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)
23 def_bool $(success,$(LD) -v | head -n 1 | grep -q LLD)
27 default $(shell,$(srctree)/scripts/clang-version.sh $(CC))
29 config CC_HAS_ASM_GOTO
30 def_bool $(success,$(srctree)/scripts/gcc-goto.sh $(CC))
39 config BUILDTIME_EXTABLE_SORT
42 config THREAD_INFO_IN_TASK
45 Select this to move thread_info off the stack into task_struct. To
46 make this work, an arch will need to remove all thread_info fields
47 except flags and fix any runtime bugs.
49 One subtle change that will be needed is to use try_get_task_stack()
50 and put_task_stack() in save_thread_stack_tsk() and get_wchan().
59 depends on BROKEN || !SMP
62 config INIT_ENV_ARG_LIMIT
67 Maximum of each of the number of arguments and environment
68 variables passed to init from the kernel command line.
71 bool "Compile also drivers which will not load"
74 Some drivers can be compiled on a different platform than they are
75 intended to be run on. Despite they cannot be loaded there (or even
76 when they load they cannot be used due to missing HW support),
77 developers still, opposing to distributors, might want to build such
78 drivers to compile-test them.
80 If you are a developer and want to build everything available, say Y
81 here. If you are a user/distributor, say N here to exclude useless
82 drivers to be distributed.
85 string "Local version - append to kernel release"
87 Append an extra string to the end of your kernel version.
88 This will show up when you type uname, for example.
89 The string you set here will be appended after the contents of
90 any files with a filename matching localversion* in your
91 object and source tree, in that order. Your total string can
92 be a maximum of 64 characters.
94 config LOCALVERSION_AUTO
95 bool "Automatically append version information to the version string"
97 depends on !COMPILE_TEST
99 This will try to automatically determine if the current tree is a
100 release tree by looking for git tags that belong to the current
101 top of tree revision.
103 A string of the format -gxxxxxxxx will be added to the localversion
104 if a git-based tree is found. The string generated by this will be
105 appended after any matching localversion* files, and after the value
106 set in CONFIG_LOCALVERSION.
108 (The actual string used here is the first eight characters produced
109 by running the command:
111 $ git rev-parse --verify HEAD
113 which is done within the script "scripts/setlocalversion".)
116 string "Build ID Salt"
119 The build ID is used to link binaries and their debug info. Setting
120 this option will use the value in the calculation of the build id.
121 This is mostly useful for distributions which want to ensure the
122 build is unique between builds. It's safe to leave the default.
124 config HAVE_KERNEL_GZIP
127 config HAVE_KERNEL_BZIP2
130 config HAVE_KERNEL_LZMA
133 config HAVE_KERNEL_XZ
136 config HAVE_KERNEL_LZO
139 config HAVE_KERNEL_LZ4
142 config HAVE_KERNEL_UNCOMPRESSED
146 prompt "Kernel compression mode"
148 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4 || HAVE_KERNEL_UNCOMPRESSED
150 The linux kernel is a kind of self-extracting executable.
151 Several compression algorithms are available, which differ
152 in efficiency, compression and decompression speed.
153 Compression speed is only relevant when building a kernel.
154 Decompression speed is relevant at each boot.
156 If you have any problems with bzip2 or lzma compressed
157 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
158 version of this functionality (bzip2 only), for 2.4, was
159 supplied by Christian Ludwig)
161 High compression options are mostly useful for users, who
162 are low on disk space (embedded systems), but for whom ram
165 If in doubt, select 'gzip'
169 depends on HAVE_KERNEL_GZIP
171 The old and tried gzip compression. It provides a good balance
172 between compression ratio and decompression speed.
176 depends on HAVE_KERNEL_BZIP2
178 Its compression ratio and speed is intermediate.
179 Decompression speed is slowest among the choices. The kernel
180 size is about 10% smaller with bzip2, in comparison to gzip.
181 Bzip2 uses a large amount of memory. For modern kernels you
182 will need at least 8MB RAM or more for booting.
186 depends on HAVE_KERNEL_LZMA
188 This compression algorithm's ratio is best. Decompression speed
189 is between gzip and bzip2. Compression is slowest.
190 The kernel size is about 33% smaller with LZMA in comparison to gzip.
194 depends on HAVE_KERNEL_XZ
196 XZ uses the LZMA2 algorithm and instruction set specific
197 BCJ filters which can improve compression ratio of executable
198 code. The size of the kernel is about 30% smaller with XZ in
199 comparison to gzip. On architectures for which there is a BCJ
200 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
201 will create a few percent smaller kernel than plain LZMA.
203 The speed is about the same as with LZMA: The decompression
204 speed of XZ is better than that of bzip2 but worse than gzip
205 and LZO. Compression is slow.
209 depends on HAVE_KERNEL_LZO
211 Its compression ratio is the poorest among the choices. The kernel
212 size is about 10% bigger than gzip; however its speed
213 (both compression and decompression) is the fastest.
217 depends on HAVE_KERNEL_LZ4
219 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
220 A preliminary version of LZ4 de/compression tool is available at
221 <https://code.google.com/p/lz4/>.
223 Its compression ratio is worse than LZO. The size of the kernel
224 is about 8% bigger than LZO. But the decompression speed is
227 config KERNEL_UNCOMPRESSED
229 depends on HAVE_KERNEL_UNCOMPRESSED
231 Produce uncompressed kernel image. This option is usually not what
232 you want. It is useful for debugging the kernel in slow simulation
233 environments, where decompressing and moving the kernel is awfully
234 slow. This option allows early boot code to skip the decompressor
235 and jump right at uncompressed kernel image.
239 config DEFAULT_HOSTNAME
240 string "Default hostname"
243 This option determines the default system hostname before userspace
244 calls sethostname(2). The kernel traditionally uses "(none)" here,
245 but you may wish to use a different default here to make a minimal
246 system more usable with less configuration.
249 # For some reason microblaze and nios2 hard code SWAP=n. Hopefully we can
250 # add proper SWAP support to them, in which case this can be remove.
256 bool "Support for paging of anonymous memory (swap)"
257 depends on MMU && BLOCK && !ARCH_NO_SWAP
260 This option allows you to choose whether you want to have support
261 for so called swap devices or swap files in your kernel that are
262 used to provide more virtual memory than the actual RAM present
263 in your computer. If unsure say Y.
268 Inter Process Communication is a suite of library functions and
269 system calls which let processes (running programs) synchronize and
270 exchange information. It is generally considered to be a good thing,
271 and some programs won't run unless you say Y here. In particular, if
272 you want to run the DOS emulator dosemu under Linux (read the
273 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
274 you'll need to say Y here.
276 You can find documentation about IPC with "info ipc" and also in
277 section 6.4 of the Linux Programmer's Guide, available from
278 <http://www.tldp.org/guides.html>.
280 config SYSVIPC_SYSCTL
287 bool "POSIX Message Queues"
290 POSIX variant of message queues is a part of IPC. In POSIX message
291 queues every message has a priority which decides about succession
292 of receiving it by a process. If you want to compile and run
293 programs written e.g. for Solaris with use of its POSIX message
294 queues (functions mq_*) say Y here.
296 POSIX message queues are visible as a filesystem called 'mqueue'
297 and can be mounted somewhere if you want to do filesystem
298 operations on message queues.
302 config POSIX_MQUEUE_SYSCTL
304 depends on POSIX_MQUEUE
308 config CROSS_MEMORY_ATTACH
309 bool "Enable process_vm_readv/writev syscalls"
313 Enabling this option adds the system calls process_vm_readv and
314 process_vm_writev which allow a process with the correct privileges
315 to directly read from or write to another process' address space.
316 See the man page for more details.
319 bool "uselib syscall"
320 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
322 This option enables the uselib syscall, a system call used in the
323 dynamic linker from libc5 and earlier. glibc does not use this
324 system call. If you intend to run programs built on libc5 or
325 earlier, you may need to enable this syscall. Current systems
326 running glibc can safely disable this.
329 bool "Auditing support"
332 Enable auditing infrastructure that can be used with another
333 kernel subsystem, such as SELinux (which requires this for
334 logging of avc messages output). System call auditing is included
335 on architectures which support it.
337 config HAVE_ARCH_AUDITSYSCALL
342 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
346 depends on AUDITSYSCALL
351 depends on AUDITSYSCALL
354 source "kernel/irq/Kconfig"
355 source "kernel/time/Kconfig"
356 source "kernel/Kconfig.preempt"
358 menu "CPU/Task time and stats accounting"
360 config VIRT_CPU_ACCOUNTING
364 prompt "Cputime accounting"
365 default TICK_CPU_ACCOUNTING if !PPC64
366 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
368 # Kind of a stub config for the pure tick based cputime accounting
369 config TICK_CPU_ACCOUNTING
370 bool "Simple tick based cputime accounting"
371 depends on !S390 && !NO_HZ_FULL
373 This is the basic tick based cputime accounting that maintains
374 statistics about user, system and idle time spent on per jiffies
379 config VIRT_CPU_ACCOUNTING_NATIVE
380 bool "Deterministic task and CPU time accounting"
381 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
382 select VIRT_CPU_ACCOUNTING
384 Select this option to enable more accurate task and CPU time
385 accounting. This is done by reading a CPU counter on each
386 kernel entry and exit and on transitions within the kernel
387 between system, softirq and hardirq state, so there is a
388 small performance impact. In the case of s390 or IBM POWER > 5,
389 this also enables accounting of stolen time on logically-partitioned
392 config VIRT_CPU_ACCOUNTING_GEN
393 bool "Full dynticks CPU time accounting"
394 depends on HAVE_CONTEXT_TRACKING
395 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
396 select VIRT_CPU_ACCOUNTING
397 select CONTEXT_TRACKING
399 Select this option to enable task and CPU time accounting on full
400 dynticks systems. This accounting is implemented by watching every
401 kernel-user boundaries using the context tracking subsystem.
402 The accounting is thus performed at the expense of some significant
405 For now this is only useful if you are working on the full
406 dynticks subsystem development.
412 config IRQ_TIME_ACCOUNTING
413 bool "Fine granularity task level IRQ time accounting"
414 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
416 Select this option to enable fine granularity task irq time
417 accounting. This is done by reading a timestamp on each
418 transitions between softirq and hardirq state, so there can be a
419 small performance impact.
421 If in doubt, say N here.
423 config HAVE_SCHED_AVG_IRQ
425 depends on IRQ_TIME_ACCOUNTING || PARAVIRT_TIME_ACCOUNTING
428 config BSD_PROCESS_ACCT
429 bool "BSD Process Accounting"
432 If you say Y here, a user level program will be able to instruct the
433 kernel (via a special system call) to write process accounting
434 information to a file: whenever a process exits, information about
435 that process will be appended to the file by the kernel. The
436 information includes things such as creation time, owning user,
437 command name, memory usage, controlling terminal etc. (the complete
438 list is in the struct acct in <file:include/linux/acct.h>). It is
439 up to the user level program to do useful things with this
440 information. This is generally a good idea, so say Y.
442 config BSD_PROCESS_ACCT_V3
443 bool "BSD Process Accounting version 3 file format"
444 depends on BSD_PROCESS_ACCT
447 If you say Y here, the process accounting information is written
448 in a new file format that also logs the process IDs of each
449 process and its parent. Note that this file format is incompatible
450 with previous v0/v1/v2 file formats, so you will need updated tools
451 for processing it. A preliminary version of these tools is available
452 at <http://www.gnu.org/software/acct/>.
455 bool "Export task/process statistics through netlink"
460 Export selected statistics for tasks/processes through the
461 generic netlink interface. Unlike BSD process accounting, the
462 statistics are available during the lifetime of tasks/processes as
463 responses to commands. Like BSD accounting, they are sent to user
468 config TASK_DELAY_ACCT
469 bool "Enable per-task delay accounting"
473 Collect information on time spent by a task waiting for system
474 resources like cpu, synchronous block I/O completion and swapping
475 in pages. Such statistics can help in setting a task's priorities
476 relative to other tasks for cpu, io, rss limits etc.
481 bool "Enable extended accounting over taskstats"
484 Collect extended task accounting data and send the data
485 to userland for processing over the taskstats interface.
489 config TASK_IO_ACCOUNTING
490 bool "Enable per-task storage I/O accounting"
491 depends on TASK_XACCT
493 Collect information on the number of bytes of storage I/O which this
498 endmenu # "CPU/Task time and stats accounting"
502 depends on SMP || COMPILE_TEST
505 Make sure that CPUs running critical tasks are not disturbed by
506 any source of "noise" such as unbound workqueues, timers, kthreads...
507 Unbound jobs get offloaded to housekeeping CPUs. This is driven by
508 the "isolcpus=" boot parameter.
512 source "kernel/rcu/Kconfig"
519 tristate "Kernel .config support"
522 This option enables the complete Linux kernel ".config" file
523 contents to be saved in the kernel. It provides documentation
524 of which kernel options are used in a running kernel or in an
525 on-disk kernel. This information can be extracted from the kernel
526 image file with the script scripts/extract-ikconfig and used as
527 input to rebuild the current kernel or to build another kernel.
528 It can also be extracted from a running kernel by reading
529 /proc/config.gz if enabled (below).
532 bool "Enable access to .config through /proc/config.gz"
533 depends on IKCONFIG && PROC_FS
535 This option enables access to the kernel configuration file
536 through /proc/config.gz.
539 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
540 range 12 25 if !H8300
545 Select the minimal kernel log buffer size as a power of 2.
546 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
547 parameter, see below. Any higher size also might be forced
548 by "log_buf_len" boot parameter.
558 config LOG_CPU_MAX_BUF_SHIFT
559 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
562 default 12 if !BASE_SMALL
563 default 0 if BASE_SMALL
566 This option allows to increase the default ring buffer size
567 according to the number of CPUs. The value defines the contribution
568 of each CPU as a power of 2. The used space is typically only few
569 lines however it might be much more when problems are reported,
572 The increased size means that a new buffer has to be allocated and
573 the original static one is unused. It makes sense only on systems
574 with more CPUs. Therefore this value is used only when the sum of
575 contributions is greater than the half of the default kernel ring
576 buffer as defined by LOG_BUF_SHIFT. The default values are set
577 so that more than 64 CPUs are needed to trigger the allocation.
579 Also this option is ignored when "log_buf_len" kernel parameter is
580 used as it forces an exact (power of two) size of the ring buffer.
582 The number of possible CPUs is used for this computation ignoring
583 hotplugging making the computation optimal for the worst case
584 scenario while allowing a simple algorithm to be used from bootup.
586 Examples shift values and their meaning:
587 17 => 128 KB for each CPU
588 16 => 64 KB for each CPU
589 15 => 32 KB for each CPU
590 14 => 16 KB for each CPU
591 13 => 8 KB for each CPU
592 12 => 4 KB for each CPU
594 config PRINTK_SAFE_LOG_BUF_SHIFT
595 int "Temporary per-CPU printk log buffer size (12 => 4KB, 13 => 8KB)"
600 Select the size of an alternate printk per-CPU buffer where messages
601 printed from usafe contexts are temporary stored. One example would
602 be NMI messages, another one - printk recursion. The messages are
603 copied to the main log buffer in a safe context to avoid a deadlock.
604 The value defines the size as a power of 2.
606 Those messages are rare and limited. The largest one is when
607 a backtrace is printed. It usually fits into 4KB. Select
608 8KB if you want to be on the safe side.
611 17 => 128 KB for each CPU
612 16 => 64 KB for each CPU
613 15 => 32 KB for each CPU
614 14 => 16 KB for each CPU
615 13 => 8 KB for each CPU
616 12 => 4 KB for each CPU
619 # Architectures with an unreliable sched_clock() should select this:
621 config HAVE_UNSTABLE_SCHED_CLOCK
624 config GENERIC_SCHED_CLOCK
628 # For architectures that want to enable the support for NUMA-affine scheduler
631 config ARCH_SUPPORTS_NUMA_BALANCING
635 # For architectures that prefer to flush all TLBs after a number of pages
636 # are unmapped instead of sending one IPI per page to flush. The architecture
637 # must provide guarantees on what happens if a clean TLB cache entry is
638 # written after the unmap. Details are in mm/rmap.c near the check for
639 # should_defer_flush. The architecture should also consider if the full flush
640 # and the refill costs are offset by the savings of sending fewer IPIs.
641 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
645 # For architectures that know their GCC __int128 support is sound
647 config ARCH_SUPPORTS_INT128
650 # For architectures that (ab)use NUMA to represent different memory regions
651 # all cpu-local but of different latencies, such as SuperH.
653 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
656 config NUMA_BALANCING
657 bool "Memory placement aware NUMA scheduler"
658 depends on ARCH_SUPPORTS_NUMA_BALANCING
659 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
660 depends on SMP && NUMA && MIGRATION
662 This option adds support for automatic NUMA aware memory/task placement.
663 The mechanism is quite primitive and is based on migrating memory when
664 it has references to the node the task is running on.
666 This system will be inactive on UMA systems.
668 config NUMA_BALANCING_DEFAULT_ENABLED
669 bool "Automatically enable NUMA aware memory/task placement"
671 depends on NUMA_BALANCING
673 If set, automatic NUMA balancing will be enabled if running on a NUMA
677 bool "Control Group support"
680 This option adds support for grouping sets of processes together, for
681 use with process control subsystems such as Cpusets, CFS, memory
682 controls or device isolation.
684 - Documentation/scheduler/sched-design-CFS.txt (CFS)
685 - Documentation/cgroup-v1/ (features for grouping, isolation
686 and resource control)
696 bool "Memory controller"
700 Provides control over the memory footprint of tasks in a cgroup.
703 bool "Swap controller"
704 depends on MEMCG && SWAP
706 Provides control over the swap space consumed by tasks in a cgroup.
708 config MEMCG_SWAP_ENABLED
709 bool "Swap controller enabled by default"
710 depends on MEMCG_SWAP
713 Memory Resource Controller Swap Extension comes with its price in
714 a bigger memory consumption. General purpose distribution kernels
715 which want to enable the feature but keep it disabled by default
716 and let the user enable it by swapaccount=1 boot command line
717 parameter should have this option unselected.
718 For those who want to have the feature enabled by default should
719 select this option (if, for some reason, they need to disable it
720 then swapaccount=0 does the trick).
724 depends on MEMCG && !SLOB
732 Generic block IO controller cgroup interface. This is the common
733 cgroup interface which should be used by various IO controlling
736 Currently, CFQ IO scheduler uses it to recognize task groups and
737 control disk bandwidth allocation (proportional time slice allocation)
738 to such task groups. It is also used by bio throttling logic in
739 block layer to implement upper limit in IO rates on a device.
741 This option only enables generic Block IO controller infrastructure.
742 One needs to also enable actual IO controlling logic/policy. For
743 enabling proportional weight division of disk bandwidth in CFQ, set
744 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
745 CONFIG_BLK_DEV_THROTTLING=y.
747 See Documentation/cgroup-v1/blkio-controller.txt for more information.
749 config DEBUG_BLK_CGROUP
750 bool "IO controller debugging"
751 depends on BLK_CGROUP
754 Enable some debugging help. Currently it exports additional stat
755 files in a cgroup which can be useful for debugging.
757 config CGROUP_WRITEBACK
759 depends on MEMCG && BLK_CGROUP
762 menuconfig CGROUP_SCHED
763 bool "CPU controller"
766 This feature lets CPU scheduler recognize task groups and control CPU
767 bandwidth allocation to such task groups. It uses cgroups to group
771 config FAIR_GROUP_SCHED
772 bool "Group scheduling for SCHED_OTHER"
773 depends on CGROUP_SCHED
777 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
778 depends on FAIR_GROUP_SCHED
781 This option allows users to define CPU bandwidth rates (limits) for
782 tasks running within the fair group scheduler. Groups with no limit
783 set are considered to be unconstrained and will run with no
785 See Documentation/scheduler/sched-bwc.txt for more information.
787 config RT_GROUP_SCHED
788 bool "Group scheduling for SCHED_RR/FIFO"
789 depends on CGROUP_SCHED
792 This feature lets you explicitly allocate real CPU bandwidth
793 to task groups. If enabled, it will also make it impossible to
794 schedule realtime tasks for non-root users until you allocate
795 realtime bandwidth for them.
796 See Documentation/scheduler/sched-rt-group.txt for more information.
801 bool "PIDs controller"
803 Provides enforcement of process number limits in the scope of a
804 cgroup. Any attempt to fork more processes than is allowed in the
805 cgroup will fail. PIDs are fundamentally a global resource because it
806 is fairly trivial to reach PID exhaustion before you reach even a
807 conservative kmemcg limit. As a result, it is possible to grind a
808 system to halt without being limited by other cgroup policies. The
809 PIDs controller is designed to stop this from happening.
811 It should be noted that organisational operations (such as attaching
812 to a cgroup hierarchy will *not* be blocked by the PIDs controller),
813 since the PIDs limit only affects a process's ability to fork, not to
817 bool "RDMA controller"
819 Provides enforcement of RDMA resources defined by IB stack.
820 It is fairly easy for consumers to exhaust RDMA resources, which
821 can result into resource unavailability to other consumers.
822 RDMA controller is designed to stop this from happening.
823 Attaching processes with active RDMA resources to the cgroup
824 hierarchy is allowed even if can cross the hierarchy's limit.
826 config CGROUP_FREEZER
827 bool "Freezer controller"
829 Provides a way to freeze and unfreeze all tasks in a
832 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
833 controller includes important in-kernel memory consumers per default.
835 If you're using cgroup2, say N.
837 config CGROUP_HUGETLB
838 bool "HugeTLB controller"
839 depends on HUGETLB_PAGE
843 Provides a cgroup controller for HugeTLB pages.
844 When you enable this, you can put a per cgroup limit on HugeTLB usage.
845 The limit is enforced during page fault. Since HugeTLB doesn't
846 support page reclaim, enforcing the limit at page fault time implies
847 that, the application will get SIGBUS signal if it tries to access
848 HugeTLB pages beyond its limit. This requires the application to know
849 beforehand how much HugeTLB pages it would require for its use. The
850 control group is tracked in the third page lru pointer. This means
851 that we cannot use the controller with huge page less than 3 pages.
854 bool "Cpuset controller"
857 This option will let you create and manage CPUSETs which
858 allow dynamically partitioning a system into sets of CPUs and
859 Memory Nodes and assigning tasks to run only within those sets.
860 This is primarily useful on large SMP or NUMA systems.
864 config PROC_PID_CPUSET
865 bool "Include legacy /proc/<pid>/cpuset file"
870 bool "Device controller"
872 Provides a cgroup controller implementing whitelists for
873 devices which a process in the cgroup can mknod or open.
875 config CGROUP_CPUACCT
876 bool "Simple CPU accounting controller"
878 Provides a simple controller for monitoring the
879 total CPU consumed by the tasks in a cgroup.
882 bool "Perf controller"
883 depends on PERF_EVENTS
885 This option extends the perf per-cpu mode to restrict monitoring
886 to threads which belong to the cgroup specified and run on the
892 bool "Support for eBPF programs attached to cgroups"
893 depends on BPF_SYSCALL
894 select SOCK_CGROUP_DATA
896 Allow attaching eBPF programs to a cgroup using the bpf(2)
897 syscall command BPF_PROG_ATTACH.
899 In which context these programs are accessed depends on the type
900 of attachment. For instance, programs that are attached using
901 BPF_CGROUP_INET_INGRESS will be executed on the ingress path of
905 bool "Debug controller"
907 depends on DEBUG_KERNEL
909 This option enables a simple controller that exports
910 debugging information about the cgroups framework. This
911 controller is for control cgroup debugging only. Its
912 interfaces are not stable.
916 config SOCK_CGROUP_DATA
922 menuconfig NAMESPACES
923 bool "Namespaces support" if EXPERT
927 Provides the way to make tasks work with different objects using
928 the same id. For example same IPC id may refer to different objects
929 or same user id or pid may refer to different tasks when used in
930 different namespaces.
938 In this namespace tasks see different info provided with the
943 depends on (SYSVIPC || POSIX_MQUEUE)
946 In this namespace tasks work with IPC ids which correspond to
947 different IPC objects in different namespaces.
950 bool "User namespace"
953 This allows containers, i.e. vservers, to use user namespaces
954 to provide different user info for different servers.
956 When user namespaces are enabled in the kernel it is
957 recommended that the MEMCG option also be enabled and that
958 user-space use the memory control groups to limit the amount
959 of memory a memory unprivileged users can use.
964 bool "PID Namespaces"
967 Support process id namespaces. This allows having multiple
968 processes with the same pid as long as they are in different
969 pid namespaces. This is a building block of containers.
972 bool "Network namespace"
976 Allow user space to create what appear to be multiple instances
977 of the network stack.
981 config CHECKPOINT_RESTORE
982 bool "Checkpoint/restore support"
986 Enables additional kernel features in a sake of checkpoint/restore.
987 In particular it adds auxiliary prctl codes to setup process text,
988 data and heap segment sizes, and a few additional /proc filesystem
991 If unsure, say N here.
993 config SCHED_AUTOGROUP
994 bool "Automatic process group scheduling"
997 select FAIR_GROUP_SCHED
999 This option optimizes the scheduler for common desktop workloads by
1000 automatically creating and populating task groups. This separation
1001 of workloads isolates aggressive CPU burners (like build jobs) from
1002 desktop applications. Task group autogeneration is currently based
1005 config SYSFS_DEPRECATED
1006 bool "Enable deprecated sysfs features to support old userspace tools"
1010 This option adds code that switches the layout of the "block" class
1011 devices, to not show up in /sys/class/block/, but only in
1014 This switch is only active when the sysfs.deprecated=1 boot option is
1015 passed or the SYSFS_DEPRECATED_V2 option is set.
1017 This option allows new kernels to run on old distributions and tools,
1018 which might get confused by /sys/class/block/. Since 2007/2008 all
1019 major distributions and tools handle this just fine.
1021 Recent distributions and userspace tools after 2009/2010 depend on
1022 the existence of /sys/class/block/, and will not work with this
1025 Only if you are using a new kernel on an old distribution, you might
1028 config SYSFS_DEPRECATED_V2
1029 bool "Enable deprecated sysfs features by default"
1032 depends on SYSFS_DEPRECATED
1034 Enable deprecated sysfs by default.
1036 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1039 Only if you are using a new kernel on an old distribution, you might
1040 need to say Y here. Even then, odds are you would not need it
1041 enabled, you can always pass the boot option if absolutely necessary.
1044 bool "Kernel->user space relay support (formerly relayfs)"
1047 This option enables support for relay interface support in
1048 certain file systems (such as debugfs).
1049 It is designed to provide an efficient mechanism for tools and
1050 facilities to relay large amounts of data from kernel space to
1055 config BLK_DEV_INITRD
1056 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1058 The initial RAM filesystem is a ramfs which is loaded by the
1059 boot loader (loadlin or lilo) and that is mounted as root
1060 before the normal boot procedure. It is typically used to
1061 load modules needed to mount the "real" root file system,
1062 etc. See <file:Documentation/admin-guide/initrd.rst> for details.
1064 If RAM disk support (BLK_DEV_RAM) is also included, this
1065 also enables initial RAM disk (initrd) support and adds
1066 15 Kbytes (more on some other architectures) to the kernel size.
1072 source "usr/Kconfig"
1077 prompt "Compiler optimization level"
1078 default CC_OPTIMIZE_FOR_PERFORMANCE
1080 config CC_OPTIMIZE_FOR_PERFORMANCE
1081 bool "Optimize for performance"
1083 This is the default optimization level for the kernel, building
1084 with the "-O2" compiler flag for best performance and most
1085 helpful compile-time warnings.
1087 config CC_OPTIMIZE_FOR_SIZE
1088 bool "Optimize for size"
1090 Enabling this option will pass "-Os" instead of "-O2" to
1091 your compiler resulting in a smaller kernel.
1097 config HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1100 This requires that the arch annotates or otherwise protects
1101 its external entry points from being discarded. Linker scripts
1102 must also merge .text.*, .data.*, and .bss.* correctly into
1103 output sections. Care must be taken not to pull in unrelated
1104 sections (e.g., '.text.init'). Typically '.' in section names
1105 is used to distinguish them from label names / C identifiers.
1107 config LD_DEAD_CODE_DATA_ELIMINATION
1108 bool "Dead code and data elimination (EXPERIMENTAL)"
1109 depends on HAVE_LD_DEAD_CODE_DATA_ELIMINATION
1111 depends on !(FUNCTION_TRACER && CC_IS_GCC && GCC_VERSION < 40800)
1112 depends on $(cc-option,-ffunction-sections -fdata-sections)
1113 depends on $(ld-option,--gc-sections)
1115 Enable this if you want to do dead code and data elimination with
1116 the linker by compiling with -ffunction-sections -fdata-sections,
1117 and linking with --gc-sections.
1119 This can reduce on disk and in-memory size of the kernel
1120 code and static data, particularly for small configs and
1121 on small systems. This has the possibility of introducing
1122 silently broken kernel if the required annotations are not
1123 present. This option is not well tested yet, so use at your
1135 config SYSCTL_EXCEPTION_TRACE
1138 Enable support for /proc/sys/debug/exception-trace.
1140 config SYSCTL_ARCH_UNALIGN_NO_WARN
1143 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1144 Allows arch to define/use @no_unaligned_warning to possibly warn
1145 about unaligned access emulation going on under the hood.
1147 config SYSCTL_ARCH_UNALIGN_ALLOW
1150 Enable support for /proc/sys/kernel/unaligned-trap
1151 Allows arches to define/use @unaligned_enabled to runtime toggle
1152 the unaligned access emulation.
1153 see arch/parisc/kernel/unaligned.c for reference
1155 config HAVE_PCSPKR_PLATFORM
1158 # interpreter that classic socket filters depend on
1163 bool "Configure standard kernel features (expert users)"
1164 # Unhide debug options, to make the on-by-default options visible
1167 This option allows certain base kernel options and settings
1168 to be disabled or tweaked. This is for specialized
1169 environments which can tolerate a "non-standard" kernel.
1170 Only use this if you really know what you are doing.
1173 bool "Enable 16-bit UID system calls" if EXPERT
1174 depends on HAVE_UID16 && MULTIUSER
1177 This enables the legacy 16-bit UID syscall wrappers.
1180 bool "Multiple users, groups and capabilities support" if EXPERT
1183 This option enables support for non-root users, groups and
1186 If you say N here, all processes will run with UID 0, GID 0, and all
1187 possible capabilities. Saying N here also compiles out support for
1188 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1191 If unsure, say Y here.
1193 config SGETMASK_SYSCALL
1194 bool "sgetmask/ssetmask syscalls support" if EXPERT
1195 def_bool PARISC || M68K || PPC || MIPS || X86 || SPARC || MICROBLAZE || SUPERH
1197 sys_sgetmask and sys_ssetmask are obsolete system calls
1198 no longer supported in libc but still enabled by default in some
1201 If unsure, leave the default option here.
1203 config SYSFS_SYSCALL
1204 bool "Sysfs syscall support" if EXPERT
1207 sys_sysfs is an obsolete system call no longer supported in libc.
1208 Note that disabling this option is more secure but might break
1209 compatibility with some systems.
1211 If unsure say Y here.
1213 config SYSCTL_SYSCALL
1214 bool "Sysctl syscall support" if EXPERT
1215 depends on PROC_SYSCTL
1219 sys_sysctl uses binary paths that have been found challenging
1220 to properly maintain and use. The interface in /proc/sys
1221 using paths with ascii names is now the primary path to this
1224 Almost nothing using the binary sysctl interface so if you are
1225 trying to save some space it is probably safe to disable this,
1226 making your kernel marginally smaller.
1228 If unsure say N here.
1231 bool "open by fhandle syscalls" if EXPERT
1235 If you say Y here, a user level program will be able to map
1236 file names to handle and then later use the handle for
1237 different file system operations. This is useful in implementing
1238 userspace file servers, which now track files using handles instead
1239 of names. The handle would remain the same even if file names
1240 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
1244 bool "Posix Clocks & timers" if EXPERT
1247 This includes native support for POSIX timers to the kernel.
1248 Some embedded systems have no use for them and therefore they
1249 can be configured out to reduce the size of the kernel image.
1251 When this option is disabled, the following syscalls won't be
1252 available: timer_create, timer_gettime: timer_getoverrun,
1253 timer_settime, timer_delete, clock_adjtime, getitimer,
1254 setitimer, alarm. Furthermore, the clock_settime, clock_gettime,
1255 clock_getres and clock_nanosleep syscalls will be limited to
1256 CLOCK_REALTIME, CLOCK_MONOTONIC and CLOCK_BOOTTIME only.
1262 bool "Enable support for printk" if EXPERT
1265 This option enables normal printk support. Removing it
1266 eliminates most of the message strings from the kernel image
1267 and makes the kernel more or less silent. As this makes it
1268 very difficult to diagnose system problems, saying N here is
1269 strongly discouraged.
1277 bool "BUG() support" if EXPERT
1280 Disabling this option eliminates support for BUG and WARN, reducing
1281 the size of your kernel image and potentially quietly ignoring
1282 numerous fatal conditions. You should only consider disabling this
1283 option for embedded systems with no facilities for reporting errors.
1289 bool "Enable ELF core dumps" if EXPERT
1291 Enable support for generating core dumps. Disabling saves about 4k.
1294 config PCSPKR_PLATFORM
1295 bool "Enable PC-Speaker support" if EXPERT
1296 depends on HAVE_PCSPKR_PLATFORM
1300 This option allows to disable the internal PC-Speaker
1301 support, saving some memory.
1305 bool "Enable full-sized data structures for core" if EXPERT
1307 Disabling this option reduces the size of miscellaneous core
1308 kernel data structures. This saves memory on small machines,
1309 but may reduce performance.
1312 bool "Enable futex support" if EXPERT
1316 Disabling this option will cause the kernel to be built without
1317 support for "fast userspace mutexes". The resulting kernel may not
1318 run glibc-based applications correctly.
1322 depends on FUTEX && RT_MUTEXES
1325 config HAVE_FUTEX_CMPXCHG
1329 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1330 is implemented and always working. This removes a couple of runtime
1334 bool "Enable eventpoll support" if EXPERT
1338 Disabling this option will cause the kernel to be built without
1339 support for epoll family of system calls.
1342 bool "Enable signalfd() system call" if EXPERT
1346 Enable the signalfd() system call that allows to receive signals
1347 on a file descriptor.
1352 bool "Enable timerfd() system call" if EXPERT
1356 Enable the timerfd() system call that allows to receive timer
1357 events on a file descriptor.
1362 bool "Enable eventfd() system call" if EXPERT
1366 Enable the eventfd() system call that allows to receive both
1367 kernel notification (ie. KAIO) or userspace notifications.
1372 bool "Use full shmem filesystem" if EXPERT
1376 The shmem is an internal filesystem used to manage shared memory.
1377 It is backed by swap and manages resource limits. It is also exported
1378 to userspace as tmpfs if TMPFS is enabled. Disabling this
1379 option replaces shmem and tmpfs with the much simpler ramfs code,
1380 which may be appropriate on small systems without swap.
1383 bool "Enable AIO support" if EXPERT
1386 This option enables POSIX asynchronous I/O which may by used
1387 by some high performance threaded applications. Disabling
1388 this option saves about 7k.
1390 config ADVISE_SYSCALLS
1391 bool "Enable madvise/fadvise syscalls" if EXPERT
1394 This option enables the madvise and fadvise syscalls, used by
1395 applications to advise the kernel about their future memory or file
1396 usage, improving performance. If building an embedded system where no
1397 applications use these syscalls, you can disable this option to save
1401 bool "Enable membarrier() system call" if EXPERT
1404 Enable the membarrier() system call that allows issuing memory
1405 barriers across all running threads, which can be used to distribute
1406 the cost of user-space memory barriers asymmetrically by transforming
1407 pairs of memory barriers into pairs consisting of membarrier() and a
1413 bool "Load all symbols for debugging/ksymoops" if EXPERT
1416 Say Y here to let the kernel print out symbolic crash information and
1417 symbolic stack backtraces. This increases the size of the kernel
1418 somewhat, as all symbols have to be loaded into the kernel image.
1421 bool "Include all symbols in kallsyms"
1422 depends on DEBUG_KERNEL && KALLSYMS
1424 Normally kallsyms only contains the symbols of functions for nicer
1425 OOPS messages and backtraces (i.e., symbols from the text and inittext
1426 sections). This is sufficient for most cases. And only in very rare
1427 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1428 names of variables from the data sections, etc).
1430 This option makes sure that all symbols are loaded into the kernel
1431 image (i.e., symbols from all sections) in cost of increased kernel
1432 size (depending on the kernel configuration, it may be 300KiB or
1433 something like this).
1435 Say N unless you really need all symbols.
1437 config KALLSYMS_ABSOLUTE_PERCPU
1440 default X86_64 && SMP
1442 config KALLSYMS_BASE_RELATIVE
1447 Instead of emitting them as absolute values in the native word size,
1448 emit the symbol references in the kallsyms table as 32-bit entries,
1449 each containing a relative value in the range [base, base + U32_MAX]
1450 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1451 an absolute value in the range [0, S32_MAX] or a relative value in the
1452 range [base, base + S32_MAX], where base is the lowest relative symbol
1453 address encountered in the image.
1455 On 64-bit builds, this reduces the size of the address table by 50%,
1456 but more importantly, it results in entries whose values are build
1457 time constants, and no relocation pass is required at runtime to fix
1458 up the entries based on the runtime load address of the kernel.
1460 # end of the "standard kernel features (expert users)" menu
1462 # syscall, maps, verifier
1464 bool "Enable bpf() system call"
1470 Enable the bpf() system call that allows to manipulate eBPF
1471 programs and maps via file descriptors.
1473 config BPF_JIT_ALWAYS_ON
1474 bool "Permanently enable BPF JIT and remove BPF interpreter"
1475 depends on BPF_SYSCALL && HAVE_EBPF_JIT && BPF_JIT
1477 Enables BPF JIT and removes BPF interpreter to avoid
1478 speculative execution of BPF instructions by the interpreter
1480 config BPF_UNPRIV_DEFAULT_OFF
1481 bool "Disable unprivileged BPF by default"
1482 depends on BPF_SYSCALL
1484 Disables unprivileged BPF by default by setting the corresponding
1485 /proc/sys/kernel/unprivileged_bpf_disabled knob to 2. An admin can
1486 still reenable it by setting it to 0 later on, or permanently
1487 disable it by setting it to 1 (from which no other transition to
1488 0 is possible anymore).
1491 bool "Enable userfaultfd() system call"
1495 Enable the userfaultfd() system call that allows to intercept and
1496 handle page faults in userland.
1498 config ARCH_HAS_MEMBARRIER_CALLBACKS
1501 config ARCH_HAS_MEMBARRIER_SYNC_CORE
1505 bool "Enable rseq() system call" if EXPERT
1507 depends on HAVE_RSEQ
1510 Enable the restartable sequences system call. It provides a
1511 user-space cache for the current CPU number value, which
1512 speeds up getting the current CPU number from user-space,
1513 as well as an ABI to speed up user-space operations on
1520 bool "Enabled debugging of rseq() system call" if EXPERT
1521 depends on RSEQ && DEBUG_KERNEL
1523 Enable extra debugging checks for the rseq system call.
1528 bool "Embedded system"
1529 option allnoconfig_y
1532 This option should be enabled if compiling the kernel for
1533 an embedded system so certain expert options are available
1536 config HAVE_PERF_EVENTS
1539 See tools/perf/design.txt for details.
1541 config PERF_USE_VMALLOC
1544 See tools/perf/design.txt for details
1547 bool "PC/104 support" if EXPERT
1549 Expose PC/104 form factor device drivers and options available for
1550 selection and configuration. Enable this option if your target
1551 machine has a PC/104 bus.
1553 menu "Kernel Performance Events And Counters"
1556 bool "Kernel performance events and counters"
1557 default y if PROFILING
1558 depends on HAVE_PERF_EVENTS
1563 Enable kernel support for various performance events provided
1564 by software and hardware.
1566 Software events are supported either built-in or via the
1567 use of generic tracepoints.
1569 Most modern CPUs support performance events via performance
1570 counter registers. These registers count the number of certain
1571 types of hw events: such as instructions executed, cachemisses
1572 suffered, or branches mis-predicted - without slowing down the
1573 kernel or applications. These registers can also trigger interrupts
1574 when a threshold number of events have passed - and can thus be
1575 used to profile the code that runs on that CPU.
1577 The Linux Performance Event subsystem provides an abstraction of
1578 these software and hardware event capabilities, available via a
1579 system call and used by the "perf" utility in tools/perf/. It
1580 provides per task and per CPU counters, and it provides event
1581 capabilities on top of those.
1585 config DEBUG_PERF_USE_VMALLOC
1587 bool "Debug: use vmalloc to back perf mmap() buffers"
1588 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1589 select PERF_USE_VMALLOC
1591 Use vmalloc memory to back perf mmap() buffers.
1593 Mostly useful for debugging the vmalloc code on platforms
1594 that don't require it.
1600 config VM_EVENT_COUNTERS
1602 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1604 VM event counters are needed for event counts to be shown.
1605 This option allows the disabling of the VM event counters
1606 on EXPERT systems. /proc/vmstat will only show page counts
1607 if VM event counters are disabled.
1611 bool "Enable SLUB debugging support" if EXPERT
1612 depends on SLUB && SYSFS
1614 SLUB has extensive debug support features. Disabling these can
1615 result in significant savings in code size. This also disables
1616 SLUB sysfs support. /sys/slab will not exist and there will be
1617 no support for cache validation etc.
1619 config SLUB_MEMCG_SYSFS_ON
1621 bool "Enable memcg SLUB sysfs support by default" if EXPERT
1622 depends on SLUB && SYSFS && MEMCG
1624 SLUB creates a directory under /sys/kernel/slab for each
1625 allocation cache to host info and debug files. If memory
1626 cgroup is enabled, each cache can have per memory cgroup
1627 caches. SLUB can create the same sysfs directories for these
1628 caches under /sys/kernel/slab/CACHE/cgroup but it can lead
1629 to a very high number of debug files being created. This is
1630 controlled by slub_memcg_sysfs boot parameter and this
1631 config option determines the parameter's default value.
1634 bool "Disable heap randomization"
1637 Randomizing heap placement makes heap exploits harder, but it
1638 also breaks ancient binaries (including anything libc5 based).
1639 This option changes the bootup default to heap randomization
1640 disabled, and can be overridden at runtime by setting
1641 /proc/sys/kernel/randomize_va_space to 2.
1643 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1646 prompt "Choose SLAB allocator"
1649 This option allows to select a slab allocator.
1653 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1655 The regular slab allocator that is established and known to work
1656 well in all environments. It organizes cache hot objects in
1657 per cpu and per node queues.
1660 bool "SLUB (Unqueued Allocator)"
1661 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1663 SLUB is a slab allocator that minimizes cache line usage
1664 instead of managing queues of cached objects (SLAB approach).
1665 Per cpu caching is realized using slabs of objects instead
1666 of queues of objects. SLUB can use memory efficiently
1667 and has enhanced diagnostics. SLUB is the default choice for
1672 bool "SLOB (Simple Allocator)"
1674 SLOB replaces the stock allocator with a drastically simpler
1675 allocator. SLOB is generally more space efficient but
1676 does not perform as well on large systems.
1680 config SLAB_MERGE_DEFAULT
1681 bool "Allow slab caches to be merged"
1684 For reduced kernel memory fragmentation, slab caches can be
1685 merged when they share the same size and other characteristics.
1686 This carries a risk of kernel heap overflows being able to
1687 overwrite objects from merged caches (and more easily control
1688 cache layout), which makes such heap attacks easier to exploit
1689 by attackers. By keeping caches unmerged, these kinds of exploits
1690 can usually only damage objects in the same cache. To disable
1691 merging at runtime, "slab_nomerge" can be passed on the kernel
1694 config SLAB_FREELIST_RANDOM
1696 depends on SLAB || SLUB
1697 bool "SLAB freelist randomization"
1699 Randomizes the freelist order used on creating new pages. This
1700 security feature reduces the predictability of the kernel slab
1701 allocator against heap overflows.
1703 config SLAB_FREELIST_HARDENED
1704 bool "Harden slab freelist metadata"
1707 Many kernel heap attacks try to target slab cache metadata and
1708 other infrastructure. This options makes minor performance
1709 sacrifies to harden the kernel slab allocator against common
1710 freelist exploit methods.
1712 config SLUB_CPU_PARTIAL
1714 depends on SLUB && SMP
1715 bool "SLUB per cpu partial cache"
1717 Per cpu partial caches accellerate objects allocation and freeing
1718 that is local to a processor at the price of more indeterminism
1719 in the latency of the free. On overflow these caches will be cleared
1720 which requires the taking of locks that may cause latency spikes.
1721 Typically one would choose no for a realtime system.
1723 config MMAP_ALLOW_UNINITIALIZED
1724 bool "Allow mmapped anonymous memory to be uninitialized"
1725 depends on EXPERT && !MMU
1728 Normally, and according to the Linux spec, anonymous memory obtained
1729 from mmap() has its contents cleared before it is passed to
1730 userspace. Enabling this config option allows you to request that
1731 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1732 providing a huge performance boost. If this option is not enabled,
1733 then the flag will be ignored.
1735 This is taken advantage of by uClibc's malloc(), and also by
1736 ELF-FDPIC binfmt's brk and stack allocator.
1738 Because of the obvious security issues, this option should only be
1739 enabled on embedded devices where you control what is run in
1740 userspace. Since that isn't generally a problem on no-MMU systems,
1741 it is normally safe to say Y here.
1743 See Documentation/nommu-mmap.txt for more information.
1745 config SYSTEM_DATA_VERIFICATION
1747 select SYSTEM_TRUSTED_KEYRING
1751 select ASYMMETRIC_KEY_TYPE
1752 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1755 select X509_CERTIFICATE_PARSER
1756 select PKCS7_MESSAGE_PARSER
1758 Provide PKCS#7 message verification using the contents of the system
1759 trusted keyring to provide public keys. This then can be used for
1760 module verification, kexec image verification and firmware blob
1764 bool "Profiling support"
1766 Say Y here to enable the extended profiling support mechanisms used
1767 by profilers such as OProfile.
1770 # Place an empty function call at each tracepoint site. Can be
1771 # dynamically changed for a probe function.
1776 endmenu # General setup
1778 source "arch/Kconfig"
1785 default 0 if BASE_FULL
1786 default 1 if !BASE_FULL
1789 bool "Enable loadable module support"
1792 Kernel modules are small pieces of compiled code which can
1793 be inserted in the running kernel, rather than being
1794 permanently built into the kernel. You use the "modprobe"
1795 tool to add (and sometimes remove) them. If you say Y here,
1796 many parts of the kernel can be built as modules (by
1797 answering M instead of Y where indicated): this is most
1798 useful for infrequently used options which are not required
1799 for booting. For more information, see the man pages for
1800 modprobe, lsmod, modinfo, insmod and rmmod.
1802 If you say Y here, you will need to run "make
1803 modules_install" to put the modules under /lib/modules/
1804 where modprobe can find them (you may need to be root to do
1811 config MODULE_FORCE_LOAD
1812 bool "Forced module loading"
1815 Allow loading of modules without version information (ie. modprobe
1816 --force). Forced module loading sets the 'F' (forced) taint flag and
1817 is usually a really bad idea.
1819 config MODULE_UNLOAD
1820 bool "Module unloading"
1822 Without this option you will not be able to unload any
1823 modules (note that some modules may not be unloadable
1824 anyway), which makes your kernel smaller, faster
1825 and simpler. If unsure, say Y.
1827 config MODULE_FORCE_UNLOAD
1828 bool "Forced module unloading"
1829 depends on MODULE_UNLOAD
1831 This option allows you to force a module to unload, even if the
1832 kernel believes it is unsafe: the kernel will remove the module
1833 without waiting for anyone to stop using it (using the -f option to
1834 rmmod). This is mainly for kernel developers and desperate users.
1838 bool "Module versioning support"
1840 Usually, you have to use modules compiled with your kernel.
1841 Saying Y here makes it sometimes possible to use modules
1842 compiled for different kernels, by adding enough information
1843 to the modules to (hopefully) spot any changes which would
1844 make them incompatible with the kernel you are running. If
1847 config MODULE_REL_CRCS
1849 depends on MODVERSIONS
1851 config MODULE_SRCVERSION_ALL
1852 bool "Source checksum for all modules"
1854 Modules which contain a MODULE_VERSION get an extra "srcversion"
1855 field inserted into their modinfo section, which contains a
1856 sum of the source files which made it. This helps maintainers
1857 see exactly which source was used to build a module (since
1858 others sometimes change the module source without updating
1859 the version). With this option, such a "srcversion" field
1860 will be created for all modules. If unsure, say N.
1863 bool "Module signature verification"
1865 select SYSTEM_DATA_VERIFICATION
1867 Check modules for valid signatures upon load: the signature
1868 is simply appended to the module. For more information see
1869 <file:Documentation/admin-guide/module-signing.rst>.
1871 Note that this option adds the OpenSSL development packages as a
1872 kernel build dependency so that the signing tool can use its crypto
1875 !!!WARNING!!! If you enable this option, you MUST make sure that the
1876 module DOES NOT get stripped after being signed. This includes the
1877 debuginfo strip done by some packagers (such as rpmbuild) and
1878 inclusion into an initramfs that wants the module size reduced.
1880 config MODULE_SIG_FORCE
1881 bool "Require modules to be validly signed"
1882 depends on MODULE_SIG
1884 Reject unsigned modules or signed modules for which we don't have a
1885 key. Without this, such modules will simply taint the kernel.
1887 config MODULE_SIG_ALL
1888 bool "Automatically sign all modules"
1890 depends on MODULE_SIG
1892 Sign all modules during make modules_install. Without this option,
1893 modules must be signed manually, using the scripts/sign-file tool.
1895 comment "Do not forget to sign required modules with scripts/sign-file"
1896 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
1899 prompt "Which hash algorithm should modules be signed with?"
1900 depends on MODULE_SIG
1902 This determines which sort of hashing algorithm will be used during
1903 signature generation. This algorithm _must_ be built into the kernel
1904 directly so that signature verification can take place. It is not
1905 possible to load a signed module containing the algorithm to check
1906 the signature on that module.
1908 config MODULE_SIG_SHA1
1909 bool "Sign modules with SHA-1"
1912 config MODULE_SIG_SHA224
1913 bool "Sign modules with SHA-224"
1914 select CRYPTO_SHA256
1916 config MODULE_SIG_SHA256
1917 bool "Sign modules with SHA-256"
1918 select CRYPTO_SHA256
1920 config MODULE_SIG_SHA384
1921 bool "Sign modules with SHA-384"
1922 select CRYPTO_SHA512
1924 config MODULE_SIG_SHA512
1925 bool "Sign modules with SHA-512"
1926 select CRYPTO_SHA512
1930 config MODULE_SIG_HASH
1932 depends on MODULE_SIG
1933 default "sha1" if MODULE_SIG_SHA1
1934 default "sha224" if MODULE_SIG_SHA224
1935 default "sha256" if MODULE_SIG_SHA256
1936 default "sha384" if MODULE_SIG_SHA384
1937 default "sha512" if MODULE_SIG_SHA512
1939 config MODULE_COMPRESS
1940 bool "Compress modules on installation"
1944 Compresses kernel modules when 'make modules_install' is run; gzip or
1945 xz depending on "Compression algorithm" below.
1947 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
1949 Out-of-tree kernel modules installed using Kbuild will also be
1950 compressed upon installation.
1952 Note: for modules inside an initrd or initramfs, it's more efficient
1953 to compress the whole initrd or initramfs instead.
1955 Note: This is fully compatible with signed modules.
1960 prompt "Compression algorithm"
1961 depends on MODULE_COMPRESS
1962 default MODULE_COMPRESS_GZIP
1964 This determines which sort of compression will be used during
1965 'make modules_install'.
1967 GZIP (default) and XZ are supported.
1969 config MODULE_COMPRESS_GZIP
1972 config MODULE_COMPRESS_XZ
1977 config TRIM_UNUSED_KSYMS
1978 bool "Trim unused exported kernel symbols"
1979 depends on MODULES && !UNUSED_SYMBOLS
1981 The kernel and some modules make many symbols available for
1982 other modules to use via EXPORT_SYMBOL() and variants. Depending
1983 on the set of modules being selected in your kernel configuration,
1984 many of those exported symbols might never be used.
1986 This option allows for unused exported symbols to be dropped from
1987 the build. In turn, this provides the compiler more opportunities
1988 (especially when using LTO) for optimizing the code and reducing
1989 binary size. This might have some security advantages as well.
1991 If unsure, or if you need to build out-of-tree modules, say N.
1995 config MODULES_TREE_LOOKUP
1997 depends on PERF_EVENTS || TRACING
1999 config INIT_ALL_POSSIBLE
2002 Back when each arch used to define their own cpu_online_mask and
2003 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2004 with all 1s, and others with all 0s. When they were centralised,
2005 it was better to provide this option than to break all the archs
2006 and have several arch maintainers pursuing me down dark alleys.
2008 source "block/Kconfig"
2010 config PREEMPT_NOTIFIERS
2020 Build a simple ASN.1 grammar compiler that produces a bytecode output
2021 that can be interpreted by the ASN.1 stream decoder and used to
2022 inform it as to what tags are to be expected in a stream and what
2023 functions to call on what tags.
2025 source "kernel/Kconfig.locks"
2027 config ARCH_HAS_SYNC_CORE_BEFORE_USERMODE
2030 # It may be useful for an architecture to override the definitions of the
2031 # SYSCALL_DEFINE() and __SYSCALL_DEFINEx() macros in <linux/syscalls.h>
2032 # and the COMPAT_ variants in <linux/compat.h>, in particular to use a
2033 # different calling convention for syscalls. They can also override the
2034 # macros for not-implemented syscalls in kernel/sys_ni.c and
2035 # kernel/time/posix-stubs.c. All these overrides need to be available in
2036 # <asm/syscall_wrapper.h>.
2037 config ARCH_HAS_SYSCALL_WRAPPER