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)"
11 config CC_VERSION_TEXT
13 default "$(CC_VERSION_TEXT)"
15 This is used in unclear ways:
17 - Re-run Kconfig when the compiler is updated
18 The 'default' property references the environment variable,
19 CC_VERSION_TEXT so it is recorded in include/config/auto.conf.cmd.
20 When the compiler is updated, Kconfig will be invoked.
22 - Ensure full rebuild when the compier is updated
23 include/linux/kconfig.h contains this option in the comment line so
24 fixdep adds include/config/cc/version/text.h into the auto-generated
25 dependency. When the compiler is updated, syncconfig will touch it
26 and then every file will be rebuilt.
29 def_bool $(success,echo "$(CC_VERSION_TEXT)" | grep -q gcc)
33 default $(shell,$(srctree)/scripts/gcc-version.sh $(CC)) if CC_IS_GCC
38 default $(shell,$(LD) --version | $(srctree)/scripts/ld-version.sh)
41 def_bool $(success,echo "$(CC_VERSION_TEXT)" | grep -q clang)
44 def_bool $(success,$(LD) -v | head -n 1 | grep -q LLD)
48 default $(shell,$(srctree)/scripts/clang-version.sh $(CC))
52 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(m64-flag)) if 64BIT
53 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(m32-flag))
55 config CC_CAN_LINK_STATIC
57 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) -static $(m64-flag)) if 64BIT
58 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) -static $(m32-flag))
60 config CC_HAS_ASM_GOTO
61 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
63 config CC_HAS_ASM_GOTO_OUTPUT
64 depends on CC_HAS_ASM_GOTO
65 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)
67 config TOOLS_SUPPORT_RELR
68 def_bool $(success,env "CC=$(CC)" "LD=$(LD)" "NM=$(NM)" "OBJCOPY=$(OBJCOPY)" $(srctree)/scripts/tools-support-relr.sh)
70 config CC_HAS_ASM_INLINE
71 def_bool $(success,echo 'void foo(void) { asm inline (""); }' | $(CC) -x c - -c -o /dev/null)
80 config BUILDTIME_TABLE_SORT
83 config THREAD_INFO_IN_TASK
86 Select this to move thread_info off the stack into task_struct. To
87 make this work, an arch will need to remove all thread_info fields
88 except flags and fix any runtime bugs.
90 One subtle change that will be needed is to use try_get_task_stack()
91 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
100 depends on BROKEN || !SMP
103 config INIT_ENV_ARG_LIMIT
108 Maximum of each of the number of arguments and environment
109 variables passed to init from the kernel command line.
112 bool "Compile also drivers which will not load"
116 Some drivers can be compiled on a different platform than they are
117 intended to be run on. Despite they cannot be loaded there (or even
118 when they load they cannot be used due to missing HW support),
119 developers still, opposing to distributors, might want to build such
120 drivers to compile-test them.
122 If you are a developer and want to build everything available, say Y
123 here. If you are a user/distributor, say N here to exclude useless
124 drivers to be distributed.
126 config UAPI_HEADER_TEST
127 bool "Compile test UAPI headers"
128 depends on HEADERS_INSTALL && CC_CAN_LINK
130 Compile test headers exported to user-space to ensure they are
131 self-contained, i.e. compilable as standalone units.
133 If you are a developer or tester and want to ensure the exported
134 headers are self-contained, say Y here. Otherwise, choose N.
137 string "Local version - append to kernel release"
139 Append an extra string to the end of your kernel version.
140 This will show up when you type uname, for example.
141 The string you set here will be appended after the contents of
142 any files with a filename matching localversion* in your
143 object and source tree, in that order. Your total string can
144 be a maximum of 64 characters.
146 config LOCALVERSION_AUTO
147 bool "Automatically append version information to the version string"
149 depends on !COMPILE_TEST
151 This will try to automatically determine if the current tree is a
152 release tree by looking for git tags that belong to the current
153 top of tree revision.
155 A string of the format -gxxxxxxxx will be added to the localversion
156 if a git-based tree is found. The string generated by this will be
157 appended after any matching localversion* files, and after the value
158 set in CONFIG_LOCALVERSION.
160 (The actual string used here is the first eight characters produced
161 by running the command:
163 $ git rev-parse --verify HEAD
165 which is done within the script "scripts/setlocalversion".)
168 string "Build ID Salt"
171 The build ID is used to link binaries and their debug info. Setting
172 this option will use the value in the calculation of the build id.
173 This is mostly useful for distributions which want to ensure the
174 build is unique between builds. It's safe to leave the default.
176 config HAVE_KERNEL_GZIP
179 config HAVE_KERNEL_BZIP2
182 config HAVE_KERNEL_LZMA
185 config HAVE_KERNEL_XZ
188 config HAVE_KERNEL_LZO
191 config HAVE_KERNEL_LZ4
194 config HAVE_KERNEL_UNCOMPRESSED
198 prompt "Kernel compression mode"
200 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
202 The linux kernel is a kind of self-extracting executable.
203 Several compression algorithms are available, which differ
204 in efficiency, compression and decompression speed.
205 Compression speed is only relevant when building a kernel.
206 Decompression speed is relevant at each boot.
208 If you have any problems with bzip2 or lzma compressed
209 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
210 version of this functionality (bzip2 only), for 2.4, was
211 supplied by Christian Ludwig)
213 High compression options are mostly useful for users, who
214 are low on disk space (embedded systems), but for whom ram
217 If in doubt, select 'gzip'
221 depends on HAVE_KERNEL_GZIP
223 The old and tried gzip compression. It provides a good balance
224 between compression ratio and decompression speed.
228 depends on HAVE_KERNEL_BZIP2
230 Its compression ratio and speed is intermediate.
231 Decompression speed is slowest among the choices. The kernel
232 size is about 10% smaller with bzip2, in comparison to gzip.
233 Bzip2 uses a large amount of memory. For modern kernels you
234 will need at least 8MB RAM or more for booting.
238 depends on HAVE_KERNEL_LZMA
240 This compression algorithm's ratio is best. Decompression speed
241 is between gzip and bzip2. Compression is slowest.
242 The kernel size is about 33% smaller with LZMA in comparison to gzip.
246 depends on HAVE_KERNEL_XZ
248 XZ uses the LZMA2 algorithm and instruction set specific
249 BCJ filters which can improve compression ratio of executable
250 code. The size of the kernel is about 30% smaller with XZ in
251 comparison to gzip. On architectures for which there is a BCJ
252 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
253 will create a few percent smaller kernel than plain LZMA.
255 The speed is about the same as with LZMA: The decompression
256 speed of XZ is better than that of bzip2 but worse than gzip
257 and LZO. Compression is slow.
261 depends on HAVE_KERNEL_LZO
263 Its compression ratio is the poorest among the choices. The kernel
264 size is about 10% bigger than gzip; however its speed
265 (both compression and decompression) is the fastest.
269 depends on HAVE_KERNEL_LZ4
271 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
272 A preliminary version of LZ4 de/compression tool is available at
273 <https://code.google.com/p/lz4/>.
275 Its compression ratio is worse than LZO. The size of the kernel
276 is about 8% bigger than LZO. But the decompression speed is
279 config KERNEL_UNCOMPRESSED
281 depends on HAVE_KERNEL_UNCOMPRESSED
283 Produce uncompressed kernel image. This option is usually not what
284 you want. It is useful for debugging the kernel in slow simulation
285 environments, where decompressing and moving the kernel is awfully
286 slow. This option allows early boot code to skip the decompressor
287 and jump right at uncompressed kernel image.
292 string "Default init path"
295 This option determines the default init for the system if no init=
296 option is passed on the kernel command line. If the requested path is
297 not present, we will still then move on to attempting further
298 locations (e.g. /sbin/init, etc). If this is empty, we will just use
299 the fallback list when init= is not passed.
301 config DEFAULT_HOSTNAME
302 string "Default hostname"
305 This option determines the default system hostname before userspace
306 calls sethostname(2). The kernel traditionally uses "(none)" here,
307 but you may wish to use a different default here to make a minimal
308 system more usable with less configuration.
311 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
312 # add proper SWAP support to them, in which case this can be remove.
318 bool "Support for paging of anonymous memory (swap)"
319 depends on MMU && BLOCK && !ARCH_NO_SWAP
322 This option allows you to choose whether you want to have support
323 for so called swap devices or swap files in your kernel that are
324 used to provide more virtual memory than the actual RAM present
325 in your computer. If unsure say Y.
330 Inter Process Communication is a suite of library functions and
331 system calls which let processes (running programs) synchronize and
332 exchange information. It is generally considered to be a good thing,
333 and some programs won't run unless you say Y here. In particular, if
334 you want to run the DOS emulator dosemu under Linux (read the
335 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
336 you'll need to say Y here.
338 You can find documentation about IPC with "info ipc" and also in
339 section 6.4 of the Linux Programmer's Guide, available from
340 <http://www.tldp.org/guides.html>.
342 config SYSVIPC_SYSCTL
349 bool "POSIX Message Queues"
352 POSIX variant of message queues is a part of IPC. In POSIX message
353 queues every message has a priority which decides about succession
354 of receiving it by a process. If you want to compile and run
355 programs written e.g. for Solaris with use of its POSIX message
356 queues (functions mq_*) say Y here.
358 POSIX message queues are visible as a filesystem called 'mqueue'
359 and can be mounted somewhere if you want to do filesystem
360 operations on message queues.
364 config POSIX_MQUEUE_SYSCTL
366 depends on POSIX_MQUEUE
370 config CROSS_MEMORY_ATTACH
371 bool "Enable process_vm_readv/writev syscalls"
375 Enabling this option adds the system calls process_vm_readv and
376 process_vm_writev which allow a process with the correct privileges
377 to directly read from or write to another process' address space.
378 See the man page for more details.
381 bool "uselib syscall"
382 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
384 This option enables the uselib syscall, a system call used in the
385 dynamic linker from libc5 and earlier. glibc does not use this
386 system call. If you intend to run programs built on libc5 or
387 earlier, you may need to enable this syscall. Current systems
388 running glibc can safely disable this.
391 bool "Auditing support"
394 Enable auditing infrastructure that can be used with another
395 kernel subsystem, such as SELinux (which requires this for
396 logging of avc messages output). System call auditing is included
397 on architectures which support it.
399 config HAVE_ARCH_AUDITSYSCALL
404 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
407 source "kernel/irq/Kconfig"
408 source "kernel/time/Kconfig"
409 source "kernel/Kconfig.preempt"
411 menu "CPU/Task time and stats accounting"
413 config VIRT_CPU_ACCOUNTING
417 prompt "Cputime accounting"
418 default TICK_CPU_ACCOUNTING if !PPC64
419 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
421 # Kind of a stub config for the pure tick based cputime accounting
422 config TICK_CPU_ACCOUNTING
423 bool "Simple tick based cputime accounting"
424 depends on !S390 && !NO_HZ_FULL
426 This is the basic tick based cputime accounting that maintains
427 statistics about user, system and idle time spent on per jiffies
432 config VIRT_CPU_ACCOUNTING_NATIVE
433 bool "Deterministic task and CPU time accounting"
434 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
435 select VIRT_CPU_ACCOUNTING
437 Select this option to enable more accurate task and CPU time
438 accounting. This is done by reading a CPU counter on each
439 kernel entry and exit and on transitions within the kernel
440 between system, softirq and hardirq state, so there is a
441 small performance impact. In the case of s390 or IBM POWER > 5,
442 this also enables accounting of stolen time on logically-partitioned
445 config VIRT_CPU_ACCOUNTING_GEN
446 bool "Full dynticks CPU time accounting"
447 depends on HAVE_CONTEXT_TRACKING
448 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
449 depends on GENERIC_CLOCKEVENTS
450 select VIRT_CPU_ACCOUNTING
451 select CONTEXT_TRACKING
453 Select this option to enable task and CPU time accounting on full
454 dynticks systems. This accounting is implemented by watching every
455 kernel-user boundaries using the context tracking subsystem.
456 The accounting is thus performed at the expense of some significant
459 For now this is only useful if you are working on the full
460 dynticks subsystem development.
466 config IRQ_TIME_ACCOUNTING
467 bool "Fine granularity task level IRQ time accounting"
468 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
470 Select this option to enable fine granularity task irq time
471 accounting. This is done by reading a timestamp on each
472 transitions between softirq and hardirq state, so there can be a
473 small performance impact.
475 If in doubt, say N here.
477 config HAVE_SCHED_AVG_IRQ
479 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
482 config SCHED_THERMAL_PRESSURE
483 bool "Enable periodic averaging of thermal pressure"
486 config BSD_PROCESS_ACCT
487 bool "BSD Process Accounting"
490 If you say Y here, a user level program will be able to instruct the
491 kernel (via a special system call) to write process accounting
492 information to a file: whenever a process exits, information about
493 that process will be appended to the file by the kernel. The
494 information includes things such as creation time, owning user,
495 command name, memory usage, controlling terminal etc. (the complete
496 list is in the struct acct in <file:include/linux/acct.h>). It is
497 up to the user level program to do useful things with this
498 information. This is generally a good idea, so say Y.
500 config BSD_PROCESS_ACCT_V3
501 bool "BSD Process Accounting version 3 file format"
502 depends on BSD_PROCESS_ACCT
505 If you say Y here, the process accounting information is written
506 in a new file format that also logs the process IDs of each
507 process and its parent. Note that this file format is incompatible
508 with previous v0/v1/v2 file formats, so you will need updated tools
509 for processing it. A preliminary version of these tools is available
510 at <http://www.gnu.org/software/acct/>.
513 bool "Export task/process statistics through netlink"
518 Export selected statistics for tasks/processes through the
519 generic netlink interface. Unlike BSD process accounting, the
520 statistics are available during the lifetime of tasks/processes as
521 responses to commands. Like BSD accounting, they are sent to user
526 config TASK_DELAY_ACCT
527 bool "Enable per-task delay accounting"
531 Collect information on time spent by a task waiting for system
532 resources like cpu, synchronous block I/O completion and swapping
533 in pages. Such statistics can help in setting a task's priorities
534 relative to other tasks for cpu, io, rss limits etc.
539 bool "Enable extended accounting over taskstats"
542 Collect extended task accounting data and send the data
543 to userland for processing over the taskstats interface.
547 config TASK_IO_ACCOUNTING
548 bool "Enable per-task storage I/O accounting"
549 depends on TASK_XACCT
551 Collect information on the number of bytes of storage I/O which this
557 bool "Pressure stall information tracking"
559 Collect metrics that indicate how overcommitted the CPU, memory,
560 and IO capacity are in the system.
562 If you say Y here, the kernel will create /proc/pressure/ with the
563 pressure statistics files cpu, memory, and io. These will indicate
564 the share of walltime in which some or all tasks in the system are
565 delayed due to contention of the respective resource.
567 In kernels with cgroup support, cgroups (cgroup2 only) will
568 have cpu.pressure, memory.pressure, and io.pressure files,
569 which aggregate pressure stalls for the grouped tasks only.
571 For more details see Documentation/accounting/psi.rst.
575 config PSI_DEFAULT_DISABLED
576 bool "Require boot parameter to enable pressure stall information tracking"
580 If set, pressure stall information tracking will be disabled
581 per default but can be enabled through passing psi=1 on the
582 kernel commandline during boot.
584 This feature adds some code to the task wakeup and sleep
585 paths of the scheduler. The overhead is too low to affect
586 common scheduling-intense workloads in practice (such as
587 webservers, memcache), but it does show up in artificial
588 scheduler stress tests, such as hackbench.
590 If you are paranoid and not sure what the kernel will be
595 endmenu # "CPU/Task time and stats accounting"
599 depends on SMP || COMPILE_TEST
602 Make sure that CPUs running critical tasks are not disturbed by
603 any source of "noise" such as unbound workqueues, timers, kthreads...
604 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
605 the "isolcpus=" boot parameter.
609 source "kernel/rcu/Kconfig"
616 tristate "Kernel .config support"
618 This option enables the complete Linux kernel ".config" file
619 contents to be saved in the kernel. It provides documentation
620 of which kernel options are used in a running kernel or in an
621 on-disk kernel. This information can be extracted from the kernel
622 image file with the script scripts/extract-ikconfig and used as
623 input to rebuild the current kernel or to build another kernel.
624 It can also be extracted from a running kernel by reading
625 /proc/config.gz if enabled (below).
628 bool "Enable access to .config through /proc/config.gz"
629 depends on IKCONFIG && PROC_FS
631 This option enables access to the kernel configuration file
632 through /proc/config.gz.
635 tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz"
638 This option enables access to the in-kernel headers that are generated during
639 the build process. These can be used to build eBPF tracing programs,
640 or similar programs. If you build the headers as a module, a module called
641 kheaders.ko is built which can be loaded on-demand to get access to headers.
644 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
649 Select the minimal kernel log buffer size as a power of 2.
650 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
651 parameter, see below. Any higher size also might be forced
652 by "log_buf_len" boot parameter.
662 config LOG_CPU_MAX_BUF_SHIFT
663 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
666 default 12 if !BASE_SMALL
667 default 0 if BASE_SMALL
670 This option allows to increase the default ring buffer size
671 according to the number of CPUs. The value defines the contribution
672 of each CPU as a power of 2. The used space is typically only few
673 lines however it might be much more when problems are reported,
676 The increased size means that a new buffer has to be allocated and
677 the original static one is unused. It makes sense only on systems
678 with more CPUs. Therefore this value is used only when the sum of
679 contributions is greater than the half of the default kernel ring
680 buffer as defined by LOG_BUF_SHIFT. The default values are set
681 so that more than 64 CPUs are needed to trigger the allocation.
683 Also this option is ignored when "log_buf_len" kernel parameter is
684 used as it forces an exact (power of two) size of the ring buffer.
686 The number of possible CPUs is used for this computation ignoring
687 hotplugging making the computation optimal for the worst case
688 scenario while allowing a simple algorithm to be used from bootup.
690 Examples shift values and their meaning:
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
698 config PRINTK_SAFE_LOG_BUF_SHIFT
699 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
704 Select the size of an alternate printk per-CPU buffer where messages
705 printed from usafe contexts are temporary stored. One example would
706 be NMI messages, another one - printk recursion. The messages are
707 copied to the main log buffer in a safe context to avoid a deadlock.
708 The value defines the size as a power of 2.
710 Those messages are rare and limited. The largest one is when
711 a backtrace is printed. It usually fits into 4KB. Select
712 8KB if you want to be on the safe side.
715 17 => 128 KB for each CPU
716 16 => 64 KB for each CPU
717 15 => 32 KB for each CPU
718 14 => 16 KB for each CPU
719 13 => 8 KB for each CPU
720 12 => 4 KB for each CPU
723 # Architectures with an unreliable sched_clock() should select this:
725 config HAVE_UNSTABLE_SCHED_CLOCK
728 config GENERIC_SCHED_CLOCK
731 menu "Scheduler features"
734 bool "Enable utilization clamping for RT/FAIR tasks"
735 depends on CPU_FREQ_GOV_SCHEDUTIL
737 This feature enables the scheduler to track the clamped utilization
738 of each CPU based on RUNNABLE tasks scheduled on that CPU.
740 With this option, the user can specify the min and max CPU
741 utilization allowed for RUNNABLE tasks. The max utilization defines
742 the maximum frequency a task should use while the min utilization
743 defines the minimum frequency it should use.
745 Both min and max utilization clamp values are hints to the scheduler,
746 aiming at improving its frequency selection policy, but they do not
747 enforce or grant any specific bandwidth for tasks.
751 config UCLAMP_BUCKETS_COUNT
752 int "Number of supported utilization clamp buckets"
755 depends on UCLAMP_TASK
757 Defines the number of clamp buckets to use. The range of each bucket
758 will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the
759 number of clamp buckets the finer their granularity and the higher
760 the precision of clamping aggregation and tracking at run-time.
762 For example, with the minimum configuration value we will have 5
763 clamp buckets tracking 20% utilization each. A 25% boosted tasks will
764 be refcounted in the [20..39]% bucket and will set the bucket clamp
765 effective value to 25%.
766 If a second 30% boosted task should be co-scheduled on the same CPU,
767 that task will be refcounted in the same bucket of the first task and
768 it will boost the bucket clamp effective value to 30%.
769 The clamp effective value of a bucket is reset to its nominal value
770 (20% in the example above) when there are no more tasks refcounted in
773 An additional boost/capping margin can be added to some tasks. In the
774 example above the 25% task will be boosted to 30% until it exits the
775 CPU. If that should be considered not acceptable on certain systems,
776 it's always possible to reduce the margin by increasing the number of
777 clamp buckets to trade off used memory for run-time tracking
780 If in doubt, use the default value.
785 # For architectures that want to enable the support for NUMA-affine scheduler
788 config ARCH_SUPPORTS_NUMA_BALANCING
792 # For architectures that prefer to flush all TLBs after a number of pages
793 # are unmapped instead of sending one IPI per page to flush. The architecture
794 # must provide guarantees on what happens if a clean TLB cache entry is
795 # written after the unmap. Details are in mm/rmap.c near the check for
796 # should_defer_flush. The architecture should also consider if the full flush
797 # and the refill costs are offset by the savings of sending fewer IPIs.
798 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
802 def_bool !$(cc-option,$(m64-flag) -D__SIZEOF_INT128__=0) && 64BIT
805 # For architectures that know their GCC __int128 support is sound
807 config ARCH_SUPPORTS_INT128
810 # For architectures that (ab)use NUMA to represent different memory regions
811 # all cpu-local but of different latencies, such as SuperH.
813 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
816 config NUMA_BALANCING
817 bool "Memory placement aware NUMA scheduler"
818 depends on ARCH_SUPPORTS_NUMA_BALANCING
819 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
820 depends on SMP && NUMA && MIGRATION
822 This option adds support for automatic NUMA aware memory/task placement.
823 The mechanism is quite primitive and is based on migrating memory when
824 it has references to the node the task is running on.
826 This system will be inactive on UMA systems.
828 config NUMA_BALANCING_DEFAULT_ENABLED
829 bool "Automatically enable NUMA aware memory/task placement"
831 depends on NUMA_BALANCING
833 If set, automatic NUMA balancing will be enabled if running on a NUMA
837 bool "Control Group support"
840 This option adds support for grouping sets of processes together, for
841 use with process control subsystems such as Cpusets, CFS, memory
842 controls or device isolation.
844 - Documentation/scheduler/sched-design-CFS.rst (CFS)
845 - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation
846 and resource control)
856 bool "Memory controller"
860 Provides control over the memory footprint of tasks in a cgroup.
864 depends on MEMCG && SWAP
869 depends on MEMCG && !SLOB
877 Generic block IO controller cgroup interface. This is the common
878 cgroup interface which should be used by various IO controlling
881 Currently, CFQ IO scheduler uses it to recognize task groups and
882 control disk bandwidth allocation (proportional time slice allocation)
883 to such task groups. It is also used by bio throttling logic in
884 block layer to implement upper limit in IO rates on a device.
886 This option only enables generic Block IO controller infrastructure.
887 One needs to also enable actual IO controlling logic/policy. For
888 enabling proportional weight division of disk bandwidth in CFQ, set
889 CONFIG_BFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
890 CONFIG_BLK_DEV_THROTTLING=y.
892 See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information.
894 config CGROUP_WRITEBACK
896 depends on MEMCG && BLK_CGROUP
899 menuconfig CGROUP_SCHED
900 bool "CPU controller"
903 This feature lets CPU scheduler recognize task groups and control CPU
904 bandwidth allocation to such task groups. It uses cgroups to group
908 config FAIR_GROUP_SCHED
909 bool "Group scheduling for SCHED_OTHER"
910 depends on CGROUP_SCHED
914 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
915 depends on FAIR_GROUP_SCHED
918 This option allows users to define CPU bandwidth rates (limits) for
919 tasks running within the fair group scheduler. Groups with no limit
920 set are considered to be unconstrained and will run with no
922 See Documentation/scheduler/sched-bwc.rst for more information.
924 config RT_GROUP_SCHED
925 bool "Group scheduling for SCHED_RR/FIFO"
926 depends on CGROUP_SCHED
929 This feature lets you explicitly allocate real CPU bandwidth
930 to task groups. If enabled, it will also make it impossible to
931 schedule realtime tasks for non-root users until you allocate
932 realtime bandwidth for them.
933 See Documentation/scheduler/sched-rt-group.rst for more information.
937 config UCLAMP_TASK_GROUP
938 bool "Utilization clamping per group of tasks"
939 depends on CGROUP_SCHED
940 depends on UCLAMP_TASK
943 This feature enables the scheduler to track the clamped utilization
944 of each CPU based on RUNNABLE tasks currently scheduled on that CPU.
946 When this option is enabled, the user can specify a min and max
947 CPU bandwidth which is allowed for each single task in a group.
948 The max bandwidth allows to clamp the maximum frequency a task
949 can use, while the min bandwidth allows to define a minimum
950 frequency a task will always use.
952 When task group based utilization clamping is enabled, an eventually
953 specified task-specific clamp value is constrained by the cgroup
954 specified clamp value. Both minimum and maximum task clamping cannot
955 be bigger than the corresponding clamping defined at task group level.
960 bool "PIDs controller"
962 Provides enforcement of process number limits in the scope of a
963 cgroup. Any attempt to fork more processes than is allowed in the
964 cgroup will fail. PIDs are fundamentally a global resource because it
965 is fairly trivial to reach PID exhaustion before you reach even a
966 conservative kmemcg limit. As a result, it is possible to grind a
967 system to halt without being limited by other cgroup policies. The
968 PIDs controller is designed to stop this from happening.
970 It should be noted that organisational operations (such as attaching
971 to a cgroup hierarchy) will *not* be blocked by the PIDs controller,
972 since the PIDs limit only affects a process's ability to fork, not to
976 bool "RDMA controller"
978 Provides enforcement of RDMA resources defined by IB stack.
979 It is fairly easy for consumers to exhaust RDMA resources, which
980 can result into resource unavailability to other consumers.
981 RDMA controller is designed to stop this from happening.
982 Attaching processes with active RDMA resources to the cgroup
983 hierarchy is allowed even if can cross the hierarchy's limit.
985 config CGROUP_FREEZER
986 bool "Freezer controller"
988 Provides a way to freeze and unfreeze all tasks in a
991 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
992 controller includes important in-kernel memory consumers per default.
994 If you're using cgroup2, say N.
996 config CGROUP_HUGETLB
997 bool "HugeTLB controller"
998 depends on HUGETLB_PAGE
1002 Provides a cgroup controller for HugeTLB pages.
1003 When you enable this, you can put a per cgroup limit on HugeTLB usage.
1004 The limit is enforced during page fault. Since HugeTLB doesn't
1005 support page reclaim, enforcing the limit at page fault time implies
1006 that, the application will get SIGBUS signal if it tries to access
1007 HugeTLB pages beyond its limit. This requires the application to know
1008 beforehand how much HugeTLB pages it would require for its use. The
1009 control group is tracked in the third page lru pointer. This means
1010 that we cannot use the controller with huge page less than 3 pages.
1013 bool "Cpuset controller"
1016 This option will let you create and manage CPUSETs which
1017 allow dynamically partitioning a system into sets of CPUs and
1018 Memory Nodes and assigning tasks to run only within those sets.
1019 This is primarily useful on large SMP or NUMA systems.
1023 config PROC_PID_CPUSET
1024 bool "Include legacy /proc/<pid>/cpuset file"
1028 config CGROUP_DEVICE
1029 bool "Device controller"
1031 Provides a cgroup controller implementing whitelists for
1032 devices which a process in the cgroup can mknod or open.
1034 config CGROUP_CPUACCT
1035 bool "Simple CPU accounting controller"
1037 Provides a simple controller for monitoring the
1038 total CPU consumed by the tasks in a cgroup.
1041 bool "Perf controller"
1042 depends on PERF_EVENTS
1044 This option extends the perf per-cpu mode to restrict monitoring
1045 to threads which belong to the cgroup specified and run on the
1046 designated cpu. Or this can be used to have cgroup ID in samples
1047 so that it can monitor performance events among cgroups.
1052 bool "Support for eBPF programs attached to cgroups"
1053 depends on BPF_SYSCALL
1054 select SOCK_CGROUP_DATA
1056 Allow attaching eBPF programs to a cgroup using the bpf(2)
1057 syscall command BPF_PROG_ATTACH.
1059 In which context these programs are accessed depends on the type
1060 of attachment. For instance, programs that are attached using
1061 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
1065 bool "Debug controller"
1067 depends on DEBUG_KERNEL
1069 This option enables a simple controller that exports
1070 debugging information about the cgroups framework. This
1071 controller is for control cgroup debugging only. Its
1072 interfaces are not stable.
1076 config SOCK_CGROUP_DATA
1082 menuconfig NAMESPACES
1083 bool "Namespaces support" if EXPERT
1084 depends on MULTIUSER
1087 Provides the way to make tasks work with different objects using
1088 the same id. For example same IPC id may refer to different objects
1089 or same user id or pid may refer to different tasks when used in
1090 different namespaces.
1095 bool "UTS namespace"
1098 In this namespace tasks see different info provided with the
1102 bool "TIME namespace"
1103 depends on GENERIC_VDSO_TIME_NS
1106 In this namespace boottime and monotonic clocks can be set.
1107 The time will keep going with the same pace.
1110 bool "IPC namespace"
1111 depends on (SYSVIPC || POSIX_MQUEUE)
1114 In this namespace tasks work with IPC ids which correspond to
1115 different IPC objects in different namespaces.
1118 bool "User namespace"
1121 This allows containers, i.e. vservers, to use user namespaces
1122 to provide different user info for different servers.
1124 When user namespaces are enabled in the kernel it is
1125 recommended that the MEMCG option also be enabled and that
1126 user-space use the memory control groups to limit the amount
1127 of memory a memory unprivileged users can use.
1132 bool "PID Namespaces"
1135 Support process id namespaces. This allows having multiple
1136 processes with the same pid as long as they are in different
1137 pid namespaces. This is a building block of containers.
1140 bool "Network namespace"
1144 Allow user space to create what appear to be multiple instances
1145 of the network stack.
1149 config CHECKPOINT_RESTORE
1150 bool "Checkpoint/restore support"
1151 select PROC_CHILDREN
1154 Enables additional kernel features in a sake of checkpoint/restore.
1155 In particular it adds auxiliary prctl codes to setup process text,
1156 data and heap segment sizes, and a few additional /proc filesystem
1159 If unsure, say N here.
1161 config SCHED_AUTOGROUP
1162 bool "Automatic process group scheduling"
1165 select FAIR_GROUP_SCHED
1167 This option optimizes the scheduler for common desktop workloads by
1168 automatically creating and populating task groups. This separation
1169 of workloads isolates aggressive CPU burners (like build jobs) from
1170 desktop applications. Task group autogeneration is currently based
1173 config SYSFS_DEPRECATED
1174 bool "Enable deprecated sysfs features to support old userspace tools"
1178 This option adds code that switches the layout of the "block" class
1179 devices, to not show up in /sys/class/block/, but only in
1182 This switch is only active when the sysfs.deprecated=1 boot option is
1183 passed or the SYSFS_DEPRECATED_V2 option is set.
1185 This option allows new kernels to run on old distributions and tools,
1186 which might get confused by /sys/class/block/. Since 2007/2008 all
1187 major distributions and tools handle this just fine.
1189 Recent distributions and userspace tools after 2009/2010 depend on
1190 the existence of /sys/class/block/, and will not work with this
1193 Only if you are using a new kernel on an old distribution, you might
1196 config SYSFS_DEPRECATED_V2
1197 bool "Enable deprecated sysfs features by default"
1200 depends on SYSFS_DEPRECATED
1202 Enable deprecated sysfs by default.
1204 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1207 Only if you are using a new kernel on an old distribution, you might
1208 need to say Y here. Even then, odds are you would not need it
1209 enabled, you can always pass the boot option if absolutely necessary.
1212 bool "Kernel->user space relay support (formerly relayfs)"
1215 This option enables support for relay interface support in
1216 certain file systems (such as debugfs).
1217 It is designed to provide an efficient mechanism for tools and
1218 facilities to relay large amounts of data from kernel space to
1223 config BLK_DEV_INITRD
1224 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1226 The initial RAM filesystem is a ramfs which is loaded by the
1227 boot loader (loadlin or lilo) and that is mounted as root
1228 before the normal boot procedure. It is typically used to
1229 load modules needed to mount the "real" root file system,
1230 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1232 If RAM disk support (BLK_DEV_RAM) is also included, this
1233 also enables initial RAM disk (initrd) support and adds
1234 15 Kbytes (more on some other architectures) to the kernel size.
1240 source "usr/Kconfig"
1245 bool "Boot config support"
1246 select BLK_DEV_INITRD
1248 Extra boot config allows system admin to pass a config file as
1249 complemental extension of kernel cmdline when booting.
1250 The boot config file must be attached at the end of initramfs
1251 with checksum, size and magic word.
1252 See <file:Documentation/admin-guide/bootconfig.rst> for details.
1257 prompt "Compiler optimization level"
1258 default CC_OPTIMIZE_FOR_PERFORMANCE
1260 config CC_OPTIMIZE_FOR_PERFORMANCE
1261 bool "Optimize for performance (-O2)"
1263 This is the default optimization level for the kernel, building
1264 with the "-O2" compiler flag for best performance and most
1265 helpful compile-time warnings.
1267 config CC_OPTIMIZE_FOR_PERFORMANCE_O3
1268 bool "Optimize more for performance (-O3)"
1271 Choosing this option will pass "-O3" to your compiler to optimize
1272 the kernel yet more for performance.
1274 config CC_OPTIMIZE_FOR_SIZE
1275 bool "Optimize for size (-Os)"
1277 Choosing this option will pass "-Os" to your compiler resulting
1278 in a smaller kernel.
1282 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1285 This requires that the arch annotates or otherwise protects
1286 its external entry points from being discarded. Linker scripts
1287 must also merge .text.*, .data.*, and .bss.* correctly into
1288 output sections. Care must be taken not to pull in unrelated
1289 sections (e.g., '.text.init'). Typically '.' in section names
1290 is used to distinguish them from label names / C identifiers.
1292 config LD_DEAD_CODE_DATA_ELIMINATION
1293 bool "Dead code and data elimination (EXPERIMENTAL)"
1294 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1296 depends on $(cc-option,-ffunction-sections -fdata-sections)
1297 depends on $(ld-option,--gc-sections)
1299 Enable this if you want to do dead code and data elimination with
1300 the linker by compiling with -ffunction-sections -fdata-sections,
1301 and linking with --gc-sections.
1303 This can reduce on disk and in-memory size of the kernel
1304 code and static data, particularly for small configs and
1305 on small systems. This has the possibility of introducing
1306 silently broken kernel if the required annotations are not
1307 present. This option is not well tested yet, so use at your
1316 config SYSCTL_EXCEPTION_TRACE
1319 Enable support for /proc/sys/debug/exception-trace.
1321 config SYSCTL_ARCH_UNALIGN_NO_WARN
1324 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1325 Allows arch to define/use @no_unaligned_warning to possibly warn
1326 about unaligned access emulation going on under the hood.
1328 config SYSCTL_ARCH_UNALIGN_ALLOW
1331 Enable support for /proc/sys/kernel/unaligned-trap
1332 Allows arches to define/use @unaligned_enabled to runtime toggle
1333 the unaligned access emulation.
1334 see arch/parisc/kernel/unaligned.c for reference
1336 config HAVE_PCSPKR_PLATFORM
1339 # interpreter that classic socket filters depend on
1344 bool "Configure standard kernel features (expert users)"
1345 # Unhide debug options, to make the on-by-default options visible
1348 This option allows certain base kernel options and settings
1349 to be disabled or tweaked. This is for specialized
1350 environments which can tolerate a "non-standard" kernel.
1351 Only use this if you really know what you are doing.
1354 bool "Enable 16-bit UID system calls" if EXPERT
1355 depends on HAVE_UID16 && MULTIUSER
1358 This enables the legacy 16-bit UID syscall wrappers.
1361 bool "Multiple users, groups and capabilities support" if EXPERT
1364 This option enables support for non-root users, groups and
1367 If you say N here, all processes will run with UID 0, GID 0, and all
1368 possible capabilities. Saying N here also compiles out support for
1369 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1372 If unsure, say Y here.
1374 config SGETMASK_SYSCALL
1375 bool "sgetmask/ssetmask syscalls support" if EXPERT
1376 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1378 sys_sgetmask and sys_ssetmask are obsolete system calls
1379 no longer supported in libc but still enabled by default in some
1382 If unsure, leave the default option here.
1384 config SYSFS_SYSCALL
1385 bool "Sysfs syscall support" if EXPERT
1388 sys_sysfs is an obsolete system call no longer supported in libc.
1389 Note that disabling this option is more secure but might break
1390 compatibility with some systems.
1392 If unsure say Y here.
1395 bool "open by fhandle syscalls" if EXPERT
1399 If you say Y here, a user level program will be able to map
1400 file names to handle and then later use the handle for
1401 different file system operations. This is useful in implementing
1402 userspace file servers, which now track files using handles instead
1403 of names. The handle would remain the same even if file names
1404 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1408 bool "Posix Clocks & timers" if EXPERT
1411 This includes native support for POSIX timers to the kernel.
1412 Some embedded systems have no use for them and therefore they
1413 can be configured out to reduce the size of the kernel image.
1415 When this option is disabled, the following syscalls won't be
1416 available: timer_create, timer_gettime: timer_getoverrun,
1417 timer_settime, timer_delete, clock_adjtime, getitimer,
1418 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1419 clock_getres and clock_nanosleep syscalls will be limited to
1420 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1426 bool "Enable support for printk" if EXPERT
1429 This option enables normal printk support. Removing it
1430 eliminates most of the message strings from the kernel image
1431 and makes the kernel more or less silent. As this makes it
1432 very difficult to diagnose system problems, saying N here is
1433 strongly discouraged.
1441 bool "BUG() support" if EXPERT
1444 Disabling this option eliminates support for BUG and WARN, reducing
1445 the size of your kernel image and potentially quietly ignoring
1446 numerous fatal conditions. You should only consider disabling this
1447 option for embedded systems with no facilities for reporting errors.
1453 bool "Enable ELF core dumps" if EXPERT
1455 Enable support for generating core dumps. Disabling saves about 4k.
1458 config PCSPKR_PLATFORM
1459 bool "Enable PC-Speaker support" if EXPERT
1460 depends on HAVE_PCSPKR_PLATFORM
1464 This option allows to disable the internal PC-Speaker
1465 support, saving some memory.
1469 bool "Enable full-sized data structures for core" if EXPERT
1471 Disabling this option reduces the size of miscellaneous core
1472 kernel data structures. This saves memory on small machines,
1473 but may reduce performance.
1476 bool "Enable futex support" if EXPERT
1480 Disabling this option will cause the kernel to be built without
1481 support for "fast userspace mutexes". The resulting kernel may not
1482 run glibc-based applications correctly.
1486 depends on FUTEX && RT_MUTEXES
1489 config HAVE_FUTEX_CMPXCHG
1493 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1494 is implemented and always working. This removes a couple of runtime
1498 bool "Enable eventpoll support" if EXPERT
1501 Disabling this option will cause the kernel to be built without
1502 support for epoll family of system calls.
1505 bool "Enable signalfd() system call" if EXPERT
1508 Enable the signalfd() system call that allows to receive signals
1509 on a file descriptor.
1514 bool "Enable timerfd() system call" if EXPERT
1517 Enable the timerfd() system call that allows to receive timer
1518 events on a file descriptor.
1523 bool "Enable eventfd() system call" if EXPERT
1526 Enable the eventfd() system call that allows to receive both
1527 kernel notification (ie. KAIO) or userspace notifications.
1532 bool "Use full shmem filesystem" if EXPERT
1536 The shmem is an internal filesystem used to manage shared memory.
1537 It is backed by swap and manages resource limits. It is also exported
1538 to userspace as tmpfs if TMPFS is enabled. Disabling this
1539 option replaces shmem and tmpfs with the much simpler ramfs code,
1540 which may be appropriate on small systems without swap.
1543 bool "Enable AIO support" if EXPERT
1546 This option enables POSIX asynchronous I/O which may by used
1547 by some high performance threaded applications. Disabling
1548 this option saves about 7k.
1551 bool "Enable IO uring support" if EXPERT
1555 This option enables support for the io_uring interface, enabling
1556 applications to submit and complete IO through submission and
1557 completion rings that are shared between the kernel and application.
1559 config ADVISE_SYSCALLS
1560 bool "Enable madvise/fadvise syscalls" if EXPERT
1563 This option enables the madvise and fadvise syscalls, used by
1564 applications to advise the kernel about their future memory or file
1565 usage, improving performance. If building an embedded system where no
1566 applications use these syscalls, you can disable this option to save
1569 config HAVE_ARCH_USERFAULTFD_WP
1572 Arch has userfaultfd write protection support
1575 bool "Enable membarrier() system call" if EXPERT
1578 Enable the membarrier() system call that allows issuing memory
1579 barriers across all running threads, which can be used to distribute
1580 the cost of user-space memory barriers asymmetrically by transforming
1581 pairs of memory barriers into pairs consisting of membarrier() and a
1587 bool "Load all symbols for debugging/ksymoops" if EXPERT
1590 Say Y here to let the kernel print out symbolic crash information and
1591 symbolic stack backtraces. This increases the size of the kernel
1592 somewhat, as all symbols have to be loaded into the kernel image.
1595 bool "Include all symbols in kallsyms"
1596 depends on DEBUG_KERNEL && KALLSYMS
1598 Normally kallsyms only contains the symbols of functions for nicer
1599 OOPS messages and backtraces (i.e., symbols from the text and inittext
1600 sections). This is sufficient for most cases. And only in very rare
1601 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1602 names of variables from the data sections, etc).
1604 This option makes sure that all symbols are loaded into the kernel
1605 image (i.e., symbols from all sections) in cost of increased kernel
1606 size (depending on the kernel configuration, it may be 300KiB or
1607 something like this).
1609 Say N unless you really need all symbols.
1611 config KALLSYMS_ABSOLUTE_PERCPU
1614 default X86_64 && SMP
1616 config KALLSYMS_BASE_RELATIVE
1621 Instead of emitting them as absolute values in the native word size,
1622 emit the symbol references in the kallsyms table as 32-bit entries,
1623 each containing a relative value in the range [base, base + U32_MAX]
1624 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1625 an absolute value in the range [0, S32_MAX] or a relative value in the
1626 range [base, base + S32_MAX], where base is the lowest relative symbol
1627 address encountered in the image.
1629 On 64-bit builds, this reduces the size of the address table by 50%,
1630 but more importantly, it results in entries whose values are build
1631 time constants, and no relocation pass is required at runtime to fix
1632 up the entries based on the runtime load address of the kernel.
1634 # end of the "standard kernel features (expert users)" menu
1636 # syscall, maps, verifier
1639 bool "LSM Instrumentation with BPF"
1640 depends on BPF_EVENTS
1641 depends on BPF_SYSCALL
1645 Enables instrumentation of the security hooks with eBPF programs for
1646 implementing dynamic MAC and Audit Policies.
1648 If you are unsure how to answer this question, answer N.
1651 bool "Enable bpf() system call"
1656 Enable the bpf() system call that allows to manipulate eBPF
1657 programs and maps via file descriptors.
1659 config ARCH_WANT_DEFAULT_BPF_JIT
1662 config BPF_JIT_ALWAYS_ON
1663 bool "Permanently enable BPF JIT and remove BPF interpreter"
1664 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1666 Enables BPF JIT and removes BPF interpreter to avoid
1667 speculative execution of BPF instructions by the interpreter
1669 config BPF_JIT_DEFAULT_ON
1670 def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
1671 depends on HAVE_EBPF_JIT && BPF_JIT
1674 bool "Enable userfaultfd() system call"
1677 Enable the userfaultfd() system call that allows to intercept and
1678 handle page faults in userland.
1680 config ARCH_HAS_MEMBARRIER_CALLBACKS
1683 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1687 bool "Enable rseq() system call" if EXPERT
1689 depends on HAVE_RSEQ
1692 Enable the restartable sequences system call. It provides a
1693 user-space cache for the current CPU number value, which
1694 speeds up getting the current CPU number from user-space,
1695 as well as an ABI to speed up user-space operations on
1702 bool "Enabled debugging of rseq() system call" if EXPERT
1703 depends on RSEQ && DEBUG_KERNEL
1705 Enable extra debugging checks for the rseq system call.
1710 bool "Embedded system"
1711 option allnoconfig_y
1714 This option should be enabled if compiling the kernel for
1715 an embedded system so certain expert options are available
1718 config HAVE_PERF_EVENTS
1721 See tools/perf/design.txt for details.
1723 config PERF_USE_VMALLOC
1726 See tools/perf/design.txt for details
1729 bool "PC/104 support" if EXPERT
1731 Expose PC/104 form factor device drivers and options available for
1732 selection and configuration. Enable this option if your target
1733 machine has a PC/104 bus.
1735 menu "Kernel Performance Events And Counters"
1738 bool "Kernel performance events and counters"
1739 default y if PROFILING
1740 depends on HAVE_PERF_EVENTS
1744 Enable kernel support for various performance events provided
1745 by software and hardware.
1747 Software events are supported either built-in or via the
1748 use of generic tracepoints.
1750 Most modern CPUs support performance events via performance
1751 counter registers. These registers count the number of certain
1752 types of hw events: such as instructions executed, cachemisses
1753 suffered, or branches mis-predicted - without slowing down the
1754 kernel or applications. These registers can also trigger interrupts
1755 when a threshold number of events have passed - and can thus be
1756 used to profile the code that runs on that CPU.
1758 The Linux Performance Event subsystem provides an abstraction of
1759 these software and hardware event capabilities, available via a
1760 system call and used by the "perf" utility in tools/perf/. It
1761 provides per task and per CPU counters, and it provides event
1762 capabilities on top of those.
1766 config DEBUG_PERF_USE_VMALLOC
1768 bool "Debug: use vmalloc to back perf mmap() buffers"
1769 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1770 select PERF_USE_VMALLOC
1772 Use vmalloc memory to back perf mmap() buffers.
1774 Mostly useful for debugging the vmalloc code on platforms
1775 that don't require it.
1781 config VM_EVENT_COUNTERS
1783 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1785 VM event counters are needed for event counts to be shown.
1786 This option allows the disabling of the VM event counters
1787 on EXPERT systems. /proc/vmstat will only show page counts
1788 if VM event counters are disabled.
1792 bool "Enable SLUB debugging support" if EXPERT
1793 depends on SLUB && SYSFS
1795 SLUB has extensive debug support features. Disabling these can
1796 result in significant savings in code size. This also disables
1797 SLUB sysfs support. /sys/slab will not exist and there will be
1798 no support for cache validation etc.
1800 config SLUB_MEMCG_SYSFS_ON
1802 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1803 depends on SLUB && SYSFS && MEMCG
1805 SLUB creates a directory under /sys/kernel/slab for each
1806 allocation cache to host info and debug files. If memory
1807 cgroup is enabled, each cache can have per memory cgroup
1808 caches. SLUB can create the same sysfs directories for these
1809 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1810 to a very high number of debug files being created. This is
1811 controlled by slub_memcg_sysfs boot parameter and this
1812 config option determines the parameter's default value.
1815 bool "Disable heap randomization"
1818 Randomizing heap placement makes heap exploits harder, but it
1819 also breaks ancient binaries (including anything libc5 based).
1820 This option changes the bootup default to heap randomization
1821 disabled, and can be overridden at runtime by setting
1822 /proc/sys/kernel/randomize_va_space to 2.
1824 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1827 prompt "Choose SLAB allocator"
1830 This option allows to select a slab allocator.
1834 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1836 The regular slab allocator that is established and known to work
1837 well in all environments. It organizes cache hot objects in
1838 per cpu and per node queues.
1841 bool "SLUB (Unqueued Allocator)"
1842 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1844 SLUB is a slab allocator that minimizes cache line usage
1845 instead of managing queues of cached objects (SLAB approach).
1846 Per cpu caching is realized using slabs of objects instead
1847 of queues of objects. SLUB can use memory efficiently
1848 and has enhanced diagnostics. SLUB is the default choice for
1853 bool "SLOB (Simple Allocator)"
1855 SLOB replaces the stock allocator with a drastically simpler
1856 allocator. SLOB is generally more space efficient but
1857 does not perform as well on large systems.
1861 config SLAB_MERGE_DEFAULT
1862 bool "Allow slab caches to be merged"
1865 For reduced kernel memory fragmentation, slab caches can be
1866 merged when they share the same size and other characteristics.
1867 This carries a risk of kernel heap overflows being able to
1868 overwrite objects from merged caches (and more easily control
1869 cache layout), which makes such heap attacks easier to exploit
1870 by attackers. By keeping caches unmerged, these kinds of exploits
1871 can usually only damage objects in the same cache. To disable
1872 merging at runtime, "slab_nomerge" can be passed on the kernel
1875 config SLAB_FREELIST_RANDOM
1877 depends on SLAB || SLUB
1878 bool "SLAB freelist randomization"
1880 Randomizes the freelist order used on creating new pages. This
1881 security feature reduces the predictability of the kernel slab
1882 allocator against heap overflows.
1884 config SLAB_FREELIST_HARDENED
1885 bool "Harden slab freelist metadata"
1888 Many kernel heap attacks try to target slab cache metadata and
1889 other infrastructure. This options makes minor performance
1890 sacrifices to harden the kernel slab allocator against common
1891 freelist exploit methods.
1893 config SHUFFLE_PAGE_ALLOCATOR
1894 bool "Page allocator randomization"
1895 default SLAB_FREELIST_RANDOM && ACPI_NUMA
1897 Randomization of the page allocator improves the average
1898 utilization of a direct-mapped memory-side-cache. See section
1899 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
1900 6.2a specification for an example of how a platform advertises
1901 the presence of a memory-side-cache. There are also incidental
1902 security benefits as it reduces the predictability of page
1903 allocations to compliment SLAB_FREELIST_RANDOM, but the
1904 default granularity of shuffling on the "MAX_ORDER - 1" i.e,
1905 10th order of pages is selected based on cache utilization
1908 While the randomization improves cache utilization it may
1909 negatively impact workloads on platforms without a cache. For
1910 this reason, by default, the randomization is enabled only
1911 after runtime detection of a direct-mapped memory-side-cache.
1912 Otherwise, the randomization may be force enabled with the
1913 'page_alloc.shuffle' kernel command line parameter.
1917 config SLUB_CPU_PARTIAL
1919 depends on SLUB && SMP
1920 bool "SLUB per cpu partial cache"
1922 Per cpu partial caches accelerate objects allocation and freeing
1923 that is local to a processor at the price of more indeterminism
1924 in the latency of the free. On overflow these caches will be cleared
1925 which requires the taking of locks that may cause latency spikes.
1926 Typically one would choose no for a realtime system.
1928 config MMAP_ALLOW_UNINITIALIZED
1929 bool "Allow mmapped anonymous memory to be uninitialized"
1930 depends on EXPERT && !MMU
1933 Normally, and according to the Linux spec, anonymous memory obtained
1934 from mmap() has its contents cleared before it is passed to
1935 userspace. Enabling this config option allows you to request that
1936 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1937 providing a huge performance boost. If this option is not enabled,
1938 then the flag will be ignored.
1940 This is taken advantage of by uClibc's malloc(), and also by
1941 ELF-FDPIC binfmt's brk and stack allocator.
1943 Because of the obvious security issues, this option should only be
1944 enabled on embedded devices where you control what is run in
1945 userspace. Since that isn't generally a problem on no-MMU systems,
1946 it is normally safe to say Y here.
1948 See Documentation/nommu-mmap.txt for more information.
1950 config SYSTEM_DATA_VERIFICATION
1952 select SYSTEM_TRUSTED_KEYRING
1956 select ASYMMETRIC_KEY_TYPE
1957 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1960 select X509_CERTIFICATE_PARSER
1961 select PKCS7_MESSAGE_PARSER
1963 Provide PKCS#7 message verification using the contents of the system
1964 trusted keyring to provide public keys. This then can be used for
1965 module verification, kexec image verification and firmware blob
1969 bool "Profiling support"
1971 Say Y here to enable the extended profiling support mechanisms used
1972 by profilers such as OProfile.
1975 # Place an empty function call at each tracepoint site. Can be
1976 # dynamically changed for a probe function.
1981 endmenu # General setup
1983 source "arch/Kconfig"
1990 default 0 if BASE_FULL
1991 default 1 if !BASE_FULL
1993 config MODULE_SIG_FORMAT
1995 select SYSTEM_DATA_VERIFICATION
1998 bool "Enable loadable module support"
2001 Kernel modules are small pieces of compiled code which can
2002 be inserted in the running kernel, rather than being
2003 permanently built into the kernel. You use the "modprobe"
2004 tool to add (and sometimes remove) them. If you say Y here,
2005 many parts of the kernel can be built as modules (by
2006 answering M instead of Y where indicated): this is most
2007 useful for infrequently used options which are not required
2008 for booting. For more information, see the man pages for
2009 modprobe, lsmod, modinfo, insmod and rmmod.
2011 If you say Y here, you will need to run "make
2012 modules_install" to put the modules under /lib/modules/
2013 where modprobe can find them (you may need to be root to do
2020 config MODULE_FORCE_LOAD
2021 bool "Forced module loading"
2024 Allow loading of modules without version information (ie. modprobe
2025 --force). Forced module loading sets the 'F' (forced) taint flag and
2026 is usually a really bad idea.
2028 config MODULE_UNLOAD
2029 bool "Module unloading"
2031 Without this option you will not be able to unload any
2032 modules (note that some modules may not be unloadable
2033 anyway), which makes your kernel smaller, faster
2034 and simpler. If unsure, say Y.
2036 config MODULE_FORCE_UNLOAD
2037 bool "Forced module unloading"
2038 depends on MODULE_UNLOAD
2040 This option allows you to force a module to unload, even if the
2041 kernel believes it is unsafe: the kernel will remove the module
2042 without waiting for anyone to stop using it (using the -f option to
2043 rmmod). This is mainly for kernel developers and desperate users.
2047 bool "Module versioning support"
2049 Usually, you have to use modules compiled with your kernel.
2050 Saying Y here makes it sometimes possible to use modules
2051 compiled for different kernels, by adding enough information
2052 to the modules to (hopefully) spot any changes which would
2053 make them incompatible with the kernel you are running. If
2056 config ASM_MODVERSIONS
2058 default HAVE_ASM_MODVERSIONS && MODVERSIONS
2060 This enables module versioning for exported symbols also from
2061 assembly. This can be enabled only when the target architecture
2064 config MODULE_REL_CRCS
2066 depends on MODVERSIONS
2068 config MODULE_SRCVERSION_ALL
2069 bool "Source checksum for all modules"
2071 Modules which contain a MODULE_VERSION get an extra "srcversion"
2072 field inserted into their modinfo section, which contains a
2073 sum of the source files which made it. This helps maintainers
2074 see exactly which source was used to build a module (since
2075 others sometimes change the module source without updating
2076 the version). With this option, such a "srcversion" field
2077 will be created for all modules. If unsure, say N.
2080 bool "Module signature verification"
2081 select MODULE_SIG_FORMAT
2083 Check modules for valid signatures upon load: the signature
2084 is simply appended to the module. For more information see
2085 <file:Documentation/admin-guide/module-signing.rst>.
2087 Note that this option adds the OpenSSL development packages as a
2088 kernel build dependency so that the signing tool can use its crypto
2091 You should enable this option if you wish to use either
2092 CONFIG_SECURITY_LOCKDOWN_LSM or lockdown functionality imposed via
2093 another LSM - otherwise unsigned modules will be loadable regardless
2094 of the lockdown policy.
2096 !!!WARNING!!! If you enable this option, you MUST make sure that the
2097 module DOES NOT get stripped after being signed. This includes the
2098 debuginfo strip done by some packagers (such as rpmbuild) and
2099 inclusion into an initramfs that wants the module size reduced.
2101 config MODULE_SIG_FORCE
2102 bool "Require modules to be validly signed"
2103 depends on MODULE_SIG
2105 Reject unsigned modules or signed modules for which we don't have a
2106 key. Without this, such modules will simply taint the kernel.
2108 config MODULE_SIG_ALL
2109 bool "Automatically sign all modules"
2111 depends on MODULE_SIG
2113 Sign all modules during make modules_install. Without this option,
2114 modules must be signed manually, using the scripts/sign-file tool.
2116 comment "Do not forget to sign required modules with scripts/sign-file"
2117 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
2120 prompt "Which hash algorithm should modules be signed with?"
2121 depends on MODULE_SIG
2123 This determines which sort of hashing algorithm will be used during
2124 signature generation. This algorithm _must_ be built into the kernel
2125 directly so that signature verification can take place. It is not
2126 possible to load a signed module containing the algorithm to check
2127 the signature on that module.
2129 config MODULE_SIG_SHA1
2130 bool "Sign modules with SHA-1"
2133 config MODULE_SIG_SHA224
2134 bool "Sign modules with SHA-224"
2135 select CRYPTO_SHA256
2137 config MODULE_SIG_SHA256
2138 bool "Sign modules with SHA-256"
2139 select CRYPTO_SHA256
2141 config MODULE_SIG_SHA384
2142 bool "Sign modules with SHA-384"
2143 select CRYPTO_SHA512
2145 config MODULE_SIG_SHA512
2146 bool "Sign modules with SHA-512"
2147 select CRYPTO_SHA512
2151 config MODULE_SIG_HASH
2153 depends on MODULE_SIG
2154 default "sha1" if MODULE_SIG_SHA1
2155 default "sha224" if MODULE_SIG_SHA224
2156 default "sha256" if MODULE_SIG_SHA256
2157 default "sha384" if MODULE_SIG_SHA384
2158 default "sha512" if MODULE_SIG_SHA512
2160 config MODULE_COMPRESS
2161 bool "Compress modules on installation"
2164 Compresses kernel modules when 'make modules_install' is run; gzip or
2165 xz depending on "Compression algorithm" below.
2167 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
2169 Out-of-tree kernel modules installed using Kbuild will also be
2170 compressed upon installation.
2172 Note: for modules inside an initrd or initramfs, it's more efficient
2173 to compress the whole initrd or initramfs instead.
2175 Note: This is fully compatible with signed modules.
2180 prompt "Compression algorithm"
2181 depends on MODULE_COMPRESS
2182 default MODULE_COMPRESS_GZIP
2184 This determines which sort of compression will be used during
2185 'make modules_install'.
2187 GZIP (default) and XZ are supported.
2189 config MODULE_COMPRESS_GZIP
2192 config MODULE_COMPRESS_XZ
2197 config MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS
2198 bool "Allow loading of modules with missing namespace imports"
2200 Symbols exported with EXPORT_SYMBOL_NS*() are considered exported in
2201 a namespace. A module that makes use of a symbol exported with such a
2202 namespace is required to import the namespace via MODULE_IMPORT_NS().
2203 There is no technical reason to enforce correct namespace imports,
2204 but it creates consistency between symbols defining namespaces and
2205 users importing namespaces they make use of. This option relaxes this
2206 requirement and lifts the enforcement when loading a module.
2210 config UNUSED_SYMBOLS
2211 bool "Enable unused/obsolete exported symbols"
2214 Unused but exported symbols make the kernel needlessly bigger. For
2215 that reason most of these unused exports will soon be removed. This
2216 option is provided temporarily to provide a transition period in case
2217 some external kernel module needs one of these symbols anyway. If you
2218 encounter such a case in your module, consider if you are actually
2219 using the right API. (rationale: since nobody in the kernel is using
2220 this in a module, there is a pretty good chance it's actually the
2221 wrong interface to use). If you really need the symbol, please send a
2222 mail to the linux kernel mailing list mentioning the symbol and why
2223 you really need it, and what the merge plan to the mainline kernel for
2226 config TRIM_UNUSED_KSYMS
2227 bool "Trim unused exported kernel symbols"
2228 depends on !UNUSED_SYMBOLS
2230 The kernel and some modules make many symbols available for
2231 other modules to use via EXPORT_SYMBOL() and variants. Depending
2232 on the set of modules being selected in your kernel configuration,
2233 many of those exported symbols might never be used.
2235 This option allows for unused exported symbols to be dropped from
2236 the build. In turn, this provides the compiler more opportunities
2237 (especially when using LTO) for optimizing the code and reducing
2238 binary size. This might have some security advantages as well.
2240 If unsure, or if you need to build out-of-tree modules, say N.
2242 config UNUSED_KSYMS_WHITELIST
2243 string "Whitelist of symbols to keep in ksymtab"
2244 depends on TRIM_UNUSED_KSYMS
2246 By default, all unused exported symbols will be un-exported from the
2247 build when TRIM_UNUSED_KSYMS is selected.
2249 UNUSED_KSYMS_WHITELIST allows to whitelist symbols that must be kept
2250 exported at all times, even in absence of in-tree users. The value to
2251 set here is the path to a text file containing the list of symbols,
2252 one per line. The path can be absolute, or relative to the kernel
2257 config MODULES_TREE_LOOKUP
2259 depends on PERF_EVENTS || TRACING
2261 config INIT_ALL_POSSIBLE
2264 Back when each arch used to define their own cpu_online_mask and
2265 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2266 with all 1s, and others with all 0s. When they were centralised,
2267 it was better to provide this option than to break all the archs
2268 and have several arch maintainers pursuing me down dark alleys.
2270 source "block/Kconfig"
2272 config PREEMPT_NOTIFIERS
2282 Build a simple ASN.1 grammar compiler that produces a bytecode output
2283 that can be interpreted by the ASN.1 stream decoder and used to
2284 inform it as to what tags are to be expected in a stream and what
2285 functions to call on what tags.
2287 source "kernel/Kconfig.locks"
2289 config ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
2292 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2295 # It may be useful for an architecture to override the definitions of the
2296 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2297 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2298 # different calling convention for syscalls. They can also override the
2299 # macros for not-implemented syscalls in kernel/sys_ni.c and
2300 # kernel/time/posix-stubs.c. All these overrides need to be available in
2301 # <asm/syscall_wrapper.h>.
2302 config ARCH_HAS_SYSCALL_WRAPPER