1 # SPDX-License-Identifier: GPL-2.0-only
6 default "/lib/modules/$(shell,uname -r)/.config"
7 default "/etc/kernel-config"
8 default "/boot/config-$(shell,uname -r)"
9 default "arch/$(SRCARCH)/configs/$(KBUILD_DEFCONFIG)"
12 def_bool $(success,$(CC) --version | head -n 1 | grep -q gcc)
16 default $(shell,$(srctree)/scripts/gcc-version.sh $(CC)) if CC_IS_GCC
21 default $(shell,$(LD) --version | $(srctree)/scripts/ld-version.sh)
24 def_bool $(success,$(CC) --version | head -n 1 | grep -q clang)
27 def_bool $(success,$(LD) -v | head -n 1 | grep -q LLD)
31 default $(shell,$(srctree)/scripts/clang-version.sh $(CC))
34 def_bool $(success,$(srctree)/scripts/cc-can-link.sh $(CC))
36 config CC_HAS_ASM_GOTO
37 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
39 config CC_HAS_ASM_GOTO_OUTPUT
40 depends on CC_HAS_ASM_GOTO
41 def_bool $(success,echo 'int foo(int x) { asm goto ("": "=r"(x) ::: bar); return x; bar: return 0; }' | $(CC) -x c - -c -o /dev/null)
43 config TOOLS_SUPPORT_RELR
44 def_bool $(success,env "CC=$(CC)" "LD=$(LD)" "NM=$(NM)" "OBJCOPY=$(OBJCOPY)" $(srctree)/scripts/tools-support-relr.sh)
46 config CC_HAS_ASM_INLINE
47 def_bool $(success,echo 'void foo(void) { asm inline (""); }' | $(CC) -x c - -c -o /dev/null)
56 config BUILDTIME_TABLE_SORT
59 config THREAD_INFO_IN_TASK
62 Select this to move thread_info off the stack into task_struct. To
63 make this work, an arch will need to remove all thread_info fields
64 except flags and fix any runtime bugs.
66 One subtle change that will be needed is to use try_get_task_stack()
67 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
76 depends on BROKEN || !SMP
79 config INIT_ENV_ARG_LIMIT
84 Maximum of each of the number of arguments and environment
85 variables passed to init from the kernel command line.
88 bool "Compile also drivers which will not load"
92 Some drivers can be compiled on a different platform than they are
93 intended to be run on. Despite they cannot be loaded there (or even
94 when they load they cannot be used due to missing HW support),
95 developers still, opposing to distributors, might want to build such
96 drivers to compile-test them.
98 If you are a developer and want to build everything available, say Y
99 here. If you are a user/distributor, say N here to exclude useless
100 drivers to be distributed.
102 config UAPI_HEADER_TEST
103 bool "Compile test UAPI headers"
104 depends on HEADERS_INSTALL && CC_CAN_LINK
106 Compile test headers exported to user-space to ensure they are
107 self-contained, i.e. compilable as standalone units.
109 If you are a developer or tester and want to ensure the exported
110 headers are self-contained, say Y here. Otherwise, choose N.
113 string "Local version - append to kernel release"
115 Append an extra string to the end of your kernel version.
116 This will show up when you type uname, for example.
117 The string you set here will be appended after the contents of
118 any files with a filename matching localversion* in your
119 object and source tree, in that order. Your total string can
120 be a maximum of 64 characters.
122 config LOCALVERSION_AUTO
123 bool "Automatically append version information to the version string"
125 depends on !COMPILE_TEST
127 This will try to automatically determine if the current tree is a
128 release tree by looking for git tags that belong to the current
129 top of tree revision.
131 A string of the format -gxxxxxxxx will be added to the localversion
132 if a git-based tree is found. The string generated by this will be
133 appended after any matching localversion* files, and after the value
134 set in CONFIG_LOCALVERSION.
136 (The actual string used here is the first eight characters produced
137 by running the command:
139 $ git rev-parse --verify HEAD
141 which is done within the script "scripts/setlocalversion".)
144 string "Build ID Salt"
147 The build ID is used to link binaries and their debug info. Setting
148 this option will use the value in the calculation of the build id.
149 This is mostly useful for distributions which want to ensure the
150 build is unique between builds. It's safe to leave the default.
152 config HAVE_KERNEL_GZIP
155 config HAVE_KERNEL_BZIP2
158 config HAVE_KERNEL_LZMA
161 config HAVE_KERNEL_XZ
164 config HAVE_KERNEL_LZO
167 config HAVE_KERNEL_LZ4
170 config HAVE_KERNEL_UNCOMPRESSED
174 prompt "Kernel compression mode"
176 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
178 The linux kernel is a kind of self-extracting executable.
179 Several compression algorithms are available, which differ
180 in efficiency, compression and decompression speed.
181 Compression speed is only relevant when building a kernel.
182 Decompression speed is relevant at each boot.
184 If you have any problems with bzip2 or lzma compressed
185 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
186 version of this functionality (bzip2 only), for 2.4, was
187 supplied by Christian Ludwig)
189 High compression options are mostly useful for users, who
190 are low on disk space (embedded systems), but for whom ram
193 If in doubt, select 'gzip'
197 depends on HAVE_KERNEL_GZIP
199 The old and tried gzip compression. It provides a good balance
200 between compression ratio and decompression speed.
204 depends on HAVE_KERNEL_BZIP2
206 Its compression ratio and speed is intermediate.
207 Decompression speed is slowest among the choices. The kernel
208 size is about 10% smaller with bzip2, in comparison to gzip.
209 Bzip2 uses a large amount of memory. For modern kernels you
210 will need at least 8MB RAM or more for booting.
214 depends on HAVE_KERNEL_LZMA
216 This compression algorithm's ratio is best. Decompression speed
217 is between gzip and bzip2. Compression is slowest.
218 The kernel size is about 33% smaller with LZMA in comparison to gzip.
222 depends on HAVE_KERNEL_XZ
224 XZ uses the LZMA2 algorithm and instruction set specific
225 BCJ filters which can improve compression ratio of executable
226 code. The size of the kernel is about 30% smaller with XZ in
227 comparison to gzip. On architectures for which there is a BCJ
228 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
229 will create a few percent smaller kernel than plain LZMA.
231 The speed is about the same as with LZMA: The decompression
232 speed of XZ is better than that of bzip2 but worse than gzip
233 and LZO. Compression is slow.
237 depends on HAVE_KERNEL_LZO
239 Its compression ratio is the poorest among the choices. The kernel
240 size is about 10% bigger than gzip; however its speed
241 (both compression and decompression) is the fastest.
245 depends on HAVE_KERNEL_LZ4
247 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
248 A preliminary version of LZ4 de/compression tool is available at
249 <https://code.google.com/p/lz4/>.
251 Its compression ratio is worse than LZO. The size of the kernel
252 is about 8% bigger than LZO. But the decompression speed is
255 config KERNEL_UNCOMPRESSED
257 depends on HAVE_KERNEL_UNCOMPRESSED
259 Produce uncompressed kernel image. This option is usually not what
260 you want. It is useful for debugging the kernel in slow simulation
261 environments, where decompressing and moving the kernel is awfully
262 slow. This option allows early boot code to skip the decompressor
263 and jump right at uncompressed kernel image.
268 string "Default init path"
271 This option determines the default init for the system if no init=
272 option is passed on the kernel command line. If the requested path is
273 not present, we will still then move on to attempting further
274 locations (e.g. /sbin/init, etc). If this is empty, we will just use
275 the fallback list when init= is not passed.
277 config DEFAULT_HOSTNAME
278 string "Default hostname"
281 This option determines the default system hostname before userspace
282 calls sethostname(2). The kernel traditionally uses "(none)" here,
283 but you may wish to use a different default here to make a minimal
284 system more usable with less configuration.
287 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
288 # add proper SWAP support to them, in which case this can be remove.
294 bool "Support for paging of anonymous memory (swap)"
295 depends on MMU && BLOCK && !ARCH_NO_SWAP
298 This option allows you to choose whether you want to have support
299 for so called swap devices or swap files in your kernel that are
300 used to provide more virtual memory than the actual RAM present
301 in your computer. If unsure say Y.
306 Inter Process Communication is a suite of library functions and
307 system calls which let processes (running programs) synchronize and
308 exchange information. It is generally considered to be a good thing,
309 and some programs won't run unless you say Y here. In particular, if
310 you want to run the DOS emulator dosemu under Linux (read the
311 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
312 you'll need to say Y here.
314 You can find documentation about IPC with "info ipc" and also in
315 section 6.4 of the Linux Programmer's Guide, available from
316 <http://www.tldp.org/guides.html>.
318 config SYSVIPC_SYSCTL
325 bool "POSIX Message Queues"
328 POSIX variant of message queues is a part of IPC. In POSIX message
329 queues every message has a priority which decides about succession
330 of receiving it by a process. If you want to compile and run
331 programs written e.g. for Solaris with use of its POSIX message
332 queues (functions mq_*) say Y here.
334 POSIX message queues are visible as a filesystem called 'mqueue'
335 and can be mounted somewhere if you want to do filesystem
336 operations on message queues.
340 config POSIX_MQUEUE_SYSCTL
342 depends on POSIX_MQUEUE
346 config CROSS_MEMORY_ATTACH
347 bool "Enable process_vm_readv/writev syscalls"
351 Enabling this option adds the system calls process_vm_readv and
352 process_vm_writev which allow a process with the correct privileges
353 to directly read from or write to another process' address space.
354 See the man page for more details.
357 bool "uselib syscall"
358 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
360 This option enables the uselib syscall, a system call used in the
361 dynamic linker from libc5 and earlier. glibc does not use this
362 system call. If you intend to run programs built on libc5 or
363 earlier, you may need to enable this syscall. Current systems
364 running glibc can safely disable this.
367 bool "Auditing support"
370 Enable auditing infrastructure that can be used with another
371 kernel subsystem, such as SELinux (which requires this for
372 logging of avc messages output). System call auditing is included
373 on architectures which support it.
375 config HAVE_ARCH_AUDITSYSCALL
380 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
383 source "kernel/irq/Kconfig"
384 source "kernel/time/Kconfig"
385 source "kernel/Kconfig.preempt"
387 menu "CPU/Task time and stats accounting"
389 config VIRT_CPU_ACCOUNTING
393 prompt "Cputime accounting"
394 default TICK_CPU_ACCOUNTING if !PPC64
395 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
397 # Kind of a stub config for the pure tick based cputime accounting
398 config TICK_CPU_ACCOUNTING
399 bool "Simple tick based cputime accounting"
400 depends on !S390 && !NO_HZ_FULL
402 This is the basic tick based cputime accounting that maintains
403 statistics about user, system and idle time spent on per jiffies
408 config VIRT_CPU_ACCOUNTING_NATIVE
409 bool "Deterministic task and CPU time accounting"
410 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
411 select VIRT_CPU_ACCOUNTING
413 Select this option to enable more accurate task and CPU time
414 accounting. This is done by reading a CPU counter on each
415 kernel entry and exit and on transitions within the kernel
416 between system, softirq and hardirq state, so there is a
417 small performance impact. In the case of s390 or IBM POWER > 5,
418 this also enables accounting of stolen time on logically-partitioned
421 config VIRT_CPU_ACCOUNTING_GEN
422 bool "Full dynticks CPU time accounting"
423 depends on HAVE_CONTEXT_TRACKING
424 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
425 depends on GENERIC_CLOCKEVENTS
426 select VIRT_CPU_ACCOUNTING
427 select CONTEXT_TRACKING
429 Select this option to enable task and CPU time accounting on full
430 dynticks systems. This accounting is implemented by watching every
431 kernel-user boundaries using the context tracking subsystem.
432 The accounting is thus performed at the expense of some significant
435 For now this is only useful if you are working on the full
436 dynticks subsystem development.
442 config IRQ_TIME_ACCOUNTING
443 bool "Fine granularity task level IRQ time accounting"
444 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
446 Select this option to enable fine granularity task irq time
447 accounting. This is done by reading a timestamp on each
448 transitions between softirq and hardirq state, so there can be a
449 small performance impact.
451 If in doubt, say N here.
453 config HAVE_SCHED_AVG_IRQ
455 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
458 config SCHED_THERMAL_PRESSURE
459 bool "Enable periodic averaging of thermal pressure"
462 config BSD_PROCESS_ACCT
463 bool "BSD Process Accounting"
466 If you say Y here, a user level program will be able to instruct the
467 kernel (via a special system call) to write process accounting
468 information to a file: whenever a process exits, information about
469 that process will be appended to the file by the kernel. The
470 information includes things such as creation time, owning user,
471 command name, memory usage, controlling terminal etc. (the complete
472 list is in the struct acct in <file:include/linux/acct.h>). It is
473 up to the user level program to do useful things with this
474 information. This is generally a good idea, so say Y.
476 config BSD_PROCESS_ACCT_V3
477 bool "BSD Process Accounting version 3 file format"
478 depends on BSD_PROCESS_ACCT
481 If you say Y here, the process accounting information is written
482 in a new file format that also logs the process IDs of each
483 process and its parent. Note that this file format is incompatible
484 with previous v0/v1/v2 file formats, so you will need updated tools
485 for processing it. A preliminary version of these tools is available
486 at <http://www.gnu.org/software/acct/>.
489 bool "Export task/process statistics through netlink"
494 Export selected statistics for tasks/processes through the
495 generic netlink interface. Unlike BSD process accounting, the
496 statistics are available during the lifetime of tasks/processes as
497 responses to commands. Like BSD accounting, they are sent to user
502 config TASK_DELAY_ACCT
503 bool "Enable per-task delay accounting"
507 Collect information on time spent by a task waiting for system
508 resources like cpu, synchronous block I/O completion and swapping
509 in pages. Such statistics can help in setting a task's priorities
510 relative to other tasks for cpu, io, rss limits etc.
515 bool "Enable extended accounting over taskstats"
518 Collect extended task accounting data and send the data
519 to userland for processing over the taskstats interface.
523 config TASK_IO_ACCOUNTING
524 bool "Enable per-task storage I/O accounting"
525 depends on TASK_XACCT
527 Collect information on the number of bytes of storage I/O which this
533 bool "Pressure stall information tracking"
535 Collect metrics that indicate how overcommitted the CPU, memory,
536 and IO capacity are in the system.
538 If you say Y here, the kernel will create /proc/pressure/ with the
539 pressure statistics files cpu, memory, and io. These will indicate
540 the share of walltime in which some or all tasks in the system are
541 delayed due to contention of the respective resource.
543 In kernels with cgroup support, cgroups (cgroup2 only) will
544 have cpu.pressure, memory.pressure, and io.pressure files,
545 which aggregate pressure stalls for the grouped tasks only.
547 For more details see Documentation/accounting/psi.rst.
551 config PSI_DEFAULT_DISABLED
552 bool "Require boot parameter to enable pressure stall information tracking"
556 If set, pressure stall information tracking will be disabled
557 per default but can be enabled through passing psi=1 on the
558 kernel commandline during boot.
560 This feature adds some code to the task wakeup and sleep
561 paths of the scheduler. The overhead is too low to affect
562 common scheduling-intense workloads in practice (such as
563 webservers, memcache), but it does show up in artificial
564 scheduler stress tests, such as hackbench.
566 If you are paranoid and not sure what the kernel will be
571 endmenu # "CPU/Task time and stats accounting"
575 depends on SMP || COMPILE_TEST
578 Make sure that CPUs running critical tasks are not disturbed by
579 any source of "noise" such as unbound workqueues, timers, kthreads...
580 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
581 the "isolcpus=" boot parameter.
585 source "kernel/rcu/Kconfig"
592 tristate "Kernel .config support"
594 This option enables the complete Linux kernel ".config" file
595 contents to be saved in the kernel. It provides documentation
596 of which kernel options are used in a running kernel or in an
597 on-disk kernel. This information can be extracted from the kernel
598 image file with the script scripts/extract-ikconfig and used as
599 input to rebuild the current kernel or to build another kernel.
600 It can also be extracted from a running kernel by reading
601 /proc/config.gz if enabled (below).
604 bool "Enable access to .config through /proc/config.gz"
605 depends on IKCONFIG && PROC_FS
607 This option enables access to the kernel configuration file
608 through /proc/config.gz.
611 tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz"
614 This option enables access to the in-kernel headers that are generated during
615 the build process. These can be used to build eBPF tracing programs,
616 or similar programs. If you build the headers as a module, a module called
617 kheaders.ko is built which can be loaded on-demand to get access to headers.
620 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
625 Select the minimal kernel log buffer size as a power of 2.
626 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
627 parameter, see below. Any higher size also might be forced
628 by "log_buf_len" boot parameter.
638 config LOG_CPU_MAX_BUF_SHIFT
639 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
642 default 12 if !BASE_SMALL
643 default 0 if BASE_SMALL
646 This option allows to increase the default ring buffer size
647 according to the number of CPUs. The value defines the contribution
648 of each CPU as a power of 2. The used space is typically only few
649 lines however it might be much more when problems are reported,
652 The increased size means that a new buffer has to be allocated and
653 the original static one is unused. It makes sense only on systems
654 with more CPUs. Therefore this value is used only when the sum of
655 contributions is greater than the half of the default kernel ring
656 buffer as defined by LOG_BUF_SHIFT. The default values are set
657 so that more than 64 CPUs are needed to trigger the allocation.
659 Also this option is ignored when "log_buf_len" kernel parameter is
660 used as it forces an exact (power of two) size of the ring buffer.
662 The number of possible CPUs is used for this computation ignoring
663 hotplugging making the computation optimal for the worst case
664 scenario while allowing a simple algorithm to be used from bootup.
666 Examples shift values and their meaning:
667 17 => 128 KB for each CPU
668 16 => 64 KB for each CPU
669 15 => 32 KB for each CPU
670 14 => 16 KB for each CPU
671 13 => 8 KB for each CPU
672 12 => 4 KB for each CPU
674 config PRINTK_SAFE_LOG_BUF_SHIFT
675 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
680 Select the size of an alternate printk per-CPU buffer where messages
681 printed from usafe contexts are temporary stored. One example would
682 be NMI messages, another one - printk recursion. The messages are
683 copied to the main log buffer in a safe context to avoid a deadlock.
684 The value defines the size as a power of 2.
686 Those messages are rare and limited. The largest one is when
687 a backtrace is printed. It usually fits into 4KB. Select
688 8KB if you want to be on the safe side.
691 17 => 128 KB for each CPU
692 16 => 64 KB for each CPU
693 15 => 32 KB for each CPU
694 14 => 16 KB for each CPU
695 13 => 8 KB for each CPU
696 12 => 4 KB for each CPU
699 # Architectures with an unreliable sched_clock() should select this:
701 config HAVE_UNSTABLE_SCHED_CLOCK
704 config GENERIC_SCHED_CLOCK
707 menu "Scheduler features"
710 bool "Enable utilization clamping for RT/FAIR tasks"
711 depends on CPU_FREQ_GOV_SCHEDUTIL
713 This feature enables the scheduler to track the clamped utilization
714 of each CPU based on RUNNABLE tasks scheduled on that CPU.
716 With this option, the user can specify the min and max CPU
717 utilization allowed for RUNNABLE tasks. The max utilization defines
718 the maximum frequency a task should use while the min utilization
719 defines the minimum frequency it should use.
721 Both min and max utilization clamp values are hints to the scheduler,
722 aiming at improving its frequency selection policy, but they do not
723 enforce or grant any specific bandwidth for tasks.
727 config UCLAMP_BUCKETS_COUNT
728 int "Number of supported utilization clamp buckets"
731 depends on UCLAMP_TASK
733 Defines the number of clamp buckets to use. The range of each bucket
734 will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the
735 number of clamp buckets the finer their granularity and the higher
736 the precision of clamping aggregation and tracking at run-time.
738 For example, with the minimum configuration value we will have 5
739 clamp buckets tracking 20% utilization each. A 25% boosted tasks will
740 be refcounted in the [20..39]% bucket and will set the bucket clamp
741 effective value to 25%.
742 If a second 30% boosted task should be co-scheduled on the same CPU,
743 that task will be refcounted in the same bucket of the first task and
744 it will boost the bucket clamp effective value to 30%.
745 The clamp effective value of a bucket is reset to its nominal value
746 (20% in the example above) when there are no more tasks refcounted in
749 An additional boost/capping margin can be added to some tasks. In the
750 example above the 25% task will be boosted to 30% until it exits the
751 CPU. If that should be considered not acceptable on certain systems,
752 it's always possible to reduce the margin by increasing the number of
753 clamp buckets to trade off used memory for run-time tracking
756 If in doubt, use the default value.
761 # For architectures that want to enable the support for NUMA-affine scheduler
764 config ARCH_SUPPORTS_NUMA_BALANCING
768 # For architectures that prefer to flush all TLBs after a number of pages
769 # are unmapped instead of sending one IPI per page to flush. The architecture
770 # must provide guarantees on what happens if a clean TLB cache entry is
771 # written after the unmap. Details are in mm/rmap.c near the check for
772 # should_defer_flush. The architecture should also consider if the full flush
773 # and the refill costs are offset by the savings of sending fewer IPIs.
774 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
778 def_bool !$(cc-option,$(m64-flag) -D__SIZEOF_INT128__=0) && 64BIT
781 # For architectures that know their GCC __int128 support is sound
783 config ARCH_SUPPORTS_INT128
786 # For architectures that (ab)use NUMA to represent different memory regions
787 # all cpu-local but of different latencies, such as SuperH.
789 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
792 config NUMA_BALANCING
793 bool "Memory placement aware NUMA scheduler"
794 depends on ARCH_SUPPORTS_NUMA_BALANCING
795 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
796 depends on SMP && NUMA && MIGRATION
798 This option adds support for automatic NUMA aware memory/task placement.
799 The mechanism is quite primitive and is based on migrating memory when
800 it has references to the node the task is running on.
802 This system will be inactive on UMA systems.
804 config NUMA_BALANCING_DEFAULT_ENABLED
805 bool "Automatically enable NUMA aware memory/task placement"
807 depends on NUMA_BALANCING
809 If set, automatic NUMA balancing will be enabled if running on a NUMA
813 bool "Control Group support"
816 This option adds support for grouping sets of processes together, for
817 use with process control subsystems such as Cpusets, CFS, memory
818 controls or device isolation.
820 - Documentation/scheduler/sched-design-CFS.rst (CFS)
821 - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation
822 and resource control)
832 bool "Memory controller"
836 Provides control over the memory footprint of tasks in a cgroup.
840 depends on MEMCG && SWAP
845 depends on MEMCG && !SLOB
853 Generic block IO controller cgroup interface. This is the common
854 cgroup interface which should be used by various IO controlling
857 Currently, CFQ IO scheduler uses it to recognize task groups and
858 control disk bandwidth allocation (proportional time slice allocation)
859 to such task groups. It is also used by bio throttling logic in
860 block layer to implement upper limit in IO rates on a device.
862 This option only enables generic Block IO controller infrastructure.
863 One needs to also enable actual IO controlling logic/policy. For
864 enabling proportional weight division of disk bandwidth in CFQ, set
865 CONFIG_BFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
866 CONFIG_BLK_DEV_THROTTLING=y.
868 See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information.
870 config CGROUP_WRITEBACK
872 depends on MEMCG && BLK_CGROUP
875 menuconfig CGROUP_SCHED
876 bool "CPU controller"
879 This feature lets CPU scheduler recognize task groups and control CPU
880 bandwidth allocation to such task groups. It uses cgroups to group
884 config FAIR_GROUP_SCHED
885 bool "Group scheduling for SCHED_OTHER"
886 depends on CGROUP_SCHED
890 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
891 depends on FAIR_GROUP_SCHED
894 This option allows users to define CPU bandwidth rates (limits) for
895 tasks running within the fair group scheduler. Groups with no limit
896 set are considered to be unconstrained and will run with no
898 See Documentation/scheduler/sched-bwc.rst for more information.
900 config RT_GROUP_SCHED
901 bool "Group scheduling for SCHED_RR/FIFO"
902 depends on CGROUP_SCHED
905 This feature lets you explicitly allocate real CPU bandwidth
906 to task groups. If enabled, it will also make it impossible to
907 schedule realtime tasks for non-root users until you allocate
908 realtime bandwidth for them.
909 See Documentation/scheduler/sched-rt-group.rst for more information.
913 config UCLAMP_TASK_GROUP
914 bool "Utilization clamping per group of tasks"
915 depends on CGROUP_SCHED
916 depends on UCLAMP_TASK
919 This feature enables the scheduler to track the clamped utilization
920 of each CPU based on RUNNABLE tasks currently scheduled on that CPU.
922 When this option is enabled, the user can specify a min and max
923 CPU bandwidth which is allowed for each single task in a group.
924 The max bandwidth allows to clamp the maximum frequency a task
925 can use, while the min bandwidth allows to define a minimum
926 frequency a task will always use.
928 When task group based utilization clamping is enabled, an eventually
929 specified task-specific clamp value is constrained by the cgroup
930 specified clamp value. Both minimum and maximum task clamping cannot
931 be bigger than the corresponding clamping defined at task group level.
936 bool "PIDs controller"
938 Provides enforcement of process number limits in the scope of a
939 cgroup. Any attempt to fork more processes than is allowed in the
940 cgroup will fail. PIDs are fundamentally a global resource because it
941 is fairly trivial to reach PID exhaustion before you reach even a
942 conservative kmemcg limit. As a result, it is possible to grind a
943 system to halt without being limited by other cgroup policies. The
944 PIDs controller is designed to stop this from happening.
946 It should be noted that organisational operations (such as attaching
947 to a cgroup hierarchy) will *not* be blocked by the PIDs controller,
948 since the PIDs limit only affects a process's ability to fork, not to
952 bool "RDMA controller"
954 Provides enforcement of RDMA resources defined by IB stack.
955 It is fairly easy for consumers to exhaust RDMA resources, which
956 can result into resource unavailability to other consumers.
957 RDMA controller is designed to stop this from happening.
958 Attaching processes with active RDMA resources to the cgroup
959 hierarchy is allowed even if can cross the hierarchy's limit.
961 config CGROUP_FREEZER
962 bool "Freezer controller"
964 Provides a way to freeze and unfreeze all tasks in a
967 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
968 controller includes important in-kernel memory consumers per default.
970 If you're using cgroup2, say N.
972 config CGROUP_HUGETLB
973 bool "HugeTLB controller"
974 depends on HUGETLB_PAGE
978 Provides a cgroup controller for HugeTLB pages.
979 When you enable this, you can put a per cgroup limit on HugeTLB usage.
980 The limit is enforced during page fault. Since HugeTLB doesn't
981 support page reclaim, enforcing the limit at page fault time implies
982 that, the application will get SIGBUS signal if it tries to access
983 HugeTLB pages beyond its limit. This requires the application to know
984 beforehand how much HugeTLB pages it would require for its use. The
985 control group is tracked in the third page lru pointer. This means
986 that we cannot use the controller with huge page less than 3 pages.
989 bool "Cpuset controller"
992 This option will let you create and manage CPUSETs which
993 allow dynamically partitioning a system into sets of CPUs and
994 Memory Nodes and assigning tasks to run only within those sets.
995 This is primarily useful on large SMP or NUMA systems.
999 config PROC_PID_CPUSET
1000 bool "Include legacy /proc/<pid>/cpuset file"
1004 config CGROUP_DEVICE
1005 bool "Device controller"
1007 Provides a cgroup controller implementing whitelists for
1008 devices which a process in the cgroup can mknod or open.
1010 config CGROUP_CPUACCT
1011 bool "Simple CPU accounting controller"
1013 Provides a simple controller for monitoring the
1014 total CPU consumed by the tasks in a cgroup.
1017 bool "Perf controller"
1018 depends on PERF_EVENTS
1020 This option extends the perf per-cpu mode to restrict monitoring
1021 to threads which belong to the cgroup specified and run on the
1022 designated cpu. Or this can be used to have cgroup ID in samples
1023 so that it can monitor performance events among cgroups.
1028 bool "Support for eBPF programs attached to cgroups"
1029 depends on BPF_SYSCALL
1030 select SOCK_CGROUP_DATA
1032 Allow attaching eBPF programs to a cgroup using the bpf(2)
1033 syscall command BPF_PROG_ATTACH.
1035 In which context these programs are accessed depends on the type
1036 of attachment. For instance, programs that are attached using
1037 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
1041 bool "Debug controller"
1043 depends on DEBUG_KERNEL
1045 This option enables a simple controller that exports
1046 debugging information about the cgroups framework. This
1047 controller is for control cgroup debugging only. Its
1048 interfaces are not stable.
1052 config SOCK_CGROUP_DATA
1058 menuconfig NAMESPACES
1059 bool "Namespaces support" if EXPERT
1060 depends on MULTIUSER
1063 Provides the way to make tasks work with different objects using
1064 the same id. For example same IPC id may refer to different objects
1065 or same user id or pid may refer to different tasks when used in
1066 different namespaces.
1071 bool "UTS namespace"
1074 In this namespace tasks see different info provided with the
1078 bool "TIME namespace"
1079 depends on GENERIC_VDSO_TIME_NS
1082 In this namespace boottime and monotonic clocks can be set.
1083 The time will keep going with the same pace.
1086 bool "IPC namespace"
1087 depends on (SYSVIPC || POSIX_MQUEUE)
1090 In this namespace tasks work with IPC ids which correspond to
1091 different IPC objects in different namespaces.
1094 bool "User namespace"
1097 This allows containers, i.e. vservers, to use user namespaces
1098 to provide different user info for different servers.
1100 When user namespaces are enabled in the kernel it is
1101 recommended that the MEMCG option also be enabled and that
1102 user-space use the memory control groups to limit the amount
1103 of memory a memory unprivileged users can use.
1108 bool "PID Namespaces"
1111 Support process id namespaces. This allows having multiple
1112 processes with the same pid as long as they are in different
1113 pid namespaces. This is a building block of containers.
1116 bool "Network namespace"
1120 Allow user space to create what appear to be multiple instances
1121 of the network stack.
1125 config CHECKPOINT_RESTORE
1126 bool "Checkpoint/restore support"
1127 select PROC_CHILDREN
1130 Enables additional kernel features in a sake of checkpoint/restore.
1131 In particular it adds auxiliary prctl codes to setup process text,
1132 data and heap segment sizes, and a few additional /proc filesystem
1135 If unsure, say N here.
1137 config SCHED_AUTOGROUP
1138 bool "Automatic process group scheduling"
1141 select FAIR_GROUP_SCHED
1143 This option optimizes the scheduler for common desktop workloads by
1144 automatically creating and populating task groups. This separation
1145 of workloads isolates aggressive CPU burners (like build jobs) from
1146 desktop applications. Task group autogeneration is currently based
1149 config SYSFS_DEPRECATED
1150 bool "Enable deprecated sysfs features to support old userspace tools"
1154 This option adds code that switches the layout of the "block" class
1155 devices, to not show up in /sys/class/block/, but only in
1158 This switch is only active when the sysfs.deprecated=1 boot option is
1159 passed or the SYSFS_DEPRECATED_V2 option is set.
1161 This option allows new kernels to run on old distributions and tools,
1162 which might get confused by /sys/class/block/. Since 2007/2008 all
1163 major distributions and tools handle this just fine.
1165 Recent distributions and userspace tools after 2009/2010 depend on
1166 the existence of /sys/class/block/, and will not work with this
1169 Only if you are using a new kernel on an old distribution, you might
1172 config SYSFS_DEPRECATED_V2
1173 bool "Enable deprecated sysfs features by default"
1176 depends on SYSFS_DEPRECATED
1178 Enable deprecated sysfs by default.
1180 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1183 Only if you are using a new kernel on an old distribution, you might
1184 need to say Y here. Even then, odds are you would not need it
1185 enabled, you can always pass the boot option if absolutely necessary.
1188 bool "Kernel->user space relay support (formerly relayfs)"
1191 This option enables support for relay interface support in
1192 certain file systems (such as debugfs).
1193 It is designed to provide an efficient mechanism for tools and
1194 facilities to relay large amounts of data from kernel space to
1199 config BLK_DEV_INITRD
1200 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1202 The initial RAM filesystem is a ramfs which is loaded by the
1203 boot loader (loadlin or lilo) and that is mounted as root
1204 before the normal boot procedure. It is typically used to
1205 load modules needed to mount the "real" root file system,
1206 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1208 If RAM disk support (BLK_DEV_RAM) is also included, this
1209 also enables initial RAM disk (initrd) support and adds
1210 15 Kbytes (more on some other architectures) to the kernel size.
1216 source "usr/Kconfig"
1221 bool "Boot config support"
1222 select BLK_DEV_INITRD
1224 Extra boot config allows system admin to pass a config file as
1225 complemental extension of kernel cmdline when booting.
1226 The boot config file must be attached at the end of initramfs
1227 with checksum, size and magic word.
1228 See <file:Documentation/admin-guide/bootconfig.rst> for details.
1233 prompt "Compiler optimization level"
1234 default CC_OPTIMIZE_FOR_PERFORMANCE
1236 config CC_OPTIMIZE_FOR_PERFORMANCE
1237 bool "Optimize for performance (-O2)"
1239 This is the default optimization level for the kernel, building
1240 with the "-O2" compiler flag for best performance and most
1241 helpful compile-time warnings.
1243 config CC_OPTIMIZE_FOR_PERFORMANCE_O3
1244 bool "Optimize more for performance (-O3)"
1247 Choosing this option will pass "-O3" to your compiler to optimize
1248 the kernel yet more for performance.
1250 config CC_OPTIMIZE_FOR_SIZE
1251 bool "Optimize for size (-Os)"
1253 Choosing this option will pass "-Os" to your compiler resulting
1254 in a smaller kernel.
1258 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1261 This requires that the arch annotates or otherwise protects
1262 its external entry points from being discarded. Linker scripts
1263 must also merge .text.*, .data.*, and .bss.* correctly into
1264 output sections. Care must be taken not to pull in unrelated
1265 sections (e.g., '.text.init'). Typically '.' in section names
1266 is used to distinguish them from label names / C identifiers.
1268 config LD_DEAD_CODE_DATA_ELIMINATION
1269 bool "Dead code and data elimination (EXPERIMENTAL)"
1270 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1272 depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800)
1273 depends on $(cc-option,-ffunction-sections -fdata-sections)
1274 depends on $(ld-option,--gc-sections)
1276 Enable this if you want to do dead code and data elimination with
1277 the linker by compiling with -ffunction-sections -fdata-sections,
1278 and linking with --gc-sections.
1280 This can reduce on disk and in-memory size of the kernel
1281 code and static data, particularly for small configs and
1282 on small systems. This has the possibility of introducing
1283 silently broken kernel if the required annotations are not
1284 present. This option is not well tested yet, so use at your
1293 config SYSCTL_EXCEPTION_TRACE
1296 Enable support for /proc/sys/debug/exception-trace.
1298 config SYSCTL_ARCH_UNALIGN_NO_WARN
1301 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1302 Allows arch to define/use @no_unaligned_warning to possibly warn
1303 about unaligned access emulation going on under the hood.
1305 config SYSCTL_ARCH_UNALIGN_ALLOW
1308 Enable support for /proc/sys/kernel/unaligned-trap
1309 Allows arches to define/use @unaligned_enabled to runtime toggle
1310 the unaligned access emulation.
1311 see arch/parisc/kernel/unaligned.c for reference
1313 config HAVE_PCSPKR_PLATFORM
1316 # interpreter that classic socket filters depend on
1321 bool "Configure standard kernel features (expert users)"
1322 # Unhide debug options, to make the on-by-default options visible
1325 This option allows certain base kernel options and settings
1326 to be disabled or tweaked. This is for specialized
1327 environments which can tolerate a "non-standard" kernel.
1328 Only use this if you really know what you are doing.
1331 bool "Enable 16-bit UID system calls" if EXPERT
1332 depends on HAVE_UID16 && MULTIUSER
1335 This enables the legacy 16-bit UID syscall wrappers.
1338 bool "Multiple users, groups and capabilities support" if EXPERT
1341 This option enables support for non-root users, groups and
1344 If you say N here, all processes will run with UID 0, GID 0, and all
1345 possible capabilities. Saying N here also compiles out support for
1346 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1349 If unsure, say Y here.
1351 config SGETMASK_SYSCALL
1352 bool "sgetmask/ssetmask syscalls support" if EXPERT
1353 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1355 sys_sgetmask and sys_ssetmask are obsolete system calls
1356 no longer supported in libc but still enabled by default in some
1359 If unsure, leave the default option here.
1361 config SYSFS_SYSCALL
1362 bool "Sysfs syscall support" if EXPERT
1365 sys_sysfs is an obsolete system call no longer supported in libc.
1366 Note that disabling this option is more secure but might break
1367 compatibility with some systems.
1369 If unsure say Y here.
1372 bool "open by fhandle syscalls" if EXPERT
1376 If you say Y here, a user level program will be able to map
1377 file names to handle and then later use the handle for
1378 different file system operations. This is useful in implementing
1379 userspace file servers, which now track files using handles instead
1380 of names. The handle would remain the same even if file names
1381 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1385 bool "Posix Clocks & timers" if EXPERT
1388 This includes native support for POSIX timers to the kernel.
1389 Some embedded systems have no use for them and therefore they
1390 can be configured out to reduce the size of the kernel image.
1392 When this option is disabled, the following syscalls won't be
1393 available: timer_create, timer_gettime: timer_getoverrun,
1394 timer_settime, timer_delete, clock_adjtime, getitimer,
1395 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1396 clock_getres and clock_nanosleep syscalls will be limited to
1397 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1403 bool "Enable support for printk" if EXPERT
1406 This option enables normal printk support. Removing it
1407 eliminates most of the message strings from the kernel image
1408 and makes the kernel more or less silent. As this makes it
1409 very difficult to diagnose system problems, saying N here is
1410 strongly discouraged.
1418 bool "BUG() support" if EXPERT
1421 Disabling this option eliminates support for BUG and WARN, reducing
1422 the size of your kernel image and potentially quietly ignoring
1423 numerous fatal conditions. You should only consider disabling this
1424 option for embedded systems with no facilities for reporting errors.
1430 bool "Enable ELF core dumps" if EXPERT
1432 Enable support for generating core dumps. Disabling saves about 4k.
1435 config PCSPKR_PLATFORM
1436 bool "Enable PC-Speaker support" if EXPERT
1437 depends on HAVE_PCSPKR_PLATFORM
1441 This option allows to disable the internal PC-Speaker
1442 support, saving some memory.
1446 bool "Enable full-sized data structures for core" if EXPERT
1448 Disabling this option reduces the size of miscellaneous core
1449 kernel data structures. This saves memory on small machines,
1450 but may reduce performance.
1453 bool "Enable futex support" if EXPERT
1457 Disabling this option will cause the kernel to be built without
1458 support for "fast userspace mutexes". The resulting kernel may not
1459 run glibc-based applications correctly.
1463 depends on FUTEX && RT_MUTEXES
1466 config HAVE_FUTEX_CMPXCHG
1470 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1471 is implemented and always working. This removes a couple of runtime
1475 bool "Enable eventpoll support" if EXPERT
1478 Disabling this option will cause the kernel to be built without
1479 support for epoll family of system calls.
1482 bool "Enable signalfd() system call" if EXPERT
1485 Enable the signalfd() system call that allows to receive signals
1486 on a file descriptor.
1491 bool "Enable timerfd() system call" if EXPERT
1494 Enable the timerfd() system call that allows to receive timer
1495 events on a file descriptor.
1500 bool "Enable eventfd() system call" if EXPERT
1503 Enable the eventfd() system call that allows to receive both
1504 kernel notification (ie. KAIO) or userspace notifications.
1509 bool "Use full shmem filesystem" if EXPERT
1513 The shmem is an internal filesystem used to manage shared memory.
1514 It is backed by swap and manages resource limits. It is also exported
1515 to userspace as tmpfs if TMPFS is enabled. Disabling this
1516 option replaces shmem and tmpfs with the much simpler ramfs code,
1517 which may be appropriate on small systems without swap.
1520 bool "Enable AIO support" if EXPERT
1523 This option enables POSIX asynchronous I/O which may by used
1524 by some high performance threaded applications. Disabling
1525 this option saves about 7k.
1528 bool "Enable IO uring support" if EXPERT
1532 This option enables support for the io_uring interface, enabling
1533 applications to submit and complete IO through submission and
1534 completion rings that are shared between the kernel and application.
1536 config ADVISE_SYSCALLS
1537 bool "Enable madvise/fadvise syscalls" if EXPERT
1540 This option enables the madvise and fadvise syscalls, used by
1541 applications to advise the kernel about their future memory or file
1542 usage, improving performance. If building an embedded system where no
1543 applications use these syscalls, you can disable this option to save
1546 config HAVE_ARCH_USERFAULTFD_WP
1549 Arch has userfaultfd write protection support
1552 bool "Enable membarrier() system call" if EXPERT
1555 Enable the membarrier() system call that allows issuing memory
1556 barriers across all running threads, which can be used to distribute
1557 the cost of user-space memory barriers asymmetrically by transforming
1558 pairs of memory barriers into pairs consisting of membarrier() and a
1564 bool "Load all symbols for debugging/ksymoops" if EXPERT
1567 Say Y here to let the kernel print out symbolic crash information and
1568 symbolic stack backtraces. This increases the size of the kernel
1569 somewhat, as all symbols have to be loaded into the kernel image.
1572 bool "Include all symbols in kallsyms"
1573 depends on DEBUG_KERNEL && KALLSYMS
1575 Normally kallsyms only contains the symbols of functions for nicer
1576 OOPS messages and backtraces (i.e., symbols from the text and inittext
1577 sections). This is sufficient for most cases. And only in very rare
1578 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1579 names of variables from the data sections, etc).
1581 This option makes sure that all symbols are loaded into the kernel
1582 image (i.e., symbols from all sections) in cost of increased kernel
1583 size (depending on the kernel configuration, it may be 300KiB or
1584 something like this).
1586 Say N unless you really need all symbols.
1588 config KALLSYMS_ABSOLUTE_PERCPU
1591 default X86_64 && SMP
1593 config KALLSYMS_BASE_RELATIVE
1598 Instead of emitting them as absolute values in the native word size,
1599 emit the symbol references in the kallsyms table as 32-bit entries,
1600 each containing a relative value in the range [base, base + U32_MAX]
1601 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1602 an absolute value in the range [0, S32_MAX] or a relative value in the
1603 range [base, base + S32_MAX], where base is the lowest relative symbol
1604 address encountered in the image.
1606 On 64-bit builds, this reduces the size of the address table by 50%,
1607 but more importantly, it results in entries whose values are build
1608 time constants, and no relocation pass is required at runtime to fix
1609 up the entries based on the runtime load address of the kernel.
1611 # end of the "standard kernel features (expert users)" menu
1613 # syscall, maps, verifier
1616 bool "LSM Instrumentation with BPF"
1617 depends on BPF_EVENTS
1618 depends on BPF_SYSCALL
1622 Enables instrumentation of the security hooks with eBPF programs for
1623 implementing dynamic MAC and Audit Policies.
1625 If you are unsure how to answer this question, answer N.
1628 bool "Enable bpf() system call"
1633 Enable the bpf() system call that allows to manipulate eBPF
1634 programs and maps via file descriptors.
1636 config ARCH_WANT_DEFAULT_BPF_JIT
1639 config BPF_JIT_ALWAYS_ON
1640 bool "Permanently enable BPF JIT and remove BPF interpreter"
1641 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1643 Enables BPF JIT and removes BPF interpreter to avoid
1644 speculative execution of BPF instructions by the interpreter
1646 config BPF_JIT_DEFAULT_ON
1647 def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
1648 depends on HAVE_EBPF_JIT && BPF_JIT
1651 bool "Enable userfaultfd() system call"
1654 Enable the userfaultfd() system call that allows to intercept and
1655 handle page faults in userland.
1657 config ARCH_HAS_MEMBARRIER_CALLBACKS
1660 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1664 bool "Enable rseq() system call" if EXPERT
1666 depends on HAVE_RSEQ
1669 Enable the restartable sequences system call. It provides a
1670 user-space cache for the current CPU number value, which
1671 speeds up getting the current CPU number from user-space,
1672 as well as an ABI to speed up user-space operations on
1679 bool "Enabled debugging of rseq() system call" if EXPERT
1680 depends on RSEQ && DEBUG_KERNEL
1682 Enable extra debugging checks for the rseq system call.
1687 bool "Embedded system"
1688 option allnoconfig_y
1691 This option should be enabled if compiling the kernel for
1692 an embedded system so certain expert options are available
1695 config HAVE_PERF_EVENTS
1698 See tools/perf/design.txt for details.
1700 config PERF_USE_VMALLOC
1703 See tools/perf/design.txt for details
1706 bool "PC/104 support" if EXPERT
1708 Expose PC/104 form factor device drivers and options available for
1709 selection and configuration. Enable this option if your target
1710 machine has a PC/104 bus.
1712 menu "Kernel Performance Events And Counters"
1715 bool "Kernel performance events and counters"
1716 default y if PROFILING
1717 depends on HAVE_PERF_EVENTS
1721 Enable kernel support for various performance events provided
1722 by software and hardware.
1724 Software events are supported either built-in or via the
1725 use of generic tracepoints.
1727 Most modern CPUs support performance events via performance
1728 counter registers. These registers count the number of certain
1729 types of hw events: such as instructions executed, cachemisses
1730 suffered, or branches mis-predicted - without slowing down the
1731 kernel or applications. These registers can also trigger interrupts
1732 when a threshold number of events have passed - and can thus be
1733 used to profile the code that runs on that CPU.
1735 The Linux Performance Event subsystem provides an abstraction of
1736 these software and hardware event capabilities, available via a
1737 system call and used by the "perf" utility in tools/perf/. It
1738 provides per task and per CPU counters, and it provides event
1739 capabilities on top of those.
1743 config DEBUG_PERF_USE_VMALLOC
1745 bool "Debug: use vmalloc to back perf mmap() buffers"
1746 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1747 select PERF_USE_VMALLOC
1749 Use vmalloc memory to back perf mmap() buffers.
1751 Mostly useful for debugging the vmalloc code on platforms
1752 that don't require it.
1758 config VM_EVENT_COUNTERS
1760 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1762 VM event counters are needed for event counts to be shown.
1763 This option allows the disabling of the VM event counters
1764 on EXPERT systems. /proc/vmstat will only show page counts
1765 if VM event counters are disabled.
1769 bool "Enable SLUB debugging support" if EXPERT
1770 depends on SLUB && SYSFS
1772 SLUB has extensive debug support features. Disabling these can
1773 result in significant savings in code size. This also disables
1774 SLUB sysfs support. /sys/slab will not exist and there will be
1775 no support for cache validation etc.
1777 config SLUB_MEMCG_SYSFS_ON
1779 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1780 depends on SLUB && SYSFS && MEMCG
1782 SLUB creates a directory under /sys/kernel/slab for each
1783 allocation cache to host info and debug files. If memory
1784 cgroup is enabled, each cache can have per memory cgroup
1785 caches. SLUB can create the same sysfs directories for these
1786 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1787 to a very high number of debug files being created. This is
1788 controlled by slub_memcg_sysfs boot parameter and this
1789 config option determines the parameter's default value.
1792 bool "Disable heap randomization"
1795 Randomizing heap placement makes heap exploits harder, but it
1796 also breaks ancient binaries (including anything libc5 based).
1797 This option changes the bootup default to heap randomization
1798 disabled, and can be overridden at runtime by setting
1799 /proc/sys/kernel/randomize_va_space to 2.
1801 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1804 prompt "Choose SLAB allocator"
1807 This option allows to select a slab allocator.
1811 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1813 The regular slab allocator that is established and known to work
1814 well in all environments. It organizes cache hot objects in
1815 per cpu and per node queues.
1818 bool "SLUB (Unqueued Allocator)"
1819 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1821 SLUB is a slab allocator that minimizes cache line usage
1822 instead of managing queues of cached objects (SLAB approach).
1823 Per cpu caching is realized using slabs of objects instead
1824 of queues of objects. SLUB can use memory efficiently
1825 and has enhanced diagnostics. SLUB is the default choice for
1830 bool "SLOB (Simple Allocator)"
1832 SLOB replaces the stock allocator with a drastically simpler
1833 allocator. SLOB is generally more space efficient but
1834 does not perform as well on large systems.
1838 config SLAB_MERGE_DEFAULT
1839 bool "Allow slab caches to be merged"
1842 For reduced kernel memory fragmentation, slab caches can be
1843 merged when they share the same size and other characteristics.
1844 This carries a risk of kernel heap overflows being able to
1845 overwrite objects from merged caches (and more easily control
1846 cache layout), which makes such heap attacks easier to exploit
1847 by attackers. By keeping caches unmerged, these kinds of exploits
1848 can usually only damage objects in the same cache. To disable
1849 merging at runtime, "slab_nomerge" can be passed on the kernel
1852 config SLAB_FREELIST_RANDOM
1854 depends on SLAB || SLUB
1855 bool "SLAB freelist randomization"
1857 Randomizes the freelist order used on creating new pages. This
1858 security feature reduces the predictability of the kernel slab
1859 allocator against heap overflows.
1861 config SLAB_FREELIST_HARDENED
1862 bool "Harden slab freelist metadata"
1865 Many kernel heap attacks try to target slab cache metadata and
1866 other infrastructure. This options makes minor performance
1867 sacrifices to harden the kernel slab allocator against common
1868 freelist exploit methods.
1870 config SHUFFLE_PAGE_ALLOCATOR
1871 bool "Page allocator randomization"
1872 default SLAB_FREELIST_RANDOM && ACPI_NUMA
1874 Randomization of the page allocator improves the average
1875 utilization of a direct-mapped memory-side-cache. See section
1876 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
1877 6.2a specification for an example of how a platform advertises
1878 the presence of a memory-side-cache. There are also incidental
1879 security benefits as it reduces the predictability of page
1880 allocations to compliment SLAB_FREELIST_RANDOM, but the
1881 default granularity of shuffling on the "MAX_ORDER - 1" i.e,
1882 10th order of pages is selected based on cache utilization
1885 While the randomization improves cache utilization it may
1886 negatively impact workloads on platforms without a cache. For
1887 this reason, by default, the randomization is enabled only
1888 after runtime detection of a direct-mapped memory-side-cache.
1889 Otherwise, the randomization may be force enabled with the
1890 'page_alloc.shuffle' kernel command line parameter.
1894 config SLUB_CPU_PARTIAL
1896 depends on SLUB && SMP
1897 bool "SLUB per cpu partial cache"
1899 Per cpu partial caches accelerate objects allocation and freeing
1900 that is local to a processor at the price of more indeterminism
1901 in the latency of the free. On overflow these caches will be cleared
1902 which requires the taking of locks that may cause latency spikes.
1903 Typically one would choose no for a realtime system.
1905 config MMAP_ALLOW_UNINITIALIZED
1906 bool "Allow mmapped anonymous memory to be uninitialized"
1907 depends on EXPERT && !MMU
1910 Normally, and according to the Linux spec, anonymous memory obtained
1911 from mmap() has its contents cleared before it is passed to
1912 userspace. Enabling this config option allows you to request that
1913 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1914 providing a huge performance boost. If this option is not enabled,
1915 then the flag will be ignored.
1917 This is taken advantage of by uClibc's malloc(), and also by
1918 ELF-FDPIC binfmt's brk and stack allocator.
1920 Because of the obvious security issues, this option should only be
1921 enabled on embedded devices where you control what is run in
1922 userspace. Since that isn't generally a problem on no-MMU systems,
1923 it is normally safe to say Y here.
1925 See Documentation/nommu-mmap.txt for more information.
1927 config SYSTEM_DATA_VERIFICATION
1929 select SYSTEM_TRUSTED_KEYRING
1933 select ASYMMETRIC_KEY_TYPE
1934 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1937 select X509_CERTIFICATE_PARSER
1938 select PKCS7_MESSAGE_PARSER
1940 Provide PKCS#7 message verification using the contents of the system
1941 trusted keyring to provide public keys. This then can be used for
1942 module verification, kexec image verification and firmware blob
1946 bool "Profiling support"
1948 Say Y here to enable the extended profiling support mechanisms used
1949 by profilers such as OProfile.
1952 # Place an empty function call at each tracepoint site. Can be
1953 # dynamically changed for a probe function.
1958 endmenu # General setup
1960 source "arch/Kconfig"
1967 default 0 if BASE_FULL
1968 default 1 if !BASE_FULL
1970 config MODULE_SIG_FORMAT
1972 select SYSTEM_DATA_VERIFICATION
1975 bool "Enable loadable module support"
1978 Kernel modules are small pieces of compiled code which can
1979 be inserted in the running kernel, rather than being
1980 permanently built into the kernel. You use the "modprobe"
1981 tool to add (and sometimes remove) them. If you say Y here,
1982 many parts of the kernel can be built as modules (by
1983 answering M instead of Y where indicated): this is most
1984 useful for infrequently used options which are not required
1985 for booting. For more information, see the man pages for
1986 modprobe, lsmod, modinfo, insmod and rmmod.
1988 If you say Y here, you will need to run "make
1989 modules_install" to put the modules under /lib/modules/
1990 where modprobe can find them (you may need to be root to do
1997 config MODULE_FORCE_LOAD
1998 bool "Forced module loading"
2001 Allow loading of modules without version information (ie. modprobe
2002 --force). Forced module loading sets the 'F' (forced) taint flag and
2003 is usually a really bad idea.
2005 config MODULE_UNLOAD
2006 bool "Module unloading"
2008 Without this option you will not be able to unload any
2009 modules (note that some modules may not be unloadable
2010 anyway), which makes your kernel smaller, faster
2011 and simpler. If unsure, say Y.
2013 config MODULE_FORCE_UNLOAD
2014 bool "Forced module unloading"
2015 depends on MODULE_UNLOAD
2017 This option allows you to force a module to unload, even if the
2018 kernel believes it is unsafe: the kernel will remove the module
2019 without waiting for anyone to stop using it (using the -f option to
2020 rmmod). This is mainly for kernel developers and desperate users.
2024 bool "Module versioning support"
2026 Usually, you have to use modules compiled with your kernel.
2027 Saying Y here makes it sometimes possible to use modules
2028 compiled for different kernels, by adding enough information
2029 to the modules to (hopefully) spot any changes which would
2030 make them incompatible with the kernel you are running. If
2033 config ASM_MODVERSIONS
2035 default HAVE_ASM_MODVERSIONS && MODVERSIONS
2037 This enables module versioning for exported symbols also from
2038 assembly. This can be enabled only when the target architecture
2041 config MODULE_REL_CRCS
2043 depends on MODVERSIONS
2045 config MODULE_SRCVERSION_ALL
2046 bool "Source checksum for all modules"
2048 Modules which contain a MODULE_VERSION get an extra "srcversion"
2049 field inserted into their modinfo section, which contains a
2050 sum of the source files which made it. This helps maintainers
2051 see exactly which source was used to build a module (since
2052 others sometimes change the module source without updating
2053 the version). With this option, such a "srcversion" field
2054 will be created for all modules. If unsure, say N.
2057 bool "Module signature verification"
2058 select MODULE_SIG_FORMAT
2060 Check modules for valid signatures upon load: the signature
2061 is simply appended to the module. For more information see
2062 <file:Documentation/admin-guide/module-signing.rst>.
2064 Note that this option adds the OpenSSL development packages as a
2065 kernel build dependency so that the signing tool can use its crypto
2068 You should enable this option if you wish to use either
2069 CONFIG_SECURITY_LOCKDOWN_LSM or lockdown functionality imposed via
2070 another LSM - otherwise unsigned modules will be loadable regardless
2071 of the lockdown policy.
2073 !!!WARNING!!! If you enable this option, you MUST make sure that the
2074 module DOES NOT get stripped after being signed. This includes the
2075 debuginfo strip done by some packagers (such as rpmbuild) and
2076 inclusion into an initramfs that wants the module size reduced.
2078 config MODULE_SIG_FORCE
2079 bool "Require modules to be validly signed"
2080 depends on MODULE_SIG
2082 Reject unsigned modules or signed modules for which we don't have a
2083 key. Without this, such modules will simply taint the kernel.
2085 config MODULE_SIG_ALL
2086 bool "Automatically sign all modules"
2088 depends on MODULE_SIG
2090 Sign all modules during make modules_install. Without this option,
2091 modules must be signed manually, using the scripts/sign-file tool.
2093 comment "Do not forget to sign required modules with scripts/sign-file"
2094 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
2097 prompt "Which hash algorithm should modules be signed with?"
2098 depends on MODULE_SIG
2100 This determines which sort of hashing algorithm will be used during
2101 signature generation. This algorithm _must_ be built into the kernel
2102 directly so that signature verification can take place. It is not
2103 possible to load a signed module containing the algorithm to check
2104 the signature on that module.
2106 config MODULE_SIG_SHA1
2107 bool "Sign modules with SHA-1"
2110 config MODULE_SIG_SHA224
2111 bool "Sign modules with SHA-224"
2112 select CRYPTO_SHA256
2114 config MODULE_SIG_SHA256
2115 bool "Sign modules with SHA-256"
2116 select CRYPTO_SHA256
2118 config MODULE_SIG_SHA384
2119 bool "Sign modules with SHA-384"
2120 select CRYPTO_SHA512
2122 config MODULE_SIG_SHA512
2123 bool "Sign modules with SHA-512"
2124 select CRYPTO_SHA512
2128 config MODULE_SIG_HASH
2130 depends on MODULE_SIG
2131 default "sha1" if MODULE_SIG_SHA1
2132 default "sha224" if MODULE_SIG_SHA224
2133 default "sha256" if MODULE_SIG_SHA256
2134 default "sha384" if MODULE_SIG_SHA384
2135 default "sha512" if MODULE_SIG_SHA512
2137 config MODULE_COMPRESS
2138 bool "Compress modules on installation"
2141 Compresses kernel modules when 'make modules_install' is run; gzip or
2142 xz depending on "Compression algorithm" below.
2144 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
2146 Out-of-tree kernel modules installed using Kbuild will also be
2147 compressed upon installation.
2149 Note: for modules inside an initrd or initramfs, it's more efficient
2150 to compress the whole initrd or initramfs instead.
2152 Note: This is fully compatible with signed modules.
2157 prompt "Compression algorithm"
2158 depends on MODULE_COMPRESS
2159 default MODULE_COMPRESS_GZIP
2161 This determines which sort of compression will be used during
2162 'make modules_install'.
2164 GZIP (default) and XZ are supported.
2166 config MODULE_COMPRESS_GZIP
2169 config MODULE_COMPRESS_XZ
2174 config MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS
2175 bool "Allow loading of modules with missing namespace imports"
2177 Symbols exported with EXPORT_SYMBOL_NS*() are considered exported in
2178 a namespace. A module that makes use of a symbol exported with such a
2179 namespace is required to import the namespace via MODULE_IMPORT_NS().
2180 There is no technical reason to enforce correct namespace imports,
2181 but it creates consistency between symbols defining namespaces and
2182 users importing namespaces they make use of. This option relaxes this
2183 requirement and lifts the enforcement when loading a module.
2187 config UNUSED_SYMBOLS
2188 bool "Enable unused/obsolete exported symbols"
2191 Unused but exported symbols make the kernel needlessly bigger. For
2192 that reason most of these unused exports will soon be removed. This
2193 option is provided temporarily to provide a transition period in case
2194 some external kernel module needs one of these symbols anyway. If you
2195 encounter such a case in your module, consider if you are actually
2196 using the right API. (rationale: since nobody in the kernel is using
2197 this in a module, there is a pretty good chance it's actually the
2198 wrong interface to use). If you really need the symbol, please send a
2199 mail to the linux kernel mailing list mentioning the symbol and why
2200 you really need it, and what the merge plan to the mainline kernel for
2203 config TRIM_UNUSED_KSYMS
2204 bool "Trim unused exported kernel symbols"
2205 depends on !UNUSED_SYMBOLS
2207 The kernel and some modules make many symbols available for
2208 other modules to use via EXPORT_SYMBOL() and variants. Depending
2209 on the set of modules being selected in your kernel configuration,
2210 many of those exported symbols might never be used.
2212 This option allows for unused exported symbols to be dropped from
2213 the build. In turn, this provides the compiler more opportunities
2214 (especially when using LTO) for optimizing the code and reducing
2215 binary size. This might have some security advantages as well.
2217 If unsure, or if you need to build out-of-tree modules, say N.
2219 config UNUSED_KSYMS_WHITELIST
2220 string "Whitelist of symbols to keep in ksymtab"
2221 depends on TRIM_UNUSED_KSYMS
2223 By default, all unused exported symbols will be un-exported from the
2224 build when TRIM_UNUSED_KSYMS is selected.
2226 UNUSED_KSYMS_WHITELIST allows to whitelist symbols that must be kept
2227 exported at all times, even in absence of in-tree users. The value to
2228 set here is the path to a text file containing the list of symbols,
2229 one per line. The path can be absolute, or relative to the kernel
2234 config MODULES_TREE_LOOKUP
2236 depends on PERF_EVENTS || TRACING
2238 config INIT_ALL_POSSIBLE
2241 Back when each arch used to define their own cpu_online_mask and
2242 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2243 with all 1s, and others with all 0s. When they were centralised,
2244 it was better to provide this option than to break all the archs
2245 and have several arch maintainers pursuing me down dark alleys.
2247 source "block/Kconfig"
2249 config PREEMPT_NOTIFIERS
2259 Build a simple ASN.1 grammar compiler that produces a bytecode output
2260 that can be interpreted by the ASN.1 stream decoder and used to
2261 inform it as to what tags are to be expected in a stream and what
2262 functions to call on what tags.
2264 source "kernel/Kconfig.locks"
2266 config ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
2269 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2272 # It may be useful for an architecture to override the definitions of the
2273 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2274 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2275 # different calling convention for syscalls. They can also override the
2276 # macros for not-implemented syscalls in kernel/sys_ni.c and
2277 # kernel/time/posix-stubs.c. All these overrides need to be available in
2278 # <asm/syscall_wrapper.h>.
2279 config ARCH_HAS_SYSCALL_WRAPPER