7 option env="KERNELVERSION"
13 default "/lib/modules/$UNAME_RELEASE/.config"
14 default "/etc/kernel-config"
15 default "/boot/config-$UNAME_RELEASE"
16 default "$ARCH_DEFCONFIG"
17 default "arch/$ARCH/defconfig"
26 config BUILDTIME_EXTABLE_SORT
36 depends on BROKEN || !SMP
39 config INIT_ENV_ARG_LIMIT
44 Maximum of each of the number of arguments and environment
45 variables passed to init from the kernel command line.
49 string "Cross-compiler tool prefix"
51 Same as running 'make CROSS_COMPILE=prefix-' but stored for
52 default make runs in this kernel build directory. You don't
53 need to set this unless you want the configured kernel build
54 directory to select the cross-compiler automatically.
57 bool "Compile also drivers which will not load"
61 Some drivers can be compiled on a different platform than they are
62 intended to be run on. Despite they cannot be loaded there (or even
63 when they load they cannot be used due to missing HW support),
64 developers still, opposing to distributors, might want to build such
65 drivers to compile-test them.
67 If you are a developer and want to build everything available, say Y
68 here. If you are a user/distributor, say N here to exclude useless
69 drivers to be distributed.
72 string "Local version - append to kernel release"
74 Append an extra string to the end of your kernel version.
75 This will show up when you type uname, for example.
76 The string you set here will be appended after the contents of
77 any files with a filename matching localversion* in your
78 object and source tree, in that order. Your total string can
79 be a maximum of 64 characters.
81 config LOCALVERSION_AUTO
82 bool "Automatically append version information to the version string"
84 depends on !COMPILE_TEST
86 This will try to automatically determine if the current tree is a
87 release tree by looking for git tags that belong to the current
90 A string of the format -gxxxxxxxx will be added to the localversion
91 if a git-based tree is found. The string generated by this will be
92 appended after any matching localversion* files, and after the value
93 set in CONFIG_LOCALVERSION.
95 (The actual string used here is the first eight characters produced
96 by running the command:
98 $ git rev-parse --verify HEAD
100 which is done within the script "scripts/setlocalversion".)
102 config HAVE_KERNEL_GZIP
105 config HAVE_KERNEL_BZIP2
108 config HAVE_KERNEL_LZMA
111 config HAVE_KERNEL_XZ
114 config HAVE_KERNEL_LZO
117 config HAVE_KERNEL_LZ4
121 prompt "Kernel compression mode"
123 depends on HAVE_KERNEL_GZIP || HAVE_KERNEL_BZIP2 || HAVE_KERNEL_LZMA || HAVE_KERNEL_XZ || HAVE_KERNEL_LZO || HAVE_KERNEL_LZ4
125 The linux kernel is a kind of self-extracting executable.
126 Several compression algorithms are available, which differ
127 in efficiency, compression and decompression speed.
128 Compression speed is only relevant when building a kernel.
129 Decompression speed is relevant at each boot.
131 If you have any problems with bzip2 or lzma compressed
132 kernels, mail me (Alain Knaff) <alain@knaff.lu>. (An older
133 version of this functionality (bzip2 only), for 2.4, was
134 supplied by Christian Ludwig)
136 High compression options are mostly useful for users, who
137 are low on disk space (embedded systems), but for whom ram
140 If in doubt, select 'gzip'
144 depends on HAVE_KERNEL_GZIP
146 The old and tried gzip compression. It provides a good balance
147 between compression ratio and decompression speed.
151 depends on HAVE_KERNEL_BZIP2
153 Its compression ratio and speed is intermediate.
154 Decompression speed is slowest among the choices. The kernel
155 size is about 10% smaller with bzip2, in comparison to gzip.
156 Bzip2 uses a large amount of memory. For modern kernels you
157 will need at least 8MB RAM or more for booting.
161 depends on HAVE_KERNEL_LZMA
163 This compression algorithm's ratio is best. Decompression speed
164 is between gzip and bzip2. Compression is slowest.
165 The kernel size is about 33% smaller with LZMA in comparison to gzip.
169 depends on HAVE_KERNEL_XZ
171 XZ uses the LZMA2 algorithm and instruction set specific
172 BCJ filters which can improve compression ratio of executable
173 code. The size of the kernel is about 30% smaller with XZ in
174 comparison to gzip. On architectures for which there is a BCJ
175 filter (i386, x86_64, ARM, IA-64, PowerPC, and SPARC), XZ
176 will create a few percent smaller kernel than plain LZMA.
178 The speed is about the same as with LZMA: The decompression
179 speed of XZ is better than that of bzip2 but worse than gzip
180 and LZO. Compression is slow.
184 depends on HAVE_KERNEL_LZO
186 Its compression ratio is the poorest among the choices. The kernel
187 size is about 10% bigger than gzip; however its speed
188 (both compression and decompression) is the fastest.
192 depends on HAVE_KERNEL_LZ4
194 LZ4 is an LZ77-type compressor with a fixed, byte-oriented encoding.
195 A preliminary version of LZ4 de/compression tool is available at
196 <https://code.google.com/p/lz4/>.
198 Its compression ratio is worse than LZO. The size of the kernel
199 is about 8% bigger than LZO. But the decompression speed is
204 config DEFAULT_HOSTNAME
205 string "Default hostname"
208 This option determines the default system hostname before userspace
209 calls sethostname(2). The kernel traditionally uses "(none)" here,
210 but you may wish to use a different default here to make a minimal
211 system more usable with less configuration.
214 bool "Support for paging of anonymous memory (swap)"
215 depends on MMU && BLOCK
218 This option allows you to choose whether you want to have support
219 for so called swap devices or swap files in your kernel that are
220 used to provide more virtual memory than the actual RAM present
221 in your computer. If unsure say Y.
226 Inter Process Communication is a suite of library functions and
227 system calls which let processes (running programs) synchronize and
228 exchange information. It is generally considered to be a good thing,
229 and some programs won't run unless you say Y here. In particular, if
230 you want to run the DOS emulator dosemu under Linux (read the
231 DOSEMU-HOWTO, available from <http://www.tldp.org/docs.html#howto>),
232 you'll need to say Y here.
234 You can find documentation about IPC with "info ipc" and also in
235 section 6.4 of the Linux Programmer's Guide, available from
236 <http://www.tldp.org/guides.html>.
238 config SYSVIPC_SYSCTL
245 bool "POSIX Message Queues"
248 POSIX variant of message queues is a part of IPC. In POSIX message
249 queues every message has a priority which decides about succession
250 of receiving it by a process. If you want to compile and run
251 programs written e.g. for Solaris with use of its POSIX message
252 queues (functions mq_*) say Y here.
254 POSIX message queues are visible as a filesystem called 'mqueue'
255 and can be mounted somewhere if you want to do filesystem
256 operations on message queues.
260 config POSIX_MQUEUE_SYSCTL
262 depends on POSIX_MQUEUE
266 config CROSS_MEMORY_ATTACH
267 bool "Enable process_vm_readv/writev syscalls"
271 Enabling this option adds the system calls process_vm_readv and
272 process_vm_writev which allow a process with the correct privileges
273 to directly read from or write to another process' address space.
274 See the man page for more details.
277 bool "open by fhandle syscalls" if EXPERT
281 If you say Y here, a user level program will be able to map
282 file names to handle and then later use the handle for
283 different file system operations. This is useful in implementing
284 userspace file servers, which now track files using handles instead
285 of names. The handle would remain the same even if file names
286 get renamed. Enables open_by_handle_at(2) and name_to_handle_at(2)
290 bool "uselib syscall"
291 def_bool ALPHA || M68K || SPARC || X86_32 || IA32_EMULATION
293 This option enables the uselib syscall, a system call used in the
294 dynamic linker from libc5 and earlier. glibc does not use this
295 system call. If you intend to run programs built on libc5 or
296 earlier, you may need to enable this syscall. Current systems
297 running glibc can safely disable this.
300 bool "Auditing support"
303 Enable auditing infrastructure that can be used with another
304 kernel subsystem, such as SELinux (which requires this for
305 logging of avc messages output). System call auditing is included
306 on architectures which support it.
308 config HAVE_ARCH_AUDITSYSCALL
313 depends on AUDIT && HAVE_ARCH_AUDITSYSCALL
317 depends on AUDITSYSCALL
322 depends on AUDITSYSCALL
325 source "kernel/irq/Kconfig"
326 source "kernel/time/Kconfig"
328 menu "CPU/Task time and stats accounting"
330 config VIRT_CPU_ACCOUNTING
334 prompt "Cputime accounting"
335 default TICK_CPU_ACCOUNTING if !PPC64
336 default VIRT_CPU_ACCOUNTING_NATIVE if PPC64
338 # Kind of a stub config for the pure tick based cputime accounting
339 config TICK_CPU_ACCOUNTING
340 bool "Simple tick based cputime accounting"
341 depends on !S390 && !NO_HZ_FULL
343 This is the basic tick based cputime accounting that maintains
344 statistics about user, system and idle time spent on per jiffies
349 config VIRT_CPU_ACCOUNTING_NATIVE
350 bool "Deterministic task and CPU time accounting"
351 depends on HAVE_VIRT_CPU_ACCOUNTING && !NO_HZ_FULL
352 select VIRT_CPU_ACCOUNTING
354 Select this option to enable more accurate task and CPU time
355 accounting. This is done by reading a CPU counter on each
356 kernel entry and exit and on transitions within the kernel
357 between system, softirq and hardirq state, so there is a
358 small performance impact. In the case of s390 or IBM POWER > 5,
359 this also enables accounting of stolen time on logically-partitioned
362 config VIRT_CPU_ACCOUNTING_GEN
363 bool "Full dynticks CPU time accounting"
364 depends on HAVE_CONTEXT_TRACKING
365 depends on HAVE_VIRT_CPU_ACCOUNTING_GEN
366 select VIRT_CPU_ACCOUNTING
367 select CONTEXT_TRACKING
369 Select this option to enable task and CPU time accounting on full
370 dynticks systems. This accounting is implemented by watching every
371 kernel-user boundaries using the context tracking subsystem.
372 The accounting is thus performed at the expense of some significant
375 For now this is only useful if you are working on the full
376 dynticks subsystem development.
382 config IRQ_TIME_ACCOUNTING
383 bool "Fine granularity task level IRQ time accounting"
384 depends on HAVE_IRQ_TIME_ACCOUNTING && !VIRT_CPU_ACCOUNTING_NATIVE
386 Select this option to enable fine granularity task irq time
387 accounting. This is done by reading a timestamp on each
388 transitions between softirq and hardirq state, so there can be a
389 small performance impact.
391 If in doubt, say N here.
393 config BSD_PROCESS_ACCT
394 bool "BSD Process Accounting"
397 If you say Y here, a user level program will be able to instruct the
398 kernel (via a special system call) to write process accounting
399 information to a file: whenever a process exits, information about
400 that process will be appended to the file by the kernel. The
401 information includes things such as creation time, owning user,
402 command name, memory usage, controlling terminal etc. (the complete
403 list is in the struct acct in <file:include/linux/acct.h>). It is
404 up to the user level program to do useful things with this
405 information. This is generally a good idea, so say Y.
407 config BSD_PROCESS_ACCT_V3
408 bool "BSD Process Accounting version 3 file format"
409 depends on BSD_PROCESS_ACCT
412 If you say Y here, the process accounting information is written
413 in a new file format that also logs the process IDs of each
414 process and it's parent. Note that this file format is incompatible
415 with previous v0/v1/v2 file formats, so you will need updated tools
416 for processing it. A preliminary version of these tools is available
417 at <http://www.gnu.org/software/acct/>.
420 bool "Export task/process statistics through netlink"
425 Export selected statistics for tasks/processes through the
426 generic netlink interface. Unlike BSD process accounting, the
427 statistics are available during the lifetime of tasks/processes as
428 responses to commands. Like BSD accounting, they are sent to user
433 config TASK_DELAY_ACCT
434 bool "Enable per-task delay accounting"
438 Collect information on time spent by a task waiting for system
439 resources like cpu, synchronous block I/O completion and swapping
440 in pages. Such statistics can help in setting a task's priorities
441 relative to other tasks for cpu, io, rss limits etc.
446 bool "Enable extended accounting over taskstats"
449 Collect extended task accounting data and send the data
450 to userland for processing over the taskstats interface.
454 config TASK_IO_ACCOUNTING
455 bool "Enable per-task storage I/O accounting"
456 depends on TASK_XACCT
458 Collect information on the number of bytes of storage I/O which this
463 endmenu # "CPU/Task time and stats accounting"
469 default y if !PREEMPT && SMP
471 This option selects the RCU implementation that is
472 designed for very large SMP system with hundreds or
473 thousands of CPUs. It also scales down nicely to
480 This option selects the RCU implementation that is
481 designed for very large SMP systems with hundreds or
482 thousands of CPUs, but for which real-time response
483 is also required. It also scales down nicely to
486 Select this option if you are unsure.
490 default y if !PREEMPT && !SMP
492 This option selects the RCU implementation that is
493 designed for UP systems from which real-time response
494 is not required. This option greatly reduces the
495 memory footprint of RCU.
498 bool "Make expert-level adjustments to RCU configuration"
501 This option needs to be enabled if you wish to make
502 expert-level adjustments to RCU configuration. By default,
503 no such adjustments can be made, which has the often-beneficial
504 side-effect of preventing "make oldconfig" from asking you all
505 sorts of detailed questions about how you would like numerous
506 obscure RCU options to be set up.
508 Say Y if you need to make expert-level adjustments to RCU.
510 Say N if you are unsure.
515 This option selects the sleepable version of RCU. This version
516 permits arbitrary sleeping or blocking within RCU read-side critical
525 This option enables a task-based RCU implementation that uses
526 only voluntary context switch (not preemption!), idle, and
527 user-mode execution as quiescent states.
529 config RCU_STALL_COMMON
530 def_bool ( TREE_RCU || PREEMPT_RCU || RCU_TRACE )
532 This option enables RCU CPU stall code that is common between
533 the TINY and TREE variants of RCU. The purpose is to allow
534 the tiny variants to disable RCU CPU stall warnings, while
535 making these warnings mandatory for the tree variants.
537 config CONTEXT_TRACKING
540 config CONTEXT_TRACKING_FORCE
541 bool "Force context tracking"
542 depends on CONTEXT_TRACKING
543 default y if !NO_HZ_FULL
545 The major pre-requirement for full dynticks to work is to
546 support the context tracking subsystem. But there are also
547 other dependencies to provide in order to make the full
550 This option stands for testing when an arch implements the
551 context tracking backend but doesn't yet fullfill all the
552 requirements to make the full dynticks feature working.
553 Without the full dynticks, there is no way to test the support
554 for context tracking and the subsystems that rely on it: RCU
555 userspace extended quiescent state and tickless cputime
556 accounting. This option copes with the absence of the full
557 dynticks subsystem by forcing the context tracking on all
560 Say Y only if you're working on the development of an
561 architecture backend for the context tracking.
563 Say N otherwise, this option brings an overhead that you
564 don't want in production.
568 int "Tree-based hierarchical RCU fanout value"
571 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
575 This option controls the fanout of hierarchical implementations
576 of RCU, allowing RCU to work efficiently on machines with
577 large numbers of CPUs. This value must be at least the fourth
578 root of NR_CPUS, which allows NR_CPUS to be insanely large.
579 The default value of RCU_FANOUT should be used for production
580 systems, but if you are stress-testing the RCU implementation
581 itself, small RCU_FANOUT values allow you to test large-system
582 code paths on small(er) systems.
584 Select a specific number if testing RCU itself.
585 Take the default if unsure.
587 config RCU_FANOUT_LEAF
588 int "Tree-based hierarchical RCU leaf-level fanout value"
591 depends on (TREE_RCU || PREEMPT_RCU) && RCU_EXPERT
594 This option controls the leaf-level fanout of hierarchical
595 implementations of RCU, and allows trading off cache misses
596 against lock contention. Systems that synchronize their
597 scheduling-clock interrupts for energy-efficiency reasons will
598 want the default because the smaller leaf-level fanout keeps
599 lock contention levels acceptably low. Very large systems
600 (hundreds or thousands of CPUs) will instead want to set this
601 value to the maximum value possible in order to reduce the
602 number of cache misses incurred during RCU's grace-period
603 initialization. These systems tend to run CPU-bound, and thus
604 are not helped by synchronized interrupts, and thus tend to
605 skew them, which reduces lock contention enough that large
606 leaf-level fanouts work well.
608 Select a specific number if testing RCU itself.
610 Select the maximum permissible value for large systems.
612 Take the default if unsure.
614 config RCU_FAST_NO_HZ
615 bool "Accelerate last non-dyntick-idle CPU's grace periods"
616 depends on NO_HZ_COMMON && SMP && RCU_EXPERT
619 This option permits CPUs to enter dynticks-idle state even if
620 they have RCU callbacks queued, and prevents RCU from waking
621 these CPUs up more than roughly once every four jiffies (by
622 default, you can adjust this using the rcutree.rcu_idle_gp_delay
623 parameter), thus improving energy efficiency. On the other
624 hand, this option increases the duration of RCU grace periods,
625 for example, slowing down synchronize_rcu().
627 Say Y if energy efficiency is critically important, and you
628 don't care about increased grace-period durations.
630 Say N if you are unsure.
632 config TREE_RCU_TRACE
633 def_bool RCU_TRACE && ( TREE_RCU || PREEMPT_RCU )
636 This option provides tracing for the TREE_RCU and
637 PREEMPT_RCU implementations, permitting Makefile to
638 trivially select kernel/rcutree_trace.c.
641 bool "Enable RCU priority boosting"
642 depends on RT_MUTEXES && PREEMPT_RCU && RCU_EXPERT
645 This option boosts the priority of preempted RCU readers that
646 block the current preemptible RCU grace period for too long.
647 This option also prevents heavy loads from blocking RCU
648 callback invocation for all flavors of RCU.
650 Say Y here if you are working with real-time apps or heavy loads
651 Say N here if you are unsure.
653 config RCU_KTHREAD_PRIO
654 int "Real-time priority to use for RCU worker threads"
655 range 1 99 if RCU_BOOST
656 range 0 99 if !RCU_BOOST
657 default 1 if RCU_BOOST
658 default 0 if !RCU_BOOST
659 depends on RCU_EXPERT
661 This option specifies the SCHED_FIFO priority value that will be
662 assigned to the rcuc/n and rcub/n threads and is also the value
663 used for RCU_BOOST (if enabled). If you are working with a
664 real-time application that has one or more CPU-bound threads
665 running at a real-time priority level, you should set
666 RCU_KTHREAD_PRIO to a priority higher than the highest-priority
667 real-time CPU-bound application thread. The default RCU_KTHREAD_PRIO
668 value of 1 is appropriate in the common case, which is real-time
669 applications that do not have any CPU-bound threads.
671 Some real-time applications might not have a single real-time
672 thread that saturates a given CPU, but instead might have
673 multiple real-time threads that, taken together, fully utilize
674 that CPU. In this case, you should set RCU_KTHREAD_PRIO to
675 a priority higher than the lowest-priority thread that is
676 conspiring to prevent the CPU from running any non-real-time
677 tasks. For example, if one thread at priority 10 and another
678 thread at priority 5 are between themselves fully consuming
679 the CPU time on a given CPU, then RCU_KTHREAD_PRIO should be
680 set to priority 6 or higher.
682 Specify the real-time priority, or take the default if unsure.
684 config RCU_BOOST_DELAY
685 int "Milliseconds to delay boosting after RCU grace-period start"
690 This option specifies the time to wait after the beginning of
691 a given grace period before priority-boosting preempted RCU
692 readers blocking that grace period. Note that any RCU reader
693 blocking an expedited RCU grace period is boosted immediately.
695 Accept the default if unsure.
698 bool "Offload RCU callback processing from boot-selected CPUs"
699 depends on TREE_RCU || PREEMPT_RCU
700 depends on RCU_EXPERT || NO_HZ_FULL
703 Use this option to reduce OS jitter for aggressive HPC or
704 real-time workloads. It can also be used to offload RCU
705 callback invocation to energy-efficient CPUs in battery-powered
706 asymmetric multiprocessors.
708 This option offloads callback invocation from the set of
709 CPUs specified at boot time by the rcu_nocbs parameter.
710 For each such CPU, a kthread ("rcuox/N") will be created to
711 invoke callbacks, where the "N" is the CPU being offloaded,
712 and where the "x" is "b" for RCU-bh, "p" for RCU-preempt, and
713 "s" for RCU-sched. Nothing prevents this kthread from running
714 on the specified CPUs, but (1) the kthreads may be preempted
715 between each callback, and (2) affinity or cgroups can be used
716 to force the kthreads to run on whatever set of CPUs is desired.
718 Say Y here if you want to help to debug reduced OS jitter.
719 Say N here if you are unsure.
722 prompt "Build-forced no-CBs CPUs"
723 default RCU_NOCB_CPU_NONE
724 depends on RCU_NOCB_CPU
726 This option allows no-CBs CPUs (whose RCU callbacks are invoked
727 from kthreads rather than from softirq context) to be specified
728 at build time. Additional no-CBs CPUs may be specified by
729 the rcu_nocbs= boot parameter.
731 config RCU_NOCB_CPU_NONE
732 bool "No build_forced no-CBs CPUs"
734 This option does not force any of the CPUs to be no-CBs CPUs.
735 Only CPUs designated by the rcu_nocbs= boot parameter will be
736 no-CBs CPUs, whose RCU callbacks will be invoked by per-CPU
737 kthreads whose names begin with "rcuo". All other CPUs will
738 invoke their own RCU callbacks in softirq context.
740 Select this option if you want to choose no-CBs CPUs at
741 boot time, for example, to allow testing of different no-CBs
742 configurations without having to rebuild the kernel each time.
744 config RCU_NOCB_CPU_ZERO
745 bool "CPU 0 is a build_forced no-CBs CPU"
747 This option forces CPU 0 to be a no-CBs CPU, so that its RCU
748 callbacks are invoked by a per-CPU kthread whose name begins
749 with "rcuo". Additional CPUs may be designated as no-CBs
750 CPUs using the rcu_nocbs= boot parameter will be no-CBs CPUs.
751 All other CPUs will invoke their own RCU callbacks in softirq
754 Select this if CPU 0 needs to be a no-CBs CPU for real-time
755 or energy-efficiency reasons, but the real reason it exists
756 is to ensure that randconfig testing covers mixed systems.
758 config RCU_NOCB_CPU_ALL
759 bool "All CPUs are build_forced no-CBs CPUs"
761 This option forces all CPUs to be no-CBs CPUs. The rcu_nocbs=
762 boot parameter will be ignored. All CPUs' RCU callbacks will
763 be executed in the context of per-CPU rcuo kthreads created for
764 this purpose. Assuming that the kthreads whose names start with
765 "rcuo" are bound to "housekeeping" CPUs, this reduces OS jitter
766 on the remaining CPUs, but might decrease memory locality during
767 RCU-callback invocation, thus potentially degrading throughput.
769 Select this if all CPUs need to be no-CBs CPUs for real-time
770 or energy-efficiency reasons.
774 config RCU_EXPEDITE_BOOT
778 This option enables expedited grace periods at boot time,
779 as if rcu_expedite_gp() had been invoked early in boot.
780 The corresponding rcu_unexpedite_gp() is invoked from
781 rcu_end_inkernel_boot(), which is intended to be invoked
782 at the end of the kernel-only boot sequence, just before
785 Accept the default if unsure.
787 endmenu # "RCU Subsystem"
794 tristate "Kernel .config support"
797 This option enables the complete Linux kernel ".config" file
798 contents to be saved in the kernel. It provides documentation
799 of which kernel options are used in a running kernel or in an
800 on-disk kernel. This information can be extracted from the kernel
801 image file with the script scripts/extract-ikconfig and used as
802 input to rebuild the current kernel or to build another kernel.
803 It can also be extracted from a running kernel by reading
804 /proc/config.gz if enabled (below).
807 bool "Enable access to .config through /proc/config.gz"
808 depends on IKCONFIG && PROC_FS
810 This option enables access to the kernel configuration file
811 through /proc/config.gz.
814 int "Kernel log buffer size (16 => 64KB, 17 => 128KB)"
819 Select the minimal kernel log buffer size as a power of 2.
820 The final size is affected by LOG_CPU_MAX_BUF_SHIFT config
821 parameter, see below. Any higher size also might be forced
822 by "log_buf_len" boot parameter.
832 config LOG_CPU_MAX_BUF_SHIFT
833 int "CPU kernel log buffer size contribution (13 => 8 KB, 17 => 128KB)"
836 default 12 if !BASE_SMALL
837 default 0 if BASE_SMALL
840 This option allows to increase the default ring buffer size
841 according to the number of CPUs. The value defines the contribution
842 of each CPU as a power of 2. The used space is typically only few
843 lines however it might be much more when problems are reported,
846 The increased size means that a new buffer has to be allocated and
847 the original static one is unused. It makes sense only on systems
848 with more CPUs. Therefore this value is used only when the sum of
849 contributions is greater than the half of the default kernel ring
850 buffer as defined by LOG_BUF_SHIFT. The default values are set
851 so that more than 64 CPUs are needed to trigger the allocation.
853 Also this option is ignored when "log_buf_len" kernel parameter is
854 used as it forces an exact (power of two) size of the ring buffer.
856 The number of possible CPUs is used for this computation ignoring
857 hotplugging making the computation optimal for the worst case
858 scenario while allowing a simple algorithm to be used from bootup.
860 Examples shift values and their meaning:
861 17 => 128 KB for each CPU
862 16 => 64 KB for each CPU
863 15 => 32 KB for each CPU
864 14 => 16 KB for each CPU
865 13 => 8 KB for each CPU
866 12 => 4 KB for each CPU
868 config NMI_LOG_BUF_SHIFT
869 int "Temporary per-CPU NMI log buffer size (12 => 4KB, 13 => 8KB)"
872 depends on PRINTK_NMI
874 Select the size of a per-CPU buffer where NMI messages are temporary
875 stored. They are copied to the main log buffer in a safe context
876 to avoid a deadlock. The value defines the size as a power of 2.
878 NMI messages are rare and limited. The largest one is when
879 a backtrace is printed. It usually fits into 4KB. Select
880 8KB if you want to be on the safe side.
883 17 => 128 KB for each CPU
884 16 => 64 KB for each CPU
885 15 => 32 KB for each CPU
886 14 => 16 KB for each CPU
887 13 => 8 KB for each CPU
888 12 => 4 KB for each CPU
891 # Architectures with an unreliable sched_clock() should select this:
893 config HAVE_UNSTABLE_SCHED_CLOCK
896 config GENERIC_SCHED_CLOCK
900 # For architectures that want to enable the support for NUMA-affine scheduler
903 config ARCH_SUPPORTS_NUMA_BALANCING
907 # For architectures that prefer to flush all TLBs after a number of pages
908 # are unmapped instead of sending one IPI per page to flush. The architecture
909 # must provide guarantees on what happens if a clean TLB cache entry is
910 # written after the unmap. Details are in mm/rmap.c near the check for
911 # should_defer_flush. The architecture should also consider if the full flush
912 # and the refill costs are offset by the savings of sending fewer IPIs.
913 config ARCH_WANT_BATCHED_UNMAP_TLB_FLUSH
917 # For architectures that know their GCC __int128 support is sound
919 config ARCH_SUPPORTS_INT128
922 # For architectures that (ab)use NUMA to represent different memory regions
923 # all cpu-local but of different latencies, such as SuperH.
925 config ARCH_WANT_NUMA_VARIABLE_LOCALITY
928 config NUMA_BALANCING
929 bool "Memory placement aware NUMA scheduler"
930 depends on ARCH_SUPPORTS_NUMA_BALANCING
931 depends on !ARCH_WANT_NUMA_VARIABLE_LOCALITY
932 depends on SMP && NUMA && MIGRATION
934 This option adds support for automatic NUMA aware memory/task placement.
935 The mechanism is quite primitive and is based on migrating memory when
936 it has references to the node the task is running on.
938 This system will be inactive on UMA systems.
940 config NUMA_BALANCING_DEFAULT_ENABLED
941 bool "Automatically enable NUMA aware memory/task placement"
943 depends on NUMA_BALANCING
945 If set, automatic NUMA balancing will be enabled if running on a NUMA
949 bool "Control Group support"
952 This option adds support for grouping sets of processes together, for
953 use with process control subsystems such as Cpusets, CFS, memory
954 controls or device isolation.
956 - Documentation/scheduler/sched-design-CFS.txt (CFS)
957 - Documentation/cgroup-v1/ (features for grouping, isolation
958 and resource control)
968 bool "Memory controller"
972 Provides control over the memory footprint of tasks in a cgroup.
975 bool "Swap controller"
976 depends on MEMCG && SWAP
978 Provides control over the swap space consumed by tasks in a cgroup.
980 config MEMCG_SWAP_ENABLED
981 bool "Swap controller enabled by default"
982 depends on MEMCG_SWAP
985 Memory Resource Controller Swap Extension comes with its price in
986 a bigger memory consumption. General purpose distribution kernels
987 which want to enable the feature but keep it disabled by default
988 and let the user enable it by swapaccount=1 boot command line
989 parameter should have this option unselected.
990 For those who want to have the feature enabled by default should
991 select this option (if, for some reason, they need to disable it
992 then swapaccount=0 does the trick).
999 Generic block IO controller cgroup interface. This is the common
1000 cgroup interface which should be used by various IO controlling
1003 Currently, CFQ IO scheduler uses it to recognize task groups and
1004 control disk bandwidth allocation (proportional time slice allocation)
1005 to such task groups. It is also used by bio throttling logic in
1006 block layer to implement upper limit in IO rates on a device.
1008 This option only enables generic Block IO controller infrastructure.
1009 One needs to also enable actual IO controlling logic/policy. For
1010 enabling proportional weight division of disk bandwidth in CFQ, set
1011 CONFIG_CFQ_GROUP_IOSCHED=y; for enabling throttling policy, set
1012 CONFIG_BLK_DEV_THROTTLING=y.
1014 See Documentation/cgroup-v1/blkio-controller.txt for more information.
1016 config DEBUG_BLK_CGROUP
1017 bool "IO controller debugging"
1018 depends on BLK_CGROUP
1021 Enable some debugging help. Currently it exports additional stat
1022 files in a cgroup which can be useful for debugging.
1024 config CGROUP_WRITEBACK
1026 depends on MEMCG && BLK_CGROUP
1029 menuconfig CGROUP_SCHED
1030 bool "CPU controller"
1033 This feature lets CPU scheduler recognize task groups and control CPU
1034 bandwidth allocation to such task groups. It uses cgroups to group
1038 config FAIR_GROUP_SCHED
1039 bool "Group scheduling for SCHED_OTHER"
1040 depends on CGROUP_SCHED
1041 default CGROUP_SCHED
1043 config CFS_BANDWIDTH
1044 bool "CPU bandwidth provisioning for FAIR_GROUP_SCHED"
1045 depends on FAIR_GROUP_SCHED
1048 This option allows users to define CPU bandwidth rates (limits) for
1049 tasks running within the fair group scheduler. Groups with no limit
1050 set are considered to be unconstrained and will run with no
1052 See tip/Documentation/scheduler/sched-bwc.txt for more information.
1054 config RT_GROUP_SCHED
1055 bool "Group scheduling for SCHED_RR/FIFO"
1056 depends on CGROUP_SCHED
1059 This feature lets you explicitly allocate real CPU bandwidth
1060 to task groups. If enabled, it will also make it impossible to
1061 schedule realtime tasks for non-root users until you allocate
1062 realtime bandwidth for them.
1063 See Documentation/scheduler/sched-rt-group.txt for more information.
1068 bool "PIDs controller"
1070 Provides enforcement of process number limits in the scope of a
1071 cgroup. Any attempt to fork more processes than is allowed in the
1072 cgroup will fail. PIDs are fundamentally a global resource because it
1073 is fairly trivial to reach PID exhaustion before you reach even a
1074 conservative kmemcg limit. As a result, it is possible to grind a
1075 system to halt without being limited by other cgroup policies. The
1076 PIDs controller is designed to stop this from happening.
1078 It should be noted that organisational operations (such as attaching
1079 to a cgroup hierarchy will *not* be blocked by the PIDs controller),
1080 since the PIDs limit only affects a process's ability to fork, not to
1083 config CGROUP_FREEZER
1084 bool "Freezer controller"
1086 Provides a way to freeze and unfreeze all tasks in a
1089 This option affects the ORIGINAL cgroup interface. The cgroup2 memory
1090 controller includes important in-kernel memory consumers per default.
1092 If you're using cgroup2, say N.
1094 config CGROUP_HUGETLB
1095 bool "HugeTLB controller"
1096 depends on HUGETLB_PAGE
1100 Provides a cgroup controller for HugeTLB pages.
1101 When you enable this, you can put a per cgroup limit on HugeTLB usage.
1102 The limit is enforced during page fault. Since HugeTLB doesn't
1103 support page reclaim, enforcing the limit at page fault time implies
1104 that, the application will get SIGBUS signal if it tries to access
1105 HugeTLB pages beyond its limit. This requires the application to know
1106 beforehand how much HugeTLB pages it would require for its use. The
1107 control group is tracked in the third page lru pointer. This means
1108 that we cannot use the controller with huge page less than 3 pages.
1111 bool "Cpuset controller"
1113 This option will let you create and manage CPUSETs which
1114 allow dynamically partitioning a system into sets of CPUs and
1115 Memory Nodes and assigning tasks to run only within those sets.
1116 This is primarily useful on large SMP or NUMA systems.
1120 config PROC_PID_CPUSET
1121 bool "Include legacy /proc/<pid>/cpuset file"
1125 config CGROUP_DEVICE
1126 bool "Device controller"
1128 Provides a cgroup controller implementing whitelists for
1129 devices which a process in the cgroup can mknod or open.
1131 config CGROUP_CPUACCT
1132 bool "Simple CPU accounting controller"
1134 Provides a simple controller for monitoring the
1135 total CPU consumed by the tasks in a cgroup.
1138 bool "Perf controller"
1139 depends on PERF_EVENTS
1141 This option extends the perf per-cpu mode to restrict monitoring
1142 to threads which belong to the cgroup specified and run on the
1148 bool "Example controller"
1151 This option enables a simple controller that exports
1152 debugging information about the cgroups framework.
1158 config CHECKPOINT_RESTORE
1159 bool "Checkpoint/restore support" if EXPERT
1160 select PROC_CHILDREN
1163 Enables additional kernel features in a sake of checkpoint/restore.
1164 In particular it adds auxiliary prctl codes to setup process text,
1165 data and heap segment sizes, and a few additional /proc filesystem
1168 If unsure, say N here.
1170 menuconfig NAMESPACES
1171 bool "Namespaces support" if EXPERT
1172 depends on MULTIUSER
1175 Provides the way to make tasks work with different objects using
1176 the same id. For example same IPC id may refer to different objects
1177 or same user id or pid may refer to different tasks when used in
1178 different namespaces.
1183 bool "UTS namespace"
1186 In this namespace tasks see different info provided with the
1190 bool "IPC namespace"
1191 depends on (SYSVIPC || POSIX_MQUEUE)
1194 In this namespace tasks work with IPC ids which correspond to
1195 different IPC objects in different namespaces.
1198 bool "User namespace"
1201 This allows containers, i.e. vservers, to use user namespaces
1202 to provide different user info for different servers.
1204 When user namespaces are enabled in the kernel it is
1205 recommended that the MEMCG option also be enabled and that
1206 user-space use the memory control groups to limit the amount
1207 of memory a memory unprivileged users can use.
1212 bool "PID Namespaces"
1215 Support process id namespaces. This allows having multiple
1216 processes with the same pid as long as they are in different
1217 pid namespaces. This is a building block of containers.
1220 bool "Network namespace"
1224 Allow user space to create what appear to be multiple instances
1225 of the network stack.
1229 config SCHED_AUTOGROUP
1230 bool "Automatic process group scheduling"
1233 select FAIR_GROUP_SCHED
1235 This option optimizes the scheduler for common desktop workloads by
1236 automatically creating and populating task groups. This separation
1237 of workloads isolates aggressive CPU burners (like build jobs) from
1238 desktop applications. Task group autogeneration is currently based
1241 config SYSFS_DEPRECATED
1242 bool "Enable deprecated sysfs features to support old userspace tools"
1246 This option adds code that switches the layout of the "block" class
1247 devices, to not show up in /sys/class/block/, but only in
1250 This switch is only active when the sysfs.deprecated=1 boot option is
1251 passed or the SYSFS_DEPRECATED_V2 option is set.
1253 This option allows new kernels to run on old distributions and tools,
1254 which might get confused by /sys/class/block/. Since 2007/2008 all
1255 major distributions and tools handle this just fine.
1257 Recent distributions and userspace tools after 2009/2010 depend on
1258 the existence of /sys/class/block/, and will not work with this
1261 Only if you are using a new kernel on an old distribution, you might
1264 config SYSFS_DEPRECATED_V2
1265 bool "Enable deprecated sysfs features by default"
1268 depends on SYSFS_DEPRECATED
1270 Enable deprecated sysfs by default.
1272 See the CONFIG_SYSFS_DEPRECATED option for more details about this
1275 Only if you are using a new kernel on an old distribution, you might
1276 need to say Y here. Even then, odds are you would not need it
1277 enabled, you can always pass the boot option if absolutely necessary.
1280 bool "Kernel->user space relay support (formerly relayfs)"
1282 This option enables support for relay interface support in
1283 certain file systems (such as debugfs).
1284 It is designed to provide an efficient mechanism for tools and
1285 facilities to relay large amounts of data from kernel space to
1290 config BLK_DEV_INITRD
1291 bool "Initial RAM filesystem and RAM disk (initramfs/initrd) support"
1292 depends on BROKEN || !FRV
1294 The initial RAM filesystem is a ramfs which is loaded by the
1295 boot loader (loadlin or lilo) and that is mounted as root
1296 before the normal boot procedure. It is typically used to
1297 load modules needed to mount the "real" root file system,
1298 etc. See <file:Documentation/initrd.txt> for details.
1300 If RAM disk support (BLK_DEV_RAM) is also included, this
1301 also enables initial RAM disk (initrd) support and adds
1302 15 Kbytes (more on some other architectures) to the kernel size.
1308 source "usr/Kconfig"
1313 prompt "Compiler optimization level"
1314 default CONFIG_CC_OPTIMIZE_FOR_PERFORMANCE
1316 config CC_OPTIMIZE_FOR_PERFORMANCE
1317 bool "Optimize for performance"
1319 This is the default optimization level for the kernel, building
1320 with the "-O2" compiler flag for best performance and most
1321 helpful compile-time warnings.
1323 config CC_OPTIMIZE_FOR_SIZE
1324 bool "Optimize for size"
1326 Enabling this option will pass "-Os" instead of "-O2" to
1327 your compiler resulting in a smaller kernel.
1342 config SYSCTL_EXCEPTION_TRACE
1345 Enable support for /proc/sys/debug/exception-trace.
1347 config SYSCTL_ARCH_UNALIGN_NO_WARN
1350 Enable support for /proc/sys/kernel/ignore-unaligned-usertrap
1351 Allows arch to define/use @no_unaligned_warning to possibly warn
1352 about unaligned access emulation going on under the hood.
1354 config SYSCTL_ARCH_UNALIGN_ALLOW
1357 Enable support for /proc/sys/kernel/unaligned-trap
1358 Allows arches to define/use @unaligned_enabled to runtime toggle
1359 the unaligned access emulation.
1360 see arch/parisc/kernel/unaligned.c for reference
1362 config HAVE_PCSPKR_PLATFORM
1365 # interpreter that classic socket filters depend on
1370 bool "Configure standard kernel features (expert users)"
1371 # Unhide debug options, to make the on-by-default options visible
1374 This option allows certain base kernel options and settings
1375 to be disabled or tweaked. This is for specialized
1376 environments which can tolerate a "non-standard" kernel.
1377 Only use this if you really know what you are doing.
1380 bool "Enable 16-bit UID system calls" if EXPERT
1381 depends on HAVE_UID16 && MULTIUSER
1384 This enables the legacy 16-bit UID syscall wrappers.
1387 bool "Multiple users, groups and capabilities support" if EXPERT
1390 This option enables support for non-root users, groups and
1393 If you say N here, all processes will run with UID 0, GID 0, and all
1394 possible capabilities. Saying N here also compiles out support for
1395 system calls related to UIDs, GIDs, and capabilities, such as setuid,
1398 If unsure, say Y here.
1400 config SGETMASK_SYSCALL
1401 bool "sgetmask/ssetmask syscalls support" if EXPERT
1402 def_bool PARISC || MN10300 || BLACKFIN || M68K || PPC || MIPS || X86 || SPARC || CRIS || MICROBLAZE || SUPERH
1404 sys_sgetmask and sys_ssetmask are obsolete system calls
1405 no longer supported in libc but still enabled by default in some
1408 If unsure, leave the default option here.
1410 config SYSFS_SYSCALL
1411 bool "Sysfs syscall support" if EXPERT
1414 sys_sysfs is an obsolete system call no longer supported in libc.
1415 Note that disabling this option is more secure but might break
1416 compatibility with some systems.
1418 If unsure say Y here.
1420 config SYSCTL_SYSCALL
1421 bool "Sysctl syscall support" if EXPERT
1422 depends on PROC_SYSCTL
1426 sys_sysctl uses binary paths that have been found challenging
1427 to properly maintain and use. The interface in /proc/sys
1428 using paths with ascii names is now the primary path to this
1431 Almost nothing using the binary sysctl interface so if you are
1432 trying to save some space it is probably safe to disable this,
1433 making your kernel marginally smaller.
1435 If unsure say N here.
1438 bool "Load all symbols for debugging/ksymoops" if EXPERT
1441 Say Y here to let the kernel print out symbolic crash information and
1442 symbolic stack backtraces. This increases the size of the kernel
1443 somewhat, as all symbols have to be loaded into the kernel image.
1446 bool "Include all symbols in kallsyms"
1447 depends on DEBUG_KERNEL && KALLSYMS
1449 Normally kallsyms only contains the symbols of functions for nicer
1450 OOPS messages and backtraces (i.e., symbols from the text and inittext
1451 sections). This is sufficient for most cases. And only in very rare
1452 cases (e.g., when a debugger is used) all symbols are required (e.g.,
1453 names of variables from the data sections, etc).
1455 This option makes sure that all symbols are loaded into the kernel
1456 image (i.e., symbols from all sections) in cost of increased kernel
1457 size (depending on the kernel configuration, it may be 300KiB or
1458 something like this).
1460 Say N unless you really need all symbols.
1462 config KALLSYMS_ABSOLUTE_PERCPU
1465 default X86_64 && SMP
1467 config KALLSYMS_BASE_RELATIVE
1470 default !IA64 && !(TILE && 64BIT)
1472 Instead of emitting them as absolute values in the native word size,
1473 emit the symbol references in the kallsyms table as 32-bit entries,
1474 each containing a relative value in the range [base, base + U32_MAX]
1475 or, when KALLSYMS_ABSOLUTE_PERCPU is in effect, each containing either
1476 an absolute value in the range [0, S32_MAX] or a relative value in the
1477 range [base, base + S32_MAX], where base is the lowest relative symbol
1478 address encountered in the image.
1480 On 64-bit builds, this reduces the size of the address table by 50%,
1481 but more importantly, it results in entries whose values are build
1482 time constants, and no relocation pass is required at runtime to fix
1483 up the entries based on the runtime load address of the kernel.
1487 bool "Enable support for printk" if EXPERT
1490 This option enables normal printk support. Removing it
1491 eliminates most of the message strings from the kernel image
1492 and makes the kernel more or less silent. As this makes it
1493 very difficult to diagnose system problems, saying N here is
1494 strongly discouraged.
1502 bool "BUG() support" if EXPERT
1505 Disabling this option eliminates support for BUG and WARN, reducing
1506 the size of your kernel image and potentially quietly ignoring
1507 numerous fatal conditions. You should only consider disabling this
1508 option for embedded systems with no facilities for reporting errors.
1514 bool "Enable ELF core dumps" if EXPERT
1516 Enable support for generating core dumps. Disabling saves about 4k.
1519 config PCSPKR_PLATFORM
1520 bool "Enable PC-Speaker support" if EXPERT
1521 depends on HAVE_PCSPKR_PLATFORM
1525 This option allows to disable the internal PC-Speaker
1526 support, saving some memory.
1530 bool "Enable full-sized data structures for core" if EXPERT
1532 Disabling this option reduces the size of miscellaneous core
1533 kernel data structures. This saves memory on small machines,
1534 but may reduce performance.
1537 bool "Enable futex support" if EXPERT
1541 Disabling this option will cause the kernel to be built without
1542 support for "fast userspace mutexes". The resulting kernel may not
1543 run glibc-based applications correctly.
1545 config HAVE_FUTEX_CMPXCHG
1549 Architectures should select this if futex_atomic_cmpxchg_inatomic()
1550 is implemented and always working. This removes a couple of runtime
1554 bool "Enable eventpoll support" if EXPERT
1558 Disabling this option will cause the kernel to be built without
1559 support for epoll family of system calls.
1562 bool "Enable signalfd() system call" if EXPERT
1566 Enable the signalfd() system call that allows to receive signals
1567 on a file descriptor.
1572 bool "Enable timerfd() system call" if EXPERT
1576 Enable the timerfd() system call that allows to receive timer
1577 events on a file descriptor.
1582 bool "Enable eventfd() system call" if EXPERT
1586 Enable the eventfd() system call that allows to receive both
1587 kernel notification (ie. KAIO) or userspace notifications.
1591 # syscall, maps, verifier
1593 bool "Enable bpf() system call"
1598 Enable the bpf() system call that allows to manipulate eBPF
1599 programs and maps via file descriptors.
1602 bool "Use full shmem filesystem" if EXPERT
1606 The shmem is an internal filesystem used to manage shared memory.
1607 It is backed by swap and manages resource limits. It is also exported
1608 to userspace as tmpfs if TMPFS is enabled. Disabling this
1609 option replaces shmem and tmpfs with the much simpler ramfs code,
1610 which may be appropriate on small systems without swap.
1613 bool "Enable AIO support" if EXPERT
1616 This option enables POSIX asynchronous I/O which may by used
1617 by some high performance threaded applications. Disabling
1618 this option saves about 7k.
1620 config ADVISE_SYSCALLS
1621 bool "Enable madvise/fadvise syscalls" if EXPERT
1624 This option enables the madvise and fadvise syscalls, used by
1625 applications to advise the kernel about their future memory or file
1626 usage, improving performance. If building an embedded system where no
1627 applications use these syscalls, you can disable this option to save
1631 bool "Enable userfaultfd() system call"
1635 Enable the userfaultfd() system call that allows to intercept and
1636 handle page faults in userland.
1640 bool "Enable PCI quirk workarounds" if EXPERT
1643 This enables workarounds for various PCI chipset
1644 bugs/quirks. Disable this only if your target machine is
1645 unaffected by PCI quirks.
1648 bool "Enable membarrier() system call" if EXPERT
1651 Enable the membarrier() system call that allows issuing memory
1652 barriers across all running threads, which can be used to distribute
1653 the cost of user-space memory barriers asymmetrically by transforming
1654 pairs of memory barriers into pairs consisting of membarrier() and a
1660 bool "Enable rseq() system call" if EXPERT
1662 depends on HAVE_RSEQ
1664 Enable the restartable sequences system call. It provides a
1665 user-space cache for the current CPU number value, which
1666 speeds up getting the current CPU number from user-space,
1667 as well as an ABI to speed up user-space operations on
1673 bool "Embedded system"
1674 option allnoconfig_y
1677 This option should be enabled if compiling the kernel for
1678 an embedded system so certain expert options are available
1681 config HAVE_PERF_EVENTS
1684 See tools/perf/design.txt for details.
1686 config PERF_USE_VMALLOC
1689 See tools/perf/design.txt for details
1691 menu "Kernel Performance Events And Counters"
1694 bool "Kernel performance events and counters"
1695 default y if PROFILING
1696 depends on HAVE_PERF_EVENTS
1701 Enable kernel support for various performance events provided
1702 by software and hardware.
1704 Software events are supported either built-in or via the
1705 use of generic tracepoints.
1707 Most modern CPUs support performance events via performance
1708 counter registers. These registers count the number of certain
1709 types of hw events: such as instructions executed, cachemisses
1710 suffered, or branches mis-predicted - without slowing down the
1711 kernel or applications. These registers can also trigger interrupts
1712 when a threshold number of events have passed - and can thus be
1713 used to profile the code that runs on that CPU.
1715 The Linux Performance Event subsystem provides an abstraction of
1716 these software and hardware event capabilities, available via a
1717 system call and used by the "perf" utility in tools/perf/. It
1718 provides per task and per CPU counters, and it provides event
1719 capabilities on top of those.
1723 config DEBUG_PERF_USE_VMALLOC
1725 bool "Debug: use vmalloc to back perf mmap() buffers"
1726 depends on PERF_EVENTS && DEBUG_KERNEL && !PPC
1727 select PERF_USE_VMALLOC
1729 Use vmalloc memory to back perf mmap() buffers.
1731 Mostly useful for debugging the vmalloc code on platforms
1732 that don't require it.
1738 config VM_EVENT_COUNTERS
1740 bool "Enable VM event counters for /proc/vmstat" if EXPERT
1742 VM event counters are needed for event counts to be shown.
1743 This option allows the disabling of the VM event counters
1744 on EXPERT systems. /proc/vmstat will only show page counts
1745 if VM event counters are disabled.
1749 bool "Enable SLUB debugging support" if EXPERT
1750 depends on SLUB && SYSFS
1752 SLUB has extensive debug support features. Disabling these can
1753 result in significant savings in code size. This also disables
1754 SLUB sysfs support. /sys/slab will not exist and there will be
1755 no support for cache validation etc.
1758 bool "Disable heap randomization"
1761 Randomizing heap placement makes heap exploits harder, but it
1762 also breaks ancient binaries (including anything libc5 based).
1763 This option changes the bootup default to heap randomization
1764 disabled, and can be overridden at runtime by setting
1765 /proc/sys/kernel/randomize_va_space to 2.
1767 On non-ancient distros (post-2000 ones) N is usually a safe choice.
1770 prompt "Choose SLAB allocator"
1773 This option allows to select a slab allocator.
1777 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1779 The regular slab allocator that is established and known to work
1780 well in all environments. It organizes cache hot objects in
1781 per cpu and per node queues.
1784 bool "SLUB (Unqueued Allocator)"
1785 select HAVE_HARDENED_USERCOPY_ALLOCATOR
1787 SLUB is a slab allocator that minimizes cache line usage
1788 instead of managing queues of cached objects (SLAB approach).
1789 Per cpu caching is realized using slabs of objects instead
1790 of queues of objects. SLUB can use memory efficiently
1791 and has enhanced diagnostics. SLUB is the default choice for
1796 bool "SLOB (Simple Allocator)"
1798 SLOB replaces the stock allocator with a drastically simpler
1799 allocator. SLOB is generally more space efficient but
1800 does not perform as well on large systems.
1804 config SLAB_FREELIST_RANDOM
1806 depends on SLAB || SLUB
1807 bool "SLAB freelist randomization"
1809 Randomizes the freelist order used on creating new pages. This
1810 security feature reduces the predictability of the kernel slab
1811 allocator against heap overflows.
1813 config SLUB_CPU_PARTIAL
1815 depends on SLUB && SMP
1816 bool "SLUB per cpu partial cache"
1818 Per cpu partial caches accellerate objects allocation and freeing
1819 that is local to a processor at the price of more indeterminism
1820 in the latency of the free. On overflow these caches will be cleared
1821 which requires the taking of locks that may cause latency spikes.
1822 Typically one would choose no for a realtime system.
1824 config MMAP_ALLOW_UNINITIALIZED
1825 bool "Allow mmapped anonymous memory to be uninitialized"
1826 depends on EXPERT && !MMU
1829 Normally, and according to the Linux spec, anonymous memory obtained
1830 from mmap() has it's contents cleared before it is passed to
1831 userspace. Enabling this config option allows you to request that
1832 mmap() skip that if it is given an MAP_UNINITIALIZED flag, thus
1833 providing a huge performance boost. If this option is not enabled,
1834 then the flag will be ignored.
1836 This is taken advantage of by uClibc's malloc(), and also by
1837 ELF-FDPIC binfmt's brk and stack allocator.
1839 Because of the obvious security issues, this option should only be
1840 enabled on embedded devices where you control what is run in
1841 userspace. Since that isn't generally a problem on no-MMU systems,
1842 it is normally safe to say Y here.
1844 See Documentation/nommu-mmap.txt for more information.
1846 config SYSTEM_DATA_VERIFICATION
1848 select SYSTEM_TRUSTED_KEYRING
1852 select ASYMMETRIC_KEY_TYPE
1853 select ASYMMETRIC_PUBLIC_KEY_SUBTYPE
1856 select X509_CERTIFICATE_PARSER
1857 select PKCS7_MESSAGE_PARSER
1859 Provide PKCS#7 message verification using the contents of the system
1860 trusted keyring to provide public keys. This then can be used for
1861 module verification, kexec image verification and firmware blob
1865 bool "Profiling support"
1867 Say Y here to enable the extended profiling support mechanisms used
1868 by profilers such as OProfile.
1871 # Place an empty function call at each tracepoint site. Can be
1872 # dynamically changed for a probe function.
1877 source "arch/Kconfig"
1879 endmenu # General setup
1881 config HAVE_GENERIC_DMA_COHERENT
1888 depends on SLAB || SLUB_DEBUG
1896 default 0 if BASE_FULL
1897 default 1 if !BASE_FULL
1900 bool "Enable loadable module support"
1903 Kernel modules are small pieces of compiled code which can
1904 be inserted in the running kernel, rather than being
1905 permanently built into the kernel. You use the "modprobe"
1906 tool to add (and sometimes remove) them. If you say Y here,
1907 many parts of the kernel can be built as modules (by
1908 answering M instead of Y where indicated): this is most
1909 useful for infrequently used options which are not required
1910 for booting. For more information, see the man pages for
1911 modprobe, lsmod, modinfo, insmod and rmmod.
1913 If you say Y here, you will need to run "make
1914 modules_install" to put the modules under /lib/modules/
1915 where modprobe can find them (you may need to be root to do
1922 config MODULE_FORCE_LOAD
1923 bool "Forced module loading"
1926 Allow loading of modules without version information (ie. modprobe
1927 --force). Forced module loading sets the 'F' (forced) taint flag and
1928 is usually a really bad idea.
1930 config MODULE_UNLOAD
1931 bool "Module unloading"
1933 Without this option you will not be able to unload any
1934 modules (note that some modules may not be unloadable
1935 anyway), which makes your kernel smaller, faster
1936 and simpler. If unsure, say Y.
1938 config MODULE_FORCE_UNLOAD
1939 bool "Forced module unloading"
1940 depends on MODULE_UNLOAD
1942 This option allows you to force a module to unload, even if the
1943 kernel believes it is unsafe: the kernel will remove the module
1944 without waiting for anyone to stop using it (using the -f option to
1945 rmmod). This is mainly for kernel developers and desperate users.
1949 bool "Module versioning support"
1951 Usually, you have to use modules compiled with your kernel.
1952 Saying Y here makes it sometimes possible to use modules
1953 compiled for different kernels, by adding enough information
1954 to the modules to (hopefully) spot any changes which would
1955 make them incompatible with the kernel you are running. If
1958 config MODULE_SRCVERSION_ALL
1959 bool "Source checksum for all modules"
1961 Modules which contain a MODULE_VERSION get an extra "srcversion"
1962 field inserted into their modinfo section, which contains a
1963 sum of the source files which made it. This helps maintainers
1964 see exactly which source was used to build a module (since
1965 others sometimes change the module source without updating
1966 the version). With this option, such a "srcversion" field
1967 will be created for all modules. If unsure, say N.
1970 bool "Module signature verification"
1972 select SYSTEM_DATA_VERIFICATION
1974 Check modules for valid signatures upon load: the signature
1975 is simply appended to the module. For more information see
1976 Documentation/module-signing.txt.
1978 Note that this option adds the OpenSSL development packages as a
1979 kernel build dependency so that the signing tool can use its crypto
1982 !!!WARNING!!! If you enable this option, you MUST make sure that the
1983 module DOES NOT get stripped after being signed. This includes the
1984 debuginfo strip done by some packagers (such as rpmbuild) and
1985 inclusion into an initramfs that wants the module size reduced.
1987 config MODULE_SIG_FORCE
1988 bool "Require modules to be validly signed"
1989 depends on MODULE_SIG
1991 Reject unsigned modules or signed modules for which we don't have a
1992 key. Without this, such modules will simply taint the kernel.
1994 config MODULE_SIG_ALL
1995 bool "Automatically sign all modules"
1997 depends on MODULE_SIG
1999 Sign all modules during make modules_install. Without this option,
2000 modules must be signed manually, using the scripts/sign-file tool.
2002 comment "Do not forget to sign required modules with scripts/sign-file"
2003 depends on MODULE_SIG_FORCE && !MODULE_SIG_ALL
2006 prompt "Which hash algorithm should modules be signed with?"
2007 depends on MODULE_SIG
2009 This determines which sort of hashing algorithm will be used during
2010 signature generation. This algorithm _must_ be built into the kernel
2011 directly so that signature verification can take place. It is not
2012 possible to load a signed module containing the algorithm to check
2013 the signature on that module.
2015 config MODULE_SIG_SHA1
2016 bool "Sign modules with SHA-1"
2019 config MODULE_SIG_SHA224
2020 bool "Sign modules with SHA-224"
2021 select CRYPTO_SHA256
2023 config MODULE_SIG_SHA256
2024 bool "Sign modules with SHA-256"
2025 select CRYPTO_SHA256
2027 config MODULE_SIG_SHA384
2028 bool "Sign modules with SHA-384"
2029 select CRYPTO_SHA512
2031 config MODULE_SIG_SHA512
2032 bool "Sign modules with SHA-512"
2033 select CRYPTO_SHA512
2037 config MODULE_SIG_HASH
2039 depends on MODULE_SIG
2040 default "sha1" if MODULE_SIG_SHA1
2041 default "sha224" if MODULE_SIG_SHA224
2042 default "sha256" if MODULE_SIG_SHA256
2043 default "sha384" if MODULE_SIG_SHA384
2044 default "sha512" if MODULE_SIG_SHA512
2046 config MODULE_COMPRESS
2047 bool "Compress modules on installation"
2051 Compresses kernel modules when 'make modules_install' is run; gzip or
2052 xz depending on "Compression algorithm" below.
2054 module-init-tools MAY support gzip, and kmod MAY support gzip and xz.
2056 Out-of-tree kernel modules installed using Kbuild will also be
2057 compressed upon installation.
2059 Note: for modules inside an initrd or initramfs, it's more efficient
2060 to compress the whole initrd or initramfs instead.
2062 Note: This is fully compatible with signed modules.
2067 prompt "Compression algorithm"
2068 depends on MODULE_COMPRESS
2069 default MODULE_COMPRESS_GZIP
2071 This determines which sort of compression will be used during
2072 'make modules_install'.
2074 GZIP (default) and XZ are supported.
2076 config MODULE_COMPRESS_GZIP
2079 config MODULE_COMPRESS_XZ
2084 config TRIM_UNUSED_KSYMS
2085 bool "Trim unused exported kernel symbols"
2086 depends on MODULES && !UNUSED_SYMBOLS
2088 The kernel and some modules make many symbols available for
2089 other modules to use via EXPORT_SYMBOL() and variants. Depending
2090 on the set of modules being selected in your kernel configuration,
2091 many of those exported symbols might never be used.
2093 This option allows for unused exported symbols to be dropped from
2094 the build. In turn, this provides the compiler more opportunities
2095 (especially when using LTO) for optimizing the code and reducing
2096 binary size. This might have some security advantages as well.
2098 If unsure, or if you need to build out-of-tree modules, say N.
2102 config MODULES_TREE_LOOKUP
2104 depends on PERF_EVENTS || TRACING
2106 config INIT_ALL_POSSIBLE
2109 Back when each arch used to define their own cpu_online_mask and
2110 cpu_possible_mask, some of them chose to initialize cpu_possible_mask
2111 with all 1s, and others with all 0s. When they were centralised,
2112 it was better to provide this option than to break all the archs
2113 and have several arch maintainers pursuing me down dark alleys.
2115 source "block/Kconfig"
2117 config PREEMPT_NOTIFIERS
2124 # Can be selected by architectures with broken toolchains
2125 # that get confused by correct const<->read_only section
2127 config BROKEN_RODATA
2133 Build a simple ASN.1 grammar compiler that produces a bytecode output
2134 that can be interpreted by the ASN.1 stream decoder and used to
2135 inform it as to what tags are to be expected in a stream and what
2136 functions to call on what tags.
2138 source "kernel/Kconfig.locks"
projects / deliverable / linux.git / blob