SLasse Collin XZ uses the LZMA2 algorithm and instruction set specific 2373ebe1243SLasse Collin BCJ filters which can improve compression ratio of executable 2383ebe1243SLasse Collin code. The size of the kernel is about 30% smaller with XZ in 2393ebe1243SLasse Collin comparison to gzip. On architectures for which there is a BCJ 2403ebe1243SLasse Collin filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ 2413ebe1243SLasse Collin will create a few percent smaller kernel than plain LZMA. 2423ebe1243SLasse Collin 2433ebe1243SLasse Collin The speed is about the same as with LZMA: The decompression 2443ebe1243SLasse Collin speed of XZ is better than that of bzip2 but worse than gzip 2453ebe1243SLasse Collin and LZO. Compression is slow. 2463ebe1243SLasse Collin 2477dd65febSAlbin Tonnerreconfig KERNEL_LZO 2487dd65febSAlbin Tonnerre bool "LZO" 2497dd65febSAlbin Tonnerre depends on HAVE_KERNEL_LZO 2507dd65febSAlbin Tonnerre help 2510a4dd35cSRandy Dunlap Its compression ratio is the poorest among the choices. The kernel 252681b3049SStephan Sperber size is about 10% bigger than gzip; however its speed 2537dd65febSAlbin Tonnerre (both compression and decompression) is the fastest. 2547dd65febSAlbin Tonnerre 255e76e1fdfSKyungsik Leeconfig KERNEL_LZ4 256e76e1fdfSKyungsik Lee bool "LZ4" 257e76e1fdfSKyungsik Lee depends on HAVE_KERNEL_LZ4 258e76e1fdfSKyungsik Lee help 259e76e1fdfSKyungsik Lee LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding. 260e76e1fdfSKyungsik Lee A preliminary version of LZ4 de/compression tool is available at 261e76e1fdfSKyungsik Lee <https://code.google.com/p/lz4/>. 262e76e1fdfSKyungsik Lee 263e76e1fdfSKyungsik Lee Its compression ratio is worse than LZO. The size of the kernel 264e76e1fdfSKyungsik Lee is about 8% bigger than LZO. But the decompression speed is 265e76e1fdfSKyungsik Lee faster than LZO. 266e76e1fdfSKyungsik Lee 267f16466afSVasily Gorbikconfig KERNEL_UNCOMPRESSED 268f16466afSVasily Gorbik bool "None" 269f16466afSVasily Gorbik depends on HAVE_KERNEL_UNCOMPRESSED 270f16466afSVasily Gorbik help 271f16466afSVasily Gorbik Produce uncompressed kernel image. This option is usually not what 272f16466afSVasily Gorbik you want. It is useful for debugging the kernel in slow simulation 273f16466afSVasily Gorbik environments, where decompressing and moving the kernel is awfully 274f16466afSVasily Gorbik slow. This option allows early boot code to skip the decompressor 275f16466afSVasily Gorbik and jump right at uncompressed kernel image. 276f16466afSVasily Gorbik 27730d65dbfSAlain Knaffendchoice 27830d65dbfSAlain Knaff 279bd5dc17bSJosh Triplettconfig DEFAULT_HOSTNAME 280bd5dc17bSJosh Triplett string "Default hostname" 281bd5dc17bSJosh Triplett default "(none)" 282bd5dc17bSJosh Triplett help 283bd5dc17bSJosh Triplett This option determines the default system hostname before userspace 284bd5dc17bSJosh Triplett calls sethostname(2). The kernel traditionally uses "(none)" here, 285bd5dc17bSJosh Triplett but you may wish to use a different default here to make a minimal 286bd5dc17bSJosh Triplett system more usable with less configuration. 287bd5dc17bSJosh Triplett 28817c46a6aSChristoph Hellwig# 28917c46a6aSChristoph Hellwig# For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can 29017c46a6aSChristoph Hellwig# add proper SWAP support to them, in which case this can be remove. 29117c46a6aSChristoph Hellwig# 29217c46a6aSChristoph Hellwigconfig ARCH_NO_SWAP 29317c46a6aSChristoph Hellwig bool 29417c46a6aSChristoph Hellwig 2951da177e4SLinus Torvaldsconfig SWAP 2961da177e4SLinus Torvalds bool "Support for paging of anonymous memory (swap)" 29717c46a6aSChristoph Hellwig depends on MMU && BLOCK && !ARCH_NO_SWAP 2981da177e4SLinus Torvalds default y 2991da177e4SLinus Torvalds help 3001da177e4SLinus Torvalds This option allows you to choose whether you want to have support 3011da177e4SLinus Torvalds for so called swap devices or swap files in your kernel that are 3021da177e4SLinus Torvalds used to provide more virtual memory than the actual RAM present 3031da177e4SLinus Torvalds in your computer. If unsure say Y. 3041da177e4SLinus Torvalds 3051da177e4SLinus Torvaldsconfig SYSVIPC 3061da177e4SLinus Torvalds bool "System V IPC" 3071da177e4SLinus Torvalds ---help--- 3081da177e4SLinus Torvalds Inter Process Communication is a suite of library functions and 3091da177e4SLinus Torvalds system calls which let processes (running programs) synchronize and 3101da177e4SLinus Torvalds exchange information. It is generally considered to be a good thing, 3111da177e4SLinus Torvalds and some programs won't run unless you say Y here. In particular, if 3121da177e4SLinus Torvalds you want to run the DOS emulator dosemu under Linux (read the 3131da177e4SLinus Torvalds DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>), 3141da177e4SLinus Torvalds you'll need to say Y here. 3151da177e4SLinus Torvalds 3161da177e4SLinus Torvalds You can find documentation about IPC with "info ipc" and also in 3171da177e4SLinus Torvalds section 6.4 of the Linux Programmer's Guide, available from 3181da177e4SLinus Torvalds <http://www.tldp.org/guides.html>. 3191da177e4SLinus Torvalds 320a5494dcdSEric W. Biedermanconfig SYSVIPC_SYSCTL 321a5494dcdSEric W. Biederman bool 322a5494dcdSEric W. Biederman depends on SYSVIPC 323a5494dcdSEric W. Biederman depends on SYSCTL 324a5494dcdSEric W. Biederman default y 325a5494dcdSEric W. Biederman 3261da177e4SLinus Torvaldsconfig POSIX_MQUEUE 3271da177e4SLinus Torvalds bool "POSIX Message Queues" 32819c92399SKees Cook depends on NET 3291da177e4SLinus Torvalds ---help--- 3301da177e4SLinus Torvalds POSIX variant of message queues is a part of IPC. In POSIX message 3311da177e4SLinus Torvalds queues every message has a priority which decides about succession 3321da177e4SLinus Torvalds of receiving it by a process. If you want to compile and run 3331da177e4SLinus Torvalds programs written e.g. for Solaris with use of its POSIX message 334b0e37650SRobert P. J. Day queues (functions mq_*) say Y here. 3351da177e4SLinus Torvalds 3361da177e4SLinus Torvalds POSIX message queues are visible as a filesystem called 'mqueue' 3371da177e4SLinus Torvalds and can be mounted somewhere if you want to do filesystem 3381da177e4SLinus Torvalds operations on message queues. 3391da177e4SLinus Torvalds 3401da177e4SLinus Torvalds If unsure, say Y. 3411da177e4SLinus Torvalds 342bdc8e5f8SSerge E. Hallynconfig POSIX_MQUEUE_SYSCTL 343bdc8e5f8SSerge E. Hallyn bool 344bdc8e5f8SSerge E. Hallyn depends on POSIX_MQUEUE 345bdc8e5f8SSerge E. Hallyn depends on SYSCTL 346bdc8e5f8SSerge E. Hallyn default y 347bdc8e5f8SSerge E. Hallyn 348226b4ccdSKonstantin Khlebnikovconfig CROSS_MEMORY_ATTACH 349226b4ccdSKonstantin Khlebnikov bool "Enable process_vm_readv/writev syscalls" 350226b4ccdSKonstantin Khlebnikov depends on MMU 351226b4ccdSKonstantin Khlebnikov default y 352226b4ccdSKonstantin Khlebnikov help 353226b4ccdSKonstantin Khlebnikov Enabling this option adds the system calls process_vm_readv and 354226b4ccdSKonstantin Khlebnikov process_vm_writev which allow a process with the correct privileges 355a2a368d9SGeert Uytterhoeven to directly read from or write to another process' address space. 356226b4ccdSKonstantin Khlebnikov See the man page for more details. 357226b4ccdSKonstantin Khlebnikov 35869369a70SJosh Triplettconfig USELIB 35969369a70SJosh Triplett bool "uselib syscall" 360b2113a41SRiku Voipio def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION 36169369a70SJosh Triplett help 36269369a70SJosh Triplett This option enables the uselib syscall, a system call used in the 36369369a70SJosh Triplett dynamic linker from libc5 and earlier. glibc does not use this 36469369a70SJosh Triplett system call. If you intend to run programs built on libc5 or 36569369a70SJosh Triplett earlier, you may need to enable this syscall. Current systems 36669369a70SJosh Triplett running glibc can safely disable this. 36769369a70SJosh Triplett 3681da177e4SLinus Torvaldsconfig AUDIT 3691da177e4SLinus Torvalds bool "Auditing support" 370804a6a49SChris Wright depends on NET 3711da177e4SLinus Torvalds help 3721da177e4SLinus Torvalds Enable auditing infrastructure that can be used with another 3731da177e4SLinus Torvalds kernel subsystem, such as SELinux (which requires this for 374cb74ed27SPaul Moore logging of avc messages output). System call auditing is included 375cb74ed27SPaul Moore on architectures which support it. 3761da177e4SLinus Torvalds 3777a017721SAKASHI Takahiroconfig HAVE_ARCH_AUDITSYSCALL 3787a017721SAKASHI Takahiro bool 3797a017721SAKASHI Takahiro 3801da177e4SLinus Torvaldsconfig AUDITSYSCALL 381cb74ed27SPaul Moore def_bool y 3827a017721SAKASHI Takahiro depends on AUDIT && HAVE_ARCH_AUDITSYSCALL 38328a3a7ebSEric Paris select FSNOTIFY 38474c3cbe3SAl Viro 385d9817ebeSThomas Gleixnersource "kernel/irq/Kconfig" 386764e0da1SThomas Gleixnersource "kernel/time/Kconfig" 38787a4c375SChristoph Hellwigsource "kernel/Kconfig.preempt" 388d9817ebeSThomas Gleixner 389391dc69cSFrederic Weisbeckermenu "CPU/Task time and stats accounting" 390391dc69cSFrederic Weisbecker 391abf917cdSFrederic Weisbeckerconfig VIRT_CPU_ACCOUNTING 392abf917cdSFrederic Weisbecker bool 393abf917cdSFrederic Weisbecker 394fdf9c356SFrederic Weisbeckerchoice 395fdf9c356SFrederic Weisbecker prompt "Cputime accounting" 396fdf9c356SFrederic Weisbecker default TICK_CPU_ACCOUNTING if !PPC64 39702fc8d37SStephen Rothwell default VIRT_CPU_ACCOUNTING_NATIVE if PPC64 398fdf9c356SFrederic Weisbecker 399fdf9c356SFrederic Weisbecker# Kind of a stub config for the pure tick based cputime accounting 400fdf9c356SFrederic Weisbeckerconfig TICK_CPU_ACCOUNTING 401fdf9c356SFrederic Weisbecker bool "Simple tick based cputime accounting" 402c58b0df1SFrederic Weisbecker depends on !S390 && !NO_HZ_FULL 403fdf9c356SFrederic Weisbecker help 404fdf9c356SFrederic Weisbecker This is the basic tick based cputime accounting that maintains 405fdf9c356SFrederic Weisbecker statistics about user, system and idle time spent on per jiffies 406fdf9c356SFrederic Weisbecker granularity. 407fdf9c356SFrederic Weisbecker 408fdf9c356SFrederic Weisbecker If unsure, say Y. 409fdf9c356SFrederic Weisbecker 410abf917cdSFrederic Weisbeckerconfig VIRT_CPU_ACCOUNTING_NATIVE 411391dc69cSFrederic Weisbecker bool "Deterministic task and CPU time accounting" 412c58b0df1SFrederic Weisbecker depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL 413abf917cdSFrederic Weisbecker select VIRT_CPU_ACCOUNTING 414391dc69cSFrederic Weisbecker help 415391dc69cSFrederic Weisbecker Select this option to enable more accurate task and CPU time 416391dc69cSFrederic Weisbecker accounting. This is done by reading a CPU counter on each 417391dc69cSFrederic Weisbecker kernel entry and exit and on transitions within the kernel 418391dc69cSFrederic Weisbecker between system, softirq and hardirq state, so there is a 419391dc69cSFrederic Weisbecker small performance impact. In the case of s390 or IBM POWER > 5, 420391dc69cSFrederic Weisbecker this also enables accounting of stolen time on logically-partitioned 421391dc69cSFrederic Weisbecker systems. 422391dc69cSFrederic Weisbecker 423abf917cdSFrederic Weisbeckerconfig VIRT_CPU_ACCOUNTING_GEN 424abf917cdSFrederic Weisbecker bool "Full dynticks CPU time accounting" 425ff3fb254SKevin Hilman depends on HAVE_CONTEXT_TRACKING 426554b0004SKevin Hilman depends on HAVE_VIRT_CPU_ACCOUNTING_GEN 427041a1574SArnd Bergmann depends on GENERIC_CLOCKEVENTS 428abf917cdSFrederic Weisbecker select VIRT_CPU_ACCOUNTING 429abf917cdSFrederic Weisbecker select CONTEXT_TRACKING 430abf917cdSFrederic Weisbecker help 431abf917cdSFrederic Weisbecker Select this option to enable task and CPU time accounting on full 432abf917cdSFrederic Weisbecker dynticks systems. This accounting is implemented by watching every 433abf917cdSFrederic Weisbecker kernel-user boundaries using the context tracking subsystem. 434abf917cdSFrederic Weisbecker The accounting is thus performed at the expense of some significant 435abf917cdSFrederic Weisbecker overhead. 436abf917cdSFrederic Weisbecker 437abf917cdSFrederic Weisbecker For now this is only useful if you are working on the full 438abf917cdSFrederic Weisbecker dynticks subsystem development. 439abf917cdSFrederic Weisbecker 440abf917cdSFrederic Weisbecker If unsure, say N. 441abf917cdSFrederic Weisbecker 442b58c3584SRik van Rielendchoice 443b58c3584SRik van Riel 444fdf9c356SFrederic Weisbeckerconfig IRQ_TIME_ACCOUNTING 445fdf9c356SFrederic Weisbecker bool "Fine granularity task level IRQ time accounting" 446b58c3584SRik van Riel depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE 447fdf9c356SFrederic Weisbecker help 448fdf9c356SFrederic Weisbecker Select this option to enable fine granularity task irq time 449fdf9c356SFrederic Weisbecker accounting. This is done by reading a timestamp on each 450fdf9c356SFrederic Weisbecker transitions between softirq and hardirq state, so there can be a 451fdf9c356SFrederic Weisbecker small performance impact. 452fdf9c356SFrederic Weisbecker 453fdf9c356SFrederic Weisbecker If in doubt, say N here. 454fdf9c356SFrederic Weisbecker 45511d4afd4SVincent Guittotconfig HAVE_SCHED_AVG_IRQ 45611d4afd4SVincent Guittot def_bool y 45711d4afd4SVincent Guittot depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING 45811d4afd4SVincent Guittot depends on SMP 45911d4afd4SVincent Guittot 46076504793SThara Gopinathconfig SCHED_THERMAL_PRESSURE 46176504793SThara Gopinath bool "Enable periodic averaging of thermal pressure" 46276504793SThara Gopinath depends on SMP 46376504793SThara Gopinath 464391dc69cSFrederic Weisbeckerconfig BSD_PROCESS_ACCT 465391dc69cSFrederic Weisbecker bool "BSD Process Accounting" 4662813893fSIulia Manda depends on MULTIUSER 467391dc69cSFrederic Weisbecker help 468391dc69cSFrederic Weisbecker If you say Y here, a user level program will be able to instruct the 469391dc69cSFrederic Weisbecker kernel (via a special system call) to write process accounting 470391dc69cSFrederic Weisbecker information to a file: whenever a process exits, information about 471391dc69cSFrederic Weisbecker that process will be appended to the file by the kernel. The 472391dc69cSFrederic Weisbecker information includes things such as creation time, owning user, 473391dc69cSFrederic Weisbecker command name, memory usage, controlling terminal etc. (the complete 474391dc69cSFrederic Weisbecker list is in the struct acct in <file:include/linux/acct.h>). It is 475391dc69cSFrederic Weisbecker up to the user level program to do useful things with this 476391dc69cSFrederic Weisbecker information. This is generally a good idea, so say Y. 477391dc69cSFrederic Weisbecker 478391dc69cSFrederic Weisbeckerconfig BSD_PROCESS_ACCT_V3 479391dc69cSFrederic Weisbecker bool "BSD Process Accounting version 3 file format" 480391dc69cSFrederic Weisbecker depends on BSD_PROCESS_ACCT 481391dc69cSFrederic Weisbecker default n 482391dc69cSFrederic Weisbecker help 483391dc69cSFrederic Weisbecker If you say Y here, the process accounting information is written 484391dc69cSFrederic Weisbecker in a new file format that also logs the process IDs of each 4853903bf94SRandy Dunlap process and its parent. Note that this file format is incompatible 486391dc69cSFrederic Weisbecker with previous v0/v1/v2 file formats, so you will need updated tools 487391dc69cSFrederic Weisbecker for processing it. A preliminary version of these tools is available 488391dc69cSFrederic Weisbecker at <http://www.gnu.org/software/acct/>. 489391dc69cSFrederic Weisbecker 490391dc69cSFrederic Weisbeckerconfig TASKSTATS 49119c92399SKees Cook bool "Export task/process statistics through netlink" 492391dc69cSFrederic Weisbecker depends on NET 4932813893fSIulia Manda depends on MULTIUSER 494391dc69cSFrederic Weisbecker default n 495391dc69cSFrederic Weisbecker help 496391dc69cSFrederic Weisbecker Export selected statistics for tasks/processes through the 497391dc69cSFrederic Weisbecker generic netlink interface. Unlike BSD process accounting, the 498391dc69cSFrederic Weisbecker statistics are available during the lifetime of tasks/processes as 499391dc69cSFrederic Weisbecker responses to commands. Like BSD accounting, they are sent to user 500391dc69cSFrederic Weisbecker space on task exit. 501391dc69cSFrederic Weisbecker 502391dc69cSFrederic Weisbecker Say N if unsure. 503391dc69cSFrederic Weisbecker 504391dc69cSFrederic Weisbeckerconfig TASK_DELAY_ACCT 50519c92399SKees Cook bool "Enable per-task delay accounting" 506391dc69cSFrederic Weisbecker depends on TASKSTATS 507f6db8347SNaveen N. Rao select SCHED_INFO 508391dc69cSFrederic Weisbecker help 509391dc69cSFrederic Weisbecker Collect information on time spent by a task waiting for system 510391dc69cSFrederic Weisbecker resources like cpu, synchronous block I/O completion and swapping 511391dc69cSFrederic Weisbecker in pages. Such statistics can help in setting a task's priorities 512391dc69cSFrederic Weisbecker relative to other tasks for cpu, io, rss limits etc. 513391dc69cSFrederic Weisbecker 514391dc69cSFrederic Weisbecker Say N if unsure. 515391dc69cSFrederic Weisbecker 516391dc69cSFrederic Weisbeckerconfig TASK_XACCT 51719c92399SKees Cook bool "Enable extended accounting over taskstats" 518391dc69cSFrederic Weisbecker depends on TASKSTATS 519391dc69cSFrederic Weisbecker help 520391dc69cSFrederic Weisbecker Collect extended task accounting data and send the data 521391dc69cSFrederic Weisbecker to userland for processing over the taskstats interface. 522391dc69cSFrederic Weisbecker 523391dc69cSFrederic Weisbecker Say N if unsure. 524391dc69cSFrederic Weisbecker 525391dc69cSFrederic Weisbeckerconfig TASK_IO_ACCOUNTING 52619c92399SKees Cook bool "Enable per-task storage I/O accounting" 527391dc69cSFrederic Weisbecker depends on TASK_XACCT 528391dc69cSFrederic Weisbecker help 529391dc69cSFrederic Weisbecker Collect information on the number of bytes of storage I/O which this 530391dc69cSFrederic Weisbecker task has caused. 531391dc69cSFrederic Weisbecker 532391dc69cSFrederic Weisbecker Say N if unsure. 533391dc69cSFrederic Weisbecker 534eb414681SJohannes Weinerconfig PSI 535eb414681SJohannes Weiner bool "Pressure stall information tracking" 536eb414681SJohannes Weiner help 537eb414681SJohannes Weiner Collect metrics that indicate how overcommitted the CPU, memory, 538eb414681SJohannes Weiner and IO capacity are in the system. 539eb414681SJohannes Weiner 540eb414681SJohannes Weiner If you say Y here, the kernel will create /proc/pressure/ with the 541eb414681SJohannes Weiner pressure statistics files cpu, memory, and io. These will indicate 542eb414681SJohannes Weiner the share of walltime in which some or all tasks in the system are 543eb414681SJohannes Weiner delayed due to contention of the respective resource. 544eb414681SJohannes Weiner 5452ce7135aSJohannes Weiner In kernels with cgroup support, cgroups (cgroup2 only) will 5462ce7135aSJohannes Weiner have cpu.pressure, memory.pressure, and io.pressure files, 5472ce7135aSJohannes Weiner which aggregate pressure stalls for the grouped tasks only. 5482ce7135aSJohannes Weiner 549c3123552SMauro Carvalho Chehab For more details see Documentation/accounting/psi.rst. 550eb414681SJohannes Weiner 551eb414681SJohannes Weiner Say N if unsure. 552eb414681SJohannes Weiner 553e0c27447SJohannes Weinerconfig PSI_DEFAULT_DISABLED 554e0c27447SJohannes Weiner bool "Require boot parameter to enable pressure stall information tracking" 555e0c27447SJohannes Weiner default n 556e0c27447SJohannes Weiner depends on PSI 557e0c27447SJohannes Weiner help 558e0c27447SJohannes Weiner If set, pressure stall information tracking will be disabled 559428a1cb4SBaruch Siach per default but can be enabled through passing psi=1 on the 560428a1cb4SBaruch Siach kernel commandline during boot. 561e0c27447SJohannes Weiner 5627b2489d3SJohannes Weiner This feature adds some code to the task wakeup and sleep 5637b2489d3SJohannes Weiner paths of the scheduler. The overhead is too low to affect 5647b2489d3SJohannes Weiner common scheduling-intense workloads in practice (such as 5657b2489d3SJohannes Weiner webservers, memcache), but it does show up in artificial 5667b2489d3SJohannes Weiner scheduler stress tests, such as hackbench. 5677b2489d3SJohannes Weiner 5687b2489d3SJohannes Weiner If you are paranoid and not sure what the kernel will be 5697b2489d3SJohannes Weiner used for, say Y. 5707b2489d3SJohannes Weiner 5717b2489d3SJohannes Weiner Say N if unsure. 5727b2489d3SJohannes Weiner 573391dc69cSFrederic Weisbeckerendmenu # "CPU/Task time and stats accounting" 574391dc69cSFrederic Weisbecker 5755c4991e2SFrederic Weisbeckerconfig CPU_ISOLATION 5765c4991e2SFrederic Weisbecker bool "CPU isolation" 577414a2dc1SGeert Uytterhoeven depends on SMP || COMPILE_TEST 5782c43838cSFrederic Weisbecker default y 5795c4991e2SFrederic Weisbecker help 5805c4991e2SFrederic Weisbecker Make sure that CPUs running critical tasks are not disturbed by 5815c4991e2SFrederic Weisbecker any source of "noise" such as unbound workqueues, timers, kthreads... 5822c43838cSFrederic Weisbecker Unbound jobs get offloaded to housekeeping CPUs. This is driven by 5832c43838cSFrederic Weisbecker the "isolcpus=" boot parameter. 5842c43838cSFrederic Weisbecker 5852c43838cSFrederic Weisbecker Say Y if unsure. 5865c4991e2SFrederic Weisbecker 5870af92d46SPaul E. McKenneysource "kernel/rcu/Kconfig" 588c903ff83SMike Travis 589de5b56baSVivek Goyalconfig BUILD_BIN2C 590de5b56baSVivek Goyal bool 591de5b56baSVivek Goyal default n 592de5b56baSVivek Goyal 5931da177e4SLinus Torvaldsconfig IKCONFIG 594f2443ab6SRoss Biro tristate "Kernel .config support" 5951da177e4SLinus Torvalds ---help--- 5961da177e4SLinus Torvalds This option enables the complete Linux kernel ".config" file 5971da177e4SLinus Torvalds contents to be saved in the kernel. It provides documentation 5981da177e4SLinus Torvalds of which kernel options are used in a running kernel or in an 5991da177e4SLinus Torvalds on-disk kernel. This information can be extracted from the kernel 6001da177e4SLinus Torvalds image file with the script scripts/extract-ikconfig and used as 6011da177e4SLinus Torvalds input to rebuild the current kernel or to build another kernel. 6021da177e4SLinus Torvalds It can also be extracted from a running kernel by reading 6031da177e4SLinus Torvalds /proc/config.gz if enabled (below). 6041da177e4SLinus Torvalds 6051da177e4SLinus Torvaldsconfig IKCONFIG_PROC 6061da177e4SLinus Torvalds bool "Enable access to .config through /proc/config.gz" 6071da177e4SLinus Torvalds depends on IKCONFIG && PROC_FS 6081da177e4SLinus Torvalds ---help--- 6091da177e4SLinus Torvalds This option enables access to the kernel configuration file 6101da177e4SLinus Torvalds through /proc/config.gz. 6111da177e4SLinus Torvalds 612f7b101d3SJoel Fernandes (Google)config IKHEADERS 613f7b101d3SJoel Fernandes (Google) tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz" 614f7b101d3SJoel Fernandes (Google) depends on SYSFS 61543d8ce9dSJoel Fernandes (Google) help 616f7b101d3SJoel Fernandes (Google) This option enables access to the in-kernel headers that are generated during 617f7b101d3SJoel Fernandes (Google) the build process. These can be used to build eBPF tracing programs, 618f7b101d3SJoel Fernandes (Google) or similar programs. If you build the headers as a module, a module called 619f7b101d3SJoel Fernandes (Google) kheaders.ko is built which can be loaded on-demand to get access to headers. 62043d8ce9dSJoel Fernandes (Google) 621794543a2SAlistair John Strachanconfig LOG_BUF_SHIFT 622794543a2SAlistair John Strachan int "Kernel log buffer size (16 => 64KB, 17 => 128KB)" 623fb39f98dSIngo Molnar range 12 25 624f17a32e9SAdrian Bunk default 17 625361e9dfbSJosh Triplett depends on PRINTK 626794543a2SAlistair John Strachan help 62723b2899fSLuis R. Rodriguez Select the minimal kernel log buffer size as a power of 2. 62823b2899fSLuis R. Rodriguez The final size is affected by LOG_CPU_MAX_BUF_SHIFT config 62923b2899fSLuis R. Rodriguez parameter, see below. Any higher size also might be forced 63023b2899fSLuis R. Rodriguez by "log_buf_len" boot parameter. 63123b2899fSLuis R. Rodriguez 632f17a32e9SAdrian Bunk Examples: 633f17a32e9SAdrian Bunk 17 => 128 KB 634f17a32e9SAdrian Bunk 16 => 64 KB 635f17a32e9SAdrian Bunk 15 => 32 KB 636f17a32e9SAdrian Bunk 14 => 16 KB 637794543a2SAlistair John Strachan 13 => 8 KB 638794543a2SAlistair John Strachan 12 => 4 KB 639794543a2SAlistair John Strachan 64023b2899fSLuis R. Rodriguezconfig LOG_CPU_MAX_BUF_SHIFT 64123b2899fSLuis R. Rodriguez int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)" 6422240a31dSGeert Uytterhoeven depends on SMP 64323b2899fSLuis R. Rodriguez range 0 21 64423b2899fSLuis R. Rodriguez default 12 if !BASE_SMALL 64523b2899fSLuis R. Rodriguez default 0 if BASE_SMALL 646361e9dfbSJosh Triplett depends on PRINTK 64723b2899fSLuis R. Rodriguez help 64823b2899fSLuis R. Rodriguez This option allows to increase the default ring buffer size 64923b2899fSLuis R. Rodriguez according to the number of CPUs. The value defines the contribution 65023b2899fSLuis R. Rodriguez of each CPU as a power of 2. The used space is typically only few 65123b2899fSLuis R. Rodriguez lines however it might be much more when problems are reported, 65223b2899fSLuis R. Rodriguez e.g. backtraces. 65323b2899fSLuis R. Rodriguez 65423b2899fSLuis R. Rodriguez The increased size means that a new buffer has to be allocated and 65523b2899fSLuis R. Rodriguez the original static one is unused. It makes sense only on systems 65623b2899fSLuis R. Rodriguez with more CPUs. Therefore this value is used only when the sum of 65723b2899fSLuis R. Rodriguez contributions is greater than the half of the default kernel ring 65823b2899fSLuis R. Rodriguez buffer as defined by LOG_BUF_SHIFT. The default values are set 65923b2899fSLuis R. Rodriguez so that more than 64 CPUs are needed to trigger the allocation. 66023b2899fSLuis R. Rodriguez 66123b2899fSLuis R. Rodriguez Also this option is ignored when "log_buf_len" kernel parameter is 66223b2899fSLuis R. Rodriguez used as it forces an exact (power of two) size of the ring buffer. 66323b2899fSLuis R. Rodriguez 66423b2899fSLuis R. Rodriguez The number of possible CPUs is used for this computation ignoring 6655e0d8d59SGeert Uytterhoeven hotplugging making the computation optimal for the worst case 6665e0d8d59SGeert Uytterhoeven scenario while allowing a simple algorithm to be used from bootup. 66723b2899fSLuis R. Rodriguez 66823b2899fSLuis R. Rodriguez Examples shift values and their meaning: 66923b2899fSLuis R. Rodriguez 17 => 128 KB for each CPU 67023b2899fSLuis R. Rodriguez 16 => 64 KB for each CPU 67123b2899fSLuis R. Rodriguez 15 => 32 KB for each CPU 67223b2899fSLuis R. Rodriguez 14 => 16 KB for each CPU 67323b2899fSLuis R. Rodriguez 13 => 8 KB for each CPU 67423b2899fSLuis R. Rodriguez 12 => 4 KB for each CPU 67523b2899fSLuis R. Rodriguez 676f92bac3bSSergey Senozhatskyconfig PRINTK_SAFE_LOG_BUF_SHIFT 677f92bac3bSSergey Senozhatsky int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)" 678427934b8SPetr Mladek range 10 21 679427934b8SPetr Mladek default 13 680f92bac3bSSergey Senozhatsky depends on PRINTK 681427934b8SPetr Mladek help 682f92bac3bSSergey Senozhatsky Select the size of an alternate printk per-CPU buffer where messages 683f92bac3bSSergey Senozhatsky printed from usafe contexts are temporary stored. One example would 684f92bac3bSSergey Senozhatsky be NMI messages, another one - printk recursion. The messages are 685f92bac3bSSergey Senozhatsky copied to the main log buffer in a safe context to avoid a deadlock. 686f92bac3bSSergey Senozhatsky The value defines the size as a power of 2. 687427934b8SPetr Mladek 688f92bac3bSSergey Senozhatsky Those messages are rare and limited. The largest one is when 689427934b8SPetr Mladek a backtrace is printed. It usually fits into 4KB. Select 690427934b8SPetr Mladek 8KB if you want to be on the safe side. 691427934b8SPetr Mladek 692427934b8SPetr Mladek Examples: 693427934b8SPetr Mladek 17 => 128 KB for each CPU 694427934b8SPetr Mladek 16 => 64 KB for each CPU 695427934b8SPetr Mladek 15 => 32 KB for each CPU 696427934b8SPetr Mladek 14 => 16 KB for each CPU 697427934b8SPetr Mladek 13 => 8 KB for each CPU 698427934b8SPetr Mladek 12 => 4 KB for each CPU 699427934b8SPetr Mladek 7005cdc38f9SKAMEZAWA Hiroyuki# 7015cdc38f9SKAMEZAWA Hiroyuki# Architectures with an unreliable sched_clock() should select this: 7025cdc38f9SKAMEZAWA Hiroyuki# 7035cdc38f9SKAMEZAWA Hiroyukiconfig HAVE_UNSTABLE_SCHED_CLOCK 7045cdc38f9SKAMEZAWA Hiroyuki bool 7055cdc38f9SKAMEZAWA Hiroyuki 70638ff87f7SStephen Boydconfig GENERIC_SCHED_CLOCK 70738ff87f7SStephen Boyd bool 70838ff87f7SStephen Boyd 70969842cbaSPatrick Bellasimenu "Scheduler features" 71069842cbaSPatrick Bellasi 71169842cbaSPatrick Bellasiconfig UCLAMP_TASK 71269842cbaSPatrick Bellasi bool "Enable utilization clamping for RT/FAIR tasks" 71369842cbaSPatrick Bellasi depends on CPU_FREQ_GOV_SCHEDUTIL 71469842cbaSPatrick Bellasi help 71569842cbaSPatrick Bellasi This feature enables the scheduler to track the clamped utilization 71669842cbaSPatrick Bellasi of each CPU based on RUNNABLE tasks scheduled on that CPU. 71769842cbaSPatrick Bellasi 71869842cbaSPatrick Bellasi With this option, the user can specify the min and max CPU 71969842cbaSPatrick Bellasi utilization allowed for RUNNABLE tasks. The max utilization defines 72069842cbaSPatrick Bellasi the maximum frequency a task should use while the min utilization 72169842cbaSPatrick Bellasi defines the minimum frequency it should use. 72269842cbaSPatrick Bellasi 72369842cbaSPatrick Bellasi Both min and max utilization clamp values are hints to the scheduler, 72469842cbaSPatrick Bellasi aiming at improving its frequency selection policy, but they do not 72569842cbaSPatrick Bellasi enforce or grant any specific bandwidth for tasks. 72669842cbaSPatrick Bellasi 72769842cbaSPatrick Bellasi If in doubt, say N. 72869842cbaSPatrick Bellasi 72969842cbaSPatrick Bellasiconfig UCLAMP_BUCKETS_COUNT 73069842cbaSPatrick Bellasi int "Number of supported utilization clamp buckets" 73169842cbaSPatrick Bellasi range 5 20 73269842cbaSPatrick Bellasi default 5 73369842cbaSPatrick Bellasi depends on UCLAMP_TASK 73469842cbaSPatrick Bellasi help 73569842cbaSPatrick Bellasi Defines the number of clamp buckets to use. The range of each bucket 73669842cbaSPatrick Bellasi will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the 73769842cbaSPatrick Bellasi number of clamp buckets the finer their granularity and the higher 73869842cbaSPatrick Bellasi the precision of clamping aggregation and tracking at run-time. 73969842cbaSPatrick Bellasi 74069842cbaSPatrick Bellasi For example, with the minimum configuration value we will have 5 74169842cbaSPatrick Bellasi clamp buckets tracking 20% utilization each. A 25% boosted tasks will 74269842cbaSPatrick Bellasi be refcounted in the [20..39]% bucket and will set the bucket clamp 74369842cbaSPatrick Bellasi effective value to 25%. 74469842cbaSPatrick Bellasi If a second 30% boosted task should be co-scheduled on the same CPU, 74569842cbaSPatrick Bellasi that task will be refcounted in the same bucket of the first task and 74669842cbaSPatrick Bellasi it will boost the bucket clamp effective value to 30%. 74769842cbaSPatrick Bellasi The clamp effective value of a bucket is reset to its nominal value 74869842cbaSPatrick Bellasi (20% in the example above) when there are no more tasks refcounted in 74969842cbaSPatrick Bellasi that bucket. 75069842cbaSPatrick Bellasi 75169842cbaSPatrick Bellasi An additional boost/capping margin can be added to some tasks. In the 75269842cbaSPatrick Bellasi example above the 25% task will be boosted to 30% until it exits the 75369842cbaSPatrick Bellasi CPU. If that should be considered not acceptable on certain systems, 75469842cbaSPatrick Bellasi it's always possible to reduce the margin by increasing the number of 75569842cbaSPatrick Bellasi clamp buckets to trade off used memory for run-time tracking 75669842cbaSPatrick Bellasi precision. 75769842cbaSPatrick Bellasi 75869842cbaSPatrick Bellasi If in doubt, use the default value. 75969842cbaSPatrick Bellasi 76069842cbaSPatrick Bellasiendmenu 76169842cbaSPatrick Bellasi 762be3a7284SAndrea Arcangeli# 763be3a7284SAndrea Arcangeli# For architectures that want to enable the support for NUMA-affine scheduler 764be3a7284SAndrea Arcangeli# balancing logic: 765be3a7284SAndrea Arcangeli# 766be3a7284SAndrea Arcangeliconfig ARCH_SUPPORTS_NUMA_BALANCING 767be3a7284SAndrea Arcangeli bool 768be3a7284SAndrea Arcangeli 769be5e610cSPeter Zijlstra# 77072b252aeSMel Gorman# For architectures that prefer to flush all TLBs after a number of pages 77172b252aeSMel Gorman# are unmapped instead of sending one IPI per page to flush. The architecture 77272b252aeSMel Gorman# must provide guarantees on what happens if a clean TLB cache entry is 77372b252aeSMel Gorman# written after the unmap. Details are in mm/rmap.c near the check for 77472b252aeSMel Gorman# should_defer_flush. The architecture should also consider if the full flush 77572b252aeSMel Gorman# and the refill costs are offset by the savings of sending fewer IPIs. 77672b252aeSMel Gormanconfig ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH 77772b252aeSMel Gorman bool 77872b252aeSMel Gorman 779c12d3362SArd Biesheuvelconfig CC_HAS_INT128 7803a7c7331SMasahiro Yamada def_bool !$(cc-option,$(m64-flag) -D__SIZEOF_INT128__=0) && 64BIT 781c12d3362SArd Biesheuvel 78272b252aeSMel Gorman# 783be5e610cSPeter Zijlstra# For architectures that know their GCC __int128 support is sound 784be5e610cSPeter Zijlstra# 785be5e610cSPeter Zijlstraconfig ARCH_SUPPORTS_INT128 786be5e610cSPeter Zijlstra bool 787be5e610cSPeter Zijlstra 788be3a7284SAndrea Arcangeli# For architectures that (ab)use NUMA to represent different memory regions 789be3a7284SAndrea Arcangeli# all cpu-local but of different latencies, such as SuperH. 790be3a7284SAndrea Arcangeli# 791be3a7284SAndrea Arcangeliconfig ARCH_WANT_NUMA_VARIABLE_LOCALITY 792be3a7284SAndrea Arcangeli bool 793be3a7284SAndrea Arcangeli 794be3a7284SAndrea Arcangeliconfig NUMA_BALANCING 795be3a7284SAndrea Arcangeli bool "Memory placement aware NUMA scheduler" 796be3a7284SAndrea Arcangeli depends on ARCH_SUPPORTS_NUMA_BALANCING 797be3a7284SAndrea Arcangeli depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY 798be3a7284SAndrea Arcangeli depends on SMP && NUMA && MIGRATION 799be3a7284SAndrea Arcangeli help 800be3a7284SAndrea Arcangeli This option adds support for automatic NUMA aware memory/task placement. 801be3a7284SAndrea Arcangeli The mechanism is quite primitive and is based on migrating memory when 8026d56a410SPaul Gortmaker it has references to the node the task is running on. 803be3a7284SAndrea Arcangeli 804be3a7284SAndrea Arcangeli This system will be inactive on UMA systems. 805be3a7284SAndrea Arcangeli 8066f7c97e8SAneesh Kumar K.Vconfig NUMA_BALANCING_DEFAULT_ENABLED 8076f7c97e8SAneesh Kumar K.V bool "Automatically enable NUMA aware memory/task placement" 8086f7c97e8SAneesh Kumar K.V default y 8096f7c97e8SAneesh Kumar K.V depends on NUMA_BALANCING 8106f7c97e8SAneesh Kumar K.V help 8116f7c97e8SAneesh Kumar K.V If set, automatic NUMA balancing will be enabled if running on a NUMA 8126f7c97e8SAneesh Kumar K.V machine. 8136f7c97e8SAneesh Kumar K.V 81423964d2dSLi Zefanmenuconfig CGROUPS 8156341e62bSChristoph Jaeger bool "Control Group support" 8162bd59d48STejun Heo select KERNFS 817ddbcc7e8SPaul Menage help 81823964d2dSLi Zefan This option adds support for grouping sets of processes together, for 8195cdc38f9SKAMEZAWA Hiroyuki use with process control subsystems such as Cpusets, CFS, memory 8205cdc38f9SKAMEZAWA Hiroyuki controls or device isolation. 8215cdc38f9SKAMEZAWA Hiroyuki See 822d6a3b247SMauro Carvalho Chehab - Documentation/scheduler/sched-design-CFS.rst (CFS) 823da82c92fSMauro Carvalho Chehab - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation 82445ce80fbSLi Zefan and resource control) 825ddbcc7e8SPaul Menage 826ddbcc7e8SPaul Menage Say N if unsure. 827ddbcc7e8SPaul Menage 82823964d2dSLi Zefanif CGROUPS 82923964d2dSLi Zefan 8303e32cb2eSJohannes Weinerconfig PAGE_COUNTER 8313e32cb2eSJohannes Weiner bool 8323e32cb2eSJohannes Weiner 833c255a458SAndrew Mortonconfig MEMCG 834a0166ec4SJohannes Weiner bool "Memory controller" 8353e32cb2eSJohannes Weiner select PAGE_COUNTER 83679bd9814STejun Heo select EVENTFD 83700f0b825SBalbir Singh help 838a0166ec4SJohannes Weiner Provides control over the memory footprint of tasks in a cgroup. 83900f0b825SBalbir Singh 840c255a458SAndrew Mortonconfig MEMCG_SWAP 841a0166ec4SJohannes Weiner bool "Swap controller" 842c255a458SAndrew Morton depends on MEMCG && SWAP 843c077719bSKAMEZAWA Hiroyuki help 844a0166ec4SJohannes Weiner Provides control over the swap space consumed by tasks in a cgroup. 845a0166ec4SJohannes Weiner 846c255a458SAndrew Mortonconfig MEMCG_SWAP_ENABLED 847a0166ec4SJohannes Weiner bool "Swap controller enabled by default" 848c255a458SAndrew Morton depends on MEMCG_SWAP 849a42c390cSMichal Hocko default y 850a42c390cSMichal Hocko help 851a42c390cSMichal Hocko Memory Resource Controller Swap Extension comes with its price in 852a42c390cSMichal Hocko a bigger memory consumption. General purpose distribution kernels 85343d547f9SJim Cromie which want to enable the feature but keep it disabled by default 85407555ac1SMichal Hocko and let the user enable it by swapaccount=1 boot command line 855a42c390cSMichal Hocko parameter should have this option unselected. 856a42c390cSMichal Hocko For those who want to have the feature enabled by default should 857a42c390cSMichal Hocko select this option (if, for some reason, they need to disable it 85800a66d29SWANG Cong then swapaccount=0 does the trick). 859c077719bSKAMEZAWA Hiroyuki 86084c07d11SKirill Tkhaiconfig MEMCG_KMEM 86184c07d11SKirill Tkhai bool 86284c07d11SKirill Tkhai depends on MEMCG && !SLOB 86384c07d11SKirill Tkhai default y 86484c07d11SKirill Tkhai 8656bf024e6SJohannes Weinerconfig BLK_CGROUP 8666bf024e6SJohannes Weiner bool "IO controller" 8676bf024e6SJohannes Weiner depends on BLOCK 8682bc64a20SAneesh Kumar K.V default n 8696bf024e6SJohannes Weiner ---help--- 8706bf024e6SJohannes Weiner Generic block IO controller cgroup interface. This is the common 8716bf024e6SJohannes Weiner cgroup interface which should be used by various IO controlling 8726bf024e6SJohannes Weiner policies. 8732bc64a20SAneesh Kumar K.V 8746bf024e6SJohannes Weiner Currently, CFQ IO scheduler uses it to recognize task groups and 8756bf024e6SJohannes Weiner control disk bandwidth allocation (proportional time slice allocation) 8766bf024e6SJohannes Weiner to such task groups. It is also used by bio throttling logic in 8776bf024e6SJohannes Weiner block layer to implement upper limit in IO rates on a device. 878e5d1367fSStephane Eranian 8796bf024e6SJohannes Weiner This option only enables generic Block IO controller infrastructure. 8806bf024e6SJohannes Weiner One needs to also enable actual IO controlling logic/policy. For 8816bf024e6SJohannes Weiner enabling proportional weight division of disk bandwidth in CFQ, set 8827baf2199SKrzysztof Kozlowski CONFIG_BFQ_GROUP_IOSCHED=y; for enabling throttling policy, set 8836bf024e6SJohannes Weiner CONFIG_BLK_DEV_THROTTLING=y. 8846bf024e6SJohannes Weiner 885da82c92fSMauro Carvalho Chehab See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information. 8866bf024e6SJohannes Weiner 8876bf024e6SJohannes Weinerconfig CGROUP_WRITEBACK 8886bf024e6SJohannes Weiner bool 8896bf024e6SJohannes Weiner depends on MEMCG && BLK_CGROUP 8906bf024e6SJohannes Weiner default y 891e5d1367fSStephane Eranian 8927c941438SDhaval Gianimenuconfig CGROUP_SCHED 893a0166ec4SJohannes Weiner bool "CPU controller" 8947c941438SDhaval Giani default n 8957c941438SDhaval Giani help 8967c941438SDhaval Giani This feature lets CPU scheduler recognize task groups and control CPU 8977c941438SDhaval Giani bandwidth allocation to such task groups. It uses cgroups to group 8987c941438SDhaval Giani tasks. 8997c941438SDhaval Giani 9007c941438SDhaval Gianiif CGROUP_SCHED 9017c941438SDhaval Gianiconfig FAIR_GROUP_SCHED 9027c941438SDhaval Giani bool "Group scheduling for SCHED_OTHER" 9037c941438SDhaval Giani depends on CGROUP_SCHED 9047c941438SDhaval Giani default CGROUP_SCHED 9057c941438SDhaval Giani 906ab84d31eSPaul Turnerconfig CFS_BANDWIDTH 907ab84d31eSPaul Turner bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED" 908ab84d31eSPaul Turner depends on FAIR_GROUP_SCHED 909ab84d31eSPaul Turner default n 910ab84d31eSPaul Turner help 911ab84d31eSPaul Turner This option allows users to define CPU bandwidth rates (limits) for 912ab84d31eSPaul Turner tasks running within the fair group scheduler. Groups with no limit 913ab84d31eSPaul Turner set are considered to be unconstrained and will run with no 914ab84d31eSPaul Turner restriction. 915d6a3b247SMauro Carvalho Chehab See Documentation/scheduler/sched-bwc.rst for more information. 916ab84d31eSPaul Turner 9177c941438SDhaval Gianiconfig RT_GROUP_SCHED 9187c941438SDhaval Giani bool "Group scheduling for SCHED_RR/FIFO" 9197c941438SDhaval Giani depends on CGROUP_SCHED 9207c941438SDhaval Giani default n 9217c941438SDhaval Giani help 9227c941438SDhaval Giani This feature lets you explicitly allocate real CPU bandwidth 92332bd7eb5SLi Zefan to task groups. If enabled, it will also make it impossible to 9247c941438SDhaval Giani schedule realtime tasks for non-root users until you allocate 9257c941438SDhaval Giani realtime bandwidth for them. 926d6a3b247SMauro Carvalho Chehab See Documentation/scheduler/sched-rt-group.rst for more information. 9277c941438SDhaval Giani 9287c941438SDhaval Gianiendif #CGROUP_SCHED 9297c941438SDhaval Giani 9302480c093SPatrick Bellasiconfig UCLAMP_TASK_GROUP 9312480c093SPatrick Bellasi bool "Utilization clamping per group of tasks" 9322480c093SPatrick Bellasi depends on CGROUP_SCHED 9332480c093SPatrick Bellasi depends on UCLAMP_TASK 9342480c093SPatrick Bellasi default n 9352480c093SPatrick Bellasi help 9362480c093SPatrick Bellasi This feature enables the scheduler to track the clamped utilization 9372480c093SPatrick Bellasi of each CPU based on RUNNABLE tasks currently scheduled on that CPU. 9382480c093SPatrick Bellasi 9392480c093SPatrick Bellasi When this option is enabled, the user can specify a min and max 9402480c093SPatrick Bellasi CPU bandwidth which is allowed for each single task in a group. 9412480c093SPatrick Bellasi The max bandwidth allows to clamp the maximum frequency a task 9422480c093SPatrick Bellasi can use, while the min bandwidth allows to define a minimum 9432480c093SPatrick Bellasi frequency a task will always use. 9442480c093SPatrick Bellasi 9452480c093SPatrick Bellasi When task group based utilization clamping is enabled, an eventually 9462480c093SPatrick Bellasi specified task-specific clamp value is constrained by the cgroup 9472480c093SPatrick Bellasi specified clamp value. Both minimum and maximum task clamping cannot 9482480c093SPatrick Bellasi be bigger than the corresponding clamping defined at task group level. 9492480c093SPatrick Bellasi 9502480c093SPatrick Bellasi If in doubt, say N. 9512480c093SPatrick Bellasi 9526bf024e6SJohannes Weinerconfig CGROUP_PIDS 9536bf024e6SJohannes Weiner bool "PIDs controller" 9546bf024e6SJohannes Weiner help 9556bf024e6SJohannes Weiner Provides enforcement of process number limits in the scope of a 9566bf024e6SJohannes Weiner cgroup. Any attempt to fork more processes than is allowed in the 9576bf024e6SJohannes Weiner cgroup will fail. PIDs are fundamentally a global resource because it 9586bf024e6SJohannes Weiner is fairly trivial to reach PID exhaustion before you reach even a 9596bf024e6SJohannes Weiner conservative kmemcg limit. As a result, it is possible to grind a 9606bf024e6SJohannes Weiner system to halt without being limited by other cgroup policies. The 9616cc578dfSParav Pandit PIDs controller is designed to stop this from happening. 9626bf024e6SJohannes Weiner 9636bf024e6SJohannes Weiner It should be noted that organisational operations (such as attaching 96498076833SJonathan Neuschäfer to a cgroup hierarchy) will *not* be blocked by the PIDs controller, 9656bf024e6SJohannes Weiner since the PIDs limit only affects a process's ability to fork, not to 9666bf024e6SJohannes Weiner attach to a cgroup. 9676bf024e6SJohannes Weiner 96839d3e758SParav Panditconfig CGROUP_RDMA 96939d3e758SParav Pandit bool "RDMA controller" 97039d3e758SParav Pandit help 97139d3e758SParav Pandit Provides enforcement of RDMA resources defined by IB stack. 97239d3e758SParav Pandit It is fairly easy for consumers to exhaust RDMA resources, which 97339d3e758SParav Pandit can result into resource unavailability to other consumers. 97439d3e758SParav Pandit RDMA controller is designed to stop this from happening. 97539d3e758SParav Pandit Attaching processes with active RDMA resources to the cgroup 97639d3e758SParav Pandit hierarchy is allowed even if can cross the hierarchy's limit. 97739d3e758SParav Pandit 9786bf024e6SJohannes Weinerconfig CGROUP_FREEZER 9796bf024e6SJohannes Weiner bool "Freezer controller" 9806bf024e6SJohannes Weiner help 9816bf024e6SJohannes Weiner Provides a way to freeze and unfreeze all tasks in a 9826bf024e6SJohannes Weiner cgroup. 9836bf024e6SJohannes Weiner 984489c2a20SJohannes Weiner This option affects the ORIGINAL cgroup interface. The cgroup2 memory 985489c2a20SJohannes Weiner controller includes important in-kernel memory consumers per default. 986489c2a20SJohannes Weiner 987489c2a20SJohannes Weiner If you're using cgroup2, say N. 988489c2a20SJohannes Weiner 9896bf024e6SJohannes Weinerconfig CGROUP_HUGETLB 9906bf024e6SJohannes Weiner bool "HugeTLB controller" 9916bf024e6SJohannes Weiner depends on HUGETLB_PAGE 9926bf024e6SJohannes Weiner select PAGE_COUNTER 993afc24d49SVivek Goyal default n 9946bf024e6SJohannes Weiner help 9956bf024e6SJohannes Weiner Provides a cgroup controller for HugeTLB pages. 9966bf024e6SJohannes Weiner When you enable this, you can put a per cgroup limit on HugeTLB usage. 9976bf024e6SJohannes Weiner The limit is enforced during page fault. Since HugeTLB doesn't 9986bf024e6SJohannes Weiner support page reclaim, enforcing the limit at page fault time implies 9996bf024e6SJohannes Weiner that, the application will get SIGBUS signal if it tries to access 10006bf024e6SJohannes Weiner HugeTLB pages beyond its limit. This requires the application to know 10016bf024e6SJohannes Weiner beforehand how much HugeTLB pages it would require for its use. The 10026bf024e6SJohannes Weiner control group is tracked in the third page lru pointer. This means 10036bf024e6SJohannes Weiner that we cannot use the controller with huge page less than 3 pages. 1004afc24d49SVivek Goyal 10056bf024e6SJohannes Weinerconfig CPUSETS 10066bf024e6SJohannes Weiner bool "Cpuset controller" 1007e1d4eeecSNicolas Pitre depends on SMP 10086bf024e6SJohannes Weiner help 10096bf024e6SJohannes Weiner This option will let you create and manage CPUSETs which 10106bf024e6SJohannes Weiner allow dynamically partitioning a system into sets of CPUs and 10116bf024e6SJohannes Weiner Memory Nodes and assigning tasks to run only within those sets. 10126bf024e6SJohannes Weiner This is primarily useful on large SMP or NUMA systems. 1013afc24d49SVivek Goyal 10146bf024e6SJohannes Weiner Say N if unsure. 1015afc24d49SVivek Goyal 10166bf024e6SJohannes Weinerconfig PROC_PID_CPUSET 10176bf024e6SJohannes Weiner bool "Include legacy /proc/<pid>/cpuset file" 10186bf024e6SJohannes Weiner depends on CPUSETS 101989e9b9e0STejun Heo default y 102089e9b9e0STejun Heo 10216bf024e6SJohannes Weinerconfig CGROUP_DEVICE 10226bf024e6SJohannes Weiner bool "Device controller" 10236bf024e6SJohannes Weiner help 10246bf024e6SJohannes Weiner Provides a cgroup controller implementing whitelists for 10256bf024e6SJohannes Weiner devices which a process in the cgroup can mknod or open. 10266bf024e6SJohannes Weiner 10276bf024e6SJohannes Weinerconfig CGROUP_CPUACCT 10286bf024e6SJohannes Weiner bool "Simple CPU accounting controller" 10296bf024e6SJohannes Weiner help 10306bf024e6SJohannes Weiner Provides a simple controller for monitoring the 10316bf024e6SJohannes Weiner total CPU consumed by the tasks in a cgroup. 10326bf024e6SJohannes Weiner 10336bf024e6SJohannes Weinerconfig CGROUP_PERF 10346bf024e6SJohannes Weiner bool "Perf controller" 10356bf024e6SJohannes Weiner depends on PERF_EVENTS 10366bf024e6SJohannes Weiner help 10376bf024e6SJohannes Weiner This option extends the perf per-cpu mode to restrict monitoring 10386bf024e6SJohannes Weiner to threads which belong to the cgroup specified and run on the 10396546b19fSNamhyung Kim designated cpu. Or this can be used to have cgroup ID in samples 10406546b19fSNamhyung Kim so that it can monitor performance events among cgroups. 10416bf024e6SJohannes Weiner 10426bf024e6SJohannes Weiner Say N if unsure. 10436bf024e6SJohannes Weiner 104430070984SDaniel Mackconfig CGROUP_BPF 104530070984SDaniel Mack bool "Support for eBPF programs attached to cgroups" 1046483c4933SAndy Lutomirski depends on BPF_SYSCALL 1047483c4933SAndy Lutomirski select SOCK_CGROUP_DATA 104830070984SDaniel Mack help 104930070984SDaniel Mack Allow attaching eBPF programs to a cgroup using the bpf(2) 105030070984SDaniel Mack syscall command BPF_PROG_ATTACH. 105130070984SDaniel Mack 105230070984SDaniel Mack In which context these programs are accessed depends on the type 105330070984SDaniel Mack of attachment. For instance, programs that are attached using 105430070984SDaniel Mack BPF_CGROUP_INET_INGRESS will be executed on the ingress path of 105530070984SDaniel Mack inet sockets. 105630070984SDaniel Mack 10576bf024e6SJohannes Weinerconfig CGROUP_DEBUG 105823b0be48SWaiman Long bool "Debug controller" 10596bf024e6SJohannes Weiner default n 106023b0be48SWaiman Long depends on DEBUG_KERNEL 10616bf024e6SJohannes Weiner help 10626bf024e6SJohannes Weiner This option enables a simple controller that exports 106323b0be48SWaiman Long debugging information about the cgroups framework. This 106423b0be48SWaiman Long controller is for control cgroup debugging only. Its 106523b0be48SWaiman Long interfaces are not stable. 10666bf024e6SJohannes Weiner 10676bf024e6SJohannes Weiner Say N. 10686bf024e6SJohannes Weiner 106973b35147SArnd Bergmannconfig SOCK_CGROUP_DATA 107073b35147SArnd Bergmann bool 107173b35147SArnd Bergmann default n 107273b35147SArnd Bergmann 107323964d2dSLi Zefanendif # CGROUPS 1074c077719bSKAMEZAWA Hiroyuki 10758dd2a82cSDaniel Lezcanomenuconfig NAMESPACES 10766a108a14SDavid Rientjes bool "Namespaces support" if EXPERT 10772813893fSIulia Manda depends on MULTIUSER 10786a108a14SDavid Rientjes default !EXPERT 1079c5289a69SPavel Emelyanov help 1080c5289a69SPavel Emelyanov Provides the way to make tasks work with different objects using 1081c5289a69SPavel Emelyanov the same id. For example same IPC id may refer to different objects 1082c5289a69SPavel Emelyanov or same user id or pid may refer to different tasks when used in 1083c5289a69SPavel Emelyanov different namespaces. 1084c5289a69SPavel Emelyanov 10858dd2a82cSDaniel Lezcanoif NAMESPACES 10868dd2a82cSDaniel Lezcano 108758bfdd6dSPavel Emelyanovconfig UTS_NS 108858bfdd6dSPavel Emelyanov bool "UTS namespace" 108917a6d441SDaniel Lezcano default y 109058bfdd6dSPavel Emelyanov help 109158bfdd6dSPavel Emelyanov In this namespace tasks see different info provided with the 109258bfdd6dSPavel Emelyanov uname() system call 109358bfdd6dSPavel Emelyanov 1094769071acSAndrei Vaginconfig TIME_NS 1095769071acSAndrei Vagin bool "TIME namespace" 1096660fd04fSThomas Gleixner depends on GENERIC_VDSO_TIME_NS 1097769071acSAndrei Vagin default y 1098769071acSAndrei Vagin help 1099769071acSAndrei Vagin In this namespace boottime and monotonic clocks can be set. 1100769071acSAndrei Vagin The time will keep going with the same pace. 1101769071acSAndrei Vagin 1102ae5e1b22SPavel Emelyanovconfig IPC_NS 1103ae5e1b22SPavel Emelyanov bool "IPC namespace" 11048dd2a82cSDaniel Lezcano depends on (SYSVIPC || POSIX_MQUEUE) 110517a6d441SDaniel Lezcano default y 1106ae5e1b22SPavel Emelyanov help 1107ae5e1b22SPavel Emelyanov In this namespace tasks work with IPC ids which correspond to 1108614b84cfSSerge E. Hallyn different IPC objects in different namespaces. 1109ae5e1b22SPavel Emelyanov 1110aee16ce7SPavel Emelyanovconfig USER_NS 111119c92399SKees Cook bool "User namespace" 11125673a94cSEric W. Biederman default n 1113aee16ce7SPavel Emelyanov help 1114aee16ce7SPavel Emelyanov This allows containers, i.e. vservers, to use user namespaces 1115aee16ce7SPavel Emelyanov to provide different user info for different servers. 1116e11f0ae3SEric W. Biederman 1117e11f0ae3SEric W. Biederman When user namespaces are enabled in the kernel it is 1118d886f4e4SJohannes Weiner recommended that the MEMCG option also be enabled and that 1119d886f4e4SJohannes Weiner user-space use the memory control groups to limit the amount 1120d886f4e4SJohannes Weiner of memory a memory unprivileged users can use. 1121e11f0ae3SEric W. Biederman 1122aee16ce7SPavel Emelyanov If unsure, say N. 1123aee16ce7SPavel Emelyanov 112474bd59bbSPavel Emelyanovconfig PID_NS 11259bd38c2cSDaniel Lezcano bool "PID Namespaces" 112617a6d441SDaniel Lezcano default y 112774bd59bbSPavel Emelyanov help 112812d2b8f9SHeikki Orsila Support process id namespaces. This allows having multiple 1129692105b8SMatt LaPlante processes with the same pid as long as they are in different 113074bd59bbSPavel Emelyanov pid namespaces. This is a building block of containers. 113174bd59bbSPavel Emelyanov 1132d6eb633fSMatt Helsleyconfig NET_NS 1133d6eb633fSMatt Helsley bool "Network namespace" 11348dd2a82cSDaniel Lezcano depends on NET 113517a6d441SDaniel Lezcano default y 1136d6eb633fSMatt Helsley help 1137d6eb633fSMatt Helsley Allow user space to create what appear to be multiple instances 1138d6eb633fSMatt Helsley of the network stack. 1139d6eb633fSMatt Helsley 11408dd2a82cSDaniel Lezcanoendif # NAMESPACES 11418dd2a82cSDaniel Lezcano 11425cb366bbSAdrian Reberconfig CHECKPOINT_RESTORE 11435cb366bbSAdrian Reber bool "Checkpoint/restore support" 11445cb366bbSAdrian Reber select PROC_CHILDREN 11455cb366bbSAdrian Reber default n 11465cb366bbSAdrian Reber help 11475cb366bbSAdrian Reber Enables additional kernel features in a sake of checkpoint/restore. 11485cb366bbSAdrian Reber In particular it adds auxiliary prctl codes to setup process text, 11495cb366bbSAdrian Reber data and heap segment sizes, and a few additional /proc filesystem 11505cb366bbSAdrian Reber entries. 11515cb366bbSAdrian Reber 11525cb366bbSAdrian Reber If unsure, say N here. 11535cb366bbSAdrian Reber 11545091faa4SMike Galbraithconfig SCHED_AUTOGROUP 11555091faa4SMike Galbraith bool "Automatic process group scheduling" 11565091faa4SMike Galbraith select CGROUPS 11575091faa4SMike Galbraith select CGROUP_SCHED 11585091faa4SMike Galbraith select FAIR_GROUP_SCHED 11595091faa4SMike Galbraith help 11605091faa4SMike Galbraith This option optimizes the scheduler for common desktop workloads by 11615091faa4SMike Galbraith automatically creating and populating task groups. This separation 11625091faa4SMike Galbraith of workloads isolates aggressive CPU burners (like build jobs) from 11635091faa4SMike Galbraith desktop applications. Task group autogeneration is currently based 11645091faa4SMike Galbraith upon task session. 11655091faa4SMike Galbraith 11667af37becSDaniel Lezcanoconfig SYSFS_DEPRECATED 11675d6a4ea5SFerenc Wagner bool "Enable deprecated sysfs features to support old userspace tools" 11687af37becSDaniel Lezcano depends on SYSFS 11697af37becSDaniel Lezcano default n 11707af37becSDaniel Lezcano help 11717af37becSDaniel Lezcano This option adds code that switches the layout of the "block" class 11727af37becSDaniel Lezcano devices, to not show up in /sys/class/block/, but only in 11737af37becSDaniel Lezcano /sys/block/. 11747af37becSDaniel Lezcano 11757af37becSDaniel Lezcano This switch is only active when the sysfs.deprecated=1 boot option is 11767af37becSDaniel Lezcano passed or the SYSFS_DEPRECATED_V2 option is set. 11777af37becSDaniel Lezcano 11787af37becSDaniel Lezcano This option allows new kernels to run on old distributions and tools, 11797af37becSDaniel Lezcano which might get confused by /sys/class/block/. Since 2007/2008 all 11807af37becSDaniel Lezcano major distributions and tools handle this just fine. 11817af37becSDaniel Lezcano 11827af37becSDaniel Lezcano Recent distributions and userspace tools after 2009/2010 depend on 11837af37becSDaniel Lezcano the existence of /sys/class/block/, and will not work with this 11847af37becSDaniel Lezcano option enabled. 11857af37becSDaniel Lezcano 11867af37becSDaniel Lezcano Only if you are using a new kernel on an old distribution, you might 11877af37becSDaniel Lezcano need to say Y here. 11887af37becSDaniel Lezcano 11897af37becSDaniel Lezcanoconfig SYSFS_DEPRECATED_V2 11905d6a4ea5SFerenc Wagner bool "Enable deprecated sysfs features by default" 11917af37becSDaniel Lezcano default n 11927af37becSDaniel Lezcano depends on SYSFS 11937af37becSDaniel Lezcano depends on SYSFS_DEPRECATED 11947af37becSDaniel Lezcano help 11957af37becSDaniel Lezcano Enable deprecated sysfs by default. 11967af37becSDaniel Lezcano 11977af37becSDaniel Lezcano See the CONFIG_SYSFS_DEPRECATED option for more details about this 11987af37becSDaniel Lezcano option. 11997af37becSDaniel Lezcano 12007af37becSDaniel Lezcano Only if you are using a new kernel on an old distribution, you might 12017af37becSDaniel Lezcano need to say Y here. Even then, odds are you would not need it 12027af37becSDaniel Lezcano enabled, you can always pass the boot option if absolutely necessary. 12037af37becSDaniel Lezcano 12047af37becSDaniel Lezcanoconfig RELAY 12057af37becSDaniel Lezcano bool "Kernel->user space relay support (formerly relayfs)" 120626b5679eSPeter Zijlstra select IRQ_WORK 12077af37becSDaniel Lezcano help 12087af37becSDaniel Lezcano This option enables support for relay interface support in 12097af37becSDaniel Lezcano certain file systems (such as debugfs). 12107af37becSDaniel Lezcano It is designed to provide an efficient mechanism for tools and 12117af37becSDaniel Lezcano facilities to relay large amounts of data from kernel space to 12127af37becSDaniel Lezcano user space. 12137af37becSDaniel Lezcano 12147af37becSDaniel Lezcano If unsure, say N. 12157af37becSDaniel Lezcano 1216f991633dSDimitri Gorokhovikconfig BLK_DEV_INITRD 1217f991633dSDimitri Gorokhovik bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support" 1218f991633dSDimitri Gorokhovik help 1219f991633dSDimitri Gorokhovik The initial RAM filesystem is a ramfs which is loaded by the 1220f991633dSDimitri Gorokhovik boot loader (loadlin or lilo) and that is mounted as root 1221f991633dSDimitri Gorokhovik before the normal boot procedure. It is typically used to 1222f991633dSDimitri Gorokhovik load modules needed to mount the "real" root file system, 12238c27ceffSMauro Carvalho Chehab etc. See <file:Documentation/admin-guide/initrd.rst> for details. 1224f991633dSDimitri Gorokhovik 1225f991633dSDimitri Gorokhovik If RAM disk support (BLK_DEV_RAM) is also included, this 1226f991633dSDimitri Gorokhovik also enables initial RAM disk (initrd) support and adds 1227f991633dSDimitri Gorokhovik 15 Kbytes (more on some other architectures) to the kernel size. 1228f991633dSDimitri Gorokhovik 1229f991633dSDimitri Gorokhovik If unsure say Y. 1230f991633dSDimitri Gorokhovik 1231c33df4eaSJean-Paul Samanif BLK_DEV_INITRD 1232c33df4eaSJean-Paul Saman 1233dbec4866SSam Ravnborgsource "usr/Kconfig" 1234dbec4866SSam Ravnborg 1235c33df4eaSJean-Paul Samanendif 1236c33df4eaSJean-Paul Saman 123776db5a27SMasami Hiramatsuconfig BOOT_CONFIG 123876db5a27SMasami Hiramatsu bool "Boot config support" 12392910b5aaSMasami Hiramatsu select BLK_DEV_INITRD 124076db5a27SMasami Hiramatsu help 124176db5a27SMasami Hiramatsu Extra boot config allows system admin to pass a config file as 124276db5a27SMasami Hiramatsu complemental extension of kernel cmdline when booting. 12430947db01SMasami Hiramatsu The boot config file must be attached at the end of initramfs 124485c46b78SMasami Hiramatsu with checksum, size and magic word. 12450947db01SMasami Hiramatsu See <file:Documentation/admin-guide/bootconfig.rst> for details. 124676db5a27SMasami Hiramatsu 124776db5a27SMasami Hiramatsu If unsure, say Y. 124876db5a27SMasami Hiramatsu 1249877417e6SArnd Bergmannchoice 1250877417e6SArnd Bergmann prompt "Compiler optimization level" 12512cc3ce24SUlf Magnusson default CC_OPTIMIZE_FOR_PERFORMANCE 1252877417e6SArnd Bergmann 1253877417e6SArnd Bergmannconfig CC_OPTIMIZE_FOR_PERFORMANCE 125415f5db60SMasahiro Yamada bool "Optimize for performance (-O2)" 1255877417e6SArnd Bergmann help 1256877417e6SArnd Bergmann This is the default optimization level for the kernel, building 1257877417e6SArnd Bergmann with the "-O2" compiler flag for best performance and most 1258877417e6SArnd Bergmann helpful compile-time warnings. 1259877417e6SArnd Bergmann 126015f5db60SMasahiro Yamadaconfig CC_OPTIMIZE_FOR_PERFORMANCE_O3 126115f5db60SMasahiro Yamada bool "Optimize more for performance (-O3)" 126215f5db60SMasahiro Yamada depends on ARC 1263b303c6dfSMasahiro Yamada imply CC_DISABLE_WARN_MAYBE_UNINITIALIZED # avoid false positives 1264c45b4f1fSLinus Torvalds help 126515f5db60SMasahiro Yamada Choosing this option will pass "-O3" to your compiler to optimize 126615f5db60SMasahiro Yamada the kernel yet more for performance. 1267c45b4f1fSLinus Torvalds 12685d20ee31SNicholas Pigginconfig CC_OPTIMIZE_FOR_SIZE 126915f5db60SMasahiro Yamada bool "Optimize for size (-Os)" 12705d20ee31SNicholas Piggin imply CC_DISABLE_WARN_MAYBE_UNINITIALIZED # avoid false positives 12715d20ee31SNicholas Piggin help 1272ce3b487fSMasahiro Yamada Choosing this option will pass "-Os" to your compiler resulting 1273ce3b487fSMasahiro Yamada in a smaller kernel. 1274c45b4f1fSLinus Torvalds 1275877417e6SArnd Bergmannendchoice 1276877417e6SArnd Bergmann 12775d20ee31SNicholas Pigginconfig HAVE_LD_DEAD_CODE_DATA_ELIMINATION 12785d20ee31SNicholas Piggin bool 12795d20ee31SNicholas Piggin help 12805d20ee31SNicholas Piggin This requires that the arch annotates or otherwise protects 12815d20ee31SNicholas Piggin its external entry points from being discarded. Linker scripts 12825d20ee31SNicholas Piggin must also merge .text.*, .data.*, and .bss.* correctly into 12835d20ee31SNicholas Piggin output sections. Care must be taken not to pull in unrelated 12845d20ee31SNicholas Piggin sections (e.g., '.text.init'). Typically '.' in section names 12855d20ee31SNicholas Piggin is used to distinguish them from label names / C identifiers. 12865d20ee31SNicholas Piggin 12875d20ee31SNicholas Pigginconfig LD_DEAD_CODE_DATA_ELIMINATION 12885d20ee31SNicholas Piggin bool "Dead code and data elimination (EXPERIMENTAL)" 12895d20ee31SNicholas Piggin depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION 12905d20ee31SNicholas Piggin depends on EXPERT 129116fd20aaSPaul Burton depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800) 1292e85d1d65SMasahiro Yamada depends on $(cc-option,-ffunction-sections -fdata-sections) 1293e85d1d65SMasahiro Yamada depends on $(ld-option,--gc-sections) 12945d20ee31SNicholas Piggin help 12958b9d2712SMasahiro Yamada Enable this if you want to do dead code and data elimination with 12968b9d2712SMasahiro Yamada the linker by compiling with -ffunction-sections -fdata-sections, 12978b9d2712SMasahiro Yamada and linking with --gc-sections. 12985d20ee31SNicholas Piggin 12995d20ee31SNicholas Piggin This can reduce on disk and in-memory size of the kernel 13005d20ee31SNicholas Piggin code and static data, particularly for small configs and 13015d20ee31SNicholas Piggin on small systems. This has the possibility of introducing 13025d20ee31SNicholas Piggin silently broken kernel if the required annotations are not 13035d20ee31SNicholas Piggin present. This option is not well tested yet, so use at your 13045d20ee31SNicholas Piggin own risk. 13055d20ee31SNicholas Piggin 13060847062aSRandy Dunlapconfig SYSCTL 13070847062aSRandy Dunlap bool 13080847062aSRandy Dunlap 1309657a5209SMike Frysingerconfig HAVE_UID16 1310657a5209SMike Frysinger bool 1311657a5209SMike Frysinger 1312657a5209SMike Frysingerconfig SYSCTL_EXCEPTION_TRACE 1313657a5209SMike Frysinger bool 1314657a5209SMike Frysinger help 1315657a5209SMike Frysinger Enable support for /proc/sys/debug/exception-trace. 1316657a5209SMike Frysinger 1317657a5209SMike Frysingerconfig SYSCTL_ARCH_UNALIGN_NO_WARN 1318657a5209SMike Frysinger bool 1319657a5209SMike Frysinger help 1320657a5209SMike Frysinger Enable support for /proc/sys/kernel/ignore-unaligned-usertrap 1321657a5209SMike Frysinger Allows arch to define/use @no_unaligned_warning to possibly warn 1322657a5209SMike Frysinger about unaligned access emulation going on under the hood. 1323657a5209SMike Frysinger 1324657a5209SMike Frysingerconfig SYSCTL_ARCH_UNALIGN_ALLOW 1325657a5209SMike Frysinger bool 1326657a5209SMike Frysinger help 1327657a5209SMike Frysinger Enable support for /proc/sys/kernel/unaligned-trap 1328657a5209SMike Frysinger Allows arches to define/use @unaligned_enabled to runtime toggle 1329657a5209SMike Frysinger the unaligned access emulation. 1330657a5209SMike Frysinger see arch/parisc/kernel/unaligned.c for reference 1331657a5209SMike Frysinger 1332657a5209SMike Frysingerconfig HAVE_PCSPKR_PLATFORM 1333657a5209SMike Frysinger bool 1334657a5209SMike Frysinger 1335f89b7755SAlexei Starovoitov# interpreter that classic socket filters depend on 1336f89b7755SAlexei Starovoitovconfig BPF 1337f89b7755SAlexei Starovoitov bool 1338f89b7755SAlexei Starovoitov 13396a108a14SDavid Rientjesmenuconfig EXPERT 13406a108a14SDavid Rientjes bool "Configure standard kernel features (expert users)" 1341f505c553SJosh Triplett # Unhide debug options, to make the on-by-default options visible 1342f505c553SJosh Triplett select DEBUG_KERNEL 13431da177e4SLinus Torvalds help 13441da177e4SLinus Torvalds This option allows certain base kernel options and settings 13451da177e4SLinus Torvalds to be disabled or tweaked. This is for specialized 13461da177e4SLinus Torvalds environments which can tolerate a "non-standard" kernel. 13471da177e4SLinus Torvalds Only use this if you really know what you are doing. 13481da177e4SLinus Torvalds 1349ae81f9e3SChuck Ebbertconfig UID16 13506a108a14SDavid Rientjes bool "Enable 16-bit UID system calls" if EXPERT 13512813893fSIulia Manda depends on HAVE_UID16 && MULTIUSER 1352ae81f9e3SChuck Ebbert default y 1353ae81f9e3SChuck Ebbert help 1354ae81f9e3SChuck Ebbert This enables the legacy 16-bit UID syscall wrappers. 1355ae81f9e3SChuck Ebbert 13562813893fSIulia Mandaconfig MULTIUSER 13572813893fSIulia Manda bool "Multiple users, groups and capabilities support" if EXPERT 13582813893fSIulia Manda default y 13592813893fSIulia Manda help 13602813893fSIulia Manda This option enables support for non-root users, groups and 13612813893fSIulia Manda capabilities. 13622813893fSIulia Manda 13632813893fSIulia Manda If you say N here, all processes will run with UID 0, GID 0, and all 13642813893fSIulia Manda possible capabilities. Saying N here also compiles out support for 13652813893fSIulia Manda system calls related to UIDs, GIDs, and capabilities, such as setuid, 13662813893fSIulia Manda setgid, and capset. 13672813893fSIulia Manda 13682813893fSIulia Manda If unsure, say Y here. 13692813893fSIulia Manda 1370f6187769SFabian Frederickconfig SGETMASK_SYSCALL 1371f6187769SFabian Frederick bool "sgetmask/ssetmask syscalls support" if EXPERT 1372a687a533SArnd Bergmann def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH 1373f6187769SFabian Frederick ---help--- 1374f6187769SFabian Frederick sys_sgetmask and sys_ssetmask are obsolete system calls 1375f6187769SFabian Frederick no longer supported in libc but still enabled by default in some 1376f6187769SFabian Frederick architectures. 1377f6187769SFabian Frederick 1378f6187769SFabian Frederick If unsure, leave the default option here. 1379f6187769SFabian Frederick 13806af9f7bfSFabian Frederickconfig SYSFS_SYSCALL 13816af9f7bfSFabian Frederick bool "Sysfs syscall support" if EXPERT 13826af9f7bfSFabian Frederick default y 13836af9f7bfSFabian Frederick ---help--- 13846af9f7bfSFabian Frederick sys_sysfs is an obsolete system call no longer supported in libc. 13856af9f7bfSFabian Frederick Note that disabling this option is more secure but might break 13866af9f7bfSFabian Frederick compatibility with some systems. 13876af9f7bfSFabian Frederick 13886af9f7bfSFabian Frederick If unsure say Y here. 13896af9f7bfSFabian Frederick 1390d1b069f5SRandy Dunlapconfig FHANDLE 1391d1b069f5SRandy Dunlap bool "open by fhandle syscalls" if EXPERT 1392d1b069f5SRandy Dunlap select EXPORTFS 1393d1b069f5SRandy Dunlap default y 1394d1b069f5SRandy Dunlap help 1395d1b069f5SRandy Dunlap If you say Y here, a user level program will be able to map 1396d1b069f5SRandy Dunlap file names to handle and then later use the handle for 1397d1b069f5SRandy Dunlap different file system operations. This is useful in implementing 1398d1b069f5SRandy Dunlap userspace file servers, which now track files using handles instead 1399d1b069f5SRandy Dunlap of names. The handle would remain the same even if file names 1400d1b069f5SRandy Dunlap get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2) 1401d1b069f5SRandy Dunlap syscalls. 1402d1b069f5SRandy Dunlap 1403baa73d9eSNicolas Pitreconfig POSIX_TIMERS 1404baa73d9eSNicolas Pitre bool "Posix Clocks & timers" if EXPERT 1405baa73d9eSNicolas Pitre default y 1406baa73d9eSNicolas Pitre help 1407baa73d9eSNicolas Pitre This includes native support for POSIX timers to the kernel. 1408baa73d9eSNicolas Pitre Some embedded systems have no use for them and therefore they 1409baa73d9eSNicolas Pitre can be configured out to reduce the size of the kernel image. 1410baa73d9eSNicolas Pitre 1411baa73d9eSNicolas Pitre When this option is disabled, the following syscalls won't be 1412baa73d9eSNicolas Pitre available: timer_create, timer_gettime: timer_getoverrun, 1413baa73d9eSNicolas Pitre timer_settime, timer_delete, clock_adjtime, getitimer, 1414baa73d9eSNicolas Pitre setitimer, alarm. Furthermore, the clock_settime, clock_gettime, 1415baa73d9eSNicolas Pitre clock_getres and clock_nanosleep syscalls will be limited to 1416baa73d9eSNicolas Pitre CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only. 1417baa73d9eSNicolas Pitre 1418baa73d9eSNicolas Pitre If unsure say y. 1419baa73d9eSNicolas Pitre 1420d59745ceSMatt Mackallconfig PRINTK 1421d59745ceSMatt Mackall default y 14226a108a14SDavid Rientjes bool "Enable support for printk" if EXPERT 142374876a98SFrederic Weisbecker select IRQ_WORK 1424d59745ceSMatt Mackall help 1425d59745ceSMatt Mackall This option enables normal printk support. Removing it 1426d59745ceSMatt Mackall eliminates most of the message strings from the kernel image 1427d59745ceSMatt Mackall and makes the kernel more or less silent. As this makes it 1428d59745ceSMatt Mackall very difficult to diagnose system problems, saying N here is 1429d59745ceSMatt Mackall strongly discouraged. 1430d59745ceSMatt Mackall 143142a0bb3fSPetr Mladekconfig PRINTK_NMI 143242a0bb3fSPetr Mladek def_bool y 143342a0bb3fSPetr Mladek depends on PRINTK 143442a0bb3fSPetr Mladek depends on HAVE_NMI 143542a0bb3fSPetr Mladek 1436c8538a7aSMatt Mackallconfig BUG 14376a108a14SDavid Rientjes bool "BUG() support" if EXPERT 1438c8538a7aSMatt Mackall default y 1439c8538a7aSMatt Mackall help 1440c8538a7aSMatt Mackall Disabling this option eliminates support for BUG and WARN, reducing 1441c8538a7aSMatt Mackall the size of your kernel image and potentially quietly ignoring 1442c8538a7aSMatt Mackall numerous fatal conditions. You should only consider disabling this 1443c8538a7aSMatt Mackall option for embedded systems with no facilities for reporting errors. 1444c8538a7aSMatt Mackall Just say Y. 1445c8538a7aSMatt Mackall 1446708e9a79SMatt Mackallconfig ELF_CORE 1447046d662fSAlex Kelly depends on COREDUMP 1448708e9a79SMatt Mackall default y 14496a108a14SDavid Rientjes bool "Enable ELF core dumps" if EXPERT 1450708e9a79SMatt Mackall help 1451708e9a79SMatt Mackall Enable support for generating core dumps. Disabling saves about 4k. 1452708e9a79SMatt Mackall 14538761f1abSRalf Baechle 1454e5e1d3cbSStas Sergeevconfig PCSPKR_PLATFORM 14556a108a14SDavid Rientjes bool "Enable PC-Speaker support" if EXPERT 14568761f1abSRalf Baechle depends on HAVE_PCSPKR_PLATFORM 145715f304b6SRalf Baechle select I8253_LOCK 1458e5e1d3cbSStas Sergeev default y 1459e5e1d3cbSStas Sergeev help 1460e5e1d3cbSStas Sergeev This option allows to disable the internal PC-Speaker 1461e5e1d3cbSStas Sergeev support, saving some memory. 1462e5e1d3cbSStas Sergeev 14631da177e4SLinus Torvaldsconfig BASE_FULL 14641da177e4SLinus Torvalds default y 14656a108a14SDavid Rientjes bool "Enable full-sized data structures for core" if EXPERT 14661da177e4SLinus Torvalds help 14671da177e4SLinus Torvalds Disabling this option reduces the size of miscellaneous core 14681da177e4SLinus Torvalds kernel data structures. This saves memory on small machines, 14691da177e4SLinus Torvalds but may reduce performance. 14701da177e4SLinus Torvalds 14711da177e4SLinus Torvaldsconfig FUTEX 14726a108a14SDavid Rientjes bool "Enable futex support" if EXPERT 14731da177e4SLinus Torvalds default y 1474bc2eecd7SNicolas Pitre imply RT_MUTEXES 14751da177e4SLinus Torvalds help 14761da177e4SLinus Torvalds Disabling this option will cause the kernel to be built without 14771da177e4SLinus Torvalds support for "fast userspace mutexes". The resulting kernel may not 14781da177e4SLinus Torvalds run glibc-based applications correctly. 14791da177e4SLinus Torvalds 1480bc2eecd7SNicolas Pitreconfig FUTEX_PI 1481bc2eecd7SNicolas Pitre bool 1482bc2eecd7SNicolas Pitre depends on FUTEX && RT_MUTEXES 1483bc2eecd7SNicolas Pitre default y 1484bc2eecd7SNicolas Pitre 148503b8c7b6SHeiko Carstensconfig HAVE_FUTEX_CMPXCHG 148603b8c7b6SHeiko Carstens bool 148762b4d204SJosh Triplett depends on FUTEX 148803b8c7b6SHeiko Carstens help 148903b8c7b6SHeiko Carstens Architectures should select this if futex_atomic_cmpxchg_inatomic() 149003b8c7b6SHeiko Carstens is implemented and always working. This removes a couple of runtime 149103b8c7b6SHeiko Carstens checks. 149203b8c7b6SHeiko Carstens 14931da177e4SLinus Torvaldsconfig EPOLL 14946a108a14SDavid Rientjes bool "Enable eventpoll support" if EXPERT 14951da177e4SLinus Torvalds default y 14961da177e4SLinus Torvalds help 14971da177e4SLinus Torvalds Disabling this option will cause the kernel to be built without 14981da177e4SLinus Torvalds support for epoll family of system calls. 14991da177e4SLinus Torvalds 1500fba2afaaSDavide Libenziconfig SIGNALFD 15016a108a14SDavid Rientjes bool "Enable signalfd() system call" if EXPERT 1502fba2afaaSDavide Libenzi default y 1503fba2afaaSDavide Libenzi help 1504fba2afaaSDavide Libenzi Enable the signalfd() system call that allows to receive signals 1505fba2afaaSDavide Libenzi on a file descriptor. 1506fba2afaaSDavide Libenzi 1507fba2afaaSDavide Libenzi If unsure, say Y. 1508fba2afaaSDavide Libenzi 1509b215e283SDavide Libenziconfig TIMERFD 15106a108a14SDavid Rientjes bool "Enable timerfd() system call" if EXPERT 1511b215e283SDavide Libenzi default y 1512b215e283SDavide Libenzi help 1513b215e283SDavide Libenzi Enable the timerfd() system call that allows to receive timer 1514b215e283SDavide Libenzi events on a file descriptor. 1515b215e283SDavide Libenzi 1516b215e283SDavide Libenzi If unsure, say Y. 1517b215e283SDavide Libenzi 1518e1ad7468SDavide Libenziconfig EVENTFD 15196a108a14SDavid Rientjes bool "Enable eventfd() system call" if EXPERT 1520e1ad7468SDavide Libenzi default y 1521e1ad7468SDavide Libenzi help 1522e1ad7468SDavide Libenzi Enable the eventfd() system call that allows to receive both 1523e1ad7468SDavide Libenzi kernel notification (ie. KAIO) or userspace notifications. 1524e1ad7468SDavide Libenzi 1525e1ad7468SDavide Libenzi If unsure, say Y. 1526e1ad7468SDavide Libenzi 15271da177e4SLinus Torvaldsconfig SHMEM 15286a108a14SDavid Rientjes bool "Use full shmem filesystem" if EXPERT 15291da177e4SLinus Torvalds default y 15301da177e4SLinus Torvalds depends on MMU 15311da177e4SLinus Torvalds help 15321da177e4SLinus Torvalds The shmem is an internal filesystem used to manage shared memory. 15331da177e4SLinus Torvalds It is backed by swap and manages resource limits. It is also exported 15341da177e4SLinus Torvalds to userspace as tmpfs if TMPFS is enabled. Disabling this 15351da177e4SLinus Torvalds option replaces shmem and tmpfs with the much simpler ramfs code, 15361da177e4SLinus Torvalds which may be appropriate on small systems without swap. 15371da177e4SLinus Torvalds 1538ebf3f09cSThomas Petazzoniconfig AIO 15396a108a14SDavid Rientjes bool "Enable AIO support" if EXPERT 1540ebf3f09cSThomas Petazzoni default y 1541ebf3f09cSThomas Petazzoni help 1542ebf3f09cSThomas Petazzoni This option enables POSIX asynchronous I/O which may by used 1543ebf3f09cSThomas Petazzoni by some high performance threaded applications. Disabling 1544ebf3f09cSThomas Petazzoni this option saves about 7k. 1545ebf3f09cSThomas Petazzoni 15462b188cc1SJens Axboeconfig IO_URING 15472b188cc1SJens Axboe bool "Enable IO uring support" if EXPERT 1548561fb04aSJens Axboe select IO_WQ 15492b188cc1SJens Axboe default y 15502b188cc1SJens Axboe help 15512b188cc1SJens Axboe This option enables support for the io_uring interface, enabling 15522b188cc1SJens Axboe applications to submit and complete IO through submission and 15532b188cc1SJens Axboe completion rings that are shared between the kernel and application. 15542b188cc1SJens Axboe 1555d3ac21caSJosh Triplettconfig ADVISE_SYSCALLS 1556d3ac21caSJosh Triplett bool "Enable madvise/fadvise syscalls" if EXPERT 1557d3ac21caSJosh Triplett default y 1558d3ac21caSJosh Triplett help 1559d3ac21caSJosh Triplett This option enables the madvise and fadvise syscalls, used by 1560d3ac21caSJosh Triplett applications to advise the kernel about their future memory or file 1561d3ac21caSJosh Triplett usage, improving performance. If building an embedded system where no 1562d3ac21caSJosh Triplett applications use these syscalls, you can disable this option to save 1563d3ac21caSJosh Triplett space. 1564d3ac21caSJosh Triplett 15655a281062SAndrea Arcangeliconfig HAVE_ARCH_USERFAULTFD_WP 15665a281062SAndrea Arcangeli bool 15675a281062SAndrea Arcangeli help 15685a281062SAndrea Arcangeli Arch has userfaultfd write protection support 15695a281062SAndrea Arcangeli 15705b25b13aSMathieu Desnoyersconfig MEMBARRIER 15715b25b13aSMathieu Desnoyers bool "Enable membarrier() system call" if EXPERT 15725b25b13aSMathieu Desnoyers default y 15735b25b13aSMathieu Desnoyers help 15745b25b13aSMathieu Desnoyers Enable the membarrier() system call that allows issuing memory 15755b25b13aSMathieu Desnoyers barriers across all running threads, which can be used to distribute 15765b25b13aSMathieu Desnoyers the cost of user-space memory barriers asymmetrically by transforming 15775b25b13aSMathieu Desnoyers pairs of memory barriers into pairs consisting of membarrier() and a 15785b25b13aSMathieu Desnoyers compiler barrier. 15795b25b13aSMathieu Desnoyers 15805b25b13aSMathieu Desnoyers If unsure, say Y. 15815b25b13aSMathieu Desnoyers 1582d1b069f5SRandy Dunlapconfig KALLSYMS 1583d1b069f5SRandy Dunlap bool "Load all symbols for debugging/ksymoops" if EXPERT 1584d1b069f5SRandy Dunlap default y 1585d1b069f5SRandy Dunlap help 1586d1b069f5SRandy Dunlap Say Y here to let the kernel print out symbolic crash information and 1587d1b069f5SRandy Dunlap symbolic stack backtraces. This increases the size of the kernel 1588d1b069f5SRandy Dunlap somewhat, as all symbols have to be loaded into the kernel image. 1589d1b069f5SRandy Dunlap 1590d1b069f5SRandy Dunlapconfig KALLSYMS_ALL 1591d1b069f5SRandy Dunlap bool "Include all symbols in kallsyms" 1592d1b069f5SRandy Dunlap depends on DEBUG_KERNEL && KALLSYMS 1593d1b069f5SRandy Dunlap help 1594d1b069f5SRandy Dunlap Normally kallsyms only contains the symbols of functions for nicer 1595d1b069f5SRandy Dunlap OOPS messages and backtraces (i.e., symbols from the text and inittext 1596d1b069f5SRandy Dunlap sections). This is sufficient for most cases. And only in very rare 1597d1b069f5SRandy Dunlap cases (e.g., when a debugger is used) all symbols are required (e.g., 1598d1b069f5SRandy Dunlap names of variables from the data sections, etc). 1599d1b069f5SRandy Dunlap 1600d1b069f5SRandy Dunlap This option makes sure that all symbols are loaded into the kernel 1601d1b069f5SRandy Dunlap image (i.e., symbols from all sections) in cost of increased kernel 1602d1b069f5SRandy Dunlap size (depending on the kernel configuration, it may be 300KiB or 1603d1b069f5SRandy Dunlap something like this). 1604d1b069f5SRandy Dunlap 1605d1b069f5SRandy Dunlap Say N unless you really need all symbols. 1606d1b069f5SRandy Dunlap 1607d1b069f5SRandy Dunlapconfig KALLSYMS_ABSOLUTE_PERCPU 1608d1b069f5SRandy Dunlap bool 1609d1b069f5SRandy Dunlap depends on KALLSYMS 1610d1b069f5SRandy Dunlap default X86_64 && SMP 1611d1b069f5SRandy Dunlap 1612d1b069f5SRandy Dunlapconfig KALLSYMS_BASE_RELATIVE 1613d1b069f5SRandy Dunlap bool 1614d1b069f5SRandy Dunlap depends on KALLSYMS 1615a687a533SArnd Bergmann default !IA64 1616d1b069f5SRandy Dunlap help 1617d1b069f5SRandy Dunlap Instead of emitting them as absolute values in the native word size, 1618d1b069f5SRandy Dunlap emit the symbol references in the kallsyms table as 32-bit entries, 1619d1b069f5SRandy Dunlap each containing a relative value in the range [base, base + U32_MAX] 1620d1b069f5SRandy Dunlap or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either 1621d1b069f5SRandy Dunlap an absolute value in the range [0, S32_MAX] or a relative value in the 1622d1b069f5SRandy Dunlap range [base, base + S32_MAX], where base is the lowest relative symbol 1623d1b069f5SRandy Dunlap address encountered in the image. 1624d1b069f5SRandy Dunlap 1625d1b069f5SRandy Dunlap On 64-bit builds, this reduces the size of the address table by 50%, 1626d1b069f5SRandy Dunlap but more importantly, it results in entries whose values are build 1627d1b069f5SRandy Dunlap time constants, and no relocation pass is required at runtime to fix 1628d1b069f5SRandy Dunlap up the entries based on the runtime load address of the kernel. 1629d1b069f5SRandy Dunlap 1630d1b069f5SRandy Dunlap# end of the "standard kernel features (expert users)" menu 1631d1b069f5SRandy Dunlap 1632d1b069f5SRandy Dunlap# syscall, maps, verifier 1633fc611f47SKP Singh 1634fc611f47SKP Singhconfig BPF_LSM 1635fc611f47SKP Singh bool "LSM Instrumentation with BPF" 16364edf16b7SKP Singh depends on BPF_EVENTS 1637fc611f47SKP Singh depends on BPF_SYSCALL 1638fc611f47SKP Singh depends on SECURITY 1639fc611f47SKP Singh depends on BPF_JIT 1640fc611f47SKP Singh help 1641fc611f47SKP Singh Enables instrumentation of the security hooks with eBPF programs for 1642fc611f47SKP Singh implementing dynamic MAC and Audit Policies. 1643fc611f47SKP Singh 1644fc611f47SKP Singh If you are unsure how to answer this question, answer N. 1645fc611f47SKP Singh 1646d1b069f5SRandy Dunlapconfig BPF_SYSCALL 1647d1b069f5SRandy Dunlap bool "Enable bpf() system call" 1648d1b069f5SRandy Dunlap select BPF 1649bae77c5eSSong Liu select IRQ_WORK 1650d1b069f5SRandy Dunlap default n 1651d1b069f5SRandy Dunlap help 1652d1b069f5SRandy Dunlap Enable the bpf() system call that allows to manipulate eBPF 1653d1b069f5SRandy Dunlap programs and maps via file descriptors. 1654d1b069f5SRandy Dunlap 165581c22041SDaniel Borkmannconfig ARCH_WANT_DEFAULT_BPF_JIT 165681c22041SDaniel Borkmann bool 165781c22041SDaniel Borkmann 1658290af866SAlexei Starovoitovconfig BPF_JIT_ALWAYS_ON 1659290af866SAlexei Starovoitov bool "Permanently enable BPF JIT and remove BPF interpreter" 1660290af866SAlexei Starovoitov depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT 1661290af866SAlexei Starovoitov help 1662290af866SAlexei Starovoitov Enables BPF JIT and removes BPF interpreter to avoid 1663290af866SAlexei Starovoitov speculative execution of BPF instructions by the interpreter 1664290af866SAlexei Starovoitov 166581c22041SDaniel Borkmannconfig BPF_JIT_DEFAULT_ON 166681c22041SDaniel Borkmann def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON 166781c22041SDaniel Borkmann depends on HAVE_EBPF_JIT && BPF_JIT 166881c22041SDaniel Borkmann 1669d1b069f5SRandy Dunlapconfig USERFAULTFD 1670d1b069f5SRandy Dunlap bool "Enable userfaultfd() system call" 1671d1b069f5SRandy Dunlap depends on MMU 1672d1b069f5SRandy Dunlap help 1673d1b069f5SRandy Dunlap Enable the userfaultfd() system call that allows to intercept and 1674d1b069f5SRandy Dunlap handle page faults in userland. 1675d1b069f5SRandy Dunlap 16763ccfebedSMathieu Desnoyersconfig ARCH_HAS_MEMBARRIER_CALLBACKS 16773ccfebedSMathieu Desnoyers bool 16783ccfebedSMathieu Desnoyers 167970216e18SMathieu Desnoyersconfig ARCH_HAS_MEMBARRIER_SYNC_CORE 168070216e18SMathieu Desnoyers bool 168170216e18SMathieu Desnoyers 1682d7822b1eSMathieu Desnoyersconfig RSEQ 1683d7822b1eSMathieu Desnoyers bool "Enable rseq() system call" if EXPERT 1684d7822b1eSMathieu Desnoyers default y 1685d7822b1eSMathieu Desnoyers depends on HAVE_RSEQ 1686d7822b1eSMathieu Desnoyers select MEMBARRIER 1687d7822b1eSMathieu Desnoyers help 1688d7822b1eSMathieu Desnoyers Enable the restartable sequences system call. It provides a 1689d7822b1eSMathieu Desnoyers user-space cache for the current CPU number value, which 1690d7822b1eSMathieu Desnoyers speeds up getting the current CPU number from user-space, 1691d7822b1eSMathieu Desnoyers as well as an ABI to speed up user-space operations on 1692d7822b1eSMathieu Desnoyers per-CPU data. 1693d7822b1eSMathieu Desnoyers 1694d7822b1eSMathieu Desnoyers If unsure, say Y. 1695d7822b1eSMathieu Desnoyers 1696d7822b1eSMathieu Desnoyersconfig DEBUG_RSEQ 1697d7822b1eSMathieu Desnoyers default n 1698d7822b1eSMathieu Desnoyers bool "Enabled debugging of rseq() system call" if EXPERT 1699d7822b1eSMathieu Desnoyers depends on RSEQ && DEBUG_KERNEL 1700d7822b1eSMathieu Desnoyers help 1701d7822b1eSMathieu Desnoyers Enable extra debugging checks for the rseq system call. 1702d7822b1eSMathieu Desnoyers 1703d7822b1eSMathieu Desnoyers If unsure, say N. 1704d7822b1eSMathieu Desnoyers 17056befe5f6SRandy Dunlapconfig EMBEDDED 17066befe5f6SRandy Dunlap bool "Embedded system" 17075d2acfc7SJosh Triplett option allnoconfig_y 17086befe5f6SRandy Dunlap select EXPERT 17096befe5f6SRandy Dunlap help 17106befe5f6SRandy Dunlap This option should be enabled if compiling the kernel for 17116befe5f6SRandy Dunlap an embedded system so certain expert options are available 17126befe5f6SRandy Dunlap for configuration. 17136befe5f6SRandy Dunlap 1714cdd6c482SIngo Molnarconfig HAVE_PERF_EVENTS 17150793a61dSThomas Gleixner bool 1716018df72dSMike Frysinger help 1717018df72dSMike Frysinger See tools/perf/design.txt for details. 17180793a61dSThomas Gleixner 1719906010b2SPeter Zijlstraconfig PERF_USE_VMALLOC 1720906010b2SPeter Zijlstra bool 1721906010b2SPeter Zijlstra help 1722906010b2SPeter Zijlstra See tools/perf/design.txt for details 1723906010b2SPeter Zijlstra 1724ad90a3deSWilliam Breathitt Grayconfig PC104 1725424529fbSWilliam Breathitt Gray bool "PC/104 support" if EXPERT 1726ad90a3deSWilliam Breathitt Gray help 1727ad90a3deSWilliam Breathitt Gray Expose PC/104 form factor device drivers and options available for 1728ad90a3deSWilliam Breathitt Gray selection and configuration. Enable this option if your target 1729ad90a3deSWilliam Breathitt Gray machine has a PC/104 bus. 1730ad90a3deSWilliam Breathitt Gray 173157c0c15bSIngo Molnarmenu "Kernel Performance Events And Counters" 17320793a61dSThomas Gleixner 1733cdd6c482SIngo Molnarconfig PERF_EVENTS 173457c0c15bSIngo Molnar bool "Kernel performance events and counters" 1735392d65a9SRobert Richter default y if PROFILING 1736cdd6c482SIngo Molnar depends on HAVE_PERF_EVENTS 1737e360adbeSPeter Zijlstra select IRQ_WORK 173883fe27eaSPranith Kumar select SRCU 17390793a61dSThomas Gleixner help 174057c0c15bSIngo Molnar Enable kernel support for various performance events provided 174157c0c15bSIngo Molnar by software and hardware. 17420793a61dSThomas Gleixner 1743dd77038dSThadeu Lima de Souza Cascardo Software events are supported either built-in or via the 174457c0c15bSIngo Molnar use of generic tracepoints. 174557c0c15bSIngo Molnar 174657c0c15bSIngo Molnar Most modern CPUs support performance events via performance 174757c0c15bSIngo Molnar counter registers. These registers count the number of certain 17480793a61dSThomas Gleixner types of hw events: such as instructions executed, cachemisses 17490793a61dSThomas Gleixner suffered, or branches mis-predicted - without slowing down the 17500793a61dSThomas Gleixner kernel or applications. These registers can also trigger interrupts 17510793a61dSThomas Gleixner when a threshold number of events have passed - and can thus be 17520793a61dSThomas Gleixner used to profile the code that runs on that CPU. 17530793a61dSThomas Gleixner 175457c0c15bSIngo Molnar The Linux Performance Event subsystem provides an abstraction of 1755dd77038dSThadeu Lima de Souza Cascardo these software and hardware event capabilities, available via a 175657c0c15bSIngo Molnar system call and used by the "perf" utility in tools/perf/. It 17570793a61dSThomas Gleixner provides per task and per CPU counters, and it provides event 17580793a61dSThomas Gleixner capabilities on top of those. 17590793a61dSThomas Gleixner 17600793a61dSThomas Gleixner Say Y if unsure. 17610793a61dSThomas Gleixner 1762906010b2SPeter Zijlstraconfig DEBUG_PERF_USE_VMALLOC 1763906010b2SPeter Zijlstra default n 1764906010b2SPeter Zijlstra bool "Debug: use vmalloc to back perf mmap() buffers" 1765cb307113SMichael Ellerman depends on PERF_EVENTS && DEBUG_KERNEL && !PPC 1766906010b2SPeter Zijlstra select PERF_USE_VMALLOC 1767906010b2SPeter Zijlstra help 1768906010b2SPeter Zijlstra Use vmalloc memory to back perf mmap() buffers. 1769906010b2SPeter Zijlstra 1770906010b2SPeter Zijlstra Mostly useful for debugging the vmalloc code on platforms 1771906010b2SPeter Zijlstra that don't require it. 1772906010b2SPeter Zijlstra 1773906010b2SPeter Zijlstra Say N if unsure. 1774906010b2SPeter Zijlstra 17750793a61dSThomas Gleixnerendmenu 17760793a61dSThomas Gleixner 1777f8891e5eSChristoph Lameterconfig VM_EVENT_COUNTERS 1778f8891e5eSChristoph Lameter default y 17796a108a14SDavid Rientjes bool "Enable VM event counters for /proc/vmstat" if EXPERT 1780f8891e5eSChristoph Lameter help 17812aea4fb6SPaul Jackson VM event counters are needed for event counts to be shown. 17822aea4fb6SPaul Jackson This option allows the disabling of the VM event counters 17836a108a14SDavid Rientjes on EXPERT systems. /proc/vmstat will only show page counts 17842aea4fb6SPaul Jackson if VM event counters are disabled. 1785f8891e5eSChristoph Lameter 178641ecc55bSChristoph Lameterconfig SLUB_DEBUG 178741ecc55bSChristoph Lameter default y 17886a108a14SDavid Rientjes bool "Enable SLUB debugging support" if EXPERT 1789f6acb635SChristoph Lameter depends on SLUB && SYSFS 179041ecc55bSChristoph Lameter help 179141ecc55bSChristoph Lameter SLUB has extensive debug support features. Disabling these can 179241ecc55bSChristoph Lameter result in significant savings in code size. This also disables 179341ecc55bSChristoph Lameter SLUB sysfs support. /sys/slab will not exist and there will be 179441ecc55bSChristoph Lameter no support for cache validation etc. 179541ecc55bSChristoph Lameter 17961663f26dSTejun Heoconfig SLUB_MEMCG_SYSFS_ON 17971663f26dSTejun Heo default n 17981663f26dSTejun Heo bool "Enable memcg SLUB sysfs support by default" if EXPERT 17991663f26dSTejun Heo depends on SLUB && SYSFS && MEMCG 18001663f26dSTejun Heo help 18011663f26dSTejun Heo SLUB creates a directory under /sys/kernel/slab for each 18021663f26dSTejun Heo allocation cache to host info and debug files. If memory 18031663f26dSTejun Heo cgroup is enabled, each cache can have per memory cgroup 18041663f26dSTejun Heo caches. SLUB can create the same sysfs directories for these 18051663f26dSTejun Heo caches under /sys/kernel/slab/CACHE/cgroup but it can lead 18061663f26dSTejun Heo to a very high number of debug files being created. This is 18071663f26dSTejun Heo controlled by slub_memcg_sysfs boot parameter and this 18081663f26dSTejun Heo config option determines the parameter's default value. 18091663f26dSTejun Heo 1810b943c460SRandy Dunlapconfig COMPAT_BRK 1811b943c460SRandy Dunlap bool "Disable heap randomization" 1812b943c460SRandy Dunlap default y 1813b943c460SRandy Dunlap help 1814b943c460SRandy Dunlap Randomizing heap placement makes heap exploits harder, but it 1815b943c460SRandy Dunlap also breaks ancient binaries (including anything libc5 based). 1816b943c460SRandy Dunlap This option changes the bootup default to heap randomization 1817692105b8SMatt LaPlante disabled, and can be overridden at runtime by setting 1818b943c460SRandy Dunlap /proc/sys/kernel/randomize_va_space to 2. 1819b943c460SRandy Dunlap 1820b943c460SRandy Dunlap On non-ancient distros (post-2000 ones) N is usually a safe choice. 1821b943c460SRandy Dunlap 182281819f0fSChristoph Lameterchoice 182381819f0fSChristoph Lameter prompt "Choose SLAB allocator" 1824a0acd820SChristoph Lameter default SLUB 182581819f0fSChristoph Lameter help 182681819f0fSChristoph Lameter This option allows to select a slab allocator. 182781819f0fSChristoph Lameter 182881819f0fSChristoph Lameterconfig SLAB 182981819f0fSChristoph Lameter bool "SLAB" 183004385fc5SKees Cook select HAVE_HARDENED_USERCOPY_ALLOCATOR 183181819f0fSChristoph Lameter help 183281819f0fSChristoph Lameter The regular slab allocator that is established and known to work 183334013886SChristoph Lameter well in all environments. It organizes cache hot objects in 183402f56210SSimon Arlott per cpu and per node queues. 183581819f0fSChristoph Lameter 183681819f0fSChristoph Lameterconfig SLUB 183781819f0fSChristoph Lameter bool "SLUB (Unqueued Allocator)" 1838ed18adc1SKees Cook select HAVE_HARDENED_USERCOPY_ALLOCATOR 183981819f0fSChristoph Lameter help 184081819f0fSChristoph Lameter SLUB is a slab allocator that minimizes cache line usage 184181819f0fSChristoph Lameter instead of managing queues of cached objects (SLAB approach). 184281819f0fSChristoph Lameter Per cpu caching is realized using slabs of objects instead 184381819f0fSChristoph Lameter of queues of objects. SLUB can use memory efficiently 184402f56210SSimon Arlott and has enhanced diagnostics. SLUB is the default choice for 184502f56210SSimon Arlott a slab allocator. 184681819f0fSChristoph Lameter 184781819f0fSChristoph Lameterconfig SLOB 18486a108a14SDavid Rientjes depends on EXPERT 184981819f0fSChristoph Lameter bool "SLOB (Simple Allocator)" 185081819f0fSChristoph Lameter help 185137291458SMatt Mackall SLOB replaces the stock allocator with a drastically simpler 185237291458SMatt Mackall allocator. SLOB is generally more space efficient but 185337291458SMatt Mackall does not perform as well on large systems. 185481819f0fSChristoph Lameter 185581819f0fSChristoph Lameterendchoice 185681819f0fSChristoph Lameter 18577660a6fdSKees Cookconfig SLAB_MERGE_DEFAULT 18587660a6fdSKees Cook bool "Allow slab caches to be merged" 18597660a6fdSKees Cook default y 18607660a6fdSKees Cook help 18617660a6fdSKees Cook For reduced kernel memory fragmentation, slab caches can be 18627660a6fdSKees Cook merged when they share the same size and other characteristics. 18637660a6fdSKees Cook This carries a risk of kernel heap overflows being able to 18647660a6fdSKees Cook overwrite objects from merged caches (and more easily control 18657660a6fdSKees Cook cache layout), which makes such heap attacks easier to exploit 18667660a6fdSKees Cook by attackers. By keeping caches unmerged, these kinds of exploits 18677660a6fdSKees Cook can usually only damage objects in the same cache. To disable 18687660a6fdSKees Cook merging at runtime, "slab_nomerge" can be passed on the kernel 18697660a6fdSKees Cook command line. 18707660a6fdSKees Cook 1871c7ce4f60SThomas Garnierconfig SLAB_FREELIST_RANDOM 1872c7ce4f60SThomas Garnier default n 1873210e7a43SThomas Garnier depends on SLAB || SLUB 1874c7ce4f60SThomas Garnier bool "SLAB freelist randomization" 1875c7ce4f60SThomas Garnier help 1876210e7a43SThomas Garnier Randomizes the freelist order used on creating new pages. This 1877c7ce4f60SThomas Garnier security feature reduces the predictability of the kernel slab 1878c7ce4f60SThomas Garnier allocator against heap overflows. 1879c7ce4f60SThomas Garnier 18802482ddecSKees Cookconfig SLAB_FREELIST_HARDENED 18812482ddecSKees Cook bool "Harden slab freelist metadata" 18822482ddecSKees Cook depends on SLUB 18832482ddecSKees Cook help 18842482ddecSKees Cook Many kernel heap attacks try to target slab cache metadata and 18852482ddecSKees Cook other infrastructure. This options makes minor performance 188692bae787SKees Cook sacrifices to harden the kernel slab allocator against common 18872482ddecSKees Cook freelist exploit methods. 18882482ddecSKees Cook 1889e900a918SDan Williamsconfig SHUFFLE_PAGE_ALLOCATOR 1890e900a918SDan Williams bool "Page allocator randomization" 1891e900a918SDan Williams default SLAB_FREELIST_RANDOM && ACPI_NUMA 1892e900a918SDan Williams help 1893e900a918SDan Williams Randomization of the page allocator improves the average 1894e900a918SDan Williams utilization of a direct-mapped memory-side-cache. See section 1895e900a918SDan Williams 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI 1896e900a918SDan Williams 6.2a specification for an example of how a platform advertises 1897e900a918SDan Williams the presence of a memory-side-cache. There are also incidental 1898e900a918SDan Williams security benefits as it reduces the predictability of page 1899e900a918SDan Williams allocations to compliment SLAB_FREELIST_RANDOM, but the 1900e900a918SDan Williams default granularity of shuffling on the "MAX_ORDER - 1" i.e, 1901e900a918SDan Williams 10th order of pages is selected based on cache utilization 1902e900a918SDan Williams benefits on x86. 1903e900a918SDan Williams 1904e900a918SDan Williams While the randomization improves cache utilization it may 1905e900a918SDan Williams negatively impact workloads on platforms without a cache. For 1906e900a918SDan Williams this reason, by default, the randomization is enabled only 1907e900a918SDan Williams after runtime detection of a direct-mapped memory-side-cache. 1908e900a918SDan Williams Otherwise, the randomization may be force enabled with the 1909e900a918SDan Williams 'page_alloc.shuffle' kernel command line parameter. 1910e900a918SDan Williams 1911e900a918SDan Williams Say Y if unsure. 1912e900a918SDan Williams 1913345c905dSJoonsoo Kimconfig SLUB_CPU_PARTIAL 1914345c905dSJoonsoo Kim default y 1915b39ffbf8SUwe Kleine-König depends on SLUB && SMP 1916345c905dSJoonsoo Kim bool "SLUB per cpu partial cache" 1917345c905dSJoonsoo Kim help 191892bae787SKees Cook Per cpu partial caches accelerate objects allocation and freeing 1919345c905dSJoonsoo Kim that is local to a processor at the price of more indeterminism 1920345c905dSJoonsoo Kim in the latency of the free. On overflow these caches will be cleared 1921345c905dSJoonsoo Kim which requires the taking of locks that may cause latency spikes. 1922345c905dSJoonsoo Kim Typically one would choose no for a realtime system. 1923345c905dSJoonsoo Kim 1924ea637639SJie Zhangconfig MMAP_ALLOW_UNINITIALIZED 1925ea637639SJie Zhang bool "Allow mmapped anonymous memory to be uninitialized" 19266a108a14SDavid Rientjes depends on EXPERT && !MMU 1927ea637639SJie Zhang default n 1928ea637639SJie Zhang help 1929ea637639SJie Zhang Normally, and according to the Linux spec, anonymous memory obtained 19303903bf94SRandy Dunlap from mmap() has its contents cleared before it is passed to 1931ea637639SJie Zhang userspace. Enabling this config option allows you to request that 1932ea637639SJie Zhang mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus 1933ea637639SJie Zhang providing a huge performance boost. If this option is not enabled, 1934ea637639SJie Zhang then the flag will be ignored. 1935ea637639SJie Zhang 1936ea637639SJie Zhang This is taken advantage of by uClibc's malloc(), and also by 1937ea637639SJie Zhang ELF-FDPIC binfmt's brk and stack allocator. 1938ea637639SJie Zhang 1939ea637639SJie Zhang Because of the obvious security issues, this option should only be 1940ea637639SJie Zhang enabled on embedded devices where you control what is run in 1941ea637639SJie Zhang userspace. Since that isn't generally a problem on no-MMU systems, 1942ea637639SJie Zhang it is normally safe to say Y here. 1943ea637639SJie Zhang 1944ea637639SJie Zhang See Documentation/nommu-mmap.txt for more information. 1945ea637639SJie Zhang 1946091f6e26SDavid Howellsconfig SYSTEM_DATA_VERIFICATION 1947091f6e26SDavid Howells def_bool n 1948091f6e26SDavid Howells select SYSTEM_TRUSTED_KEYRING 1949091f6e26SDavid Howells select KEYS 1950091f6e26SDavid Howells select CRYPTO 1951d43de6c7SDavid Howells select CRYPTO_RSA 1952091f6e26SDavid Howells select ASYMMETRIC_KEY_TYPE 1953091f6e26SDavid Howells select ASYMMETRIC_PUBLIC_KEY_SUBTYPE 1954091f6e26SDavid Howells select ASN1 1955091f6e26SDavid Howells select OID_REGISTRY 1956091f6e26SDavid Howells select X509_CERTIFICATE_PARSER 1957091f6e26SDavid Howells select PKCS7_MESSAGE_PARSER 195882c04ff8SPeter Foley help 1959091f6e26SDavid Howells Provide PKCS#7 message verification using the contents of the system 1960091f6e26SDavid Howells trusted keyring to provide public keys. This then can be used for 1961091f6e26SDavid Howells module verification, kexec image verification and firmware blob 1962091f6e26SDavid Howells verification. 196382c04ff8SPeter Foley 1964125e5645SMathieu Desnoyersconfig PROFILING 1965b309a294SRobert Richter bool "Profiling support" 1966125e5645SMathieu Desnoyers help 1967125e5645SMathieu Desnoyers Say Y here to enable the extended profiling support mechanisms used 1968125e5645SMathieu Desnoyers by profilers such as OProfile. 1969125e5645SMathieu Desnoyers 19705f87f112SIngo Molnar# 19715f87f112SIngo Molnar# Place an empty function call at each tracepoint site. Can be 19725f87f112SIngo Molnar# dynamically changed for a probe function. 19735f87f112SIngo Molnar# 197497e1c18eSMathieu Desnoyersconfig TRACEPOINTS 19755f87f112SIngo Molnar bool 197697e1c18eSMathieu Desnoyers 19771da177e4SLinus Torvaldsendmenu # General setup 19781da177e4SLinus Torvalds 19791572497cSChristoph Hellwigsource "arch/Kconfig" 19801572497cSChristoph Hellwig 1981ae81f9e3SChuck Ebbertconfig RT_MUTEXES 19826341e62bSChristoph Jaeger bool 1983ae81f9e3SChuck Ebbert 19841da177e4SLinus Torvaldsconfig BASE_SMALL 19851da177e4SLinus Torvalds int 19861da177e4SLinus Torvalds default 0 if BASE_FULL 19871da177e4SLinus Torvalds default 1 if !BASE_FULL 19881da177e4SLinus Torvalds 1989c8424e77SThiago Jung Bauermannconfig MODULE_SIG_FORMAT 1990c8424e77SThiago Jung Bauermann def_bool n 1991c8424e77SThiago Jung Bauermann select SYSTEM_DATA_VERIFICATION 1992c8424e77SThiago Jung Bauermann 199366da5733SJan Engelhardtmenuconfig MODULES 19941da177e4SLinus Torvalds bool "Enable loadable module support" 199511097a03SYann E. MORIN option modules 19961da177e4SLinus Torvalds help 19971da177e4SLinus Torvalds Kernel modules are small pieces of compiled code which can 19981da177e4SLinus Torvalds be inserted in the running kernel, rather than being 19991da177e4SLinus Torvalds permanently built into the kernel. You use the "modprobe" 20001da177e4SLinus Torvalds tool to add (and sometimes remove) them. If you say Y here, 20011da177e4SLinus Torvalds many parts of the kernel can be built as modules (by 20021da177e4SLinus Torvalds answering M instead of Y where indicated): this is most 20031da177e4SLinus Torvalds useful for infrequently used options which are not required 20041da177e4SLinus Torvalds for booting. For more information, see the man pages for 20051da177e4SLinus Torvalds modprobe, lsmod, modinfo, insmod and rmmod. 20061da177e4SLinus Torvalds 20071da177e4SLinus Torvalds If you say Y here, you will need to run "make 20081da177e4SLinus Torvalds modules_install" to put the modules under /lib/modules/ 20091da177e4SLinus Torvalds where modprobe can find them (you may need to be root to do 20101da177e4SLinus Torvalds this). 20111da177e4SLinus Torvalds 20121da177e4SLinus Torvalds If unsure, say Y. 20131da177e4SLinus Torvalds 20140b0de144SRobert P. J. Dayif MODULES 20150b0de144SRobert P. J. Day 2016826e4506SLinus Torvaldsconfig MODULE_FORCE_LOAD 2017826e4506SLinus Torvalds bool "Forced module loading" 2018826e4506SLinus Torvalds default n 2019826e4506SLinus Torvalds help 202091e37a79SRusty Russell Allow loading of modules without version information (ie. modprobe 202191e37a79SRusty Russell --force). Forced module loading sets the 'F' (forced) taint flag and 202291e37a79SRusty Russell is usually a really bad idea. 2023826e4506SLinus Torvalds 20241da177e4SLinus Torvaldsconfig MODULE_UNLOAD 20251da177e4SLinus Torvalds bool "Module unloading" 20261da177e4SLinus Torvalds help 20271da177e4SLinus Torvalds Without this option you will not be able to unload any 20281da177e4SLinus Torvalds modules (note that some modules may not be unloadable 2029f7f5b675SDenys Vlasenko anyway), which makes your kernel smaller, faster 2030f7f5b675SDenys Vlasenko and simpler. If unsure, say Y. 20311da177e4SLinus Torvalds 20321da177e4SLinus Torvaldsconfig MODULE_FORCE_UNLOAD 20331da177e4SLinus Torvalds bool "Forced module unloading" 203419c92399SKees Cook depends on MODULE_UNLOAD 20351da177e4SLinus Torvalds help 20361da177e4SLinus Torvalds This option allows you to force a module to unload, even if the 20371da177e4SLinus Torvalds kernel believes it is unsafe: the kernel will remove the module 20381da177e4SLinus Torvalds without waiting for anyone to stop using it (using the -f option to 20391da177e4SLinus Torvalds rmmod). This is mainly for kernel developers and desperate users. 20401da177e4SLinus Torvalds If unsure, say N. 20411da177e4SLinus Torvalds 20421da177e4SLinus Torvaldsconfig MODVERSIONS 20430d541643SSam Ravnborg bool "Module versioning support" 20441da177e4SLinus Torvalds help 20451da177e4SLinus Torvalds Usually, you have to use modules compiled with your kernel. 20461da177e4SLinus Torvalds Saying Y here makes it sometimes possible to use modules 20471da177e4SLinus Torvalds compiled for different kernels, by adding enough information 20481da177e4SLinus Torvalds to the modules to (hopefully) spot any changes which would 20491da177e4SLinus Torvalds make them incompatible with the kernel you are running. If 20501da177e4SLinus Torvalds unsure, say N. 20511da177e4SLinus Torvalds 20522ff2b7ecSMasahiro Yamadaconfig ASM_MODVERSIONS 20532ff2b7ecSMasahiro Yamada bool 20542ff2b7ecSMasahiro Yamada default HAVE_ASM_MODVERSIONS && MODVERSIONS 20552ff2b7ecSMasahiro Yamada help 20562ff2b7ecSMasahiro Yamada This enables module versioning for exported symbols also from 20572ff2b7ecSMasahiro Yamada assembly. This can be enabled only when the target architecture 20582ff2b7ecSMasahiro Yamada supports it. 20592ff2b7ecSMasahiro Yamada 206056067812SArd Biesheuvelconfig MODULE_REL_CRCS 206156067812SArd Biesheuvel bool 206256067812SArd Biesheuvel depends on MODVERSIONS 206356067812SArd Biesheuvel 20641da177e4SLinus Torvaldsconfig MODULE_SRCVERSION_ALL 20651da177e4SLinus Torvalds bool "Source checksum for all modules" 20661da177e4SLinus Torvalds help 20671da177e4SLinus Torvalds Modules which contain a MODULE_VERSION get an extra "srcversion" 20681da177e4SLinus Torvalds field inserted into their modinfo section, which contains a 20691da177e4SLinus Torvalds sum of the source files which made it. This helps maintainers 20701da177e4SLinus Torvalds see exactly which source was used to build a module (since 20711da177e4SLinus Torvalds others sometimes change the module source without updating 20721da177e4SLinus Torvalds the version). With this option, such a "srcversion" field 20731da177e4SLinus Torvalds will be created for all modules. If unsure, say N. 20741da177e4SLinus Torvalds 2075106a4ee2SRusty Russellconfig MODULE_SIG 2076106a4ee2SRusty Russell bool "Module signature verification" 2077c8424e77SThiago Jung Bauermann select MODULE_SIG_FORMAT 2078106a4ee2SRusty Russell help 2079106a4ee2SRusty Russell Check modules for valid signatures upon load: the signature 2080106a4ee2SRusty Russell is simply appended to the module. For more information see 2081cbdc8217SNathan Chancellor <file:Documentation/admin-guide/module-signing.rst>. 2082106a4ee2SRusty Russell 2083228c37ffSDavid Howells Note that this option adds the OpenSSL development packages as a 2084228c37ffSDavid Howells kernel build dependency so that the signing tool can use its crypto 2085228c37ffSDavid Howells library. 2086228c37ffSDavid Howells 208749fcf732SDavid Howells You should enable this option if you wish to use either 208849fcf732SDavid Howells CONFIG_SECURITY_LOCKDOWN_LSM or lockdown functionality imposed via 208949fcf732SDavid Howells another LSM - otherwise unsigned modules will be loadable regardless 209049fcf732SDavid Howells of the lockdown policy. 209149fcf732SDavid Howells 2092ea0b6dcfSDavid Howells !!!WARNING!!! If you enable this option, you MUST make sure that the 2093ea0b6dcfSDavid Howells module DOES NOT get stripped after being signed. This includes the 2094ea0b6dcfSDavid Howells debuginfo strip done by some packagers (such as rpmbuild) and 2095ea0b6dcfSDavid Howells inclusion into an initramfs that wants the module size reduced. 2096ea0b6dcfSDavid Howells 2097106a4ee2SRusty Russellconfig MODULE_SIG_FORCE 2098106a4ee2SRusty Russell bool "Require modules to be validly signed" 2099106a4ee2SRusty Russell depends on MODULE_SIG 2100106a4ee2SRusty Russell help 2101106a4ee2SRusty Russell Reject unsigned modules or signed modules for which we don't have a 2102106a4ee2SRusty Russell key. Without this, such modules will simply taint the kernel. 2103ea0b6dcfSDavid Howells 2104d9d8d7edSMichal Marekconfig MODULE_SIG_ALL 2105d9d8d7edSMichal Marek bool "Automatically sign all modules" 2106d9d8d7edSMichal Marek default y 2107d9d8d7edSMichal Marek depends on MODULE_SIG 2108d9d8d7edSMichal Marek help 2109d9d8d7edSMichal Marek Sign all modules during make modules_install. Without this option, 2110d9d8d7edSMichal Marek modules must be signed manually, using the scripts/sign-file tool. 2111d9d8d7edSMichal Marek 2112d9d8d7edSMichal Marekcomment "Do not forget to sign required modules with scripts/sign-file" 2113d9d8d7edSMichal Marek depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL 2114d9d8d7edSMichal Marek 2115ea0b6dcfSDavid Howellschoice 2116ea0b6dcfSDavid Howells prompt "Which hash algorithm should modules be signed with?" 2117ea0b6dcfSDavid Howells depends on MODULE_SIG 2118ea0b6dcfSDavid Howells help 2119ea0b6dcfSDavid Howells This determines which sort of hashing algorithm will be used during 2120ea0b6dcfSDavid Howells signature generation. This algorithm _must_ be built into the kernel 2121ea0b6dcfSDavid Howells directly so that signature verification can take place. It is not 2122ea0b6dcfSDavid Howells possible to load a signed module containing the algorithm to check 2123ea0b6dcfSDavid Howells the signature on that module. 2124ea0b6dcfSDavid Howells 2125ea0b6dcfSDavid Howellsconfig MODULE_SIG_SHA1 2126ea0b6dcfSDavid Howells bool "Sign modules with SHA-1" 2127ea0b6dcfSDavid Howells select CRYPTO_SHA1 2128ea0b6dcfSDavid Howells 2129ea0b6dcfSDavid Howellsconfig MODULE_SIG_SHA224 2130ea0b6dcfSDavid Howells bool "Sign modules with SHA-224" 2131ea0b6dcfSDavid Howells select CRYPTO_SHA256 2132ea0b6dcfSDavid Howells 2133ea0b6dcfSDavid Howellsconfig MODULE_SIG_SHA256 2134ea0b6dcfSDavid Howells bool "Sign modules with SHA-256" 2135ea0b6dcfSDavid Howells select CRYPTO_SHA256 2136ea0b6dcfSDavid Howells 2137ea0b6dcfSDavid Howellsconfig MODULE_SIG_SHA384 2138ea0b6dcfSDavid Howells bool "Sign modules with SHA-384" 2139ea0b6dcfSDavid Howells select CRYPTO_SHA512 2140ea0b6dcfSDavid Howells 2141ea0b6dcfSDavid Howellsconfig MODULE_SIG_SHA512 2142ea0b6dcfSDavid Howells bool "Sign modules with SHA-512" 2143ea0b6dcfSDavid Howells select CRYPTO_SHA512 2144ea0b6dcfSDavid Howells 2145ea0b6dcfSDavid Howellsendchoice 2146ea0b6dcfSDavid Howells 214722753674SMichal Marekconfig MODULE_SIG_HASH 214822753674SMichal Marek string 214922753674SMichal Marek depends on MODULE_SIG 215022753674SMichal Marek default "sha1" if MODULE_SIG_SHA1 215122753674SMichal Marek default "sha224" if MODULE_SIG_SHA224 215222753674SMichal Marek default "sha256" if MODULE_SIG_SHA256 215322753674SMichal Marek default "sha384" if MODULE_SIG_SHA384 215422753674SMichal Marek default "sha512" if MODULE_SIG_SHA512 215522753674SMichal Marek 2156beb50df3SBertrand Jacquinconfig MODULE_COMPRESS 2157beb50df3SBertrand Jacquin bool "Compress modules on installation" 2158beb50df3SBertrand Jacquin help 2159beb50df3SBertrand Jacquin 2160b6c09b51SRusty Russell Compresses kernel modules when 'make modules_install' is run; gzip or 2161b6c09b51SRusty Russell xz depending on "Compression algorithm" below. 2162beb50df3SBertrand Jacquin 2163b6c09b51SRusty Russell module-init-tools MAY support gzip, and kmod MAY support gzip and xz. 2164beb50df3SBertrand Jacquin 2165b6c09b51SRusty Russell Out-of-tree kernel modules installed using Kbuild will also be 2166b6c09b51SRusty Russell compressed upon installation. 2167beb50df3SBertrand Jacquin 2168b6c09b51SRusty Russell Note: for modules inside an initrd or initramfs, it's more efficient 2169b6c09b51SRusty Russell to compress the whole initrd or initramfs instead. 2170beb50df3SBertrand Jacquin 2171b6c09b51SRusty Russell Note: This is fully compatible with signed modules. 2172b6c09b51SRusty Russell 2173b6c09b51SRusty Russell If in doubt, say N. 2174beb50df3SBertrand Jacquin 2175beb50df3SBertrand Jacquinchoice 2176beb50df3SBertrand Jacquin prompt "Compression algorithm" 2177beb50df3SBertrand Jacquin depends on MODULE_COMPRESS 2178beb50df3SBertrand Jacquin default MODULE_COMPRESS_GZIP 2179beb50df3SBertrand Jacquin help 2180beb50df3SBertrand Jacquin This determines which sort of compression will be used during 2181beb50df3SBertrand Jacquin 'make modules_install'. 2182beb50df3SBertrand Jacquin 2183beb50df3SBertrand Jacquin GZIP (default) and XZ are supported. 2184beb50df3SBertrand Jacquin 2185beb50df3SBertrand Jacquinconfig MODULE_COMPRESS_GZIP 2186beb50df3SBertrand Jacquin bool "GZIP" 2187beb50df3SBertrand Jacquin 2188beb50df3SBertrand Jacquinconfig MODULE_COMPRESS_XZ 2189beb50df3SBertrand Jacquin bool "XZ" 2190beb50df3SBertrand Jacquin 2191beb50df3SBertrand Jacquinendchoice 2192beb50df3SBertrand Jacquin 21933d52ec5eSMatthias Maennichconfig MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS 21943d52ec5eSMatthias Maennich bool "Allow loading of modules with missing namespace imports" 21953d52ec5eSMatthias Maennich help 21963d52ec5eSMatthias Maennich Symbols exported with EXPORT_SYMBOL_NS*() are considered exported in 21973d52ec5eSMatthias Maennich a namespace. A module that makes use of a symbol exported with such a 21983d52ec5eSMatthias Maennich namespace is required to import the namespace via MODULE_IMPORT_NS(). 21993d52ec5eSMatthias Maennich There is no technical reason to enforce correct namespace imports, 22003d52ec5eSMatthias Maennich but it creates consistency between symbols defining namespaces and 22013d52ec5eSMatthias Maennich users importing namespaces they make use of. This option relaxes this 22023d52ec5eSMatthias Maennich requirement and lifts the enforcement when loading a module. 22033d52ec5eSMatthias Maennich 22043d52ec5eSMatthias Maennich If unsure, say N. 22053d52ec5eSMatthias Maennich 2206efd9763dSMasahiro Yamadaconfig UNUSED_SYMBOLS 2207efd9763dSMasahiro Yamada bool "Enable unused/obsolete exported symbols" 2208efd9763dSMasahiro Yamada default y if X86 2209efd9763dSMasahiro Yamada help 2210efd9763dSMasahiro Yamada Unused but exported symbols make the kernel needlessly bigger. For 2211efd9763dSMasahiro Yamada that reason most of these unused exports will soon be removed. This 2212efd9763dSMasahiro Yamada option is provided temporarily to provide a transition period in case 2213efd9763dSMasahiro Yamada some external kernel module needs one of these symbols anyway. If you 2214efd9763dSMasahiro Yamada encounter such a case in your module, consider if you are actually 2215efd9763dSMasahiro Yamada using the right API. (rationale: since nobody in the kernel is using 2216efd9763dSMasahiro Yamada this in a module, there is a pretty good chance it's actually the 2217efd9763dSMasahiro Yamada wrong interface to use). If you really need the symbol, please send a 2218efd9763dSMasahiro Yamada mail to the linux kernel mailing list mentioning the symbol and why 2219efd9763dSMasahiro Yamada you really need it, and what the merge plan to the mainline kernel for 2220efd9763dSMasahiro Yamada your module is. 2221efd9763dSMasahiro Yamada 2222dbacb0efSNicolas Pitreconfig TRIM_UNUSED_KSYMS 2223dbacb0efSNicolas Pitre bool "Trim unused exported kernel symbols" 2224d189c2a4SMasahiro Yamada depends on !UNUSED_SYMBOLS 2225dbacb0efSNicolas Pitre help 2226dbacb0efSNicolas Pitre The kernel and some modules make many symbols available for 2227dbacb0efSNicolas Pitre other modules to use via EXPORT_SYMBOL() and variants. Depending 2228dbacb0efSNicolas Pitre on the set of modules being selected in your kernel configuration, 2229dbacb0efSNicolas Pitre many of those exported symbols might never be used. 2230dbacb0efSNicolas Pitre 2231dbacb0efSNicolas Pitre This option allows for unused exported symbols to be dropped from 2232dbacb0efSNicolas Pitre the build. In turn, this provides the compiler more opportunities 2233dbacb0efSNicolas Pitre (especially when using LTO) for optimizing the code and reducing 2234dbacb0efSNicolas Pitre binary size. This might have some security advantages as well. 2235dbacb0efSNicolas Pitre 2236f1cb637eSValdis Kletnieks If unsure, or if you need to build out-of-tree modules, say N. 2237dbacb0efSNicolas Pitre 22381518c633SQuentin Perretconfig UNUSED_KSYMS_WHITELIST 22391518c633SQuentin Perret string "Whitelist of symbols to keep in ksymtab" 22401518c633SQuentin Perret depends on TRIM_UNUSED_KSYMS 22411518c633SQuentin Perret help 22421518c633SQuentin Perret By default, all unused exported symbols will be un-exported from the 22431518c633SQuentin Perret build when TRIM_UNUSED_KSYMS is selected. 22441518c633SQuentin Perret 22451518c633SQuentin Perret UNUSED_KSYMS_WHITELIST allows to whitelist symbols that must be kept 22461518c633SQuentin Perret exported at all times, even in absence of in-tree users. The value to 22471518c633SQuentin Perret set here is the path to a text file containing the list of symbols, 22481518c633SQuentin Perret one per line. The path can be absolute, or relative to the kernel 22491518c633SQuentin Perret source tree. 22501518c633SQuentin Perret 22510b0de144SRobert P. J. Dayendif # MODULES 22520b0de144SRobert P. J. Day 22536c9692e2SPeter Zijlstraconfig MODULES_TREE_LOOKUP 22546c9692e2SPeter Zijlstra def_bool y 22556c9692e2SPeter Zijlstra depends on PERF_EVENTS || TRACING 22566c9692e2SPeter Zijlstra 225798a79d6aSRusty Russellconfig INIT_ALL_POSSIBLE 225898a79d6aSRusty Russell bool 225998a79d6aSRusty Russell help 22605f054e31SRusty Russell Back when each arch used to define their own cpu_online_mask and 22615f054e31SRusty Russell cpu_possible_mask, some of them chose to initialize cpu_possible_mask 226298a79d6aSRusty Russell with all 1s, and others with all 0s. When they were centralised, 226398a79d6aSRusty Russell it was better to provide this option than to break all the archs 2264692105b8SMatt LaPlante and have several arch maintainers pursuing me down dark alleys. 226598a79d6aSRusty Russell 22663a65dfe8SJens Axboesource "block/Kconfig" 2267e98c3202SAvi Kivity 2268e98c3202SAvi Kivityconfig PREEMPT_NOTIFIERS 2269e98c3202SAvi Kivity bool 2270e260be67SPaul E. McKenney 227116295becSSteffen Klassertconfig PADATA 227216295becSSteffen Klassert depends on SMP 227316295becSSteffen Klassert bool 227416295becSSteffen Klassert 22754520c6a4SDavid Howellsconfig ASN1 22764520c6a4SDavid Howells tristate 22774520c6a4SDavid Howells help 22784520c6a4SDavid Howells Build a simple ASN.1 grammar compiler that produces a bytecode output 22794520c6a4SDavid Howells that can be interpreted by the ASN.1 stream decoder and used to 22804520c6a4SDavid Howells inform it as to what tags are to be expected in a stream and what 22814520c6a4SDavid Howells functions to call on what tags. 22824520c6a4SDavid Howells 22836beb0009SThomas Gleixnersource "kernel/Kconfig.locks" 2284e61938a9SMathieu Desnoyers 2285e61938a9SMathieu Desnoyersconfig ARCH_HAS_SYNC_CORE_BEFORE_USERMODE 2286e61938a9SMathieu Desnoyers bool 22871bd21c6cSDominik Brodowski 22881bd21c6cSDominik Brodowski# It may be useful for an architecture to override the definitions of the 22897303e30eSDominik Brodowski# SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h> 22907303e30eSDominik Brodowski# and the COMPAT_ variants in <linux/compat.h>, in particular to use a 22917303e30eSDominik Brodowski# different calling convention for syscalls. They can also override the 22927303e30eSDominik Brodowski# macros for not-implemented syscalls in kernel/sys_ni.c and 22937303e30eSDominik Brodowski# kernel/time/posix-stubs.c. All these overrides need to be available in 22947303e30eSDominik Brodowski# <asm/syscall_wrapper.h>. 22951bd21c6cSDominik Brodowskiconfig ARCH_HAS_SYSCALL_WRAPPER 22961bd21c6cSDominik Brodowski def_bool n 2297