5 default "/lib/modules/$(shell,uname -r)/.config"
6 default "/etc/kernel-config"
7 default "/boot/config-$(shell,uname -r)"
9 default "arch/$(ARCH)/defconfig"
12 def_bool $(success,$(CC) --version | head -n 1 | grep -q gcc)
16 default $(shell,$(srctree)/scripts/gcc-version.sh -p $(CC) | sed 's/^0*//') if CC_IS_GCC
20 def_bool $(success,$(CC) --version | head -n 1 | grep -q clang)
24 default $(shell,$(srctree)/scripts/clang-version.sh $(CC))
26 config CC_HAS_ASM_GOTO
27 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
36 config BUILDTIME_EXTABLE_SORT
39 config THREAD_INFO_IN_TASK
42 Select this to move thread_info off the stack into task_struct. To
43 make this work, an arch will need to remove all thread_info fields
44 except flags and fix any runtime bugs.
46 One subtle change that will be needed is to use try_get_task_stack()
47 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
56 depends on BROKEN || !SMP
59 config INIT_ENV_ARG_LIMIT
64 Maximum of each of the number of arguments and environment
65 variables passed to init from the kernel command line.
68 bool "Compile also drivers which will not load"
72 Some drivers can be compiled on a different platform than they are
73 intended to be run on. Despite they cannot be loaded there (or even
74 when they load they cannot be used due to missing HW support),
75 developers still, opposing to distributors, might want to build such
76 drivers to compile-test them.
78 If you are a developer and want to build everything available, say Y
79 here. If you are a user/distributor, say N here to exclude useless
80 drivers to be distributed.
83 string "Local version - append to kernel release"
85 Append an extra string to the end of your kernel version.
86 This will show up when you type uname, for example.
87 The string you set here will be appended after the contents of
88 any files with a filename matching localversion* in your
89 object and source tree, in that order. Your total string can
90 be a maximum of 64 characters.
92 config LOCALVERSION_AUTO
93 bool "Automatically append version information to the version string"
95 depends on !COMPILE_TEST
97 This will try to automatically determine if the current tree is a
98 release tree by looking for git tags that belong to the current
101 A string of the format -gxxxxxxxx will be added to the localversion
102 if a git-based tree is found. The string generated by this will be
103 appended after any matching localversion* files, and after the value
104 set in CONFIG_LOCALVERSION.
106 (The actual string used here is the first eight characters produced
107 by running the command:
109 $ git rev-parse --verify HEAD
111 which is done within the script "scripts/setlocalversion".)
114 string "Build ID Salt"
117 The build ID is used to link binaries and their debug info. Setting
118 this option will use the value in the calculation of the build id.
119 This is mostly useful for distributions which want to ensure the
120 build is unique between builds. It's safe to leave the default.
122 config HAVE_KERNEL_GZIP
125 config HAVE_KERNEL_BZIP2
128 config HAVE_KERNEL_LZMA
131 config HAVE_KERNEL_XZ
134 config HAVE_KERNEL_LZO
137 config HAVE_KERNEL_LZ4
140 config HAVE_KERNEL_UNCOMPRESSED
144 prompt "Kernel compression mode"
146 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
148 The linux kernel is a kind of self-extracting executable.
149 Several compression algorithms are available, which differ
150 in efficiency, compression and decompression speed.
151 Compression speed is only relevant when building a kernel.
152 Decompression speed is relevant at each boot.
154 If you have any problems with bzip2 or lzma compressed
155 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
156 version of this functionality (bzip2 only), for 2.4, was
157 supplied by Christian Ludwig)
159 High compression options are mostly useful for users, who
160 are low on disk space (embedded systems), but for whom ram
163 If in doubt, select 'gzip'
167 depends on HAVE_KERNEL_GZIP
169 The old and tried gzip compression. It provides a good balance
170 between compression ratio and decompression speed.
174 depends on HAVE_KERNEL_BZIP2
176 Its compression ratio and speed is intermediate.
177 Decompression speed is slowest among the choices. The kernel
178 size is about 10% smaller with bzip2, in comparison to gzip.
179 Bzip2 uses a large amount of memory. For modern kernels you
180 will need at least 8MB RAM or more for booting.
184 depends on HAVE_KERNEL_LZMA
186 This compression algorithm's ratio is best. Decompression speed
187 is between gzip and bzip2. Compression is slowest.
188 The kernel size is about 33% smaller with LZMA in comparison to gzip.
192 depends on HAVE_KERNEL_XZ
194 XZ uses the LZMA2 algorithm and instruction set specific
195 BCJ filters which can improve compression ratio of executable
196 code. The size of the kernel is about 30% smaller with XZ in
197 comparison to gzip. On architectures for which there is a BCJ
198 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
199 will create a few percent smaller kernel than plain LZMA.
201 The speed is about the same as with LZMA: The decompression
202 speed of XZ is better than that of bzip2 but worse than gzip
203 and LZO. Compression is slow.
207 depends on HAVE_KERNEL_LZO
209 Its compression ratio is the poorest among the choices. The kernel
210 size is about 10% bigger than gzip; however its speed
211 (both compression and decompression) is the fastest.
215 depends on HAVE_KERNEL_LZ4
217 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
218 A preliminary version of LZ4 de/compression tool is available at
219 <https://code.google.com/p/lz4/>.
221 Its compression ratio is worse than LZO. The size of the kernel
222 is about 8% bigger than LZO. But the decompression speed is
225 config KERNEL_UNCOMPRESSED
227 depends on HAVE_KERNEL_UNCOMPRESSED
229 Produce uncompressed kernel image. This option is usually not what
230 you want. It is useful for debugging the kernel in slow simulation
231 environments, where decompressing and moving the kernel is awfully
232 slow. This option allows early boot code to skip the decompressor
233 and jump right at uncompressed kernel image.
237 config DEFAULT_HOSTNAME
238 string "Default hostname"
241 This option determines the default system hostname before userspace
242 calls sethostname(2). The kernel traditionally uses "(none)" here,
243 but you may wish to use a different default here to make a minimal
244 system more usable with less configuration.
247 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
248 # add proper SWAP support to them, in which case this can be remove.
254 bool "Support for paging of anonymous memory (swap)"
255 depends on MMU && BLOCK && !ARCH_NO_SWAP
258 This option allows you to choose whether you want to have support
259 for so called swap devices or swap files in your kernel that are
260 used to provide more virtual memory than the actual RAM present
261 in your computer. If unsure say Y.
266 Inter Process Communication is a suite of library functions and
267 system calls which let processes (running programs) synchronize and
268 exchange information. It is generally considered to be a good thing,
269 and some programs won't run unless you say Y here. In particular, if
270 you want to run the DOS emulator dosemu under Linux (read the
271 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
272 you'll need to say Y here.
274 You can find documentation about IPC with "info ipc" and also in
275 section 6.4 of the Linux Programmer's Guide, available from
276 <http://www.tldp.org/guides.html>.
278 config SYSVIPC_SYSCTL
285 bool "POSIX Message Queues"
288 POSIX variant of message queues is a part of IPC. In POSIX message
289 queues every message has a priority which decides about succession
290 of receiving it by a process. If you want to compile and run
291 programs written e.g. for Solaris with use of its POSIX message
292 queues (functions mq_*) say Y here.
294 POSIX message queues are visible as a filesystem called 'mqueue'
295 and can be mounted somewhere if you want to do filesystem
296 operations on message queues.
300 config POSIX_MQUEUE_SYSCTL
302 depends on POSIX_MQUEUE
306 config CROSS_MEMORY_ATTACH
307 bool "Enable process_vm_readv/writev syscalls"
311 Enabling this option adds the system calls process_vm_readv and
312 process_vm_writev which allow a process with the correct privileges
313 to directly read from or write to another process' address space.
314 See the man page for more details.
317 bool "uselib syscall"
318 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
320 This option enables the uselib syscall, a system call used in the
321 dynamic linker from libc5 and earlier. glibc does not use this
322 system call. If you intend to run programs built on libc5 or
323 earlier, you may need to enable this syscall. Current systems
324 running glibc can safely disable this.
327 bool "Auditing support"
330 Enable auditing infrastructure that can be used with another
331 kernel subsystem, such as SELinux (which requires this for
332 logging of avc messages output). System call auditing is included
333 on architectures which support it.
335 config HAVE_ARCH_AUDITSYSCALL
340 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
343 source "kernel/irq/Kconfig"
344 source "kernel/time/Kconfig"
345 source "kernel/Kconfig.preempt"
347 menu "CPU/Task time and stats accounting"
349 config VIRT_CPU_ACCOUNTING
353 prompt "Cputime accounting"
354 default TICK_CPU_ACCOUNTING if !PPC64
355 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
357 # Kind of a stub config for the pure tick based cputime accounting
358 config TICK_CPU_ACCOUNTING
359 bool "Simple tick based cputime accounting"
360 depends on !S390 && !NO_HZ_FULL
362 This is the basic tick based cputime accounting that maintains
363 statistics about user, system and idle time spent on per jiffies
368 config VIRT_CPU_ACCOUNTING_NATIVE
369 bool "Deterministic task and CPU time accounting"
370 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
371 select VIRT_CPU_ACCOUNTING
373 Select this option to enable more accurate task and CPU time
374 accounting. This is done by reading a CPU counter on each
375 kernel entry and exit and on transitions within the kernel
376 between system, softirq and hardirq state, so there is a
377 small performance impact. In the case of s390 or IBM POWER > 5,
378 this also enables accounting of stolen time on logically-partitioned
381 config VIRT_CPU_ACCOUNTING_GEN
382 bool "Full dynticks CPU time accounting"
383 depends on HAVE_CONTEXT_TRACKING
384 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
385 depends on GENERIC_CLOCKEVENTS
386 select VIRT_CPU_ACCOUNTING
387 select CONTEXT_TRACKING
389 Select this option to enable task and CPU time accounting on full
390 dynticks systems. This accounting is implemented by watching every
391 kernel-user boundaries using the context tracking subsystem.
392 The accounting is thus performed at the expense of some significant
395 For now this is only useful if you are working on the full
396 dynticks subsystem development.
402 config IRQ_TIME_ACCOUNTING
403 bool "Fine granularity task level IRQ time accounting"
404 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
406 Select this option to enable fine granularity task irq time
407 accounting. This is done by reading a timestamp on each
408 transitions between softirq and hardirq state, so there can be a
409 small performance impact.
411 If in doubt, say N here.
413 config HAVE_SCHED_AVG_IRQ
415 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
418 config BSD_PROCESS_ACCT
419 bool "BSD Process Accounting"
422 If you say Y here, a user level program will be able to instruct the
423 kernel (via a special system call) to write process accounting
424 information to a file: whenever a process exits, information about
425 that process will be appended to the file by the kernel. The
426 information includes things such as creation time, owning user,
427 command name, memory usage, controlling terminal etc. (the complete
428 list is in the struct acct in <file:include/linux/acct.h>). It is
429 up to the user level program to do useful things with this
430 information. This is generally a good idea, so say Y.
432 config BSD_PROCESS_ACCT_V3
433 bool "BSD Process Accounting version 3 file format"
434 depends on BSD_PROCESS_ACCT
437 If you say Y here, the process accounting information is written
438 in a new file format that also logs the process IDs of each
439 process and its parent. Note that this file format is incompatible
440 with previous v0/v1/v2 file formats, so you will need updated tools
441 for processing it. A preliminary version of these tools is available
442 at <http://www.gnu.org/software/acct/>.
445 bool "Export task/process statistics through netlink"
450 Export selected statistics for tasks/processes through the
451 generic netlink interface. Unlike BSD process accounting, the
452 statistics are available during the lifetime of tasks/processes as
453 responses to commands. Like BSD accounting, they are sent to user
458 config TASK_DELAY_ACCT
459 bool "Enable per-task delay accounting"
463 Collect information on time spent by a task waiting for system
464 resources like cpu, synchronous block I/O completion and swapping
465 in pages. Such statistics can help in setting a task's priorities
466 relative to other tasks for cpu, io, rss limits etc.
471 bool "Enable extended accounting over taskstats"
474 Collect extended task accounting data and send the data
475 to userland for processing over the taskstats interface.
479 config TASK_IO_ACCOUNTING
480 bool "Enable per-task storage I/O accounting"
481 depends on TASK_XACCT
483 Collect information on the number of bytes of storage I/O which this
489 bool "Pressure stall information tracking"
491 Collect metrics that indicate how overcommitted the CPU, memory,
492 and IO capacity are in the system.
494 If you say Y here, the kernel will create /proc/pressure/ with the
495 pressure statistics files cpu, memory, and io. These will indicate
496 the share of walltime in which some or all tasks in the system are
497 delayed due to contention of the respective resource.
499 In kernels with cgroup support, cgroups (cgroup2 only) will
500 have cpu.pressure, memory.pressure, and io.pressure files,
501 which aggregate pressure stalls for the grouped tasks only.
503 For more details see Documentation/accounting/psi.txt.
507 config PSI_DEFAULT_DISABLED
508 bool "Require boot parameter to enable pressure stall information tracking"
512 If set, pressure stall information tracking will be disabled
513 per default but can be enabled through passing psi=1 on the
514 kernel commandline during boot.
516 This feature adds some code to the task wakeup and sleep
517 paths of the scheduler. The overhead is too low to affect
518 common scheduling-intense workloads in practice (such as
519 webservers, memcache), but it does show up in artificial
520 scheduler stress tests, such as hackbench.
522 If you are paranoid and not sure what the kernel will be
527 endmenu # "CPU/Task time and stats accounting"
531 depends on SMP || COMPILE_TEST
534 Make sure that CPUs running critical tasks are not disturbed by
535 any source of "noise" such as unbound workqueues, timers, kthreads...
536 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
537 the "isolcpus=" boot parameter.
541 source "kernel/rcu/Kconfig"
548 tristate "Kernel .config support"
551 This option enables the complete Linux kernel ".config" file
552 contents to be saved in the kernel. It provides documentation
553 of which kernel options are used in a running kernel or in an
554 on-disk kernel. This information can be extracted from the kernel
555 image file with the script scripts/extract-ikconfig and used as
556 input to rebuild the current kernel or to build another kernel.
557 It can also be extracted from a running kernel by reading
558 /proc/config.gz if enabled (below).
561 bool "Enable access to .config through /proc/config.gz"
562 depends on IKCONFIG && PROC_FS
564 This option enables access to the kernel configuration file
565 through /proc/config.gz.
568 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
573 Select the minimal kernel log buffer size as a power of 2.
574 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
575 parameter, see below. Any higher size also might be forced
576 by "log_buf_len" boot parameter.
586 config LOG_CPU_MAX_BUF_SHIFT
587 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
590 default 12 if !BASE_SMALL
591 default 0 if BASE_SMALL
594 This option allows to increase the default ring buffer size
595 according to the number of CPUs. The value defines the contribution
596 of each CPU as a power of 2. The used space is typically only few
597 lines however it might be much more when problems are reported,
600 The increased size means that a new buffer has to be allocated and
601 the original static one is unused. It makes sense only on systems
602 with more CPUs. Therefore this value is used only when the sum of
603 contributions is greater than the half of the default kernel ring
604 buffer as defined by LOG_BUF_SHIFT. The default values are set
605 so that more than 64 CPUs are needed to trigger the allocation.
607 Also this option is ignored when "log_buf_len" kernel parameter is
608 used as it forces an exact (power of two) size of the ring buffer.
610 The number of possible CPUs is used for this computation ignoring
611 hotplugging making the computation optimal for the worst case
612 scenario while allowing a simple algorithm to be used from bootup.
614 Examples shift values and their meaning:
615 17 => 128 KB for each CPU
616 16 => 64 KB for each CPU
617 15 => 32 KB for each CPU
618 14 => 16 KB for each CPU
619 13 => 8 KB for each CPU
620 12 => 4 KB for each CPU
622 config PRINTK_SAFE_LOG_BUF_SHIFT
623 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
628 Select the size of an alternate printk per-CPU buffer where messages
629 printed from usafe contexts are temporary stored. One example would
630 be NMI messages, another one - printk recursion. The messages are
631 copied to the main log buffer in a safe context to avoid a deadlock.
632 The value defines the size as a power of 2.
634 Those messages are rare and limited. The largest one is when
635 a backtrace is printed. It usually fits into 4KB. Select
636 8KB if you want to be on the safe side.
639 17 => 128 KB for each CPU
640 16 => 64 KB for each CPU
641 15 => 32 KB for each CPU
642 14 => 16 KB for each CPU
643 13 => 8 KB for each CPU
644 12 => 4 KB for each CPU
647 # Architectures with an unreliable sched_clock() should select this:
649 config HAVE_UNSTABLE_SCHED_CLOCK
652 config GENERIC_SCHED_CLOCK
656 # For architectures that want to enable the support for NUMA-affine scheduler
659 config ARCH_SUPPORTS_NUMA_BALANCING
663 # For architectures that prefer to flush all TLBs after a number of pages
664 # are unmapped instead of sending one IPI per page to flush. The architecture
665 # must provide guarantees on what happens if a clean TLB cache entry is
666 # written after the unmap. Details are in mm/rmap.c near the check for
667 # should_defer_flush. The architecture should also consider if the full flush
668 # and the refill costs are offset by the savings of sending fewer IPIs.
669 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
673 # For architectures that know their GCC __int128 support is sound
675 config ARCH_SUPPORTS_INT128
678 # For architectures that (ab)use NUMA to represent different memory regions
679 # all cpu-local but of different latencies, such as SuperH.
681 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
684 config NUMA_BALANCING
685 bool "Memory placement aware NUMA scheduler"
686 depends on ARCH_SUPPORTS_NUMA_BALANCING
687 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
688 depends on SMP && NUMA && MIGRATION
690 This option adds support for automatic NUMA aware memory/task placement.
691 The mechanism is quite primitive and is based on migrating memory when
692 it has references to the node the task is running on.
694 This system will be inactive on UMA systems.
696 config NUMA_BALANCING_DEFAULT_ENABLED
697 bool "Automatically enable NUMA aware memory/task placement"
699 depends on NUMA_BALANCING
701 If set, automatic NUMA balancing will be enabled if running on a NUMA
705 bool "Control Group support"
708 This option adds support for grouping sets of processes together, for
709 use with process control subsystems such as Cpusets, CFS, memory
710 controls or device isolation.
712 - Documentation/scheduler/sched-design-CFS.txt (CFS)
713 - Documentation/cgroup-v1/ (features for grouping, isolation
714 and resource control)
724 bool "Memory controller"
728 Provides control over the memory footprint of tasks in a cgroup.
731 bool "Swap controller"
732 depends on MEMCG && SWAP
734 Provides control over the swap space consumed by tasks in a cgroup.
736 config MEMCG_SWAP_ENABLED
737 bool "Swap controller enabled by default"
738 depends on MEMCG_SWAP
741 Memory Resource Controller Swap Extension comes with its price in
742 a bigger memory consumption. General purpose distribution kernels
743 which want to enable the feature but keep it disabled by default
744 and let the user enable it by swapaccount=1 boot command line
745 parameter should have this option unselected.
746 For those who want to have the feature enabled by default should
747 select this option (if, for some reason, they need to disable it
748 then swapaccount=0 does the trick).
752 depends on MEMCG && !SLOB
760 Generic block IO controller cgroup interface. This is the common
761 cgroup interface which should be used by various IO controlling
764 Currently, CFQ IO scheduler uses it to recognize task groups and
765 control disk bandwidth allocation (proportional time slice allocation)
766 to such task groups. It is also used by bio throttling logic in
767 block layer to implement upper limit in IO rates on a device.
769 This option only enables generic Block IO controller infrastructure.
770 One needs to also enable actual IO controlling logic/policy. For
771 enabling proportional weight division of disk bandwidth in CFQ, set
772 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
773 CONFIG_BLK_DEV_THROTTLING=y.
775 See Documentation/cgroup-v1/blkio-controller.txt for more information.
777 config DEBUG_BLK_CGROUP
778 bool "IO controller debugging"
779 depends on BLK_CGROUP
782 Enable some debugging help. Currently it exports additional stat
783 files in a cgroup which can be useful for debugging.
785 config CGROUP_WRITEBACK
787 depends on MEMCG && BLK_CGROUP
790 menuconfig CGROUP_SCHED
791 bool "CPU controller"
794 This feature lets CPU scheduler recognize task groups and control CPU
795 bandwidth allocation to such task groups. It uses cgroups to group
799 config FAIR_GROUP_SCHED
800 bool "Group scheduling for SCHED_OTHER"
801 depends on CGROUP_SCHED
805 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
806 depends on FAIR_GROUP_SCHED
809 This option allows users to define CPU bandwidth rates (limits) for
810 tasks running within the fair group scheduler. Groups with no limit
811 set are considered to be unconstrained and will run with no
813 See Documentation/scheduler/sched-bwc.txt for more information.
815 config RT_GROUP_SCHED
816 bool "Group scheduling for SCHED_RR/FIFO"
817 depends on CGROUP_SCHED
820 This feature lets you explicitly allocate real CPU bandwidth
821 to task groups. If enabled, it will also make it impossible to
822 schedule realtime tasks for non-root users until you allocate
823 realtime bandwidth for them.
824 See Documentation/scheduler/sched-rt-group.txt for more information.
829 bool "PIDs controller"
831 Provides enforcement of process number limits in the scope of a
832 cgroup. Any attempt to fork more processes than is allowed in the
833 cgroup will fail. PIDs are fundamentally a global resource because it
834 is fairly trivial to reach PID exhaustion before you reach even a
835 conservative kmemcg limit. As a result, it is possible to grind a
836 system to halt without being limited by other cgroup policies. The
837 PIDs controller is designed to stop this from happening.
839 It should be noted that organisational operations (such as attaching
840 to a cgroup hierarchy) will *not* be blocked by the PIDs controller,
841 since the PIDs limit only affects a process's ability to fork, not to
845 bool "RDMA controller"
847 Provides enforcement of RDMA resources defined by IB stack.
848 It is fairly easy for consumers to exhaust RDMA resources, which
849 can result into resource unavailability to other consumers.
850 RDMA controller is designed to stop this from happening.
851 Attaching processes with active RDMA resources to the cgroup
852 hierarchy is allowed even if can cross the hierarchy's limit.
854 config CGROUP_FREEZER
855 bool "Freezer controller"
857 Provides a way to freeze and unfreeze all tasks in a
860 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
861 controller includes important in-kernel memory consumers per default.
863 If you're using cgroup2, say N.
865 config CGROUP_HUGETLB
866 bool "HugeTLB controller"
867 depends on HUGETLB_PAGE
871 Provides a cgroup controller for HugeTLB pages.
872 When you enable this, you can put a per cgroup limit on HugeTLB usage.
873 The limit is enforced during page fault. Since HugeTLB doesn't
874 support page reclaim, enforcing the limit at page fault time implies
875 that, the application will get SIGBUS signal if it tries to access
876 HugeTLB pages beyond its limit. This requires the application to know
877 beforehand how much HugeTLB pages it would require for its use. The
878 control group is tracked in the third page lru pointer. This means
879 that we cannot use the controller with huge page less than 3 pages.
882 bool "Cpuset controller"
885 This option will let you create and manage CPUSETs which
886 allow dynamically partitioning a system into sets of CPUs and
887 Memory Nodes and assigning tasks to run only within those sets.
888 This is primarily useful on large SMP or NUMA systems.
892 config PROC_PID_CPUSET
893 bool "Include legacy /proc/<pid>/cpuset file"
898 bool "Device controller"
900 Provides a cgroup controller implementing whitelists for
901 devices which a process in the cgroup can mknod or open.
903 config CGROUP_CPUACCT
904 bool "Simple CPU accounting controller"
906 Provides a simple controller for monitoring the
907 total CPU consumed by the tasks in a cgroup.
910 bool "Perf controller"
911 depends on PERF_EVENTS
913 This option extends the perf per-cpu mode to restrict monitoring
914 to threads which belong to the cgroup specified and run on the
920 bool "Support for eBPF programs attached to cgroups"
921 depends on BPF_SYSCALL
922 select SOCK_CGROUP_DATA
924 Allow attaching eBPF programs to a cgroup using the bpf(2)
925 syscall command BPF_PROG_ATTACH.
927 In which context these programs are accessed depends on the type
928 of attachment. For instance, programs that are attached using
929 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
933 bool "Debug controller"
935 depends on DEBUG_KERNEL
937 This option enables a simple controller that exports
938 debugging information about the cgroups framework. This
939 controller is for control cgroup debugging only. Its
940 interfaces are not stable.
944 config SOCK_CGROUP_DATA
950 menuconfig NAMESPACES
951 bool "Namespaces support" if EXPERT
955 Provides the way to make tasks work with different objects using
956 the same id. For example same IPC id may refer to different objects
957 or same user id or pid may refer to different tasks when used in
958 different namespaces.
966 In this namespace tasks see different info provided with the
971 depends on (SYSVIPC || POSIX_MQUEUE)
974 In this namespace tasks work with IPC ids which correspond to
975 different IPC objects in different namespaces.
978 bool "User namespace"
981 This allows containers, i.e. vservers, to use user namespaces
982 to provide different user info for different servers.
984 When user namespaces are enabled in the kernel it is
985 recommended that the MEMCG option also be enabled and that
986 user-space use the memory control groups to limit the amount
987 of memory a memory unprivileged users can use.
992 bool "PID Namespaces"
995 Support process id namespaces. This allows having multiple
996 processes with the same pid as long as they are in different
997 pid namespaces. This is a building block of containers.
1000 bool "Network namespace"
1004 Allow user space to create what appear to be multiple instances
1005 of the network stack.
1009 config CHECKPOINT_RESTORE
1010 bool "Checkpoint/restore support"
1011 select PROC_CHILDREN
1014 Enables additional kernel features in a sake of checkpoint/restore.
1015 In particular it adds auxiliary prctl codes to setup process text,
1016 data and heap segment sizes, and a few additional /proc filesystem
1019 If unsure, say N here.
1021 config SCHED_AUTOGROUP
1022 bool "Automatic process group scheduling"
1025 select FAIR_GROUP_SCHED
1027 This option optimizes the scheduler for common desktop workloads by
1028 automatically creating and populating task groups. This separation
1029 of workloads isolates aggressive CPU burners (like build jobs) from
1030 desktop applications. Task group autogeneration is currently based
1033 config SYSFS_DEPRECATED
1034 bool "Enable deprecated sysfs features to support old userspace tools"
1038 This option adds code that switches the layout of the "block" class
1039 devices, to not show up in /sys/class/block/, but only in
1042 This switch is only active when the sysfs.deprecated=1 boot option is
1043 passed or the SYSFS_DEPRECATED_V2 option is set.
1045 This option allows new kernels to run on old distributions and tools,
1046 which might get confused by /sys/class/block/. Since 2007/2008 all
1047 major distributions and tools handle this just fine.
1049 Recent distributions and userspace tools after 2009/2010 depend on
1050 the existence of /sys/class/block/, and will not work with this
1053 Only if you are using a new kernel on an old distribution, you might
1056 config SYSFS_DEPRECATED_V2
1057 bool "Enable deprecated sysfs features by default"
1060 depends on SYSFS_DEPRECATED
1062 Enable deprecated sysfs by default.
1064 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1067 Only if you are using a new kernel on an old distribution, you might
1068 need to say Y here. Even then, odds are you would not need it
1069 enabled, you can always pass the boot option if absolutely necessary.
1072 bool "Kernel->user space relay support (formerly relayfs)"
1075 This option enables support for relay interface support in
1076 certain file systems (such as debugfs).
1077 It is designed to provide an efficient mechanism for tools and
1078 facilities to relay large amounts of data from kernel space to
1083 config BLK_DEV_INITRD
1084 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1086 The initial RAM filesystem is a ramfs which is loaded by the
1087 boot loader (loadlin or lilo) and that is mounted as root
1088 before the normal boot procedure. It is typically used to
1089 load modules needed to mount the "real" root file system,
1090 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1092 If RAM disk support (BLK_DEV_RAM) is also included, this
1093 also enables initial RAM disk (initrd) support and adds
1094 15 Kbytes (more on some other architectures) to the kernel size.
1100 source "usr/Kconfig"
1105 prompt "Compiler optimization level"
1106 default CC_OPTIMIZE_FOR_PERFORMANCE
1108 config CC_OPTIMIZE_FOR_PERFORMANCE
1109 bool "Optimize for performance"
1111 This is the default optimization level for the kernel, building
1112 with the "-O2" compiler flag for best performance and most
1113 helpful compile-time warnings.
1115 config CC_OPTIMIZE_FOR_SIZE
1116 bool "Optimize for size"
1118 Enabling this option will pass "-Os" instead of "-O2" to
1119 your compiler resulting in a smaller kernel.
1125 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1128 This requires that the arch annotates or otherwise protects
1129 its external entry points from being discarded. Linker scripts
1130 must also merge .text.*, .data.*, and .bss.* correctly into
1131 output sections. Care must be taken not to pull in unrelated
1132 sections (e.g., '.text.init'). Typically '.' in section names
1133 is used to distinguish them from label names / C identifiers.
1135 config LD_DEAD_CODE_DATA_ELIMINATION
1136 bool "Dead code and data elimination (EXPERIMENTAL)"
1137 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1139 depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800)
1140 depends on $(cc-option,-ffunction-sections -fdata-sections)
1141 depends on $(ld-option,--gc-sections)
1143 Enable this if you want to do dead code and data elimination with
1144 the linker by compiling with -ffunction-sections -fdata-sections,
1145 and linking with --gc-sections.
1147 This can reduce on disk and in-memory size of the kernel
1148 code and static data, particularly for small configs and
1149 on small systems. This has the possibility of introducing
1150 silently broken kernel if the required annotations are not
1151 present. This option is not well tested yet, so use at your
1163 config SYSCTL_EXCEPTION_TRACE
1166 Enable support for /proc/sys/debug/exception-trace.
1168 config SYSCTL_ARCH_UNALIGN_NO_WARN
1171 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1172 Allows arch to define/use @no_unaligned_warning to possibly warn
1173 about unaligned access emulation going on under the hood.
1175 config SYSCTL_ARCH_UNALIGN_ALLOW
1178 Enable support for /proc/sys/kernel/unaligned-trap
1179 Allows arches to define/use @unaligned_enabled to runtime toggle
1180 the unaligned access emulation.
1181 see arch/parisc/kernel/unaligned.c for reference
1183 config HAVE_PCSPKR_PLATFORM
1186 # interpreter that classic socket filters depend on
1191 bool "Configure standard kernel features (expert users)"
1192 # Unhide debug options, to make the on-by-default options visible
1195 This option allows certain base kernel options and settings
1196 to be disabled or tweaked. This is for specialized
1197 environments which can tolerate a "non-standard" kernel.
1198 Only use this if you really know what you are doing.
1201 bool "Enable 16-bit UID system calls" if EXPERT
1202 depends on HAVE_UID16 && MULTIUSER
1205 This enables the legacy 16-bit UID syscall wrappers.
1208 bool "Multiple users, groups and capabilities support" if EXPERT
1211 This option enables support for non-root users, groups and
1214 If you say N here, all processes will run with UID 0, GID 0, and all
1215 possible capabilities. Saying N here also compiles out support for
1216 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1219 If unsure, say Y here.
1221 config SGETMASK_SYSCALL
1222 bool "sgetmask/ssetmask syscalls support" if EXPERT
1223 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1225 sys_sgetmask and sys_ssetmask are obsolete system calls
1226 no longer supported in libc but still enabled by default in some
1229 If unsure, leave the default option here.
1231 config SYSFS_SYSCALL
1232 bool "Sysfs syscall support" if EXPERT
1235 sys_sysfs is an obsolete system call no longer supported in libc.
1236 Note that disabling this option is more secure but might break
1237 compatibility with some systems.
1239 If unsure say Y here.
1241 config SYSCTL_SYSCALL
1242 bool "Sysctl syscall support" if EXPERT
1243 depends on PROC_SYSCTL
1247 sys_sysctl uses binary paths that have been found challenging
1248 to properly maintain and use. The interface in /proc/sys
1249 using paths with ascii names is now the primary path to this
1252 Almost nothing using the binary sysctl interface so if you are
1253 trying to save some space it is probably safe to disable this,
1254 making your kernel marginally smaller.
1256 If unsure say N here.
1259 bool "open by fhandle syscalls" if EXPERT
1263 If you say Y here, a user level program will be able to map
1264 file names to handle and then later use the handle for
1265 different file system operations. This is useful in implementing
1266 userspace file servers, which now track files using handles instead
1267 of names. The handle would remain the same even if file names
1268 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1272 bool "Posix Clocks & timers" if EXPERT
1275 This includes native support for POSIX timers to the kernel.
1276 Some embedded systems have no use for them and therefore they
1277 can be configured out to reduce the size of the kernel image.
1279 When this option is disabled, the following syscalls won't be
1280 available: timer_create, timer_gettime: timer_getoverrun,
1281 timer_settime, timer_delete, clock_adjtime, getitimer,
1282 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1283 clock_getres and clock_nanosleep syscalls will be limited to
1284 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1290 bool "Enable support for printk" if EXPERT
1293 This option enables normal printk support. Removing it
1294 eliminates most of the message strings from the kernel image
1295 and makes the kernel more or less silent. As this makes it
1296 very difficult to diagnose system problems, saying N here is
1297 strongly discouraged.
1305 bool "BUG() support" if EXPERT
1308 Disabling this option eliminates support for BUG and WARN, reducing
1309 the size of your kernel image and potentially quietly ignoring
1310 numerous fatal conditions. You should only consider disabling this
1311 option for embedded systems with no facilities for reporting errors.
1317 bool "Enable ELF core dumps" if EXPERT
1319 Enable support for generating core dumps. Disabling saves about 4k.
1322 config PCSPKR_PLATFORM
1323 bool "Enable PC-Speaker support" if EXPERT
1324 depends on HAVE_PCSPKR_PLATFORM
1328 This option allows to disable the internal PC-Speaker
1329 support, saving some memory.
1333 bool "Enable full-sized data structures for core" if EXPERT
1335 Disabling this option reduces the size of miscellaneous core
1336 kernel data structures. This saves memory on small machines,
1337 but may reduce performance.
1340 bool "Enable futex support" if EXPERT
1344 Disabling this option will cause the kernel to be built without
1345 support for "fast userspace mutexes". The resulting kernel may not
1346 run glibc-based applications correctly.
1350 depends on FUTEX && RT_MUTEXES
1353 config HAVE_FUTEX_CMPXCHG
1357 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1358 is implemented and always working. This removes a couple of runtime
1362 bool "Enable eventpoll support" if EXPERT
1366 Disabling this option will cause the kernel to be built without
1367 support for epoll family of system calls.
1370 bool "Enable signalfd() system call" if EXPERT
1374 Enable the signalfd() system call that allows to receive signals
1375 on a file descriptor.
1380 bool "Enable timerfd() system call" if EXPERT
1384 Enable the timerfd() system call that allows to receive timer
1385 events on a file descriptor.
1390 bool "Enable eventfd() system call" if EXPERT
1394 Enable the eventfd() system call that allows to receive both
1395 kernel notification (ie. KAIO) or userspace notifications.
1400 bool "Use full shmem filesystem" if EXPERT
1404 The shmem is an internal filesystem used to manage shared memory.
1405 It is backed by swap and manages resource limits. It is also exported
1406 to userspace as tmpfs if TMPFS is enabled. Disabling this
1407 option replaces shmem and tmpfs with the much simpler ramfs code,
1408 which may be appropriate on small systems without swap.
1411 bool "Enable AIO support" if EXPERT
1414 This option enables POSIX asynchronous I/O which may by used
1415 by some high performance threaded applications. Disabling
1416 this option saves about 7k.
1419 bool "Enable IO uring support" if EXPERT
1423 This option enables support for the io_uring interface, enabling
1424 applications to submit and complete IO through submission and
1425 completion rings that are shared between the kernel and application.
1427 config ADVISE_SYSCALLS
1428 bool "Enable madvise/fadvise syscalls" if EXPERT
1431 This option enables the madvise and fadvise syscalls, used by
1432 applications to advise the kernel about their future memory or file
1433 usage, improving performance. If building an embedded system where no
1434 applications use these syscalls, you can disable this option to save
1438 bool "Enable membarrier() system call" if EXPERT
1441 Enable the membarrier() system call that allows issuing memory
1442 barriers across all running threads, which can be used to distribute
1443 the cost of user-space memory barriers asymmetrically by transforming
1444 pairs of memory barriers into pairs consisting of membarrier() and a
1450 bool "Load all symbols for debugging/ksymoops" if EXPERT
1453 Say Y here to let the kernel print out symbolic crash information and
1454 symbolic stack backtraces. This increases the size of the kernel
1455 somewhat, as all symbols have to be loaded into the kernel image.
1458 bool "Include all symbols in kallsyms"
1459 depends on DEBUG_KERNEL && KALLSYMS
1461 Normally kallsyms only contains the symbols of functions for nicer
1462 OOPS messages and backtraces (i.e., symbols from the text and inittext
1463 sections). This is sufficient for most cases. And only in very rare
1464 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1465 names of variables from the data sections, etc).
1467 This option makes sure that all symbols are loaded into the kernel
1468 image (i.e., symbols from all sections) in cost of increased kernel
1469 size (depending on the kernel configuration, it may be 300KiB or
1470 something like this).
1472 Say N unless you really need all symbols.
1474 config KALLSYMS_ABSOLUTE_PERCPU
1477 default X86_64 && SMP
1479 config KALLSYMS_BASE_RELATIVE
1484 Instead of emitting them as absolute values in the native word size,
1485 emit the symbol references in the kallsyms table as 32-bit entries,
1486 each containing a relative value in the range [base, base + U32_MAX]
1487 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1488 an absolute value in the range [0, S32_MAX] or a relative value in the
1489 range [base, base + S32_MAX], where base is the lowest relative symbol
1490 address encountered in the image.
1492 On 64-bit builds, this reduces the size of the address table by 50%,
1493 but more importantly, it results in entries whose values are build
1494 time constants, and no relocation pass is required at runtime to fix
1495 up the entries based on the runtime load address of the kernel.
1497 # end of the "standard kernel features (expert users)" menu
1499 # syscall, maps, verifier
1501 bool "Enable bpf() system call"
1507 Enable the bpf() system call that allows to manipulate eBPF
1508 programs and maps via file descriptors.
1510 config BPF_JIT_ALWAYS_ON
1511 bool "Permanently enable BPF JIT and remove BPF interpreter"
1512 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1514 Enables BPF JIT and removes BPF interpreter to avoid
1515 speculative execution of BPF instructions by the interpreter
1518 bool "Enable userfaultfd() system call"
1522 Enable the userfaultfd() system call that allows to intercept and
1523 handle page faults in userland.
1525 config ARCH_HAS_MEMBARRIER_CALLBACKS
1528 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1532 bool "Enable rseq() system call" if EXPERT
1534 depends on HAVE_RSEQ
1537 Enable the restartable sequences system call. It provides a
1538 user-space cache for the current CPU number value, which
1539 speeds up getting the current CPU number from user-space,
1540 as well as an ABI to speed up user-space operations on
1547 bool "Enabled debugging of rseq() system call" if EXPERT
1548 depends on RSEQ && DEBUG_KERNEL
1550 Enable extra debugging checks for the rseq system call.
1555 bool "Embedded system"
1556 option allnoconfig_y
1559 This option should be enabled if compiling the kernel for
1560 an embedded system so certain expert options are available
1563 config HAVE_PERF_EVENTS
1566 See tools/perf/design.txt for details.
1568 config PERF_USE_VMALLOC
1571 See tools/perf/design.txt for details
1574 bool "PC/104 support" if EXPERT
1576 Expose PC/104 form factor device drivers and options available for
1577 selection and configuration. Enable this option if your target
1578 machine has a PC/104 bus.
1580 menu "Kernel Performance Events And Counters"
1583 bool "Kernel performance events and counters"
1584 default y if PROFILING
1585 depends on HAVE_PERF_EVENTS
1590 Enable kernel support for various performance events provided
1591 by software and hardware.
1593 Software events are supported either built-in or via the
1594 use of generic tracepoints.
1596 Most modern CPUs support performance events via performance
1597 counter registers. These registers count the number of certain
1598 types of hw events: such as instructions executed, cachemisses
1599 suffered, or branches mis-predicted - without slowing down the
1600 kernel or applications. These registers can also trigger interrupts
1601 when a threshold number of events have passed - and can thus be
1602 used to profile the code that runs on that CPU.
1604 The Linux Performance Event subsystem provides an abstraction of
1605 these software and hardware event capabilities, available via a
1606 system call and used by the "perf" utility in tools/perf/. It
1607 provides per task and per CPU counters, and it provides event
1608 capabilities on top of those.
1612 config DEBUG_PERF_USE_VMALLOC
1614 bool "Debug: use vmalloc to back perf mmap() buffers"
1615 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1616 select PERF_USE_VMALLOC
1618 Use vmalloc memory to back perf mmap() buffers.
1620 Mostly useful for debugging the vmalloc code on platforms
1621 that don't require it.
1627 config VM_EVENT_COUNTERS
1629 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1631 VM event counters are needed for event counts to be shown.
1632 This option allows the disabling of the VM event counters
1633 on EXPERT systems. /proc/vmstat will only show page counts
1634 if VM event counters are disabled.
1638 bool "Enable SLUB debugging support" if EXPERT
1639 depends on SLUB && SYSFS
1641 SLUB has extensive debug support features. Disabling these can
1642 result in significant savings in code size. This also disables
1643 SLUB sysfs support. /sys/slab will not exist and there will be
1644 no support for cache validation etc.
1646 config SLUB_MEMCG_SYSFS_ON
1648 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1649 depends on SLUB && SYSFS && MEMCG
1651 SLUB creates a directory under /sys/kernel/slab for each
1652 allocation cache to host info and debug files. If memory
1653 cgroup is enabled, each cache can have per memory cgroup
1654 caches. SLUB can create the same sysfs directories for these
1655 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1656 to a very high number of debug files being created. This is
1657 controlled by slub_memcg_sysfs boot parameter and this
1658 config option determines the parameter's default value.
1661 bool "Disable heap randomization"
1664 Randomizing heap placement makes heap exploits harder, but it
1665 also breaks ancient binaries (including anything libc5 based).
1666 This option changes the bootup default to heap randomization
1667 disabled, and can be overridden at runtime by setting
1668 /proc/sys/kernel/randomize_va_space to 2.
1670 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1673 prompt "Choose SLAB allocator"
1676 This option allows to select a slab allocator.
1680 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1682 The regular slab allocator that is established and known to work
1683 well in all environments. It organizes cache hot objects in
1684 per cpu and per node queues.
1687 bool "SLUB (Unqueued Allocator)"
1688 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1690 SLUB is a slab allocator that minimizes cache line usage
1691 instead of managing queues of cached objects (SLAB approach).
1692 Per cpu caching is realized using slabs of objects instead
1693 of queues of objects. SLUB can use memory efficiently
1694 and has enhanced diagnostics. SLUB is the default choice for
1699 bool "SLOB (Simple Allocator)"
1701 SLOB replaces the stock allocator with a drastically simpler
1702 allocator. SLOB is generally more space efficient but
1703 does not perform as well on large systems.
1707 config SLAB_MERGE_DEFAULT
1708 bool "Allow slab caches to be merged"
1711 For reduced kernel memory fragmentation, slab caches can be
1712 merged when they share the same size and other characteristics.
1713 This carries a risk of kernel heap overflows being able to
1714 overwrite objects from merged caches (and more easily control
1715 cache layout), which makes such heap attacks easier to exploit
1716 by attackers. By keeping caches unmerged, these kinds of exploits
1717 can usually only damage objects in the same cache. To disable
1718 merging at runtime, "slab_nomerge" can be passed on the kernel
1721 config SLAB_FREELIST_RANDOM
1723 depends on SLAB || SLUB
1724 bool "SLAB freelist randomization"
1726 Randomizes the freelist order used on creating new pages. This
1727 security feature reduces the predictability of the kernel slab
1728 allocator against heap overflows.
1730 config SLAB_FREELIST_HARDENED
1731 bool "Harden slab freelist metadata"
1734 Many kernel heap attacks try to target slab cache metadata and
1735 other infrastructure. This options makes minor performance
1736 sacrifies to harden the kernel slab allocator against common
1737 freelist exploit methods.
1739 config SLUB_CPU_PARTIAL
1741 depends on SLUB && SMP
1742 bool "SLUB per cpu partial cache"
1744 Per cpu partial caches accellerate objects allocation and freeing
1745 that is local to a processor at the price of more indeterminism
1746 in the latency of the free. On overflow these caches will be cleared
1747 which requires the taking of locks that may cause latency spikes.
1748 Typically one would choose no for a realtime system.
1750 config MMAP_ALLOW_UNINITIALIZED
1751 bool "Allow mmapped anonymous memory to be uninitialized"
1752 depends on EXPERT && !MMU
1755 Normally, and according to the Linux spec, anonymous memory obtained
1756 from mmap() has its contents cleared before it is passed to
1757 userspace. Enabling this config option allows you to request that
1758 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1759 providing a huge performance boost. If this option is not enabled,
1760 then the flag will be ignored.
1762 This is taken advantage of by uClibc's malloc(), and also by
1763 ELF-FDPIC binfmt's brk and stack allocator.
1765 Because of the obvious security issues, this option should only be
1766 enabled on embedded devices where you control what is run in
1767 userspace. Since that isn't generally a problem on no-MMU systems,
1768 it is normally safe to say Y here.
1770 See Documentation/nommu-mmap.txt for more information.
1772 config SYSTEM_DATA_VERIFICATION
1774 select SYSTEM_TRUSTED_KEYRING
1778 select ASYMMETRIC_KEY_TYPE
1779 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1782 select X509_CERTIFICATE_PARSER
1783 select PKCS7_MESSAGE_PARSER
1785 Provide PKCS#7 message verification using the contents of the system
1786 trusted keyring to provide public keys. This then can be used for
1787 module verification, kexec image verification and firmware blob
1791 bool "Profiling support"
1793 Say Y here to enable the extended profiling support mechanisms used
1794 by profilers such as OProfile.
1797 # Place an empty function call at each tracepoint site. Can be
1798 # dynamically changed for a probe function.
1803 endmenu # General setup
1805 source "arch/Kconfig"
1812 default 0 if BASE_FULL
1813 default 1 if !BASE_FULL
1816 bool "Enable loadable module support"
1819 Kernel modules are small pieces of compiled code which can
1820 be inserted in the running kernel, rather than being
1821 permanently built into the kernel. You use the "modprobe"
1822 tool to add (and sometimes remove) them. If you say Y here,
1823 many parts of the kernel can be built as modules (by
1824 answering M instead of Y where indicated): this is most
1825 useful for infrequently used options which are not required
1826 for booting. For more information, see the man pages for
1827 modprobe, lsmod, modinfo, insmod and rmmod.
1829 If you say Y here, you will need to run "make
1830 modules_install" to put the modules under /lib/modules/
1831 where modprobe can find them (you may need to be root to do
1838 config MODULE_FORCE_LOAD
1839 bool "Forced module loading"
1842 Allow loading of modules without version information (ie. modprobe
1843 --force). Forced module loading sets the 'F' (forced) taint flag and
1844 is usually a really bad idea.
1846 config MODULE_UNLOAD
1847 bool "Module unloading"
1849 Without this option you will not be able to unload any
1850 modules (note that some modules may not be unloadable
1851 anyway), which makes your kernel smaller, faster
1852 and simpler. If unsure, say Y.
1854 config MODULE_FORCE_UNLOAD
1855 bool "Forced module unloading"
1856 depends on MODULE_UNLOAD
1858 This option allows you to force a module to unload, even if the
1859 kernel believes it is unsafe: the kernel will remove the module
1860 without waiting for anyone to stop using it (using the -f option to
1861 rmmod). This is mainly for kernel developers and desperate users.
1865 bool "Module versioning support"
1867 Usually, you have to use modules compiled with your kernel.
1868 Saying Y here makes it sometimes possible to use modules
1869 compiled for different kernels, by adding enough information
1870 to the modules to (hopefully) spot any changes which would
1871 make them incompatible with the kernel you are running. If
1874 config MODULE_REL_CRCS
1876 depends on MODVERSIONS
1878 config MODULE_SRCVERSION_ALL
1879 bool "Source checksum for all modules"
1881 Modules which contain a MODULE_VERSION get an extra "srcversion"
1882 field inserted into their modinfo section, which contains a
1883 sum of the source files which made it. This helps maintainers
1884 see exactly which source was used to build a module (since
1885 others sometimes change the module source without updating
1886 the version). With this option, such a "srcversion" field
1887 will be created for all modules. If unsure, say N.
1890 bool "Module signature verification"
1892 select SYSTEM_DATA_VERIFICATION
1894 Check modules for valid signatures upon load: the signature
1895 is simply appended to the module. For more information see
1896 <file:Documentation/admin-guide/module-signing.rst>.
1898 Note that this option adds the OpenSSL development packages as a
1899 kernel build dependency so that the signing tool can use its crypto
1902 !!!WARNING!!! If you enable this option, you MUST make sure that the
1903 module DOES NOT get stripped after being signed. This includes the
1904 debuginfo strip done by some packagers (such as rpmbuild) and
1905 inclusion into an initramfs that wants the module size reduced.
1907 config MODULE_SIG_FORCE
1908 bool "Require modules to be validly signed"
1909 depends on MODULE_SIG
1911 Reject unsigned modules or signed modules for which we don't have a
1912 key. Without this, such modules will simply taint the kernel.
1914 config MODULE_SIG_ALL
1915 bool "Automatically sign all modules"
1917 depends on MODULE_SIG
1919 Sign all modules during make modules_install. Without this option,
1920 modules must be signed manually, using the scripts/sign-file tool.
1922 comment "Do not forget to sign required modules with scripts/sign-file"
1923 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
1926 prompt "Which hash algorithm should modules be signed with?"
1927 depends on MODULE_SIG
1929 This determines which sort of hashing algorithm will be used during
1930 signature generation. This algorithm _must_ be built into the kernel
1931 directly so that signature verification can take place. It is not
1932 possible to load a signed module containing the algorithm to check
1933 the signature on that module.
1935 config MODULE_SIG_SHA1
1936 bool "Sign modules with SHA-1"
1939 config MODULE_SIG_SHA224
1940 bool "Sign modules with SHA-224"
1941 select CRYPTO_SHA256
1943 config MODULE_SIG_SHA256
1944 bool "Sign modules with SHA-256"
1945 select CRYPTO_SHA256
1947 config MODULE_SIG_SHA384
1948 bool "Sign modules with SHA-384"
1949 select CRYPTO_SHA512
1951 config MODULE_SIG_SHA512
1952 bool "Sign modules with SHA-512"
1953 select CRYPTO_SHA512
1957 config MODULE_SIG_HASH
1959 depends on MODULE_SIG
1960 default "sha1" if MODULE_SIG_SHA1
1961 default "sha224" if MODULE_SIG_SHA224
1962 default "sha256" if MODULE_SIG_SHA256
1963 default "sha384" if MODULE_SIG_SHA384
1964 default "sha512" if MODULE_SIG_SHA512
1966 config MODULE_COMPRESS
1967 bool "Compress modules on installation"
1971 Compresses kernel modules when 'make modules_install' is run; gzip or
1972 xz depending on "Compression algorithm" below.
1974 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
1976 Out-of-tree kernel modules installed using Kbuild will also be
1977 compressed upon installation.
1979 Note: for modules inside an initrd or initramfs, it's more efficient
1980 to compress the whole initrd or initramfs instead.
1982 Note: This is fully compatible with signed modules.
1987 prompt "Compression algorithm"
1988 depends on MODULE_COMPRESS
1989 default MODULE_COMPRESS_GZIP
1991 This determines which sort of compression will be used during
1992 'make modules_install'.
1994 GZIP (default) and XZ are supported.
1996 config MODULE_COMPRESS_GZIP
1999 config MODULE_COMPRESS_XZ
2004 config TRIM_UNUSED_KSYMS
2005 bool "Trim unused exported kernel symbols"
2006 depends on MODULES && !UNUSED_SYMBOLS
2008 The kernel and some modules make many symbols available for
2009 other modules to use via EXPORT_SYMBOL() and variants. Depending
2010 on the set of modules being selected in your kernel configuration,
2011 many of those exported symbols might never be used.
2013 This option allows for unused exported symbols to be dropped from
2014 the build. In turn, this provides the compiler more opportunities
2015 (especially when using LTO) for optimizing the code and reducing
2016 binary size. This might have some security advantages as well.
2018 If unsure, or if you need to build out-of-tree modules, say N.
2022 config MODULES_TREE_LOOKUP
2024 depends on PERF_EVENTS || TRACING
2026 config INIT_ALL_POSSIBLE
2029 Back when each arch used to define their own cpu_online_mask and
2030 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2031 with all 1s, and others with all 0s. When they were centralised,
2032 it was better to provide this option than to break all the archs
2033 and have several arch maintainers pursuing me down dark alleys.
2035 source "block/Kconfig"
2037 config PREEMPT_NOTIFIERS
2047 Build a simple ASN.1 grammar compiler that produces a bytecode output
2048 that can be interpreted by the ASN.1 stream decoder and used to
2049 inform it as to what tags are to be expected in a stream and what
2050 functions to call on what tags.
2052 source "kernel/Kconfig.locks"
2054 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2057 # It may be useful for an architecture to override the definitions of the
2058 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2059 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2060 # different calling convention for syscalls. They can also override the
2061 # macros for not-implemented syscalls in kernel/sys_ni.c and
2062 # kernel/time/posix-stubs.c. All these overrides need to be available in
2063 # <asm/syscall_wrapper.h>.
2064 config ARCH_HAS_SYSCALL_WRAPPER