1 # SPDX-License-Identifier: GPL-2.0-only
4 default "$(CC_VERSION_TEXT)"
6 This is used in unclear ways:
8 - Re-run Kconfig when the compiler is updated
9 The 'default' property references the environment variable,
10 CC_VERSION_TEXT so it is recorded in include/config/auto.conf.cmd.
11 When the compiler is updated, Kconfig will be invoked.
13 - Ensure full rebuild when the compiler is updated
14 include/linux/compiler-version.h contains this option in the comment
15 line so fixdep adds include/config/CC_VERSION_TEXT into the
16 auto-generated dependency. When the compiler is updated, syncconfig
17 will touch it and then every file will be rebuilt.
20 def_bool $(success,test "$(cc-name)" = GCC)
24 default $(cc-version) if CC_IS_GCC
28 def_bool $(success,test "$(cc-name)" = Clang)
32 default $(cc-version) if CC_IS_CLANG
36 def_bool $(success,test "$(as-name)" = GNU)
39 def_bool $(success,test "$(as-name)" = LLVM)
43 # Use clang version if this is the integrated assembler
44 default CLANG_VERSION if AS_IS_LLVM
48 def_bool $(success,test "$(ld-name)" = BFD)
52 default $(ld-version) if LD_IS_BFD
56 def_bool $(success,test "$(ld-name)" = LLD)
60 default $(ld-version) if LD_IS_LLD
65 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m64-flag)) if 64BIT
66 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m32-flag))
68 config CC_CAN_LINK_STATIC
70 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m64-flag) -static) if 64BIT
71 default $(success,$(srctree)/scripts/cc-can-link.sh $(CC) $(CLANG_FLAGS) $(USERCFLAGS) $(USERLDFLAGS) $(m32-flag) -static)
73 config CC_HAS_ASM_GOTO
74 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
76 config CC_HAS_ASM_GOTO_OUTPUT
77 depends on CC_HAS_ASM_GOTO
78 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)
80 config TOOLS_SUPPORT_RELR
81 def_bool $(success,env "CC=$(CC)" "LD=$(LD)" "NM=$(NM)" "OBJCOPY=$(OBJCOPY)" $(srctree)/scripts/tools-support-relr.sh)
83 config CC_HAS_ASM_INLINE
84 def_bool $(success,echo 'void foo(void) { asm inline (""); }' | $(CC) -x c - -c -o /dev/null)
86 config CC_HAS_NO_PROFILE_FN_ATTR
87 def_bool $(success,echo '__attribute__((no_profile_instrument_function)) int x();' | $(CC) -x c - -c -o /dev/null -Werror)
91 default $(shell,$(srctree)/scripts/pahole-version.sh $(PAHOLE))
99 config BUILDTIME_TABLE_SORT
102 config THREAD_INFO_IN_TASK
105 Select this to move thread_info off the stack into task_struct. To
106 make this work, an arch will need to remove all thread_info fields
107 except flags and fix any runtime bugs.
109 One subtle change that will be needed is to use try_get_task_stack()
110 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
119 depends on BROKEN || !SMP
122 config INIT_ENV_ARG_LIMIT
127 Maximum of each of the number of arguments and environment
128 variables passed to init from the kernel command line.
131 bool "Compile also drivers which will not load"
134 Some drivers can be compiled on a different platform than they are
135 intended to be run on. Despite they cannot be loaded there (or even
136 when they load they cannot be used due to missing HW support),
137 developers still, opposing to distributors, might want to build such
138 drivers to compile-test them.
140 If you are a developer and want to build everything available, say Y
141 here. If you are a user/distributor, say N here to exclude useless
142 drivers to be distributed.
145 bool "Compile the kernel with warnings as errors"
148 A kernel build should not cause any compiler warnings, and this
149 enables the '-Werror' flag to enforce that rule by default.
151 However, if you have a new (or very old) compiler with odd and
152 unusual warnings, or you have some architecture with problems,
153 you may need to disable this config option in order to
154 successfully build the kernel.
158 config UAPI_HEADER_TEST
159 bool "Compile test UAPI headers"
160 depends on HEADERS_INSTALL && CC_CAN_LINK
162 Compile test headers exported to user-space to ensure they are
163 self-contained, i.e. compilable as standalone units.
165 If you are a developer or tester and want to ensure the exported
166 headers are self-contained, say Y here. Otherwise, choose N.
169 string "Local version - append to kernel release"
171 Append an extra string to the end of your kernel version.
172 This will show up when you type uname, for example.
173 The string you set here will be appended after the contents of
174 any files with a filename matching localversion* in your
175 object and source tree, in that order. Your total string can
176 be a maximum of 64 characters.
178 config LOCALVERSION_AUTO
179 bool "Automatically append version information to the version string"
181 depends on !COMPILE_TEST
183 This will try to automatically determine if the current tree is a
184 release tree by looking for git tags that belong to the current
185 top of tree revision.
187 A string of the format -gxxxxxxxx will be added to the localversion
188 if a git-based tree is found. The string generated by this will be
189 appended after any matching localversion* files, and after the value
190 set in CONFIG_LOCALVERSION.
192 (The actual string used here is the first eight characters produced
193 by running the command:
195 $ git rev-parse --verify HEAD
197 which is done within the script "scripts/setlocalversion".)
200 string "Build ID Salt"
203 The build ID is used to link binaries and their debug info. Setting
204 this option will use the value in the calculation of the build id.
205 This is mostly useful for distributions which want to ensure the
206 build is unique between builds. It's safe to leave the default.
208 config HAVE_KERNEL_GZIP
211 config HAVE_KERNEL_BZIP2
214 config HAVE_KERNEL_LZMA
217 config HAVE_KERNEL_XZ
220 config HAVE_KERNEL_LZO
223 config HAVE_KERNEL_LZ4
226 config HAVE_KERNEL_ZSTD
229 config HAVE_KERNEL_UNCOMPRESSED
233 prompt "Kernel compression mode"
235 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_ZSTD || HAVE_KERNEL_UNCOMPRESSED
237 The linux kernel is a kind of self-extracting executable.
238 Several compression algorithms are available, which differ
239 in efficiency, compression and decompression speed.
240 Compression speed is only relevant when building a kernel.
241 Decompression speed is relevant at each boot.
243 If you have any problems with bzip2 or lzma compressed
244 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
245 version of this functionality (bzip2 only), for 2.4, was
246 supplied by Christian Ludwig)
248 High compression options are mostly useful for users, who
249 are low on disk space (embedded systems), but for whom ram
252 If in doubt, select 'gzip'
256 depends on HAVE_KERNEL_GZIP
258 The old and tried gzip compression. It provides a good balance
259 between compression ratio and decompression speed.
263 depends on HAVE_KERNEL_BZIP2
265 Its compression ratio and speed is intermediate.
266 Decompression speed is slowest among the choices. The kernel
267 size is about 10% smaller with bzip2, in comparison to gzip.
268 Bzip2 uses a large amount of memory. For modern kernels you
269 will need at least 8MB RAM or more for booting.
273 depends on HAVE_KERNEL_LZMA
275 This compression algorithm's ratio is best. Decompression speed
276 is between gzip and bzip2. Compression is slowest.
277 The kernel size is about 33% smaller with LZMA in comparison to gzip.
281 depends on HAVE_KERNEL_XZ
283 XZ uses the LZMA2 algorithm and instruction set specific
284 BCJ filters which can improve compression ratio of executable
285 code. The size of the kernel is about 30% smaller with XZ in
286 comparison to gzip. On architectures for which there is a BCJ
287 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
288 will create a few percent smaller kernel than plain LZMA.
290 The speed is about the same as with LZMA: The decompression
291 speed of XZ is better than that of bzip2 but worse than gzip
292 and LZO. Compression is slow.
296 depends on HAVE_KERNEL_LZO
298 Its compression ratio is the poorest among the choices. The kernel
299 size is about 10% bigger than gzip; however its speed
300 (both compression and decompression) is the fastest.
304 depends on HAVE_KERNEL_LZ4
306 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
307 A preliminary version of LZ4 de/compression tool is available at
308 <https://code.google.com/p/lz4/>.
310 Its compression ratio is worse than LZO. The size of the kernel
311 is about 8% bigger than LZO. But the decompression speed is
316 depends on HAVE_KERNEL_ZSTD
318 ZSTD is a compression algorithm targeting intermediate compression
319 with fast decompression speed. It will compress better than GZIP and
320 decompress around the same speed as LZO, but slower than LZ4. You
321 will need at least 192 KB RAM or more for booting. The zstd command
322 line tool is required for compression.
324 config KERNEL_UNCOMPRESSED
326 depends on HAVE_KERNEL_UNCOMPRESSED
328 Produce uncompressed kernel image. This option is usually not what
329 you want. It is useful for debugging the kernel in slow simulation
330 environments, where decompressing and moving the kernel is awfully
331 slow. This option allows early boot code to skip the decompressor
332 and jump right at uncompressed kernel image.
337 string "Default init path"
340 This option determines the default init for the system if no init=
341 option is passed on the kernel command line. If the requested path is
342 not present, we will still then move on to attempting further
343 locations (e.g. /sbin/init, etc). If this is empty, we will just use
344 the fallback list when init= is not passed.
346 config DEFAULT_HOSTNAME
347 string "Default hostname"
350 This option determines the default system hostname before userspace
351 calls sethostname(2). The kernel traditionally uses "(none)" here,
352 but you may wish to use a different default here to make a minimal
353 system more usable with less configuration.
358 Inter Process Communication is a suite of library functions and
359 system calls which let processes (running programs) synchronize and
360 exchange information. It is generally considered to be a good thing,
361 and some programs won't run unless you say Y here. In particular, if
362 you want to run the DOS emulator dosemu under Linux (read the
363 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
364 you'll need to say Y here.
366 You can find documentation about IPC with "info ipc" and also in
367 section 6.4 of the Linux Programmer's Guide, available from
368 <http://www.tldp.org/guides.html>.
370 config SYSVIPC_SYSCTL
377 bool "POSIX Message Queues"
380 POSIX variant of message queues is a part of IPC. In POSIX message
381 queues every message has a priority which decides about succession
382 of receiving it by a process. If you want to compile and run
383 programs written e.g. for Solaris with use of its POSIX message
384 queues (functions mq_*) say Y here.
386 POSIX message queues are visible as a filesystem called 'mqueue'
387 and can be mounted somewhere if you want to do filesystem
388 operations on message queues.
392 config POSIX_MQUEUE_SYSCTL
394 depends on POSIX_MQUEUE
399 bool "General notification queue"
403 This is a general notification queue for the kernel to pass events to
404 userspace by splicing them into pipes. It can be used in conjunction
405 with watches for key/keyring change notifications and device
408 See Documentation/watch_queue.rst
410 config CROSS_MEMORY_ATTACH
411 bool "Enable process_vm_readv/writev syscalls"
415 Enabling this option adds the system calls process_vm_readv and
416 process_vm_writev which allow a process with the correct privileges
417 to directly read from or write to another process' address space.
418 See the man page for more details.
421 bool "uselib syscall"
422 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
424 This option enables the uselib syscall, a system call used in the
425 dynamic linker from libc5 and earlier. glibc does not use this
426 system call. If you intend to run programs built on libc5 or
427 earlier, you may need to enable this syscall. Current systems
428 running glibc can safely disable this.
431 bool "Auditing support"
434 Enable auditing infrastructure that can be used with another
435 kernel subsystem, such as SELinux (which requires this for
436 logging of avc messages output). System call auditing is included
437 on architectures which support it.
439 config HAVE_ARCH_AUDITSYSCALL
444 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
447 source "kernel/irq/Kconfig"
448 source "kernel/time/Kconfig"
449 source "kernel/bpf/Kconfig"
450 source "kernel/Kconfig.preempt"
452 menu "CPU/Task time and stats accounting"
454 config VIRT_CPU_ACCOUNTING
458 prompt "Cputime accounting"
459 default TICK_CPU_ACCOUNTING if !PPC64
460 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
462 # Kind of a stub config for the pure tick based cputime accounting
463 config TICK_CPU_ACCOUNTING
464 bool "Simple tick based cputime accounting"
465 depends on !S390 && !NO_HZ_FULL
467 This is the basic tick based cputime accounting that maintains
468 statistics about user, system and idle time spent on per jiffies
473 config VIRT_CPU_ACCOUNTING_NATIVE
474 bool "Deterministic task and CPU time accounting"
475 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
476 select VIRT_CPU_ACCOUNTING
478 Select this option to enable more accurate task and CPU time
479 accounting. This is done by reading a CPU counter on each
480 kernel entry and exit and on transitions within the kernel
481 between system, softirq and hardirq state, so there is a
482 small performance impact. In the case of s390 or IBM POWER > 5,
483 this also enables accounting of stolen time on logically-partitioned
486 config VIRT_CPU_ACCOUNTING_GEN
487 bool "Full dynticks CPU time accounting"
488 depends on HAVE_CONTEXT_TRACKING
489 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
490 depends on GENERIC_CLOCKEVENTS
491 select VIRT_CPU_ACCOUNTING
492 select CONTEXT_TRACKING
494 Select this option to enable task and CPU time accounting on full
495 dynticks systems. This accounting is implemented by watching every
496 kernel-user boundaries using the context tracking subsystem.
497 The accounting is thus performed at the expense of some significant
500 For now this is only useful if you are working on the full
501 dynticks subsystem development.
507 config IRQ_TIME_ACCOUNTING
508 bool "Fine granularity task level IRQ time accounting"
509 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
511 Select this option to enable fine granularity task irq time
512 accounting. This is done by reading a timestamp on each
513 transitions between softirq and hardirq state, so there can be a
514 small performance impact.
516 If in doubt, say N here.
518 config HAVE_SCHED_AVG_IRQ
520 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
523 config SCHED_THERMAL_PRESSURE
525 default y if ARM && ARM_CPU_TOPOLOGY
528 depends on CPU_FREQ_THERMAL
530 Select this option to enable thermal pressure accounting in the
531 scheduler. Thermal pressure is the value conveyed to the scheduler
532 that reflects the reduction in CPU compute capacity resulted from
533 thermal throttling. Thermal throttling occurs when the performance of
534 a CPU is capped due to high operating temperatures.
536 If selected, the scheduler will be able to balance tasks accordingly,
537 i.e. put less load on throttled CPUs than on non/less throttled ones.
539 This requires the architecture to implement
540 arch_update_thermal_pressure() and arch_scale_thermal_pressure().
542 config BSD_PROCESS_ACCT
543 bool "BSD Process Accounting"
546 If you say Y here, a user level program will be able to instruct the
547 kernel (via a special system call) to write process accounting
548 information to a file: whenever a process exits, information about
549 that process will be appended to the file by the kernel. The
550 information includes things such as creation time, owning user,
551 command name, memory usage, controlling terminal etc. (the complete
552 list is in the struct acct in <file:include/linux/acct.h>). It is
553 up to the user level program to do useful things with this
554 information. This is generally a good idea, so say Y.
556 config BSD_PROCESS_ACCT_V3
557 bool "BSD Process Accounting version 3 file format"
558 depends on BSD_PROCESS_ACCT
561 If you say Y here, the process accounting information is written
562 in a new file format that also logs the process IDs of each
563 process and its parent. Note that this file format is incompatible
564 with previous v0/v1/v2 file formats, so you will need updated tools
565 for processing it. A preliminary version of these tools is available
566 at <http://www.gnu.org/software/acct/>.
569 bool "Export task/process statistics through netlink"
574 Export selected statistics for tasks/processes through the
575 generic netlink interface. Unlike BSD process accounting, the
576 statistics are available during the lifetime of tasks/processes as
577 responses to commands. Like BSD accounting, they are sent to user
582 config TASK_DELAY_ACCT
583 bool "Enable per-task delay accounting"
587 Collect information on time spent by a task waiting for system
588 resources like cpu, synchronous block I/O completion and swapping
589 in pages. Such statistics can help in setting a task's priorities
590 relative to other tasks for cpu, io, rss limits etc.
595 bool "Enable extended accounting over taskstats"
598 Collect extended task accounting data and send the data
599 to userland for processing over the taskstats interface.
603 config TASK_IO_ACCOUNTING
604 bool "Enable per-task storage I/O accounting"
605 depends on TASK_XACCT
607 Collect information on the number of bytes of storage I/O which this
613 bool "Pressure stall information tracking"
615 Collect metrics that indicate how overcommitted the CPU, memory,
616 and IO capacity are in the system.
618 If you say Y here, the kernel will create /proc/pressure/ with the
619 pressure statistics files cpu, memory, and io. These will indicate
620 the share of walltime in which some or all tasks in the system are
621 delayed due to contention of the respective resource.
623 In kernels with cgroup support, cgroups (cgroup2 only) will
624 have cpu.pressure, memory.pressure, and io.pressure files,
625 which aggregate pressure stalls for the grouped tasks only.
627 For more details see Documentation/accounting/psi.rst.
631 config PSI_DEFAULT_DISABLED
632 bool "Require boot parameter to enable pressure stall information tracking"
636 If set, pressure stall information tracking will be disabled
637 per default but can be enabled through passing psi=1 on the
638 kernel commandline during boot.
640 This feature adds some code to the task wakeup and sleep
641 paths of the scheduler. The overhead is too low to affect
642 common scheduling-intense workloads in practice (such as
643 webservers, memcache), but it does show up in artificial
644 scheduler stress tests, such as hackbench.
646 If you are paranoid and not sure what the kernel will be
651 endmenu # "CPU/Task time and stats accounting"
655 depends on SMP || COMPILE_TEST
658 Make sure that CPUs running critical tasks are not disturbed by
659 any source of "noise" such as unbound workqueues, timers, kthreads...
660 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
661 the "isolcpus=" boot parameter.
665 source "kernel/rcu/Kconfig"
672 tristate "Kernel .config support"
674 This option enables the complete Linux kernel ".config" file
675 contents to be saved in the kernel. It provides documentation
676 of which kernel options are used in a running kernel or in an
677 on-disk kernel. This information can be extracted from the kernel
678 image file with the script scripts/extract-ikconfig and used as
679 input to rebuild the current kernel or to build another kernel.
680 It can also be extracted from a running kernel by reading
681 /proc/config.gz if enabled (below).
684 bool "Enable access to .config through /proc/config.gz"
685 depends on IKCONFIG && PROC_FS
687 This option enables access to the kernel configuration file
688 through /proc/config.gz.
691 tristate "Enable kernel headers through /sys/kernel/kheaders.tar.xz"
694 This option enables access to the in-kernel headers that are generated during
695 the build process. These can be used to build eBPF tracing programs,
696 or similar programs. If you build the headers as a module, a module called
697 kheaders.ko is built which can be loaded on-demand to get access to headers.
700 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
701 range 12 25 if !H8300
706 Select the minimal kernel log buffer size as a power of 2.
707 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
708 parameter, see below. Any higher size also might be forced
709 by "log_buf_len" boot parameter.
719 config LOG_CPU_MAX_BUF_SHIFT
720 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
723 default 12 if !BASE_SMALL
724 default 0 if BASE_SMALL
727 This option allows to increase the default ring buffer size
728 according to the number of CPUs. The value defines the contribution
729 of each CPU as a power of 2. The used space is typically only few
730 lines however it might be much more when problems are reported,
733 The increased size means that a new buffer has to be allocated and
734 the original static one is unused. It makes sense only on systems
735 with more CPUs. Therefore this value is used only when the sum of
736 contributions is greater than the half of the default kernel ring
737 buffer as defined by LOG_BUF_SHIFT. The default values are set
738 so that more than 16 CPUs are needed to trigger the allocation.
740 Also this option is ignored when "log_buf_len" kernel parameter is
741 used as it forces an exact (power of two) size of the ring buffer.
743 The number of possible CPUs is used for this computation ignoring
744 hotplugging making the computation optimal for the worst case
745 scenario while allowing a simple algorithm to be used from bootup.
747 Examples shift values and their meaning:
748 17 => 128 KB for each CPU
749 16 => 64 KB for each CPU
750 15 => 32 KB for each CPU
751 14 => 16 KB for each CPU
752 13 => 8 KB for each CPU
753 12 => 4 KB for each CPU
755 config PRINTK_SAFE_LOG_BUF_SHIFT
756 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
761 Select the size of an alternate printk per-CPU buffer where messages
762 printed from usafe contexts are temporary stored. One example would
763 be NMI messages, another one - printk recursion. The messages are
764 copied to the main log buffer in a safe context to avoid a deadlock.
765 The value defines the size as a power of 2.
767 Those messages are rare and limited. The largest one is when
768 a backtrace is printed. It usually fits into 4KB. Select
769 8KB if you want to be on the safe side.
772 17 => 128 KB for each CPU
773 16 => 64 KB for each CPU
774 15 => 32 KB for each CPU
775 14 => 16 KB for each CPU
776 13 => 8 KB for each CPU
777 12 => 4 KB for each CPU
780 bool "Printk indexing debugfs interface"
781 depends on PRINTK && DEBUG_FS
783 Add support for indexing of all printk formats known at compile time
784 at <debugfs>/printk/index/<module>.
786 This can be used as part of maintaining daemons which monitor
787 /dev/kmsg, as it permits auditing the printk formats present in a
788 kernel, allowing detection of cases where monitored printks are
789 changed or no longer present.
791 There is no additional runtime cost to printk with this enabled.
794 # Architectures with an unreliable sched_clock() should select this:
796 config HAVE_UNSTABLE_SCHED_CLOCK
799 config GENERIC_SCHED_CLOCK
802 menu "Scheduler features"
805 bool "Enable utilization clamping for RT/FAIR tasks"
806 depends on CPU_FREQ_GOV_SCHEDUTIL
808 This feature enables the scheduler to track the clamped utilization
809 of each CPU based on RUNNABLE tasks scheduled on that CPU.
811 With this option, the user can specify the min and max CPU
812 utilization allowed for RUNNABLE tasks. The max utilization defines
813 the maximum frequency a task should use while the min utilization
814 defines the minimum frequency it should use.
816 Both min and max utilization clamp values are hints to the scheduler,
817 aiming at improving its frequency selection policy, but they do not
818 enforce or grant any specific bandwidth for tasks.
822 config UCLAMP_BUCKETS_COUNT
823 int "Number of supported utilization clamp buckets"
826 depends on UCLAMP_TASK
828 Defines the number of clamp buckets to use. The range of each bucket
829 will be SCHED_CAPACITY_SCALE/UCLAMP_BUCKETS_COUNT. The higher the
830 number of clamp buckets the finer their granularity and the higher
831 the precision of clamping aggregation and tracking at run-time.
833 For example, with the minimum configuration value we will have 5
834 clamp buckets tracking 20% utilization each. A 25% boosted tasks will
835 be refcounted in the [20..39]% bucket and will set the bucket clamp
836 effective value to 25%.
837 If a second 30% boosted task should be co-scheduled on the same CPU,
838 that task will be refcounted in the same bucket of the first task and
839 it will boost the bucket clamp effective value to 30%.
840 The clamp effective value of a bucket is reset to its nominal value
841 (20% in the example above) when there are no more tasks refcounted in
844 An additional boost/capping margin can be added to some tasks. In the
845 example above the 25% task will be boosted to 30% until it exits the
846 CPU. If that should be considered not acceptable on certain systems,
847 it's always possible to reduce the margin by increasing the number of
848 clamp buckets to trade off used memory for run-time tracking
851 If in doubt, use the default value.
856 # For architectures that want to enable the support for NUMA-affine scheduler
859 config ARCH_SUPPORTS_NUMA_BALANCING
863 # For architectures that prefer to flush all TLBs after a number of pages
864 # are unmapped instead of sending one IPI per page to flush. The architecture
865 # must provide guarantees on what happens if a clean TLB cache entry is
866 # written after the unmap. Details are in mm/rmap.c near the check for
867 # should_defer_flush. The architecture should also consider if the full flush
868 # and the refill costs are offset by the savings of sending fewer IPIs.
869 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
873 def_bool !$(cc-option,$(m64-flag) -D__SIZEOF_INT128__=0) && 64BIT
875 config CC_IMPLICIT_FALLTHROUGH
877 default "-Wimplicit-fallthrough=5" if CC_IS_GCC && $(cc-option,-Wimplicit-fallthrough=5)
878 default "-Wimplicit-fallthrough" if CC_IS_CLANG && $(cc-option,-Wunreachable-code-fallthrough)
881 # For architectures that know their GCC __int128 support is sound
883 config ARCH_SUPPORTS_INT128
886 # For architectures that (ab)use NUMA to represent different memory regions
887 # all cpu-local but of different latencies, such as SuperH.
889 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
892 config NUMA_BALANCING
893 bool "Memory placement aware NUMA scheduler"
894 depends on ARCH_SUPPORTS_NUMA_BALANCING
895 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
896 depends on SMP && NUMA && MIGRATION && !PREEMPT_RT
898 This option adds support for automatic NUMA aware memory/task placement.
899 The mechanism is quite primitive and is based on migrating memory when
900 it has references to the node the task is running on.
902 This system will be inactive on UMA systems.
904 config NUMA_BALANCING_DEFAULT_ENABLED
905 bool "Automatically enable NUMA aware memory/task placement"
907 depends on NUMA_BALANCING
909 If set, automatic NUMA balancing will be enabled if running on a NUMA
913 bool "Control Group support"
916 This option adds support for grouping sets of processes together, for
917 use with process control subsystems such as Cpusets, CFS, memory
918 controls or device isolation.
920 - Documentation/scheduler/sched-design-CFS.rst (CFS)
921 - Documentation/admin-guide/cgroup-v1/ (features for grouping, isolation
922 and resource control)
932 bool "Memory controller"
936 Provides control over the memory footprint of tasks in a cgroup.
940 depends on MEMCG && SWAP
945 depends on MEMCG && !SLOB
953 Generic block IO controller cgroup interface. This is the common
954 cgroup interface which should be used by various IO controlling
957 Currently, CFQ IO scheduler uses it to recognize task groups and
958 control disk bandwidth allocation (proportional time slice allocation)
959 to such task groups. It is also used by bio throttling logic in
960 block layer to implement upper limit in IO rates on a device.
962 This option only enables generic Block IO controller infrastructure.
963 One needs to also enable actual IO controlling logic/policy. For
964 enabling proportional weight division of disk bandwidth in CFQ, set
965 CONFIG_BFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
966 CONFIG_BLK_DEV_THROTTLING=y.
968 See Documentation/admin-guide/cgroup-v1/blkio-controller.rst for more information.
970 config CGROUP_WRITEBACK
972 depends on MEMCG && BLK_CGROUP
975 menuconfig CGROUP_SCHED
976 bool "CPU controller"
979 This feature lets CPU scheduler recognize task groups and control CPU
980 bandwidth allocation to such task groups. It uses cgroups to group
984 config FAIR_GROUP_SCHED
985 bool "Group scheduling for SCHED_OTHER"
986 depends on CGROUP_SCHED
990 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
991 depends on FAIR_GROUP_SCHED
994 This option allows users to define CPU bandwidth rates (limits) for
995 tasks running within the fair group scheduler. Groups with no limit
996 set are considered to be unconstrained and will run with no
998 See Documentation/scheduler/sched-bwc.rst for more information.
1000 config RT_GROUP_SCHED
1001 bool "Group scheduling for SCHED_RR/FIFO"
1002 depends on CGROUP_SCHED
1005 This feature lets you explicitly allocate real CPU bandwidth
1006 to task groups. If enabled, it will also make it impossible to
1007 schedule realtime tasks for non-root users until you allocate
1008 realtime bandwidth for them.
1009 See Documentation/scheduler/sched-rt-group.rst for more information.
1013 config UCLAMP_TASK_GROUP
1014 bool "Utilization clamping per group of tasks"
1015 depends on CGROUP_SCHED
1016 depends on UCLAMP_TASK
1019 This feature enables the scheduler to track the clamped utilization
1020 of each CPU based on RUNNABLE tasks currently scheduled on that CPU.
1022 When this option is enabled, the user can specify a min and max
1023 CPU bandwidth which is allowed for each single task in a group.
1024 The max bandwidth allows to clamp the maximum frequency a task
1025 can use, while the min bandwidth allows to define a minimum
1026 frequency a task will always use.
1028 When task group based utilization clamping is enabled, an eventually
1029 specified task-specific clamp value is constrained by the cgroup
1030 specified clamp value. Both minimum and maximum task clamping cannot
1031 be bigger than the corresponding clamping defined at task group level.
1036 bool "PIDs controller"
1038 Provides enforcement of process number limits in the scope of a
1039 cgroup. Any attempt to fork more processes than is allowed in the
1040 cgroup will fail. PIDs are fundamentally a global resource because it
1041 is fairly trivial to reach PID exhaustion before you reach even a
1042 conservative kmemcg limit. As a result, it is possible to grind a
1043 system to halt without being limited by other cgroup policies. The
1044 PIDs controller is designed to stop this from happening.
1046 It should be noted that organisational operations (such as attaching
1047 to a cgroup hierarchy) will *not* be blocked by the PIDs controller,
1048 since the PIDs limit only affects a process's ability to fork, not to
1052 bool "RDMA controller"
1054 Provides enforcement of RDMA resources defined by IB stack.
1055 It is fairly easy for consumers to exhaust RDMA resources, which
1056 can result into resource unavailability to other consumers.
1057 RDMA controller is designed to stop this from happening.
1058 Attaching processes with active RDMA resources to the cgroup
1059 hierarchy is allowed even if can cross the hierarchy's limit.
1061 config CGROUP_FREEZER
1062 bool "Freezer controller"
1064 Provides a way to freeze and unfreeze all tasks in a
1067 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
1068 controller includes important in-kernel memory consumers per default.
1070 If you're using cgroup2, say N.
1072 config CGROUP_HUGETLB
1073 bool "HugeTLB controller"
1074 depends on HUGETLB_PAGE
1078 Provides a cgroup controller for HugeTLB pages.
1079 When you enable this, you can put a per cgroup limit on HugeTLB usage.
1080 The limit is enforced during page fault. Since HugeTLB doesn't
1081 support page reclaim, enforcing the limit at page fault time implies
1082 that, the application will get SIGBUS signal if it tries to access
1083 HugeTLB pages beyond its limit. This requires the application to know
1084 beforehand how much HugeTLB pages it would require for its use. The
1085 control group is tracked in the third page lru pointer. This means
1086 that we cannot use the controller with huge page less than 3 pages.
1089 bool "Cpuset controller"
1092 This option will let you create and manage CPUSETs which
1093 allow dynamically partitioning a system into sets of CPUs and
1094 Memory Nodes and assigning tasks to run only within those sets.
1095 This is primarily useful on large SMP or NUMA systems.
1099 config PROC_PID_CPUSET
1100 bool "Include legacy /proc/<pid>/cpuset file"
1104 config CGROUP_DEVICE
1105 bool "Device controller"
1107 Provides a cgroup controller implementing whitelists for
1108 devices which a process in the cgroup can mknod or open.
1110 config CGROUP_CPUACCT
1111 bool "Simple CPU accounting controller"
1113 Provides a simple controller for monitoring the
1114 total CPU consumed by the tasks in a cgroup.
1117 bool "Perf controller"
1118 depends on PERF_EVENTS
1120 This option extends the perf per-cpu mode to restrict monitoring
1121 to threads which belong to the cgroup specified and run on the
1122 designated cpu. Or this can be used to have cgroup ID in samples
1123 so that it can monitor performance events among cgroups.
1128 bool "Support for eBPF programs attached to cgroups"
1129 depends on BPF_SYSCALL
1130 select SOCK_CGROUP_DATA
1132 Allow attaching eBPF programs to a cgroup using the bpf(2)
1133 syscall command BPF_PROG_ATTACH.
1135 In which context these programs are accessed depends on the type
1136 of attachment. For instance, programs that are attached using
1137 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
1141 bool "Misc resource controller"
1144 Provides a controller for miscellaneous resources on a host.
1146 Miscellaneous scalar resources are the resources on the host system
1147 which cannot be abstracted like the other cgroups. This controller
1148 tracks and limits the miscellaneous resources used by a process
1149 attached to a cgroup hierarchy.
1151 For more information, please check misc cgroup section in
1152 /Documentation/admin-guide/cgroup-v2.rst.
1155 bool "Debug controller"
1157 depends on DEBUG_KERNEL
1159 This option enables a simple controller that exports
1160 debugging information about the cgroups framework. This
1161 controller is for control cgroup debugging only. Its
1162 interfaces are not stable.
1166 config SOCK_CGROUP_DATA
1172 menuconfig NAMESPACES
1173 bool "Namespaces support" if EXPERT
1174 depends on MULTIUSER
1177 Provides the way to make tasks work with different objects using
1178 the same id. For example same IPC id may refer to different objects
1179 or same user id or pid may refer to different tasks when used in
1180 different namespaces.
1185 bool "UTS namespace"
1188 In this namespace tasks see different info provided with the
1192 bool "TIME namespace"
1193 depends on GENERIC_VDSO_TIME_NS
1196 In this namespace boottime and monotonic clocks can be set.
1197 The time will keep going with the same pace.
1200 bool "IPC namespace"
1201 depends on (SYSVIPC || POSIX_MQUEUE)
1204 In this namespace tasks work with IPC ids which correspond to
1205 different IPC objects in different namespaces.
1208 bool "User namespace"
1211 This allows containers, i.e. vservers, to use user namespaces
1212 to provide different user info for different servers.
1214 When user namespaces are enabled in the kernel it is
1215 recommended that the MEMCG option also be enabled and that
1216 user-space use the memory control groups to limit the amount
1217 of memory a memory unprivileged users can use.
1222 bool "PID Namespaces"
1225 Support process id namespaces. This allows having multiple
1226 processes with the same pid as long as they are in different
1227 pid namespaces. This is a building block of containers.
1230 bool "Network namespace"
1234 Allow user space to create what appear to be multiple instances
1235 of the network stack.
1239 config CHECKPOINT_RESTORE
1240 bool "Checkpoint/restore support"
1241 select PROC_CHILDREN
1245 Enables additional kernel features in a sake of checkpoint/restore.
1246 In particular it adds auxiliary prctl codes to setup process text,
1247 data and heap segment sizes, and a few additional /proc filesystem
1250 If unsure, say N here.
1252 config SCHED_AUTOGROUP
1253 bool "Automatic process group scheduling"
1256 select FAIR_GROUP_SCHED
1258 This option optimizes the scheduler for common desktop workloads by
1259 automatically creating and populating task groups. This separation
1260 of workloads isolates aggressive CPU burners (like build jobs) from
1261 desktop applications. Task group autogeneration is currently based
1264 config SYSFS_DEPRECATED
1265 bool "Enable deprecated sysfs features to support old userspace tools"
1269 This option adds code that switches the layout of the "block" class
1270 devices, to not show up in /sys/class/block/, but only in
1273 This switch is only active when the sysfs.deprecated=1 boot option is
1274 passed or the SYSFS_DEPRECATED_V2 option is set.
1276 This option allows new kernels to run on old distributions and tools,
1277 which might get confused by /sys/class/block/. Since 2007/2008 all
1278 major distributions and tools handle this just fine.
1280 Recent distributions and userspace tools after 2009/2010 depend on
1281 the existence of /sys/class/block/, and will not work with this
1284 Only if you are using a new kernel on an old distribution, you might
1287 config SYSFS_DEPRECATED_V2
1288 bool "Enable deprecated sysfs features by default"
1291 depends on SYSFS_DEPRECATED
1293 Enable deprecated sysfs by default.
1295 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1298 Only if you are using a new kernel on an old distribution, you might
1299 need to say Y here. Even then, odds are you would not need it
1300 enabled, you can always pass the boot option if absolutely necessary.
1303 bool "Kernel->user space relay support (formerly relayfs)"
1306 This option enables support for relay interface support in
1307 certain file systems (such as debugfs).
1308 It is designed to provide an efficient mechanism for tools and
1309 facilities to relay large amounts of data from kernel space to
1314 config BLK_DEV_INITRD
1315 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1317 The initial RAM filesystem is a ramfs which is loaded by the
1318 boot loader (loadlin or lilo) and that is mounted as root
1319 before the normal boot procedure. It is typically used to
1320 load modules needed to mount the "real" root file system,
1321 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1323 If RAM disk support (BLK_DEV_RAM) is also included, this
1324 also enables initial RAM disk (initrd) support and adds
1325 15 Kbytes (more on some other architectures) to the kernel size.
1331 source "usr/Kconfig"
1336 bool "Boot config support"
1337 select BLK_DEV_INITRD
1339 Extra boot config allows system admin to pass a config file as
1340 complemental extension of kernel cmdline when booting.
1341 The boot config file must be attached at the end of initramfs
1342 with checksum, size and magic word.
1343 See <file:Documentation/admin-guide/bootconfig.rst> for details.
1348 prompt "Compiler optimization level"
1349 default CC_OPTIMIZE_FOR_PERFORMANCE
1351 config CC_OPTIMIZE_FOR_PERFORMANCE
1352 bool "Optimize for performance (-O2)"
1354 This is the default optimization level for the kernel, building
1355 with the "-O2" compiler flag for best performance and most
1356 helpful compile-time warnings.
1358 config CC_OPTIMIZE_FOR_PERFORMANCE_O3
1359 bool "Optimize more for performance (-O3)"
1362 Choosing this option will pass "-O3" to your compiler to optimize
1363 the kernel yet more for performance.
1365 config CC_OPTIMIZE_FOR_SIZE
1366 bool "Optimize for size (-Os)"
1368 Choosing this option will pass "-Os" to your compiler resulting
1369 in a smaller kernel.
1373 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1376 This requires that the arch annotates or otherwise protects
1377 its external entry points from being discarded. Linker scripts
1378 must also merge .text.*, .data.*, and .bss.* correctly into
1379 output sections. Care must be taken not to pull in unrelated
1380 sections (e.g., '.text.init'). Typically '.' in section names
1381 is used to distinguish them from label names / C identifiers.
1383 config LD_DEAD_CODE_DATA_ELIMINATION
1384 bool "Dead code and data elimination (EXPERIMENTAL)"
1385 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1387 depends on $(cc-option,-ffunction-sections -fdata-sections)
1388 depends on $(ld-option,--gc-sections)
1390 Enable this if you want to do dead code and data elimination with
1391 the linker by compiling with -ffunction-sections -fdata-sections,
1392 and linking with --gc-sections.
1394 This can reduce on disk and in-memory size of the kernel
1395 code and static data, particularly for small configs and
1396 on small systems. This has the possibility of introducing
1397 silently broken kernel if the required annotations are not
1398 present. This option is not well tested yet, so use at your
1401 config LD_ORPHAN_WARN
1403 depends on ARCH_WANT_LD_ORPHAN_WARN
1404 depends on $(ld-option,--orphan-handling=warn)
1412 config SYSCTL_EXCEPTION_TRACE
1415 Enable support for /proc/sys/debug/exception-trace.
1417 config SYSCTL_ARCH_UNALIGN_NO_WARN
1420 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1421 Allows arch to define/use @no_unaligned_warning to possibly warn
1422 about unaligned access emulation going on under the hood.
1424 config SYSCTL_ARCH_UNALIGN_ALLOW
1427 Enable support for /proc/sys/kernel/unaligned-trap
1428 Allows arches to define/use @unaligned_enabled to runtime toggle
1429 the unaligned access emulation.
1430 see arch/parisc/kernel/unaligned.c for reference
1432 config HAVE_PCSPKR_PLATFORM
1435 # interpreter that classic socket filters depend on
1440 bool "Configure standard kernel features (expert users)"
1441 # Unhide debug options, to make the on-by-default options visible
1444 This option allows certain base kernel options and settings
1445 to be disabled or tweaked. This is for specialized
1446 environments which can tolerate a "non-standard" kernel.
1447 Only use this if you really know what you are doing.
1450 bool "Enable 16-bit UID system calls" if EXPERT
1451 depends on HAVE_UID16 && MULTIUSER
1454 This enables the legacy 16-bit UID syscall wrappers.
1457 bool "Multiple users, groups and capabilities support" if EXPERT
1460 This option enables support for non-root users, groups and
1463 If you say N here, all processes will run with UID 0, GID 0, and all
1464 possible capabilities. Saying N here also compiles out support for
1465 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1468 If unsure, say Y here.
1470 config SGETMASK_SYSCALL
1471 bool "sgetmask/ssetmask syscalls support" if EXPERT
1472 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1474 sys_sgetmask and sys_ssetmask are obsolete system calls
1475 no longer supported in libc but still enabled by default in some
1478 If unsure, leave the default option here.
1480 config SYSFS_SYSCALL
1481 bool "Sysfs syscall support" if EXPERT
1484 sys_sysfs is an obsolete system call no longer supported in libc.
1485 Note that disabling this option is more secure but might break
1486 compatibility with some systems.
1488 If unsure say Y here.
1491 bool "open by fhandle syscalls" if EXPERT
1495 If you say Y here, a user level program will be able to map
1496 file names to handle and then later use the handle for
1497 different file system operations. This is useful in implementing
1498 userspace file servers, which now track files using handles instead
1499 of names. The handle would remain the same even if file names
1500 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1504 bool "Posix Clocks & timers" if EXPERT
1507 This includes native support for POSIX timers to the kernel.
1508 Some embedded systems have no use for them and therefore they
1509 can be configured out to reduce the size of the kernel image.
1511 When this option is disabled, the following syscalls won't be
1512 available: timer_create, timer_gettime: timer_getoverrun,
1513 timer_settime, timer_delete, clock_adjtime, getitimer,
1514 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1515 clock_getres and clock_nanosleep syscalls will be limited to
1516 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1522 bool "Enable support for printk" if EXPERT
1525 This option enables normal printk support. Removing it
1526 eliminates most of the message strings from the kernel image
1527 and makes the kernel more or less silent. As this makes it
1528 very difficult to diagnose system problems, saying N here is
1529 strongly discouraged.
1532 bool "BUG() support" if EXPERT
1535 Disabling this option eliminates support for BUG and WARN, reducing
1536 the size of your kernel image and potentially quietly ignoring
1537 numerous fatal conditions. You should only consider disabling this
1538 option for embedded systems with no facilities for reporting errors.
1544 bool "Enable ELF core dumps" if EXPERT
1546 Enable support for generating core dumps. Disabling saves about 4k.
1549 config PCSPKR_PLATFORM
1550 bool "Enable PC-Speaker support" if EXPERT
1551 depends on HAVE_PCSPKR_PLATFORM
1555 This option allows to disable the internal PC-Speaker
1556 support, saving some memory.
1560 bool "Enable full-sized data structures for core" if EXPERT
1562 Disabling this option reduces the size of miscellaneous core
1563 kernel data structures. This saves memory on small machines,
1564 but may reduce performance.
1567 bool "Enable futex support" if EXPERT
1568 depends on !(SPARC32 && SMP)
1572 Disabling this option will cause the kernel to be built without
1573 support for "fast userspace mutexes". The resulting kernel may not
1574 run glibc-based applications correctly.
1578 depends on FUTEX && RT_MUTEXES
1582 bool "Enable eventpoll support" if EXPERT
1585 Disabling this option will cause the kernel to be built without
1586 support for epoll family of system calls.
1589 bool "Enable signalfd() system call" if EXPERT
1592 Enable the signalfd() system call that allows to receive signals
1593 on a file descriptor.
1598 bool "Enable timerfd() system call" if EXPERT
1601 Enable the timerfd() system call that allows to receive timer
1602 events on a file descriptor.
1607 bool "Enable eventfd() system call" if EXPERT
1610 Enable the eventfd() system call that allows to receive both
1611 kernel notification (ie. KAIO) or userspace notifications.
1616 bool "Use full shmem filesystem" if EXPERT
1620 The shmem is an internal filesystem used to manage shared memory.
1621 It is backed by swap and manages resource limits. It is also exported
1622 to userspace as tmpfs if TMPFS is enabled. Disabling this
1623 option replaces shmem and tmpfs with the much simpler ramfs code,
1624 which may be appropriate on small systems without swap.
1627 bool "Enable AIO support" if EXPERT
1630 This option enables POSIX asynchronous I/O which may by used
1631 by some high performance threaded applications. Disabling
1632 this option saves about 7k.
1635 bool "Enable IO uring support" if EXPERT
1639 This option enables support for the io_uring interface, enabling
1640 applications to submit and complete IO through submission and
1641 completion rings that are shared between the kernel and application.
1643 config ADVISE_SYSCALLS
1644 bool "Enable madvise/fadvise syscalls" if EXPERT
1647 This option enables the madvise and fadvise syscalls, used by
1648 applications to advise the kernel about their future memory or file
1649 usage, improving performance. If building an embedded system where no
1650 applications use these syscalls, you can disable this option to save
1654 bool "Enable membarrier() system call" if EXPERT
1657 Enable the membarrier() system call that allows issuing memory
1658 barriers across all running threads, which can be used to distribute
1659 the cost of user-space memory barriers asymmetrically by transforming
1660 pairs of memory barriers into pairs consisting of membarrier() and a
1666 bool "Load all symbols for debugging/ksymoops" if EXPERT
1669 Say Y here to let the kernel print out symbolic crash information and
1670 symbolic stack backtraces. This increases the size of the kernel
1671 somewhat, as all symbols have to be loaded into the kernel image.
1674 bool "Include all symbols in kallsyms"
1675 depends on DEBUG_KERNEL && KALLSYMS
1677 Normally kallsyms only contains the symbols of functions for nicer
1678 OOPS messages and backtraces (i.e., symbols from the text and inittext
1679 sections). This is sufficient for most cases. And only in very rare
1680 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1681 names of variables from the data sections, etc).
1683 This option makes sure that all symbols are loaded into the kernel
1684 image (i.e., symbols from all sections) in cost of increased kernel
1685 size (depending on the kernel configuration, it may be 300KiB or
1686 something like this).
1688 Say N unless you really need all symbols.
1690 config KALLSYMS_ABSOLUTE_PERCPU
1693 default X86_64 && SMP
1695 config KALLSYMS_BASE_RELATIVE
1700 Instead of emitting them as absolute values in the native word size,
1701 emit the symbol references in the kallsyms table as 32-bit entries,
1702 each containing a relative value in the range [base, base + U32_MAX]
1703 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1704 an absolute value in the range [0, S32_MAX] or a relative value in the
1705 range [base, base + S32_MAX], where base is the lowest relative symbol
1706 address encountered in the image.
1708 On 64-bit builds, this reduces the size of the address table by 50%,
1709 but more importantly, it results in entries whose values are build
1710 time constants, and no relocation pass is required at runtime to fix
1711 up the entries based on the runtime load address of the kernel.
1713 # end of the "standard kernel features (expert users)" menu
1715 # syscall, maps, verifier
1717 config ARCH_HAS_MEMBARRIER_CALLBACKS
1720 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1724 bool "Enable kcmp() system call" if EXPERT
1726 Enable the kernel resource comparison system call. It provides
1727 user-space with the ability to compare two processes to see if they
1728 share a common resource, such as a file descriptor or even virtual
1734 bool "Enable rseq() system call" if EXPERT
1736 depends on HAVE_RSEQ
1739 Enable the restartable sequences system call. It provides a
1740 user-space cache for the current CPU number value, which
1741 speeds up getting the current CPU number from user-space,
1742 as well as an ABI to speed up user-space operations on
1749 bool "Enabled debugging of rseq() system call" if EXPERT
1750 depends on RSEQ && DEBUG_KERNEL
1752 Enable extra debugging checks for the rseq system call.
1757 bool "Embedded system"
1760 This option should be enabled if compiling the kernel for
1761 an embedded system so certain expert options are available
1764 config HAVE_PERF_EVENTS
1767 See tools/perf/design.txt for details.
1769 config GUEST_PERF_EVENTS
1771 depends on HAVE_PERF_EVENTS
1773 config PERF_USE_VMALLOC
1776 See tools/perf/design.txt for details
1779 bool "PC/104 support" if EXPERT
1781 Expose PC/104 form factor device drivers and options available for
1782 selection and configuration. Enable this option if your target
1783 machine has a PC/104 bus.
1785 menu "Kernel Performance Events And Counters"
1788 bool "Kernel performance events and counters"
1789 default y if PROFILING
1790 depends on HAVE_PERF_EVENTS
1794 Enable kernel support for various performance events provided
1795 by software and hardware.
1797 Software events are supported either built-in or via the
1798 use of generic tracepoints.
1800 Most modern CPUs support performance events via performance
1801 counter registers. These registers count the number of certain
1802 types of hw events: such as instructions executed, cachemisses
1803 suffered, or branches mis-predicted - without slowing down the
1804 kernel or applications. These registers can also trigger interrupts
1805 when a threshold number of events have passed - and can thus be
1806 used to profile the code that runs on that CPU.
1808 The Linux Performance Event subsystem provides an abstraction of
1809 these software and hardware event capabilities, available via a
1810 system call and used by the "perf" utility in tools/perf/. It
1811 provides per task and per CPU counters, and it provides event
1812 capabilities on top of those.
1816 config DEBUG_PERF_USE_VMALLOC
1818 bool "Debug: use vmalloc to back perf mmap() buffers"
1819 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1820 select PERF_USE_VMALLOC
1822 Use vmalloc memory to back perf mmap() buffers.
1824 Mostly useful for debugging the vmalloc code on platforms
1825 that don't require it.
1831 config SYSTEM_DATA_VERIFICATION
1833 select SYSTEM_TRUSTED_KEYRING
1837 select ASYMMETRIC_KEY_TYPE
1838 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1841 select X509_CERTIFICATE_PARSER
1842 select PKCS7_MESSAGE_PARSER
1844 Provide PKCS#7 message verification using the contents of the system
1845 trusted keyring to provide public keys. This then can be used for
1846 module verification, kexec image verification and firmware blob
1850 bool "Profiling support"
1852 Say Y here to enable the extended profiling support mechanisms used
1856 # Place an empty function call at each tracepoint site. Can be
1857 # dynamically changed for a probe function.
1862 endmenu # General setup
1864 source "arch/Kconfig"
1868 default y if PREEMPT_RT
1872 default 0 if BASE_FULL
1873 default 1 if !BASE_FULL
1875 config MODULE_SIG_FORMAT
1877 select SYSTEM_DATA_VERIFICATION
1880 bool "Enable loadable module support"
1883 Kernel modules are small pieces of compiled code which can
1884 be inserted in the running kernel, rather than being
1885 permanently built into the kernel. You use the "modprobe"
1886 tool to add (and sometimes remove) them. If you say Y here,
1887 many parts of the kernel can be built as modules (by
1888 answering M instead of Y where indicated): this is most
1889 useful for infrequently used options which are not required
1890 for booting. For more information, see the man pages for
1891 modprobe, lsmod, modinfo, insmod and rmmod.
1893 If you say Y here, you will need to run "make
1894 modules_install" to put the modules under /lib/modules/
1895 where modprobe can find them (you may need to be root to do
1902 config MODULE_FORCE_LOAD
1903 bool "Forced module loading"
1906 Allow loading of modules without version information (ie. modprobe
1907 --force). Forced module loading sets the 'F' (forced) taint flag and
1908 is usually a really bad idea.
1910 config MODULE_UNLOAD
1911 bool "Module unloading"
1913 Without this option you will not be able to unload any
1914 modules (note that some modules may not be unloadable
1915 anyway), which makes your kernel smaller, faster
1916 and simpler. If unsure, say Y.
1918 config MODULE_FORCE_UNLOAD
1919 bool "Forced module unloading"
1920 depends on MODULE_UNLOAD
1922 This option allows you to force a module to unload, even if the
1923 kernel believes it is unsafe: the kernel will remove the module
1924 without waiting for anyone to stop using it (using the -f option to
1925 rmmod). This is mainly for kernel developers and desperate users.
1929 bool "Module versioning support"
1931 Usually, you have to use modules compiled with your kernel.
1932 Saying Y here makes it sometimes possible to use modules
1933 compiled for different kernels, by adding enough information
1934 to the modules to (hopefully) spot any changes which would
1935 make them incompatible with the kernel you are running. If
1938 config ASM_MODVERSIONS
1940 default HAVE_ASM_MODVERSIONS && MODVERSIONS
1942 This enables module versioning for exported symbols also from
1943 assembly. This can be enabled only when the target architecture
1946 config MODULE_REL_CRCS
1948 depends on MODVERSIONS
1950 config MODULE_SRCVERSION_ALL
1951 bool "Source checksum for all modules"
1953 Modules which contain a MODULE_VERSION get an extra "srcversion"
1954 field inserted into their modinfo section, which contains a
1955 sum of the source files which made it. This helps maintainers
1956 see exactly which source was used to build a module (since
1957 others sometimes change the module source without updating
1958 the version). With this option, such a "srcversion" field
1959 will be created for all modules. If unsure, say N.
1962 bool "Module signature verification"
1963 select MODULE_SIG_FORMAT
1965 Check modules for valid signatures upon load: the signature
1966 is simply appended to the module. For more information see
1967 <file:Documentation/admin-guide/module-signing.rst>.
1969 Note that this option adds the OpenSSL development packages as a
1970 kernel build dependency so that the signing tool can use its crypto
1973 You should enable this option if you wish to use either
1974 CONFIG_SECURITY_LOCKDOWN_LSM or lockdown functionality imposed via
1975 another LSM - otherwise unsigned modules will be loadable regardless
1976 of the lockdown policy.
1978 !!!WARNING!!! If you enable this option, you MUST make sure that the
1979 module DOES NOT get stripped after being signed. This includes the
1980 debuginfo strip done by some packagers (such as rpmbuild) and
1981 inclusion into an initramfs that wants the module size reduced.
1983 config MODULE_SIG_FORCE
1984 bool "Require modules to be validly signed"
1985 depends on MODULE_SIG
1987 Reject unsigned modules or signed modules for which we don't have a
1988 key. Without this, such modules will simply taint the kernel.
1990 config MODULE_SIG_ALL
1991 bool "Automatically sign all modules"
1993 depends on MODULE_SIG || IMA_APPRAISE_MODSIG
1995 Sign all modules during make modules_install. Without this option,
1996 modules must be signed manually, using the scripts/sign-file tool.
1998 comment "Do not forget to sign required modules with scripts/sign-file"
1999 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
2002 prompt "Which hash algorithm should modules be signed with?"
2003 depends on MODULE_SIG || IMA_APPRAISE_MODSIG
2005 This determines which sort of hashing algorithm will be used during
2006 signature generation. This algorithm _must_ be built into the kernel
2007 directly so that signature verification can take place. It is not
2008 possible to load a signed module containing the algorithm to check
2009 the signature on that module.
2011 config MODULE_SIG_SHA1
2012 bool "Sign modules with SHA-1"
2015 config MODULE_SIG_SHA224
2016 bool "Sign modules with SHA-224"
2017 select CRYPTO_SHA256
2019 config MODULE_SIG_SHA256
2020 bool "Sign modules with SHA-256"
2021 select CRYPTO_SHA256
2023 config MODULE_SIG_SHA384
2024 bool "Sign modules with SHA-384"
2025 select CRYPTO_SHA512
2027 config MODULE_SIG_SHA512
2028 bool "Sign modules with SHA-512"
2029 select CRYPTO_SHA512
2033 config MODULE_SIG_HASH
2035 depends on MODULE_SIG || IMA_APPRAISE_MODSIG
2036 default "sha1" if MODULE_SIG_SHA1
2037 default "sha224" if MODULE_SIG_SHA224
2038 default "sha256" if MODULE_SIG_SHA256
2039 default "sha384" if MODULE_SIG_SHA384
2040 default "sha512" if MODULE_SIG_SHA512
2043 prompt "Module compression mode"
2045 This option allows you to choose the algorithm which will be used to
2046 compress modules when 'make modules_install' is run. (or, you can
2047 choose to not compress modules at all.)
2049 External modules will also be compressed in the same way during the
2052 For modules inside an initrd or initramfs, it's more efficient to
2053 compress the whole initrd or initramfs instead.
2055 This is fully compatible with signed modules.
2057 Please note that the tool used to load modules needs to support the
2058 corresponding algorithm. module-init-tools MAY support gzip, and kmod
2059 MAY support gzip, xz and zstd.
2061 Your build system needs to provide the appropriate compression tool
2062 to compress the modules.
2064 If in doubt, select 'None'.
2066 config MODULE_COMPRESS_NONE
2069 Do not compress modules. The installed modules are suffixed
2072 config MODULE_COMPRESS_GZIP
2075 Compress modules with GZIP. The installed modules are suffixed
2078 config MODULE_COMPRESS_XZ
2081 Compress modules with XZ. The installed modules are suffixed
2084 config MODULE_COMPRESS_ZSTD
2087 Compress modules with ZSTD. The installed modules are suffixed
2092 config MODULE_DECOMPRESS
2093 bool "Support in-kernel module decompression"
2094 depends on MODULE_COMPRESS_GZIP || MODULE_COMPRESS_XZ
2095 select ZLIB_INFLATE if MODULE_COMPRESS_GZIP
2096 select XZ_DEC if MODULE_COMPRESS_XZ
2099 Support for decompressing kernel modules by the kernel itself
2100 instead of relying on userspace to perform this task. Useful when
2101 load pinning security policy is enabled.
2105 config MODULE_ALLOW_MISSING_NAMESPACE_IMPORTS
2106 bool "Allow loading of modules with missing namespace imports"
2108 Symbols exported with EXPORT_SYMBOL_NS*() are considered exported in
2109 a namespace. A module that makes use of a symbol exported with such a
2110 namespace is required to import the namespace via MODULE_IMPORT_NS().
2111 There is no technical reason to enforce correct namespace imports,
2112 but it creates consistency between symbols defining namespaces and
2113 users importing namespaces they make use of. This option relaxes this
2114 requirement and lifts the enforcement when loading a module.
2118 config MODPROBE_PATH
2119 string "Path to modprobe binary"
2120 default "/sbin/modprobe"
2122 When kernel code requests a module, it does so by calling
2123 the "modprobe" userspace utility. This option allows you to
2124 set the path where that binary is found. This can be changed
2125 at runtime via the sysctl file
2126 /proc/sys/kernel/modprobe. Setting this to the empty string
2127 removes the kernel's ability to request modules (but
2128 userspace can still load modules explicitly).
2130 config TRIM_UNUSED_KSYMS
2131 bool "Trim unused exported kernel symbols" if EXPERT
2132 depends on !COMPILE_TEST
2134 The kernel and some modules make many symbols available for
2135 other modules to use via EXPORT_SYMBOL() and variants. Depending
2136 on the set of modules being selected in your kernel configuration,
2137 many of those exported symbols might never be used.
2139 This option allows for unused exported symbols to be dropped from
2140 the build. In turn, this provides the compiler more opportunities
2141 (especially when using LTO) for optimizing the code and reducing
2142 binary size. This might have some security advantages as well.
2144 If unsure, or if you need to build out-of-tree modules, say N.
2146 config UNUSED_KSYMS_WHITELIST
2147 string "Whitelist of symbols to keep in ksymtab"
2148 depends on TRIM_UNUSED_KSYMS
2150 By default, all unused exported symbols will be un-exported from the
2151 build when TRIM_UNUSED_KSYMS is selected.
2153 UNUSED_KSYMS_WHITELIST allows to whitelist symbols that must be kept
2154 exported at all times, even in absence of in-tree users. The value to
2155 set here is the path to a text file containing the list of symbols,
2156 one per line. The path can be absolute, or relative to the kernel
2161 config MODULES_TREE_LOOKUP
2163 depends on PERF_EVENTS || TRACING || CFI_CLANG
2165 config INIT_ALL_POSSIBLE
2168 Back when each arch used to define their own cpu_online_mask and
2169 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2170 with all 1s, and others with all 0s. When they were centralised,
2171 it was better to provide this option than to break all the archs
2172 and have several arch maintainers pursuing me down dark alleys.
2174 source "block/Kconfig"
2176 config PREEMPT_NOTIFIERS
2186 Build a simple ASN.1 grammar compiler that produces a bytecode output
2187 that can be interpreted by the ASN.1 stream decoder and used to
2188 inform it as to what tags are to be expected in a stream and what
2189 functions to call on what tags.
2191 source "kernel/Kconfig.locks"
2193 config ARCH_HAS_NON_OVERLAPPING_ADDRESS_SPACE
2196 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2199 # It may be useful for an architecture to override the definitions of the
2200 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2201 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2202 # different calling convention for syscalls. They can also override the
2203 # macros for not-implemented syscalls in kernel/sys_ni.c and
2204 # kernel/time/posix-stubs.c. All these overrides need to be available in
2205 # <asm/syscall_wrapper.h>.
2206 config ARCH_HAS_SYSCALL_WRAPPER