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)
45 default $(shell,$(srctree)/scripts/clang-version.sh $(CC))
49 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(m64-flag)) if 64BIT
50 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(m32-flag))
52 config CC_CAN_LINK_STATIC
54 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) -static $(m64-flag)) if 64BIT
55 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) -static $(m32-flag))
57 config CC_HAS_ASM_GOTO
58 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
60 config TOOLS_SUPPORT_RELR
61 def_bool $(success,env "CC=$(CC)" "LD=$(LD)" "NM=$(NM)" "OBJCOPY=$(OBJCOPY)" $(srctree)/scripts/tools-support-relr.sh)
63 config CC_HAS_ASM_INLINE
64 def_bool $(success,echo 'void foo(void) { asm inline (""); }' | $(CC) -x c - -c -o /dev/null)
73 config BUILDTIME_TABLE_SORT
76 config THREAD_INFO_IN_TASK
79 Select this to move thread_info off the stack into task_struct. To
80 make this work, an arch will need to remove all thread_info fields
81 except flags and fix any runtime bugs.
83 One subtle change that will be needed is to use try_get_task_stack()
84 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
93 depends on BROKEN || !SMP
96 config INIT_ENV_ARG_LIMIT
101 Maximum of each of the number of arguments and environment
102 variables passed to init from the kernel command line.
105 bool "Compile also drivers which will not load"
109 Some drivers can be compiled on a different platform than they are
110 intended to be run on. Despite they cannot be loaded there (or even
111 when they load they cannot be used due to missing HW support),
112 developers still, opposing to distributors, might want to build such
113 drivers to compile-test them.
115 If you are a developer and want to build everything available, say Y
116 here. If you are a user/distributor, say N here to exclude useless
117 drivers to be distributed.
119 config UAPI_HEADER_TEST
120 bool "Compile test UAPI headers"
121 depends on HEADERS_INSTALL && CC_CAN_LINK
123 Compile test headers exported to user-space to ensure they are
124 self-contained, i.e. compilable as standalone units.
126 If you are a developer or tester and want to ensure the exported
127 headers are self-contained, say Y here. Otherwise, choose N.
130 string "Local version - append to kernel release"
132 Append an extra string to the end of your kernel version.
133 This will show up when you type uname, for example.
134 The string you set here will be appended after the contents of
135 any files with a filename matching localversion* in your
136 object and source tree, in that order. Your total string can
137 be a maximum of 64 characters.
139 config LOCALVERSION_AUTO
140 bool "Automatically append version information to the version string"
142 depends on !COMPILE_TEST
144 This will try to automatically determine if the current tree is a
145 release tree by looking for git tags that belong to the current
146 top of tree revision.
148 A string of the format -gxxxxxxxx will be added to the localversion
149 if a git-based tree is found. The string generated by this will be
150 appended after any matching localversion* files, and after the value
151 set in CONFIG_LOCALVERSION.
153 (The actual string used here is the first eight characters produced
154 by running the command:
156 $ git rev-parse --verify HEAD
158 which is done within the script "scripts/setlocalversion".)
161 string "Build ID Salt"
164 The build ID is used to link binaries and their debug info. Setting
165 this option will use the value in the calculation of the build id.
166 This is mostly useful for distributions which want to ensure the
167 build is unique between builds. It's safe to leave the default.
169 config HAVE_KERNEL_GZIP
172 config HAVE_KERNEL_BZIP2
175 config HAVE_KERNEL_LZMA
178 config HAVE_KERNEL_XZ
181 config HAVE_KERNEL_LZO
184 config HAVE_KERNEL_LZ4
187 config HAVE_KERNEL_UNCOMPRESSED
191 prompt "Kernel compression mode"
193 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
195 The linux kernel is a kind of self-extracting executable.
196 Several compression algorithms are available, which differ
197 in efficiency, compression and decompression speed.
198 Compression speed is only relevant when building a kernel.
199 Decompression speed is relevant at each boot.
201 If you have any problems with bzip2 or lzma compressed
203 version of this functionality (bzip2 only), for 2.4, was
204 supplied by Christian Ludwig)
206 High compression options are mostly useful for users, who
207 are low on disk space (embedded systems), but for whom ram
210 If in doubt, select 'gzip'
214 depends on HAVE_KERNEL_GZIP
216 The old and tried gzip compression. It provides a good balance
217 between compression ratio and decompression speed.
221 depends on HAVE_KERNEL_BZIP2
223 Its compression ratio and speed is intermediate.
224 Decompression speed is slowest among the choices. The kernel
225 size is about 10% smaller with bzip2, in comparison to gzip.
226 Bzip2 uses a large amount of memory. For modern kernels you
227 will need at least 8MB RAM or more for booting.
231 depends on HAVE_KERNEL_LZMA
233 This compression algorithm's ratio is best. Decompression speed
234 is between gzip and bzip2. Compression is slowest.
235 The kernel size is about 33% smaller with LZMA in comparison to gzip.
239 depends on HAVE_KERNEL_XZ
241 XZ uses the LZMA2 algorithm and instruction set specific
242 BCJ filters which can improve compression ratio of executable
243 code. The size of the kernel is about 30% smaller with XZ in
244 comparison to gzip. On architectures for which there is a BCJ
245 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
246 will create a few percent smaller kernel than plain LZMA.
248 The speed is about the same as with LZMA: The decompression
249 speed of XZ is better than that of bzip2 but worse than gzip
250 and LZO. Compression is slow.
254 depends on HAVE_KERNEL_LZO
256 Its compression ratio is the poorest among the choices. The kernel
257 size is about 10% bigger than gzip; however its speed
258 (both compression and decompression) is the fastest.
262 depends on HAVE_KERNEL_LZ4
264 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
265 A preliminary version of LZ4 de/compression tool is available at
266 <https://code.google.com/p/lz4/>.
268 Its compression ratio is worse than LZO. The size of the kernel
269 is about 8% bigger than LZO. But the decompression speed is
272 config KERNEL_UNCOMPRESSED
274 depends on HAVE_KERNEL_UNCOMPRESSED
276 Produce uncompressed kernel image. This option is usually not what
277 you want. It is useful for debugging the kernel in slow simulation
278 environments, where decompressing and moving the kernel is awfully
279 slow. This option allows early boot code to skip the decompressor
280 and jump right at uncompressed kernel image.
284 config DEFAULT_HOSTNAME
285 string "Default hostname"
288 This option determines the default system hostname before userspace
289 calls sethostname(2). The kernel traditionally uses "(none)" here,
290 but you may wish to use a different default here to make a minimal
291 system more usable with less configuration.
294 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
295 # add proper SWAP support to them, in which case this can be remove.
301 bool "Support for paging of anonymous memory (swap)"
302 depends on MMU && BLOCK && !ARCH_NO_SWAP
305 This option allows you to choose whether you want to have support
306 for so called swap devices or swap files in your kernel that are
307 used to provide more virtual memory than the actual RAM present
308 in your computer. If unsure say Y.
313 Inter Process Communication is a suite of library functions and
314 system calls which let processes (running programs) synchronize and
315 exchange information. It is generally considered to be a good thing,
316 and some programs won't run unless you say Y here. In particular, if
317 you want to run the DOS emulator dosemu under Linux (read the
318 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
319 you'll need to say Y here.
321 You can find documentation about IPC with "info ipc" and also in
322 section 6.4 of the Linux Programmer's Guide, available from
323 <http://www.tldp.org/guides.html>.
325 config SYSVIPC_SYSCTL
332 bool "POSIX Message Queues"
335 POSIX variant of message queues is a part of IPC. In POSIX message
336 queues every message has a priority which decides about succession
337 of receiving it by a process. If you want to compile and run
338 programs written e.g. for Solaris with use of its POSIX message
339 queues (functions mq_*) say Y here.
341 POSIX message queues are visible as a filesystem called 'mqueue'
342 and can be mounted somewhere if you want to do filesystem
343 operations on message queues.
347 config POSIX_MQUEUE_SYSCTL
349 depends on POSIX_MQUEUE
353 config CROSS_MEMORY_ATTACH
354 bool "Enable process_vm_readv/writev syscalls"
358 Enabling this option adds the system calls process_vm_readv and
359 process_vm_writev which allow a process with the correct privileges
360 to directly read from or write to another process' address space.
361 See the man page for more details.
364 bool "uselib syscall"
365 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
367 This option enables the uselib syscall, a system call used in the
368 dynamic linker from libc5 and earlier. glibc does not use this
369 system call. If you intend to run programs built on libc5 or
370 earlier, you may need to enable this syscall. Current systems
371 running glibc can safely disable this.
374 bool "Auditing support"
377 Enable auditing infrastructure that can be used with another
378 kernel subsystem, such as SELinux (which requires this for
379 logging of avc messages output). System call auditing is included
380 on architectures which support it.
382 config HAVE_ARCH_AUDITSYSCALL
387 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
390 source "kernel/irq/Kconfig"
391 source "kernel/time/Kconfig"
392 source "kernel/Kconfig.preempt"
394 menu "CPU/Task time and stats accounting"
396 config VIRT_CPU_ACCOUNTING
400 prompt "Cputime accounting"
401 default TICK_CPU_ACCOUNTING if !PPC64
402 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
404 # Kind of a stub config for the pure tick based cputime accounting
405 config TICK_CPU_ACCOUNTING
406 bool "Simple tick based cputime accounting"
407 depends on !S390 && !NO_HZ_FULL
409 This is the basic tick based cputime accounting that maintains
410 statistics about user, system and idle time spent on per jiffies
415 config VIRT_CPU_ACCOUNTING_NATIVE
416 bool "Deterministic task and CPU time accounting"
417 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
418 select VIRT_CPU_ACCOUNTING
420 Select this option to enable more accurate task and CPU time
421 accounting. This is done by reading a CPU counter on each
422 kernel entry and exit and on transitions within the kernel
423 between system, softirq and hardirq state, so there is a
424 small performance impact. In the case of s390 or IBM POWER > 5,
425 this also enables accounting of stolen time on logically-partitioned
428 config VIRT_CPU_ACCOUNTING_GEN
429 bool "Full dynticks CPU time accounting"
430 depends on HAVE_CONTEXT_TRACKING
431 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
432 depends on GENERIC_CLOCKEVENTS
433 select VIRT_CPU_ACCOUNTING
434 select CONTEXT_TRACKING
436 Select this option to enable task and CPU time accounting on full
437 dynticks systems. This accounting is implemented by watching every
438 kernel-user boundaries using the context tracking subsystem.
439 The accounting is thus performed at the expense of some significant
442 For now this is only useful if you are working on the full
443 dynticks subsystem development.
449 config IRQ_TIME_ACCOUNTING
450 bool "Fine granularity task level IRQ time accounting"
451 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
453 Select this option to enable fine granularity task irq time
454 accounting. This is done by reading a timestamp on each
455 transitions between softirq and hardirq state, so there can be a
456 small performance impact.
458 If in doubt, say N here.
460 config HAVE_SCHED_AVG_IRQ
462 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
465 config SCHED_THERMAL_PRESSURE
466 bool "Enable periodic averaging of thermal pressure"
469 config BSD_PROCESS_ACCT
470 bool "BSD Process Accounting"
473 If you say Y here, a user level program will be able to instruct the
474 kernel (via a special system call) to write process accounting
475 information to a file: whenever a process exits, information about
476 that process will be appended to the file by the kernel. The
477 information includes things such as creation time, owning user,
478 command name, memory usage, controlling terminal etc. (the complete
479 list is in the struct acct in <file:include/linux/acct.h>). It is
480 up to the user level program to do useful things with this
481 information. This is generally a good idea, so say Y.
483 config BSD_PROCESS_ACCT_V3
484 bool "BSD Process Accounting version 3 file format"
485 depends on BSD_PROCESS_ACCT
488 If you say Y here, the process accounting information is written
489 in a new file format that also logs the process IDs of each
490 process and its parent. Note that this file format is incompatible
491 with previous v0/v1/v2 file formats, so you will need updated tools
492 for processing it. A preliminary version of these tools is available
493 at <http://www.gnu.org/software/acct/>.
496 bool "Export task/process statistics through netlink"
501 Export selected statistics for tasks/processes through the
502 generic netlink interface. Unlike BSD process accounting, the
503 statistics are available during the lifetime of tasks/processes as
504 responses to commands. Like BSD accounting, they are sent to user
509 config TASK_DELAY_ACCT
510 bool "Enable per-task delay accounting"
514 Collect information on time spent by a task waiting for system
515 resources like cpu, synchronous block I/O completion and swapping
516 in pages. Such statistics can help in setting a task's priorities
517 relative to other tasks for cpu, io, rss limits etc.
522 bool "Enable extended accounting over taskstats"
525 Collect extended task accounting data and send the data
526 to userland for processing over the taskstats interface.
530 config TASK_IO_ACCOUNTING
531 bool "Enable per-task storage I/O accounting"
532 depends on TASK_XACCT
534 Collect information on the number of bytes of storage I/O which this
540 bool "Pressure stall information tracking"
542 Collect metrics that indicate how overcommitted the CPU, memory,
543 and IO capacity are in the system.
545 If you say Y here, the kernel will create /proc/pressure/ with the
546 pressure statistics files cpu, memory, and io. These will indicate
547 the share of walltime in which some or all tasks in the system are
548 delayed due to contention of the respective resource.
550 In kernels with cgroup support, cgroups (cgroup2 only) will
551 have cpu.pressure, memory.pressure, and io.pressure files,
552 which aggregate pressure stalls for the grouped tasks only.
554 For more details see Documentation/accounting/psi.rst.
558 config PSI_DEFAULT_DISABLED
559 bool "Require boot parameter to enable pressure stall information tracking"
563 If set, pressure stall information tracking will be disabled
564 per default but can be enabled through passing psi=1 on the
565 kernel commandline during boot.
567 This feature adds some code to the task wakeup and sleep
568 paths of the scheduler. The overhead is too low to affect
569 common scheduling-intense workloads in practice (such as
570 webservers, memcache), but it does show up in artificial
571 scheduler stress tests, such as hackbench.
573 If you are paranoid and not sure what the kernel will be
578 endmenu # "CPU/Task time and stats accounting"
582 depends on SMP || COMPILE_TEST
585 Make sure that CPUs running critical tasks are not disturbed by
586 any source of "noise" such as unbound workqueues, timers, kthreads...
587 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
588 the "isolcpus=" boot parameter.
592 source "kernel/rcu/Kconfig"
599 tristate "Kernel .config support"
601 This option enables the complete Linux kernel ".config" file
602 contents to be saved in the kernel. It provides documentation
603 of which kernel options are used in a running kernel or in an
604 on-disk kernel. This information can be extracted from the kernel
605 image file with the script scripts/extract-ikconfig and used as
606 input to rebuild the current kernel or to build another kernel.
607 It can also be extracted from a running kernel by reading
608 /proc/config.gz if enabled (below).
611 bool "Enable access to .config through /proc/config.gz"
612 depends on IKCONFIG && PROC_FS
614 This option enables access to the kernel configuration file
615 through /proc/config.gz.
618 tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz"
621 This option enables access to the in-kernel headers that are generated during
622 the build process. These can be used to build eBPF tracing programs,
623 or similar programs. If you build the headers as a module, a module called
624 kheaders.ko is built which can be loaded on-demand to get access to headers.
627 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
632 Select the minimal kernel log buffer size as a power of 2.
633 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
634 parameter, see below. Any higher size also might be forced
635 by "log_buf_len" boot parameter.
645 config LOG_CPU_MAX_BUF_SHIFT
646 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
649 default 12 if !BASE_SMALL
650 default 0 if BASE_SMALL
653 This option allows to increase the default ring buffer size
654 according to the number of CPUs. The value defines the contribution
655 of each CPU as a power of 2. The used space is typically only few
656 lines however it might be much more when problems are reported,
659 The increased size means that a new buffer has to be allocated and
660 the original static one is unused. It makes sense only on systems
661 with more CPUs. Therefore this value is used only when the sum of
662 contributions is greater than the half of the default kernel ring
663 buffer as defined by LOG_BUF_SHIFT. The default values are set
664 so that more than 64 CPUs are needed to trigger the allocation.
666 Also this option is ignored when "log_buf_len" kernel parameter is
667 used as it forces an exact (power of two) size of the ring buffer.
669 The number of possible CPUs is used for this computation ignoring
670 hotplugging making the computation optimal for the worst case
671 scenario while allowing a simple algorithm to be used from bootup.
673 Examples shift values and their meaning:
674 17 => 128 KB for each CPU
675 16 => 64 KB for each CPU
676 15 => 32 KB for each CPU
677 14 => 16 KB for each CPU
678 13 => 8 KB for each CPU
679 12 => 4 KB for each CPU
681 config PRINTK_SAFE_LOG_BUF_SHIFT
682 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
687 Select the size of an alternate printk per-CPU buffer where messages
688 printed from usafe contexts are temporary stored. One example would
689 be NMI messages, another one - printk recursion. The messages are
690 copied to the main log buffer in a safe context to avoid a deadlock.
691 The value defines the size as a power of 2.
693 Those messages are rare and limited. The largest one is when
694 a backtrace is printed. It usually fits into 4KB. Select
695 8KB if you want to be on the safe side.
698 17 => 128 KB for each CPU
699 16 => 64 KB for each CPU
700 15 => 32 KB for each CPU
701 14 => 16 KB for each CPU
702 13 => 8 KB for each CPU
703 12 => 4 KB for each CPU
706 # Architectures with an unreliable sched_clock() should select this:
708 config HAVE_UNSTABLE_SCHED_CLOCK
711 config GENERIC_SCHED_CLOCK
714 menu "Scheduler features"
717 bool "Enable utilization clamping for RT/FAIR tasks"
718 depends on CPU_FREQ_GOV_SCHEDUTIL
720 This feature enables the scheduler to track the clamped utilization
721 of each CPU based on RUNNABLE tasks scheduled on that CPU.
723 With this option, the user can specify the min and max CPU
724 utilization allowed for RUNNABLE tasks. The max utilization defines
725 the maximum frequency a task should use while the min utilization
726 defines the minimum frequency it should use.
728 Both min and max utilization clamp values are hints to the scheduler,
729 aiming at improving its frequency selection policy, but they do not
730 enforce or grant any specific bandwidth for tasks.
734 config UCLAMP_BUCKETS_COUNT
735 int "Number of supported utilization clamp buckets"
738 depends on UCLAMP_TASK
740 Defines the number of clamp buckets to use. The range of each bucket
741 will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the
742 number of clamp buckets the finer their granularity and the higher
743 the precision of clamping aggregation and tracking at run-time.
745 For example, with the minimum configuration value we will have 5
746 clamp buckets tracking 20% utilization each. A 25% boosted tasks will
747 be refcounted in the [20..39]% bucket and will set the bucket clamp
748 effective value to 25%.
749 If a second 30% boosted task should be co-scheduled on the same CPU,
750 that task will be refcounted in the same bucket of the first task and
751 it will boost the bucket clamp effective value to 30%.
752 The clamp effective value of a bucket is reset to its nominal value
753 (20% in the example above) when there are no more tasks refcounted in
756 An additional boost/capping margin can be added to some tasks. In the
757 example above the 25% task will be boosted to 30% until it exits the
758 CPU. If that should be considered not acceptable on certain systems,
759 it's always possible to reduce the margin by increasing the number of
760 clamp buckets to trade off used memory for run-time tracking
763 If in doubt, use the default value.
768 # For architectures that want to enable the support for NUMA-affine scheduler
771 config ARCH_SUPPORTS_NUMA_BALANCING
775 # For architectures that prefer to flush all TLBs after a number of pages
776 # are unmapped instead of sending one IPI per page to flush. The architecture
777 # must provide guarantees on what happens if a clean TLB cache entry is
778 # written after the unmap. Details are in mm/rmap.c near the check for
779 # should_defer_flush. The architecture should also consider if the full flush
780 # and the refill costs are offset by the savings of sending fewer IPIs.
781 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
785 def_bool !$(cc-option,$(m64-flag) -D__SIZEOF_INT128__=0) && 64BIT
788 # For architectures that know their GCC __int128 support is sound
790 config ARCH_SUPPORTS_INT128
793 # For architectures that (ab)use NUMA to represent different memory regions
794 # all cpu-local but of different latencies, such as SuperH.
796 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
799 config NUMA_BALANCING
800 bool "Memory placement aware NUMA scheduler"
801 depends on ARCH_SUPPORTS_NUMA_BALANCING
802 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
803 depends on SMP && NUMA && MIGRATION
805 This option adds support for automatic NUMA aware memory/task placement.
806 The mechanism is quite primitive and is based on migrating memory when
807 it has references to the node the task is running on.
809 This system will be inactive on UMA systems.
811 config NUMA_BALANCING_DEFAULT_ENABLED
812 bool "Automatically enable NUMA aware memory/task placement"
814 depends on NUMA_BALANCING
816 If set, automatic NUMA balancing will be enabled if running on a NUMA
820 bool "Control Group support"
823 This option adds support for grouping sets of processes together, for
824 use with process control subsystems such as Cpusets, CFS, memory
825 controls or device isolation.
827 - Documentation/scheduler/sched-design-CFS.rst (CFS)
828 - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation
829 and resource control)
839 bool "Memory controller"
843 Provides control over the memory footprint of tasks in a cgroup.
846 bool "Swap controller"
847 depends on MEMCG && SWAP
849 Provides control over the swap space consumed by tasks in a cgroup.
851 config MEMCG_SWAP_ENABLED
852 bool "Swap controller enabled by default"
853 depends on MEMCG_SWAP
856 Memory Resource Controller Swap Extension comes with its price in
857 a bigger memory consumption. General purpose distribution kernels
858 which want to enable the feature but keep it disabled by default
859 and let the user enable it by swapaccount=1 boot command line
860 parameter should have this option unselected.
861 For those who want to have the feature enabled by default should
862 select this option (if, for some reason, they need to disable it
863 then swapaccount=0 does the trick).
867 depends on MEMCG && !SLOB
875 Generic block IO controller cgroup interface. This is the common
876 cgroup interface which should be used by various IO controlling
879 Currently, CFQ IO scheduler uses it to recognize task groups and
880 control disk bandwidth allocation (proportional time slice allocation)
881 to such task groups. It is also used by bio throttling logic in
882 block layer to implement upper limit in IO rates on a device.
884 This option only enables generic Block IO controller infrastructure.
885 One needs to also enable actual IO controlling logic/policy. For
886 enabling proportional weight division of disk bandwidth in CFQ, set
887 CONFIG_BFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
888 CONFIG_BLK_DEV_THROTTLING=y.
890 See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information.
892 config CGROUP_WRITEBACK
894 depends on MEMCG && BLK_CGROUP
897 menuconfig CGROUP_SCHED
898 bool "CPU controller"
901 This feature lets CPU scheduler recognize task groups and control CPU
902 bandwidth allocation to such task groups. It uses cgroups to group
906 config FAIR_GROUP_SCHED
907 bool "Group scheduling for SCHED_OTHER"
908 depends on CGROUP_SCHED
912 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
913 depends on FAIR_GROUP_SCHED
916 This option allows users to define CPU bandwidth rates (limits) for
917 tasks running within the fair group scheduler. Groups with no limit
918 set are considered to be unconstrained and will run with no
920 See Documentation/scheduler/sched-bwc.rst for more information.
922 config RT_GROUP_SCHED
923 bool "Group scheduling for SCHED_RR/FIFO"
924 depends on CGROUP_SCHED
927 This feature lets you explicitly allocate real CPU bandwidth
928 to task groups. If enabled, it will also make it impossible to
929 schedule realtime tasks for non-root users until you allocate
930 realtime bandwidth for them.
931 See Documentation/scheduler/sched-rt-group.rst for more information.
935 config UCLAMP_TASK_GROUP
936 bool "Utilization clamping per group of tasks"
937 depends on CGROUP_SCHED
938 depends on UCLAMP_TASK
941 This feature enables the scheduler to track the clamped utilization
942 of each CPU based on RUNNABLE tasks currently scheduled on that CPU.
944 When this option is enabled, the user can specify a min and max
945 CPU bandwidth which is allowed for each single task in a group.
946 The max bandwidth allows to clamp the maximum frequency a task
947 can use, while the min bandwidth allows to define a minimum
948 frequency a task will always use.
950 When task group based utilization clamping is enabled, an eventually
951 specified task-specific clamp value is constrained by the cgroup
952 specified clamp value. Both minimum and maximum task clamping cannot
953 be bigger than the corresponding clamping defined at task group level.
958 bool "PIDs controller"
960 Provides enforcement of process number limits in the scope of a
961 cgroup. Any attempt to fork more processes than is allowed in the
962 cgroup will fail. PIDs are fundamentally a global resource because it
963 is fairly trivial to reach PID exhaustion before you reach even a
964 conservative kmemcg limit. As a result, it is possible to grind a
965 system to halt without being limited by other cgroup policies. The
966 PIDs controller is designed to stop this from happening.
968 It should be noted that organisational operations (such as attaching
969 to a cgroup hierarchy) will *not* be blocked by the PIDs controller,
970 since the PIDs limit only affects a process's ability to fork, not to
974 bool "RDMA controller"
976 Provides enforcement of RDMA resources defined by IB stack.
977 It is fairly easy for consumers to exhaust RDMA resources, which
978 can result into resource unavailability to other consumers.
979 RDMA controller is designed to stop this from happening.
980 Attaching processes with active RDMA resources to the cgroup
981 hierarchy is allowed even if can cross the hierarchy's limit.
983 config CGROUP_FREEZER
984 bool "Freezer controller"
986 Provides a way to freeze and unfreeze all tasks in a
989 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
990 controller includes important in-kernel memory consumers per default.
992 If you're using cgroup2, say N.
994 config CGROUP_HUGETLB
995 bool "HugeTLB controller"
996 depends on HUGETLB_PAGE
1000 Provides a cgroup controller for HugeTLB pages.
1001 When you enable this, you can put a per cgroup limit on HugeTLB usage.
1002 The limit is enforced during page fault. Since HugeTLB doesn't
1003 support page reclaim, enforcing the limit at page fault time implies
1004 that, the application will get SIGBUS signal if it tries to access
1005 HugeTLB pages beyond its limit. This requires the application to know
1006 beforehand how much HugeTLB pages it would require for its use. The
1007 control group is tracked in the third page lru pointer. This means
1008 that we cannot use the controller with huge page less than 3 pages.
1011 bool "Cpuset controller"
1014 This option will let you create and manage CPUSETs which
1015 allow dynamically partitioning a system into sets of CPUs and
1016 Memory Nodes and assigning tasks to run only within those sets.
1017 This is primarily useful on large SMP or NUMA systems.
1021 config PROC_PID_CPUSET
1022 bool "Include legacy /proc/<pid>/cpuset file"
1026 config CGROUP_DEVICE
1027 bool "Device controller"
1029 Provides a cgroup controller implementing whitelists for
1030 devices which a process in the cgroup can mknod or open.
1032 config CGROUP_CPUACCT
1033 bool "Simple CPU accounting controller"
1035 Provides a simple controller for monitoring the
1036 total CPU consumed by the tasks in a cgroup.
1039 bool "Perf controller"
1040 depends on PERF_EVENTS
1042 This option extends the perf per-cpu mode to restrict monitoring
1043 to threads which belong to the cgroup specified and run on the
1044 designated cpu. Or this can be used to have cgroup ID in samples
1045 so that it can monitor performance events among cgroups.
1050 bool "Support for eBPF programs attached to cgroups"
1051 depends on BPF_SYSCALL
1052 select SOCK_CGROUP_DATA
1054 Allow attaching eBPF programs to a cgroup using the bpf(2)
1055 syscall command BPF_PROG_ATTACH.
1057 In which context these programs are accessed depends on the type
1058 of attachment. For instance, programs that are attached using
1059 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
1063 bool "Debug controller"
1065 depends on DEBUG_KERNEL
1067 This option enables a simple controller that exports
1068 debugging information about the cgroups framework. This
1069 controller is for control cgroup debugging only. Its
1070 interfaces are not stable.
1074 config SOCK_CGROUP_DATA
1080 menuconfig NAMESPACES
1081 bool "Namespaces support" if EXPERT
1082 depends on MULTIUSER
1085 Provides the way to make tasks work with different objects using
1086 the same id. For example same IPC id may refer to different objects
1087 or same user id or pid may refer to different tasks when used in
1088 different namespaces.
1093 bool "UTS namespace"
1096 In this namespace tasks see different info provided with the
1100 bool "TIME namespace"
1101 depends on GENERIC_VDSO_TIME_NS
1104 In this namespace boottime and monotonic clocks can be set.
1105 The time will keep going with the same pace.
1108 bool "IPC namespace"
1109 depends on (SYSVIPC || POSIX_MQUEUE)
1112 In this namespace tasks work with IPC ids which correspond to
1113 different IPC objects in different namespaces.
1116 bool "User namespace"
1119 This allows containers, i.e. vservers, to use user namespaces
1120 to provide different user info for different servers.
1122 When user namespaces are enabled in the kernel it is
1123 recommended that the MEMCG option also be enabled and that
1124 user-space use the memory control groups to limit the amount
1125 of memory a memory unprivileged users can use.
1130 bool "PID Namespaces"
1133 Support process id namespaces. This allows having multiple
1134 processes with the same pid as long as they are in different
1135 pid namespaces. This is a building block of containers.
1138 bool "Network namespace"
1142 Allow user space to create what appear to be multiple instances
1143 of the network stack.
1147 config CHECKPOINT_RESTORE
1148 bool "Checkpoint/restore support"
1149 select PROC_CHILDREN
1152 Enables additional kernel features in a sake of checkpoint/restore.
1153 In particular it adds auxiliary prctl codes to setup process text,
1154 data and heap segment sizes, and a few additional /proc filesystem
1157 If unsure, say N here.
1159 config SCHED_AUTOGROUP
1160 bool "Automatic process group scheduling"
1163 select FAIR_GROUP_SCHED
1165 This option optimizes the scheduler for common desktop workloads by
1166 automatically creating and populating task groups. This separation
1167 of workloads isolates aggressive CPU burners (like build jobs) from
1168 desktop applications. Task group autogeneration is currently based
1171 config SYSFS_DEPRECATED
1172 bool "Enable deprecated sysfs features to support old userspace tools"
1176 This option adds code that switches the layout of the "block" class
1177 devices, to not show up in /sys/class/block/, but only in
1180 This switch is only active when the sysfs.deprecated=1 boot option is
1181 passed or the SYSFS_DEPRECATED_V2 option is set.
1183 This option allows new kernels to run on old distributions and tools,
1184 which might get confused by /sys/class/block/. Since 2007/2008 all
1185 major distributions and tools handle this just fine.
1187 Recent distributions and userspace tools after 2009/2010 depend on
1188 the existence of /sys/class/block/, and will not work with this
1191 Only if you are using a new kernel on an old distribution, you might
1194 config SYSFS_DEPRECATED_V2
1195 bool "Enable deprecated sysfs features by default"
1198 depends on SYSFS_DEPRECATED
1200 Enable deprecated sysfs by default.
1202 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1205 Only if you are using a new kernel on an old distribution, you might
1206 need to say Y here. Even then, odds are you would not need it
1207 enabled, you can always pass the boot option if absolutely necessary.
1210 bool "Kernel->user space relay support (formerly relayfs)"
1213 This option enables support for relay interface support in
1214 certain file systems (such as debugfs).
1215 It is designed to provide an efficient mechanism for tools and
1216 facilities to relay large amounts of data from kernel space to
1221 config BLK_DEV_INITRD
1222 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1224 The initial RAM filesystem is a ramfs which is loaded by the
1225 boot loader (loadlin or lilo) and that is mounted as root
1226 before the normal boot procedure. It is typically used to
1227 load modules needed to mount the "real" root file system,
1228 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1230 If RAM disk support (BLK_DEV_RAM) is also included, this
1231 also enables initial RAM disk (initrd) support and adds
1232 15 Kbytes (more on some other architectures) to the kernel size.
1238 source "usr/Kconfig"
1243 bool "Boot config support"
1244 select BLK_DEV_INITRD
1246 Extra boot config allows system admin to pass a config file as
1247 complemental extension of kernel cmdline when booting.
1248 The boot config file must be attached at the end of initramfs
1249 with checksum, size and magic word.
1250 See <file:Documentation/admin-guide/bootconfig.rst> for details.
1255 prompt "Compiler optimization level"
1256 default CC_OPTIMIZE_FOR_PERFORMANCE
1258 config CC_OPTIMIZE_FOR_PERFORMANCE
1259 bool "Optimize for performance (-O2)"
1261 This is the default optimization level for the kernel, building
1262 with the "-O2" compiler flag for best performance and most
1263 helpful compile-time warnings.
1265 config CC_OPTIMIZE_FOR_PERFORMANCE_O3
1266 bool "Optimize more for performance (-O3)"
1269 Choosing this option will pass "-O3" to your compiler to optimize
1270 the kernel yet more for performance.
1272 config CC_OPTIMIZE_FOR_SIZE
1273 bool "Optimize for size (-Os)"
1275 Choosing this option will pass "-Os" to your compiler resulting
1276 in a smaller kernel.
1280 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1283 This requires that the arch annotates or otherwise protects
1284 its external entry points from being discarded. Linker scripts
1285 must also merge .text.*, .data.*, and .bss.* correctly into
1286 output sections. Care must be taken not to pull in unrelated
1287 sections (e.g., '.text.init'). Typically '.' in section names
1288 is used to distinguish them from label names / C identifiers.
1290 config LD_DEAD_CODE_DATA_ELIMINATION
1291 bool "Dead code and data elimination (EXPERIMENTAL)"
1292 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1294 depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800)
1295 depends on $(cc-option,-ffunction-sections -fdata-sections)
1296 depends on $(ld-option,--gc-sections)
1298 Enable this if you want to do dead code and data elimination with
1299 the linker by compiling with -ffunction-sections -fdata-sections,
1300 and linking with --gc-sections.
1302 This can reduce on disk and in-memory size of the kernel
1303 code and static data, particularly for small configs and
1304 on small systems. This has the possibility of introducing
1305 silently broken kernel if the required annotations are not
1306 present. This option is not well tested yet, so use at your
1315 config SYSCTL_EXCEPTION_TRACE
1318 Enable support for /proc/sys/debug/exception-trace.
1320 config SYSCTL_ARCH_UNALIGN_NO_WARN
1323 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1324 Allows arch to define/use @no_unaligned_warning to possibly warn
1325 about unaligned access emulation going on under the hood.
1327 config SYSCTL_ARCH_UNALIGN_ALLOW
1330 Enable support for /proc/sys/kernel/unaligned-trap
1331 Allows arches to define/use @unaligned_enabled to runtime toggle
1332 the unaligned access emulation.
1333 see arch/parisc/kernel/unaligned.c for reference
1335 config HAVE_PCSPKR_PLATFORM
1338 # interpreter that classic socket filters depend on
1343 bool "Configure standard kernel features (expert users)"
1344 # Unhide debug options, to make the on-by-default options visible
1347 This option allows certain base kernel options and settings
1348 to be disabled or tweaked. This is for specialized
1349 environments which can tolerate a "non-standard" kernel.
1350 Only use this if you really know what you are doing.
1353 bool "Enable 16-bit UID system calls" if EXPERT
1354 depends on HAVE_UID16 && MULTIUSER
1357 This enables the legacy 16-bit UID syscall wrappers.
1360 bool "Multiple users, groups and capabilities support" if EXPERT
1363 This option enables support for non-root users, groups and
1366 If you say N here, all processes will run with UID 0, GID 0, and all
1367 possible capabilities. Saying N here also compiles out support for
1368 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1371 If unsure, say Y here.
1373 config SGETMASK_SYSCALL
1374 bool "sgetmask/ssetmask syscalls support" if EXPERT
1375 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1377 sys_sgetmask and sys_ssetmask are obsolete system calls
1378 no longer supported in libc but still enabled by default in some
1381 If unsure, leave the default option here.
1383 config SYSFS_SYSCALL
1384 bool "Sysfs syscall support" if EXPERT
1387 sys_sysfs is an obsolete system call no longer supported in libc.
1388 Note that disabling this option is more secure but might break
1389 compatibility with some systems.
1391 If unsure say Y here.
1394 bool "open by fhandle syscalls" if EXPERT
1398 If you say Y here, a user level program will be able to map
1399 file names to handle and then later use the handle for
1400 different file system operations. This is useful in implementing
1401 userspace file servers, which now track files using handles instead
1402 of names. The handle would remain the same even if file names
1403 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1407 bool "Posix Clocks & timers" if EXPERT
1410 This includes native support for POSIX timers to the kernel.
1411 Some embedded systems have no use for them and therefore they
1412 can be configured out to reduce the size of the kernel image.
1414 When this option is disabled, the following syscalls won't be
1415 available: timer_create, timer_gettime: timer_getoverrun,
1416 timer_settime, timer_delete, clock_adjtime, getitimer,
1417 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1418 clock_getres and clock_nanosleep syscalls will be limited to
1419 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1425 bool "Enable support for printk" if EXPERT
1428 This option enables normal printk support. Removing it
1429 eliminates most of the message strings from the kernel image
1430 and makes the kernel more or less silent. As this makes it
1431 very difficult to diagnose system problems, saying N here is
1432 strongly discouraged.
1440 bool "BUG() support" if EXPERT
1443 Disabling this option eliminates support for BUG and WARN, reducing
1444 the size of your kernel image and potentially quietly ignoring
1445 numerous fatal conditions. You should only consider disabling this
1446 option for embedded systems with no facilities for reporting errors.
1452 bool "Enable ELF core dumps" if EXPERT
1454 Enable support for generating core dumps. Disabling saves about 4k.
1457 config PCSPKR_PLATFORM
1458 bool "Enable PC-Speaker support" if EXPERT
1459 depends on HAVE_PCSPKR_PLATFORM
1463 This option allows to disable the internal PC-Speaker
1464 support, saving some memory.
1468 bool "Enable full-sized data structures for core" if EXPERT
1470 Disabling this option reduces the size of miscellaneous core
1471 kernel data structures. This saves memory on small machines,
1472 but may reduce performance.
1475 bool "Enable futex support" if EXPERT
1479 Disabling this option will cause the kernel to be built without
1480 support for "fast userspace mutexes". The resulting kernel may not
1481 run glibc-based applications correctly.
1485 depends on FUTEX && RT_MUTEXES
1488 config HAVE_FUTEX_CMPXCHG
1492 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1493 is implemented and always working. This removes a couple of runtime
1497 bool "Enable eventpoll support" if EXPERT
1500 Disabling this option will cause the kernel to be built without
1501 support for epoll family of system calls.
1504 bool "Enable signalfd() system call" if EXPERT
1507 Enable the signalfd() system call that allows to receive signals
1508 on a file descriptor.
1513 bool "Enable timerfd() system call" if EXPERT
1516 Enable the timerfd() system call that allows to receive timer
1517 events on a file descriptor.
1522 bool "Enable eventfd() system call" if EXPERT
1525 Enable the eventfd() system call that allows to receive both
1526 kernel notification (ie. KAIO) or userspace notifications.
1531 bool "Use full shmem filesystem" if EXPERT
1535 The shmem is an internal filesystem used to manage shared memory.
1536 It is backed by swap and manages resource limits. It is also exported
1537 to userspace as tmpfs if TMPFS is enabled. Disabling this
1538 option replaces shmem and tmpfs with the much simpler ramfs code,
1539 which may be appropriate on small systems without swap.
1542 bool "Enable AIO support" if EXPERT
1545 This option enables POSIX asynchronous I/O which may by used
1546 by some high performance threaded applications. Disabling
1547 this option saves about 7k.
1550 bool "Enable IO uring support" if EXPERT
1554 This option enables support for the io_uring interface, enabling
1555 applications to submit and complete IO through submission and
1556 completion rings that are shared between the kernel and application.
1558 config ADVISE_SYSCALLS
1559 bool "Enable madvise/fadvise syscalls" if EXPERT
1562 This option enables the madvise and fadvise syscalls, used by
1563 applications to advise the kernel about their future memory or file
1564 usage, improving performance. If building an embedded system where no
1565 applications use these syscalls, you can disable this option to save
1568 config HAVE_ARCH_USERFAULTFD_WP
1571 Arch has userfaultfd write protection support
1574 bool "Enable membarrier() system call" if EXPERT
1577 Enable the membarrier() system call that allows issuing memory
1578 barriers across all running threads, which can be used to distribute
1579 the cost of user-space memory barriers asymmetrically by transforming
1580 pairs of memory barriers into pairs consisting of membarrier() and a
1586 bool "Load all symbols for debugging/ksymoops" if EXPERT
1589 Say Y here to let the kernel print out symbolic crash information and
1590 symbolic stack backtraces. This increases the size of the kernel
1591 somewhat, as all symbols have to be loaded into the kernel image.
1594 bool "Include all symbols in kallsyms"
1595 depends on DEBUG_KERNEL && KALLSYMS
1597 Normally kallsyms only contains the symbols of functions for nicer
1598 OOPS messages and backtraces (i.e., symbols from the text and inittext
1599 sections). This is sufficient for most cases. And only in very rare
1600 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1601 names of variables from the data sections, etc).
1603 This option makes sure that all symbols are loaded into the kernel
1604 image (i.e., symbols from all sections) in cost of increased kernel
1605 size (depending on the kernel configuration, it may be 300KiB or
1606 something like this).
1608 Say N unless you really need all symbols.
1610 config KALLSYMS_ABSOLUTE_PERCPU
1613 default X86_64 && SMP
1615 config KALLSYMS_BASE_RELATIVE
1620 Instead of emitting them as absolute values in the native word size,
1621 emit the symbol references in the kallsyms table as 32-bit entries,
1622 each containing a relative value in the range [base, base + U32_MAX]
1623 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1624 an absolute value in the range [0, S32_MAX] or a relative value in the
1625 range [base, base + S32_MAX], where base is the lowest relative symbol
1626 address encountered in the image.
1628 On 64-bit builds, this reduces the size of the address table by 50%,
1629 but more importantly, it results in entries whose values are build
1630 time constants, and no relocation pass is required at runtime to fix
1631 up the entries based on the runtime load address of the kernel.
1633 # end of the "standard kernel features (expert users)" menu
1635 # syscall, maps, verifier
1638 bool "LSM Instrumentation with BPF"
1639 depends on BPF_EVENTS
1640 depends on BPF_SYSCALL
1644 Enables instrumentation of the security hooks with eBPF programs for
1645 implementing dynamic MAC and Audit Policies.
1647 If you are unsure how to answer this question, answer N.
1650 bool "Enable bpf() system call"
1655 Enable the bpf() system call that allows to manipulate eBPF
1656 programs and maps via file descriptors.
1658 config ARCH_WANT_DEFAULT_BPF_JIT
1661 config BPF_JIT_ALWAYS_ON
1662 bool "Permanently enable BPF JIT and remove BPF interpreter"
1663 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1665 Enables BPF JIT and removes BPF interpreter to avoid
1666 speculative execution of BPF instructions by the interpreter
1668 config BPF_JIT_DEFAULT_ON
1669 def_bool ARCH_WANT_DEFAULT_BPF_JIT || BPF_JIT_ALWAYS_ON
1670 depends on HAVE_EBPF_JIT && BPF_JIT
1673 bool "Enable userfaultfd() system call"
1676 Enable the userfaultfd() system call that allows to intercept and
1677 handle page faults in userland.
1679 config ARCH_HAS_MEMBARRIER_CALLBACKS
1682 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1686 bool "Enable rseq() system call" if EXPERT
1688 depends on HAVE_RSEQ
1691 Enable the restartable sequences system call. It provides a
1692 user-space cache for the current CPU number value, which
1693 speeds up getting the current CPU number from user-space,
1694 as well as an ABI to speed up user-space operations on
1701 bool "Enabled debugging of rseq() system call" if EXPERT
1702 depends on RSEQ && DEBUG_KERNEL
1704 Enable extra debugging checks for the rseq system call.
1709 bool "Embedded system"
1710 option allnoconfig_y
1713 This option should be enabled if compiling the kernel for
1714 an embedded system so certain expert options are available
1717 config HAVE_PERF_EVENTS
1720 See tools/perf/design.txt for details.
1722 config PERF_USE_VMALLOC
1725 See tools/perf/design.txt for details
1728 bool "PC/104 support" if EXPERT
1730 Expose PC/104 form factor device drivers and options available for
1731 selection and configuration. Enable this option if your target
1732 machine has a PC/104 bus.
1734 menu "Kernel Performance Events And Counters"
1737 bool "Kernel performance events and counters"
1738 default y if PROFILING
1739 depends on HAVE_PERF_EVENTS
1743 Enable kernel support for various performance events provided
1744 by software and hardware.
1746 Software events are supported either built-in or via the
1747 use of generic tracepoints.
1749 Most modern CPUs support performance events via performance
1750 counter registers. These registers count the number of certain
1751 types of hw events: such as instructions executed, cachemisses
1752 suffered, or branches mis-predicted - without slowing down the
1753 kernel or applications. These registers can also trigger interrupts
1754 when a threshold number of events have passed - and can thus be
1755 used to profile the code that runs on that CPU.
1757 The Linux Performance Event subsystem provides an abstraction of
1758 these software and hardware event capabilities, available via a
1759 system call and used by the "perf" utility in tools/perf/. It
1760 provides per task and per CPU counters, and it provides event
1761 capabilities on top of those.
1765 config DEBUG_PERF_USE_VMALLOC
1767 bool "Debug: use vmalloc to back perf mmap() buffers"
1768 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1769 select PERF_USE_VMALLOC
1771 Use vmalloc memory to back perf mmap() buffers.
1773 Mostly useful for debugging the vmalloc code on platforms
1774 that don't require it.
1780 config VM_EVENT_COUNTERS
1782 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1784 VM event counters are needed for event counts to be shown.
1785 This option allows the disabling of the VM event counters
1786 on EXPERT systems. /proc/vmstat will only show page counts
1787 if VM event counters are disabled.
1791 bool "Enable SLUB debugging support" if EXPERT
1792 depends on SLUB && SYSFS
1794 SLUB has extensive debug support features. Disabling these can
1795 result in significant savings in code size. This also disables
1796 SLUB sysfs support. /sys/slab will not exist and there will be
1797 no support for cache validation etc.
1799 config SLUB_MEMCG_SYSFS_ON
1801 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1802 depends on SLUB && SYSFS && MEMCG
1804 SLUB creates a directory under /sys/kernel/slab for each
1805 allocation cache to host info and debug files. If memory
1806 cgroup is enabled, each cache can have per memory cgroup
1807 caches. SLUB can create the same sysfs directories for these
1808 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1809 to a very high number of debug files being created. This is
1810 controlled by slub_memcg_sysfs boot parameter and this
1811 config option determines the parameter's default value.
1814 bool "Disable heap randomization"
1817 Randomizing heap placement makes heap exploits harder, but it
1818 also breaks ancient binaries (including anything libc5 based).
1819 This option changes the bootup default to heap randomization
1820 disabled, and can be overridden at runtime by setting
1821 /proc/sys/kernel/randomize_va_space to 2.
1823 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1826 prompt "Choose SLAB allocator"
1829 This option allows to select a slab allocator.
1833 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1835 The regular slab allocator that is established and known to work
1836 well in all environments. It organizes cache hot objects in
1837 per cpu and per node queues.
1840 bool "SLUB (Unqueued Allocator)"
1841 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1843 SLUB is a slab allocator that minimizes cache line usage
1844 instead of managing queues of cached objects (SLAB approach).
1845 Per cpu caching is realized using slabs of objects instead
1846 of queues of objects. SLUB can use memory efficiently
1847 and has enhanced diagnostics. SLUB is the default choice for
1852 bool "SLOB (Simple Allocator)"
1854 SLOB replaces the stock allocator with a drastically simpler
1855 allocator. SLOB is generally more space efficient but
1856 does not perform as well on large systems.
1860 config SLAB_MERGE_DEFAULT
1861 bool "Allow slab caches to be merged"
1864 For reduced kernel memory fragmentation, slab caches can be
1865 merged when they share the same size and other characteristics.
1866 This carries a risk of kernel heap overflows being able to
1867 overwrite objects from merged caches (and more easily control
1868 cache layout), which makes such heap attacks easier to exploit
1869 by attackers. By keeping caches unmerged, these kinds of exploits
1870 can usually only damage objects in the same cache. To disable
1871 merging at runtime, "slab_nomerge" can be passed on the kernel
1874 config SLAB_FREELIST_RANDOM
1876 depends on SLAB || SLUB
1877 bool "SLAB freelist randomization"
1879 Randomizes the freelist order used on creating new pages. This
1880 security feature reduces the predictability of the kernel slab
1881 allocator against heap overflows.
1883 config SLAB_FREELIST_HARDENED
1884 bool "Harden slab freelist metadata"
1887 Many kernel heap attacks try to target slab cache metadata and
1888 other infrastructure. This options makes minor performance
1889 sacrifices to harden the kernel slab allocator against common
1890 freelist exploit methods.
1892 config SHUFFLE_PAGE_ALLOCATOR
1893 bool "Page allocator randomization"
1894 default SLAB_FREELIST_RANDOM && ACPI_NUMA
1896 Randomization of the page allocator improves the average
1897 utilization of a direct-mapped memory-side-cache. See section
1898 5.2.27 Heterogeneous Memory Attribute Table (HMAT) in the ACPI
1899 6.2a specification for an example of how a platform advertises
1900 the presence of a memory-side-cache. There are also incidental
1901 security benefits as it reduces the predictability of page
1902 allocations to compliment SLAB_FREELIST_RANDOM, but the
1903 default granularity of shuffling on the "MAX_ORDER - 1" i.e,
1904 10th order of pages is selected based on cache utilization
1907 While the randomization improves cache utilization it may
1908 negatively impact workloads on platforms without a cache. For
1909 this reason, by default, the randomization is enabled only
1910 after runtime detection of a direct-mapped memory-side-cache.
1911 Otherwise, the randomization may be force enabled with the
1912 'page_alloc.shuffle' kernel command line parameter.
1916 config SLUB_CPU_PARTIAL
1918 depends on SLUB && SMP
1919 bool "SLUB per cpu partial cache"
1921 Per cpu partial caches accelerate objects allocation and freeing
1922 that is local to a processor at the price of more indeterminism
1923 in the latency of the free. On overflow these caches will be cleared
1924 which requires the taking of locks that may cause latency spikes.
1925 Typically one would choose no for a realtime system.
1927 config MMAP_ALLOW_UNINITIALIZED
1928 bool "Allow mmapped anonymous memory to be uninitialized"
1929 depends on EXPERT && !MMU
1932 Normally, and according to the Linux spec, anonymous memory obtained
1933 from mmap() has its contents cleared before it is passed to
1934 userspace. Enabling this config option allows you to request that
1935 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1936 providing a huge performance boost. If this option is not enabled,
1937 then the flag will be ignored.
1939 This is taken advantage of by uClibc's malloc(), and also by
1940 ELF-FDPIC binfmt's brk and stack allocator.
1942 Because of the obvious security issues, this option should only be
1943 enabled on embedded devices where you control what is run in
1944 userspace. Since that isn't generally a problem on no-MMU systems,
1945 it is normally safe to say Y here.
1947 See Documentation/nommu-mmap.txt for more information.
1949 config SYSTEM_DATA_VERIFICATION
1951 select SYSTEM_TRUSTED_KEYRING
1955 select ASYMMETRIC_KEY_TYPE
1956 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1959 select X509_CERTIFICATE_PARSER
1960 select PKCS7_MESSAGE_PARSER
1962 Provide PKCS#7 message verification using the contents of the system
1963 trusted keyring to provide public keys. This then can be used for
1964 module verification, kexec image verification and firmware blob
1968 bool "Profiling support"
1970 Say Y here to enable the extended profiling support mechanisms used
1971 by profilers such as OProfile.
1974 # Place an empty function call at each tracepoint site. Can be
1975 # dynamically changed for a probe function.
1980 endmenu # General setup
1982 source "arch/Kconfig"
1989 default 0 if BASE_FULL
1990 default 1 if !BASE_FULL
1992 config MODULE_SIG_FORMAT
1994 select SYSTEM_DATA_VERIFICATION
1997 bool "Enable loadable module support"
2000 Kernel modules are small pieces of compiled code which can
2001 be inserted in the running kernel, rather than being
2002 permanently built into the kernel. You use the "modprobe"
2003 tool to add (and sometimes remove) them. If you say Y here,
2004 many parts of the kernel can be built as modules (by
2005 answering M instead of Y where indicated): this is most
2006 useful for infrequently used options which are not required
2007 for booting. For more information, see the man pages for
2008 modprobe, lsmod, modinfo, insmod and rmmod.
2010 If you say Y here, you will need to run "make
2011 modules_install" to put the modules under /lib/modules/
2012 where modprobe can find them (you may need to be root to do
2019 config MODULE_FORCE_LOAD
2020 bool "Forced module loading"
2023 Allow loading of modules without version information (ie. modprobe
2024 --force). Forced module loading sets the 'F' (forced) taint flag and
2025 is usually a really bad idea.
2027 config MODULE_UNLOAD
2028 bool "Module unloading"
2030 Without this option you will not be able to unload any
2031 modules (note that some modules may not be unloadable
2032 anyway), which makes your kernel smaller, faster
2033 and simpler. If unsure, say Y.
2035 config MODULE_FORCE_UNLOAD
2036 bool "Forced module unloading"
2037 depends on MODULE_UNLOAD
2039 This option allows you to force a module to unload, even if the
2040 kernel believes it is unsafe: the kernel will remove the module
2041 without waiting for anyone to stop using it (using the -f option to
2042 rmmod). This is mainly for kernel developers and desperate users.
2046 bool "Module versioning support"
2048 Usually, you have to use modules compiled with your kernel.
2049 Saying Y here makes it sometimes possible to use modules
2050 compiled for different kernels, by adding enough information
2051 to the modules to (hopefully) spot any changes which would
2052 make them incompatible with the kernel you are running. If
2055 config ASM_MODVERSIONS
2057 default HAVE_ASM_MODVERSIONS && MODVERSIONS
2059 This enables module versioning for exported symbols also from
2060 assembly. This can be enabled only when the target architecture
2063 config MODULE_REL_CRCS
2065 depends on MODVERSIONS
2067 config MODULE_SRCVERSION_ALL
2068 bool "Source checksum for all modules"
2070 Modules which contain a MODULE_VERSION get an extra "srcversion"
2071 field inserted into their modinfo section, which contains a
2072 sum of the source files which made it. This helps maintainers
2073 see exactly which source was used to build a module (since
2074 others sometimes change the module source without updating
2075 the version). With this option, such a "srcversion" field
2076 will be created for all modules. If unsure, say N.
2079 bool "Module signature verification"
2080 select MODULE_SIG_FORMAT
2082 Check modules for valid signatures upon load: the signature
2083 is simply appended to the module. For more information see
2084 <file:Documentation/admin-guide/module-signing.rst>.
2086 Note that this option adds the OpenSSL development packages as a
2087 kernel build dependency so that the signing tool can use its crypto
2090 You should enable this option if you wish to use either
2091 CONFIG_SECURITY_LOCKDOWN_LSM or lockdown functionality imposed via
2092 another LSM - otherwise unsigned modules will be loadable regardless
2093 of the lockdown policy.
2095 !!!WARNING!!! If you enable this option, you MUST make sure that the
2096 module DOES NOT get stripped after being signed. This includes the
2097 debuginfo strip done by some packagers (such as rpmbuild) and
2098 inclusion into an initramfs that wants the module size reduced.
2100 config MODULE_SIG_FORCE
2101 bool "Require modules to be validly signed"
2102 depends on MODULE_SIG
2104 Reject unsigned modules or signed modules for which we don't have a
2105 key. Without this, such modules will simply taint the kernel.
2107 config MODULE_SIG_ALL
2108 bool "Automatically sign all modules"
2110 depends on MODULE_SIG
2112 Sign all modules during make modules_install. Without this option,
2113 modules must be signed manually, using the scripts/sign-file tool.
2115 comment "Do not forget to sign required modules with scripts/sign-file"
2116 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
2119 prompt "Which hash algorithm should modules be signed with?"
2120 depends on MODULE_SIG
2122 This determines which sort of hashing algorithm will be used during
2123 signature generation. This algorithm _must_ be built into the kernel
2124 directly so that signature verification can take place. It is not
2125 possible to load a signed module containing the algorithm to check
2126 the signature on that module.
2128 config MODULE_SIG_SHA1
2129 bool "Sign modules with SHA-1"
2132 config MODULE_SIG_SHA224
2133 bool "Sign modules with SHA-224"
2134 select CRYPTO_SHA256
2136 config MODULE_SIG_SHA256
2137 bool "Sign modules with SHA-256"
2138 select CRYPTO_SHA256
2140 config MODULE_SIG_SHA384
2141 bool "Sign modules with SHA-384"
2142 select CRYPTO_SHA512
2144 config MODULE_SIG_SHA512
2145 bool "Sign modules with SHA-512"
2146 select CRYPTO_SHA512
2150 config MODULE_SIG_HASH
2152 depends on MODULE_SIG
2153 default "sha1" if MODULE_SIG_SHA1
2154 default "sha224" if MODULE_SIG_SHA224
2155 default "sha256" if MODULE_SIG_SHA256
2156 default "sha384" if MODULE_SIG_SHA384
2157 default "sha512" if MODULE_SIG_SHA512
2159 config MODULE_COMPRESS
2160 bool "Compress modules on installation"
2163 Compresses kernel modules when 'make modules_install' is run; gzip or
2164 xz depending on "Compression algorithm" below.
2166 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
2168 Out-of-tree kernel modules installed using Kbuild will also be
2169 compressed upon installation.
2171 Note: for modules inside an initrd or initramfs, it's more efficient
2172 to compress the whole initrd or initramfs instead.
2174 Note: This is fully compatible with signed modules.
2179 prompt "Compression algorithm"
2180 depends on MODULE_COMPRESS
2181 default MODULE_COMPRESS_GZIP
2183 This determines which sort of compression will be used during
2184 'make modules_install'.
2186 GZIP (default) and XZ are supported.
2188 config MODULE_COMPRESS_GZIP
2191 config MODULE_COMPRESS_XZ
2196 config MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS
2197 bool "Allow loading of modules with missing namespace imports"
2199 Symbols exported with EXPORT_SYMBOL_NS*() are considered exported in
2200 a namespace. A module that makes use of a symbol exported with such a
2201 namespace is required to import the namespace via MODULE_IMPORT_NS().
2202 There is no technical reason to enforce correct namespace imports,
2203 but it creates consistency between symbols defining namespaces and
2204 users importing namespaces they make use of. This option relaxes this
2205 requirement and lifts the enforcement when loading a module.
2209 config UNUSED_SYMBOLS
2210 bool "Enable unused/obsolete exported symbols"
2213 Unused but exported symbols make the kernel needlessly bigger. For
2214 that reason most of these unused exports will soon be removed. This
2215 option is provided temporarily to provide a transition period in case
2216 some external kernel module needs one of these symbols anyway. If you
2217 encounter such a case in your module, consider if you are actually
2218 using the right API. (rationale: since nobody in the kernel is using
2219 this in a module, there is a pretty good chance it's actually the
2220 wrong interface to use). If you really need the symbol, please send a
2221 mail to the linux kernel mailing list mentioning the symbol and why
2222 you really need it, and what the merge plan to the mainline kernel for
2225 config TRIM_UNUSED_KSYMS
2226 bool "Trim unused exported kernel symbols"
2227 depends on !UNUSED_SYMBOLS
2229 The kernel and some modules make many symbols available for
2230 other modules to use via EXPORT_SYMBOL() and variants. Depending
2231 on the set of modules being selected in your kernel configuration,
2232 many of those exported symbols might never be used.
2234 This option allows for unused exported symbols to be dropped from
2235 the build. In turn, this provides the compiler more opportunities
2236 (especially when using LTO) for optimizing the code and reducing
2237 binary size. This might have some security advantages as well.
2239 If unsure, or if you need to build out-of-tree modules, say N.
2241 config UNUSED_KSYMS_WHITELIST
2242 string "Whitelist of symbols to keep in ksymtab"
2243 depends on TRIM_UNUSED_KSYMS
2245 By default, all unused exported symbols will be un-exported from the
2246 build when TRIM_UNUSED_KSYMS is selected.
2248 UNUSED_KSYMS_WHITELIST allows to whitelist symbols that must be kept
2249 exported at all times, even in absence of in-tree users. The value to
2250 set here is the path to a text file containing the list of symbols,
2251 one per line. The path can be absolute, or relative to the kernel
2256 config MODULES_TREE_LOOKUP
2258 depends on PERF_EVENTS || TRACING
2260 config INIT_ALL_POSSIBLE
2263 Back when each arch used to define their own cpu_online_mask and
2264 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2265 with all 1s, and others with all 0s. When they were centralised,
2266 it was better to provide this option than to break all the archs
2267 and have several arch maintainers pursuing me down dark alleys.
2269 source "block/Kconfig"
2271 config PREEMPT_NOTIFIERS
2281 Build a simple ASN.1 grammar compiler that produces a bytecode output
2282 that can be interpreted by the ASN.1 stream decoder and used to
2283 inform it as to what tags are to be expected in a stream and what
2284 functions to call on what tags.
2286 source "kernel/Kconfig.locks"
2288 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2291 # It may be useful for an architecture to override the definitions of the
2292 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2293 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2294 # different calling convention for syscalls. They can also override the
2295 # macros for not-implemented syscalls in kernel/sys_ni.c and
2296 # kernel/time/posix-stubs.c. All these overrides need to be available in
2297 # <asm/syscall_wrapper.h>.
2298 config ARCH_HAS_SYSCALL_WRAPPER
projects / linux.git / blob