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))
33 config BUILDTIME_EXTABLE_SORT
36 config THREAD_INFO_IN_TASK
39 Select this to move thread_info off the stack into task_struct. To
40 make this work, an arch will need to remove all thread_info fields
41 except flags and fix any runtime bugs.
43 One subtle change that will be needed is to use try_get_task_stack()
44 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
53 depends on BROKEN || !SMP
56 config INIT_ENV_ARG_LIMIT
61 Maximum of each of the number of arguments and environment
62 variables passed to init from the kernel command line.
65 bool "Compile also drivers which will not load"
69 Some drivers can be compiled on a different platform than they are
70 intended to be run on. Despite they cannot be loaded there (or even
71 when they load they cannot be used due to missing HW support),
72 developers still, opposing to distributors, might want to build such
73 drivers to compile-test them.
75 If you are a developer and want to build everything available, say Y
76 here. If you are a user/distributor, say N here to exclude useless
77 drivers to be distributed.
80 string "Local version - append to kernel release"
82 Append an extra string to the end of your kernel version.
83 This will show up when you type uname, for example.
84 The string you set here will be appended after the contents of
85 any files with a filename matching localversion* in your
86 object and source tree, in that order. Your total string can
87 be a maximum of 64 characters.
89 config LOCALVERSION_AUTO
90 bool "Automatically append version information to the version string"
92 depends on !COMPILE_TEST
94 This will try to automatically determine if the current tree is a
95 release tree by looking for git tags that belong to the current
98 A string of the format -gxxxxxxxx will be added to the localversion
99 if a git-based tree is found. The string generated by this will be
100 appended after any matching localversion* files, and after the value
101 set in CONFIG_LOCALVERSION.
103 (The actual string used here is the first eight characters produced
104 by running the command:
106 $ git rev-parse --verify HEAD
108 which is done within the script "scripts/setlocalversion".)
111 string "Build ID Salt"
114 The build ID is used to link binaries and their debug info. Setting
115 this option will use the value in the calculation of the build id.
116 This is mostly useful for distributions which want to ensure the
117 build is unique between builds. It's safe to leave the default.
119 config HAVE_KERNEL_GZIP
122 config HAVE_KERNEL_BZIP2
125 config HAVE_KERNEL_LZMA
128 config HAVE_KERNEL_XZ
131 config HAVE_KERNEL_LZO
134 config HAVE_KERNEL_LZ4
137 config HAVE_KERNEL_UNCOMPRESSED
141 prompt "Kernel compression mode"
143 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
145 The linux kernel is a kind of self-extracting executable.
146 Several compression algorithms are available, which differ
147 in efficiency, compression and decompression speed.
148 Compression speed is only relevant when building a kernel.
149 Decompression speed is relevant at each boot.
151 If you have any problems with bzip2 or lzma compressed
152 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
153 version of this functionality (bzip2 only), for 2.4, was
154 supplied by Christian Ludwig)
156 High compression options are mostly useful for users, who
157 are low on disk space (embedded systems), but for whom ram
160 If in doubt, select 'gzip'
164 depends on HAVE_KERNEL_GZIP
166 The old and tried gzip compression. It provides a good balance
167 between compression ratio and decompression speed.
171 depends on HAVE_KERNEL_BZIP2
173 Its compression ratio and speed is intermediate.
174 Decompression speed is slowest among the choices. The kernel
175 size is about 10% smaller with bzip2, in comparison to gzip.
176 Bzip2 uses a large amount of memory. For modern kernels you
177 will need at least 8MB RAM or more for booting.
181 depends on HAVE_KERNEL_LZMA
183 This compression algorithm's ratio is best. Decompression speed
184 is between gzip and bzip2. Compression is slowest.
185 The kernel size is about 33% smaller with LZMA in comparison to gzip.
189 depends on HAVE_KERNEL_XZ
191 XZ uses the LZMA2 algorithm and instruction set specific
192 BCJ filters which can improve compression ratio of executable
193 code. The size of the kernel is about 30% smaller with XZ in
194 comparison to gzip. On architectures for which there is a BCJ
195 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
196 will create a few percent smaller kernel than plain LZMA.
198 The speed is about the same as with LZMA: The decompression
199 speed of XZ is better than that of bzip2 but worse than gzip
200 and LZO. Compression is slow.
204 depends on HAVE_KERNEL_LZO
206 Its compression ratio is the poorest among the choices. The kernel
207 size is about 10% bigger than gzip; however its speed
208 (both compression and decompression) is the fastest.
212 depends on HAVE_KERNEL_LZ4
214 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
215 A preliminary version of LZ4 de/compression tool is available at
216 <https://code.google.com/p/lz4/>.
218 Its compression ratio is worse than LZO. The size of the kernel
219 is about 8% bigger than LZO. But the decompression speed is
222 config KERNEL_UNCOMPRESSED
224 depends on HAVE_KERNEL_UNCOMPRESSED
226 Produce uncompressed kernel image. This option is usually not what
227 you want. It is useful for debugging the kernel in slow simulation
228 environments, where decompressing and moving the kernel is awfully
229 slow. This option allows early boot code to skip the decompressor
230 and jump right at uncompressed kernel image.
234 config DEFAULT_HOSTNAME
235 string "Default hostname"
238 This option determines the default system hostname before userspace
239 calls sethostname(2). The kernel traditionally uses "(none)" here,
240 but you may wish to use a different default here to make a minimal
241 system more usable with less configuration.
244 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
245 # add proper SWAP support to them, in which case this can be remove.
251 bool "Support for paging of anonymous memory (swap)"
252 depends on MMU && BLOCK && !ARCH_NO_SWAP
255 This option allows you to choose whether you want to have support
256 for so called swap devices or swap files in your kernel that are
257 used to provide more virtual memory than the actual RAM present
258 in your computer. If unsure say Y.
263 Inter Process Communication is a suite of library functions and
264 system calls which let processes (running programs) synchronize and
265 exchange information. It is generally considered to be a good thing,
266 and some programs won't run unless you say Y here. In particular, if
267 you want to run the DOS emulator dosemu under Linux (read the
268 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
269 you'll need to say Y here.
271 You can find documentation about IPC with "info ipc" and also in
272 section 6.4 of the Linux Programmer's Guide, available from
273 <http://www.tldp.org/guides.html>.
275 config SYSVIPC_SYSCTL
282 bool "POSIX Message Queues"
285 POSIX variant of message queues is a part of IPC. In POSIX message
286 queues every message has a priority which decides about succession
287 of receiving it by a process. If you want to compile and run
288 programs written e.g. for Solaris with use of its POSIX message
289 queues (functions mq_*) say Y here.
291 POSIX message queues are visible as a filesystem called 'mqueue'
292 and can be mounted somewhere if you want to do filesystem
293 operations on message queues.
297 config POSIX_MQUEUE_SYSCTL
299 depends on POSIX_MQUEUE
303 config CROSS_MEMORY_ATTACH
304 bool "Enable process_vm_readv/writev syscalls"
308 Enabling this option adds the system calls process_vm_readv and
309 process_vm_writev which allow a process with the correct privileges
310 to directly read from or write to another process' address space.
311 See the man page for more details.
314 bool "uselib syscall"
315 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
317 This option enables the uselib syscall, a system call used in the
318 dynamic linker from libc5 and earlier. glibc does not use this
319 system call. If you intend to run programs built on libc5 or
320 earlier, you may need to enable this syscall. Current systems
321 running glibc can safely disable this.
324 bool "Auditing support"
327 Enable auditing infrastructure that can be used with another
328 kernel subsystem, such as SELinux (which requires this for
329 logging of avc messages output). System call auditing is included
330 on architectures which support it.
332 config HAVE_ARCH_AUDITSYSCALL
337 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
341 depends on AUDITSYSCALL
346 depends on AUDITSYSCALL
349 source "kernel/irq/Kconfig"
350 source "kernel/time/Kconfig"
351 source "kernel/Kconfig.preempt"
353 menu "CPU/Task time and stats accounting"
355 config VIRT_CPU_ACCOUNTING
359 prompt "Cputime accounting"
360 default TICK_CPU_ACCOUNTING if !PPC64
361 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
363 # Kind of a stub config for the pure tick based cputime accounting
364 config TICK_CPU_ACCOUNTING
365 bool "Simple tick based cputime accounting"
366 depends on !S390 && !NO_HZ_FULL
368 This is the basic tick based cputime accounting that maintains
369 statistics about user, system and idle time spent on per jiffies
374 config VIRT_CPU_ACCOUNTING_NATIVE
375 bool "Deterministic task and CPU time accounting"
376 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
377 select VIRT_CPU_ACCOUNTING
379 Select this option to enable more accurate task and CPU time
380 accounting. This is done by reading a CPU counter on each
381 kernel entry and exit and on transitions within the kernel
382 between system, softirq and hardirq state, so there is a
383 small performance impact. In the case of s390 or IBM POWER > 5,
384 this also enables accounting of stolen time on logically-partitioned
387 config VIRT_CPU_ACCOUNTING_GEN
388 bool "Full dynticks CPU time accounting"
389 depends on HAVE_CONTEXT_TRACKING
390 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
391 select VIRT_CPU_ACCOUNTING
392 select CONTEXT_TRACKING
394 Select this option to enable task and CPU time accounting on full
395 dynticks systems. This accounting is implemented by watching every
396 kernel-user boundaries using the context tracking subsystem.
397 The accounting is thus performed at the expense of some significant
400 For now this is only useful if you are working on the full
401 dynticks subsystem development.
407 config IRQ_TIME_ACCOUNTING
408 bool "Fine granularity task level IRQ time accounting"
409 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
411 Select this option to enable fine granularity task irq time
412 accounting. This is done by reading a timestamp on each
413 transitions between softirq and hardirq state, so there can be a
414 small performance impact.
416 If in doubt, say N here.
418 config HAVE_SCHED_AVG_IRQ
420 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
423 config BSD_PROCESS_ACCT
424 bool "BSD Process Accounting"
427 If you say Y here, a user level program will be able to instruct the
428 kernel (via a special system call) to write process accounting
429 information to a file: whenever a process exits, information about
430 that process will be appended to the file by the kernel. The
431 information includes things such as creation time, owning user,
432 command name, memory usage, controlling terminal etc. (the complete
433 list is in the struct acct in <file:include/linux/acct.h>). It is
434 up to the user level program to do useful things with this
435 information. This is generally a good idea, so say Y.
437 config BSD_PROCESS_ACCT_V3
438 bool "BSD Process Accounting version 3 file format"
439 depends on BSD_PROCESS_ACCT
442 If you say Y here, the process accounting information is written
443 in a new file format that also logs the process IDs of each
444 process and its parent. Note that this file format is incompatible
445 with previous v0/v1/v2 file formats, so you will need updated tools
446 for processing it. A preliminary version of these tools is available
447 at <http://www.gnu.org/software/acct/>.
450 bool "Export task/process statistics through netlink"
455 Export selected statistics for tasks/processes through the
456 generic netlink interface. Unlike BSD process accounting, the
457 statistics are available during the lifetime of tasks/processes as
458 responses to commands. Like BSD accounting, they are sent to user
463 config TASK_DELAY_ACCT
464 bool "Enable per-task delay accounting"
468 Collect information on time spent by a task waiting for system
469 resources like cpu, synchronous block I/O completion and swapping
470 in pages. Such statistics can help in setting a task's priorities
471 relative to other tasks for cpu, io, rss limits etc.
476 bool "Enable extended accounting over taskstats"
479 Collect extended task accounting data and send the data
480 to userland for processing over the taskstats interface.
484 config TASK_IO_ACCOUNTING
485 bool "Enable per-task storage I/O accounting"
486 depends on TASK_XACCT
488 Collect information on the number of bytes of storage I/O which this
494 bool "Pressure stall information tracking"
496 Collect metrics that indicate how overcommitted the CPU, memory,
497 and IO capacity are in the system.
499 If you say Y here, the kernel will create /proc/pressure/ with the
500 pressure statistics files cpu, memory, and io. These will indicate
501 the share of walltime in which some or all tasks in the system are
502 delayed due to contention of the respective resource.
504 For more details see Documentation/accounting/psi.txt.
508 endmenu # "CPU/Task time and stats accounting"
512 depends on SMP || COMPILE_TEST
515 Make sure that CPUs running critical tasks are not disturbed by
516 any source of "noise" such as unbound workqueues, timers, kthreads...
517 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
518 the "isolcpus=" boot parameter.
522 source "kernel/rcu/Kconfig"
529 tristate "Kernel .config support"
532 This option enables the complete Linux kernel ".config" file
533 contents to be saved in the kernel. It provides documentation
534 of which kernel options are used in a running kernel or in an
535 on-disk kernel. This information can be extracted from the kernel
536 image file with the script scripts/extract-ikconfig and used as
537 input to rebuild the current kernel or to build another kernel.
538 It can also be extracted from a running kernel by reading
539 /proc/config.gz if enabled (below).
542 bool "Enable access to .config through /proc/config.gz"
543 depends on IKCONFIG && PROC_FS
545 This option enables access to the kernel configuration file
546 through /proc/config.gz.
549 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
554 Select the minimal kernel log buffer size as a power of 2.
555 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
556 parameter, see below. Any higher size also might be forced
557 by "log_buf_len" boot parameter.
567 config LOG_CPU_MAX_BUF_SHIFT
568 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
571 default 12 if !BASE_SMALL
572 default 0 if BASE_SMALL
575 This option allows to increase the default ring buffer size
576 according to the number of CPUs. The value defines the contribution
577 of each CPU as a power of 2. The used space is typically only few
578 lines however it might be much more when problems are reported,
581 The increased size means that a new buffer has to be allocated and
582 the original static one is unused. It makes sense only on systems
583 with more CPUs. Therefore this value is used only when the sum of
584 contributions is greater than the half of the default kernel ring
585 buffer as defined by LOG_BUF_SHIFT. The default values are set
586 so that more than 64 CPUs are needed to trigger the allocation.
588 Also this option is ignored when "log_buf_len" kernel parameter is
589 used as it forces an exact (power of two) size of the ring buffer.
591 The number of possible CPUs is used for this computation ignoring
592 hotplugging making the computation optimal for the worst case
593 scenario while allowing a simple algorithm to be used from bootup.
595 Examples shift values and their meaning:
596 17 => 128 KB for each CPU
597 16 => 64 KB for each CPU
598 15 => 32 KB for each CPU
599 14 => 16 KB for each CPU
600 13 => 8 KB for each CPU
601 12 => 4 KB for each CPU
603 config PRINTK_SAFE_LOG_BUF_SHIFT
604 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
609 Select the size of an alternate printk per-CPU buffer where messages
610 printed from usafe contexts are temporary stored. One example would
611 be NMI messages, another one - printk recursion. The messages are
612 copied to the main log buffer in a safe context to avoid a deadlock.
613 The value defines the size as a power of 2.
615 Those messages are rare and limited. The largest one is when
616 a backtrace is printed. It usually fits into 4KB. Select
617 8KB if you want to be on the safe side.
620 17 => 128 KB for each CPU
621 16 => 64 KB for each CPU
622 15 => 32 KB for each CPU
623 14 => 16 KB for each CPU
624 13 => 8 KB for each CPU
625 12 => 4 KB for each CPU
628 # Architectures with an unreliable sched_clock() should select this:
630 config HAVE_UNSTABLE_SCHED_CLOCK
633 config GENERIC_SCHED_CLOCK
637 # For architectures that want to enable the support for NUMA-affine scheduler
640 config ARCH_SUPPORTS_NUMA_BALANCING
644 # For architectures that prefer to flush all TLBs after a number of pages
645 # are unmapped instead of sending one IPI per page to flush. The architecture
646 # must provide guarantees on what happens if a clean TLB cache entry is
647 # written after the unmap. Details are in mm/rmap.c near the check for
648 # should_defer_flush. The architecture should also consider if the full flush
649 # and the refill costs are offset by the savings of sending fewer IPIs.
650 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
654 # For architectures that know their GCC __int128 support is sound
656 config ARCH_SUPPORTS_INT128
659 # For architectures that (ab)use NUMA to represent different memory regions
660 # all cpu-local but of different latencies, such as SuperH.
662 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
665 config NUMA_BALANCING
666 bool "Memory placement aware NUMA scheduler"
667 depends on ARCH_SUPPORTS_NUMA_BALANCING
668 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
669 depends on SMP && NUMA && MIGRATION
671 This option adds support for automatic NUMA aware memory/task placement.
672 The mechanism is quite primitive and is based on migrating memory when
673 it has references to the node the task is running on.
675 This system will be inactive on UMA systems.
677 config NUMA_BALANCING_DEFAULT_ENABLED
678 bool "Automatically enable NUMA aware memory/task placement"
680 depends on NUMA_BALANCING
682 If set, automatic NUMA balancing will be enabled if running on a NUMA
686 bool "Control Group support"
689 This option adds support for grouping sets of processes together, for
690 use with process control subsystems such as Cpusets, CFS, memory
691 controls or device isolation.
693 - Documentation/scheduler/sched-design-CFS.txt (CFS)
694 - Documentation/cgroup-v1/ (features for grouping, isolation
695 and resource control)
705 bool "Memory controller"
709 Provides control over the memory footprint of tasks in a cgroup.
712 bool "Swap controller"
713 depends on MEMCG && SWAP
715 Provides control over the swap space consumed by tasks in a cgroup.
717 config MEMCG_SWAP_ENABLED
718 bool "Swap controller enabled by default"
719 depends on MEMCG_SWAP
722 Memory Resource Controller Swap Extension comes with its price in
723 a bigger memory consumption. General purpose distribution kernels
724 which want to enable the feature but keep it disabled by default
725 and let the user enable it by swapaccount=1 boot command line
726 parameter should have this option unselected.
727 For those who want to have the feature enabled by default should
728 select this option (if, for some reason, they need to disable it
729 then swapaccount=0 does the trick).
733 depends on MEMCG && !SLOB
741 Generic block IO controller cgroup interface. This is the common
742 cgroup interface which should be used by various IO controlling
745 Currently, CFQ IO scheduler uses it to recognize task groups and
746 control disk bandwidth allocation (proportional time slice allocation)
747 to such task groups. It is also used by bio throttling logic in
748 block layer to implement upper limit in IO rates on a device.
750 This option only enables generic Block IO controller infrastructure.
751 One needs to also enable actual IO controlling logic/policy. For
752 enabling proportional weight division of disk bandwidth in CFQ, set
753 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
754 CONFIG_BLK_DEV_THROTTLING=y.
756 See Documentation/cgroup-v1/blkio-controller.txt for more information.
758 config DEBUG_BLK_CGROUP
759 bool "IO controller debugging"
760 depends on BLK_CGROUP
763 Enable some debugging help. Currently it exports additional stat
764 files in a cgroup which can be useful for debugging.
766 config CGROUP_WRITEBACK
768 depends on MEMCG && BLK_CGROUP
771 menuconfig CGROUP_SCHED
772 bool "CPU controller"
775 This feature lets CPU scheduler recognize task groups and control CPU
776 bandwidth allocation to such task groups. It uses cgroups to group
780 config FAIR_GROUP_SCHED
781 bool "Group scheduling for SCHED_OTHER"
782 depends on CGROUP_SCHED
786 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
787 depends on FAIR_GROUP_SCHED
790 This option allows users to define CPU bandwidth rates (limits) for
791 tasks running within the fair group scheduler. Groups with no limit
792 set are considered to be unconstrained and will run with no
794 See Documentation/scheduler/sched-bwc.txt for more information.
796 config RT_GROUP_SCHED
797 bool "Group scheduling for SCHED_RR/FIFO"
798 depends on CGROUP_SCHED
801 This feature lets you explicitly allocate real CPU bandwidth
802 to task groups. If enabled, it will also make it impossible to
803 schedule realtime tasks for non-root users until you allocate
804 realtime bandwidth for them.
805 See Documentation/scheduler/sched-rt-group.txt for more information.
810 bool "PIDs controller"
812 Provides enforcement of process number limits in the scope of a
813 cgroup. Any attempt to fork more processes than is allowed in the
814 cgroup will fail. PIDs are fundamentally a global resource because it
815 is fairly trivial to reach PID exhaustion before you reach even a
816 conservative kmemcg limit. As a result, it is possible to grind a
817 system to halt without being limited by other cgroup policies. The
818 PIDs controller is designed to stop this from happening.
820 It should be noted that organisational operations (such as attaching
821 to a cgroup hierarchy will *not* be blocked by the PIDs controller),
822 since the PIDs limit only affects a process's ability to fork, not to
826 bool "RDMA controller"
828 Provides enforcement of RDMA resources defined by IB stack.
829 It is fairly easy for consumers to exhaust RDMA resources, which
830 can result into resource unavailability to other consumers.
831 RDMA controller is designed to stop this from happening.
832 Attaching processes with active RDMA resources to the cgroup
833 hierarchy is allowed even if can cross the hierarchy's limit.
835 config CGROUP_FREEZER
836 bool "Freezer controller"
838 Provides a way to freeze and unfreeze all tasks in a
841 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
842 controller includes important in-kernel memory consumers per default.
844 If you're using cgroup2, say N.
846 config CGROUP_HUGETLB
847 bool "HugeTLB controller"
848 depends on HUGETLB_PAGE
852 Provides a cgroup controller for HugeTLB pages.
853 When you enable this, you can put a per cgroup limit on HugeTLB usage.
854 The limit is enforced during page fault. Since HugeTLB doesn't
855 support page reclaim, enforcing the limit at page fault time implies
856 that, the application will get SIGBUS signal if it tries to access
857 HugeTLB pages beyond its limit. This requires the application to know
858 beforehand how much HugeTLB pages it would require for its use. The
859 control group is tracked in the third page lru pointer. This means
860 that we cannot use the controller with huge page less than 3 pages.
863 bool "Cpuset controller"
866 This option will let you create and manage CPUSETs which
867 allow dynamically partitioning a system into sets of CPUs and
868 Memory Nodes and assigning tasks to run only within those sets.
869 This is primarily useful on large SMP or NUMA systems.
873 config PROC_PID_CPUSET
874 bool "Include legacy /proc/<pid>/cpuset file"
879 bool "Device controller"
881 Provides a cgroup controller implementing whitelists for
882 devices which a process in the cgroup can mknod or open.
884 config CGROUP_CPUACCT
885 bool "Simple CPU accounting controller"
887 Provides a simple controller for monitoring the
888 total CPU consumed by the tasks in a cgroup.
891 bool "Perf controller"
892 depends on PERF_EVENTS
894 This option extends the perf per-cpu mode to restrict monitoring
895 to threads which belong to the cgroup specified and run on the
901 bool "Support for eBPF programs attached to cgroups"
902 depends on BPF_SYSCALL
903 select SOCK_CGROUP_DATA
905 Allow attaching eBPF programs to a cgroup using the bpf(2)
906 syscall command BPF_PROG_ATTACH.
908 In which context these programs are accessed depends on the type
909 of attachment. For instance, programs that are attached using
910 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
914 bool "Debug controller"
916 depends on DEBUG_KERNEL
918 This option enables a simple controller that exports
919 debugging information about the cgroups framework. This
920 controller is for control cgroup debugging only. Its
921 interfaces are not stable.
925 config SOCK_CGROUP_DATA
931 menuconfig NAMESPACES
932 bool "Namespaces support" if EXPERT
936 Provides the way to make tasks work with different objects using
937 the same id. For example same IPC id may refer to different objects
938 or same user id or pid may refer to different tasks when used in
939 different namespaces.
947 In this namespace tasks see different info provided with the
952 depends on (SYSVIPC || POSIX_MQUEUE)
955 In this namespace tasks work with IPC ids which correspond to
956 different IPC objects in different namespaces.
959 bool "User namespace"
962 This allows containers, i.e. vservers, to use user namespaces
963 to provide different user info for different servers.
965 When user namespaces are enabled in the kernel it is
966 recommended that the MEMCG option also be enabled and that
967 user-space use the memory control groups to limit the amount
968 of memory a memory unprivileged users can use.
973 bool "PID Namespaces"
976 Support process id namespaces. This allows having multiple
977 processes with the same pid as long as they are in different
978 pid namespaces. This is a building block of containers.
981 bool "Network namespace"
985 Allow user space to create what appear to be multiple instances
986 of the network stack.
990 config CHECKPOINT_RESTORE
991 bool "Checkpoint/restore support"
995 Enables additional kernel features in a sake of checkpoint/restore.
996 In particular it adds auxiliary prctl codes to setup process text,
997 data and heap segment sizes, and a few additional /proc filesystem
1000 If unsure, say N here.
1002 config SCHED_AUTOGROUP
1003 bool "Automatic process group scheduling"
1006 select FAIR_GROUP_SCHED
1008 This option optimizes the scheduler for common desktop workloads by
1009 automatically creating and populating task groups. This separation
1010 of workloads isolates aggressive CPU burners (like build jobs) from
1011 desktop applications. Task group autogeneration is currently based
1014 config SYSFS_DEPRECATED
1015 bool "Enable deprecated sysfs features to support old userspace tools"
1019 This option adds code that switches the layout of the "block" class
1020 devices, to not show up in /sys/class/block/, but only in
1023 This switch is only active when the sysfs.deprecated=1 boot option is
1024 passed or the SYSFS_DEPRECATED_V2 option is set.
1026 This option allows new kernels to run on old distributions and tools,
1027 which might get confused by /sys/class/block/. Since 2007/2008 all
1028 major distributions and tools handle this just fine.
1030 Recent distributions and userspace tools after 2009/2010 depend on
1031 the existence of /sys/class/block/, and will not work with this
1034 Only if you are using a new kernel on an old distribution, you might
1037 config SYSFS_DEPRECATED_V2
1038 bool "Enable deprecated sysfs features by default"
1041 depends on SYSFS_DEPRECATED
1043 Enable deprecated sysfs by default.
1045 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1048 Only if you are using a new kernel on an old distribution, you might
1049 need to say Y here. Even then, odds are you would not need it
1050 enabled, you can always pass the boot option if absolutely necessary.
1053 bool "Kernel->user space relay support (formerly relayfs)"
1056 This option enables support for relay interface support in
1057 certain file systems (such as debugfs).
1058 It is designed to provide an efficient mechanism for tools and
1059 facilities to relay large amounts of data from kernel space to
1064 config BLK_DEV_INITRD
1065 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1067 The initial RAM filesystem is a ramfs which is loaded by the
1068 boot loader (loadlin or lilo) and that is mounted as root
1069 before the normal boot procedure. It is typically used to
1070 load modules needed to mount the "real" root file system,
1071 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1073 If RAM disk support (BLK_DEV_RAM) is also included, this
1074 also enables initial RAM disk (initrd) support and adds
1075 15 Kbytes (more on some other architectures) to the kernel size.
1081 source "usr/Kconfig"
1086 prompt "Compiler optimization level"
1087 default CC_OPTIMIZE_FOR_PERFORMANCE
1089 config CC_OPTIMIZE_FOR_PERFORMANCE
1090 bool "Optimize for performance"
1092 This is the default optimization level for the kernel, building
1093 with the "-O2" compiler flag for best performance and most
1094 helpful compile-time warnings.
1096 config CC_OPTIMIZE_FOR_SIZE
1097 bool "Optimize for size"
1099 Enabling this option will pass "-Os" instead of "-O2" to
1100 your compiler resulting in a smaller kernel.
1106 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1109 This requires that the arch annotates or otherwise protects
1110 its external entry points from being discarded. Linker scripts
1111 must also merge .text.*, .data.*, and .bss.* correctly into
1112 output sections. Care must be taken not to pull in unrelated
1113 sections (e.g., '.text.init'). Typically '.' in section names
1114 is used to distinguish them from label names / C identifiers.
1116 config LD_DEAD_CODE_DATA_ELIMINATION
1117 bool "Dead code and data elimination (EXPERIMENTAL)"
1118 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1120 depends on $(cc-option,-ffunction-sections -fdata-sections)
1121 depends on $(ld-option,--gc-sections)
1123 Enable this if you want to do dead code and data elimination with
1124 the linker by compiling with -ffunction-sections -fdata-sections,
1125 and linking with --gc-sections.
1127 This can reduce on disk and in-memory size of the kernel
1128 code and static data, particularly for small configs and
1129 on small systems. This has the possibility of introducing
1130 silently broken kernel if the required annotations are not
1131 present. This option is not well tested yet, so use at your
1143 config SYSCTL_EXCEPTION_TRACE
1146 Enable support for /proc/sys/debug/exception-trace.
1148 config SYSCTL_ARCH_UNALIGN_NO_WARN
1151 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1152 Allows arch to define/use @no_unaligned_warning to possibly warn
1153 about unaligned access emulation going on under the hood.
1155 config SYSCTL_ARCH_UNALIGN_ALLOW
1158 Enable support for /proc/sys/kernel/unaligned-trap
1159 Allows arches to define/use @unaligned_enabled to runtime toggle
1160 the unaligned access emulation.
1161 see arch/parisc/kernel/unaligned.c for reference
1163 config HAVE_PCSPKR_PLATFORM
1166 # interpreter that classic socket filters depend on
1171 bool "Configure standard kernel features (expert users)"
1172 # Unhide debug options, to make the on-by-default options visible
1175 This option allows certain base kernel options and settings
1176 to be disabled or tweaked. This is for specialized
1177 environments which can tolerate a "non-standard" kernel.
1178 Only use this if you really know what you are doing.
1181 bool "Enable 16-bit UID system calls" if EXPERT
1182 depends on HAVE_UID16 && MULTIUSER
1185 This enables the legacy 16-bit UID syscall wrappers.
1188 bool "Multiple users, groups and capabilities support" if EXPERT
1191 This option enables support for non-root users, groups and
1194 If you say N here, all processes will run with UID 0, GID 0, and all
1195 possible capabilities. Saying N here also compiles out support for
1196 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1199 If unsure, say Y here.
1201 config SGETMASK_SYSCALL
1202 bool "sgetmask/ssetmask syscalls support" if EXPERT
1203 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1205 sys_sgetmask and sys_ssetmask are obsolete system calls
1206 no longer supported in libc but still enabled by default in some
1209 If unsure, leave the default option here.
1211 config SYSFS_SYSCALL
1212 bool "Sysfs syscall support" if EXPERT
1215 sys_sysfs is an obsolete system call no longer supported in libc.
1216 Note that disabling this option is more secure but might break
1217 compatibility with some systems.
1219 If unsure say Y here.
1221 config SYSCTL_SYSCALL
1222 bool "Sysctl syscall support" if EXPERT
1223 depends on PROC_SYSCTL
1227 sys_sysctl uses binary paths that have been found challenging
1228 to properly maintain and use. The interface in /proc/sys
1229 using paths with ascii names is now the primary path to this
1232 Almost nothing using the binary sysctl interface so if you are
1233 trying to save some space it is probably safe to disable this,
1234 making your kernel marginally smaller.
1236 If unsure say N here.
1239 bool "open by fhandle syscalls" if EXPERT
1243 If you say Y here, a user level program will be able to map
1244 file names to handle and then later use the handle for
1245 different file system operations. This is useful in implementing
1246 userspace file servers, which now track files using handles instead
1247 of names. The handle would remain the same even if file names
1248 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1252 bool "Posix Clocks & timers" if EXPERT
1255 This includes native support for POSIX timers to the kernel.
1256 Some embedded systems have no use for them and therefore they
1257 can be configured out to reduce the size of the kernel image.
1259 When this option is disabled, the following syscalls won't be
1260 available: timer_create, timer_gettime: timer_getoverrun,
1261 timer_settime, timer_delete, clock_adjtime, getitimer,
1262 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1263 clock_getres and clock_nanosleep syscalls will be limited to
1264 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1270 bool "Enable support for printk" if EXPERT
1273 This option enables normal printk support. Removing it
1274 eliminates most of the message strings from the kernel image
1275 and makes the kernel more or less silent. As this makes it
1276 very difficult to diagnose system problems, saying N here is
1277 strongly discouraged.
1285 bool "BUG() support" if EXPERT
1288 Disabling this option eliminates support for BUG and WARN, reducing
1289 the size of your kernel image and potentially quietly ignoring
1290 numerous fatal conditions. You should only consider disabling this
1291 option for embedded systems with no facilities for reporting errors.
1297 bool "Enable ELF core dumps" if EXPERT
1299 Enable support for generating core dumps. Disabling saves about 4k.
1302 config PCSPKR_PLATFORM
1303 bool "Enable PC-Speaker support" if EXPERT
1304 depends on HAVE_PCSPKR_PLATFORM
1308 This option allows to disable the internal PC-Speaker
1309 support, saving some memory.
1313 bool "Enable full-sized data structures for core" if EXPERT
1315 Disabling this option reduces the size of miscellaneous core
1316 kernel data structures. This saves memory on small machines,
1317 but may reduce performance.
1320 bool "Enable futex support" if EXPERT
1324 Disabling this option will cause the kernel to be built without
1325 support for "fast userspace mutexes". The resulting kernel may not
1326 run glibc-based applications correctly.
1330 depends on FUTEX && RT_MUTEXES
1333 config HAVE_FUTEX_CMPXCHG
1337 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1338 is implemented and always working. This removes a couple of runtime
1342 bool "Enable eventpoll support" if EXPERT
1346 Disabling this option will cause the kernel to be built without
1347 support for epoll family of system calls.
1350 bool "Enable signalfd() system call" if EXPERT
1354 Enable the signalfd() system call that allows to receive signals
1355 on a file descriptor.
1360 bool "Enable timerfd() system call" if EXPERT
1364 Enable the timerfd() system call that allows to receive timer
1365 events on a file descriptor.
1370 bool "Enable eventfd() system call" if EXPERT
1374 Enable the eventfd() system call that allows to receive both
1375 kernel notification (ie. KAIO) or userspace notifications.
1380 bool "Use full shmem filesystem" if EXPERT
1384 The shmem is an internal filesystem used to manage shared memory.
1385 It is backed by swap and manages resource limits. It is also exported
1386 to userspace as tmpfs if TMPFS is enabled. Disabling this
1387 option replaces shmem and tmpfs with the much simpler ramfs code,
1388 which may be appropriate on small systems without swap.
1391 bool "Enable AIO support" if EXPERT
1394 This option enables POSIX asynchronous I/O which may by used
1395 by some high performance threaded applications. Disabling
1396 this option saves about 7k.
1398 config ADVISE_SYSCALLS
1399 bool "Enable madvise/fadvise syscalls" if EXPERT
1402 This option enables the madvise and fadvise syscalls, used by
1403 applications to advise the kernel about their future memory or file
1404 usage, improving performance. If building an embedded system where no
1405 applications use these syscalls, you can disable this option to save
1409 bool "Enable membarrier() system call" if EXPERT
1412 Enable the membarrier() system call that allows issuing memory
1413 barriers across all running threads, which can be used to distribute
1414 the cost of user-space memory barriers asymmetrically by transforming
1415 pairs of memory barriers into pairs consisting of membarrier() and a
1421 bool "Load all symbols for debugging/ksymoops" if EXPERT
1424 Say Y here to let the kernel print out symbolic crash information and
1425 symbolic stack backtraces. This increases the size of the kernel
1426 somewhat, as all symbols have to be loaded into the kernel image.
1429 bool "Include all symbols in kallsyms"
1430 depends on DEBUG_KERNEL && KALLSYMS
1432 Normally kallsyms only contains the symbols of functions for nicer
1433 OOPS messages and backtraces (i.e., symbols from the text and inittext
1434 sections). This is sufficient for most cases. And only in very rare
1435 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1436 names of variables from the data sections, etc).
1438 This option makes sure that all symbols are loaded into the kernel
1439 image (i.e., symbols from all sections) in cost of increased kernel
1440 size (depending on the kernel configuration, it may be 300KiB or
1441 something like this).
1443 Say N unless you really need all symbols.
1445 config KALLSYMS_ABSOLUTE_PERCPU
1448 default X86_64 && SMP
1450 config KALLSYMS_BASE_RELATIVE
1455 Instead of emitting them as absolute values in the native word size,
1456 emit the symbol references in the kallsyms table as 32-bit entries,
1457 each containing a relative value in the range [base, base + U32_MAX]
1458 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1459 an absolute value in the range [0, S32_MAX] or a relative value in the
1460 range [base, base + S32_MAX], where base is the lowest relative symbol
1461 address encountered in the image.
1463 On 64-bit builds, this reduces the size of the address table by 50%,
1464 but more importantly, it results in entries whose values are build
1465 time constants, and no relocation pass is required at runtime to fix
1466 up the entries based on the runtime load address of the kernel.
1468 # end of the "standard kernel features (expert users)" menu
1470 # syscall, maps, verifier
1472 bool "Enable bpf() system call"
1478 Enable the bpf() system call that allows to manipulate eBPF
1479 programs and maps via file descriptors.
1481 config BPF_JIT_ALWAYS_ON
1482 bool "Permanently enable BPF JIT and remove BPF interpreter"
1483 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1485 Enables BPF JIT and removes BPF interpreter to avoid
1486 speculative execution of BPF instructions by the interpreter
1489 bool "Enable userfaultfd() system call"
1493 Enable the userfaultfd() system call that allows to intercept and
1494 handle page faults in userland.
1496 config ARCH_HAS_MEMBARRIER_CALLBACKS
1499 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1503 bool "Enable rseq() system call" if EXPERT
1505 depends on HAVE_RSEQ
1508 Enable the restartable sequences system call. It provides a
1509 user-space cache for the current CPU number value, which
1510 speeds up getting the current CPU number from user-space,
1511 as well as an ABI to speed up user-space operations on
1518 bool "Enabled debugging of rseq() system call" if EXPERT
1519 depends on RSEQ && DEBUG_KERNEL
1521 Enable extra debugging checks for the rseq system call.
1526 bool "Embedded system"
1527 option allnoconfig_y
1530 This option should be enabled if compiling the kernel for
1531 an embedded system so certain expert options are available
1534 config HAVE_PERF_EVENTS
1537 See tools/perf/design.txt for details.
1539 config PERF_USE_VMALLOC
1542 See tools/perf/design.txt for details
1545 bool "PC/104 support" if EXPERT
1547 Expose PC/104 form factor device drivers and options available for
1548 selection and configuration. Enable this option if your target
1549 machine has a PC/104 bus.
1551 menu "Kernel Performance Events And Counters"
1554 bool "Kernel performance events and counters"
1555 default y if PROFILING
1556 depends on HAVE_PERF_EVENTS
1561 Enable kernel support for various performance events provided
1562 by software and hardware.
1564 Software events are supported either built-in or via the
1565 use of generic tracepoints.
1567 Most modern CPUs support performance events via performance
1568 counter registers. These registers count the number of certain
1569 types of hw events: such as instructions executed, cachemisses
1570 suffered, or branches mis-predicted - without slowing down the
1571 kernel or applications. These registers can also trigger interrupts
1572 when a threshold number of events have passed - and can thus be
1573 used to profile the code that runs on that CPU.
1575 The Linux Performance Event subsystem provides an abstraction of
1576 these software and hardware event capabilities, available via a
1577 system call and used by the "perf" utility in tools/perf/. It
1578 provides per task and per CPU counters, and it provides event
1579 capabilities on top of those.
1583 config DEBUG_PERF_USE_VMALLOC
1585 bool "Debug: use vmalloc to back perf mmap() buffers"
1586 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1587 select PERF_USE_VMALLOC
1589 Use vmalloc memory to back perf mmap() buffers.
1591 Mostly useful for debugging the vmalloc code on platforms
1592 that don't require it.
1598 config VM_EVENT_COUNTERS
1600 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1602 VM event counters are needed for event counts to be shown.
1603 This option allows the disabling of the VM event counters
1604 on EXPERT systems. /proc/vmstat will only show page counts
1605 if VM event counters are disabled.
1609 bool "Enable SLUB debugging support" if EXPERT
1610 depends on SLUB && SYSFS
1612 SLUB has extensive debug support features. Disabling these can
1613 result in significant savings in code size. This also disables
1614 SLUB sysfs support. /sys/slab will not exist and there will be
1615 no support for cache validation etc.
1617 config SLUB_MEMCG_SYSFS_ON
1619 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1620 depends on SLUB && SYSFS && MEMCG
1622 SLUB creates a directory under /sys/kernel/slab for each
1623 allocation cache to host info and debug files. If memory
1624 cgroup is enabled, each cache can have per memory cgroup
1625 caches. SLUB can create the same sysfs directories for these
1626 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1627 to a very high number of debug files being created. This is
1628 controlled by slub_memcg_sysfs boot parameter and this
1629 config option determines the parameter's default value.
1632 bool "Disable heap randomization"
1635 Randomizing heap placement makes heap exploits harder, but it
1636 also breaks ancient binaries (including anything libc5 based).
1637 This option changes the bootup default to heap randomization
1638 disabled, and can be overridden at runtime by setting
1639 /proc/sys/kernel/randomize_va_space to 2.
1641 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1644 prompt "Choose SLAB allocator"
1647 This option allows to select a slab allocator.
1651 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1653 The regular slab allocator that is established and known to work
1654 well in all environments. It organizes cache hot objects in
1655 per cpu and per node queues.
1658 bool "SLUB (Unqueued Allocator)"
1659 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1661 SLUB is a slab allocator that minimizes cache line usage
1662 instead of managing queues of cached objects (SLAB approach).
1663 Per cpu caching is realized using slabs of objects instead
1664 of queues of objects. SLUB can use memory efficiently
1665 and has enhanced diagnostics. SLUB is the default choice for
1670 bool "SLOB (Simple Allocator)"
1672 SLOB replaces the stock allocator with a drastically simpler
1673 allocator. SLOB is generally more space efficient but
1674 does not perform as well on large systems.
1678 config SLAB_MERGE_DEFAULT
1679 bool "Allow slab caches to be merged"
1682 For reduced kernel memory fragmentation, slab caches can be
1683 merged when they share the same size and other characteristics.
1684 This carries a risk of kernel heap overflows being able to
1685 overwrite objects from merged caches (and more easily control
1686 cache layout), which makes such heap attacks easier to exploit
1687 by attackers. By keeping caches unmerged, these kinds of exploits
1688 can usually only damage objects in the same cache. To disable
1689 merging at runtime, "slab_nomerge" can be passed on the kernel
1692 config SLAB_FREELIST_RANDOM
1694 depends on SLAB || SLUB
1695 bool "SLAB freelist randomization"
1697 Randomizes the freelist order used on creating new pages. This
1698 security feature reduces the predictability of the kernel slab
1699 allocator against heap overflows.
1701 config SLAB_FREELIST_HARDENED
1702 bool "Harden slab freelist metadata"
1705 Many kernel heap attacks try to target slab cache metadata and
1706 other infrastructure. This options makes minor performance
1707 sacrifies to harden the kernel slab allocator against common
1708 freelist exploit methods.
1710 config SLUB_CPU_PARTIAL
1712 depends on SLUB && SMP
1713 bool "SLUB per cpu partial cache"
1715 Per cpu partial caches accellerate objects allocation and freeing
1716 that is local to a processor at the price of more indeterminism
1717 in the latency of the free. On overflow these caches will be cleared
1718 which requires the taking of locks that may cause latency spikes.
1719 Typically one would choose no for a realtime system.
1721 config MMAP_ALLOW_UNINITIALIZED
1722 bool "Allow mmapped anonymous memory to be uninitialized"
1723 depends on EXPERT && !MMU
1726 Normally, and according to the Linux spec, anonymous memory obtained
1727 from mmap() has its contents cleared before it is passed to
1728 userspace. Enabling this config option allows you to request that
1729 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1730 providing a huge performance boost. If this option is not enabled,
1731 then the flag will be ignored.
1733 This is taken advantage of by uClibc's malloc(), and also by
1734 ELF-FDPIC binfmt's brk and stack allocator.
1736 Because of the obvious security issues, this option should only be
1737 enabled on embedded devices where you control what is run in
1738 userspace. Since that isn't generally a problem on no-MMU systems,
1739 it is normally safe to say Y here.
1741 See Documentation/nommu-mmap.txt for more information.
1743 config SYSTEM_DATA_VERIFICATION
1745 select SYSTEM_TRUSTED_KEYRING
1749 select ASYMMETRIC_KEY_TYPE
1750 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1753 select X509_CERTIFICATE_PARSER
1754 select PKCS7_MESSAGE_PARSER
1756 Provide PKCS#7 message verification using the contents of the system
1757 trusted keyring to provide public keys. This then can be used for
1758 module verification, kexec image verification and firmware blob
1762 bool "Profiling support"
1764 Say Y here to enable the extended profiling support mechanisms used
1765 by profilers such as OProfile.
1768 # Place an empty function call at each tracepoint site. Can be
1769 # dynamically changed for a probe function.
1774 endmenu # General setup
1776 source "arch/Kconfig"
1783 default 0 if BASE_FULL
1784 default 1 if !BASE_FULL
1787 bool "Enable loadable module support"
1790 Kernel modules are small pieces of compiled code which can
1791 be inserted in the running kernel, rather than being
1792 permanently built into the kernel. You use the "modprobe"
1793 tool to add (and sometimes remove) them. If you say Y here,
1794 many parts of the kernel can be built as modules (by
1795 answering M instead of Y where indicated): this is most
1796 useful for infrequently used options which are not required
1797 for booting. For more information, see the man pages for
1798 modprobe, lsmod, modinfo, insmod and rmmod.
1800 If you say Y here, you will need to run "make
1801 modules_install" to put the modules under /lib/modules/
1802 where modprobe can find them (you may need to be root to do
1809 config MODULE_FORCE_LOAD
1810 bool "Forced module loading"
1813 Allow loading of modules without version information (ie. modprobe
1814 --force). Forced module loading sets the 'F' (forced) taint flag and
1815 is usually a really bad idea.
1817 config MODULE_UNLOAD
1818 bool "Module unloading"
1820 Without this option you will not be able to unload any
1821 modules (note that some modules may not be unloadable
1822 anyway), which makes your kernel smaller, faster
1823 and simpler. If unsure, say Y.
1825 config MODULE_FORCE_UNLOAD
1826 bool "Forced module unloading"
1827 depends on MODULE_UNLOAD
1829 This option allows you to force a module to unload, even if the
1830 kernel believes it is unsafe: the kernel will remove the module
1831 without waiting for anyone to stop using it (using the -f option to
1832 rmmod). This is mainly for kernel developers and desperate users.
1836 bool "Module versioning support"
1838 Usually, you have to use modules compiled with your kernel.
1839 Saying Y here makes it sometimes possible to use modules
1840 compiled for different kernels, by adding enough information
1841 to the modules to (hopefully) spot any changes which would
1842 make them incompatible with the kernel you are running. If
1845 config MODULE_REL_CRCS
1847 depends on MODVERSIONS
1849 config MODULE_SRCVERSION_ALL
1850 bool "Source checksum for all modules"
1852 Modules which contain a MODULE_VERSION get an extra "srcversion"
1853 field inserted into their modinfo section, which contains a
1854 sum of the source files which made it. This helps maintainers
1855 see exactly which source was used to build a module (since
1856 others sometimes change the module source without updating
1857 the version). With this option, such a "srcversion" field
1858 will be created for all modules. If unsure, say N.
1861 bool "Module signature verification"
1863 select SYSTEM_DATA_VERIFICATION
1865 Check modules for valid signatures upon load: the signature
1866 is simply appended to the module. For more information see
1867 <file:Documentation/admin-guide/module-signing.rst>.
1869 Note that this option adds the OpenSSL development packages as a
1870 kernel build dependency so that the signing tool can use its crypto
1873 !!!WARNING!!! If you enable this option, you MUST make sure that the
1874 module DOES NOT get stripped after being signed. This includes the
1875 debuginfo strip done by some packagers (such as rpmbuild) and
1876 inclusion into an initramfs that wants the module size reduced.
1878 config MODULE_SIG_FORCE
1879 bool "Require modules to be validly signed"
1880 depends on MODULE_SIG
1882 Reject unsigned modules or signed modules for which we don't have a
1883 key. Without this, such modules will simply taint the kernel.
1885 config MODULE_SIG_ALL
1886 bool "Automatically sign all modules"
1888 depends on MODULE_SIG
1890 Sign all modules during make modules_install. Without this option,
1891 modules must be signed manually, using the scripts/sign-file tool.
1893 comment "Do not forget to sign required modules with scripts/sign-file"
1894 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
1897 prompt "Which hash algorithm should modules be signed with?"
1898 depends on MODULE_SIG
1900 This determines which sort of hashing algorithm will be used during
1901 signature generation. This algorithm _must_ be built into the kernel
1902 directly so that signature verification can take place. It is not
1903 possible to load a signed module containing the algorithm to check
1904 the signature on that module.
1906 config MODULE_SIG_SHA1
1907 bool "Sign modules with SHA-1"
1910 config MODULE_SIG_SHA224
1911 bool "Sign modules with SHA-224"
1912 select CRYPTO_SHA256
1914 config MODULE_SIG_SHA256
1915 bool "Sign modules with SHA-256"
1916 select CRYPTO_SHA256
1918 config MODULE_SIG_SHA384
1919 bool "Sign modules with SHA-384"
1920 select CRYPTO_SHA512
1922 config MODULE_SIG_SHA512
1923 bool "Sign modules with SHA-512"
1924 select CRYPTO_SHA512
1928 config MODULE_SIG_HASH
1930 depends on MODULE_SIG
1931 default "sha1" if MODULE_SIG_SHA1
1932 default "sha224" if MODULE_SIG_SHA224
1933 default "sha256" if MODULE_SIG_SHA256
1934 default "sha384" if MODULE_SIG_SHA384
1935 default "sha512" if MODULE_SIG_SHA512
1937 config MODULE_COMPRESS
1938 bool "Compress modules on installation"
1942 Compresses kernel modules when 'make modules_install' is run; gzip or
1943 xz depending on "Compression algorithm" below.
1945 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
1947 Out-of-tree kernel modules installed using Kbuild will also be
1948 compressed upon installation.
1950 Note: for modules inside an initrd or initramfs, it's more efficient
1951 to compress the whole initrd or initramfs instead.
1953 Note: This is fully compatible with signed modules.
1958 prompt "Compression algorithm"
1959 depends on MODULE_COMPRESS
1960 default MODULE_COMPRESS_GZIP
1962 This determines which sort of compression will be used during
1963 'make modules_install'.
1965 GZIP (default) and XZ are supported.
1967 config MODULE_COMPRESS_GZIP
1970 config MODULE_COMPRESS_XZ
1975 config TRIM_UNUSED_KSYMS
1976 bool "Trim unused exported kernel symbols"
1977 depends on MODULES && !UNUSED_SYMBOLS
1979 The kernel and some modules make many symbols available for
1980 other modules to use via EXPORT_SYMBOL() and variants. Depending
1981 on the set of modules being selected in your kernel configuration,
1982 many of those exported symbols might never be used.
1984 This option allows for unused exported symbols to be dropped from
1985 the build. In turn, this provides the compiler more opportunities
1986 (especially when using LTO) for optimizing the code and reducing
1987 binary size. This might have some security advantages as well.
1989 If unsure, or if you need to build out-of-tree modules, say N.
1993 config MODULES_TREE_LOOKUP
1995 depends on PERF_EVENTS || TRACING
1997 config INIT_ALL_POSSIBLE
2000 Back when each arch used to define their own cpu_online_mask and
2001 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2002 with all 1s, and others with all 0s. When they were centralised,
2003 it was better to provide this option than to break all the archs
2004 and have several arch maintainers pursuing me down dark alleys.
2006 source "block/Kconfig"
2008 config PREEMPT_NOTIFIERS
2018 Build a simple ASN.1 grammar compiler that produces a bytecode output
2019 that can be interpreted by the ASN.1 stream decoder and used to
2020 inform it as to what tags are to be expected in a stream and what
2021 functions to call on what tags.
2023 source "kernel/Kconfig.locks"
2025 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2028 # It may be useful for an architecture to override the definitions of the
2029 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2030 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2031 # different calling convention for syscalls. They can also override the
2032 # macros for not-implemented syscalls in kernel/sys_ni.c and
2033 # kernel/time/posix-stubs.c. All these overrides need to be available in
2034 # <asm/syscall_wrapper.h>.
2035 config ARCH_HAS_SYSCALL_WRAPPER